CANINE	
  LOWER	
  URINARY	
  TRACT	
  UROTHELIAL	
  CARCINOMA:	
  RELEVANCE	
  AS	
  AN	
  ANIMAL	
  
MODEL	
  
By	
  
Dodd	
  Sledge	
  
	
  
	
  
	
  
	
  
	
  
A	
  DISSERTATION	
  
Submitted	
  to	
  
Michigan	
  State	
  University	
  
in	
  partial	
  fulfillment	
  of	
  the	
  requirements	
  
for	
  the	
  degree	
  of	
  
	
  
Pathobiology	
  –	
  Doctor	
  of	
  Philosophy	
  	
  
2014	
  
	
  
	
  

	
  

	
  

ABSTRACT	
  
CANINE	
  LOWER	
  URINARY	
  TRACT	
  UROTHELIAL	
  CARCINOMA:	
  RELEVANCE	
  AS	
  AN	
  ANIMAL	
  
MODEL	
  
By	
  
Dodd	
  Sledge	
  
Lower	
  urinary	
  tract	
  urothelial	
  carcinomas	
  are	
  commonly	
  diagnosed	
  neoplasms	
  in	
  humans	
  
and	
  dogs.	
  	
  Similarities	
  between	
  human	
  and	
  canine	
  urothelial	
  carcinomas	
  have	
  been	
  well	
  
described	
  in	
  terms	
  of	
  histomorphology	
  and	
  clinical	
  progression.	
  	
  Further	
  due	
  to	
  strong	
  
breed	
  predispositions	
  for	
  tumor	
  development,	
  there	
  are	
  likely	
  heritable	
  factors	
  that	
  
regulate	
  carcinogenesis	
  in	
  the	
  lower	
  urinary	
  tract	
  of	
  subsets	
  of	
  dogs.	
  This	
  has	
  led	
  to	
  the	
  
suggestion	
  that	
  canine	
  urothelial	
  carcinomas	
  could	
  be	
  used	
  as	
  a	
  naturally	
  occurring	
  animal	
  
model.	
  	
  However,	
  it	
  is	
  unclear	
  what	
  features	
  in	
  canine	
  urothelial	
  carcinomas	
  are	
  associated	
  
with	
  prognostic	
  significance	
  or	
  treatment	
  response.	
  	
  Further,	
  while	
  some	
  similarities	
  
between	
  human	
  and	
  canine	
  urothelial	
  carcinomas	
  are	
  known,	
  what	
  similarities	
  or	
  
differences	
  may	
  exist	
  in	
  terms	
  of	
  molecular	
  features	
  that	
  drive	
  carcinogenesis	
  are	
  largely	
  
unknown.	
  This	
  dissertation	
  first	
  examines	
  correlations	
  of	
  common	
  urothelial	
  carcinoma	
  
markers	
  to	
  histologic	
  classification,	
  grading,	
  and	
  degree	
  of	
  bladder	
  wall	
  invasion	
  in	
  dogs	
  
relative	
  to	
  the	
  histologic	
  classification	
  scheme	
  accepted	
  for	
  humans	
  in	
  order	
  strengthen	
  the	
  
stance	
  that	
  there	
  are	
  biologic	
  differences	
  between	
  proliferative	
  urothelial	
  lesions	
  and	
  
histologic	
  grades.	
  	
  Then,	
  specific	
  carcinogenesis	
  pathways	
  that	
  have	
  been	
  suggested	
  to	
  play	
  
roles	
  in	
  epithelial-­‐to-­‐mesenchymal	
  transition	
  and	
  that	
  have	
  prognostic	
  significance	
  in	
  
human	
  urinary	
  bladder	
  urothelial	
  carcinomas	
  were	
  evaluated.	
  These	
  included	
  pathways	
  
that	
  govern	
  prostaglandin	
  E2	
  regulation,	
  cadherin	
  swithching,	
  and	
  Wnt	
  signaling.	
  	
  Finally,	
  
	
  

the	
  role	
  of	
  defective	
  DNA	
  mismatch	
  repair	
  (MMR)	
  was	
  examined.	
  	
  In	
  urothelial	
  carcinomas,	
  
evidence	
  of	
  MMR	
  repair	
  dysfunction	
  was	
  found	
  and	
  was	
  correlated	
  with	
  genetic	
  
background	
  of	
  the	
  dogs	
  from	
  which	
  tumors	
  originated.	
  	
  Additionally,	
  canine	
  lower	
  urinary	
  
tract	
  urothelial	
  carcinoma	
  cell	
  lines	
  were	
  established	
  which	
  differential	
  MMR	
  proficiency	
  
and	
  which	
  had	
  also	
  had	
  differential	
  response	
  to	
  treatment	
  similar	
  to	
  that	
  described	
  in	
  
humans.	
  	
  These	
  studies	
  combine	
  to	
  show	
  that	
  while	
  there	
  are	
  many	
  similarities	
  between	
  
urothelial	
  carcinomas	
  in	
  dogs	
  and	
  humans;	
  there	
  are	
  also	
  many	
  differences	
  suggesting	
  that	
  
while	
  further	
  study	
  of	
  canine	
  urothelial	
  carcinoma	
  is	
  warranted,	
  canine	
  urothelial	
  
carcinoma	
  do	
  not	
  perfectly	
  recapitulate	
  similar	
  disease	
  in	
  humans

	
  

	
  
ACKNOWLEDGMENTS	
  
	
  
I	
  thank	
  Bristol-­‐Meyers-­‐Squibb	
  and	
  the	
  American	
  College	
  of	
  Veterinary	
  
Pathologists/Society	
  of	
  Toxicologic	
  Pathologists	
  coalition	
  for	
  funding	
  my	
  graduate	
  stipend.	
  
In	
  addition,	
  I	
  also	
  thank	
  all	
  those	
  that	
  contributed	
  to	
  the	
  Michigan	
  State	
  University	
  College	
  
of	
  Veterinary	
  Medicine	
  companion	
  animal	
  fund,	
  which	
  provided	
  funding	
  for	
  portions	
  of	
  the	
  
presented	
  work.	
  
	
  

I	
  recognize	
  and	
  thank	
  the	
  members	
  of	
  the	
  dissertation	
  committee,	
  Dr.	
  Matti	
  Kiupel,	
  

Dr.	
  Elizabeth	
  McNiel,	
  Dr.	
  Monica	
  Liebert,	
  Dr.	
  James	
  Resau,	
  and	
  Dr.	
  Richard	
  Westhouse,	
  for	
  
their	
  training,	
  assistance,	
  patience,	
  time,	
  and	
  knowledgeable	
  input	
  throughout	
  the	
  process	
  
of	
  compiling	
  this	
  dissertation.	
  	
  I	
  also	
  recognize	
  and	
  thank	
  the	
  many	
  scientists,	
  veterinary	
  
pathologists	
  and	
  veterinary	
  residents	
  from	
  the	
  Michigan	
  State	
  University	
  Department	
  of	
  
Pathobiology	
  and	
  Diagnostic	
  Investigation	
  and	
  Diagnostic	
  Center	
  for	
  Population	
  and	
  
Animal	
  Health	
  for	
  their	
  support.	
  	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

iv	
  

	
  
TABLE	
  OF	
  CONTENTS	
  
	
  
LIST	
  OF	
  TABLES…………………………………………………………………………...........………………………..viii	
  
	
  
LIST	
  OF	
  FIGURES………………………………………………………………..……………….………………….............x	
  
	
  
KEY	
  TO	
  ABBREVIATIONS………………………………………………………….……………………………...…..xiii	
  	
  
	
  
CHAPTER	
  1	
  Introduction	
  and	
  Literature	
  Review…………………………………………….………………..1	
  
	
  
Introduction:	
  The	
  dog	
  as	
  a	
  model	
  for	
  urothelial	
  carcinogenesis…………………….………..2	
  
	
  
Prognostication	
  in	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas……………………...6	
  	
  
	
  
Outline	
  of	
  studies……………………………………………………………………………………………….11	
  
	
  
	
  
Chapter	
  2.......................................................................................................................................12	
  
	
  
	
  
Chapter	
  3.......................................................................................................................................14	
  
	
  
	
  
Chapter	
  4.......................................................................................................................................17	
  
	
  
	
  
Chapter	
  5.......................................................................................................................................19	
  
	
  
REFERENCES..……………………………………………………………………………………………………22	
  
	
  
CHAPTER	
  2	
  Differences	
  in	
  Expression	
  of	
  Uroplakin	
  III,	
  Cytokeratin-­‐7,	
  and	
  COX-­‐2	
  in	
  Canine	
  
Proliferative	
  Urothelial	
  Lesions	
  of	
  the	
  Urinary	
  Bladder……………………………………………...…...35	
  
	
  
Abstract……………………………………………………………………………………………………………..36	
  
	
  
Introduction…………………………………………………………………..…………………………………..37	
  
	
  
Materials	
  and	
  methods……………………………………………………………………………………….40	
  
	
  
	
  
Selection	
  of	
  Cases	
  and	
  Histologic	
  Classification.............................................................40	
  
Immunohistochemistry............................................................................................................41	
  
Statistical	
  Analysis.....................................................................................................................42	
  
	
  
Results……………………………………………………………...…...……………………………………..……44	
  
	
  
	
  
Immunohistochemistry:	
  UPIII	
  and	
  CK7.............................................................................44	
  
Immunohistochemistry:	
  COX-­‐2.............................................................................................44	
  
Immunohistochemistry:	
  Activated	
  caspase	
  3..................................................................45	
  
Statistical	
  Analysis.....................................................................................................................46	
  
	
  
Discussion………………………………………………………………………………………………………….47	
  
	
  
Acknowledgements.................................................................................................................................49	
  
	
  
APPENDIX...................................................................................................................................................50	
  
	
  
REFERENCES.............................................................................................................................................58	
  
	
  
CHAPTER	
  3	
  Evaluation	
  of	
  15-­‐hydroxyprostaglandin	
  dehydrogenase	
  (HPGD),	
  
cyclooxygenase-­‐2	
  (COX-­‐2),	
  cadherin,	
  and	
  β-­‐catenin	
  expression	
  in	
  canine	
  urinary	
  bladder	
  
urothelial	
  carcinomas...........................................................................................................................................65	
  
	
  
Abstract........................................................................................................................................................66	
  	
  
	
  
Introduction...............................................................................................................................................67	
  
Materials	
  and	
  methods..........................................................................................................................70	
  

v	
  

	
  

	
  
	
  
	
  

Selection	
  of	
  Cases	
  and	
  Histologic	
  Classification.............................................................70	
  
Immunohistochemistry............................................................................................................70	
  
Statistical	
  Analysis.....................................................................................................................73	
  
Western	
  blots	
  ..............................................................................................................................73	
  
Results..........................................................................................................................................................75	
  
Demographic	
  information,	
  histologic	
  classification,	
  and	
  grading	
  of	
  urothelial	
  	
  
carcinomas.....................................................................................................................75	
  
Immunohistochemistry:	
  HPGD..............................................................................................76	
  
Immunohistochemistry:	
  COX-­‐2.............................................................................................76	
  
Immunohistochemistry:	
  Cadherins	
  ....................................................................................77	
  
Immunohistochemistry:	
  β-­‐catenin.......................................................................................79	
  
Survival	
  analysis.........................................................................................................................80	
  
Western	
  blot	
  analysis	
  of	
  human	
  and	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines......81	
  
Discussion...................................................................................................................................................83	
  
Acknowledgments...................................................................................................................................88	
  
APPENDIX...................................................................................................................................................89	
  
REFERENCES.............................................................................................................................................99	
  

CHAPTER	
  4	
  Evaluation	
  of	
  microsatellite	
  instability	
  and	
  DNA	
  mismatch	
  repair	
  protein	
  

expression	
  in	
  canine	
  urothelial	
  carcinomas.............................................................................................104	
  
	
  
Abstract.....................................................................................................................................................105	
  
	
  
Introduction.............................................................................................................................................106	
  
	
  
Materials	
  and	
  methods.......................................................................................................................109	
  
Evaluation	
  of	
  Microsatellite	
  Instability	
  in	
  Canine	
  Tumors......................................109	
  
	
  
	
  
Evaluation	
  of	
  MMR	
  protein	
  expression	
  in	
  Canine	
  Urothelial	
  Carcinomas.........110	
  
	
  
	
  
Statistical	
  analysis...................................................................................................................113	
  
	
  
Results........................................................................................................................................................114	
  
	
  
	
  
Evaluation	
  of	
  MSI.....................................................................................................................114	
  
Demographic	
  information,	
  histologic	
  classification,	
  and	
  grading	
  of	
  urothelial	
  	
  
carcinomas..................................................................................................................115	
  
MLH1,	
  MSH2,	
  MSH3,	
  and	
  MSH6	
  Immunohistochemistry.........................................116	
  
Survival	
  analysis......................................................................................................................117	
  
Discussion.................................................................................................................................................119	
  
Acknowledgements..............................................................................................................................125	
  
	
  
APPENDIX.................................................................................................................................................127	
  
	
  
REFERENCES...........................................................................................................................................137	
  
	
  
CHAPTER	
  5	
  Evaluation	
  of	
  DNA	
  mismatch	
  repair	
  in	
  novel	
  canine	
  lower	
  urinary	
  tract	
  
urothelial	
  carcinoma	
  cell	
  lines.......................................................................................................................143	
  
	
  
Abstract.....................................................................................................................................................144	
  
	
  
Introduction.............................................................................................................................................145	
  
	
  
Materials	
  and	
  methods.......................................................................................................................147	
  
Characterization	
  of	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  	
  
lines................................................................................................................................147	
  
Development	
  of	
  xenografts..................................................................................................148	
  
Immunophenotyping	
  of	
  primary	
  tumors,	
  cell	
  lines,	
  and	
  xenografts....................149	
  
vi	
  

	
  

	
  

Evaluation	
  of	
  MMR	
  protein	
  expression	
  by	
  Western	
  blots........................................149	
  
	
  
qPCR	
  for	
  MMR	
  genes..............................................................................................................150	
  	
  
Immunohistochemistry	
  and	
  morphometric	
  analysis	
  for	
  MSH2	
  and	
  MSH6	
  in	
  	
  
xenografts....................................................................................................................152	
  
Effect	
  of	
  chemotherapeutics	
  on	
  cell	
  survival.................................................................153	
  
Results........................................................................................................................................................155	
  
Characterization	
  of	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines	
  	
  
and	
  xenografts...........................................................................................................155	
  
Characterization	
  of	
  xenografts..........................................................................................156	
  
Evaluation	
  of	
  MMR	
  protein	
  expression	
  by	
  Western	
  blots........................................157	
  
qPCR	
  for	
  MMR	
  gene	
  expression..........................................................................................157	
  
Immunohistochemistry	
  and	
  morphometric	
  analysis	
  for	
  MSH2	
  and	
  MSH6	
  in	
  	
  
xenografts....................................................................................................................158	
  
Survival	
  assays..........................................................................................................................158	
  
Discussion.................................................................................................................................................160	
  
Acknowledgments................................................................................................................................164	
  
APPENDIX.................................................................................................................................................165	
  
REFERENCES...........................................................................................................................................189	
  

	
  
	
  
	
  
CHAPTER	
  6	
  Conclusions...................................................................................................................................193	
  
	
  
Summary	
  of	
  findings............................................................................................................................194	
  
	
  
Limitations	
  of	
  studies	
  and	
  unanswered	
  questions.................................................................197	
  
	
  
Future	
  directions...................................................................................................................................201	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

vii	
  

	
  
LIST	
  OF	
  TABLES	
  
	
  
Table	
  1:	
  Histologic	
  Features	
  of	
  Proliferative	
  Urothelial	
  Lesions	
  according	
  to	
  the	
  WHO/ISUP	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Consensus	
  Classification	
  System......................................................................................................54	
  
	
  
Table	
  2:	
  Immunohistochemical	
  Scoring	
  of	
  Uroplakin	
  III	
  Expression	
  in	
  Canine	
  Proliferative	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Urothelial	
  Lesions	
  by	
  Overall	
  Pattern...........................................................................................55	
  
	
  
Table	
  3:	
  Immunohistochemical	
  Scoring	
  of	
  Cytokeratin	
  7	
  Expression	
  in	
  Canine	
  Proliferative	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Urothelial	
  Lesions	
  by	
  Overall	
  Pattern...........................................................................................55	
  
	
  
Table	
  4:	
  Immunohistochemical	
  Scoring	
  of	
  COX-­‐2	
  Expression	
  in	
  Canine	
  Proliferative	
  Urothelial	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Lesions	
  by	
  Overall	
  Pattern.................................................................................................................56	
  
	
  
Table	
  5:	
  Immunohistochemical	
  Scoring	
  of	
  Activated	
  Caspase-­‐3	
  in	
  Canine	
  Proliferative	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Urothelial	
  Lesions	
  by	
  Overall	
  Pattern...........................................................................................56	
  
	
  
Table	
  6:	
  Summary	
  of	
  Patterns	
  of	
  Immunoreactivity	
  of	
  Uroplakin	
  III,	
  Cytokeratin	
  7,	
  and	
  COX-­‐2	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  in	
  Proliferative	
  Urothelial	
  Lesions..................................................................................................57	
  
	
  
Table	
  7:	
  Expression	
  of	
  HPGD	
  in	
  canine	
  urothelial	
  carcinomas	
  and	
  normal	
  urothelium	
  with	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  respect	
  to	
  lesion	
  classification	
  and	
  degree	
  of	
  invasion...........................................................96	
  
	
  
Table	
  8:	
  Expression	
  of	
  P-­‐cadherin	
  and	
  β-­‐catenin	
  in	
  canine	
  urothelial	
  carcinomas	
  and	
  normal	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  urothelium	
  with	
  respect	
  to	
  lesion	
  classification	
  and	
  degree	
  of	
  invasion.........................97	
  
	
  
Table	
  9:	
  Significant	
  Correlative	
  Results	
  of	
  Univariate	
  and	
  Multivariate	
  Analysis	
  of	
  Examined	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Variables	
  with	
  Respect	
  to	
  Survival	
  Time	
  as	
  Determined	
  by	
  p-­‐values	
  less	
  than	
  0.05...98	
  
	
  
Table	
  10:	
  Distribution	
  of	
  microsatellite	
  instability	
  in	
  canine	
  epithelial	
  tumors.........................134	
  
	
  
Table11:	
  Distribution	
  of	
  microsatellite	
  instability	
  in	
  canine	
  urothelial	
  carcinomas	
  of	
  the	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  urinary	
  bladder	
  classified	
  by	
  breed...........................................................................................134	
  
	
  
Table	
  12:	
  Distribution	
  of	
  microsatellite	
  instability	
  in	
  canine	
  urothelial	
  carcinomas	
  of	
  the	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  urinary	
  bladder	
  by	
  breed	
  and	
  repeat	
  motif..........................................................................135	
  
	
  
Table	
  13:	
  Distribution	
  of	
  %MSI	
  and	
  immunohistochemically	
  evaluated	
  MMR	
  protein	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  expression	
  in	
  canine	
  normal	
  urinary	
  bladders	
  and	
  urothelial	
  carcinomas	
  of	
  the	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  urinary	
  bladder................................................................................................................................136	
  
	
  
Table	
  14:	
  TaqMan®	
  gene	
  expression	
  assays	
  used	
  for	
  qPCR...............................................................187	
  
	
  
	
  
viii	
  

	
  
Table	
  15:	
  Distribution	
  of	
  MSH2	
  expression	
  in	
  canine	
  urothelial	
  carcinoma	
  xenogafts	
  as	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  evaluated	
  by	
  morphometric	
  analysis	
  of	
  immunohistochemistry..................................187	
  
	
  
Table	
  16:	
  Distribution	
  of	
  MSH6	
  expression	
  in	
  canine	
  urothelial	
  carcinoma	
  xenogafts	
  as	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  evaluated	
  by	
  morphometric	
  analysis	
  of	
  immunohistochemistry..................................188	
  
	
  
	
  

	
  

ix	
  

	
  
LIST	
  OF	
  FIGURES	
  
	
  
Figure	
  1:	
  Urinary	
  Bladder;	
  Dog.	
  Hyperplastic	
  urotheluim	
  with	
  Uroplakin	
  III	
  (UPIII)	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Pattern	
  1.................................................................................................................................................51	
  
	
  
Figure	
  2:	
  Urinary	
  Bladder;	
  Dog.	
  Hyperplastic	
  urothelium	
  with	
  UPIII	
  IHC	
  Pattern	
  2..................51	
  
	
  
Figure	
  3:	
  Urinary	
  Bladder;	
  Dog.	
  Papillary	
  urothelial	
  carcinoma	
  grade	
  II	
  with	
  UPIII	
  IHC	
  	
  
	
  
	
  	
  	
  	
  	
  Pattern	
  4.................................................................................................................................................51	
  
	
  
Figure	
  4:	
  Urinary	
  Bladder;	
  Dog.	
  Hyperplastic	
  urotheluim	
  with	
  Cytokeratin	
  7	
  (CK7)	
  IHC	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Pattern	
  1................................................................................................................................................52	
  
	
  
Figure	
  5:	
  Urinary	
  Bladder;	
  Dog.	
  Papillary	
  urothelial	
  carcinoma	
  grade	
  II	
  with	
  CK7	
  IHC	
  	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Pattern	
  	
  3...............................................................................................................................................52	
  
	
  
Figure	
  6:	
  Urinary	
  Bladder;	
  Dog.	
  Papillary	
  urothelial	
  carcinoma	
  grade	
  II	
  with	
  CK7	
  IHC	
  	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Pattern	
  4................................................................................................................................................52	
  
	
  
Figure	
  7:	
  Urinary	
  Bladder;	
  Dog.	
  Hyperplastic	
  urotheluim	
  with	
  Cyclooxygenase-­‐2	
  (COX-­‐2)	
  IHC	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Pattern	
  1................................................................................................................................................53	
  
	
  
Figure	
  8:	
  Urinary	
  Bladder;	
  Dog.	
  Hyperplastic	
  urotheluim	
  with	
  COX-­‐2	
  IHC	
  Pattern	
  2I..............53	
  
	
  
Figure	
  9:	
  Urinary	
  Bladder;	
  Dog.	
  Papillary	
  urothelial	
  carcinoma	
  grade	
  II	
  with	
  COX-­‐2	
  IHC	
  	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Pattern	
  3................................................................................................................................................53	
  
	
  
Figure	
  10:	
  Urinary	
  bladder,	
  Dog,	
  15-­‐hydroxyprostaglandin	
  dehydrogenase	
  (HPGD)	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  immunohistochemistry	
  (IHC)......................................................................................................90	
  
	
  
Figure	
  11:	
  Urinary	
  bladder,	
  Dog,	
  Cyclooxygenase-­‐2	
  (COX-­‐2)	
  IHC.......................................................91	
  
	
  
Figure	
  12:	
  Urinary	
  bladder,	
  Dog,	
  E-­‐cadherin	
  IHC.....................................................................................92	
  
	
  
Figure	
  13:	
  Urinary	
  bladder,	
  Dog,	
  P-­‐cadherin	
  IHC.....................................................................................93	
  
	
  
Figure	
  14:	
  Urinary	
  bladder,	
  Dog,	
  β-­‐catenin	
  IHC........................................................................................94	
  
	
  
Figure	
  15:	
  Western	
  blot	
  comparing	
  expression	
  of	
  HPGD,	
  COX-­‐2,	
  E-­‐cadherin,	
  P-­‐cadherin,	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  N-­‐cadherin,	
  and	
  β-­‐catenin	
  in	
  human	
  (RT4	
  and	
  UC3)	
  and	
  canine	
  urothelial	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  carcinoma	
  cell	
  lines	
  (AXA,	
  AXB,	
  AXC,	
  and	
  NK)......................................................................95	
  
	
  
Figure	
  16:	
  Unmixed	
  composite	
  images	
  and	
  unmixed	
  pseudofluorecent	
  images	
  derived	
  from	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  multispectral	
  imaging	
  of	
  MLH1,	
  MSH2,	
  MSH3,	
  and	
  MSH6	
  immunohistochemically	
  	
  
x	
  

	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  labeled	
  urothelium	
  from	
  a	
  normal	
  urinary	
  bladder	
  control	
  and	
  a	
  grade	
  II	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  papillary	
  urothelial	
  carcinoma................................................................................................128	
  
	
  
Figure	
  17:	
  Scatter	
  plot	
  depicting	
  the	
  frequency	
  of	
  microsatellite	
  aberrations	
  identified	
  in	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  canine	
  gastric	
  carcinoma	
  tumor	
  (N=15),	
  mammary	
  tumors	
  (N=35),	
  and	
  urothelial	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  carcinomas	
  of	
  the	
  urinary	
  bladder	
  (N=46)..........................................................................129	
  
	
  
Figure	
  18:	
  Scatter	
  plot	
  depicting	
  the	
  frequency	
  of	
  microsatellite	
  aberrations	
  identified	
  in	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  canine	
  urothelial	
  carcinomas	
  stratified	
  by	
  breed.............................................................130	
  
	
  
Figure	
  19:	
  Scatter	
  plot	
  depicting	
  the	
  frequency	
  of	
  microsatellite	
  aberrations	
  identified	
  in	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  canine	
  urothelial	
  carcinomas	
  stratified	
  by	
  phylogenetic	
  clade..................................131	
  
	
  
Figure	
  20:	
  Scatter	
  plots	
  depicting	
  immunoreactivity	
  of	
  MSH2	
  and	
  MSH6	
  in	
  canine	
  normal	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  urinary	
  bladder	
  urothelium	
  and	
  urothelial	
  carcinomas	
  as	
  determined	
  by	
  	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  morphometric	
  analysis	
  of	
  multispectral	
  imaging............................................................132	
  
	
  
Figure	
  21:	
  Scatter	
  plot	
  depicting	
  ratio	
  of	
  MSH2	
  and	
  MSH6	
  immunoreactivity	
  in	
  individual	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  cases	
  of	
  canine	
  normal	
  urinary	
  bladder	
  urothelium	
  and	
  urothelial	
  carcinomas	
  as	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  determined	
  by	
  morphometric	
  analysis	
  of	
  multispectral	
  imaging..............................133	
  
	
  
Figure	
  22.	
  Multispectral	
  imaging	
  of	
  MSH2	
  immunohistochemistry	
  in	
  canine	
  lower	
  urinary	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  tract	
  urothelial	
  carcinoma	
  xenografts.................................................................................166	
  
	
  
Figure	
  23:	
  Multispectral	
  imaging	
  of	
  MSH6	
  immunohistochemistry	
  in	
  canine	
  lower	
  urinary	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  tract	
  urothelial	
  carcinoma	
  xenografts.................................................................................168	
  
	
  
Figure	
  24:	
  Photomicrographs	
  from	
  the	
  initial	
  diagnostic	
  biopsy	
  and	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  immunohistochemistry	
  (IHC)	
  on	
  the	
  papillary	
  grade	
  II	
  urothelial	
  from	
  which	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  TYLER	
  cell	
  lines	
  were	
  derived...................................................................................................170	
  
	
  
Figure	
  25:	
  	
  Phase	
  contrast	
  photomicrographs	
  of	
  ANGUS,	
  KINSEY,	
  original	
  TYLER	
  cell	
  	
  	
  	
  	
  	
  	
  	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  Lines..................................................................................................................................................171	
  
	
  
Figure	
  26:	
  Phase	
  contrast	
  photomicrographs	
  of	
  TYLER1	
  and	
  TYLER2	
  cell	
  lines	
  and	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  photomicrographs	
  of	
  IHC	
  for	
  differentiation	
  markers	
  in	
  TYLER1	
  and	
  TYLER2	
  cell	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  lines.....................................................................................................................................................172	
  
	
  
Figure	
  27:	
  Photomicrographs	
  from	
  a	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  xenograft	
  derived	
  from	
  the	
  TYLER2	
  cell	
  line	
  showing	
  two	
  morphologically	
  distinct	
  	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  cell	
  populations,	
  hematoxylin	
  and	
  eosin	
  stain...................................................................173	
  
	
  
Figure	
  28:	
  Photomicrographs	
  of	
  IHC	
  for	
  differentiation	
  markers	
  in	
  a	
  canine	
  lower	
  urinary	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  tract	
  urothelial	
  carcinoma	
  xenograft	
  derived	
  from	
  the	
  KINSEY	
  cell	
  line................174	
  
	
  
Figure	
  29:	
  Photomicrographs	
  of	
  IHC	
  for	
  differentiation	
  markers	
  in	
  a	
  canine	
  lower	
  urinary	
  	
  
xi	
  

	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  tract	
  urothelial	
  carcinoma	
  xenograft	
  derived	
  from	
  the	
  TYLER2	
  cell	
  line...............175	
  
	
  
Figure	
  30:	
  Western	
  blots	
  show	
  relative	
  decreased	
  expression	
  of	
  MSH2	
  and	
  MSH6	
  in	
  the	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  TYLER2	
  cell	
  line	
  in	
  comparison	
  to	
  that	
  of	
  ANGUS,	
  KINSEY,	
  and	
  TYLER1...............176	
  
	
  
Figure	
  31:	
  Graph	
  depicting	
  results	
  of	
  qPCR	
  for	
  MLH1,	
  MSH2,	
  MSH3,	
  MSH6,	
  and	
  PMS2	
  in	
  the	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines,	
  ANGUS,	
  KINSEY,	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  TYLER1,	
  and	
  TYLER2	
  cell	
  lines.................................................................................................177	
  
	
  
Figure	
  32:	
  Graph	
  depicting	
  results	
  of	
  qPCR	
  for	
  MSH2	
  and	
  MSH6	
  in	
  canine	
  lower	
  urinary	
  tract	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  urothelial	
  carcinomas	
  xenografts	
  derived	
  from	
  TYLER1,	
  and	
  TYLER2	
  cell	
  lines	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  laser	
  capture	
  microdissection	
  separated	
  from	
  mouse	
  tissues.....................................179	
  
	
  
Figure	
  33:	
  Graph	
  depicting	
  results	
  of	
  qPCR	
  for	
  MSH2	
  in	
  laser	
  capture	
  microdissection	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  separated	
  epitheliod	
  and	
  discrete	
  cell	
  populations	
  of	
  canine	
  lower	
  urinary	
  tract	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  urothelial	
  carcinomas	
  xenografts	
  derived	
  from	
  TYLER1,	
  and	
  TYLER2	
  cell	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  lines.....................................................................................................................................................181	
  
	
  
Figure	
  34:	
  Graph	
  depicting	
  results	
  of	
  morphometric	
  analysis	
  of	
  multispectral	
  imaging	
  of	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  immunoreactivity	
  of	
  MSH2	
  and	
  MSH6	
  in	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  carcinomas	
  xenografts	
  derived	
  from	
  KINSEY,	
  TYLER1,	
  and	
  TYLER2	
  cell	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  lines.....................................................................................................................................................183	
  
	
  
Figure	
  35:	
  Results	
  of	
  XTT	
  survival	
  assays	
  comparing	
  survival	
  of	
  canine	
  urothelial	
  carcinoma	
  	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  cell	
  lines	
  upon	
  exposure	
  to	
  specific	
  chemotherapeutics.................................................185	
  
	
  
	
  

	
  

xii	
  

	
  
KEY	
  TO	
  ABBREVIATIONS	
  
	
  
CK7:	
  Cytokeratin	
  7	
  
COX-­‐2:	
  Cyclooxygenase-­‐2	
  	
  
DAB:	
  3,3’-­‐Diaminobenzidine	
  
GAPDH:	
  Glyceraldehyde	
  3-­‐phosphate	
  dehydrogenase	
  
HPGD:15-­‐hydroxyprostaglandin	
  dehydrogenase	
  	
  
IHC:	
  Immunohistochemistry	
  	
  
MMR:	
  DNA	
  mismatch	
  repair	
  
MS:	
  Microsatellite	
  
MSI:	
  Microsatellite	
  instability	
  
PAP:	
  Prostatic	
  acid	
  phosphatase	
  
PGE2:	
  Prostaglandin	
  E2	
  	
  
PUNLMP:	
  Papillary	
  urothelial	
  neoplasm	
  of	
  low	
  malignant	
  potential	
  	
  
qPCR:	
  quantitative	
  real-­‐time	
  PCR	
  
SD:	
  Standard	
  deviation	
  
SDS-­‐PAGE:	
  Sodium	
  dodecyl	
  sulfate	
  polyacrylamide	
  gel	
  electrophoresis	
  
SE:	
  Standard	
  error	
  	
  
TBST:	
  Tris-­‐buffered	
  saline	
  with	
  0.1%	
  Tween	
  20	
  
UPIII:	
  Uroplakin	
  III	
  
XTT:	
  2,3-­‐Bis-­‐(2-­‐Methoxy-­‐4-­‐Nitro-­‐5-­‐Sulfophenyl)-­‐2H-­‐Tetrazolium-­‐5-­‐Carboxanilide	
  
	
  
	
  

xiii	
  

	
  
CHAPTER	
  1	
  
	
  
	
  
Introduction	
  and	
  Literature	
  Review	
  
	
  

	
  
Dodd	
  Sledge	
  
	
  
Department	
  of	
  Pathobiology	
  and	
  Diagnostic	
  Investigation,	
  College	
  of	
  Veterinary	
  Medicine,	
  
Michigan	
  State	
  University,	
  East	
  Lansing	
  MI	
  

	
   1	
  

	
  
Introduction:	
  The	
  dog	
  as	
  a	
  model	
  for	
  urothelial	
  carcinogenesis	
  
Bladder	
  cancer	
  is	
  a	
  significant	
  problem	
  in	
  humans.	
  Based	
  on	
  current	
  trends,	
  2.4%	
  of	
  
people	
   in	
   the	
   United	
   States	
   are	
   predicted	
   to	
   develop	
   bladder	
   cancer	
   during	
   the	
   course	
   of	
  
their	
  life.1	
  An	
  estimated	
  74,690	
  people	
  have	
  or	
  will	
  be	
  diagnosed	
  and	
  15,580	
  are	
  expected	
  
die	
   due	
   to	
   the	
   disease	
   in	
   the	
   United	
   States	
   during	
   2014.1	
   In	
   2008,	
   bladder	
   cancer	
   was	
  
reported	
  as	
  the	
  fourth	
  most	
  common	
  cancer	
  diagnosis	
  in	
  men,	
  and	
  the	
  ninth	
  most	
  common	
  
in	
   women.2	
   Although	
   the	
   majority	
   (70-­‐80%)	
   of	
   bladder	
   cancers	
   in	
   humans	
   are	
   organ-­‐
confined	
  and	
  considered	
  “superficial,”	
  the	
  recurrence	
  rate	
  may	
  be	
  as	
  high	
  as	
  50%,	
  and	
  as	
  
many	
   as	
   25%	
   of	
   these	
   patients	
   will	
   progress	
   to	
   invasive	
   disease,	
   in	
   spite	
   of	
   treatment.3	
  
Another	
   20-­‐30%	
   of	
   patients	
   have	
   advanced	
   bladder	
   cancer	
   at	
   diagnosis.4	
   Therapies	
   for	
  
invasive	
  bladder	
  cancer	
  are	
  minimally	
  effective.	
  Even	
  with	
  treatment,	
  the	
  median	
  survival	
  
time	
   is	
   less	
   than	
   2	
   years	
   from	
   initial	
   diagnosis.4	
   Little	
   improvement	
   in	
   survival	
   has	
   been	
  
observed	
  in	
  the	
  past	
  30	
  years,	
  as	
  evidenced	
  by	
  slowly	
  increasing	
  numbers	
  of	
  new	
  cases	
  and	
  
deaths	
  per	
  100,000	
  people	
  over	
  this	
  time	
  period.1	
  From	
  a	
  financial	
  view,	
  bladder	
  cancer	
  is	
  
among	
   the	
   costliest	
   cancers	
   to	
   manage	
   and	
   treat.5	
   For	
   these	
   reasons,	
   an	
   improved	
  
understanding	
   of	
   urothelial	
   carcinogenesis	
   and	
   novel,	
   cost-­‐effective	
   therapeutic	
  
approaches	
  are	
  urgently	
  needed.	
  
Establishment	
  and	
  utilization	
  of	
  preclinical	
  animal	
  models	
  beyond	
  the	
  typical	
  rodent	
  
model	
  that	
  reflect	
  the	
  biology	
  of	
  human	
  cancers	
  would	
  have	
  great	
  impact	
  on	
  drug	
  
development.	
  Chemically	
  induced,	
  transgenic,	
  and	
  xenograft	
  rodent	
  models	
  have	
  been	
  well	
  
described	
  and	
  are	
  commonly	
  used	
  for	
  the	
  study	
  of	
  cancer;	
  however,	
  such	
  models	
  of	
  often	
  
fail	
  predict	
  treatment	
  success	
  and	
  to	
  guide	
  the	
  rational	
  development	
  of	
  early	
  phase	
  human	
  
trials.6-­‐8	
  In	
  general,	
  dogs	
  offer	
  a	
  cost-­‐effective,	
  naturally	
  occurring	
  model	
  for	
  cancer	
  in	
  

	
   2	
  

	
  
which	
  clinical	
  trials	
  are	
  easily	
  preformed.	
  	
  Among	
  other	
  factors,	
  dogs	
  are	
  attractive	
  for	
  such	
  
studies	
  because	
  the	
  general	
  types	
  and	
  histomorphology	
  of	
  canine	
  tumors	
  are	
  similar	
  to	
  
those	
  of	
  humans,	
  cancer	
  is	
  naturally	
  occurring	
  in	
  dogs,	
  drug	
  metabolism	
  is	
  similar,	
  dogs	
  
share	
  environments	
  with	
  their	
  owners	
  and	
  often	
  have	
  similar	
  general	
  lifestyles,	
  the	
  larger	
  
size	
  of	
  dogs	
  in	
  comparison	
  to	
  rodents	
  allows	
  for	
  easier	
  performance	
  of	
  many	
  clinical	
  
procedures,	
  and	
  lifespans	
  of	
  dogs	
  are	
  longer	
  than	
  rodents,	
  yet	
  still	
  relatively	
  short	
  in	
  
comparison	
  to	
  humans.9,	
  10	
  In	
  addition,	
  the	
  genetic	
  diversity	
  of	
  dogs	
  more	
  closely	
  mirrors	
  
that	
  of	
  humans	
  than	
  that	
  seen	
  in	
  laboratory	
  rodents,	
  while	
  individual	
  breeds	
  provide	
  
genetic	
  similarities	
  and	
  often	
  pedigree	
  information.11	
  Further,	
  findings	
  of	
  studies	
  in	
  dogs	
  
can	
  be	
  applied	
  to	
  improving	
  clinical	
  outcome	
  not	
  only	
  in	
  human	
  medicine,	
  but	
  also	
  can	
  be	
  
used	
  in	
  veterinary	
  medicine.	
  
Many	
  features	
  make	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas	
  attractive	
  for	
  
study.	
  Similar	
  to	
  humans,	
  bladder	
  cancer	
  is	
  common	
  in	
  dogs.	
  A	
  reported	
  2%	
  of	
  malignant	
  
tumors	
   of	
   the	
   dog	
   originate	
   from	
   the	
   bladder,	
   and	
   most	
   of	
   these	
   are	
   derived	
   from	
   the	
  
urothelium.10,	
   12	
   Such	
   tumors	
   occur	
   spontaneously	
   rather	
   than	
   necessitating	
   exposure	
   to	
  
carcinogens	
   or	
   genetic	
   manipulation.13,	
   14	
   Clinical	
   signs	
   including	
   hematuria,	
   increased	
  
frequency	
   of	
   urination,	
   and	
   stranguria	
   are	
   common	
   to	
   both	
   dogs	
   and	
   humans,	
   as	
   are	
  
secondary	
   bacterial	
   urinary	
   tract	
   infections.15	
   Macroscopic	
   and	
   microscopic	
   findings	
   are	
  
similar	
   between	
   dogs	
   and	
   humans.16,	
   17	
   Canine	
   bladder	
   cancer	
   typically	
   follows	
   a	
   similar	
  
clinical	
  course	
  to	
  that	
  reported	
  in	
  invasive	
  bladder	
  carcinomas	
  of	
  humans	
  suggesting	
  that	
  
the	
   pathways	
   involved	
   in	
   the	
   development	
   and	
   progression	
   of	
   these	
   tumors	
   may	
   be	
  
homologous.13,	
   15	
   In	
   addition,	
   there	
   are	
   strong	
   breed	
   predispositions	
   for	
   development	
   of	
  
these	
   cancers	
   in	
   dogs	
   with	
   Scottish	
   terriers,	
   West	
   Highland	
   white	
   terriers,	
   beagles,	
   and	
  

	
   3	
  

	
  
Shetland	
   sheepdogs	
   being	
   overrepresented.13,	
  

14	
  

This	
   suggests	
   a	
   potential	
   hereditary	
  

component	
   to	
   the	
   development	
   of	
   these	
   tumors	
   in	
   a	
   subset	
   of	
   dogs	
   and	
   thus,	
   a	
   common	
  
molecular	
  pathogenesis	
  for	
  cancer	
  development.	
  
Studies	
  of	
  molecular	
   features	
   associated	
   with	
   carcinogenesis	
   are	
   severely	
   lacking	
  in	
  
the	
  dog;	
  however,	
  few	
  studies	
  in	
  canine	
  urothelial	
  carcinomas	
  evaluating	
  potential	
  cancer	
  
markers	
  have	
  yielded	
  results	
  similar	
  to	
  that	
  in	
  humans.	
  	
  COX-­‐2	
  overexpression	
  is	
  common	
  
in	
  both	
  human	
  and	
  canine	
  bladder	
  cancer,	
  as	
  is	
  the	
  expression	
  of	
  prostaglandin	
  E2.10,	
   15,	
   18	
  
High	
   levels	
   of	
   p53	
   expression	
   has	
   been	
   reported	
   in	
   a	
   small	
   number	
   of	
   canine	
   urothelial	
  
carcinomas,	
   similar	
   to	
   reports	
   in	
   humans.10	
   Basic	
   fibroblast	
   growth	
   factor,	
   a	
   proangiotic	
  
factor,	
  has	
  been	
  reported	
  in	
  the	
  urine	
  of	
  affected	
  dogs	
  and	
  is	
  expressed	
  in	
  human	
  urothelial	
  
carcinomas.10	
  Nuclear	
  expression	
  of	
  survivin,	
  an	
  apoptosis-­‐inhibiting	
  protein,	
  is	
  described	
  
in	
  many	
  human	
  and	
  canine	
  urothelial	
  carcinomas,	
  but	
  is	
  not	
  expressed	
  in	
  normal	
  bladder.19,	
  
20	
   Telomerase	
   activity	
   is	
   reported	
   in	
   90%	
   of	
   human	
   urothelial	
   carcinomas,	
   and	
   has	
   been	
  

described	
   in	
   canine	
   urothelial	
   carcinoma	
   cell	
   lines	
   and	
   in	
   urine	
   samples	
   from	
   affected	
  
dogs.10	
   	
   	
   Expression	
   of	
   the	
   receptor	
   tyrosine	
   kinase,	
   epidermal	
   growth	
   factor	
   receptor,	
  
which	
  is	
  commonly	
  observed	
  in	
  human	
  urinary	
  bladder	
  cancer,	
  was	
  significantly	
  higher	
  in	
  
canine	
   urothelial	
   carcinomas	
   in	
   comparison	
   to	
   normal	
   urotheluim	
   and	
   cases	
   of	
   polypoid	
  
cystitis.21	
   β-­‐catenin	
   and	
   p63	
   expression	
   were	
   significantly	
   lower	
   and	
   Ki67	
   expression	
  
higher	
   in	
   urothelial	
   carcinomas	
   in	
   comparison	
   to	
   normal	
   urinary	
   bladder	
   and	
   cases	
   of	
  
polypoid	
  cystitis.22	
  	
  
While	
   there	
   are	
   numerous	
   similarities	
   in	
   dogs	
   and	
   humans	
   in	
   terms	
   of	
   urothelial	
  
carcinomas,	
   there	
   are	
   also	
   differences.	
   	
   In	
   humans,	
   men	
   develop	
   urothelial	
   carcinomas	
  
twice	
   as	
   often	
   as	
   women;	
   however	
   in	
   dogs,	
   females	
   are	
   nearly	
   twice	
   as	
   likely	
   to	
   be	
   affected	
  

	
   4	
  

	
  
as	
   males.1,	
   10,	
   12,	
   15	
   In	
   dogs,	
   the	
   vast	
   majority	
   of	
   bladder	
   cancers	
   occur	
   within	
   the	
   trigone	
  
area,	
  while	
  in	
  humans	
  occurrence	
  is	
  more	
  randomly	
  distributed	
  throughout	
  the	
  bladder.15	
  
This	
   difference	
   is	
   interesting,	
   as	
   it	
   has	
   classically	
   been	
   proposed	
   that	
   the	
   trigone	
   is	
  
embryologically	
   distinct	
   from	
   the	
   remainder	
   of	
   the	
   urinary	
   bladder	
   originating	
   from	
   the	
  
mesoderm	
  derived	
  Wolffian	
  ducts;	
  however,	
  more	
  recent	
  studies	
  have	
  suggested	
  that	
  the	
  
trigone	
  develops	
  from	
  endoderm	
  similar	
  to	
  the	
  rest	
  of	
  the	
  bladder.23,	
   24	
  More	
  importantly,	
  
in	
   situ	
   carcinomas	
   are	
   only	
   rarely	
   identified	
   in	
   dogs,	
   while	
   they	
   are	
   by	
   far	
   the	
   most	
  
commonly	
   recognized	
   form	
   of	
   urothelial	
   carcinoma	
   in	
   humans.13,	
   14,	
   25	
   This	
   fact	
   limits	
   the	
  
dog	
  as	
  a	
  model	
  suggesting	
  that	
  not	
  all	
  pathways	
  leading	
  to	
  urothelial	
  carcinomas	
  in	
  humans	
  
are	
   active	
   in	
   the	
   dog;	
   however,	
   the	
   dog	
   can	
   still	
   serve	
   as	
   an	
   excellent	
   model	
   for	
   invasive	
  
human	
  urothelial	
  carcinomas.	
  
	
  
	
  

	
  

	
   5	
  

	
  
Prognostication	
  in	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas	
  
Naturally	
   occurring	
   lower	
   tract	
   urothelial	
   carcinomas	
   comprise	
   a	
   significant	
  
proportion	
   of	
   canine	
   neoplasms;	
   however,	
   discrete	
   criteria	
   for	
   prognostic	
   evaluation	
   and	
  
targeted	
   treatment	
   options	
   are	
   lacking.	
   	
   Up	
   to	
   90%	
   of	
   canine	
   urothelial	
   tumors	
   of	
   the	
  
urinary	
   bladder	
   are	
   diagnosed	
   as	
   malignant,	
   most	
   are	
   highly	
   invasive,	
   and	
   many	
  
metastasize	
  to	
  regional	
  and	
  distant	
  sites.26	
  Radical	
  cystectomy	
  is	
  considered	
  the	
  treatment	
  
of	
  choice	
  for	
  invasive	
  human	
  urothelial	
  carcinomas,	
  which	
  are	
  similar	
  to	
  those	
  occurring	
  in	
  
the	
   dog.10	
   This,	
   however,	
   is	
   rarely	
   feasible	
   in	
   pet	
   dogs.	
   Because	
   complete	
   surgical	
   resection	
  
is	
  rarely	
  achievable	
  in	
  dogs	
  and	
  30-­‐50%	
  of	
  these	
  cancers	
  metastasize,	
  most	
  animals	
  die	
  of	
  
their	
   disease.14	
   Although	
   affected	
   dogs	
   are	
   frequently	
   treated	
   with	
   chemotherapy,	
   these	
  
cancers	
   respond	
   inconsistently.	
   However,	
   the	
   choice	
   of	
   chemotherapeutic	
   for	
   urothelial	
  
carcinomas	
   has	
   traditionally	
   been	
   empirical	
   rather	
   than	
   being	
   targeted	
   and	
   based	
   on	
  
scientific	
   evidence.14,	
   27	
   Most	
   dogs	
   are	
   treated	
   with	
   chemotherapy	
   and	
   nonsteroidal	
   anti-­‐
inflammatory	
  drugs,	
  which	
  tend	
  to	
  be	
  palliative	
  and	
  not	
  curative.	
  Recently,	
  multimodality,	
  
bladder-­‐preserving	
   strategies	
   for	
   invasive	
   human	
   bladder	
   cancer	
   have	
   been	
   shown	
   to	
   be	
  
very	
  effective	
  in	
  certain	
  patients,	
  although	
  such	
  approaches	
  have	
  not	
  been	
  optimized.28	
  	
  
The	
  prognostic	
  significance	
  of	
  clinico-­‐demographic	
  features	
  has	
  been	
  evaluated	
  in	
  
multiple	
  studies.	
  In	
  general,	
  correlations	
  of	
  individual	
  with	
  prognostic	
  measures	
  in	
  such	
  
studies	
  have	
  been	
  inconsistent	
  and	
  sometimes	
  contradictory.	
  This,	
  however,	
  may	
  be	
  due	
  to	
  
low	
  sample	
  sizes	
  within	
  individual	
  studies,	
  differences	
  in	
  treatment,	
  differences	
  in	
  what	
  
prognostic	
  measure	
  was	
  used,	
  or	
  differences	
  in	
  grouping	
  of	
  cases	
  for	
  analysis.	
  	
  Most	
  
commonly,	
  studies	
  have	
  employed	
  survival	
  time	
  as	
  the	
  main	
  prognostic	
  indicator,	
  but	
  

	
   6	
  

	
  
survival	
  time	
  can	
  be	
  highly	
  influenced	
  by	
  treatment,	
  euthanasia,	
  and	
  other	
  factors.	
  	
  Few	
  
studies	
  have	
  examined	
  other	
  measures	
  of	
  prognosis	
  such	
  as	
  progression	
  free	
  interval.	
  
In	
  general,	
  basic	
  demographic	
  information	
  such	
  as	
  age,	
  sex,	
  and	
  breed	
  are	
  unlikely	
  
to	
  influence	
  prognosis;	
  however,	
  reports	
  vary.	
  In	
  one	
  study,	
  sex	
  was	
  associated	
  with	
  
survival	
  time	
  with	
  spayed	
  females	
  surviving	
  significantly	
  longer	
  than	
  castrated	
  males.29	
  In	
  
another	
  study,	
  there	
  was	
  no	
  difference	
  in	
  survival	
  time	
  when	
  sex.30	
  One	
  study	
  found	
  a	
  
negative	
  correlation	
  between	
  breed	
  and	
  survival	
  time	
  with	
  at-­‐risk	
  breeds	
  having	
  shorter	
  
survival.31	
  Other	
  studies	
  have	
  found	
  no	
  significant	
  association	
  between	
  survival	
  time	
  and	
  
breed.15,	
  29,	
  32	
  Age	
  at	
  diagnosis	
  has	
  been	
  suggested	
  to	
  not	
  be	
  associated	
  with	
  survival	
  time	
  in	
  
multiple	
  studies.	
  29,	
  30	
  In	
  addition,	
  patient	
  weight	
  at	
  the	
  time	
  of	
  diagnosis	
  was	
  not	
  associated	
  
with	
  survival	
  in	
  at	
  least	
  one	
  study.32	
  
Features	
  such	
  as	
  tumor	
  location	
  as	
  it	
  relates	
  to	
  potential	
  uretheral	
  obstruction,	
  
clinical	
  stage,	
  and	
  surgical	
  respectability	
  are	
  likely	
  to	
  influence	
  clinical	
  outcome,	
  but	
  
correlation	
  of	
  prognostic	
  measures	
  with	
  such	
  factors	
  has	
  varied	
  between	
  study.	
  
Ultrasonographic	
  evidence	
  of	
  wall	
  involvement,	
  heterogeneity	
  of	
  masses,	
  and	
  location	
  of	
  
tumors	
  within	
  the	
  trigone	
  had	
  significant	
  negative	
  associations	
  with	
  survival	
  time.33	
  In	
  
addition,	
  urethral	
  involvement	
  has	
  been	
  suggested	
  to	
  have	
  a	
  significant	
  negative	
  
association	
  with	
  survival	
  in	
  multiple	
  studies.31,	
  34	
  However,	
  other	
  studies	
  have	
  found	
  no	
  
association	
  between	
  survival	
  time	
  and	
  location	
  of	
  the	
  tumor	
  or	
  urethral	
  involvement.	
  35,	
  36	
  
One	
  study	
  found	
  negative	
  associations	
  between	
  progression	
  free	
  interval	
  and	
  presence	
  of	
  
metastasis	
  at	
  the	
  time	
  of	
  presentation	
  and	
  negative	
  associations	
  between	
  survival	
  time	
  and	
  
the	
  presence	
  of	
  distant	
  metastasis.31	
  Other	
  studies	
  in	
  which	
  clinical	
  staging	
  was	
  preformed	
  
found	
  no	
  statistical	
  association	
  between	
  survival	
  time	
  and	
  variable	
  features	
  including	
  T	
  

	
   7	
  

	
  
stage,	
  N	
  stage,	
  M	
  stage,	
  or	
  presence	
  of	
  nodal	
  or	
  other	
  metastasis.15,	
  30,	
  32,	
  37	
  In	
  terms	
  of	
  
surgical	
  resection,	
  a	
  statistically	
  significant	
  longer	
  survival	
  time	
  was	
  associated	
  with	
  
complete	
  resection	
  compared	
  to	
  only	
  incomplete	
  resection,	
  and	
  in	
  another	
  study,	
  surgical	
  
debulking	
  prolonged	
  survival,	
  but	
  did	
  not	
  prolong	
  progression	
  free	
  survival.34,	
  38	
  
Multiple	
  histologic	
  grading	
  schemes	
  have	
  been	
  proposed	
  largely	
  based	
  on	
  systems	
  
developed	
   for	
   use	
   in	
   humans.	
   In	
   1995,	
   Valli,	
   et	
   al.,	
   described	
   a	
   classification	
   and	
   grading	
  
system	
   for	
   use	
   in	
   canine	
   bladder	
   and	
   urethral	
   cancer.	
   17	
   This	
   scheme	
   was	
   based	
   on	
   the	
  
1986	
   version	
   of	
   the	
   World	
   Health	
   Organization	
   classification	
   and	
   grading	
   scheme	
   for	
  
proliferative	
  urothelial	
  in	
  human	
  bladder	
  cancer.39	
  In	
  2006,	
  Patrick,	
  et	
  al.,	
  showed	
  that	
  the	
  
canine	
   urothelial	
   carcinomas	
   could	
   be	
   classified	
   according	
   to	
   the	
   updated	
   World	
   Health	
  
Organization/International	
   Society	
   for	
   Urologic	
   Pathology	
   consensus	
   system	
   used	
   for	
  
classifying	
  and	
  grading	
  proliferative	
  urothelial	
  lesions	
  in	
  humans,	
  as	
  reported	
  in	
  1998	
  and	
  
updated	
  in	
  2004.16,	
  40,	
  41	
  More	
  recently,	
  Knapp	
  et	
  al.,	
  suggested	
  a	
  simplified	
  two-­‐tier	
  version	
  
of	
  grading	
  classifying	
  urothelial	
  carcinomas	
  as	
  either	
  high	
  or	
  low	
  grade.15	
  	
  	
  
While	
   histologic	
   classification	
   and	
   grading	
   likely	
   have	
   relevance,	
   the	
   prognostic	
  
significance	
   is	
   unclear.	
   	
   In	
   application,	
   Valli	
   et	
   al.	
   found	
   that	
   there	
   were	
   significant	
  
correlations	
   with	
   tumor	
   grade	
   and	
   depth	
   of	
   invasion,	
   tumor	
   grade	
   and	
   presence	
   of	
  
metastases,	
  and	
  peritumoral	
  desmoplasia	
  and	
  metastases.17	
  Regarding	
  survival	
  time,	
  Valli,	
  
et	
   al.,	
   also	
   found	
   significant	
   correlations	
   between	
   survival	
   and	
   tumor	
   grade	
   when	
  
comparing	
   grade	
   2	
   and	
   3	
   to	
   grade	
   1	
   tumors,	
   but	
   no	
   significant	
   difference	
   in	
   survival	
  
between	
   grades	
   2	
   and	
   3.17	
   Further,	
   there	
   were	
   also	
   no	
   significant	
   differences	
   between	
  
survival	
  and	
  type	
  of	
  invasion	
  (tentacular	
  or	
  broad)	
  or	
  macroscopic	
  tumor	
  architecture	
  (flat	
  
or	
   papillary).17	
   Two	
   independent	
   studies	
   employing	
   Valli’s	
   system	
   of	
   grading	
   found	
   no	
  

	
   8	
  

	
  
significance	
   between	
   histologic	
   grades	
   of	
   urothelial	
   carcinoma;	
   however,	
   overall	
  
architecture	
  was	
  not	
  specifically	
  examined	
  in	
  either	
  study,	
  analysis	
  in	
  one	
  study	
  compared	
  
a	
   combination	
   of	
   grades	
   1	
   and	
   2	
   to	
   grade	
   3,	
   and	
   the	
   other	
   study	
   only	
   included	
   and	
   thus,	
  
only	
   compared	
   grades	
   2	
   and	
   3.29,	
   35	
   No	
   studies	
   have	
   specifically	
   evaluated	
   the	
   prognostic	
  
significance	
  of	
  histologic	
  classification	
  and	
  grading	
  as	
  described	
  by	
  Patrick,	
  et	
  al.,	
  or	
  Knapp,	
  
et	
  al.	
  
Aside	
  from	
  grading,	
  individual	
  histologic	
  factors	
  have	
  rarely	
  been	
  evaluated	
  with	
  
respect	
  to	
  prognostic	
  significance.	
  	
  The	
  presence	
  of	
  necrosis	
  within	
  urothelial	
  carcinomas	
  
was	
  negatively	
  statistically	
  associated	
  with	
  survival	
  time	
  in	
  one	
  study.33	
  Other	
  studies	
  have	
  
found	
  no	
  significance	
  correlation	
  between	
  survival	
  time	
  and	
  mitotic	
  index	
  or	
  lymphatic	
  
invasion	
  29,	
  33	
  
Few	
  studies	
  have	
  evaluated	
  the	
  prognostic	
  significance	
  of	
  molecular	
  markers	
  with	
  
regards	
  to	
  canine	
  urothelial	
  carcinomas,	
  and	
  the	
  few	
  such	
  studies	
  that	
  have	
  attempted	
  to	
  
correlate	
  molecular	
  markers	
  with	
  prognosis	
  have	
  often	
  not	
  found	
  statistical	
  significance.	
  	
  
The	
  expression	
  of	
  p63	
  is	
  an	
  exception	
  as	
  low	
  p63	
  levels	
  have	
  been	
  associated	
  with	
  vessel	
  
invasion,	
  metastasis,	
  and	
  short	
  survival	
  time.22	
  No	
  statistical	
  significance	
  was	
  detected	
  
between	
  survival	
  time	
  and	
  immunoreactivity	
  for	
  the	
  chemotherapy	
  resistance	
  markers	
  P-­‐
glycoprotein	
  and	
  glutathione-­‐S-­‐transferase	
  π,	
  or	
  factor	
  VIII-­‐related	
  antigen,	
  which	
  was	
  
used	
  as	
  a	
  marker	
  for	
  angiogenesis.29	
  Expression	
  of	
  epidermal	
  growth	
  factor	
  receptor	
  was	
  
not	
  significantly	
  associated	
  with	
  vessel	
  invasion,	
  lymph	
  node	
  metastasis,	
  or	
  survival	
  time	
  in	
  
canine	
  urothelial	
  carcinomas.21	
  	
  
Based	
  on	
  the	
  relatively	
  sparse	
  and	
  disparate	
  results	
  of	
  studies	
  examining	
  prognosis	
  
with	
  regards	
  to	
  demographic,	
  clinical,	
  histopathologic	
  and	
  molecular	
  features,	
  it	
  is	
  clear	
  

	
   9	
  

	
  
that	
  additional	
  studies	
  are	
  needed.	
  	
  Prognostication	
  of	
  cancer,	
  especially	
  in	
  veterinary	
  
medicine,	
  from	
  a	
  diagnostic	
  and	
  pathologic	
  perspective	
  has	
  classically	
  relied	
  on	
  
classification	
  and	
  grading	
  based	
  on	
  histopathologic	
  criteria.	
  	
  More	
  recently,	
  there	
  has	
  been	
  
a	
  paradigm	
  shift	
  to	
  move	
  beyond	
  H&E,	
  and	
  to	
  improve	
  general	
  prognostication	
  and	
  
prediction	
  of	
  treatment	
  response	
  by	
  exploiting	
  the	
  molecular	
  constitution	
  of	
  a	
  given	
  tumor.	
  	
  
Molecular	
  features	
  that	
  could	
  be	
  exploited	
  to	
  better	
  prognosticate	
  canine	
  urothelial	
  
carcinomas	
  likely	
  exist;	
  however,	
  a	
  better	
  understanding	
  of	
  the	
  pathways	
  that	
  drive	
  
carcinogenesis	
  is	
  obviously	
  required.	
  
	
  
	
  

	
  

	
  10	
  

	
  
Outline	
  of	
  studies	
  
Optimization	
  of	
  therapy	
  for	
  canine	
  urothelial	
  carcinomas	
  and	
  acceptance	
  of	
  the	
  dog	
  
as	
   a	
   model	
   for	
   urothelial	
   carcinogenesis	
   is	
   hampered	
   by	
   a	
   lack	
   of	
   understanding	
   of	
   the	
  
inherent	
   features	
   that	
   affect	
   prognosis	
   and	
   knowledge	
   of	
   the	
   molecular	
   constitution	
   of	
  
individual	
   tumors.	
   Numerous	
   studies	
   have	
   examined	
   associations	
   of	
   canine	
   urothelial	
  
carcinomas	
   with	
   risk	
   factors	
   for	
   development	
   and	
   response	
   to	
   specific	
   treatments;	
  
however,	
   relatively	
   few	
   have	
   examined	
   markers	
   of	
   prognostication	
   or	
   molecular	
   features	
  
that	
   underlie	
   carcinogenesis	
   of	
   these	
   cancers.	
   While	
   it	
   is	
   clear	
   that	
   similarities	
   between	
  
lower	
   urinary	
   tract	
   urothelial	
   carcinomas	
   in	
   humans	
   and	
   dogs	
   exist	
   in	
   terms	
   of	
   general	
  
histomorphology	
  and	
  clinical	
  progression,	
  it	
  is	
  less	
  clear	
  whether	
  the	
  molecular	
  pathways	
  
that	
  drive	
  carcinogenesis	
  are	
  the	
  same.	
  An	
  understanding	
  of	
  the	
  pathologic	
  basis	
  for	
  lower	
  
urinary	
   tract	
   cancer	
   in	
   dogs	
   is	
   needed	
   from	
   a	
   prognostic	
   and	
   treatment	
   perspective	
   in	
  
veterinary	
   medicine	
   and	
   to	
   further	
   establish	
   the	
   dog	
   as	
   a	
   relevant	
   model	
   for	
   human	
  
disease.	
  	
  	
  
	
  

	
  
Given	
   these	
   facts	
   the	
   main	
   aims	
   of	
   the	
   studies	
   presented	
   in	
   this	
   dissertation	
   are	
   as	
  

follows:	
  
•

To	
   demonstrate	
   biologic	
   and	
   prognostic	
   significance	
   for	
   histomorphologic	
  
classification	
  and	
  grading	
  of	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas	
  

•

To	
   evaluate	
   components	
   of	
   specific	
   carcinogenesis	
   pathways	
   in	
   canine	
   lower	
  
urinary	
  tract	
  urothelial	
  carcinomas	
  with	
  the	
  goal	
  of	
  identifying	
  molecular	
  prognostic	
  
markers	
  and	
  therapeutic	
  targets	
  

	
  11	
  

	
  
To	
  compare	
  and	
  contrast	
  urothelial	
  carcinogenesis	
  in	
  the	
  lower	
  urinary	
  tract	
  of	
  dogs	
  

•

to	
   that	
   reported	
   in	
   humans	
   with	
   the	
   goal	
   of	
   better	
   establishing	
   the	
   dog	
   as	
   a	
   relevant	
  
model	
  for	
  human	
  disease	
  
	
  
Studies	
  are	
  laid	
  out	
  in	
  four	
  chapters	
  followed	
  by	
  a	
  conclusion	
  that	
  presents	
  a	
  
summary	
  of	
  findings	
  and	
  discussion	
  of	
  limitations	
  and	
  future	
  directions.	
  The	
  background	
  
and	
  goals	
  of	
  each	
  chapter	
  are	
  briefly	
  described	
  here,	
  and	
  expanded	
  upon	
  within	
  the	
  
chapters	
  themselves.	
  	
  
	
  
Chapter	
  2	
  	
  
Chapter	
   2	
   examines	
   the	
   biologic	
   significance	
   of	
   the	
   histologic	
   classification	
   and	
  
grading	
  of	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  with	
  respect	
  to	
  the	
  distribution	
  
of	
   a	
   set	
   of	
   immunohistochemical	
   markers	
   that	
   have	
   been	
   associated	
   with	
   differentiation	
  
and/or	
   prognosis	
   in	
   humans.	
   The	
   central	
   hypothesis	
   for	
   this	
   work	
   was	
   that	
   there	
   are	
  
significant	
   differences	
   in	
   the	
   expression	
   of	
   the	
   evaluated	
   markers	
   between	
   histologic	
  
classifications	
   and	
   grades	
   of	
   proliferative	
   urothelial	
   lesions.	
   	
   The	
   goal	
   in	
   examining	
   this	
  
hypothesis	
   was	
   to	
   provide	
   rationale	
   for	
   use	
   of	
   this	
   classification	
   and	
   grading	
   system	
   in	
  
dogs;	
   however,	
   this	
   study	
   was	
   limited	
   by	
   the	
   fact	
   that	
   follow-­‐up	
   information	
   regarding	
  
outcome	
  was	
  not	
  available.	
  
The	
  World	
  Health	
  Organization	
  (WHO)/International	
  Society	
  of	
  Urologic	
  Pathology	
  
(ISUP)	
   Consensus	
   Classification	
   System	
   published	
   in	
   1998	
   and	
   updated	
   in	
   2004	
   has	
   been	
  
demonstrated	
   to	
   be	
   significantly	
   associated	
   with	
   clinical	
   outcome	
   in	
   humans.42-­‐50	
   While	
  
Patrick,	
   et	
   al.,	
   demonstrated	
   that	
   canine	
   proliferative	
   urothelial	
   lesions	
   could	
   easily	
   be	
  

	
  12	
  

	
  
classified	
   according	
   to	
   the	
   system	
   based	
   on	
   histomorphology,	
   the	
   significance	
   of	
   this	
  
system	
   in	
   terms	
   of	
   prognostication	
   in	
   dogs	
   remains	
   unclear.16	
   To	
   further	
   characterize	
  
differences	
   between	
   histologic	
   classification	
   and	
   grade	
   beyond	
   histomorphology,	
   we	
  
examined	
   the	
   expression	
   of	
   uroplakin	
   III,	
   cytokeratin	
   7,	
   COX-­‐2,	
   and	
   caspase	
   3	
   in	
   a	
   set	
   of	
  
canine	
  proliferative	
  urothelial	
  carcinomas	
  including	
  neoplastic	
  and	
  non-­‐neoplastic	
  lesions.	
  
	
  

Uroplakin	
  III	
  and	
  cytokeratin	
  7	
  are	
  expressed	
  by	
  urothelium	
  and	
  used	
  as	
  

differentiation	
  markers	
  for	
  diagnostic	
  purposes	
  in	
  dogs.51-­‐60	
  In	
  humans,	
  loss	
  of	
  UPIII	
  
expression	
  is	
  frequently	
  seen	
  in	
  metastatic	
  sites	
  of	
  high-­‐grade	
  tumors	
  and	
  is	
  associated	
  
with	
  lymphovascular	
  invasion,	
  stage,	
  and	
  grade.60,	
  61	
  No	
  such	
  associations	
  have	
  previously	
  
been	
  found	
  in	
  dogs;	
  however,	
  studies	
  have	
  not	
  focused	
  on	
  the	
  distribution	
  pattern	
  of	
  
expression	
  as	
  was	
  evaluated	
  in	
  this	
  case	
  series.	
  	
  
	
  

The	
   expression	
   of	
   the	
   inducible	
   enzyme	
   COX-­‐2	
   and	
   the	
   subsequent	
   production	
   of	
  

prostaglandin	
   E2	
   have	
   significant	
   roles	
   in	
   carcinogenesis,	
   including	
   immunosuppression,	
  
inhibition	
   of	
   apoptosis,	
   increasing	
   the	
   metastatic	
   potential	
   of	
   neoplastic	
   epithelial	
   cells,	
  
promoting	
   drug	
   resistance,	
   and	
   stimulating	
   angiogenesis.62-­‐67	
   Numerous	
   studies	
   have	
  
variably	
   shown	
   significant	
   correlations	
   between	
   COX-­‐2	
   expression	
   and	
   tumor	
   grade,	
  
invasion,	
  metastasis,	
  and	
  survival.68-­‐75	
  	
  
Caspase-­‐3	
   is	
   an	
   executioner	
   protein	
   that	
   is	
   activated	
   by	
   both	
   the	
   intrinsic	
   and	
  
extrinsic	
   pathways	
   of	
   apoptosis.76-­‐78	
   Due	
   to	
   the	
   fact	
   that	
   this	
   protein	
   is	
   activated	
   late	
   in	
   the	
  
apoptotic	
   pathway,	
   immunohistochemical	
   detection	
   of	
   activated	
   caspase-­‐3	
   has	
   been	
   used	
  
to	
  evaluate	
  apoptotic	
  rate.79,	
  80	
  In	
  human	
  bladder	
  cancers,	
  expression	
  of	
  caspase-­‐3	
  has	
  been	
  
suggested	
  as	
  to	
  have	
  prognostic	
  significance.81-­‐83	
  	
  
	
  

	
  13	
  

	
  
Chapter	
  3	
  
Chapter	
   3	
   investigates	
   the	
   potential	
   relationship	
   between	
   mechanisms	
   governing	
  
prostaglandin	
  E2	
   regulation	
  and	
  epithelial-­‐to-­‐mesenchymal	
  transition	
  in	
  terms	
  of	
  cadherin	
  
expression	
   and	
   Wnt	
   signaling	
   in	
   canine	
   lower	
   urinary	
   tract	
   urothelial	
   carcinomas.	
   As	
  
canine	
  urothelial	
  carcinomas	
  are	
  often	
  highly	
  invasive	
  and	
  metastasis	
  is	
  common,	
  it	
  is	
  likely	
  
that	
  these	
  tumors	
  often	
  undergo	
  epithelial-­‐to-­‐mesenchymal	
  transition	
  during	
  progression,	
  
gaining	
  the	
  ability	
  to	
  invade.	
  Based	
  on	
  studies	
  in	
  human	
  urothelial	
  carcinomas,	
  it	
  has	
  been	
  
suggested	
   that	
   loss	
   expression	
   of	
   15-­‐hydroxyprostaglandin	
   dehydrogenase	
   (HPGD)	
   is	
  
associated	
  with	
  loss	
  of	
  E-­‐cadherin	
  expression	
  and	
  development	
  of	
  an	
  invasive	
  phenotype.	
  
Our	
   central	
   hypothesis	
   for	
   this	
   study	
   was	
   that	
   altered	
   expression	
   of	
   COX-­‐2	
   and	
   PGDH	
   in	
  
canine	
  urothelial	
  carcinomas	
  is	
  associated	
  with	
  aberrant	
  expression	
  of	
  adhesion-­‐associated	
  
cadherin	
  proteins,	
  higher	
  histologic	
  grade,	
  invasion,	
  and	
  shorter	
  survival	
  times.	
  	
  
Prostaglandins	
   are	
   inflammatory	
   mediators	
   implicated	
   in	
   avoidance	
   of	
   apoptosis,	
  
angiogenesis,	
   cellular	
   proliferation,	
   invasion	
   and	
   metastasis	
   in	
   cancer.66,	
   84-­‐88	
   COX-­‐2	
   is	
   an	
  
inducible	
  enzyme	
  that	
  is	
  up	
  regulated	
  in	
  a	
  variety	
  of	
  inflammatory	
  and	
  neoplastic	
  processes	
  
and	
   that	
   acts	
   in	
   the	
   production	
   of	
   prostaglandins.	
   Increased	
   expression	
   of	
   COX-­‐2	
   by	
  
neoplastic	
   cells	
   has	
   been	
   associated	
   with	
   worsening	
   prognosis.66,	
   89	
   As	
   numerous	
   COX	
  
inhibitors	
   are	
   available,	
   COX-­‐2	
   is	
   a	
   prime	
   therapeutic	
   target	
   in	
   many	
   cancers.	
  
Unfortunately,	
   COX-­‐2	
   inhibition	
   has	
   been	
   associated	
   with	
   side	
   effects	
   including	
   life-­‐
threatening	
  cardiovascular	
  effects,	
  best	
  recognized	
  in	
  association	
  with	
  the	
  infamous	
  drug,	
  
Vioxx.90	
  HPGD,	
  on	
  the	
  other	
  hand,	
  inactivates	
  prostaglandin	
  E2.10	
  Decreased	
  expression	
  of	
  
HPGD	
  results	
  in	
  increased	
  amounts	
  of	
  prostaglandin	
  within	
  tissues	
  and	
  has	
  been	
  associated	
  
with	
   worsening	
   prognosis	
   in	
   cancers.91,	
   92	
   Further,	
   a	
   recent	
   study	
   showed	
   specific	
   single-­‐

	
  14	
  

	
  
nucleotide	
  polymorphisms	
  within	
  the	
  HPGD	
  gene	
  were	
  associated	
  with	
  both	
  increased	
  risk	
  
of	
   colon	
   cancer	
   development	
   and	
   decreased	
   expression	
   of	
   HPGD	
   in	
   the	
   colon	
   overall.93	
  
Developing	
   a	
   better	
   view	
   of	
   the	
   prostaglandin	
   pathways,	
   the	
   consequences	
   of	
   altered	
  
expression,	
   and	
   how	
   this	
   impacts	
   treatment	
   response	
   could	
   lead	
   to	
   the	
   development	
   of	
  
better	
  clinical	
  strategies	
  to	
  address	
  canine	
  urothelial	
  cancer.	
  	
  
The	
   cadherins	
   are	
   a	
   group	
   of	
   membrane-­‐associated	
   molecules	
   involved	
   in	
   cell-­‐cell	
  
adhesion.	
   As	
   such,	
   these	
   proteins	
   play	
   integral	
   roles	
   in	
   embryogenesis	
   and	
   development,	
  
cellular	
   polarity,	
   and	
   carcinogenesis.94	
   E-­‐cadherin	
   is	
   a	
   primary	
   mediator	
   of	
   cell-­‐cell	
  
adhesion	
  in	
  epithelial	
  cells	
  and	
  is	
  strongly	
   expressed	
  in	
  the	
  urothelium	
  of	
  the	
  bladder.	
  N-­‐
cadherin	
   is	
   predominately	
   expressed	
   by	
   mesenchymal	
   cells	
   in	
   adults,	
   but	
   is	
   expressed	
   by	
  
epithelial	
   cells	
   during	
   embryogenesis.	
   P-­‐cadherin	
   is	
   most	
   prominently	
   expressed	
   in	
   the	
  
placenta,	
  but	
  is	
  also	
  expressed	
  by	
  basal	
  cells	
  of	
  stratified	
  epithelia.	
  Decreased	
  expression	
  of	
  
E-­‐cadherin	
  and	
  increased	
  expression	
  of	
  other	
  cadherins	
  (termed	
  cadherin	
  switching)	
  result	
  
in	
   decreased	
   strength	
   of	
   cell-­‐cell	
   adhesions	
   and	
   an	
   increased	
   propensity	
   for	
   cellular	
  
migration.94-­‐96	
   In	
   terms	
   of	
   carcinogenesis,	
   cadherin	
   switching	
   is	
   observed	
   in	
   epithelial	
   to	
  
mesenchymal	
   transitions	
   and	
   has	
   been	
   associated	
   with	
   increased	
   invasiveness	
   and	
  
metastasis.	
  94-­‐97	
  	
  	
  
Because	
   cadherins	
   form	
   associations	
   with	
   other	
   proteins	
   within	
   cells,	
   they	
   are	
  
integrally	
   linked	
   to	
   intracellular	
   signaling	
   and	
   trafficking.	
   β-­‐catenin	
   binds	
   to	
   the	
  
cytoplasmic	
  tail	
  of	
  cadherin	
  molecules.	
  In	
  this	
  capacity	
  it	
  functions	
  in	
  cell-­‐cell	
  adhesion,	
  but	
  
it	
   is	
   also	
   involved	
   in	
   a	
   variety	
   of	
   signaling	
   pathways	
   including	
   some	
   that	
   are	
   involved	
   in	
  
carcinogenesis,	
  such	
  as	
  the	
  Wnt	
  pathways.95,	
   98,	
   99	
  While	
  loss	
  of	
  cadherin	
  expression	
  alone	
  
has	
   not	
   been	
   shown	
   to	
   directly	
   result	
   in	
   intracellular	
   signaling,	
   altered	
   expression	
   of	
  

	
  15	
  

	
  
cadherins	
  has	
  been	
  shown	
  to	
  amplify	
  or	
  buffer	
  the	
  effects	
  of	
  such	
  pathways.95,	
  98,	
  99	
  This	
  is	
  
thought	
   to	
   be	
   due	
   to	
   loss	
   of	
   the	
   association	
   and	
   sequestration	
   of	
   β-­‐catenin	
   in	
   cadherin	
  
junctional	
  complexes	
  as	
  cadherin	
  expression	
  is	
  lost.	
  	
  
It	
  has	
  recently	
  been	
  shown	
  that	
  cadherin	
  expression	
  is	
  integrally	
  linked	
  to	
  HPGD.	
  HPGD	
  
expression	
   was	
   shown	
   to	
   increase	
   in	
   expression	
   with	
   urothelial	
   differentiation	
   and	
  
inhibition	
   of	
   PGDH	
   expression	
   resulted	
   in	
   disruption	
   of	
   E-­‐cadherin	
   expression	
   at	
   cell-­‐cell	
  
junction	
   in	
   cell	
   lines.100	
   In	
   non-­‐small	
   cell	
   lung	
   cancer	
   of	
   humans,	
   it	
   was	
   shown	
   that	
  
exogenous	
   prostaglandin	
   can	
   decrease	
   E-­‐cadherin	
   expression,	
   and	
   that	
   such	
   changes	
   are	
  
mediated	
   by	
   the	
   specific	
   transition	
   repressors,	
   ZEB1	
   and	
   Snail.101	
   In	
   squamous	
   cell	
  
carcinomas,	
  administration	
  of	
  prostaglandin	
  E2	
  or	
  prostaglandin	
  receptor	
  agonists	
  lead	
  to	
  
decreased	
   expression	
   of	
   E-­‐cadherin	
   and	
   internalization	
   of	
   this	
   molecule	
   into	
   the	
  
cytoplasm.102	
  	
  
Changes	
   in	
   the	
   expression	
   of	
   all	
   of	
   the	
   above	
   mentioned	
   molecules	
   have	
   been	
   reported	
  
in	
   urothelial	
   carcinomas	
   and	
   many	
   are	
   targets	
   for	
   treatment	
   making	
   them	
   of	
   particular	
  
interest	
   for	
   further	
   study.	
   COX-­‐2	
   over	
   expression	
   has	
   been	
   described	
   in	
   a	
   significant	
  
number	
   of	
   urothelial	
   carcinomas	
   of	
   humans	
   and	
   increasing	
   expression	
   is	
   associated	
   with	
  
invasiveness,	
   metastasis,	
   and	
   increased	
   mortality.84,	
  

89	
  

Previous	
   studies	
   of	
   COX-­‐2	
  

expression	
   have	
   yielded	
   similar	
   results	
   in	
   dogs.86	
   Treatment	
   studies	
   with	
   COX-­‐2	
   inhibitors,	
  
however,	
   have	
   met	
   mixed	
   results.	
   In	
   vitro	
   studies	
   have	
   suggested	
   that	
   treatment	
   of	
  
urothelial	
  carcinoma	
  cell	
  lines	
  with	
  such	
  compounds	
  can	
  result	
  in	
  decreased	
  invasiveness	
  
and	
   decreased	
   tumor	
   grade.103	
   Results	
   of	
   in	
   vivo	
   studies	
   vary	
   with	
   only	
   a	
   proportion	
   of	
  
studies	
  showing	
  a	
  decreased	
  risk	
  for	
  tumor	
  development.87,	
  103-­‐105	
  PGDH	
  has	
  been	
  shown	
  to	
  
be	
  important	
  in	
  urothelial	
  differentiation	
  and	
  its	
  expression	
  is	
  decreased	
  in	
  malignancies.	
  

	
  16	
  

	
  
E-­‐cadherin	
  and	
  B-­‐catenin	
  expression	
  is	
  often	
  down	
  regulated	
  in	
  urothelial	
  carcinomas.106,	
  
107	
  

Such	
   down	
   regulation	
   is	
   associated	
   with	
   higher	
   degrees	
   of	
   neoplastic	
   infiltration.107	
  

Mutations	
   in	
   E-­‐cadherin	
   genes	
   are	
   rarely	
   implicated	
   in	
   carcinogenesis.98	
   Rather,	
   silencing	
  
of	
  E-­‐cadherin	
  expression	
  is	
  most	
  often	
  an	
  epigenetic	
  change.	
  As	
  such,	
  numerous	
  classes	
  of	
  
drugs	
  have	
  been	
  suggested	
  to	
  up	
  regulate	
  E-­‐cadherin	
  expression.98	
  N-­‐cadherin	
  expression	
  
has	
   been	
   shown	
   to	
   be	
   up	
   regulated	
   in	
   urothelial	
   carcinomas	
   and	
   in	
   bladder	
   cancer	
   cell	
  
lines.85	
  The	
  compound	
  (-­‐)-­‐epigallocatechin-­‐3-­‐gallate,	
  which	
  is	
  found	
  in	
  green	
  tea,	
  has	
  been	
  
associated	
  with	
  down	
  regulation	
  of	
  N-­‐cadherin	
  and	
  decreased	
  migration	
  in	
  bladder	
  cancer	
  
cell	
  lines.85	
  Increased	
  P-­‐cadherin	
  expression	
  has	
  been	
  associated	
  with	
  a	
  significantly	
  worse	
  
bladder	
   cancer-­‐specific	
   survival	
   and	
   a	
   more	
   malignant	
   and	
   invasive	
   cancer	
   phenotype	
   in	
  
humans.106	
  	
  
	
  
Chapter	
  4	
  
Chapter	
   4	
   examines	
   the	
   potential	
   role	
   of	
   DNA	
   mismatch	
   repair	
   (MMR)	
   in	
  
carcinogenesis	
   of	
   canine	
   lower	
   urinary	
   tract	
   urothelial	
   carcinomas	
   through	
   evaluation	
   of	
  
microsatellite	
   instability	
   and	
   expression	
   of	
   MMR	
   proteins.	
   Hereditary	
   deficiency	
   of	
   MMR	
   in	
  
humans	
   is	
   associated	
   with	
   syndromes	
   of	
   cancer	
   development	
   throughout	
   the	
   body,	
   but	
  
MMR	
   is	
   often	
   defective	
   in	
   sporadic	
   cancers	
   as	
   well.	
   	
   Based	
   on	
   the	
   fact	
   that	
   MMR	
   repair	
  
deficiency	
   is	
   hereditary	
   in	
   humans	
   and	
   that	
   there	
   are	
   strong	
   breed	
   predispositions	
   for	
  
cancer	
  in	
  dogs,	
  we	
  hypothesized	
  that	
  deficiencies	
  in	
  MMR	
  contributed	
  to	
  carcinogenesis	
  of	
  
hereditary	
  cancers	
  in	
  dogs	
  including	
  urothelial	
  carcinomas.	
  To	
  evaluate	
  MMR	
  in	
  dogs,	
  we	
  
examined	
   the	
   prevalence	
   of	
   microsatellite	
   instability	
   (MSI)	
   in	
   a	
   set	
   of	
   cancer	
   types	
   that	
  
have	
   breed	
   predispositions	
   including	
   urothelial	
   carcinomas.	
   After	
   demonstrating	
   MSI	
   in	
  

	
  17	
  

	
  
urothelial	
  carcinomas,	
  we	
  subsequently	
  evaluated	
  MMR	
  protein	
  expression.	
  
The	
   DNA	
   mismatch	
   repair	
   (MMR)	
   system	
   participates	
   in	
   a	
   variety	
   of	
   cellular	
  
processes.	
   Most	
   notably,	
   this	
   system	
   is	
   responsible	
   for	
   post	
   replication	
   recognition	
   and	
  
repair	
  of	
  base-­‐base	
  mismatches	
  and	
  the	
  resolution	
  of	
  insertion	
  and	
  deletion	
  loops	
  that	
  can	
  
occur	
  in	
  repetitive	
  regions	
  of	
  DNA.108-­‐112	
  Accordingly,	
  deficiencies	
  in	
  this	
  system	
  can	
  lead	
  to	
  
increased	
   rates	
   of	
   point	
   and	
   frame	
   shift	
   mutations	
   throughout	
   the	
   genome.	
   Such	
   mutations	
  
can	
   lead	
   to	
   loss	
   of	
   function	
   changes	
   in	
   tumor	
   suppressors	
   or	
   gain	
   of	
   function	
   changes	
   in	
  
tumor	
  oncogenes.113-­‐117	
  In	
  addition,	
  the	
  MMR	
  system	
  also	
  participates	
  in	
  a	
  variety	
  of	
  other	
  
processes	
   including	
   DNA	
   damage	
   recognition	
   signaling,	
   promoting	
   cell	
   cycle	
   arrest	
   and	
  
apoptosis,	
   homologous	
   recombination,	
   meiotic	
   recombination,	
   and	
   other	
   DNA	
   repair	
  
pathways.108,	
   109,	
   111,	
   118	
   Given	
   the	
   varied	
   roles	
   played	
   by	
   MMR	
   in	
   maintenance	
   of	
   DNA	
  
integrity	
  and	
  signaling,	
  it	
  is	
  not	
  surprising	
  that	
  defects	
  in	
  MMR	
  facilitate	
  carcinogenesis	
  and	
  
affect	
  response	
  to	
  treatment.	
  	
  
MMR	
   function	
   has	
   most	
   often	
   been	
   evaluated	
   through	
   analysis	
   of	
   microsatellites.	
  
Microsatellites	
   are	
   regions	
   of	
   nucleotide	
   repeats	
   located	
   throughout	
   the	
   genome.	
   These	
  
areas	
   are	
   prone	
   to	
   polymerase	
   slippage	
   during	
   replication	
   leading	
   to	
   formation	
   of	
   small	
  
insertion	
   and	
   deletion	
   loops.119	
   If	
   not	
   recognized	
   and	
   repaired	
   by	
   MMR,	
   buildup	
   of	
   frame	
  
shift	
   mutations	
   occurs	
   within	
   microsatellites.112,	
   118	
   In	
   cancers	
   with	
   defects	
   in	
   the	
   MMR	
  
system,	
  a	
  high	
  percentage	
  of	
  microsatellites	
  have	
  recognizable	
  mutations.	
  Such	
  buildup	
  of	
  
mutations	
   within	
   microsatellites	
   is	
   termed	
   microsatellite	
   instability	
   (MSI)	
   and	
   is	
  
considered	
  a	
  “signature”	
  for	
  MMR	
  dysfuntion.118,	
  120	
  	
  
In	
  terms	
  of	
  prognosis,	
  MMR	
  deficient	
  cancers	
  are	
  generally	
  associated	
  with	
  a	
  more	
  
favorable	
   clinical	
   outcome	
   than	
   those	
   that	
   are	
   MMR	
   proficient.121-­‐125	
   In	
   colorectal	
  

	
  18	
  

	
  
carcinomas,	
  for	
  example,	
  MMR	
  deficient	
  tumors	
  are	
  associated	
  with	
  longer	
  survival	
  times	
  
and	
   a	
   decreased	
   risk	
   of	
   metastasis	
   compared	
   to	
   those	
   that	
   are	
   proficient.124	
   The	
   reason	
   for	
  
this	
  difference	
  in	
  prognosis	
  is	
  likely	
  multifactorial.	
  Inherently,	
  cancers	
  with	
  defects	
  in	
  MMR	
  
are	
  genetically	
  less	
  stable	
  that	
  those	
  with	
  intact	
  MMR	
  leading	
  to	
  an	
  increase	
  in	
  DNA	
  lesions,	
  
which	
   promotes	
   cell	
   cycle	
   arrest	
   and	
   apoptosis	
   signaling.126	
   Also,	
   there	
   is	
   evidence	
   that	
  
MMR	
   deficient	
   cancer	
   cells	
   are	
   often	
   highly	
   immunogenic.	
   This	
   is	
   proposed	
   to	
   occur	
   due	
   to	
  
production	
  of	
  atypical	
  proteins	
  generated	
  through	
  frameshift	
  mutations	
  resulting	
  in	
  T	
  cell	
  
mediated	
  immune	
  responses	
  directed	
  against	
  the	
  cancer	
  cells.127	
  Additionally,	
  it	
  has	
  been	
  
shown	
  that	
  several	
  genes	
  associated	
  with	
  antitumor	
  immune	
  responses	
  are	
  over	
  expressed	
  
in	
  MMR	
  deficient	
  cancers	
  and	
  cell	
  lines.126	
  
Urinary	
  carcinomas	
  of	
  the	
  urinary	
  bladder	
  represent	
  one	
  of	
  the	
  many	
  cancer	
  types	
  
in	
   which	
   a	
   significant	
   percentage	
   of	
   tumors	
   have	
   been	
   reported	
   to	
   show	
   defects	
   in	
   MMR	
   in	
  
humans.	
  Development	
  of	
  urothelial	
  carcinomas	
  of	
  the	
  urinary	
  bladder	
  has	
  been	
  associated	
  
with	
   both	
   hereditary	
   and	
   spontaneously	
   developing	
   MMR	
   deficiency.	
   High	
   MSI	
   and	
   a	
  
particular	
   form	
   of	
   MSI	
   known	
   as	
   elevated	
   microsatellite	
   instability	
   at	
   selected	
  
tetranucleotide	
   repeats	
   (EMAST)	
   have	
   been	
   described	
   in	
   urothelial	
   carcinomas.121	
  
Differential	
   expression	
   of	
   MMR	
   proteins	
   including	
   MSH2,	
   MSH3,	
   and	
   MLH1	
   has	
   been	
  
associated	
  with	
  urothelial	
  carcinoma	
  grade	
  and	
  clinical	
  outcome	
  in	
  humans.121,	
  125,	
  128	
  	
  
	
  
Chapter	
  5	
  
	
  

Chapter	
   5	
   describes	
   the	
   characterization	
   of	
   canine	
   lower	
   urinary	
   tract	
   urothelial	
  

carcinoma	
   cell	
   lines	
   and	
   the	
   evaluation	
   of	
   their	
   MMR	
   capacity.	
   	
   Further,	
   chapter	
   4	
   also	
  
describes	
  the	
  assessment	
  of	
  sensitivity	
  to	
  chemotherapeutic	
  therapy	
  in	
  vitro	
  as	
  influenced	
  

	
  19	
  

	
  
by	
  variable	
  MMR	
  proficiency.	
  	
  
	
  

The	
   potential	
   application	
   of	
   MMR	
   as	
   a	
   therapeutic	
   target	
   is	
   highlighted	
   by	
   the	
  

differences	
   observed	
   in	
   prognosis	
   and	
   the	
   response	
   to	
   treatment	
   between	
   cancers	
   of	
   the	
  
same	
  type	
  that	
  differ	
  in	
  MMR	
  capacity.	
  Having	
  shown	
  in	
  chapter	
  3	
  that	
  MSI	
  and	
  decreased	
  
expression	
  of	
  the	
  MMR	
  protein,	
  MSH2,	
  is	
  common	
  in	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  
carcinomas,	
   we	
   hypothesized	
   that	
   variance	
   in	
   MMR	
   would	
   be	
   reflected	
   by	
   differences	
   in	
  
response	
   to	
   chemotherapeutics.	
   To	
   investigate	
   this	
   hypothesis,	
   four	
   canine	
   urothelial	
  
carcinoma	
  cell	
  lines	
  were	
  established	
  and	
  characterized	
  with	
  the	
  goal	
  of	
  generating	
  in	
  vitro	
  
canine	
   urothelial	
   carcinoma	
   models.	
   We	
   were	
   able	
   to	
   show	
   that	
   one	
   such	
   cell	
   line	
   had	
  
decreased	
   relative	
   expression	
   of	
   MMR	
   proteins	
   and	
   genes,	
   and	
   considered	
   this	
   line	
   MMR	
  
deficient.	
   	
   Based	
   on	
   this,	
   we	
   preformed	
   survival	
   assays	
   exposing	
   MMR	
   proficient	
   and	
  
deficient	
   cell	
   lines	
   to	
   a	
   panel	
   of	
   chemotherapeutics	
   to	
   show	
   that	
   there	
   are	
   differences	
   in	
  
response	
  to	
  treatment,	
  thus	
  highlighting	
  MMR	
  proficiency	
  as	
  a	
  target	
  for	
  selective	
  therapy.	
  
The	
   effects	
   of	
   MMR	
   deficiency	
   on	
   chemotherapeutic	
   response	
   are	
   complex.	
   MMR	
  
deficiency	
   is	
   capable	
   of	
   conferring	
   either	
   drug	
   resistance	
   or	
   sensitivity	
   according	
   to	
   the	
  
drugs	
   mechanism	
   of	
   action.123,	
  

129-­‐137	
  

For	
   example,	
   MMR	
   deficiency	
   results	
   in	
   drug	
  

resistance	
   to	
   the	
   fluorinated	
   pyrimidine	
   analog	
   5-­‐FU	
   as	
   well	
   as	
   certain	
   alkylating	
   agents	
  
including	
  the	
  SN1	
  methylators,	
  temozolomide	
  and	
  dacarbazine.123,	
  129,	
  134,	
  135,	
  137	
  In	
  contrast,	
  
MMR	
  deficient	
  cells	
  are	
  highly	
  sensitive	
  to	
  many	
  of	
  the	
  interstrand	
  cross-­‐linking	
  alkylators,	
  
including	
   CCNU	
   and	
   mitomycin	
   C.129,	
   137	
   There	
   have	
   also	
   been	
   reports	
   of	
   differences	
   in	
  
response	
   of	
   cancers	
   with	
   varying	
   MMR	
   capacity	
   to	
   some	
   drugs	
   within	
   the	
   same	
   class.	
  
Resistance	
  has	
  been	
  reported	
  to	
  platinum	
  containing	
  compounds	
  cisplatin	
  and	
  carboplatin	
  
in	
  MMR	
  deficient	
  cancers,	
  while	
  no	
  such	
  resistance	
  has	
  been	
  reported	
  to	
  oxaliplatin.123,	
  134	
  	
  

	
  20	
  

	
  
The	
   differential	
   effect	
   of	
   the	
   MMR	
   system	
   in	
   response	
   to	
   drugs	
   probably	
   reflects	
   the	
  
ability	
   of	
   the	
   MMR	
   machinery	
   to	
   participate	
   in	
   a	
   number	
   of	
   alternative	
   DNA	
   damage	
  
processing/signaling	
   pathways.	
   In	
   the	
   case	
   of	
   simple	
   methylated	
   bases	
   in	
   the	
   DNA	
  
molecule,	
  it	
  has	
  been	
  proposed	
  that	
  MMR	
  may	
  be	
  involved	
  in	
  triggering	
  apoptosis	
  through	
  
either	
  futile	
  repair	
  attempts	
  or	
  through	
  conversion	
  of	
  the	
  alkylated	
  base	
  to	
  a	
  lethal	
  lesion	
  
such	
  as	
  a	
  double	
  strand	
  break.137	
  In	
  contrast,	
  an	
  increased	
  sensitivity	
  to	
  cytotoxins	
  can	
  be	
  
observed	
   if	
   defective	
   MMR	
   results	
   in	
   failure	
   to	
   repair	
   certain	
   types	
   of	
   DNA	
   lesions.	
   For	
  
example,	
   certain	
   interstrand	
   cross-­‐links	
   are	
   recognized	
   by	
   the	
   MMR	
   system	
   and	
   repair	
   is	
  
thought	
   to	
   occur	
   through	
   MMR	
   mediated	
   homologous	
   recombination.129	
   Cells	
   deficient	
   in	
  
MMR	
  do	
  not	
  effectively	
  repair	
  these	
  cross-­‐links	
  and	
  are	
  extremely	
  sensitive	
  to	
  agents	
  that	
  
induce	
  such	
  these	
  lesions.	
  	
  
Additionally,	
  there	
  has	
  been	
  a	
  great	
  deal	
  of	
  recent	
  attention	
  to	
  the	
  targeting	
  of	
  DNA	
  
repair	
   defective	
   tumor	
   cells	
   through	
   inhibition	
   of	
   coordinating	
   DNA	
   repair	
   systems.	
   Such	
  
an	
  approach	
  has	
  been	
  termed	
  synthetic	
  lethality.133	
  The	
  basis	
  for	
  this	
  paradigm	
  is	
  that	
  due	
  
to	
   molecular	
   redundancies,	
   cells	
   may	
   tolerate	
   loss	
   of	
   function	
   of	
   a	
   single	
   pathway	
   that	
  
participates	
  in	
  DNA	
  repair.	
  However,	
  when	
  another	
  coordinating	
  pathway	
  is	
  inhibited,	
  the	
  
result	
  is	
  cell	
  death.	
  New	
  evidence	
  suggests	
  that	
  inhibition	
  of	
  particular	
  DNA	
  polymerases	
  is	
  
synthetically	
  lethal	
  in	
  cells	
  that	
  have	
  MMR	
  defects.133	
  	
  Also,	
  drugs	
  that	
  are	
  known	
  to	
  cause	
  
oxidative	
  damage	
  such	
  as	
  methotrexate	
  have	
  been	
  suggested	
  to	
  have	
  a	
  potential	
  synthetic	
  
lethal	
  relationship	
  with	
  deficiencies	
  in	
  MMR.123	
  
	
  

	
  

	
  21	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
REFERENCES	
  
	
  
	
  

	
  

	
  22	
  

	
  
REFERENCES	
  
	
  
1.	
  SEER	
  stat	
  fact	
  sheets:	
  bladder	
  cancer.	
  Available	
  from	
  URL:	
  
http://seer.cancer.gov/statfacts/html/urinb.html	
  [accessed	
  Nov	
  5,	
  2014].	
  
2.	
  Jemal	
  A,	
  Siegel	
  R,	
  Ward	
  E,	
  et	
  al.	
  Cancer	
  statistics,	
  2008.	
  CA	
  Cancer	
  J	
  Clin.	
  2008;58:	
  71-­‐96.	
  
3.	
  Sugano	
  K,	
  Kakizoe	
  T.	
  Genetic	
  alterations	
  in	
  bladder	
  cancer	
  and	
  their	
  clinical	
  applications	
  
in	
  molecular	
  tumor	
  staging.	
  Nat	
  Clin	
  Pract	
  Urol.	
  2006;3:	
  642-­‐652.	
  
4.	
  Herr	
  HW,	
  Dotan	
  Z,	
  Donat	
  SM,	
  Bajorin	
  DF.	
  Defining	
  optimal	
  therapy	
  for	
  muscle	
  invasive	
  
bladder	
  cancer.	
  J	
  Urol.	
  2007;177:	
  437-­‐443.	
  
5.	
  Riley	
  GF,	
  Potosky	
  AL,	
  Lubitz	
  JD,	
  Kessler	
  LG.	
  Medicare	
  payments	
  from	
  diagnosis	
  to	
  death	
  
for	
  elderly	
  cancer	
  patients	
  by	
  stage	
  at	
  diagnosis.	
  Med	
  Care.	
  1995;33:	
  828-­‐841.	
  
6.	
  Kummar	
  S,	
  Kinders	
  R,	
  Rubinstein	
  L,	
  et	
  al.	
  Compressing	
  drug	
  development	
  timelines	
  in	
  
oncology	
  using	
  phase	
  '0'	
  trials.	
  Nat	
  Rev	
  Cancer.	
  2007;7:	
  131-­‐139.	
  
7.	
  Cekanova	
  M,	
  Rathore	
  K.	
  Animal	
  models	
  and	
  therapeutic	
  molecular	
  targets	
  of	
  cancer:	
  
utility	
  and	
  limitations.	
  Drug	
  Des	
  Devel	
  Ther.	
  2014;8:	
  1911-­‐1922.	
  
8.	
  Mak	
  IW,	
  Evaniew	
  N,	
  Ghert	
  M.	
  Lost	
  in	
  translation:	
  animal	
  models	
  and	
  clinical	
  trials	
  in	
  
cancer	
  treatment.	
  Am	
  J	
  Transl	
  Res.	
  2014;6:	
  114-­‐118.	
  
9.	
  Knapp	
  DW,	
  Waters	
  DJ.	
  Naturally	
  occurring	
  cancer	
  in	
  pet	
  dogs:	
  important	
  models	
  for	
  
developing	
  improved	
  cancer	
  therapy	
  for	
  humans.	
  Mol	
  Med	
  Today.	
  1997;3:	
  8-­‐11.	
  
10.	
  Knapp	
  DW,	
  Glickman	
  NW,	
  Denicola	
  DB,	
  Bonney	
  PL,	
  Lin	
  TL,	
  Glickman	
  LT.	
  Naturally-­‐
occurring	
  canine	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder	
  A	
  relevant	
  model	
  of	
  
human	
  invasive	
  bladder	
  cancer.	
  Urol	
  Oncol.	
  2000;5:	
  47-­‐59.	
  
11.	
  Lindblad-­‐Toh	
  K,	
  Wade	
  CM,	
  Mikkelsen	
  TS,	
  et	
  al.	
  Genome	
  sequence,	
  comparative	
  analysis	
  
and	
  haplotype	
  structure	
  of	
  the	
  domestic	
  dog.	
  Nature.	
  2005;438:	
  803-­‐819.	
  
12.	
  Mutsaers	
  AJ,	
  Widmer	
  WR,	
  Knapp	
  DW.	
  Canine	
  transitional	
  cell	
  carcinoma.	
  J	
  Vet	
  Intern	
  
Med.	
  2003;17:	
  136-­‐144.	
  
13.	
  Knapp	
  DW,	
  Glickman	
  NW,	
  Denicola	
  DB,	
  Bonney	
  PL,	
  Lin	
  TL,	
  Glickman	
  LT.	
  Naturally-­‐
occurring	
  canine	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder	
  A	
  relevant	
  model	
  of	
  
human	
  invasive	
  bladder	
  cancer.	
  Urologic	
  oncology.	
  2000;5:	
  47-­‐59.	
  
14.	
  Mutsaers	
  AJ,	
  Widmer	
  WR,	
  Knapp	
  DW.	
  Canine	
  transitional	
  cell	
  carcinoma.	
  Journal	
  of	
  
veterinary	
  internal	
  medicine	
  /	
  American	
  College	
  of	
  Veterinary	
  Internal	
  Medicine.	
  2003;17:	
  
136-­‐144.	
  

	
  23	
  

	
  
15.	
  Knapp	
  DW,	
  Ramos-­‐Vara	
  JA,	
  Moore	
  GE,	
  Dhawan	
  D,	
  Bonney	
  PL,	
  Young	
  KE.	
  Urinary	
  
bladder	
  cancer	
  in	
  dogs,	
  a	
  naturally	
  occurring	
  model	
  for	
  cancer	
  biology	
  and	
  drug	
  
development.	
  ILAR	
  J.	
  2014;55:	
  100-­‐118.	
  
16.	
  Patrick	
  DJ,	
  Fitzgerald	
  SD,	
  Sesterhenn	
  IA,	
  Davis	
  CJ,	
  Kiupel	
  M.	
  Classification	
  of	
  canine	
  
urinary	
  bladder	
  urothelial	
  tumours	
  based	
  on	
  the	
  World	
  Health	
  Organization/International	
  
Society	
  of	
  Urological	
  Pathology	
  consensus	
  classification.	
  J	
  Comp	
  Pathol.	
  2006;135:	
  190-­‐199.	
  
17.	
  Valli	
  VE,	
  Norris	
  A,	
  Jacobs	
  RM,	
  et	
  al.	
  Pathology	
  of	
  canine	
  bladder	
  and	
  urethral	
  cancer	
  and	
  
correlation	
  with	
  tumour	
  progression	
  and	
  survival.	
  J	
  Comp	
  Pathol.	
  1995;113:	
  113-­‐130.	
  
18.	
  Knottenbelt	
  C,	
  Mellor	
  D,	
  Nixon	
  C,	
  Thompson	
  H,	
  Argyle	
  DJ.	
  Cohort	
  study	
  of	
  COX-­‐1	
  and	
  
COX-­‐2	
  expression	
  in	
  canine	
  rectal	
  and	
  bladder	
  tumours.	
  J	
  Small	
  Anim	
  Pract.	
  2006;47:	
  196-­‐
200.	
  
19.	
  Rankin	
  WV,	
  Henry	
  CJ,	
  Turnquist	
  SE,	
  et	
  al.	
  Comparison	
  of	
  distributions	
  of	
  survivin	
  among	
  
tissues	
  from	
  urinary	
  bladders	
  of	
  dogs	
  with	
  cystitis,	
  transitional	
  cell	
  carcinoma,	
  or	
  
histologically	
  normal	
  urinary	
  bladders.	
  Am	
  J	
  Vet	
  Res.	
  2008;69:	
  1073-­‐1078.	
  
20.	
  Rankin	
  WV,	
  Henry	
  CJ,	
  Turnquist	
  SE,	
  et	
  al.	
  Identification	
  of	
  survivin,	
  an	
  inhibitor	
  of	
  
apoptosis,	
  in	
  canine	
  urinary	
  bladder	
  transitional	
  cell	
  carcinoma.	
  Vet	
  Comp	
  Oncol.	
  2008;6:	
  
141-­‐150.	
  
21.	
  Hanazono	
  K,	
  Fukumoto	
  S,	
  Kawamura	
  Y,	
  et	
  al.	
  Epidermal	
  Growth	
  Factor	
  Receptor	
  
Expression	
  in	
  Canine	
  Transitional	
  Cell	
  Carcinoma.	
  J	
  Vet	
  Med	
  Sci.	
  2014.	
  
22.	
  Hanazono	
  K,	
  Nishimori	
  T,	
  Fukumoto	
  S,	
  et	
  al.	
  Immunohistochemical	
  expression	
  of	
  p63,	
  
Ki67	
  and	
  beta-­‐catenin	
  in	
  canine	
  transitional	
  cell	
  carcinoma	
  and	
  polypoid	
  cystitis	
  of	
  the	
  
urinary	
  bladder.	
  Vet	
  Comp	
  Oncol.	
  2014.	
  
23.	
  Tanaka	
  ST,	
  Ishii	
  K,	
  Demarco	
  RT,	
  Pope	
  JCt,	
  Brock	
  JW,	
  3rd,	
  Hayward	
  SW.	
  Endodermal	
  
origin	
  of	
  bladder	
  trigone	
  inferred	
  from	
  mesenchymal-­‐epithelial	
  interaction.	
  J	
  Urol.	
  
2010;183:	
  386-­‐391.	
  
24.	
  Viana	
  R,	
  Batourina	
  E,	
  Huang	
  H,	
  et	
  al.	
  The	
  development	
  of	
  the	
  bladder	
  trigone,	
  the	
  center	
  
of	
  the	
  anti-­‐reflux	
  mechanism.	
  Development.	
  2007;134:	
  3763-­‐3769.	
  
25.	
  van	
  der	
  Meijden	
  AP.	
  Bladder	
  cancer.	
  BMJ.	
  1998;317:	
  1366-­‐1369.	
  
26.	
  Meuten	
  D.	
  Tumors	
  of	
  the	
  Urinary	
  System.	
  Tumors	
  in	
  Domestic	
  Animals	
  4th	
  ed.	
  Ames,	
  
IA:	
  Iowa	
  State	
  Press,	
  2002:509-­‐546.	
  
27.	
  Henry	
  CJ,	
  McCaw	
  DL,	
  Turnquist	
  SE,	
  et	
  al.	
  Clinical	
  evaluation	
  of	
  mitoxantrone	
  and	
  
piroxicam	
  in	
  a	
  canine	
  model	
  of	
  human	
  invasive	
  urinary	
  bladder	
  carcinoma.	
  Clinical	
  cancer	
  
research	
  :	
  an	
  official	
  journal	
  of	
  the	
  American	
  Association	
  for	
  Cancer	
  Research.	
  2003;9:	
  906-­‐
911.	
  

	
  24	
  

	
  
28.	
  McDougal	
  W,	
  Shipley	
  W,	
  Kaufman	
  D,	
  al	
  e.	
  Cancer	
  of	
  the	
  Bladder,	
  Ureter,	
  and	
  Renal	
  
Pelvis.	
  In:	
  Devita	
  VJ,	
  Lawrence	
  T,	
  Rosenberg	
  S,	
  editors.	
  Cancer:	
  Principles	
  and	
  Practice	
  of	
  
Oncology.	
  Philadelphia,	
  PA:	
  Lippincott	
  Williams	
  &	
  Wilkins,	
  2008:1358-­‐1384.	
  
29.	
  Rocha	
  TA,	
  Mauldin	
  GN,	
  Patnaik	
  AK,	
  Bergman	
  PJ.	
  Prognostic	
  factors	
  in	
  dogs	
  with	
  urinary	
  
bladder	
  carcinoma.	
  J	
  Vet	
  Intern	
  Med.	
  2000;14:	
  486-­‐490.	
  
30.	
  McMillan	
  SK,	
  Boria	
  P,	
  Moore	
  GE,	
  Widmer	
  WR,	
  Bonney	
  PL,	
  Knapp	
  DW.	
  Antitumor	
  effects	
  
of	
  deracoxib	
  treatment	
  in	
  26	
  dogs	
  with	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder.	
  J	
  
Am	
  Vet	
  Med	
  Assoc.	
  2011;239:	
  1084-­‐1089.	
  
31.	
  Schrempp	
  DR,	
  Childress	
  MO,	
  Stewart	
  JC,	
  et	
  al.	
  Metronomic	
  administration	
  of	
  
chlorambucil	
  for	
  treatment	
  of	
  dogs	
  with	
  urinary	
  bladder	
  transitional	
  cell	
  carcinoma.	
  J	
  Am	
  
Vet	
  Med	
  Assoc.	
  2013;242:	
  1534-­‐1538.	
  
32.	
  Henry	
  CJ,	
  McCaw	
  DL,	
  Turnquist	
  SE,	
  et	
  al.	
  Clinical	
  evaluation	
  of	
  mitoxantrone	
  and	
  
piroxicam	
  in	
  a	
  canine	
  model	
  of	
  human	
  invasive	
  urinary	
  bladder	
  carcinoma.	
  Clin	
  Cancer	
  Res.	
  
2003;9:	
  906-­‐911.	
  
33.	
  Hanazono	
  K,	
  Fukumoto	
  S,	
  Endo	
  Y,	
  Ueno	
  H,	
  Kadosawa	
  T,	
  Uchide	
  T.	
  Ultrasonographic	
  
findings	
  related	
  to	
  prognosis	
  in	
  canine	
  transitional	
  cell	
  carcinoma.	
  Vet	
  Radiol	
  Ultrasound.	
  
2014;55:	
  79-­‐84.	
  
34.	
  Norris	
  AM,	
  Laing	
  EJ,	
  Valli	
  VE,	
  et	
  al.	
  Canine	
  bladder	
  and	
  urethral	
  tumors:	
  a	
  retrospective	
  
study	
  of	
  115	
  cases	
  (1980-­‐1985).	
  J	
  Vet	
  Intern	
  Med.	
  1992;6:	
  145-­‐153.	
  
35.	
  Cerf	
  DJ,	
  Lindquist	
  EC.	
  Palliative	
  ultrasound-­‐guided	
  endoscopic	
  diode	
  laser	
  ablation	
  of	
  
transitional	
  cell	
  carcinomas	
  of	
  the	
  lower	
  urinary	
  tract	
  in	
  dogs.	
  J	
  Am	
  Vet	
  Med	
  Assoc.	
  
2012;240:	
  51-­‐60.	
  
36.	
  Nolan	
  MW,	
  Kogan	
  L,	
  Griffin	
  LR,	
  et	
  al.	
  Intensity-­‐modulated	
  and	
  image-­‐guided	
  radiation	
  
therapy	
  for	
  treatment	
  of	
  genitourinary	
  carcinomas	
  in	
  dogs.	
  J	
  Vet	
  Intern	
  Med.	
  2012;26:	
  987-­‐
995.	
  
37.	
  Arnold	
  EJ,	
  Childress	
  MO,	
  Fourez	
  LM,	
  et	
  al.	
  Clinical	
  trial	
  of	
  vinblastine	
  in	
  dogs	
  with	
  
transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder.	
  J	
  Vet	
  Intern	
  Med.	
  2011;25:	
  1385-­‐1390.	
  
38.	
  Robat	
  C,	
  Burton	
  J,	
  Thamm	
  D,	
  Vail	
  D.	
  Retrospective	
  evaluation	
  of	
  doxorubicin-­‐piroxicam	
  
combination	
  for	
  the	
  treatment	
  of	
  transitional	
  cell	
  carcinoma	
  in	
  dogs.	
  J	
  Small	
  Anim	
  Pract.	
  
2013;54:	
  67-­‐74.	
  
39.	
  Mostofi	
  K,	
  Ito	
  N,	
  Weinstein	
  R.	
  Paathology;	
  the	
  need	
  for	
  standardization	
  of	
  pathological	
  
examination	
  and	
  reporting.	
  Developments	
  in	
  Bladder	
  Cancer.	
  New	
  York,	
  NY:	
  Alan	
  R.	
  Liss,	
  
Inc,	
  1986:66-­‐83.	
  
40.	
  Epstein	
  JI,	
  Amin	
  MB,	
  Reuter	
  VR,	
  Mostofi	
  FK.	
  The	
  World	
  Health	
  
Organization/International	
  Society	
  of	
  Urological	
  Pathology	
  consensus	
  classification	
  of	
  

	
  25	
  

	
  
urothelial	
  (transitional	
  cell)	
  neoplasms	
  of	
  the	
  urinary	
  bladder.	
  Bladder	
  Consensus	
  
Conference	
  Committee.	
  The	
  American	
  journal	
  of	
  surgical	
  pathology.	
  1998;22:	
  1435-­‐1448.	
  
41.	
  Eble	
  J,	
  Sauter	
  G,	
  Epstein	
  JI,	
  Sesterhenn	
  IA.	
  World	
  Health	
  Organization	
  classification	
  of	
  
tumours.	
  Pathology	
  and	
  genetics	
  of	
  tumours	
  of	
  the	
  urinary	
  system	
  and	
  male	
  genital	
  organs.	
  
Lyon:	
  IARC	
  Press,	
  2004.	
  
42.	
  Alsheikh	
  A,	
  Mohamedali	
  Z,	
  Jones	
  E,	
  Masterson	
  J,	
  Gilks	
  CB.	
  Comparison	
  of	
  the	
  WHO/ISUP	
  
classification	
  and	
  cytokeratin	
  20	
  expression	
  in	
  predicting	
  the	
  behavior	
  of	
  low-­‐grade	
  
papillary	
  urothelial	
  tumors.	
  World/Health	
  Organization/Internattional	
  Society	
  of	
  Urologic	
  
Pathology.	
  Modern	
  pathology	
  :	
  an	
  official	
  journal	
  of	
  the	
  United	
  States	
  and	
  Canadian	
  
Academy	
  of	
  Pathology,	
  Inc.	
  2001;14:	
  267-­‐272.	
  
43.	
  Oosterhuis	
  JW,	
  Schapers	
  RF,	
  Janssen-­‐Heijnen	
  ML,	
  Pauwels	
  RP,	
  Newling	
  DW,	
  ten	
  Kate	
  F.	
  
Histological	
  grading	
  of	
  papillary	
  urothelial	
  carcinoma	
  of	
  the	
  bladder:	
  prognostic	
  value	
  of	
  
the	
  1998	
  WHO/ISUP	
  classification	
  system	
  and	
  comparison	
  with	
  conventional	
  grading	
  
systems.	
  Journal	
  of	
  clinical	
  pathology.	
  2002;55:	
  900-­‐905.	
  
44.	
  Pan	
  CC,	
  Chang	
  YH,	
  Chen	
  KK,	
  Yu	
  HJ,	
  Sun	
  CH,	
  Ho	
  DM.	
  Constructing	
  prognostic	
  model	
  
incorporating	
  the	
  2004	
  WHO/ISUP	
  classification	
  for	
  patients	
  with	
  non-­‐muscle-­‐invasive	
  
urothelial	
  tumours	
  of	
  the	
  urinary	
  bladder.	
  Journal	
  of	
  clinical	
  pathology.	
  2010;63:	
  910-­‐915.	
  
45.	
  Pan	
  CC,	
  Chang	
  YH,	
  Chen	
  KK,	
  Yu	
  HJ,	
  Sun	
  CH,	
  Ho	
  DM.	
  Prognostic	
  significance	
  of	
  the	
  2004	
  
WHO/ISUP	
  classification	
  for	
  prediction	
  of	
  recurrence,	
  progression,	
  and	
  cancer-­‐specific	
  
mortality	
  of	
  non-­‐muscle-­‐invasive	
  urothelial	
  tumors	
  of	
  the	
  urinary	
  bladder:	
  a	
  
clinicopathologic	
  study	
  of	
  1,515	
  cases.	
  American	
  journal	
  of	
  clinical	
  pathology.	
  2010;133:	
  
788-­‐795.	
  
46.	
  Pich	
  A,	
  Chiusa	
  L,	
  Formiconi	
  A,	
  Galliano	
  D,	
  Bortolin	
  P,	
  Navone	
  R.	
  Biologic	
  differences	
  
between	
  noninvasive	
  papillary	
  urothelial	
  neoplasms	
  of	
  low	
  malignant	
  potential	
  and	
  low-­‐
grade	
  (grade	
  1)	
  papillary	
  carcinomas	
  of	
  the	
  bladder.	
  The	
  American	
  journal	
  of	
  surgical	
  
pathology.	
  2001;25:	
  1528-­‐1533.	
  
47.	
  Samaratunga	
  H,	
  Makarov	
  DV,	
  Epstein	
  JI.	
  Comparison	
  of	
  WHO/ISUP	
  and	
  WHO	
  
classification	
  of	
  noninvasive	
  papillary	
  urothelial	
  neoplasms	
  for	
  risk	
  of	
  progression.	
  
Urology.	
  2002;60:	
  315-­‐319.	
  
48.	
  Schned	
  AR,	
  Andrew	
  AS,	
  Marsit	
  CJ,	
  Zens	
  MS,	
  Kelsey	
  KT,	
  Karagas	
  MR.	
  Survival	
  following	
  
the	
  diagnosis	
  of	
  noninvasive	
  bladder	
  cancer:	
  WHO/International	
  Society	
  of	
  Urological	
  
Pathology	
  versus	
  WHO	
  classification	
  systems.	
  The	
  Journal	
  of	
  urology.	
  2007;178:	
  1196-­‐
1200;	
  discussion	
  1200.	
  
49.	
  Vardar	
  E,	
  Gunlusoy	
  B,	
  Minareci	
  S,	
  Postaci	
  H,	
  Ayder	
  AR.	
  Evaluation	
  of	
  p53	
  nuclear	
  
accumulation	
  in	
  low-­‐	
  and	
  high-­‐grade	
  (WHO/ISUP	
  classification)	
  transitional	
  papillary	
  
carcinomas	
  of	
  the	
  bladder	
  for	
  tumor	
  recurrence	
  and	
  progression.	
  Urologia	
  internationalis.	
  
2006;77:	
  27-­‐33.	
  

	
  26	
  

	
  
50.	
  Yin	
  H,	
  Leong	
  AS.	
  Histologic	
  grading	
  of	
  noninvasive	
  papillary	
  urothelial	
  tumors:	
  
validation	
  of	
  the	
  1998	
  WHO/ISUP	
  system	
  by	
  immunophenotyping	
  and	
  follow-­‐up.	
  American	
  
journal	
  of	
  clinical	
  pathology.	
  2004;121:	
  679-­‐687.	
  
51.	
  Wu	
  XR,	
  Lin	
  JH,	
  Walz	
  T,	
  et	
  al.	
  Mammalian	
  uroplakins.	
  A	
  group	
  of	
  highly	
  conserved	
  
urothelial	
  differentiation-­‐related	
  membrane	
  proteins.	
  The	
  Journal	
  of	
  biological	
  chemistry.	
  
1994;269:	
  13716-­‐13724.	
  
52.	
  Wu	
  XR,	
  Kong	
  XP,	
  Pellicer	
  A,	
  Kreibich	
  G,	
  Sun	
  TT.	
  Uroplakins	
  in	
  urothelial	
  biology,	
  
function,	
  and	
  disease.	
  Kidney	
  international.	
  2009;75:	
  1153-­‐1165.	
  
53.	
  Tot	
  T.	
  Cytokeratins	
  20	
  and	
  7	
  as	
  biomarkers:	
  usefulness	
  in	
  discriminating	
  primary	
  from	
  
metastatic	
  adenocarcinoma.	
  European	
  journal	
  of	
  cancer.	
  2002;38:	
  758-­‐763.	
  
54.	
  Vojtesek	
  B,	
  Staskova	
  Z,	
  Nenutil	
  R,	
  et	
  al.	
  A	
  panel	
  of	
  monoclonal	
  antibodies	
  to	
  keratin	
  no.	
  
7:	
  characterization	
  and	
  value	
  in	
  tumor	
  diagnosis.	
  Neoplasma.	
  1990;37:	
  333-­‐342.	
  
55.	
  Soslow	
  RA,	
  Rouse	
  RV,	
  Hendrickson	
  MR,	
  Silva	
  EG,	
  Longacre	
  TA.	
  Transitional	
  cell	
  
neoplasms	
  of	
  the	
  ovary	
  and	
  urinary	
  bladder:	
  a	
  comparative	
  immunohistochemical	
  analysis.	
  
International	
  journal	
  of	
  gynecological	
  pathology	
  :	
  official	
  journal	
  of	
  the	
  International	
  
Society	
  of	
  Gynecological	
  Pathologists.	
  1996;15:	
  257-­‐265.	
  
56.	
  Espinosa	
  de	
  los	
  Monteros	
  A,	
  Fernandez	
  A,	
  Millan	
  MY,	
  Rodriguez	
  F,	
  Herraez	
  P,	
  Martin	
  de	
  
las	
  Mulas	
  J.	
  Coordinate	
  expression	
  of	
  cytokeratins	
  7	
  and	
  20	
  in	
  feline	
  and	
  canine	
  carcinomas.	
  
Veterinary	
  pathology.	
  1999;36:	
  179-­‐190.	
  
57.	
  Ramos-­‐Vara	
  JA,	
  Miller	
  MA,	
  Boucher	
  M,	
  Roudabush	
  A,	
  Johnson	
  GC.	
  Immunohistochemical	
  
detection	
  of	
  uroplakin	
  III,	
  cytokeratin	
  7,	
  and	
  cytokeratin	
  20	
  in	
  canine	
  urothelial	
  tumors.	
  
Veterinary	
  pathology.	
  2003;40:	
  55-­‐62.	
  
58.	
  Moll	
  R,	
  Laufer	
  J,	
  Wu	
  XR,	
  Sun	
  TT.	
  [Uroplakin	
  III,	
  a	
  specific	
  membrane	
  protein	
  of	
  
urothelial	
  umbrella	
  cells,	
  as	
  a	
  histological	
  markers	
  for	
  metastatic	
  transitional	
  cell	
  
carcinomas].	
  Verhandlungen	
  der	
  Deutschen	
  Gesellschaft	
  fur	
  Pathologie.	
  1993;77:	
  260-­‐265.	
  
59.	
  Parker	
  DC,	
  Folpe	
  AL,	
  Bell	
  J,	
  et	
  al.	
  Potential	
  utility	
  of	
  uroplakin	
  III,	
  thrombomodulin,	
  high	
  
molecular	
  weight	
  cytokeratin,	
  and	
  cytokeratin	
  20	
  in	
  noninvasive,	
  invasive,	
  and	
  metastatic	
  
urothelial	
  (transitional	
  cell)	
  carcinomas.	
  The	
  American	
  journal	
  of	
  surgical	
  pathology.	
  
2003;27:	
  1-­‐10.	
  
60.	
  Gruver	
  AM,	
  Amin	
  MB,	
  Luthringer	
  DJ,	
  et	
  al.	
  Selective	
  immunohistochemical	
  markers	
  to	
  
distinguish	
  between	
  metastatic	
  high-­‐grade	
  urothelial	
  carcinoma	
  and	
  primary	
  poorly	
  
differentiated	
  invasive	
  squamous	
  cell	
  carcinoma	
  of	
  the	
  lung.	
  Archives	
  of	
  pathology	
  &	
  
laboratory	
  medicine.	
  2012;136:	
  1339-­‐1346.	
  
61.	
  Matsumoto	
  K,	
  Satoh	
  T,	
  Irie	
  A,	
  et	
  al.	
  Loss	
  expression	
  of	
  uroplakin	
  III	
  is	
  associated	
  with	
  
clinicopathologic	
  features	
  of	
  aggressive	
  bladder	
  cancer.	
  Urology.	
  2008;72:	
  444-­‐449.	
  

	
  27	
  

	
  
62.	
  Gately	
  S,	
  Li	
  WW.	
  Multiple	
  roles	
  of	
  COX-­‐2	
  in	
  tumor	
  angiogenesis:	
  a	
  target	
  for	
  
antiangiogenic	
  therapy.	
  Seminars	
  in	
  oncology.	
  2004;31:	
  2-­‐11.	
  
63.	
  Wendum	
  D,	
  Masliah	
  J,	
  Trugnan	
  G,	
  Flejou	
  JF.	
  Cyclooxygenase-­‐2	
  and	
  its	
  role	
  in	
  colorectal	
  
cancer	
  development.	
  Virchows	
  Archiv	
  :	
  an	
  international	
  journal	
  of	
  pathology.	
  2004;445:	
  
327-­‐333.	
  
64.	
  Wang	
  D,	
  Dubois	
  RN.	
  Prostaglandins	
  and	
  cancer.	
  Gut.	
  2006;55:	
  115-­‐122.	
  
65.	
  Wang	
  MT,	
  Honn	
  KV,	
  Nie	
  D.	
  Cyclooxygenases,	
  prostanoids,	
  and	
  tumor	
  progression.	
  
Cancer	
  metastasis	
  reviews.	
  2007;26:	
  525-­‐534.	
  
66.	
  Greenhough	
  A,	
  Smartt	
  HJ,	
  Moore	
  AE,	
  et	
  al.	
  The	
  COX-­‐2/PGE2	
  pathway:	
  key	
  roles	
  in	
  the	
  
hallmarks	
  of	
  cancer	
  and	
  adaptation	
  to	
  the	
  tumour	
  microenvironment.	
  Carcinogenesis.	
  
2009;30:	
  377-­‐386.	
  
67.	
  Liu	
  B,	
  Qu	
  L,	
  Tao	
  H.	
  Cyclo-­‐oxygenase	
  2	
  up-­‐regulates	
  the	
  effect	
  of	
  multidrug	
  resistance.	
  
Cell	
  biology	
  international.	
  2010;34:	
  21-­‐25.	
  
68.	
  Margulis	
  V,	
  Shariat	
  SF,	
  Ashfaq	
  R,	
  et	
  al.	
  Expression	
  of	
  cyclooxygenase-­‐2	
  in	
  normal	
  
urothelium,	
  and	
  superficial	
  and	
  advanced	
  transitional	
  cell	
  carcinoma	
  of	
  bladder.	
  The	
  
Journal	
  of	
  urology.	
  2007;177:	
  1163-­‐1168.	
  
69.	
  Mohammed	
  SI,	
  Knapp	
  DW,	
  Bostwick	
  DG,	
  et	
  al.	
  Expression	
  of	
  cyclooxygenase-­‐2	
  (COX-­‐2)	
  
in	
  human	
  invasive	
  transitional	
  cell	
  carcinoma	
  (TCC)	
  of	
  the	
  urinary	
  bladder.	
  Cancer	
  
research.	
  1999;59:	
  5647-­‐5650.	
  
70.	
  Shirahama	
  T.	
  Cyclooxygenase-­‐2	
  expression	
  is	
  up-­‐regulated	
  in	
  transitional	
  cell	
  
carcinoma	
  and	
  its	
  preneoplastic	
  lesions	
  in	
  the	
  human	
  urinary	
  bladder.	
  Clinical	
  cancer	
  
research	
  :	
  an	
  official	
  journal	
  of	
  the	
  American	
  Association	
  for	
  Cancer	
  Research.	
  2000;6:	
  
2424-­‐2430.	
  
71.	
  Komhoff	
  M,	
  Guan	
  Y,	
  Shappell	
  HW,	
  et	
  al.	
  Enhanced	
  expression	
  of	
  cyclooxygenase-­‐2	
  in	
  
high	
  grade	
  human	
  transitional	
  cell	
  bladder	
  carcinomas.	
  The	
  American	
  journal	
  of	
  pathology.	
  
2000;157:	
  29-­‐35.	
  
72.	
  Ristimaki	
  A,	
  Nieminen	
  O,	
  Saukkonen	
  K,	
  Hotakainen	
  K,	
  Nordling	
  S,	
  Haglund	
  C.	
  
Expression	
  of	
  cyclooxygenase-­‐2	
  in	
  human	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  
bladder.	
  The	
  American	
  journal	
  of	
  pathology.	
  2001;158:	
  849-­‐853.	
  
73.	
  Shirahama	
  T,	
  Arima	
  J,	
  Akiba	
  S,	
  Sakakura	
  C.	
  Relation	
  between	
  cyclooxygenase-­‐2	
  
expression	
  and	
  tumor	
  invasiveness	
  and	
  patient	
  survival	
  in	
  transitional	
  cell	
  carcinoma	
  of	
  
the	
  urinary	
  bladder.	
  Cancer.	
  2001;92:	
  188-­‐193.	
  
74.	
  Shariat	
  SF,	
  Matsumoto	
  K,	
  Kim	
  J,	
  et	
  al.	
  Correlation	
  of	
  cyclooxygenase-­‐2	
  expression	
  with	
  
molecular	
  markers,	
  pathological	
  features	
  and	
  clinical	
  outcome	
  of	
  transitional	
  cell	
  
carcinoma	
  of	
  the	
  bladder.	
  The	
  Journal	
  of	
  urology.	
  2003;170:	
  985-­‐989.	
  

	
  28	
  

	
  
75.	
  Wadhwa	
  P,	
  Goswami	
  AK,	
  Joshi	
  K,	
  Sharma	
  SK.	
  Cyclooxygenase-­‐2	
  expression	
  increases	
  
with	
  the	
  stage	
  and	
  grade	
  in	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder.	
  International	
  
urology	
  and	
  nephrology.	
  2005;37:	
  47-­‐53.	
  
76.	
  Cohen	
  GM.	
  Caspases:	
  the	
  executioners	
  of	
  apoptosis.	
  The	
  Biochemical	
  journal.	
  1997;326	
  
(	
  Pt	
  1):	
  1-­‐16.	
  
77.	
  Porter	
  AG,	
  Janicke	
  RU.	
  Emerging	
  roles	
  of	
  caspase-­‐3	
  in	
  apoptosis.	
  Cell	
  death	
  and	
  
differentiation.	
  1999;6:	
  99-­‐104.	
  
78.	
  Grutter	
  MG.	
  Caspases:	
  key	
  players	
  in	
  programmed	
  cell	
  death.	
  Current	
  opinion	
  in	
  
structural	
  biology.	
  2000;10:	
  649-­‐655.	
  
79.	
  Stadelmann	
  C,	
  Lassmann	
  H.	
  Detection	
  of	
  apoptosis	
  in	
  tissue	
  sections.	
  Cell	
  and	
  tissue	
  
research.	
  2000;301:	
  19-­‐31.	
  
80.	
  Abu-­‐Qare	
  AW,	
  Abou-­‐Donia	
  MB.	
  Biomarkers	
  of	
  apoptosis:	
  release	
  of	
  cytochrome	
  c,	
  
activation	
  of	
  caspase-­‐3,	
  induction	
  of	
  8-­‐hydroxy-­‐2'-­‐deoxyguanosine,	
  increased	
  3-­‐
nitrotyrosine,	
  and	
  alteration	
  of	
  p53	
  gene.	
  Journal	
  of	
  toxicology	
  and	
  environmental	
  health.	
  
Part	
  B,	
  Critical	
  reviews.	
  2001;4:	
  313-­‐332.	
  
81.	
  Karam	
  JA,	
  Lotan	
  Y,	
  Karakiewicz	
  PI,	
  et	
  al.	
  Use	
  of	
  combined	
  apoptosis	
  biomarkers	
  for	
  
prediction	
  of	
  bladder	
  cancer	
  recurrence	
  and	
  mortality	
  after	
  radical	
  cystectomy.	
  The	
  lancet	
  
oncology.	
  2007;8:	
  128-­‐136.	
  
82.	
  Karamitopoulou	
  E,	
  Rentsch	
  CA,	
  Markwalder	
  R,	
  Vallan	
  C,	
  Thalmann	
  GN,	
  Brunner	
  T.	
  
Prognostic	
  significance	
  of	
  apoptotic	
  cell	
  death	
  in	
  bladder	
  cancer:	
  a	
  tissue	
  microarray	
  study	
  
on	
  179	
  urothelial	
  carcinomas	
  from	
  cystectomy	
  specimens.	
  Pathology.	
  2010;42:	
  37-­‐42.	
  
83.	
  Mitra	
  AP,	
  Castelao	
  JE,	
  Hawes	
  D,	
  et	
  al.	
  Combination	
  of	
  molecular	
  alterations	
  and	
  smoking	
  
intensity	
  predicts	
  bladder	
  cancer	
  outcome:	
  A	
  report	
  from	
  the	
  Los	
  Angeles	
  Cancer	
  
Surveillance	
  Program.	
  Cancer.	
  2013;119:	
  756-­‐765.	
  
84.	
  Gee	
  J,	
  Lee	
  IL,	
  Grossman	
  HB,	
  Sabichi	
  AL.	
  Forced	
  COX-­‐2	
  expression	
  induces	
  PGE(2)	
  and	
  
invasion	
  in	
  immortalized	
  urothelial	
  cells.	
  Urol	
  Oncol.	
  2008;26:	
  641-­‐645.	
  
85.	
  Muller-­‐Decker	
  K,	
  Furstenberger	
  G.	
  The	
  cyclooxygenase-­‐2-­‐mediated	
  prostaglandin	
  
signaling	
  is	
  causally	
  related	
  to	
  epithelial	
  carcinogenesis.	
  Mol	
  Carcinog.	
  2007;46:	
  705-­‐710.	
  
86.	
  Ono	
  M.	
  Molecular	
  links	
  between	
  tumor	
  angiogenesis	
  and	
  inflammation:	
  inflammatory	
  
stimuli	
  of	
  macrophages	
  and	
  cancer	
  cells	
  as	
  targets	
  for	
  therapeutic	
  strategy.	
  Cancer	
  Sci.	
  
2008;99:	
  1501-­‐1506.	
  
87.	
  Taylor	
  JA,	
  3rd,	
  Pilbeam	
  C,	
  Nisbet	
  A.	
  Role	
  of	
  the	
  prostaglandin	
  pathway	
  and	
  the	
  use	
  of	
  
NSAIDs	
  in	
  genitourinary	
  malignancies.	
  Expert	
  Rev	
  Anticancer	
  Ther.	
  2008;8:	
  1125-­‐1134.	
  

	
  29	
  

	
  
88.	
  Taylor	
  JA,	
  3rd,	
  Ristau	
  B,	
  Bonnemaison	
  M,	
  et	
  al.	
  Regulation	
  of	
  the	
  prostaglandin	
  pathway	
  
during	
  development	
  of	
  invasive	
  bladder	
  cancer	
  in	
  mice.	
  Prostaglandins	
  Other	
  Lipid	
  Mediat.	
  
2009;88:	
  36-­‐41.	
  
89.	
  Shariat	
  SF,	
  Matsumoto	
  K,	
  Kim	
  J,	
  et	
  al.	
  Correlation	
  of	
  cyclooxygenase-­‐2	
  expression	
  with	
  
molecular	
  markers,	
  pathological	
  features	
  and	
  clinical	
  outcome	
  of	
  transitional	
  cell	
  
carcinoma	
  of	
  the	
  bladder.	
  J	
  Urol.	
  2003;170:	
  985-­‐989.	
  
90.	
  Fosbol	
  EL,	
  Gislason	
  GH,	
  Jacobsen	
  S,	
  et	
  al.	
  Risk	
  of	
  myocardial	
  infarction	
  and	
  death	
  
associated	
  with	
  the	
  use	
  of	
  nonsteroidal	
  anti-­‐inflammatory	
  drugs	
  (NSAIDs)	
  among	
  healthy	
  
individuals:	
  a	
  nationwide	
  cohort	
  study.	
  Clin	
  Pharmacol	
  Ther.	
  2009;85:	
  190-­‐197.	
  
91.	
  Celis	
  JE,	
  Ostergaard	
  M,	
  Basse	
  B,	
  et	
  al.	
  Loss	
  of	
  adipocyte-­‐type	
  fatty	
  acid	
  binding	
  protein	
  
and	
  other	
  protein	
  biomarkers	
  is	
  associated	
  with	
  progression	
  of	
  human	
  bladder	
  transitional	
  
cell	
  carcinomas.	
  Cancer	
  Res.	
  1996;56:	
  4782-­‐4790.	
  
92.	
  Gee	
  JR,	
  Montoya	
  RG,	
  Khaled	
  HM,	
  Sabichi	
  AL,	
  Grossman	
  HB.	
  Cytokeratin	
  20,	
  AN43,	
  PGDH,	
  
and	
  COX-­‐2	
  expression	
  in	
  transitional	
  and	
  squamous	
  cell	
  carcinoma	
  of	
  the	
  bladder.	
  Urol	
  
Oncol.	
  2003;21:	
  266-­‐270.	
  
93.	
  Thompson	
  CL,	
  Fink	
  SP,	
  Lutterbaugh	
  JD,	
  et	
  al.	
  Genetic	
  variation	
  in	
  15-­‐
hydroxyprostaglandin	
  dehydrogenase	
  and	
  colon	
  cancer	
  susceptibility.	
  PLoS	
  One.	
  2013;8:	
  
e64122.	
  
94.	
  Stemmler	
  MP.	
  Cadherins	
  in	
  development	
  and	
  cancer.	
  Mol	
  Biosyst.	
  2008;4:	
  835-­‐850.	
  
95.	
  Jeanes	
  A,	
  Gottardi	
  CJ,	
  Yap	
  AS.	
  Cadherins	
  and	
  cancer:	
  how	
  does	
  cadherin	
  dysfunction	
  
promote	
  tumor	
  progression?	
  Oncogene.	
  2008;27:	
  6920-­‐6929.	
  
96.	
  Wheelock	
  MJ,	
  Shintani	
  Y,	
  Maeda	
  M,	
  Fukumoto	
  Y,	
  Johnson	
  KR.	
  Cadherin	
  switching.	
  J	
  Cell	
  
Sci.	
  2008;121:	
  727-­‐735.	
  
97.	
  Gloushankova	
  NA.	
  Changes	
  in	
  regulation	
  of	
  cell-­‐cell	
  adhesion	
  during	
  tumor	
  
transformation.	
  Biochemistry	
  (Mosc).	
  2008;73:	
  742-­‐750.	
  
98.	
  Howard	
  EW,	
  Camm	
  KD,	
  Wong	
  YC,	
  Wang	
  XH.	
  E-­‐cadherin	
  upregulation	
  as	
  a	
  therapeutic	
  
goal	
  in	
  cancer	
  treatment.	
  Mini	
  Rev	
  Med	
  Chem.	
  2008;8:	
  496-­‐518.	
  
99.	
  Nowak	
  M,	
  Madej	
  JA,	
  Dziegiel	
  P.	
  Expression	
  of	
  E-­‐cadherin,	
  beta-­‐catenin	
  and	
  Ki-­‐67	
  
antigen	
  and	
  their	
  reciprocal	
  relationships	
  in	
  mammary	
  adenocarcinomas	
  in	
  bitches.	
  Folia	
  
Histochem	
  Cytobiol.	
  2007;45:	
  233-­‐238.	
  
100.	
  Tseng-­‐Rogenski	
  S,	
  Lee	
  IL,	
  Gebhardt	
  D,	
  et	
  al.	
  Loss	
  of	
  15-­‐hydroxyprostaglandin	
  
dehydrogenase	
  expression	
  disrupts	
  urothelial	
  differentiation.	
  Urology.	
  2008;71:	
  346-­‐350.	
  

	
  30	
  

	
  
101.	
  Dohadwala	
  M,	
  Yang	
  SC,	
  Luo	
  J,	
  et	
  al.	
  Cyclooxygenase-­‐2-­‐dependent	
  regulation	
  of	
  E-­‐
cadherin:	
  prostaglandin	
  E(2)	
  induces	
  transcriptional	
  repressors	
  ZEB1	
  and	
  snail	
  in	
  non-­‐
small	
  cell	
  lung	
  cancer.	
  Cancer	
  Res.	
  2006;66:	
  5338-­‐5345.	
  
102.	
  Brouxhon	
  S,	
  Kyrkanides	
  S,	
  O'Banion	
  MK,	
  et	
  al.	
  Sequential	
  down-­‐regulation	
  of	
  E-­‐
cadherin	
  with	
  squamous	
  cell	
  carcinoma	
  progression:	
  loss	
  of	
  E-­‐cadherin	
  via	
  a	
  prostaglandin	
  
E2-­‐EP2	
  dependent	
  posttranslational	
  mechanism.	
  Cancer	
  Res.	
  2007;67:	
  7654-­‐7664.	
  
103.	
  Mohammed	
  SI,	
  Bennett	
  PF,	
  Craig	
  BA,	
  et	
  al.	
  Effects	
  of	
  the	
  cyclooxygenase	
  inhibitor,	
  
piroxicam,	
  on	
  tumor	
  response,	
  apoptosis,	
  and	
  angiogenesis	
  in	
  a	
  canine	
  model	
  of	
  human	
  
invasive	
  urinary	
  bladder	
  cancer.	
  Cancer	
  Res.	
  2002;62:	
  356-­‐358.	
  
104.	
  Gee	
  J,	
  Lee	
  IL,	
  Jendiroba	
  D,	
  Fischer	
  SM,	
  Grossman	
  HB,	
  Sabichi	
  AL.	
  Selective	
  
cyclooxygenase-­‐2	
  inhibitors	
  inhibit	
  growth	
  and	
  induce	
  apoptosis	
  of	
  bladder	
  cancer.	
  Oncol	
  
Rep.	
  2006;15:	
  471-­‐477.	
  
105.	
  Mutsaers	
  AJ,	
  Mohammed	
  SI,	
  DeNicola	
  DB,	
  et	
  al.	
  Pretreatment	
  tumor	
  prostaglandin	
  E2	
  
concentration	
  and	
  cyclooxygenase-­‐2	
  expression	
  are	
  not	
  associated	
  with	
  the	
  response	
  of	
  
canine	
  naturally	
  occurring	
  invasive	
  urinary	
  bladder	
  cancer	
  to	
  cyclooxygenase	
  inhibitor	
  
therapy.	
  Prostaglandins	
  Leukot	
  Essent	
  Fatty	
  Acids.	
  2005;72:	
  181-­‐186.	
  
106.	
  Bryan	
  RT,	
  Atherfold	
  PA,	
  Yeo	
  Y,	
  et	
  al.	
  Cadherin	
  switching	
  dictates	
  the	
  biology	
  of	
  
transitional	
  cell	
  carcinoma	
  of	
  the	
  bladder:	
  ex	
  vivo	
  and	
  in	
  vitro	
  studies.	
  J	
  Pathol.	
  2008;215:	
  
184-­‐194.	
  
107.	
  Kashibuchi	
  K,	
  Tomita	
  K,	
  Schalken	
  JA,	
  Kume	
  H,	
  Takeuchi	
  T,	
  Kitamura	
  T.	
  The	
  prognostic	
  
value	
  of	
  E-­‐cadherin,	
  alpha-­‐,	
  beta-­‐	
  and	
  gamma-­‐catenin	
  in	
  bladder	
  cancer	
  patients	
  who	
  
underwent	
  radical	
  cystectomy.	
  Int	
  J	
  Urol.	
  2007;14:	
  789-­‐794.	
  
108.	
  Harfe	
  BD,	
  Jinks-­‐Robertson	
  S.	
  DNA	
  mismatch	
  repair	
  and	
  genetic	
  instability.	
  Annual	
  
review	
  of	
  genetics.	
  2000;34:	
  359-­‐399.	
  
109.	
  Hsieh	
  P,	
  Yamane	
  K.	
  DNA	
  mismatch	
  repair:	
  molecular	
  mechanism,	
  cancer,	
  and	
  ageing.	
  
Mechanisms	
  of	
  ageing	
  and	
  development.	
  2008;129:	
  391-­‐407.	
  
110.	
  Iyer	
  RR,	
  Pluciennik	
  A,	
  Burdett	
  V,	
  Modrich	
  PL.	
  DNA	
  mismatch	
  repair:	
  functions	
  and	
  
mechanisms.	
  Chemical	
  reviews.	
  2006;106:	
  302-­‐323.	
  
111.	
  Jiricny	
  J.	
  The	
  multifaceted	
  mismatch-­‐repair	
  system.	
  Nature	
  reviews.	
  Molecular	
  cell	
  
biology.	
  2006;7:	
  335-­‐346.	
  
112.	
  Preston	
  BD,	
  Albertson	
  TM,	
  Herr	
  AJ.	
  DNA	
  replication	
  fidelity	
  and	
  cancer.	
  Seminars	
  in	
  
cancer	
  biology.	
  2010;20:	
  281-­‐293.	
  
113.	
  Chung	
  H,	
  Young	
  DJ,	
  Lopez	
  CG,	
  et	
  al.	
  Mutation	
  rates	
  of	
  TGFBR2	
  and	
  ACVR2	
  coding	
  
microsatellites	
  in	
  human	
  cells	
  with	
  defective	
  DNA	
  mismatch	
  repair.	
  PloS	
  one.	
  2008;3:	
  
e3463.	
  

	
  31	
  

	
  
114.	
  Kim	
  CJ,	
  Lee	
  JH,	
  Song	
  JW,	
  et	
  al.	
  Chk1	
  frameshift	
  mutation	
  in	
  sporadic	
  and	
  hereditary	
  
non-­‐polyposis	
  colorectal	
  cancers	
  with	
  microsatellite	
  instability.	
  European	
  journal	
  of	
  
surgical	
  oncology	
  :	
  the	
  journal	
  of	
  the	
  European	
  Society	
  of	
  Surgical	
  Oncology	
  and	
  the	
  British	
  
Association	
  of	
  Surgical	
  Oncology.	
  2007;33:	
  580-­‐585.	
  
115.	
  Miquel	
  C,	
  Jacob	
  S,	
  Grandjouan	
  S,	
  et	
  al.	
  Frequent	
  alteration	
  of	
  DNA	
  damage	
  signalling	
  
and	
  repair	
  pathways	
  in	
  human	
  colorectal	
  cancers	
  with	
  microsatellite	
  instability.	
  Oncogene.	
  
2007;26:	
  5919-­‐5926.	
  
116.	
  Fernandez-­‐Peralta	
  AM,	
  Nejda	
  N,	
  Oliart	
  S,	
  Medina	
  V,	
  Azcoita	
  MM,	
  Gonzalez-­‐Aguilera	
  JJ.	
  
Significance	
  of	
  mutations	
  in	
  TGFBR2	
  and	
  BAX	
  in	
  neoplastic	
  progression	
  and	
  patient	
  
outcome	
  in	
  sporadic	
  colorectal	
  tumors	
  with	
  high-­‐frequency	
  microsatellite	
  instability.	
  
Cancer	
  genetics	
  and	
  cytogenetics.	
  2005;157:	
  18-­‐24.	
  
117.	
  Hampson	
  R.	
  Selection	
  for	
  genome	
  instability	
  by	
  DNA	
  damage	
  in	
  human	
  cells:	
  unstable	
  
microsatellites	
  and	
  their	
  consequences	
  for	
  tumourigenesis.	
  Radiation	
  oncology	
  
investigations.	
  1997;5:	
  111-­‐114.	
  
118.	
  Jascur	
  T,	
  Boland	
  CR.	
  Structure	
  and	
  function	
  of	
  the	
  components	
  of	
  the	
  human	
  DNA	
  
mismatch	
  repair	
  system.	
  International	
  journal	
  of	
  cancer.	
  Journal	
  international	
  du	
  cancer.	
  
2006;119:	
  2030-­‐2035.	
  
119.	
  Eckert	
  KA,	
  Hile	
  SE.	
  Every	
  microsatellite	
  is	
  different:	
  Intrinsic	
  DNA	
  features	
  dictate	
  
mutagenesis	
  of	
  common	
  microsatellites	
  present	
  in	
  the	
  human	
  genome.	
  Molecular	
  
carcinogenesis.	
  2009;48:	
  379-­‐388.	
  
120.	
  Umar	
  A,	
  Boland	
  CR,	
  Terdiman	
  JP,	
  et	
  al.	
  Revised	
  Bethesda	
  Guidelines	
  for	
  hereditary	
  
nonpolyposis	
  colorectal	
  cancer	
  (Lynch	
  syndrome)	
  and	
  microsatellite	
  instability.	
  Journal	
  of	
  
the	
  National	
  Cancer	
  Institute.	
  2004;96:	
  261-­‐268.	
  
121.	
  Catto	
  JW,	
  Azzouzi	
  AR,	
  Amira	
  N,	
  et	
  al.	
  Distinct	
  patterns	
  of	
  microsatellite	
  instability	
  are	
  
seen	
  in	
  tumours	
  of	
  the	
  urinary	
  tract.	
  Oncogene.	
  2003;22:	
  8699-­‐8706.	
  
122.	
  D'Errico	
  M,	
  de	
  Rinaldis	
  E,	
  Blasi	
  MF,	
  et	
  al.	
  Genome-­‐wide	
  expression	
  profile	
  of	
  sporadic	
  
gastric	
  cancers	
  with	
  microsatellite	
  instability.	
  European	
  journal	
  of	
  cancer.	
  2009;45:	
  461-­‐
469.	
  
123.	
  Hewish	
  M,	
  Lord	
  CJ,	
  Martin	
  SA,	
  Cunningham	
  D,	
  Ashworth	
  A.	
  Mismatch	
  repair	
  deficient	
  
colorectal	
  cancer	
  in	
  the	
  era	
  of	
  personalized	
  treatment.	
  Nature	
  reviews.	
  Clinical	
  oncology.	
  
2010;7:	
  197-­‐208.	
  
124.	
  Jacob	
  S,	
  Praz	
  F.	
  DNA	
  mismatch	
  repair	
  defects:	
  role	
  in	
  colorectal	
  carcinogenesis.	
  
Biochimie.	
  2002;84:	
  27-­‐47.	
  
125.	
  Mylona	
  E,	
  Zarogiannos	
  A,	
  Nomikos	
  A,	
  et	
  al.	
  Prognostic	
  value	
  of	
  microsatellite	
  
instability	
  determined	
  by	
  immunohistochemical	
  staining	
  of	
  hMSH2	
  and	
  hMSH6	
  in	
  

	
  32	
  

	
  
urothelial	
  carcinoma	
  of	
  the	
  bladder.	
  Acta	
  pathologica,	
  microbiologica	
  et	
  immunologica	
  
Scandinavica.	
  2008;116:	
  59-­‐65.	
  
126.	
  di	
  Pietro	
  M,	
  Sabates	
  Bellver	
  J,	
  Menigatti	
  M,	
  et	
  al.	
  Defective	
  DNA	
  mismatch	
  repair	
  
determines	
  a	
  characteristic	
  transcriptional	
  profile	
  in	
  proximal	
  colon	
  cancers.	
  
Gastroenterology.	
  2005;129:	
  1047-­‐1059.	
  
127.	
  Schwitalle	
  Y,	
  Kloor	
  M,	
  Eiermann	
  S,	
  et	
  al.	
  Immune	
  response	
  against	
  frameshift-­‐induced	
  
neopeptides	
  in	
  HNPCC	
  patients	
  and	
  healthy	
  HNPCC	
  mutation	
  carriers.	
  Gastroenterology.	
  
2008;134:	
  988-­‐997.	
  
128.	
  Kawakami	
  T,	
  Shiina	
  H,	
  Igawa	
  M,	
  et	
  al.	
  Inactivation	
  of	
  the	
  hMSH3	
  mismatch	
  repair	
  gene	
  
in	
  bladder	
  cancer.	
  Biochemical	
  and	
  biophysical	
  research	
  communications.	
  2004;325:	
  934-­‐
942.	
  
129.	
  Casorelli	
  I,	
  Russo	
  MT,	
  Bignami	
  M.	
  Role	
  of	
  mismatch	
  repair	
  and	
  MGMT	
  in	
  response	
  to	
  
anticancer	
  therapies.	
  Anti-­‐cancer	
  agents	
  in	
  medicinal	
  chemistry.	
  2008;8:	
  368-­‐380.	
  
130.	
  Cejka	
  P,	
  Stojic	
  L,	
  Marra	
  G,	
  Jiricny	
  J.	
  Is	
  mismatch	
  repair	
  really	
  required	
  for	
  ionizing	
  
radiation-­‐induced	
  DNA	
  damage	
  signaling?	
  Nature	
  Genetics.	
  2004;36:	
  432-­‐433.	
  
131.	
  Flanagan	
  SA,	
  Robinson	
  BW,	
  Krokosky	
  CW,	
  Shewach	
  DS.	
  Mismatched	
  nucleotides	
  as	
  the	
  
lesions	
  responsible	
  for	
  radiosensitization	
  with	
  gemcitabine:	
  a	
  new	
  paradigm	
  for	
  
antimetabolite	
  radiosentizers.	
  Molecular	
  cancer	
  therapeutics.	
  2007;6:	
  1858-­‐1868.	
  
132.	
  Hart	
  JR,	
  Glebov	
  O,	
  Ernst	
  RJ,	
  Kirsch	
  IR,	
  Barton	
  JK.	
  DNA	
  mismatch-­‐specific	
  targeting	
  and	
  
hypersensitivity	
  of	
  mismatch-­‐repair-­‐deficient	
  cells	
  to	
  bulky	
  rhodium(III)	
  intercalators.	
  
Proceedings	
  of	
  the	
  National	
  Academy	
  of	
  Sciences	
  of	
  the	
  United	
  States	
  of	
  America.	
  2006;103:	
  
15359-­‐15363.	
  
133.	
  Martin	
  SA,	
  Lord	
  CJ,	
  Ashworth	
  A.	
  Therapeutic	
  targeting	
  of	
  the	
  DNA	
  mismatch	
  repair	
  
pathway.	
  Clinical	
  cancer	
  research	
  :	
  an	
  official	
  journal	
  of	
  the	
  American	
  Association	
  for	
  
Cancer	
  Research.	
  2010;16:	
  5107-­‐5113.	
  
134.	
  Pors	
  K,	
  Patterson	
  LH.	
  DNA	
  mismatch	
  repair	
  deficiency,	
  resistance	
  to	
  cancer	
  
chemotherapy	
  and	
  the	
  development	
  of	
  hypersensitive	
  agents.	
  Current	
  topics	
  in	
  medicinal	
  
chemistry.	
  2005;5:	
  1133-­‐1149.	
  
135.	
  Sargent	
  DJ,	
  Marsoni	
  S,	
  Monges	
  G,	
  et	
  al.	
  Defective	
  mismatch	
  repair	
  as	
  a	
  predictive	
  
marker	
  for	
  lack	
  of	
  efficacy	
  of	
  fluorouracil-­‐based	
  adjuvant	
  therapy	
  in	
  colon	
  cancer.	
  Journal	
  
of	
  clinical	
  oncology	
  :	
  official	
  journal	
  of	
  the	
  American	
  Society	
  of	
  Clinical	
  Oncology.	
  2010;28:	
  
3219-­‐3226.	
  
136.	
  Takahashi	
  T,	
  Min	
  Z,	
  Uchida	
  I,	
  et	
  al.	
  Hypersensitivity	
  in	
  DNA	
  mismatch	
  repair-­‐deficient	
  
colon	
  carcinoma	
  cells	
  to	
  DNA	
  polymerase	
  reaction	
  inhibitors.	
  Cancer	
  letters.	
  2005;220:	
  85-­‐
93.	
  

	
  33	
  

	
  
137.	
  Valentini	
  AM,	
  Armentano	
  R,	
  Pirrelli	
  M,	
  Caruso	
  ML.	
  Chemotherapeutic	
  agents	
  for	
  
colorectal	
  cancer	
  with	
  a	
  defective	
  mismatch	
  repair	
  system:	
  the	
  state	
  of	
  the	
  art.	
  Cancer	
  
treatment	
  reviews.	
  2006;32:	
  607-­‐618.	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  34	
  

	
  
CHAPTER	
  2	
  
	
  
	
  
Differences	
  in	
  Expression	
  of	
  Uroplakin	
  III,	
  Cytokeratin-­‐7,	
  and	
  COX-­‐2	
  in	
  Canine	
  
Proliferative	
  Urothelial	
  Lesions	
  of	
  the	
  Urinary	
  Bladder	
  
	
  

	
  
Dodd	
  Sledge1,	
  Daniel	
  J.	
  Patrick2,	
  Scott	
  D.	
  Fitzgerald1,	
  Yan	
  Xie3,	
  Matti	
  Kiupel1	
  
	
  

1:	
  Department	
  of	
  Pathobiology	
  and	
  Diagnostic	
  Investigation,	
  College	
  of	
  Veterinary	
  
Medicine,	
  Michigan	
  State	
  University,	
  East	
  Lansing	
  MI	
  	
  
2:	
  MPI	
  Research,	
  Mattawan,	
  MI	
  	
  	
  
3:	
  Pharmanet/i3,	
  Haslett,	
  MI	
  (YX)	
  
	
  
	
  
	
  
	
  
	
  

	
  35	
  

	
  
Abstract	
  
The	
  expression	
  of	
  immunohistochemical	
  markers	
  that	
  have	
  been	
  used	
  in	
  diagnosis	
  
and/or	
  prognostication	
  of	
  urothelial	
  tumors	
  in	
  humans	
  (uroplakin	
  III	
  (UPIII),	
  cytokeratin-­‐7	
  
(CK7),	
  cyclooxygenase-­‐2	
  (COX-­‐2),	
  and	
  activated	
  caspase	
  3)	
  was	
  evaluated	
  in	
  a	
  series	
  of	
  99	
  
canine	
   proliferative	
   urothelial	
   lesions	
   of	
   the	
   urinary	
   bladder	
   and	
   compared	
   to	
   the	
   lesion	
  
classification	
   and	
   grade	
   as	
   defined	
   by	
   the	
   WHO/ISUP	
   Consensus	
   System.	
   There	
   were	
  
significant	
   associations	
   between	
   tumor	
   classification	
   and	
   overall	
   UPIII	
   pattern	
   (P=1.49x10-­‐
18),	
  

loss	
   of	
   UPIII	
   (P=1.27x10-­‐4),	
   overall	
   CK7	
   pattern	
   (P=4.34x10-­‐18),	
   and	
   COX-­‐2	
   pattern	
  

(P=8.12x10-­‐25).	
  In	
  addition,	
  there	
  were	
  significant	
  associations	
  between	
  depth	
  of	
  neoplastic	
  
cell	
  infiltration	
  into	
  the	
  urinary	
  bladder	
  wall	
  and	
  overall	
  UPIII	
  pattern	
  (P=1.54x10-­‐14),	
  loss	
  
of	
   UPIII	
   (P=2.07x10-­‐4),	
   overall	
   CK7	
   pattern	
   (P=1.17x10-­‐13),	
   loss	
   of	
   CK7	
   expression	
  
(P=0.0485),	
   and	
   COX-­‐2	
   pattern	
   (P=8.23x10-­‐21).	
   There	
   were	
   no	
   significant	
   associations	
  
between	
  tumor	
  classification	
  or	
  infiltration	
  and	
  caspase-­‐3	
  expression	
  pattern.	
  	
  	
  
	
  
	
  

	
  

	
  36	
  

	
  
Introduction	
  
	
  

Proliferative	
   lesions	
   of	
   the	
   urothelium	
   of	
   the	
   canine	
   urinary	
   bladder	
   range	
   from	
  

benign	
   polyps	
   and	
   papillomas	
   to	
   carcinomas	
   with	
   varying	
   metastatic	
   potential.	
   Also,	
  
inflammatory	
   conditions	
   such	
   as	
   polypoid	
   cystitis	
   can	
   form	
   tumor-­‐like	
   masses	
   within	
   the	
  
urinary	
   bladder	
   that	
   can	
   be	
   confused	
   with	
   urothelial	
   neoplasms.	
   Classification	
   of	
   urothelial	
  
proliferative	
   lesions	
   and	
   histologic	
   grading	
   of	
   urothelial	
   carcinomas	
   is	
   key	
   to	
   accurate	
  
prognostication	
  and	
  treatment	
  selection.	
  A	
  classification	
  and	
  grading	
  scheme	
  based	
  on	
  the	
  
1986	
   World	
   Health	
   Organization	
   classification	
   scheme	
   for	
   human	
   urinary	
   bladder	
   and	
  
urethral	
   cancer	
   was	
   proposed	
   in	
   1995	
   for	
   use	
   in	
   evaluating	
   urothelial	
   neoplasms	
   in	
   dogs	
  
based	
   on	
   pattern	
   of	
   growth,	
   nuclear	
   atypia,	
   and	
   degree	
   of	
   infiltration	
   into	
   the	
   urinary	
  
bladder	
  wall.50	
  In	
  the	
  initial	
  description	
  of	
  this	
  system’s	
  use	
  in	
  dogs,	
  significant	
  correlations	
  
between	
   tumor	
   grade	
   and	
   depth	
   of	
   infiltration,	
   the	
   presence	
   of	
   metastases,	
   and	
   survival	
  
time	
   were	
   found.	
   However,	
   a	
   subsequent	
   smaller	
   study	
   failed	
   to	
   find	
   a	
   significant	
  
correlation	
   between	
   grade	
   and	
   prognosis.41	
   To	
   our	
   knowledge,	
   this	
   scheme	
   is	
   not	
   widely	
  
used	
   in	
   dogs	
   and	
   has	
   been	
   replaced	
   in	
   human	
   medicine	
   by	
   more	
   modern	
   classification	
  
schemes.	
   There	
   remains	
   a	
   need	
   for	
   a	
   clear,	
   reproducible,	
   and	
   well-­‐accepted	
   classification	
  
and	
  grading	
  scheme	
  for	
  urothelial	
  proliferative	
  lesions	
  in	
  dogs.	
  
Currently,	
   the	
   most	
   widely	
   accepted	
   scheme	
   for	
   classification	
   and	
   grading	
   of	
  
proliferative	
   urothelial	
   lesions	
   in	
   humans	
   is	
   the	
   World	
   Health	
   Organization	
  
(WHO)/International	
  Society	
  of	
  Urologic	
  Pathology	
  (ISUP)	
  Consensus	
  Classification	
  System	
  
published	
   in	
   1998	
   and	
   updated	
   in	
   2004.6,7	
   In	
   multiple	
   studies,	
   the	
   WHO/ISUP	
   Consensus	
  
Classification	
   system	
   has	
   been	
   demonstrated	
   to	
   be	
   significantly	
   associated	
   with	
   clinical	
  
outcome.2,32-­‐34,37,42,43,51,59	
   It	
   has	
   long	
   been	
   known	
   that	
   canine	
   urothelial	
   neoplasms	
   are	
  

	
  37	
  

	
  
similar	
  to	
  urothelial	
  neoplasms	
  in	
  humans	
  in	
  terms	
  of	
  morphology,	
  biologic	
  behavior	
  and	
  
response	
  to	
  chemotherapy.14,18,19,27,28,59	
  Recently,	
  Patrick	
  et	
  al.	
  examined	
  the	
  potential	
  use	
  
of	
   the	
   WHO/ISUP	
   Consensus	
   Classification	
   System	
   in	
   classifying	
   canine	
   proliferative	
  
urothelial	
   lesions.36	
   In	
   that	
   study,	
   the	
   authors	
   demonstrated	
   that	
   the	
   histomorphology	
   of	
  
proliferative	
  urothelial	
  lesions	
  was	
  homologous	
  between	
  dogs	
  and	
  humans,	
  and	
  that	
  canine	
  
lesions	
   could	
   easily	
   be	
   classified	
   according	
   to	
   the	
   system;	
   however,	
   data	
   regarding	
   the	
  
prognostic	
   relevance	
   of	
   classification	
   system	
   remains	
   lacking.	
   Additional	
   study	
   of	
   the	
  
biologic	
   differences	
   between	
   classifications	
   and	
   grades	
  of	
   proliferative	
   urothelial	
   lesions	
   in	
  
dogs	
  is	
  needed.	
  
Uroplakins	
   comprise	
   a	
   group	
   of	
   membrane-­‐associated	
   proteins	
   expressed	
   by	
  
urothelial	
   cells	
   that	
   are	
   important	
   for	
   cell-­‐to-­‐cell	
   adhesion	
   and	
   maintenance	
   of	
   water	
  
impermeability.57,58	
   These	
   proteins	
   form	
   a	
   plaque-­‐like	
   complex	
   along	
   the	
   apical	
   membrane	
  
of	
   the	
   umbrella	
   cells	
   that	
   form	
   the	
   most	
   superficial	
   layer	
   of	
   the	
   urothelium.57	
   Cytokeratin	
   7	
  
(CK7)	
  is	
  a	
  cytokeratin	
  expressed	
  by	
  simple	
  epithelium	
  as	
  well	
  as	
  differentiated	
  urothelial	
  
cells	
   and	
   a	
   variety	
   of	
   carcinomas	
   in	
   humans	
   including	
   those	
   of	
   urothelial,	
   pancreatic,	
  
cholangiolar,	
   and	
   ovarian	
   origin.49,52	
   Uroplakin	
   III	
   (UPIII)	
   and	
   CK7	
   are	
   used	
   in	
   both	
   dogs	
  
and	
   humans	
   as	
   diagnostic	
   markers	
   of	
   urothelial	
   differentiation	
   in	
   primary	
   tumors	
   and	
  
metastases.8,13,31,35,39,47	
  In	
  humans,	
  loss	
  of	
  UPIII	
  has	
  been	
  associated	
  with	
  several	
  prognostic	
  
features.13,25	
  
	
  

The	
  inducible	
  enzyme	
  COX-­‐2	
  and	
  the	
  resulting	
  production	
  of	
  prostaglandin	
  E2	
   have	
  

been	
  ascribed	
  significant	
  roles	
  in	
  carcinogenesis,	
  including	
  immunosuppression,	
  inhibition	
  
of	
  apoptosis,	
  increased	
  metastatic	
  potential	
  of	
  neoplastic	
  epithelial	
  cells,	
  promotion	
  of	
  drug	
  
resistance,	
   and	
   stimulation	
   of	
   angiogenesis.9,11,23,54-­‐56	
   Numerous	
   studies	
   have	
   shown	
  

	
  38	
  

	
  
significant	
   correlations	
   between	
   COX-­‐2	
   expression	
   and	
   tumor	
   grade,	
   infiltration,	
  
metastasis,	
  and	
  survival.21,24,30,40,44-­‐46,53	
  Concordantly,	
  a	
  decreased	
  risk	
  for	
  urinary	
  bladder	
  
cancer	
   development	
   has	
   been	
   seen	
   in	
   humans	
   undergoing	
   long-­‐term	
   non-­‐steroidal	
   anti-­‐
inflammatory	
  therapy	
  and	
  in	
  vitro	
  and	
  in	
  vivo	
  studies	
  have	
  suggested	
  potential	
  use	
  of	
  COX-­‐
2	
  inhibitors	
  in	
  treatment.3,5,10,29,53	
  In	
  humans	
  and	
  dogs,	
  COX-­‐2	
  is	
  not	
  expressed	
  by	
  normal	
  
urothelium	
  of	
  the	
  urinary	
  bladder.17,24	
  Substantial	
  expression	
  of	
  COX-­‐2,	
  however,	
  has	
  been	
  
observed	
  in	
  transitional	
  cell	
  carcinomas.17,20,22	
  	
  
Caspase	
  3	
  is	
  an	
  effector	
  or	
  executioner	
  caspase	
  that	
  is	
  activated	
  by	
  both	
  intrinsic	
  and	
  
extrinsic	
   apoptosis	
   signaling	
   pathways	
   to	
   cleave	
   multiple	
   cellular	
   structural	
   and	
   repair	
  
proteins.4,12,38	
   Due	
   to	
   the	
   fact	
   that	
   this	
   protein	
   is	
   activated	
   late	
   in	
   the	
   apoptotic	
   pathway,	
  
immunohistochemical	
  detection	
  of	
  activated	
  caspase-­‐3	
  has	
  been	
  used	
  to	
  evaluate	
  apoptotic	
  
rate.1,48	
  In	
  human	
  urinary	
  bladder	
  cancers,	
  expression	
  of	
  caspase-­‐3	
  has	
  been	
  suggested	
  to	
  
have	
  prognostic	
  significance.15,16,26	
  
The	
  goals	
  of	
  the	
  current	
  study	
  were	
  twofold:	
  1)	
  to	
  evaluate	
  the	
  expression	
  of	
  UPIII,	
  
CK7,	
   COX-­‐2,	
   and	
   caspase	
   3	
   in	
   non-­‐neoplastic	
   and	
   neoplastic	
   proliferative	
   lesions	
   of	
   the	
  
canine	
   urothelium	
   of	
   the	
   urinary	
   bladder,	
   and	
   2)	
   to	
   correlate	
   the	
   observed	
   patterns	
   of	
  
expression	
   of	
   each	
   of	
   these	
   markers	
   with	
   specific	
   lesion	
   classification	
   and	
   grade	
   as	
   defined	
  
by	
  the	
  WHO/ISUP	
  Consensus	
  Classification	
  System.	
  	
  
	
  
	
  
	
  
	
  
	
  

	
  39	
  

	
  
Materials	
  and	
  methods	
  
Selection	
  of	
  Cases	
  and	
  Histologic	
  Classification	
  
A	
   series	
   of	
   99	
   formalin-­‐fixed,	
   paraffin-­‐embedded	
   proliferative	
   urothelial	
   lesions	
  
from	
   99	
   dogs	
   that	
   had	
   been	
   submitted	
   as	
   diagnostic	
   cases	
   were	
   selected	
   from	
   the	
   tissue	
  
archives	
   of	
   the	
   Michigan	
   State	
   University,	
   Diagnostic	
   Center	
   for	
   Population	
   and	
   Animal	
  
Health	
   (DCPAH).	
   Of	
   these	
   99,	
   93	
   had	
   previously	
   been	
   classified	
   and	
   where	
   applicable	
  
graded	
   using	
   the	
   WHO/ISUP	
   Consensus	
   Classification	
   System.36	
   Six	
   additional	
   diagnostic	
  
cases	
  that	
  presented	
  to	
  DCPAH,	
  including	
  additional	
  low-­‐grade	
  neoplasms,	
  were	
  included	
  
in	
   the	
   study	
   set.	
   For	
   grading	
   of	
   these	
   additional	
   samples,	
   5	
   µm	
   sections	
   of	
   all	
   samples	
   were	
  
routinely	
  processed	
  and	
  stained	
  with	
  hematoxylin	
  and	
  eosin	
  for	
  microscopic	
  examination.	
  
Proliferative	
   urothelial	
   lesions	
   were	
   categorized	
   according	
   to	
   the	
   WHO/ISUP	
   Consensus	
  
Classification	
  System	
  as	
  previously	
  described	
  and	
  summarized	
  in	
  Table	
  1.36	
  For	
  urothelial	
  
carcinomas,	
   the	
   degree	
   of	
   infiltration	
   into	
   the	
   urinary	
   bladder	
   wall	
   was	
   scored	
   as	
   no	
  
infiltration,	
   infiltration	
   into	
   the	
   substantia	
   propria,	
   or	
   infiltration	
   into	
   the	
   tunica	
  
muscularis.	
  
Of	
   the	
   99	
   proliferative	
   urothelial	
   lesions	
   examined,	
   44	
   were	
   non-­‐neoplastic	
   and	
  
categorized	
  as	
  either	
  urothelial	
  polyps	
  or	
  polypoid	
  cystitis.	
  Of	
  the	
  55	
  neoplasms,	
  there	
  were	
  
two	
  urothelial	
  papillomas	
  and	
  one	
  papillary	
  urothelial	
  neoplasm	
  of	
  low	
  malignant	
  potential	
  
(PUNLMP).	
   The	
   remainders	
   of	
   the	
   urothelial	
   neoplasms	
   were	
   papillary	
   urothelial	
  
carcinomas	
   of	
   varying	
   grade.	
   Low-­‐grade	
   (grade	
   1)	
   papillary	
   urothelial	
   carcinomas	
   were	
  
rare	
  with	
  only	
  2	
  being	
  included	
  in	
  the	
  set.	
  Both	
  of	
  these	
  had	
  some	
  degree	
  of	
  infiltration	
  into	
  
the	
   urinary	
   bladder	
   wall	
   with	
   one	
   sample	
   having	
   infiltrative	
   clusters	
   of	
   neoplastic	
  

	
  40	
  

	
  
urothelial	
   cells	
   within	
   the	
   muscularis.	
   Papillary	
   carcinomas	
   infiltrated	
   at	
   least	
   into	
   the	
  
substantia	
  propria	
  in	
  49/52	
  (94%)	
  and	
  into	
  the	
  muscularis	
  in	
  21/52	
  (40%)	
  of	
  cases.	
  	
  
Immunohistochemistry	
  
Five	
   µm	
   sections	
   of	
   all	
   samples	
   were	
   processed	
   for	
   immunohistochemistry	
   and	
  
labeled	
  with	
  a	
  mouse	
  monoclonal	
  anti-­‐UPIII	
  antibody	
  (1:5,	
  RDI,	
  Fitzgerald	
  Industries	
  Intl,	
  
Concord,	
   MA,	
   USA),	
   a	
   mouse	
   monoclonal	
   anti-­‐CK7	
   antibody	
   (1:75,	
   Dako	
   Cytomation,	
  
Carpentaria,	
   CA,	
   USA),	
   a	
   rabbit	
   polyclonal	
   anti-­‐COX-­‐2	
   antibody	
   (1:100.	
   Cayman	
   Chemical	
  
Company,	
   Ann	
   Arbor,	
   MI,	
   USA),	
   or	
   rabbit	
   polyclonal	
   anti-­‐activated	
   caspase-­‐3	
   antibody	
  
(1:5,000,	
   RDI,	
   Fitzgerald	
   Industries	
   Intl,	
   Concord,	
   MA,	
   USA).	
   Deparaffinization,	
   antigen	
  
retrieval,	
  immunohistochemical	
  labeling	
  with	
  3,3’-­‐diaminobenzidine	
  (DAB)	
  chromogen	
  and	
  
counterstaining	
   with	
   hematoxylin	
   were	
   performed	
   on	
   the	
   Bond	
   maXTM	
   Automated	
  
Staining	
   System	
   (Vision	
   BioSystemsTM,	
   Leica,	
   Bannockburn,	
   IL,	
   USA)	
   using	
   the	
   BondTM	
  
Polymer	
  Detection	
  System	
  (Vision	
  BioSystemsTM,	
  Leica,	
  Bannockburn,	
  IL,	
  USA).	
  Sections	
  of	
  
normal	
  canine	
  urothelium	
  were	
  similarly	
  labeled	
  as	
  positive	
  controls	
  for	
  UPIII	
  and	
  CK7.	
  A	
  
canine	
   squamous	
   cell	
   carcinoma	
   known	
   to	
   express	
   COX-­‐2	
   and	
   a	
   lymph	
   node	
   with	
   large	
  
numbers	
   of	
   activated	
   caspase-­‐3-­‐positive	
   cells	
   were	
   respectively	
   used	
   as	
   positive	
   controls	
  
for	
   these	
   antibodies.	
   For	
   negative	
   controls,	
   homologous	
   non-­‐immune	
   sera	
   or	
   buffer	
  
replaced	
  primary	
  antibodies.	
  
Immunoreactivity	
  for	
  UPIII	
  and	
  CK7	
  was	
  scored	
  according	
  to	
  overall	
  pattern	
  within	
  
the	
  urothelium	
  and	
  partial	
  loss	
  of	
  immunoreactivity.	
  Immunoreactivity	
  for	
  UPIII	
  and	
  CK7	
  in	
  
positively	
  labeled	
  cells	
  was	
  variably	
  perimembranous,	
  predominately	
  cytoplasmic	
  without	
  
distinct	
  perimembrane	
  labeling,	
  or	
  associated	
  with	
  both	
  the	
  membrane	
  the	
  cell	
  membrane	
  
and	
   the	
   cytoplasm.	
   Specifically,	
   for	
   overall	
   pattern	
   of	
   UPIII	
   and	
   CK7	
   immunoreactivity	
  

	
  41	
  

	
  
within	
   the	
   urothelium:	
   pattern	
   1	
   was	
   defined	
   by	
   labeling	
   limited	
   to	
   the	
   most	
   superficial	
  
layer	
  of	
  cells	
  (Fig.	
  1	
  and	
  4),	
  pattern	
  2	
  was	
  defined	
  by	
  labeling	
  extending	
  to	
  the	
  middle	
  layer	
  
of	
   the	
   urothelium	
   (Fig.	
   2),	
   pattern	
   3	
   was	
   defined	
   by	
   cells	
   labeling	
   throughout	
   the	
   full	
  
thickness	
   of	
   the	
   urothelium	
   (Fig.	
   5),	
   and	
   pattern	
   4	
   was	
   defined	
   by	
   immunoreactivity	
   that	
  
was	
  patchy	
  and	
  randomly	
  distributed	
  (Fig.	
  3	
  and	
  6).	
  Partial	
  loss	
  of	
  immunoreactivity	
  was	
  
noted	
   when	
   greater	
   than	
   50%	
   of	
   epithelial	
   cells	
   were	
   immunonegative	
   and	
   there	
   were	
  
areas	
   within	
   the	
   proliferative	
   urothelium	
   with	
   no	
   immunopositive	
   cells	
   in	
   at	
   least	
   2	
  
contiguous	
  high	
  power	
  (400X)	
  fields.	
  	
  	
  
Immunoreactivity	
   for	
   COX-­‐2	
   and	
   activated	
   caspase	
   3	
   was	
   classified	
   according	
   to	
  
overall	
   pattern	
   within	
   the	
   urothelium.	
   Cells	
   immunoreactive	
   for	
   COX-­‐2	
   had	
   diffuse	
  
cytoplasmic	
   to	
   perinuclear	
   labeling.	
   Cells	
   immunoreactive	
   for	
   activated	
   caspase	
   3	
   had	
  
diffuse	
   or,	
   more	
   often,	
   finely	
   granular	
   labeling	
   within	
   the	
   cytoplasm	
   and/or	
   nucleus.	
   In	
  
pattern	
   1	
   for	
   these	
   markers,	
   immunoreactivity	
   was	
   limited	
   to	
   the	
   superficial	
   1-­‐3	
   cell	
   layers	
  
(Fig.	
  7).	
  In	
  Pattern	
  2,	
  cells	
  were	
  labeled	
  throughout	
  the	
  full	
  thickness	
  of	
  the	
  urothelium	
  (Fig.	
  
8).	
   In	
   pattern	
   3,	
   immunoreactivity	
   was	
   patchy	
   and	
   randomly	
   distributed	
   throughout	
   the	
  
proliferative	
   urothelium,	
   but	
   greater	
   than	
   15%	
   of	
   neoplastic	
   cells	
   were	
   positively	
   labeled	
  
(Fig.	
  9).	
  	
  
For	
   each	
   marker,	
   any	
   section	
   that	
   completely	
   lacked	
   immunoreactivity	
   was	
  
excluded	
   from	
   analysis,	
   as	
   it	
   could	
   not	
   be	
   determined	
   whether	
   this	
   was	
   true	
   loss	
   of	
  
expression	
  or	
  was	
  artifactual.	
  	
  
Statistical	
  Analysis	
  
Statistical	
  Analysis	
  Software	
  (SAS)	
  version	
  9.1.3	
  (2002,	
  SAS	
  Institute	
  Inc,	
  Cary,	
  NC)	
  
was	
  used	
  for	
  the	
  data	
  analysis.	
  Fisher	
  exact	
  test	
  was	
  used	
  to	
  test	
  the	
  association	
  between	
  

	
  42	
  

	
  
grade	
   or	
   degree	
   of	
   infiltration	
   and	
   the	
   pattern	
   of	
   immunoreactivity	
   for	
   all	
   evaluated	
  
markers	
  and	
  loss	
  of	
  immunoreactivity	
  for	
  UPIII	
  and	
  CK7.	
  For	
  all	
  statistical	
  analyses,	
  lesions	
  
were	
   categorized	
   as	
   follows:	
   non-­‐neoplastic	
   lesions	
   (urothelial	
   polyps	
   and	
   polypoid	
  
cystitis),	
   low-­‐grade	
   neoplasms	
   (urothelial	
   papillomas,	
   papillary	
   urothelial	
   neoplasms	
   of	
  
low	
  malignant	
  potential,	
  and	
  grade	
  1	
  urothelial	
  carcinomas),	
  grade	
  2	
  urothelial	
  carcinomas,	
  
and	
  grade	
  3	
  urothelial	
  carcinomas.	
  Significance	
  was	
  set	
  at	
  P=0.05.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  43	
  

	
  
Results	
  
Immunohistochemistry:	
  UPIII	
  and	
  CK7	
  
The	
   urothelium	
   of	
   control	
   urinary	
   bladders	
   and	
   areas	
   of	
   normal	
   urothelium	
   in	
  
tumor	
  samples	
  demonstrated	
  strong	
  expression	
  of	
  UPIII	
  and	
  CK7	
  diffusely	
  throughout	
  the	
  
superficial	
   layers	
   (umbrella	
   cells).	
   This	
   distribution	
   of	
   immunolabeling	
   typified	
   UPIII	
   and	
  
CK7	
  pattern	
  1.	
  	
  
In	
   the	
   hyperplastic	
   urothelium	
   of	
   polyps	
   and	
   polypoid	
   cystitis,	
   expression	
   of	
   UPIII	
  
and	
  CK7	
  was	
  generally	
  limited	
  to	
  the	
  superficial	
  (umbrella)	
  cell	
  layers	
  (UPIII	
  pattern	
  1;	
  CK7	
  
pattern	
   1)	
   or	
   extended	
   only	
   to	
   the	
   mid-­‐portion	
   of	
   the	
   urothelium	
   (UPIII	
   pattern	
   2).	
   Only	
  
one	
   urothelial	
   polyp	
   had	
   UPIII	
   pattern	
   4	
   and	
   none	
   had	
   CK7	
   pattern	
   4.	
   The	
   PUNLMP	
   and	
  
urothelial	
   papillomas	
   demonstrated	
   UPIII	
   pattern	
   3	
   and	
   CK7	
   pattern	
   2	
   or	
   3.	
   With	
   few	
  
exceptions,	
   papillary	
   carcinomas	
   of	
   all	
   grades	
   demonstrated	
   patchy	
   randomly	
   distributed	
  
UPIII	
  expression	
  consistent	
  with	
  pattern	
  4.	
  Partial	
  loss	
  of	
  UPIII	
  expression	
  was	
  not	
  detected	
  
in	
  any	
  grade	
  I	
  papillary	
  carcinoma	
  and	
  in	
  only	
  one	
  grade	
  II	
  papillary	
  carcinoma.	
  In	
  contrast,	
  
14/33	
   (42%)	
   grade	
   III	
   papillary	
   carcinomas	
   had	
   partial	
   loss	
   of	
   expression	
   of	
   UPIII.	
   More	
  
grade	
   II	
   and	
   III	
   carcinomas	
   had	
   a	
   CK7	
   pattern	
   3,	
   than	
   had	
   a	
   UPIII	
   pattern	
   3	
   (11/50	
   CK7	
  
pattern	
   3	
   cases	
   compared	
   to	
   3/50	
   UPIII	
   pattern	
   3	
   cases).	
   The	
   majority	
   (71%)	
   of	
   grade	
   II	
  
and	
   III	
   urothelial	
   carcinomas,	
   however,	
   had	
   a	
   CK7	
   pattern	
   4	
   similar	
   to	
   that	
   observed	
   for	
  
UPIII.	
   Only	
   grades	
   II	
   and	
   III	
   urothelial	
   carcinomas	
   had	
   significant	
   loss	
   of	
   CK7	
   expression	
  
(Tables	
  2	
  and	
  3).	
  
Immunohistochemistry:	
  COX-­‐2	
  	
  
The	
   squamous	
   cell	
   carcinoma	
   from	
   the	
   digit	
   of	
   a	
   dog	
   used	
   as	
   a	
   positive	
   control	
  
demonstrated	
   strong	
   positive	
   cytoplasmic	
   and	
   mainly	
   perinuclear	
   immunoreactivity	
   for	
  

	
  44	
  

	
  
COX-­‐2	
  in	
  30%	
  of	
  neoplastic	
  cells.	
  There	
  was	
  no	
  positive	
  immunoreactivity	
  for	
  COX-­‐2	
  in	
  any	
  
part	
   of	
   the	
   normal	
   urothelium	
   from	
   control	
   urinary	
   bladders;	
   however,	
   in	
   the	
   superficial	
  
layers	
  of	
  non-­‐proliferative	
  urothelium	
  adjacent	
  to	
  proliferative	
  lesions,	
  there	
  were	
  often	
  a	
  
small	
  percentage	
  (less	
  than	
  10%)	
  of	
  COX-­‐2	
  positive	
  cells.	
  	
  
In	
   all	
   proliferative	
   lesions,	
   at	
   least	
   10%	
   of	
   proliferative	
   urothelial	
   cells	
   exhibited	
  
positive	
   COX-­‐2	
   immunoreactivity.	
   In	
   urothelial	
   polyps	
   and	
   polypoid	
   cystitis,	
   COX-­‐2	
  
expression	
   in	
   the	
   proliferative	
   urothelium	
   was	
   usually	
   restricted	
   to	
   the	
   superficial	
   layers	
  
(COX-­‐2	
  pattern	
  1)	
  or	
  was	
  less	
  commonly	
  full	
  thickness	
  (COX-­‐2	
  pattern	
  2).	
  Only	
  one	
  polyp	
  
had	
   randomly	
   distributed	
   and	
   patchy	
   expression	
   of	
   COX-­‐2	
   (COX-­‐2	
   pattern	
   3).	
   The	
   PUNLMP	
  
and	
   the	
   two	
   papillomas	
   demonstrated	
   COX-­‐2	
   pattern	
   1	
   while	
   the	
   grade	
   I	
   papillary	
  
carcinomas	
   exhibited	
   either	
   pattern	
   2	
   or	
   3.	
   All	
   high-­‐grade	
   (grades	
   II	
   and	
   III)	
   papillary	
  
urothelial	
  carcinomas	
  in	
  which	
  COX-­‐2	
  expression	
  was	
  detected	
  exhibited	
  COX-­‐2	
  pattern	
  3	
  
(Table	
  4).	
  
Immunohistochemistry:	
  Activated	
  caspase	
  3	
  
Approximately	
   10%	
   of	
   cells	
   in	
   the	
   lymph	
   node	
   used	
   as	
   a	
   positive	
   control	
   for	
  
activated	
  caspase	
  3	
  had	
  positive	
  finely	
  granular	
  labeling	
  of	
  the	
  cytoplasm	
  and/or	
  nucleus.	
  
In	
  the	
  normal	
  urothelium	
  of	
  control	
  urinary	
  bladders	
  and	
  adjacent	
  to	
  proliferative	
  lesions,	
  
immunoreactive	
   cells	
   comprised	
   less	
   than	
   10%	
   of	
   the	
   total	
   urothelium	
   and	
   were	
   largely	
  
limited	
   to	
   the	
   superficial	
   1-­‐2	
   cell	
   layers.	
   The	
   total	
   percentage	
   of	
   caspase	
   3	
   immunoreactive	
  
cells	
   ranged	
   from	
   10-­‐40%	
   in	
   all	
   other	
   proliferative	
   lesions;	
   however,	
   there	
   was	
   no	
  
appreciable	
   variation	
   in	
   the	
   percentage	
   of	
   positive	
   cells	
   between	
   different	
   tumor	
   types.	
  
Immunoreactivity	
   for	
   activated	
   caspase	
   3	
   was	
   most	
   commonly	
   noted	
   in	
   the	
   superficial	
  
most	
   cell	
   layers	
   of	
   all	
   proliferative	
   lesions	
   (caspase	
   3	
   pattern	
   1),	
   but	
   was	
   also	
   seen	
   both	
  

	
  45	
  

	
  
diffusely	
   throughout	
   the	
   urothelium	
   (caspase	
   3	
   pattern	
   2)	
   and	
   in	
   a	
   randomly	
   distributed	
  
patchy	
  distribution	
  (caspase	
  3	
  pattern	
  3)(Table	
  5).	
  	
  
A	
   summary	
   of	
   the	
   predominant	
   patterns	
   of	
   immunoreactivity	
   for	
   UPIII,	
   CK7,	
   and	
  
COX-­‐2	
  observed	
  in	
  each	
  proliferative	
  urothelial	
  lesion	
  classification	
  is	
  presented	
  in	
  Table	
  6.	
  
Statistical	
  Analysis	
  
Using	
   Fisher’s	
   exact	
   test	
   there	
   were	
   significant	
   associations	
   between	
   tumor	
  
classification	
   and	
   overall	
  UPIII	
   pattern	
   (P=1.49x10-­‐18),	
   loss	
   of	
   UPIII	
   (P=1.27x10-­‐4),	
   overall	
  
CK7	
  pattern	
  (P=4.34x10-­‐18),	
  and	
  COX-­‐2	
  pattern	
  (P=8.12x10-­‐25).	
  Also	
  by	
  Fisher’s	
  exact	
  test,	
  
there	
   were	
   significant	
   associations	
   between	
   depth	
   of	
   neoplastic	
   cell	
   infiltration	
   and	
   overall	
  
UPIII	
   pattern	
   (P=1.54x10-­‐14),	
   loss	
   of	
   UPIII	
   (P=2.07x10-­‐4),	
   overall	
   CK7	
   pattern	
   (P=1.17x10-­‐
13),	
   loss	
   of	
   CK7	
   expression	
   (P=0.0485),	
   and	
   COX-­‐2	
   pattern	
   (P=8.23x10-­‐21).	
   There	
   were	
   no	
  

significant	
  associations	
  between	
  tumor	
  classification	
  and	
  loss	
  of	
  CK7	
  or	
  caspase-­‐3	
  pattern.	
  	
  
There	
   were	
   also	
   no	
   significant	
   associations	
   between	
   depth	
   of	
   infiltration	
   and	
   caspase-­‐3	
  
expression.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  46	
  

	
  
Discussion	
  
In	
   the	
   current	
   study,	
   strong	
   differences	
   in	
   expression	
   of	
   UPIII,	
   CK7,	
   and	
   COX-­‐2	
   were	
  
observed	
   between	
   urothelial	
   polyps	
   and	
   cases	
   of	
   polypoid	
   cystitis,	
   urothelial	
   papillomas	
  
and	
   papillary	
   urothelial	
   neoplasms	
   of	
   low	
   malignant	
   potential,	
   and	
   papillary	
   carcinomas.	
  
However,	
  no	
  association	
  was	
  observed	
  between	
  the	
  expression	
  of	
  activated	
  caspase-­‐3	
  and	
  
tumor	
  classification,	
  grade,	
  or	
  depth	
  of	
  infiltration	
  into	
  the	
  urinary	
  bladder	
  wall.	
  
Overall,	
   urothelial	
   polyps	
   and	
   cases	
   of	
   polypoid	
   cystitis	
   predominately	
   had	
  
expression	
   of	
   UPIII	
   and	
   CK7	
   that	
   either	
   was	
   limited	
   to	
   the	
   superficial-­‐most	
   cell	
   layers	
  
(UPIII/CK7	
   pattern	
   1)	
   or	
   extended	
   to	
   the	
   middle-­‐most	
   cell	
   layer	
   (UPIII/CK7	
   pattern	
   2).	
  
Urothelial	
   papillomas	
   and	
   a	
   papillary	
   urothelial	
   neoplasm	
   of	
   low	
   malignant	
   potential	
   had	
  
UPIII	
  and	
  CK7	
  expression	
  throughout	
  the	
  full	
  thickness	
  of	
  the	
  urothelium	
  or	
  extending	
  at	
  
least	
  up	
  to	
  the	
  middle-­‐most	
  cell	
  layer.	
  The	
  majority	
  of	
  papillary	
  carcinomas,	
  in	
  contrast,	
  had	
  
randomly	
  distributed	
   patchy	
   immunoreactivity	
  for	
  UPIII	
  and	
  CK7	
  (UPIII/CK7	
  pattern	
  4).	
  In	
  
addition,	
  there	
  was	
  often	
  partial	
  loss	
  of	
  UPIII	
  and	
  CK7	
  expression	
  in	
  high-­‐grade	
  carcinomas	
  
that	
  invaded	
  into	
  the	
  urinary	
  bladder	
  wall.	
  	
  
Based	
   on	
   the	
   association	
   with	
   infiltration	
   into	
   the	
   urinary	
   bladder	
   wall,	
   loss	
   of	
   UPIII	
  
and	
   CK7	
   in	
   urothelial	
   carcinomas	
   may	
   suggest	
   a	
   lack	
   of	
   differentiation	
   or	
   epithelial-­‐
mesenchymal	
   transition	
   in	
   that	
   favors	
   infiltration.	
   In	
   humans,	
   loss	
   of	
   UPIII	
   expression	
   is	
  
frequently	
   seen	
   in	
   metastases	
   of	
   high-­‐grade	
   tumors	
   and	
   has	
   been	
   associated	
   with	
  
lymphovascular	
   infiltration,	
   stage,	
   and	
   grade.13,25	
   Loss	
   of	
   cell	
   membrane	
   adhesion	
  
molecules	
   may	
   also	
   result	
   in	
   an	
   increased	
   propensity	
   for	
   metastasis	
   in	
   dogs;	
   however,	
  
information	
   on	
   clinical	
   outcome	
   including	
   presence	
   of	
   metastasis	
   and	
   survival	
   was	
   not	
  
available	
  for	
  the	
  examined	
  canine	
  tumor	
  set.	
  The	
  observed	
  significant	
  loss	
  of	
  UPIII	
  and	
  CK7	
  

	
  47	
  

	
  
in	
   many	
   high-­‐grade	
   carcinomas	
   suggests	
   that	
   when	
   used	
   as	
   diagnostic	
   markers	
   on	
   small	
  
specimens,	
  some	
  urothelial	
  carcinomas	
  might	
  not	
  be	
  positively	
  labeled.	
  
There	
   was	
   a	
   clear	
   difference	
   in	
   the	
   location	
   of	
   COX-­‐2-­‐positive	
   cells	
   within	
   the	
  
urothelium	
   between	
   non-­‐neoplastic	
   and	
   neoplastic	
   lesions	
   as	
   well	
   as	
   between	
   papillary	
  
urothelial	
   carcinomas	
   with	
   different	
   degrees	
   of	
   infiltration	
   into	
   the	
   urinary	
   bladder	
   wall.	
  
Urothelial	
   polyps,	
   cases	
   of	
   polypoid	
   cystitis,	
   urothelial	
   papillomas,	
   and	
   the	
   one	
   PUNLMP	
  
exhibited	
   expression	
   of	
   COX-­‐2	
   in	
   the	
   superficial	
   layers	
   of	
   the	
   urothelium	
   (COX-­‐2	
   pattern	
   1)	
  
or	
   rarely	
   diffusely	
   throughout	
   the	
   urothelium	
   (COX-­‐2	
   pattern	
   2).	
   In	
   contrast,	
   papillary	
  
carcinomas	
  had	
  a	
  randomly	
  distributed,	
  patchy	
  pattern	
  of	
  COX-­‐2	
  expression	
  (COX-­‐2	
  pattern	
  
3).	
  
Previous	
  studies	
  of	
  COX-­‐2	
  in	
  canine	
  urinary	
  bladder	
  urothelium	
  have	
  focused	
  on	
  the	
  
difference	
  in	
  COX-­‐2	
  expression	
  between	
  normal	
  urinary	
  bladder	
  and	
  urothelial	
  carcinomas,	
  
and	
   reported	
   differences	
   as	
   a	
   percentage	
   of	
   COX-­‐2	
   positive	
   cells.	
   In	
   the	
   current	
   study,	
   COX-­‐
2	
  immunoreactivity	
  was	
  observed	
  in	
  the	
  majority	
  of	
  non-­‐neoplastic	
  and	
  neoplastic	
  lesions,	
  
often	
   in	
   a	
   relatively	
   high	
   percentage	
   of	
   cells,	
   precluding	
   the	
   use	
   of	
   percentage	
   of	
   positive	
  
cells	
  as	
  a	
  differentiating	
  criterion.	
  In	
  one	
  previous	
  immunohistochemical	
  study,	
  only	
  30/52	
  
(58%)	
   canine	
   TCC	
   demonstrated	
   COX-­‐2	
   expression	
   by	
   IHC.22	
   In	
   another	
   study	
   of	
  18	
   canine	
  
TCC,	
  all	
  tumors	
  expressed	
  COX-­‐2	
  and	
  three	
  showed	
  more	
  than	
  30%	
  of	
  cells	
  to	
  be	
  positive	
  
for	
   COX-­‐2.20	
   The	
   reasons	
   for	
   discrepancy	
   between	
   these	
   studies	
   in	
   terms	
   of	
   the	
   numbers	
   of	
  
cases	
  and	
  number	
  of	
  cells	
  within	
  a	
  given	
  case	
  that	
  are	
  positively	
  labeled	
  for	
  COX-­‐2	
  may	
  be	
  
due	
  to	
  the	
  low	
  number	
  of	
  cases	
  examined	
  in	
  each	
  study,	
  differing	
  methods	
  of	
  fixation	
  and	
  
processing	
  of	
  the	
  tissues,	
  differences	
  in	
  immunohistochemical	
  methods	
  including	
  the	
  anti-­‐
COX-­‐2	
  antibody	
  used,	
  or	
  inter-­‐observer	
  variation.	
  	
  

	
  48	
  

	
  
To	
   our	
   knowledge,	
   COX-­‐2	
   expression	
   has	
   not	
   been	
   previously	
   evaluated	
   in	
   non-­‐
neoplastic	
   proliferative	
   urothelial	
   lesions	
   of	
   the	
   urinary	
   bladder	
   in	
   dogs.	
   In	
   contrast	
   to	
  
normal	
  urothelium	
  from	
  urinary	
  bladders	
  used	
  as	
  controls,	
  all	
  urothelial	
  polyps	
  and	
  cases	
  
of	
   polypoid	
   cystitis	
   had	
   expression	
   of	
   COX-­‐2.	
   	
   Urothelial	
   polyps	
   and	
   polypoid	
   cystitis	
   are	
  
often	
  associated	
  with	
  mucosal	
  irritation	
  such	
  as	
  might	
  occur	
  with	
  cystoliths.	
  Expression	
  of	
  
COX-­‐2	
   often	
   throughout	
   the	
   superficial	
   portion	
   or	
   full	
   thickness	
   of	
   the	
   proliferative	
  
urothelium	
   of	
   non-­‐neoplastic	
   lesions	
   observed	
   in	
   the	
   current	
   study	
   might	
   represent	
   a	
  
consequence	
  of	
  such	
  surface	
  irritation.	
  	
  
Overall,	
   there	
   were	
   distinct	
   differences	
   in	
   the	
   patterns	
   of	
   UPIII,	
   CK7,	
   and	
   COX-­‐2	
  
expression	
  in	
  canine	
  urothelial	
  proliferative	
  lesions	
  of	
  the	
  urinary	
  bladder	
  as	
  defined	
  and	
  
categorized	
  by	
  the	
  WHO/ISUP	
  Consensus	
  Classification	
  System.	
  As	
  the	
  cases	
  included	
  in	
  the	
  
study	
  set	
  were	
  comprised	
  of	
  diagnostic	
  samples,	
  reliable	
  follow-­‐up	
  data	
  regarding	
  clinical	
  
outcome	
   was	
   not	
   available.	
   Definitive	
   prospective	
   studies	
   of	
   the	
   clinical	
   outcome	
   using	
  
these	
   markers	
   and	
   the	
   WHO/ISUP	
   Consensus	
   Classification	
   System	
   in	
   dogs	
   remain	
   lacking,	
  
but	
   this	
   study	
   encourages	
   the	
   continued	
   prognostic	
   evaluation	
   of	
   the	
   WHO/ISUP	
  
Consensus	
   Classification	
   system	
   and	
   of	
   these	
   immunohistochemical	
   markers	
   in	
   canine	
  
proliferative	
  urothelial	
  lesions.	
  
Acknowledgements	
  
The	
   work	
   described	
   in	
   this	
   manuscript	
   comprised	
   a	
   portion	
   of	
   Dr.	
   Sledge’s	
   PhD	
  
graduate	
   program	
   for	
   which	
   a	
   fellowship	
   was	
   provided	
   from	
   Bristol-­‐Meyers	
   Squibb	
  
through	
   the	
   American	
   College	
   of	
   Veterinary	
   Pathologists/Society	
   of	
   Toxicological	
  
Pathologists	
  Coalition.	
  
	
  

	
  49	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
APPENDIX	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  50	
  

	
  
Figure	
  1:	
  Urinary	
  Bladder;	
  Dog.	
  Hyperplastic	
  urotheluim	
  with	
  Uroplakin	
  III	
  (UPIII)	
  Pattern	
  1:	
  
Immunolabeling	
  (brown)	
  is	
  limited	
  to	
  the	
  superficial	
  1-­‐2	
  cell	
  layers.	
  3,3’-­‐Diaminobenzidine	
  
(DAB)	
  chromogen,	
  hematoxylin	
  counterstain.	
  
Figure	
   2:	
   Urinary	
   Bladder;	
   Dog.	
   Hyperplastic	
   urothelium	
   with	
   UPIII	
   IHC	
   Pattern	
   2:	
   UPIII	
   is	
  
expressed	
   strongly	
   by	
   all	
   but	
   the	
   most	
   basal	
   cell	
   layers.	
   DAB	
   chromogen,	
   hematoxylin	
  
counterstain.	
  
Figure	
   3:	
   Urinary	
   Bladder;	
   Dog.	
   Papillary	
   urothelial	
   carcinoma	
   grade	
   II	
   with	
   UPIII	
   IHC	
  
Pattern	
  4:	
  UPIII	
  expression	
  is	
  randomly	
  distributed	
  throughout	
  the	
  neoplasm.	
  Individual	
  or	
  
small	
   clusters	
   of	
   neoplastic	
   cells	
   strongly	
   express	
   UPIII,	
   while	
   numerous	
   islands	
   of	
  
neoplastic	
  cells	
  lack	
  expression.	
  DAB	
  chromogen,	
  hematoxylin	
  counterstain.	
  

	
  
	
  
	
  
	
  
	
  

	
  51	
  

	
  
Figure	
   4:	
   Urinary	
   Bladder;	
   Dog.	
   Hyperplastic	
   urotheluim	
   with	
   Cytokeratin	
   7	
   (CK7)	
   IHC	
  
Pattern	
   1:	
   Strong	
   expression	
   of	
   CK7	
   is	
   limited	
   to	
   the	
   superficial	
   1-­‐2	
   cell	
   layers.	
   DAB	
  
chromogen,	
  hematoxylin	
  counterstain.	
  
Figure	
  5:	
  Urinary	
  Bladder;	
  Dog.	
  Papillary	
  urothelial	
  carcinoma	
  grade	
  II	
  with	
  CK7	
  IHC	
  Pattern	
  
3:	
  CK7	
  is	
  strongly	
  expressed	
  by	
  all	
  cell	
  layers.	
  DAB	
  chromogen,	
  hematoxylin	
  counterstain.	
  
Figure	
  6:	
  Urinary	
  Bladder;	
  Dog.	
  Papillary	
  urothelial	
  carcinoma	
  grade	
  II	
  with	
  CK7	
  IHC	
  Pattern	
  
4:	
  CK7	
  expression	
  is	
  patchy.	
  While	
  most	
  cells	
  have	
  strong	
  expression	
  of	
  CK7,	
  individual	
  or	
  
small	
  groups	
  of	
  neoplastic	
  cells	
  randomly	
  distributed	
  throughout	
  the	
  mass	
  lack	
  expression.	
  
DAB	
  chromogen,	
  hematoxylin	
  counterstain.	
  
	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  52	
  

	
  
Figure	
   7:	
   Urinary	
   Bladder;	
   Dog.	
   Hyperplastic	
   urotheluim	
   with	
   Cyclooxygenase-­‐2	
   (COX-­‐2)	
   IHC	
  
Pattern	
   1:	
   Immunolabeling	
   for	
   COX-­‐2	
   is	
   limited	
   to	
   the	
   superficial	
   1-­‐2	
   cell	
   layers.	
   DAB	
  
chromogen,	
  hematoxylin	
  counterstain.	
  	
  
Figure	
   8:	
   Urinary	
   Bladder;	
   Dog.	
   Hyperplastic	
   urotheluim	
   with	
   COX-­‐2	
   IHC	
   Pattern	
   2:	
   COX-­‐2	
   is	
  
strongly	
   expressed	
   by	
   cells	
   throughout	
   all	
   cell	
   layers.	
   DAB	
   chromogen,	
   hematoxylin	
  
counterstain.	
  
Figure	
   9:	
   Urinary	
   Bladder;	
   Dog.	
   Papillary	
   urothelial	
   carcinoma	
   grade	
   II	
   with	
   COX-­‐2	
   IHC	
  
Pattern	
   3:	
   Cells	
   that	
   strongly	
   express	
   COX-­‐2	
   are	
   randomly	
   distributed	
   throughout	
   the	
  
neoplastic	
  cell	
  population.	
  DAB	
  chromogen,	
  hematoxylin	
  counterstain.	
  
	
  

	
  53	
  

	
  
Table	
   1:	
   Histologic	
   Features	
   of	
   Proliferative	
   Urothelial	
   Lesions	
   according	
   to	
   the	
   WHO/ISUP	
  
Consensus	
  Classification	
  System*	
  
Classification	
  

Mitoses	
  

Histologic	
  Characteristics	
  

Non-­‐neoplastic	
  
Lesions	
  
Polyp/Polypoid	
  
Cystitis	
  

	
  

	
  	
  

Rare	
  and	
  confined	
  to	
  
the	
  basal	
  cell	
  layers	
  

Neoplastic	
  Lesions	
  

	
  

Exophytic	
  protrusions	
  of	
  mucosa	
  and	
  
supporting	
  fibrovascular	
  stroma	
  lacking	
  true	
  
papillary	
  fronds;	
  often	
  associated	
  with	
  
stromal	
  edema	
  and	
  inflammation;	
  may	
  occur	
  
as	
  single	
  fibroepithelial	
  polyp	
  or	
  in	
  multiples	
  
as	
  polypoid	
  cystitis	
  
	
  	
  

Urothelial	
  
Papilloma	
  

Rare	
  and	
  confined	
  to	
  
the	
  basal	
  cell	
  layer	
  

PUNLMP**	
  

Rare	
  and	
  confined	
  to	
  
the	
  basal	
  cell	
  layer	
  

Papillary	
  
carcinoma	
  
	
  	
  	
  	
  	
  Grade	
  I	
  

	
  

<6	
  cell	
  layers	
  lining	
  papillary	
  fronds;	
  orderly	
  
arrangement	
  of	
  cells;	
  nuclei	
  are	
  uniform	
  in	
  
size,	
  shape,	
  and	
  chromatin	
  staining	
  
>6	
  cell	
  layers	
  lining	
  papillary	
  fronds;	
  orderly	
  
arrangement	
  of	
  cells;	
  nuclei	
  are	
  uniform	
  in	
  
size,	
  shape,	
  and	
  chromatin	
  staining	
  
	
  	
  

Infrequent	
  and	
  
Orderly	
  appearance	
  with	
  recognizable	
  
limited	
  to	
  the	
  basal	
  
variation	
  in	
  architectural	
  or	
  cytological	
  
1/2	
  of	
  the	
  epithelium	
   features	
  at	
  low	
  magnification;	
  mild	
  
anisokaryosis	
  with	
  variable	
  chromatin	
  
staining	
  
	
  	
  	
  	
  	
  Grade	
  II	
  
Low	
  to	
  moderate	
  
Overall	
  disorderly	
  appearance	
  with	
  
numbers	
  throughout	
   retainment	
  of	
  some	
  degree	
  of	
  polarity;	
  
all	
  levels	
  of	
  the	
  
irregular	
  clustering	
  and	
  disorganization	
  of	
  
urothelium	
  with	
  
cells;	
  moderate	
  anaplasia;	
  moderate	
  
possible	
  atypia	
  
anisokaryosis	
  with	
  prominent	
  nucleoli	
  and	
  
clumped	
  chromatin	
  
	
  	
  	
  	
  	
  Grade	
  III	
  
High	
  numbers	
  
Complete	
  loss	
  of	
  polarity;	
  irregular	
  clustering	
  
throughout	
  all	
  levels	
   and	
  disorganization	
  of	
  cells;	
  marked	
  
with	
  common	
  atypia	
   pleomorphism,	
  anisocytosis,	
  and	
  
anisokaryosis;	
  prominent	
  nucleoli	
  and	
  
clumped	
  chromatin	
  
*Adapted	
  from	
  descriptions	
  made	
  by	
  Patrick	
  et	
  al.	
  200636	
  
**Papillary	
  urothelial	
  neoplasm	
  of	
  low	
  malignant	
  potential	
  

	
  54	
  

	
  
Table	
  2:	
  Immunohistochemical	
  Scoring	
  of	
  Uroplakin	
  III	
  Expression	
  in	
  Canine	
  Proliferative	
  Urothelial	
  Lesions	
  by	
  Overall	
  Pattern	
  
Classification	
  

	
  

Pattern	
  1	
   Pattern	
  2	
  
Pattern	
  3	
  
19/44	
  
20/44	
  
Polyp/Polypoid	
  Cystitis	
  
(43%)	
  
(46%)	
  
3/44	
  (7%)	
  
Urothelial	
  Papilloma	
  
0	
  
0	
  
1/1	
  (100%)	
  
PUNLMP*	
  
0	
  
0	
  
1/1	
  (100%)	
  
Papillary	
  Carcinoma	
  
	
  	
  
	
  	
  
	
  	
  
	
  	
  	
  	
  Grade	
  I	
  
0	
  
0	
  
0	
  
	
  	
  	
  	
  Grade	
  II	
  
1/17	
  (6%)	
  
0	
  
1/17	
  (6%)	
  
	
  	
  	
  	
  Grade	
  III	
  
2/33	
  (6%)	
  
0	
  
2/33	
  (6%)	
  
*Papillary	
  urothelial	
  neoplasm	
  of	
  low	
  malignant	
  potential	
  

Pattern	
  4	
  

Loss	
  

1/44	
  (2%)	
  
1/44	
  (2%)	
  
0	
  
0	
  
0	
  
0	
  
	
  	
  
	
  	
  
2/2	
  (100%)	
  
0	
  
15/17	
  (88%)	
  
1/17	
  (6%)	
  
27/33	
  (82%)	
   14/33	
  (42%)	
  

	
  
	
  
	
  
Table	
  3:	
  Immunohistochemical	
  Scoring	
  of	
  Cytokeratin	
  7	
  Expression	
  in	
  Canine	
  Proliferative	
  Urothelial	
  Lesions	
  by	
  Overall	
  Pattern	
  
Classification	
  
Pattern	
  1	
  
Pattern	
  2	
  
Pattern	
  3	
  
Pattern	
  4	
  
Polyp/Polypoid	
  Cystitis	
   16/44	
  (36%)	
   25/44	
  (57%)	
  
3/44	
  (7%)	
  
0	
  
Urothelial	
  Papilloma	
  
0	
  
1/2	
  (50%)	
  
1/2	
  (50%)	
  
0	
  
PUNLMP*	
  
0	
  
0	
   1/1	
  (100%)	
  
0	
  
Papillary	
  Carcinoma	
  
	
  	
  
	
  	
  
	
  	
  
	
  	
  
	
  	
  	
  	
  Grade	
  I	
  
0	
  
1/2	
  (50%)	
  
0	
  
1/2	
  (50%)	
  
	
  	
  	
  	
  Grade	
  II	
  
2/17	
  (12%)	
  
0	
   5/17	
  (29%)	
   10/17	
  (59%)	
  
	
  	
  	
  	
  Grade	
  III	
  
0	
  
1/33	
  (3%)	
   6/33	
  (18%)	
   25/33	
  (76%)	
  
*Papillary	
  urothelial	
  neoplasm	
  of	
  low	
  malignant	
  potential	
  

	
  55	
  

Loss	
  
0	
  
0	
  
0	
  
	
  	
  
0	
  
2/17	
  (12%)	
  
1/33	
  (3%)	
  

	
  
Table	
  4:	
  Immunohistochemical	
  Scoring	
  of	
  COX-­‐2	
  Expression	
  in	
  Canine	
  Proliferative	
  Urothelial	
  
Lesions	
  by	
  Overall	
  Pattern	
  
Classification	
  
Pattern	
  1	
  
Pattern	
  2	
  
Polyp/Polypoid	
  Cystitis	
  
32/44	
  (73%)	
   11/44	
  (25%	
  
Urothelial	
  Papilloma	
  
1/2	
  (50%)	
  
1/2	
  (50%)	
  
PUNLMP*	
  
1/1	
  (100%)	
  
0	
  
Papillary	
  Carcinoma	
  
	
  	
  
	
  	
  
	
  	
  	
  	
  Grade	
  I	
  
1/2	
  (50%)	
  
0	
  
	
  	
  	
  	
  Grade	
  II	
  
0	
  
0	
  
	
  	
  	
  	
  Grade	
  III	
  
0	
  
0	
  
*Papillary	
  urothelial	
  neoplasm	
  of	
  low	
  malignant	
  potential	
  

Pattern	
  3	
  
1/44	
  (2%)	
  
0	
  
0	
  
	
  	
  
1/2	
  (50%)	
  
17/17	
  (100%)	
  
33/33	
  (100%)	
  

	
  
	
  
	
  
	
  
	
  
	
  
Table	
   5:	
   Immunohistochemical	
   Scoring	
   of	
   Activated	
   Caspase-­‐3	
   in	
   Canine	
   Proliferative	
  
Urothelial	
  Lesions	
  by	
  Overall	
  Pattern	
  

	
  

Classification	
  
Pattern	
  1	
  
Pattern	
  2	
  
Pattern	
  3	
  
Polyp/Polypoid	
  Cystitis	
   27/44	
  (61%)	
   13/44	
  (30%)	
   4/44	
  (9%)	
  
Urothelial	
  Papilloma	
  
1/2	
  (50%)	
  
0	
   1/2	
  (50%)	
  
PUNLMP*	
  
1/1	
  (100%)	
  
0	
  
0	
  
Papillary	
  Carcinoma	
  
	
  	
  
	
  	
  
	
  	
  
	
  	
  	
  	
  Grade	
  I	
  
1/2	
  (50%)	
  
0	
   1/2	
  (50%)	
  
	
  	
  	
  	
  Grade	
  II	
  
10/16	
  (63%)	
   3/16	
  (19%)	
   3/16	
  (19%)	
  
	
  	
  	
  	
  Grade	
  III	
  
22/30	
  (73%)	
   4/30	
  (13%)	
   4/30	
  (13%)	
  
*Papillary	
  urothelial	
  neoplasm	
  of	
  low	
  malignant	
  potential	
  

	
  

	
  

56	
  

	
  
Table	
  6:	
  Summary	
  of	
  Patterns	
  of	
  Immunoreactivity	
  of	
  Uroplakin	
  III,	
  Cytokeratin	
  7,	
  and	
  COX-­‐2	
  
in	
  Proliferative	
  Urothelial	
  Lesions	
  
Classification	
  
Uroplakin	
  III	
  
Polyp/Polypoi 89%	
  had	
  
d	
  Cystitis	
  
expression	
  limited	
  
to	
  cells	
  in	
  the	
  
superficial-­‐most	
  
cell	
  layer	
  (pattern	
  
1)	
  or	
  extending	
  to	
  
the	
  middle-­‐most	
  
layer	
  	
  (pattern	
  2)	
  
Urothelial	
  
All	
  cases	
  had	
  
Papilloma/PU expression	
  
NLMP*	
  
throughout	
  all	
  cell	
  
layers	
  (pattern	
  3)	
  

Papillary	
  
Carcinoma	
  

	
  	
  	
  	
  Grade	
  I	
  

Cytokeratin	
  7	
  
93%	
  had	
  expression	
  
limited	
  to	
  cells	
  in	
  the	
  
superficial-­‐most	
  cell	
  
layer	
  (pattern	
  1)	
  or	
  
extending	
  to	
  the	
  
middle-­‐most	
  layer	
  
(pattern	
  2)	
  

COX-­‐2	
  
98%	
  had	
  expression	
  
limited	
  to	
  the	
  
superficial	
  1/3	
  of	
  cell	
  
layers	
  (pattern	
  1)	
  or	
  
throughout	
  all	
  cell	
  
layers	
  (pattern	
  2)	
  

All	
  cases	
  had	
  
expression	
  that	
  
extending	
  to	
  the	
  
middle-­‐most	
  layer	
  of	
  
the	
  urotheluim	
  
(pattern	
  2)	
  or	
  
throughout	
  all	
  cell	
  
layers	
  (pattern	
  3)	
  
83%	
  had	
  randomly	
   21%	
  had	
  expression	
  
distributed,	
  patchy	
   throughout	
  all	
  	
  cell	
  
expression	
  (pattern	
   layers	
  (pattern	
  3);	
  
4)	
  
68%	
  had	
  randomly	
  
distributed,	
  patchy	
  
expression	
  (pattern	
  
4)	
  
No	
  significant	
  loss	
   No	
  significant	
  loss	
  of	
  
of	
  expression	
  was	
  
expression	
  was	
  
observed	
  
observed	
  

All	
  cases	
  	
  had	
  
expression	
  limited	
  to	
  
the	
  superficial	
  1/3	
  of	
  
cell	
  layers	
  (pattern	
  1)	
  
or	
  throughout	
  all	
  cell	
  
layers	
  (pattern	
  2)	
  

	
  	
  	
  	
  Grade	
  II	
  

6%	
  	
  had	
  significant	
   12%	
  had	
  significant	
  
areas	
  of	
  expression	
   areas	
  of	
  expression	
  
loss	
  
loss	
  

	
  	
  	
  	
  Grade	
  III	
  

42%	
  had	
  
significant	
  areas	
  of	
  
expression	
  loss	
  

3%	
  had	
  significant	
  
areas	
  of	
  expression	
  
loss	
  	
  

*Papillary	
  urothelial	
  neoplasm	
  of	
  low	
  malignant	
  potential	
  
	
  

57	
  

98%	
  had	
  randomly	
  
distributed	
  and	
  
patchy	
  expression	
  
(pattern	
  3)	
  

50%	
  had	
  expression	
  
limited	
  to	
  the	
  
superficial-­‐most	
  1/3	
  
of	
  cell	
  layers	
  (pattern	
  
1);	
  50%	
  had	
  
randomly	
  distributed,	
  
patchy	
  expression	
  
(pattern	
  3)	
  
All	
  cases	
  had	
  
randomly	
  distributed,	
  
patchy	
  expression	
  
(pattern	
  3)	
  
All	
  cases	
  had	
  
randomly	
  distributed,	
  
patchy	
  expression	
  
(pattern	
  3)	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
REFERENCES	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

58	
  

	
  
REFERENCES	
  
	
  
1.	
  
Abu-­‐Qare	
  AW,	
  Abou-­‐Donia	
  MB.	
  Biomarkers	
  of	
  apoptosis:	
  release	
  of	
  cytochrome	
  c,	
  
activation	
  of	
  caspase-­‐3,	
  induction	
  of	
  8-­‐hydroxy-­‐2'-­‐deoxyguanosine,	
  increased	
  3-­‐
nitrotyrosine,	
  and	
  alteration	
  of	
  p53	
  gene.	
  J	
  Toxicol	
  Environ	
  Health	
  B	
  Crit	
  Rev.	
  Jul-­‐Sep	
  
2001;4(3):313-­‐332.	
  
2.	
  
Alsheikh	
  A,	
  Mohamedali	
  Z,	
  Jones	
  E,	
  Masterson	
  J,	
  Gilks	
  CB.	
  Comparison	
  of	
  the	
  
WHO/ISUP	
  classification	
  and	
  cytokeratin	
  20	
  expression	
  in	
  predicting	
  the	
  behavior	
  of	
  low-­‐
grade	
  papillary	
  urothelial	
  tumors.	
  World/Health	
  Organization/Internattional	
  Society	
  of	
  
Urologic	
  Pathology.	
  Mod	
  Pathol.	
  Apr	
  2001;14(4):267-­‐272.	
  
3.	
  
Castelao	
  JE,	
  Yuan	
  JM,	
  Gago-­‐Dominguez	
  M,	
  Yu	
  MC,	
  Ross	
  RK.	
  Non-­‐steroidal	
  anti-­‐
inflammatory	
  drugs	
  and	
  bladder	
  cancer	
  prevention.	
  Br	
  J	
  Cancer.	
  Apr	
  2000;82(7):1364-­‐
1369.	
  
4.	
  
Cohen	
  GM.	
  Caspases:	
  the	
  executioners	
  of	
  apoptosis.	
  Biochem	
  J.	
  Aug	
  15	
  1997;326	
  (	
  Pt	
  
1):1-­‐16.	
  
5.	
  
Dhawan	
  D,	
  Jeffreys	
  AB,	
  Zheng	
  R,	
  Stewart	
  JC,	
  Knapp	
  DW.	
  Cyclooxygenase-­‐2	
  
dependent	
  and	
  independent	
  antitumor	
  effects	
  induced	
  by	
  celecoxib	
  in	
  urinary	
  bladder	
  
cancer	
  cells.	
  Mol	
  Cancer	
  Ther.	
  Apr	
  2008;7(4):897-­‐904.	
  
6.	
  
Eble	
  J,	
  Sauter	
  G,	
  Epstein	
  JI,	
  Sesterhenn	
  IA.	
  World	
  Health	
  Organization	
  classification	
  
of	
  tumours.	
  Pathology	
  and	
  genetics	
  of	
  tumours	
  of	
  the	
  urinary	
  system	
  and	
  male	
  genital	
  
organs.	
  Lyon:	
  IARC	
  Press;	
  2004.	
  
7.	
  
Epstein	
  JI,	
  Amin	
  MB,	
  Reuter	
  VR,	
  Mostofi	
  FK.	
  The	
  World	
  Health	
  
Organization/International	
  Society	
  of	
  Urological	
  Pathology	
  consensus	
  classification	
  of	
  
urothelial	
  (transitional	
  cell)	
  neoplasms	
  of	
  the	
  urinary	
  bladder.	
  Bladder	
  Consensus	
  
Conference	
  Committee.	
  Am	
  J	
  Surg	
  Pathol.	
  Dec	
  1998;22(12):1435-­‐1448.	
  
8.	
  
Espinosa	
  de	
  los	
  Monteros	
  A,	
  Fernandez	
  A,	
  Millan	
  MY,	
  Rodriguez	
  F,	
  Herraez	
  P,	
  Martin	
  
de	
  las	
  Mulas	
  J.	
  Coordinate	
  expression	
  of	
  cytokeratins	
  7	
  and	
  20	
  in	
  feline	
  and	
  canine	
  
carcinomas.	
  Vet	
  Pathol.	
  May	
  1999;36(3):179-­‐190.	
  
9.	
  
Gately	
  S,	
  Li	
  WW.	
  Multiple	
  roles	
  of	
  COX-­‐2	
  in	
  tumor	
  angiogenesis:	
  a	
  target	
  for	
  
antiangiogenic	
  therapy.	
  Semin	
  Oncol.	
  Apr	
  2004;31(2	
  Suppl	
  7):2-­‐11.	
  
10.	
   Gee	
  J,	
  Lee	
  IL,	
  Jendiroba	
  D,	
  Fischer	
  SM,	
  Grossman	
  HB,	
  Sabichi	
  AL.	
  Selective	
  
cyclooxygenase-­‐2	
  inhibitors	
  inhibit	
  growth	
  and	
  induce	
  apoptosis	
  of	
  bladder	
  cancer.	
  Oncol	
  
Rep.	
  Feb	
  2006;15(2):471-­‐477.	
  

	
  

59	
  

	
  
11.	
   Greenhough	
  A,	
  Smartt	
  HJ,	
  Moore	
  AE,	
  et	
  al.	
  The	
  COX-­‐2/PGE2	
  pathway:	
  key	
  roles	
  in	
  
the	
  hallmarks	
  of	
  cancer	
  and	
  adaptation	
  to	
  the	
  tumour	
  microenvironment.	
  Carcinogenesis.	
  
Mar	
  2009;30(3):377-­‐386.	
  
12.	
   Grutter	
  MG.	
  Caspases:	
  key	
  players	
  in	
  programmed	
  cell	
  death.	
  Curr	
  Opin	
  Struct	
  Biol.	
  
Dec	
  2000;10(6):649-­‐655.	
  
13.	
   Gruver	
  AM,	
  Amin	
  MB,	
  Luthringer	
  DJ,	
  et	
  al.	
  Selective	
  immunohistochemical	
  markers	
  
to	
  distinguish	
  between	
  metastatic	
  high-­‐grade	
  urothelial	
  carcinoma	
  and	
  primary	
  poorly	
  
differentiated	
  invasive	
  squamous	
  cell	
  carcinoma	
  of	
  the	
  lung.	
  Arch	
  Pathol	
  Lab	
  Med.	
  Nov	
  
2012;136(11):1339-­‐1346.	
  
14.	
   Henry	
  CJ,	
  McCaw	
  DL,	
  Turnquist	
  SE,	
  et	
  al.	
  Clinical	
  evaluation	
  of	
  mitoxantrone	
  and	
  
piroxicam	
  in	
  a	
  canine	
  model	
  of	
  human	
  invasive	
  urinary	
  bladder	
  carcinoma.	
  Clin	
  Cancer	
  Res.	
  
Feb	
  2003;9(2):906-­‐911.	
  
15.	
   Karam	
  JA,	
  Lotan	
  Y,	
  Karakiewicz	
  PI,	
  et	
  al.	
  Use	
  of	
  combined	
  apoptosis	
  biomarkers	
  for	
  
prediction	
  of	
  bladder	
  cancer	
  recurrence	
  and	
  mortality	
  after	
  radical	
  cystectomy.	
  Lancet	
  
Oncol.	
  Feb	
  2007;8(2):128-­‐136.	
  
16.	
   Karamitopoulou	
  E,	
  Rentsch	
  CA,	
  Markwalder	
  R,	
  Vallan	
  C,	
  Thalmann	
  GN,	
  Brunner	
  T.	
  
Prognostic	
  significance	
  of	
  apoptotic	
  cell	
  death	
  in	
  bladder	
  cancer:	
  a	
  tissue	
  microarray	
  study	
  
on	
  179	
  urothelial	
  carcinomas	
  from	
  cystectomy	
  specimens.	
  Pathology.	
  Jan	
  2010;42(1):37-­‐
42.	
  
17.	
   Khan	
  KN,	
  Knapp	
  DW,	
  Denicola	
  DB,	
  Harris	
  RK.	
  Expression	
  of	
  cyclooxygenase-­‐2	
  in	
  
transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder	
  in	
  dogs.	
  Am	
  J	
  Vet	
  Res.	
  May	
  
2000;61(5):478-­‐481.	
  
18.	
   Knapp	
  DW,	
  Glickman	
  NW,	
  Denicola	
  DB,	
  Bonney	
  PL,	
  Lin	
  TL,	
  Glickman	
  LT.	
  Naturally-­‐
occurring	
  canine	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder	
  A	
  relevant	
  model	
  of	
  
human	
  invasive	
  bladder	
  cancer.	
  Urol	
  Oncol.	
  Mar-­‐Apr	
  2000;5(2):47-­‐59.	
  
19.	
   Knapp	
  DW,	
  Glickman	
  NW,	
  Widmer	
  WR,	
  et	
  al.	
  Cisplatin	
  versus	
  cisplatin	
  combined	
  
with	
  piroxicam	
  in	
  a	
  canine	
  model	
  of	
  human	
  invasive	
  urinary	
  bladder	
  cancer.	
  Cancer	
  
Chemother	
  Pharmacol.	
  2000;46(3):221-­‐226.	
  
20.	
   Knottenbelt	
  C,	
  Mellor	
  D,	
  Nixon	
  C,	
  Thompson	
  H,	
  Argyle	
  DJ.	
  Cohort	
  study	
  of	
  COX-­‐1	
  and	
  
COX-­‐2	
  expression	
  in	
  canine	
  rectal	
  and	
  bladder	
  tumours.	
  J	
  Small	
  Anim	
  Pract.	
  Apr	
  
2006;47(4):196-­‐200.	
  
21.	
   Komhoff	
  M,	
  Guan	
  Y,	
  Shappell	
  HW,	
  et	
  al.	
  Enhanced	
  expression	
  of	
  cyclooxygenase-­‐2	
  in	
  
high	
  grade	
  human	
  transitional	
  cell	
  bladder	
  carcinomas.	
  Am	
  J	
  Pathol.	
  Jul	
  2000;157(1):29-­‐35.	
  
22.	
   Lee	
  JY,	
  Tanabe	
  S,	
  Shimohira	
  H,	
  et	
  al.	
  Expression	
  of	
  cyclooxygenase-­‐2,	
  P-­‐glycoprotein	
  
and	
  multi-­‐drug	
  resistance-­‐associated	
  protein	
  in	
  canine	
  transitional	
  cell	
  carcinoma.	
  Res	
  Vet	
  
Sci.	
  Oct	
  2007;83(2):210-­‐216.	
  

	
  

60	
  

	
  
23.	
   Liu	
  B,	
  Qu	
  L,	
  Tao	
  H.	
  Cyclo-­‐oxygenase	
  2	
  up-­‐regulates	
  the	
  effect	
  of	
  multidrug	
  resistance.	
  
Cell	
  Biol	
  Int.	
  Jan	
  2010;34(1):21-­‐25.	
  
24.	
   Margulis	
  V,	
  Shariat	
  SF,	
  Ashfaq	
  R,	
  et	
  al.	
  Expression	
  of	
  cyclooxygenase-­‐2	
  in	
  normal	
  
urothelium,	
  and	
  superficial	
  and	
  advanced	
  transitional	
  cell	
  carcinoma	
  of	
  bladder.	
  J	
  Urol.	
  Mar	
  
2007;177(3):1163-­‐1168.	
  
25.	
   Matsumoto	
  K,	
  Satoh	
  T,	
  Irie	
  A,	
  et	
  al.	
  Loss	
  expression	
  of	
  uroplakin	
  III	
  is	
  associated	
  with	
  
clinicopathologic	
  features	
  of	
  aggressive	
  bladder	
  cancer.	
  Urology.	
  Aug	
  2008;72(2):444-­‐449.	
  
26.	
   Mitra	
  AP,	
  Castelao	
  JE,	
  Hawes	
  D,	
  et	
  al.	
  Combination	
  of	
  molecular	
  alterations	
  and	
  
smoking	
  intensity	
  predicts	
  bladder	
  cancer	
  outcome:	
  A	
  report	
  from	
  the	
  Los	
  Angeles	
  Cancer	
  
Surveillance	
  Program.	
  Cancer.	
  Feb	
  15	
  2013;119(4):756-­‐765.	
  
27.	
   Mohammed	
  SI,	
  Bennett	
  PF,	
  Craig	
  BA,	
  et	
  al.	
  Effects	
  of	
  the	
  cyclooxygenase	
  inhibitor,	
  
piroxicam,	
  on	
  tumor	
  response,	
  apoptosis,	
  and	
  angiogenesis	
  in	
  a	
  canine	
  model	
  of	
  human	
  
invasive	
  urinary	
  bladder	
  cancer.	
  Cancer	
  Res.	
  Jan	
  15	
  2002;62(2):356-­‐358.	
  
28.	
   Mohammed	
  SI,	
  Craig	
  BA,	
  Mutsaers	
  AJ,	
  et	
  al.	
  Effects	
  of	
  the	
  cyclooxygenase	
  inhibitor,	
  
piroxicam,	
  in	
  combination	
  with	
  chemotherapy	
  on	
  tumor	
  response,	
  apoptosis,	
  and	
  
angiogenesis	
  in	
  a	
  canine	
  model	
  of	
  human	
  invasive	
  urinary	
  bladder	
  cancer.	
  Mol	
  Cancer	
  Ther.	
  
Feb	
  2003;2(2):183-­‐188.	
  
29.	
   Mohammed	
  SI,	
  Dhawan	
  D,	
  Abraham	
  S,	
  et	
  al.	
  Cyclooxygenase	
  inhibitors	
  in	
  urinary	
  
bladder	
  cancer:	
  in	
  vitro	
  and	
  in	
  vivo	
  effects.	
  Mol	
  Cancer	
  Ther.	
  Feb	
  2006;5(2):329-­‐336.	
  
30.	
   Mohammed	
  SI,	
  Knapp	
  DW,	
  Bostwick	
  DG,	
  et	
  al.	
  Expression	
  of	
  cyclooxygenase-­‐2	
  (COX-­‐
2)	
  in	
  human	
  invasive	
  transitional	
  cell	
  carcinoma	
  (TCC)	
  of	
  the	
  urinary	
  bladder.	
  Cancer	
  Res.	
  
Nov	
  15	
  1999;59(22):5647-­‐5650.	
  
31.	
   Moll	
  R,	
  Laufer	
  J,	
  Wu	
  XR,	
  Sun	
  TT.	
  [Uroplakin	
  III,	
  a	
  specific	
  membrane	
  protein	
  of	
  
urothelial	
  umbrella	
  cells,	
  as	
  a	
  histological	
  markers	
  for	
  metastatic	
  transitional	
  cell	
  
carcinomas].	
  Verh	
  Dtsch	
  Ges	
  Pathol.	
  1993;77:260-­‐265.	
  
32.	
   Oosterhuis	
  JW,	
  Schapers	
  RF,	
  Janssen-­‐Heijnen	
  ML,	
  Pauwels	
  RP,	
  Newling	
  DW,	
  ten	
  Kate	
  
F.	
  Histological	
  grading	
  of	
  papillary	
  urothelial	
  carcinoma	
  of	
  the	
  bladder:	
  prognostic	
  value	
  of	
  
the	
  1998	
  WHO/ISUP	
  classification	
  system	
  and	
  comparison	
  with	
  conventional	
  grading	
  
systems.	
  J	
  Clin	
  Pathol.	
  Dec	
  2002;55(12):900-­‐905.	
  
33.	
   Pan	
  CC,	
  Chang	
  YH,	
  Chen	
  KK,	
  Yu	
  HJ,	
  Sun	
  CH,	
  Ho	
  DM.	
  Constructing	
  prognostic	
  model	
  
incorporating	
  the	
  2004	
  WHO/ISUP	
  classification	
  for	
  patients	
  with	
  non-­‐muscle-­‐invasive	
  
urothelial	
  tumours	
  of	
  the	
  urinary	
  bladder.	
  J	
  Clin	
  Pathol.	
  Oct	
  2010;63(10):910-­‐915.	
  
34.	
   Pan	
  CC,	
  Chang	
  YH,	
  Chen	
  KK,	
  Yu	
  HJ,	
  Sun	
  CH,	
  Ho	
  DM.	
  Prognostic	
  significance	
  of	
  the	
  
2004	
  WHO/ISUP	
  classification	
  for	
  prediction	
  of	
  recurrence,	
  progression,	
  and	
  cancer-­‐
specific	
  mortality	
  of	
  non-­‐muscle-­‐invasive	
  urothelial	
  tumors	
  of	
  the	
  urinary	
  bladder:	
  a	
  
clinicopathologic	
  study	
  of	
  1,515	
  cases.	
  Am	
  J	
  Clin	
  Pathol.	
  May	
  2010;133(5):788-­‐795.	
  

	
  

61	
  

	
  
35.	
   Parker	
  DC,	
  Folpe	
  AL,	
  Bell	
  J,	
  et	
  al.	
  Potential	
  utility	
  of	
  uroplakin	
  III,	
  thrombomodulin,	
  
high	
  molecular	
  weight	
  cytokeratin,	
  and	
  cytokeratin	
  20	
  in	
  noninvasive,	
  invasive,	
  and	
  
metastatic	
  urothelial	
  (transitional	
  cell)	
  carcinomas.	
  Am	
  J	
  Surg	
  Pathol.	
  Jan	
  2003;27(1):1-­‐10.	
  
36.	
   Patrick	
  DJ,	
  Fitzgerald	
  SD,	
  Sesterhenn	
  IA,	
  Davis	
  CJ,	
  Kiupel	
  M.	
  Classification	
  of	
  canine	
  
urinary	
  bladder	
  urothelial	
  tumours	
  based	
  on	
  the	
  World	
  Health	
  Organization/International	
  
Society	
  of	
  Urological	
  Pathology	
  consensus	
  classification.	
  J	
  Comp	
  Pathol.	
  Nov	
  
2006;135(4):190-­‐199.	
  
37.	
   Pich	
  A,	
  Chiusa	
  L,	
  Formiconi	
  A,	
  Galliano	
  D,	
  Bortolin	
  P,	
  Navone	
  R.	
  Biologic	
  differences	
  
between	
  noninvasive	
  papillary	
  urothelial	
  neoplasms	
  of	
  low	
  malignant	
  potential	
  and	
  low-­‐
grade	
  (grade	
  1)	
  papillary	
  carcinomas	
  of	
  the	
  bladder.	
  Am	
  J	
  Surg	
  Pathol.	
  Dec	
  
2001;25(12):1528-­‐1533.	
  
38.	
   Porter	
  AG,	
  Janicke	
  RU.	
  Emerging	
  roles	
  of	
  caspase-­‐3	
  in	
  apoptosis.	
  Cell	
  Death	
  Differ.	
  
Feb	
  1999;6(2):99-­‐104.	
  
39.	
   Ramos-­‐Vara	
  JA,	
  Miller	
  MA,	
  Boucher	
  M,	
  Roudabush	
  A,	
  Johnson	
  GC.	
  
Immunohistochemical	
  detection	
  of	
  uroplakin	
  III,	
  cytokeratin	
  7,	
  and	
  cytokeratin	
  20	
  in	
  
canine	
  urothelial	
  tumors.	
  Vet	
  Pathol.	
  Jan	
  2003;40(1):55-­‐62.	
  
40.	
   Ristimaki	
  A,	
  Nieminen	
  O,	
  Saukkonen	
  K,	
  Hotakainen	
  K,	
  Nordling	
  S,	
  Haglund	
  C.	
  
Expression	
  of	
  cyclooxygenase-­‐2	
  in	
  human	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  
bladder.	
  Am	
  J	
  Pathol.	
  Mar	
  2001;158(3):849-­‐853.	
  
41.	
   Rocha	
  TA,	
  Mauldin	
  GN,	
  Patnaik	
  AK,	
  Bergman	
  PJ.	
  Prognostic	
  factors	
  in	
  dogs	
  with	
  
urinary	
  bladder	
  carcinoma.	
  J	
  Vet	
  Intern	
  Med.	
  Sep-­‐Oct	
  2000;14(5):486-­‐490.	
  
42.	
   Samaratunga	
  H,	
  Makarov	
  DV,	
  Epstein	
  JI.	
  Comparison	
  of	
  WHO/ISUP	
  and	
  WHO	
  
classification	
  of	
  noninvasive	
  papillary	
  urothelial	
  neoplasms	
  for	
  risk	
  of	
  progression.	
  Urology.	
  
Aug	
  2002;60(2):315-­‐319.	
  
43.	
   Schned	
  AR,	
  Andrew	
  AS,	
  Marsit	
  CJ,	
  Zens	
  MS,	
  Kelsey	
  KT,	
  Karagas	
  MR.	
  Survival	
  
following	
  the	
  diagnosis	
  of	
  noninvasive	
  bladder	
  cancer:	
  WHO/International	
  Society	
  of	
  
Urological	
  Pathology	
  versus	
  WHO	
  classification	
  systems.	
  J	
  Urol.	
  Oct	
  2007;178(4	
  Pt	
  1):1196-­‐
1200;	
  discussion	
  1200.	
  
44.	
   Shariat	
  SF,	
  Matsumoto	
  K,	
  Kim	
  J,	
  et	
  al.	
  Correlation	
  of	
  cyclooxygenase-­‐2	
  expression	
  
with	
  molecular	
  markers,	
  pathological	
  features	
  and	
  clinical	
  outcome	
  of	
  transitional	
  cell	
  
carcinoma	
  of	
  the	
  bladder.	
  J	
  Urol.	
  Sep	
  2003;170(3):985-­‐989.	
  
45.	
   Shirahama	
  T.	
  Cyclooxygenase-­‐2	
  expression	
  is	
  up-­‐regulated	
  in	
  transitional	
  cell	
  
carcinoma	
  and	
  its	
  preneoplastic	
  lesions	
  in	
  the	
  human	
  urinary	
  bladder.	
  Clin	
  Cancer	
  Res.	
  Jun	
  
2000;6(6):2424-­‐2430.	
  

	
  

62	
  

	
  
46.	
   Shirahama	
  T,	
  Arima	
  J,	
  Akiba	
  S,	
  Sakakura	
  C.	
  Relation	
  between	
  cyclooxygenase-­‐2	
  
expression	
  and	
  tumor	
  invasiveness	
  and	
  patient	
  survival	
  in	
  transitional	
  cell	
  carcinoma	
  of	
  
the	
  urinary	
  bladder.	
  Cancer.	
  Jul	
  1	
  2001;92(1):188-­‐193.	
  
47.	
   Soslow	
  RA,	
  Rouse	
  RV,	
  Hendrickson	
  MR,	
  Silva	
  EG,	
  Longacre	
  TA.	
  Transitional	
  cell	
  
neoplasms	
  of	
  the	
  ovary	
  and	
  urinary	
  bladder:	
  a	
  comparative	
  immunohistochemical	
  analysis.	
  
Int	
  J	
  Gynecol	
  Pathol.	
  Jul	
  1996;15(3):257-­‐265.	
  
48.	
   Stadelmann	
  C,	
  Lassmann	
  H.	
  Detection	
  of	
  apoptosis	
  in	
  tissue	
  sections.	
  Cell	
  Tissue	
  Res.	
  
Jul	
  2000;301(1):19-­‐31.	
  
49.	
   Tot	
  T.	
  Cytokeratins	
  20	
  and	
  7	
  as	
  biomarkers:	
  usefulness	
  in	
  discriminating	
  primary	
  
from	
  metastatic	
  adenocarcinoma.	
  Eur	
  J	
  Cancer.	
  Apr	
  2002;38(6):758-­‐763.	
  
50.	
   Valli	
  VE,	
  Norris	
  A,	
  Jacobs	
  RM,	
  et	
  al.	
  Pathology	
  of	
  canine	
  bladder	
  and	
  urethral	
  cancer	
  
and	
  correlation	
  with	
  tumour	
  progression	
  and	
  survival.	
  J	
  Comp	
  Pathol.	
  Aug	
  
1995;113(2):113-­‐130.	
  
51.	
   Vardar	
  E,	
  Gunlusoy	
  B,	
  Minareci	
  S,	
  Postaci	
  H,	
  Ayder	
  AR.	
  Evaluation	
  of	
  p53	
  nuclear	
  
accumulation	
  in	
  low-­‐	
  and	
  high-­‐grade	
  (WHO/ISUP	
  classification)	
  transitional	
  papillary	
  
carcinomas	
  of	
  the	
  bladder	
  for	
  tumor	
  recurrence	
  and	
  progression.	
  Urol	
  Int.	
  2006;77(1):27-­‐
33.	
  
52.	
   Vojtesek	
  B,	
  Staskova	
  Z,	
  Nenutil	
  R,	
  et	
  al.	
  A	
  panel	
  of	
  monoclonal	
  antibodies	
  to	
  keratin	
  
no.	
  7:	
  characterization	
  and	
  value	
  in	
  tumor	
  diagnosis.	
  Neoplasma.	
  1990;37(3):333-­‐342.	
  
53.	
   Wadhwa	
  P,	
  Goswami	
  AK,	
  Joshi	
  K,	
  Sharma	
  SK.	
  Cyclooxygenase-­‐2	
  expression	
  
increases	
  with	
  the	
  stage	
  and	
  grade	
  in	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder.	
  Int	
  
Urol	
  Nephrol.	
  2005;37(1):47-­‐53.	
  
54.	
  

Wang	
  D,	
  Dubois	
  RN.	
  Prostaglandins	
  and	
  cancer.	
  Gut.	
  Jan	
  2006;55(1):115-­‐122.	
  

55.	
   Wang	
  MT,	
  Honn	
  KV,	
  Nie	
  D.	
  Cyclooxygenases,	
  prostanoids,	
  and	
  tumor	
  progression.	
  
Cancer	
  Metastasis	
  Rev.	
  Dec	
  2007;26(3-­‐4):525-­‐534.	
  
56.	
   Wendum	
  D,	
  Masliah	
  J,	
  Trugnan	
  G,	
  Flejou	
  JF.	
  Cyclooxygenase-­‐2	
  and	
  its	
  role	
  in	
  
colorectal	
  cancer	
  development.	
  Virchows	
  Arch.	
  Oct	
  2004;445(4):327-­‐333.	
  
57.	
   Wu	
  XR,	
  Kong	
  XP,	
  Pellicer	
  A,	
  Kreibich	
  G,	
  Sun	
  TT.	
  Uroplakins	
  in	
  urothelial	
  biology,	
  
function,	
  and	
  disease.	
  Kidney	
  Int.	
  Jun	
  2009;75(11):1153-­‐1165.	
  
58.	
   Wu	
  XR,	
  Lin	
  JH,	
  Walz	
  T,	
  et	
  al.	
  Mammalian	
  uroplakins.	
  A	
  group	
  of	
  highly	
  conserved	
  
urothelial	
  differentiation-­‐related	
  membrane	
  proteins.	
  J	
  Biol	
  Chem.	
  May	
  6	
  
1994;269(18):13716-­‐13724.	
  

	
  

63	
  

	
  
59.	
   Yin	
  H,	
  Leong	
  AS.	
  Histologic	
  grading	
  of	
  noninvasive	
  papillary	
  urothelial	
  tumors:	
  
validation	
  of	
  the	
  1998	
  WHO/ISUP	
  system	
  by	
  immunophenotyping	
  and	
  follow-­‐up.	
  Am	
  J	
  Clin	
  
Pathol.	
  May	
  2004;121(5):679-­‐687.	
  
	
  
	
  	
  
	
  
	
  
	
  
	
  

	
  

	
  

64	
  

	
  
CHAPTER	
  3	
  
	
  
	
  
Evaluation	
  of	
  15-­‐hydroxyprostaglandin	
  dehydrogenase	
  (HPGD),	
  cyclooxygenase-­‐2	
  
(COX-­‐2),	
  cadherin,	
  and	
  β-­‐catenin	
  expression	
  in	
  canine	
  urinary	
  bladder	
  urothelial	
  
carcinomas	
  
	
  
Dodd	
  Sledge1,	
  Elizabeth	
  A.	
  McNiel2,	
  Nicole	
  J.	
  Madrill3,	
  Monica	
  Liebert4,	
  Matti	
  Kiupel1	
  
	
  
1	
  Department	
  of	
  Pathobiology	
  and	
  Diagnostic	
  Investigation,	
  Diagnostic	
  Center	
  for	
  
Population	
  and	
  Animal	
  Health,	
  Michigan	
  State	
  University,	
  Lansing	
  MI	
  
	
  University,	
  Lansing,	
  MI	
  	
  
2	
  Tufts	
  Cummings	
  School	
  of	
  Veterinary	
  Medicine	
  and	
  Molecular	
  Oncology	
  Research	
  
Institute,	
  Boston,	
  MA,	
  
3	
  College	
  of	
  Veterinary	
  Medicine,	
  Michigan	
  State	
  University,	
  East	
  Lansing,	
  MI	
  
4	
  Department	
  of	
  Urology,	
  University	
  of	
  Michigan,	
  Ann	
  Arbor,	
  MI	
  
	
  
	
  
	
  

	
  

	
  

65	
  

	
  
Abstract	
  
There	
  likely	
  are	
  links	
  between	
  prostaglandin	
  regulation,	
  cadherin	
  switching,	
  and	
  Wnt	
  
signaling	
  in	
  the	
  carcinogenesis	
  of	
  urinary	
  bladder	
  urothelial	
  carcinomas.	
  Cadherin	
  
switching	
  is	
  described	
  and	
  decreased	
  expression	
  of	
  15-­‐hydroxyprostaglandin	
  
dehydrogenase	
  (HPGD),	
  which	
  metabolizes	
  prostaglandin	
  E2,	
  is	
  associated	
  with	
  loss	
  of	
  E-­‐
cadherin	
  expression	
  in	
  human	
  urothelial	
  carcinomas.	
  Aberrant	
  expression	
  of	
  β-­‐catenin,	
  an	
  
integral	
  part	
  of	
  Wnt	
  signaling,	
  is	
  associated	
  with	
  changes	
  in	
  prostaglandin	
  E2	
  and	
  cadherin	
  
expression	
  in	
  other	
  carcinomas.	
  Using	
  the	
  dog	
  as	
  a	
  model,	
  the	
  expression	
  of	
  HPGD,	
  
cyclooxygenase-­‐2	
  (COX-­‐2),	
  β-­‐catenin,	
  and	
  E-­‐,	
  P-­‐,	
  and	
  N-­‐cadherin	
  was	
  evaluated	
  in	
  canine	
  
normal	
  urinary	
  bladders	
  and	
  urothelial	
  carcinomas,	
  and	
  urothelial	
  carcinoma	
  cell	
  lines.	
  
Using	
  immunohistochemistry,	
  normal	
  canine	
  urinary	
  bladders	
  and	
  low	
  grade,	
  noninvasive	
  
canine	
  urothelial	
  carcinomas	
  expressed	
  HPGD	
  in	
  superficial	
  epithelial	
  cells,	
  lacked	
  COX-­‐2,	
  
and	
  expressed	
  E-­‐	
  and	
  P-­‐cadherin	
  and	
  β-­‐catenin	
  along	
  cell	
  membranes.	
  In	
  comparison,	
  a	
  
significant	
  proportion	
  of	
  high	
  grade,	
  infiltrative	
  urothelial	
  carcinomas	
  exhibited	
  loss	
  of	
  
HPGD,	
  increased	
  COX-­‐2,	
  decreased	
  P-­‐cadherin,	
  and	
  aberrant	
  localization	
  of	
  β-­‐catenin	
  
expression	
  within	
  neoplastic	
  cells.	
  E-­‐cadherin	
  was	
  expressed	
  in	
  all	
  canine	
  urothelial	
  
carcinomas	
  regardless	
  of	
  HPGD	
  expression.	
  N-­‐cadherin	
  was	
  not	
  expressed	
  in	
  normal	
  or	
  
neoplastic	
  canine	
  urothelium.	
  Western	
  blots	
  demonstrated	
  that	
  none	
  of	
  the	
  canine	
  
urothelial	
  carcinoma	
  cell	
  lines	
  expressed	
  HPGD	
  or	
  N-­‐cadherin,	
  but	
  all	
  expressed	
  E-­‐	
  and	
  P-­‐
cadherin.	
  In	
  contrast,	
  human	
  urothelial	
  carcinoma	
  cell	
  lines	
  examined	
  in	
  parallel	
  had	
  loss	
  of	
  
HPGD	
  had	
  loss	
  of	
  E-­‐	
  and	
  P-­‐cadherin	
  expression	
  and	
  gain	
  of	
  N-­‐cadherin	
  expression.	
  These	
  
data	
  suggest	
  that	
  while	
  these	
  pathways	
  may	
  be	
  related	
  in	
  both	
  canine	
  and	
  human	
  urothelial	
  
carcinomas,	
  regulation	
  is	
  somewhat	
  different	
  between	
  dogs	
  and	
  humans.	
  

	
  

66	
  

	
  
Introduction	
  
Increased	
  expression	
  of	
  prostaglandin	
  E2	
  (PGE2)	
  has	
  been	
  associated	
  with	
  
carcinogenesis	
  in	
  many	
  epithelial	
  tissues	
  and	
  has	
  specifically	
  been	
  implicated	
  in	
  avoidance	
  
of	
  apoptosis,	
  angiogenesis,	
  cellular	
  proliferation,	
  invasion,	
  and	
  metastasis	
  in	
  cancer.1-­‐6	
  The	
  
overall	
  expression	
  of	
  PGE2	
  within	
  a	
  given	
  tissue	
  is	
  dependent	
  on	
  the	
  rates	
  of	
  its	
  synthesis	
  
from	
  arachadonic	
  acid	
  and	
  its	
  metabolism	
  by	
  15-­‐hydroxyprostaglandin	
  dehydrogenase	
  
(HPGD).	
  Carcinogenesis	
  research	
  concerning	
  PGE2	
  has	
  most	
  intensely	
  focused	
  on	
  the	
  
synthesis	
  side	
  of	
  expression.	
  Specifically,	
  cyclooxygenase-­‐2	
  (COX-­‐2),	
  an	
  inducible	
  enzyme	
  
that	
  metabolizes	
  arachidonic	
  acid	
  to	
  the	
  PGE2	
  precursor	
  prostaglandin	
  H2,	
  has	
  been	
  shown	
  
to	
  be	
  upregulated	
  in	
  carcinomas	
  arising	
  in	
  numerous	
  tissues	
  including	
  the	
  urothelium	
  of	
  
the	
  urinary	
  bladder.2,	
  7	
  More	
  recently,	
  a	
  loss	
  of	
  HPGD	
  has	
  been	
  documented	
  in	
  various	
  
carcinomas	
  suggesting	
  that	
  a	
  decrease	
  in	
  PGE2	
  degradation	
  also	
  plays	
  a	
  role	
  in	
  
carcinogenesis.8,	
  9	
  The	
  exact	
  mechanisms	
  that	
  drive	
  PGE2	
  associated	
  carcinogenesis	
  are	
  
unclear	
  and	
  likely	
  multifactorial,	
  but	
  there	
  is	
  some	
  evidence	
  that	
  changes	
  in	
  PGE2	
  
regulation	
  are	
  related	
  to	
  cadherin	
  expression	
  and	
  the	
  Wnt	
  signaling	
  pathway	
  through	
  
changes	
  in	
  β-­‐catenin	
  expression.10-­‐12	
  
Cadherins	
  are	
  membrane-­‐associated	
  proteins	
  that	
  mediate	
  cell-­‐cell	
  and/or	
  cell-­‐
matrix	
  interactions.	
  Loss	
  of	
  cadherins	
  typically	
  expressed	
  by	
  epithelial	
  cells	
  such	
  as	
  E-­‐	
  and	
  
P-­‐cadherin	
  and/or	
  switching	
  of	
  the	
  expression	
  of	
  cadherins,	
  often	
  to	
  N-­‐cadherin,	
  have	
  been	
  
associated	
  with	
  increased	
  invasiveness,	
  increased	
  metastatic	
  potential,	
  and	
  an	
  overall	
  
epithelial-­‐mesenchymal	
  transition	
  in	
  many	
  types	
  of	
  carcinomas.13-­‐16	
  Regulation	
  of	
  cadherin	
  
expression	
  has	
  recently	
  been	
  linked	
  to	
  the	
  regulation	
  of	
  PGE2	
  in	
  some	
  types	
  of	
  carcinomas	
  
including,	
  but	
  not	
  limited	
  to	
  urothelial	
  carcinomas,	
  squamous	
  cell	
  carcinomas,	
  and	
  non-­‐
	
  

67	
  

	
  
small	
  cell	
  lung	
  cancer.17-­‐19	
  
In	
  addition	
  to	
  the	
  potential	
  effects	
  altered	
  metabolism	
  of	
  PGE2	
  has	
  on	
  cadherin	
  
expression,	
  alterations	
  in	
  prostaglandin	
  regulation	
  may	
  affect	
  urothelial	
  carcinogenesis	
  
through	
  the	
  Wnt	
  signaling	
  pathway.	
  While	
  numerous	
  factors	
  can	
  affect	
  Wnt	
  signaling,	
  the	
  
end	
  result	
  of	
  Wnt	
  signaling	
  is	
  the	
  accumulation	
  and	
  translocation	
  of	
  β-­‐catenin	
  into	
  the	
  
nucleus	
  where	
  it	
  interacts	
  with	
  transcription	
  factors	
  resulting	
  in	
  the	
  expression	
  of	
  cancer	
  
associated	
  genes.	
  β-­‐catenin	
  is	
  normally	
  sequestered	
  along	
  the	
  cell	
  membrane	
  due	
  to	
  its	
  
association	
  with	
  intracellular	
  portions	
  of	
  cadherins.	
  	
  Although	
  loss	
  of	
  cadherin	
  expression	
  
alone	
  cannot	
  lead	
  to	
  Wnt	
  signaling,	
  changes	
  in	
  cadherin	
  expression	
  can	
  free	
  β-­‐catenin	
  from	
  
its	
  normal	
  membrane	
  localization	
  favoring	
  its	
  accumulation	
  within	
  the	
  cytoplasm	
  and	
  
nucleus.3,6	
  Thus,	
  aberrant	
  expression	
  of	
  PGE2	
  due	
  to	
  either	
  increased	
  production	
  or	
  
decreased	
  degradation	
  may	
  lead	
  to	
  increased	
  Wnt	
  pathway	
  signaling	
  in	
  association	
  with	
  
changes	
  in	
  cadherin	
  expression.	
  	
  
In	
  urothelial	
  carcinomas	
  of	
  humans,	
  it	
  has	
  recently	
  been	
  shown	
  that	
  decreased	
  
expression	
  of	
  HPGD	
  and	
  increased	
  PGE2	
  expression	
  are	
  associated	
  with	
  a	
  more	
  invasive	
  
phenotype	
  and	
  worse	
  prognosis	
  making	
  this	
  neoplasm	
  ideal	
  for	
  the	
  study	
  of	
  the	
  role	
  of	
  
aberrant	
  prostaglandin	
  regulation	
  in	
  carcinogenesis.12,	
  19	
  Interestingly,	
  decreased	
  
expression	
  of	
  HPGD	
  in	
  neoplastic	
  cells	
  has	
  been	
  associated	
  with	
  loss	
  of	
  E-­‐cadherin.	
  Given	
  
such	
  loss	
  of	
  E-­‐cadherin	
  in	
  urothelial	
  carcinomas,	
  it	
  is	
  reasonable	
  to	
  postulate	
  that	
  cadherin	
  
switching	
  driven	
  by	
  alterations	
  in	
  prostaglandin	
  regulation	
  occurs	
  in	
  the	
  development	
  of	
  
urothelial	
  carcinomas.	
  Also,	
  it	
  is	
  plausible	
  that	
  changes	
  in	
  both	
  prostaglandin	
  expression	
  
and	
  cadherin	
  expression	
  would	
  affect	
  Wnt	
  signaling	
  through	
  changes	
  in	
  β-­‐catenin	
  
expression.	
  	
  

	
  

68	
  

	
  
Dogs	
  are	
  recognized	
  as	
  outstanding	
  models	
  for	
  many	
  human	
  diseases	
  including	
  
some	
  cancers.20	
  Specific	
  to	
  the	
  current	
  study,	
  dogs	
  spontaneously	
  develop	
  bladder	
  cancers	
  
that	
  histologically	
  resemble	
  human	
  bladder	
  cancers	
  and	
  that	
  have	
  a	
  similar	
  clinical	
  
course.21,	
  22	
  23	
  Given	
  their	
  similarity	
  to	
  human	
  urothelial	
  carcinomas	
  in	
  both	
  morphology	
  
and	
  progression,	
  it	
  is	
  plausible	
  that	
  canine	
  urothelial	
  carcinomas	
  have	
  features	
  of	
  
carcinogenesis	
  homologous	
  to	
  those	
  in	
  humans.	
  As	
  such,	
  dogs	
  may	
  prove	
  pivotal	
  in	
  studies	
  
into	
  specific	
  pathways	
  of	
  urothelial	
  carcinogenesis	
  as	
  well	
  as	
  provide	
  a	
  nonhuman	
  model	
  
for	
  therapeutic	
  manipulation.	
  To	
  investigate	
  the	
  significance	
  of	
  prostaglandin	
  regulation	
  
pathways,	
  cadherin	
  switching,	
  and	
  Wnt	
  signaling	
  in	
  canine	
  urothelial	
  carcinomas	
  and	
  
potential	
  applicability	
  of	
  the	
  dog	
  as	
  a	
  model	
  for	
  these	
  specific	
  carcinogenesis	
  pathways,	
  we	
  
evaluated	
  the	
  expression	
  of	
  HPGD,	
  COX-­‐2,	
  E-­‐cadherin,	
  P-­‐cadherin,	
  N-­‐cadherin,	
  and	
  β-­‐
catenin	
  in	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines	
  and	
  in	
  ex	
  vivo	
  tissues.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

69	
  

	
  
Materials	
  and	
  methods	
  
Selection	
  of	
  Cases	
  and	
  Histologic	
  Classification	
  
A	
  series	
  of	
  36	
  dogs	
  that	
  were	
  diagnosed	
  with	
  urothelial	
  carcinomas	
  at	
  either	
  the	
  
Michigan	
  State	
  University	
  Veterinary	
  Teaching	
  Hospital	
  or	
  the	
  University	
  of	
  Minnesota	
  
Veterinary	
  Teaching	
  Hospital	
  through	
  biopsy	
  were	
  selected	
  for	
  inclusion	
  in	
  the	
  study	
  based	
  
on	
  owner	
  consent,	
  availability	
  of	
  clinical	
  history,	
  and	
  availability	
  of	
  paraffin-­‐embedded,	
  
formalin-­‐fixed	
  diagnostic	
  samples.	
  For	
  each	
  urothelial	
  carcinoma	
  case,	
  descriptive	
  
information	
  was	
  obtained	
  from	
  medical	
  records	
  and	
  periodic	
  follow-­‐up	
  questionnaires	
  
including	
  age	
  at	
  diagnosis,	
  breed,	
  sex,	
  survival	
  time	
  from	
  the	
  date	
  of	
  diagnosis,	
  the	
  reported	
  
cause	
  of	
  death,	
  and	
  any	
  treatment	
  employed.	
  In	
  addition,	
  bladder	
  samples	
  from	
  10	
  healthy	
  
dogs	
  that	
  were	
  used	
  for	
  veterinary	
  student	
  teaching	
  purposes	
  were	
  harvested	
  in	
  10%	
  
buffered	
  formalin	
  and	
  routinely	
  processed	
  for	
  histologic	
  examination.	
  Five	
  µm	
  sections	
  of	
  
all	
  samples	
  were	
  routinely	
  processed	
  and	
  stained	
  with	
  hematoxylin	
  and	
  eosin.	
  Urothelial	
  
proliferative	
  lesions	
  were	
  classified	
  and	
  graded	
  according	
  to	
  the	
  WHO/ISUP	
  Consensus	
  
Classification	
  System	
  as	
  previously	
  described	
  in	
  canine	
  urothelial	
  tumors.22	
  In	
  addition,	
  the	
  
degree	
  of	
  invasion	
  of	
  all	
  carcinomas	
  into	
  the	
  wall	
  of	
  the	
  urinary	
  bladder	
  was	
  scored	
  as	
  no	
  
invasion,	
  invasion	
  into	
  the	
  substantia	
  propria,	
  or	
  invasion	
  into	
  the	
  muscularis.	
  	
  
Immunohistochemistry	
  
Immunohistochemistry	
  (IHC)	
  was	
  used	
  to	
  evaluate	
  the	
  expression	
  of	
  HPGD,	
  COX-­‐2,	
  
β-­‐catenin,	
  and	
  E-­‐,	
  N-­‐,	
  and	
  P-­‐cadherin	
  in	
  normal	
  canine	
  urinary	
  bladder	
  and	
  canine	
  
urothelial	
  carcinomas.	
  Five	
  µm	
  sections	
  of	
  all	
  formalin-­‐fixed,	
  paraffin-­‐embedded	
  tissues	
  
were	
  processed	
  for	
  immunohistochemistry	
  and	
  labeled	
  with	
  a	
  rabbit	
  monoclonal	
  anti-­‐
HPGD	
  antibody	
  (1:100,	
  Sigma-­‐Aldrich,	
  St.	
  Louis,	
  MO,	
  USA),	
  a	
  rabbit	
  polyclonal	
  anti-­‐COX-­‐2	
  

	
  

70	
  

	
  
antibody	
  (1:100.	
  Cayman	
  Chemical	
  Company,	
  Ann	
  Arbor,	
  MI,	
  USA),	
  a	
  rabbit	
  monoclonal	
  
anti-­‐β-­‐catenin	
  antibody	
  (1:1000,	
  Abcam,	
  Cambridge,	
  MA,	
  USA),	
  a	
  mouse	
  polyclonal	
  anti-­‐E-­‐
cadherin	
  antibody	
  (1:300,	
  BD	
  Biosciences,	
  San	
  Jose,	
  CA,	
  USA),	
  a	
  mouse	
  monoclonal	
  anti-­‐N-­‐
cadherin	
  (1:100,	
  Invitrogen,	
  Life	
  Technologies,	
  Grand	
  Island,	
  NY,	
  USA),	
  or	
  a	
  mouse	
  
monoclonal	
  anti-­‐P-­‐cadherin	
  antibody	
  (1:100,	
  NovoCastra,	
  Leica	
  Biosystems,	
  Buffalo	
  Grove,	
  
IL	
  USA).	
  For	
  HPGD,	
  β-­‐catenin,	
  and	
  E-­‐,	
  N-­‐,	
  and	
  P-­‐cadherin,	
  deparaffinization,	
  antigen	
  
retrieval,	
  immunohistochemical	
  staining	
  and	
  counterstaining	
  was	
  performed	
  on	
  a	
  Bond	
  
maX™	
  Automated	
  Staining	
  System	
  (Leica	
  Biosystems,	
  Buffalo	
  Grove,	
  IL,	
  USA)	
  using	
  the	
  
Bond™	
  Polymer	
  Refine	
  Detection	
  System	
  (Leica	
  Biosystems,	
  Buffalo	
  Grove,	
  IL,	
  USA),	
  which	
  
employs	
  a	
  3,3’	
  diaminobenzidine	
  tetrahydrochloride	
  (DAB)	
  chromogen	
  detection	
  system.	
  
For	
  COX-­‐2,	
  deparaffinization,	
  antigen	
  retrieval,	
  immunohistochemical	
  staining	
  and	
  
counterstaining	
  was	
  preformed	
  on	
  a	
  Benchmark	
  XT™	
  autostainer	
  (Ventana,	
  Tucson,	
  AZ,	
  
USA)	
  using	
  and	
  an	
  Enhanced	
  Alkaline	
  Phosphatase	
  Red	
  Detection	
  Kit	
  (Ventana)	
  that	
  uses	
  
an	
  indirect	
  biotin	
  streptavidin	
  and	
  Fast	
  Red	
  chromogen	
  detection	
  system.	
  	
  Retrieval	
  for	
  
HPGD	
  was	
  accomplished	
  by	
  incubation	
  for	
  20	
  minutes	
  with	
  EnVision™	
  FLEX	
  Target	
  
Retrieval	
  Solution,	
  Low	
  pH	
  (Dako,	
  Carpinteria,	
  CA).	
  	
  Retrieval	
  for	
  β-­‐catenin	
  and	
  E-­‐	
  and	
  N-­‐
cadherin	
  was	
  accomplished	
  using	
  heat	
  induced	
  epitope	
  retrieval	
  and	
  incubation	
  with	
  
Bond™	
  Epitope	
  Retrieval	
  Solution	
  1	
  (Leica	
  Biosystems)	
  for	
  20	
  minutes.	
  	
  Retrieval	
  for	
  P-­‐
cadherin	
  was	
  accomplished	
  using	
  heat	
  induced	
  epitope	
  retrieval	
  and	
  incubation	
  with	
  
Bond™	
  Epitope	
  Retrieval	
  Solution	
  1	
  (Leica	
  Biosystems)	
  for	
  20	
  minutes.	
  	
  Positive	
  controls	
  
using	
  appropriate	
  canine	
  tissues	
  were	
  ran	
  in	
  parallel	
  to	
  cases	
  for	
  each	
  of	
  the	
  IHC	
  protocols	
  
as	
  follows:	
  normal	
  bladder	
  for	
  HPGD,	
  squamous	
  cell	
  carcinoma	
  for	
  COX-­‐2,	
  haired	
  skin	
  and	
  
normal	
  bladder	
  for	
  β-­‐catenin	
  and	
  E-­‐cadherin,	
  heart	
  for	
  N-­‐cadherin,	
  and	
  uterus	
  for	
  P-­‐

	
  

71	
  

	
  
cadherin.	
  For	
  negative	
  controls,	
  homologous	
  non-­‐immune	
  sera	
  or	
  buffer	
  replaced	
  primary	
  
antibodies.	
  
For	
  each	
  case,	
  immunoreactivity	
  for	
  HPGD	
  was	
  scored	
  by	
  the	
  percentage	
  of	
  
positively	
  labeled	
  cells	
  and	
  the	
  location	
  of	
  immunoreactivity	
  within	
  urothelium	
  (Fig.	
  10).	
  	
  
Specifically,	
  percentage	
  of	
  urothelium	
  expressing	
  HPGD	
  was	
  categorized	
  as	
  none,	
  <5%,	
  5-­‐
15%,	
  15-­‐30%	
  or	
  >30%.	
  Location	
  of	
  HPGD	
  was	
  categorized	
  as	
  Pattern	
  1	
  when	
  limited	
  to	
  the	
  
superficial	
  1-­‐3	
  cell	
  layers	
  and	
  diffuse	
  throughout	
  these	
  cell	
  layers;	
  Pattern	
  2	
  when	
  
superficial,	
  but	
  patchy	
  within	
  the	
  urothelium	
  with	
  <80%	
  of	
  the	
  total	
  superficial	
  urothelium	
  
being	
  labeled;	
  Pattern	
  3	
  when	
  diffuse	
  throughout	
  the	
  full	
  thickness	
  of	
  urothelium;	
  or	
  
Pattern	
  4	
  when	
  randomly	
  distributed	
  patchy	
  areas	
  of	
  immunoreactivity	
  with	
  <80%	
  of	
  the	
  
total	
  urothelium	
  being	
  labeled.	
  	
  Immunoreactivity	
  for	
  COX-­‐2	
  was	
  scored	
  by	
  percentage	
  of	
  
positive	
  cells	
  with	
  expression	
  being	
  categorized	
  as	
  none,	
  <5%,	
  5-­‐15%,	
  15-­‐30%	
  or	
  >30%	
  
(Fig.	
  11).	
  	
  For	
  E-­‐	
  and	
  P-­‐cadherin,	
  immunoreactivity	
  was	
  scored	
  by	
  location	
  within	
  
urothelium	
  and	
  within	
  cells,	
  intensity	
  of	
  immunolabeling,	
  and	
  degree	
  of	
  loss	
  (Fig.	
  12	
  and	
  
13).	
  	
  Location	
  of	
  E-­‐	
  and	
  P-­‐	
  expression	
  was	
  categorized	
  as	
  membrane	
  associated,	
  
cytoplasmic	
  only,	
  or	
  mixed	
  cytoplasmic	
  with	
  strong	
  membrane	
  labeling.	
  	
  Intensity	
  was	
  
subjectively	
  categorized	
  as	
  weak,	
  moderate,	
  or	
  strong.	
  Loss	
  of	
  E-­‐	
  and	
  P-­‐cadherin	
  expression	
  
was	
  categorized	
  as	
  no	
  loss,	
  patchy	
  loss,	
  or	
  extensive	
  loss.	
  	
  For	
  analysis,	
  aberrant	
  P-­‐cadherin	
  
expression	
  was	
  defined	
  as	
  extensive	
  loss	
  of	
  expression	
  or	
  cytoplasmic	
  expression.	
  For	
  N-­‐
cadherin,	
  only	
  presence	
  or	
  absence	
  of	
  immunoreactivity	
  within	
  the	
  urothelium	
  was	
  
evaluated.	
  	
  β-­‐catenin	
  expression	
  was	
  scored	
  by	
  percentage	
  of	
  cells	
  within	
  each	
  lesion	
  with	
  
membrane,	
  cytoplasmic,	
  mixed	
  membrane	
  and	
  cytoplasmic,	
  and	
  nuclear	
  expression	
  (Fig.	
  

	
  

72	
  

	
  
14).	
  	
  Aberrant	
  β-­‐catenin	
  expression	
  within	
  a	
  given	
  case	
  was	
  defined	
  as	
  >	
  or	
  =	
  to	
  40%	
  of	
  cell	
  
showing	
  cytoplasmic	
  labeling	
  or	
  >	
  or	
  =	
  to	
  10%	
  of	
  cells	
  showing	
  nuclear	
  labeling.	
  
Statistical	
  Analysis	
  
Chi	
  square	
  tests	
  performed	
  with	
  SPSS	
  statistical	
  software	
  (Somers,	
  NY,	
  USA)	
  was	
  
used	
  to	
  evaluate	
  correlations	
  between	
  histologic	
  classifications,	
  degree	
  of	
  anaplasia,	
  depth	
  
of	
  infiltration,	
  and	
  immunohistochemical	
  expression	
  of	
  examined	
  markers.	
  SPSS	
  statistical	
  
software	
  (Somers,	
  NY)	
  was	
  also	
  used	
  to	
  analyze	
  correlations	
  between	
  survival	
  time	
  of	
  
canine	
  urinary	
  carcinomas	
  and	
  clinico-­‐demographic	
  data,	
  histologic	
  classification	
  and	
  
grading,	
  and	
  immunohistochemical	
  expression	
  of	
  examined	
  markers	
  using	
  Kaplan-­‐Meier	
  
estimators	
  with	
  log	
  rank	
  tests,	
  and	
  univariate	
  and	
  multivariate	
  Cox	
  proportional	
  hazard	
  
modeling.	
  For	
  all	
  statistical	
  analysis,	
  significance	
  was	
  defined	
  as	
  p<0.05.	
  
Western	
  blots	
  
In	
  order	
  to	
  1)	
  validate	
  antibody	
  use	
  in	
  dogs	
  and	
  2)	
  establish	
  an	
  in	
  vitro	
  model	
  for	
  
future	
  studies,	
  IHC	
  and	
  Western	
  blotting	
  were	
  used	
  to	
  evaluate	
  expression	
  of	
  these	
  
molecules	
  in	
  a	
  variety	
  of	
  normal	
  canine	
  tissues	
  and	
  canine	
  and	
  human	
  urothelial	
  carcinoma	
  
cell	
  lines.	
  	
  Eight	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines	
  (ANGUS,	
  AXA,	
  AXC,	
  KINSEY,	
  K9TCC,	
  
NK,	
  TYLER1,	
  TYLER2)	
  and	
  seven	
  human	
  urothelial	
  carcinoma	
  cell	
  lines	
  (RT4,	
  UC2,	
  UC3,	
  
UC6,	
  UC12,	
  UC14,	
  5637)	
  were	
  cultured	
  in	
  100mm	
  plates	
  in	
  1:1	
  DMEM/F12	
  media	
  with	
  10%	
  
fetal	
  bovine	
  serum,	
  l-­‐glutamine,	
  and	
  penicillin/streptomycin.	
  	
  Cells	
  were	
  harvested	
  using	
  a	
  
cell	
  scraper	
  and	
  protein	
  was	
  subsequently	
  isolated	
  through	
  incubation	
  with	
  RIPA	
  buffer.	
  	
  
Total	
  protein	
  for	
  each	
  was	
  quantitated	
  using	
  Pierce™	
  BCA	
  Protein	
  Assay	
  Kit	
  (Thermo	
  
Scientific,	
  Rockford,	
  IL,	
  USA),	
  a	
  Victor	
  X3	
  microplate	
  reader	
  (PerkinElmer,	
  Shelton,	
  CT	
  USA)	
  
reading	
  absorbance	
  at	
  750nm,	
  and	
  comparison	
  to	
  a	
  dilution	
  series	
  of	
  bovine	
  serum	
  

	
  

73	
  

	
  
albumin.	
  	
  Equally	
  loaded	
  proteins	
  isolates	
  were	
  separated	
  using	
  SDS-­‐PAGE.	
  	
  For	
  PGDH,	
  
COX-­‐2	
  blots,	
  proteins	
  were	
  separated	
  using	
  12%	
  gels	
  (BioRad,	
  Hercules,	
  CA,	
  USA).	
  	
  For	
  E-­‐,	
  
N,	
  and	
  P-­‐cadherin	
  and	
  ß-­‐catenin	
  blots,	
  proteins	
  were	
  separated	
  using	
  7.5%	
  gels	
  (BioRad,	
  
Hercules,	
  USA).	
  	
  Percision	
  Plus	
  Protein™,	
  Dual	
  Color	
  (BioRad,	
  Hercules,	
  USA)	
  was	
  used	
  for	
  
protein	
  standards.	
  	
  Proteins	
  were	
  transferred	
  to	
  nitrocellulose	
  by	
  electroblotting.	
  	
  
Following	
  washes	
  in	
  TBST,	
  blots	
  were	
  blocked	
  using	
  3%	
  bovine	
  serum	
  albumin.	
  	
  Blots	
  were	
  
incubated	
  overnight	
  with	
  the	
  primary	
  antibodies	
  used	
  for	
  IHC	
  at	
  respective	
  concentrations	
  
of	
  1:000,	
  1:500,	
  1:200,	
  1:5000,	
  1:1000,	
  1:200	
  for	
  HPGD,	
  COX-­‐2,	
  ß-­‐catenin,	
  and	
  E-­‐,	
  N-­‐	
  and	
  P-­‐
cadherin	
  in	
  3%	
  powdered	
  milk.	
  	
  Following	
  washes	
  in	
  TBST,	
  blots	
  were	
  accordingly	
  
incubated	
  for	
  2	
  hours	
  with	
  either	
  1:2000	
  goat	
  anti-­‐rabbit	
  IgG	
  or	
  1:7500	
  goat	
  anti-­‐rabbit	
  IgG	
  
secondary	
  antibody	
  (Santa	
  Cruz	
  Biotechnology,	
  Dallas,	
  TX,	
  USA).	
  Following	
  washes	
  in	
  TBST,	
  
blots	
  were	
  developed	
  using	
  a	
  chemiluminescence	
  detection	
  system	
  (Thermo	
  Scientific,	
  
Rockford,	
  IL,	
  USA)	
  and	
  using	
  Amersham	
  Hyperfilm™	
  MP	
  autoradiography	
  film	
  (GE	
  
Healthcare,	
  Little	
  Chalfont,	
  BM,	
  UK).	
  	
  Blots	
  were	
  subsequently	
  stripped	
  for	
  1	
  hour	
  using	
  
Western	
  blot	
  stripping	
  solution	
  (Thermo	
  Scientific,	
  Rockford,	
  IL,	
  USA).	
  	
  Following	
  washes	
  
in	
  TBST	
  and	
  blocking	
  in	
  3%	
  BSA,	
  blots	
  were	
  incubated	
  overnight	
  with	
  goat	
  anti-­‐GAPDH	
  
primary	
  antibody.	
  	
  	
  Following	
  washes	
  in	
  TBST,	
  blots	
  were	
  incubated	
  for	
  2	
  hours	
  with	
  
donkey	
  anti-­‐goat	
  secondary	
  antibody	
  (Santa	
  Cruz	
  Biotechnology,	
  Dallas,	
  TX,	
  USA).	
  
Following	
  washes	
  in	
  TBST,	
  blots	
  were	
  developed	
  using	
  a	
  chemiluminescence	
  detection	
  
system	
  (Thermo	
  Scientific,	
  Rockford,	
  IL,	
  USA)	
  and	
  chemilumiescence	
  was	
  detected	
  using	
  
Amersham	
  Hyperfilm™	
  MP	
  radiographic	
  film	
  (GE	
  Healthcare,	
  Little	
  Chalfont,	
  BM,	
  UK).	
  
	
  
	
  

	
  

74	
  

	
  
Results	
  
Demographic	
  information,	
  histologic	
  classification,	
  and	
  grading	
  of	
  urothelial	
  carcinomas	
  
For	
  dogs	
  diagnosed	
  with	
  urothelial	
  carcinomas,	
  the	
  mean	
  age	
  at	
  initial	
  diagnosis	
  
was	
  10.5	
  years	
  of	
  age	
  (2.0	
  SD)	
  with	
  a	
  range	
  of	
  6.7	
  to	
  17.1	
  years	
  of	
  age.	
  Affected	
  dogs	
  
included	
  4	
  intact	
  females,	
  21	
  spayed	
  females,	
  and	
  11	
  castrated	
  males.	
  Represented	
  breeds	
  
included	
  5	
  Scottish	
  Terriers,	
  2	
  West	
  Highland	
  White	
  Terriers,	
  7	
  Beagles,	
  7	
  Shetland	
  
Sheepdogs,	
  and	
  14	
  dogs	
  of	
  other	
  breed	
  or	
  mixed	
  breed.	
  	
  For	
  the	
  32	
  cases	
  in	
  which	
  
information	
  regarding	
  treatment	
  was	
  available,	
  7	
  had	
  no	
  treatment,	
  2	
  had	
  only	
  local	
  
resection	
  of	
  the	
  mass,	
  17	
  were	
  treated	
  with	
  only	
  piroxicam,	
  and	
  6	
  were	
  treated	
  with	
  
piroxicam,	
  surgical	
  resection,	
  and	
  systemic	
  chemotherapy	
  which	
  variably	
  included	
  
adriamycin,	
  cyclophosphamide,	
  and/or	
  mitoxantrone.	
  	
  Two	
  animals	
  were	
  alive	
  at	
  the	
  end	
  
of	
  study,	
  1	
  animal	
  died	
  from	
  causes	
  unrelated	
  to	
  the	
  urinary	
  carcinoma,	
  30	
  reportedly	
  died	
  
or	
  were	
  euthanized	
  as	
  a	
  result	
  of	
  progressive	
  disease	
  related	
  to	
  the	
  urinary	
  bladder	
  tumors,	
  
and	
  3	
  were	
  lost	
  to	
  follow-­‐up.	
  	
  	
  
Of	
  the	
  36	
  urothelial	
  carcinomas,	
  8	
  were	
  diagnosed	
  as	
  infiltrating	
  carcinomas,	
  which	
  
had	
  no	
  appreciable	
  exophytic	
  papillary	
  component,	
  but	
  instead	
  primarily	
  infiltrated	
  into	
  
and	
  expanded	
  the	
  bladder	
  wall.	
  	
  The	
  remaining	
  28	
  cases	
  were	
  papillary	
  carcinomas,	
  the	
  
vast	
  majority	
  of	
  which	
  (24	
  of	
  28)	
  were	
  high	
  grade	
  (grade	
  II	
  or	
  III).	
  	
  Invasion	
  of	
  high-­‐grade	
  
papillary	
  carcinomas	
  into	
  the	
  bladder	
  wall	
  was	
  common	
  with	
  9	
  of	
  24	
  having	
  invasion	
  into	
  
the	
  substantia	
  propria	
  and	
  7	
  of	
  24	
  invading	
  into	
  the	
  muscularis.	
  	
  For	
  all	
  carcinomas,	
  the	
  
mean	
  survival	
  time	
  was	
  326	
  days	
  with	
  a	
  standard	
  deviation	
  of	
  341	
  and	
  range	
  of	
  0	
  to	
  1225	
  
days.	
  	
  For	
  infiltrative	
  carcinomas,	
  the	
  mean	
  survival	
  time	
  was	
  173	
  days	
  with	
  a	
  standard	
  
deviation	
  of	
  223	
  days	
  and	
  range	
  of	
  0	
  to	
  365	
  days.	
  	
  For	
  papillary	
  carcinomas,	
  the	
  mean	
  

	
  

75	
  

	
  
survival	
  time	
  was	
  370	
  days	
  with	
  a	
  standard	
  deviation	
  of	
  359	
  days	
  and	
  range	
  of	
  0	
  to	
  1225	
  
days.	
  
Immunohistochemistry:	
  HPGD	
  
There	
  was	
  diffuse	
  to	
  rarely	
  patchy	
  superficial	
  immunoreactivity	
  for	
  HPGD	
  consistent	
  
with	
  a	
  HPGD	
  Pattern	
  1	
  or	
  2	
  in	
  all	
  but	
  one	
  of	
  ten	
  examined	
  control	
  bladders	
  from	
  normal	
  
dogs.	
  	
  The	
  percentage	
  of	
  urothelial	
  cells	
  that	
  were	
  immunoreactive	
  for	
  HPGD	
  was	
  highly	
  
variable	
  between	
  control	
  bladders	
  ranging	
  from	
  <5	
  to	
  >30%.	
  	
  The	
  vast	
  majority	
  of	
  
urothelial	
  carcinomas	
  had	
  either	
  extensive	
  lack	
  of	
  immunoreactivity	
  (14/36)	
  or	
  had	
  
Pattern	
  4,	
  patchy	
  immunoreactivity	
  randomly	
  scattered	
  throughout	
  the	
  tumors	
  (14/36)	
  
with	
  the	
  total	
  percent	
  of	
  cells	
  expressing	
  HPGD	
  ranging	
  from	
  <5	
  to	
  >30%,	
  similar	
  to	
  that	
  
seen	
  in	
  control	
  bladders	
  (Table	
  7).	
  	
  	
  
Comparing	
  urothelial	
  carcinomas	
  and	
  urothelium	
  from	
  normal	
  controls,	
  there	
  was	
  
no	
  significant	
  difference	
  between	
  the	
  overall	
  percent	
  of	
  cells	
  within	
  the	
  urothelium	
  
expressing	
  HPGD	
  (p=0.205),	
  but	
  there	
  was	
  a	
  significant	
  difference	
  between	
  the	
  patterns	
  of	
  
HPGD	
  expression	
  (p<0.0001).	
  	
  	
  Examining	
  only	
  urothelial	
  carcinomas,	
  there	
  was	
  no	
  
significant	
  correlation	
  between	
  percentage	
  of	
  cells	
  within	
  the	
  urothelium	
  expressing	
  HPGD	
  
and	
  classification	
  as	
  infiltrative	
  or	
  papillary	
  (p=0.094),	
  degree	
  of	
  anaplasia	
  (p=0.70),	
  or	
  
depth	
  of	
  invasion	
  into	
  the	
  bladder	
  wall	
  (p=0.547).	
  	
  There	
  was	
  also	
  no	
  significant	
  correlation	
  
between	
  pattern	
  of	
  HPGD	
  expression	
  and	
  classification	
  as	
  infiltrative	
  or	
  papillary	
  
(p=0.121),	
  degree	
  of	
  anaplasia	
  (p=0.484),	
  or	
  depth	
  of	
  invasion	
  into	
  the	
  bladder	
  wall	
  
(p=0.109).	
  
Immunohistochemistry:	
  COX-­‐2	
  

	
  

76	
  

	
  
Less	
  than	
  5%	
  of	
  cells	
  within	
  the	
  urothelium	
  expressed	
  COX-­‐2	
  in	
  all	
  of	
  the	
  normal	
  
control	
  urinary	
  bladders	
  and	
  8	
  of	
  10	
  had	
  no	
  appreciable	
  immunoreactivity	
  for	
  COX-­‐2.	
  	
  In	
  
contrast,	
  25	
  of	
  36	
  urothelial	
  carcinomas	
  had	
  5%	
  or	
  greater	
  urothelial	
  cells	
  with	
  expression	
  
of	
  COX-­‐2	
  and	
  14	
  of	
  these	
  had	
  more	
  than	
  30%	
  of	
  urothelial	
  cells	
  with	
  immunoreactivity.	
  	
  	
  
Comparing	
  urothelial	
  carcinomas	
  and	
  urothelium	
  of	
  normal	
  controls,	
  there	
  was	
  a	
  
significant	
  difference	
  between	
  the	
  overall	
  percent	
  of	
  cells	
  within	
  the	
  urothelium	
  expressing	
  
COX-­‐2	
  (p<0.001).	
  	
  Examining	
  only	
  urothelial	
  carcinomas,	
  there	
  were	
  statistically	
  significant	
  
associations	
  between	
  COX-­‐2	
  expression	
  and	
  degree	
  of	
  anaplasia	
  (p=0.048)	
  and	
  depth	
  of	
  
invasion	
  into	
  the	
  urinary	
  bladder	
  wall	
  (p=0.017),	
  with	
  COX-­‐2	
  often	
  being	
  expressed	
  in	
  high	
  
percentages	
  of	
  neoplastic	
  urothelial	
  cells	
  in	
  urothelial	
  carcinomas	
  with	
  anaplasia	
  grades	
  2	
  
and	
  3,	
  and	
  in	
  carcinomas	
  with	
  invasion	
  into	
  the	
  muscularis	
  of	
  the	
  urinary	
  bladder.	
  There	
  
was	
  no	
  significant	
  correlation	
  between	
  percentage	
  of	
  cells	
  within	
  the	
  urothelium	
  
expressing	
  COX-­‐2	
  and	
  classification	
  of	
  the	
  urothelial	
  carcinomas	
  as	
  infiltrative	
  or	
  papillary	
  
(p=0.481).	
  
Immunohistochemistry:	
  Cadherins	
  
	
  

All	
  normal	
  control	
  bladders	
  had	
  strong	
  to	
  rarely	
  moderate	
  (1/10	
  cases),	
  

membranous	
  immunoreactivity	
  for	
  E-­‐cadherin	
  and	
  no	
  significant	
  loss	
  of	
  expression.	
  The	
  
majority	
  (25/36)	
  of	
  urothelial	
  carcinomas	
  had	
  membranous	
  immunoreactivity	
  for	
  E-­‐
cadherin,	
  5/36	
  had	
  predominately	
  cytoplasmic	
  immunoreactivity,	
  and	
  2/36	
  had	
  both	
  
membranous	
  and	
  cytoplasmic	
  immunoreactivity.	
  	
  Twenty-­‐nine	
  of	
  36	
  urothelial	
  carcinomas	
  
had	
  moderate	
  to	
  strong	
  immunoreactivity,	
  and	
  3	
  had	
  only	
  weak	
  immunoreactivity.	
  	
  Six	
  
urothelial	
  carcinomas	
  had	
  significant	
  loss	
  of	
  E-­‐cadherin	
  expression	
  that	
  was	
  most	
  often	
  
patchy	
  (5/6	
  cases),	
  and	
  rarely	
  extensive	
  (1/6	
  cases).	
  	
  

	
  

77	
  

	
  
All	
  ten	
  normal	
  control	
  urinary	
  bladders	
  had	
  strong	
  membranous	
  immunoreactivity	
  
for	
  P-­‐cadherin	
  and	
  no	
  significant	
  loss	
  of	
  expression.	
  The	
  majority	
  (26/36)	
  of	
  urothelial	
  
carcinomas	
  had	
  membranous	
  immunoreactivity	
  for	
  P-­‐cadherin,	
  1/36	
  had	
  predominately	
  
cytoplasmic	
  immunoreactivity,	
  and	
  4/36	
  had	
  both	
  membranous	
  and	
  cytoplasmic	
  
immunoreactivity.	
  	
  Twenty-­‐four	
  of	
  36	
  urothelial	
  carcinomas	
  had	
  moderate	
  to	
  strong	
  
immunoreactivity,	
  and	
  7	
  had	
  only	
  weak	
  immunoreactivity.	
  	
  Twenty-­‐one	
  urothelial	
  
carcinomas	
  had	
  significant	
  loss	
  of	
  P-­‐cadherin	
  expression	
  that	
  was	
  most	
  often	
  patchy	
  
(14/21	
  cases	
  with	
  significant	
  loss),	
  and	
  rarely	
  extensive	
  (7/21	
  cases).	
  	
  Defining	
  P-­‐cadherin	
  
as	
  aberrant	
  when	
  there	
  was	
  extensive	
  loss	
  of	
  expression	
  or	
  cytoplasmic	
  expression,	
  no	
  
normal	
  control	
  urinary	
  bladders	
  had	
  aberrant	
  P-­‐cadherin	
  expression	
  (Table	
  8).	
  	
  Four	
  of	
  7	
  
(57%)	
  infiltrating	
  urothelial	
  carcinomas	
  had	
  aberrant	
  P-­‐cadherin	
  expression	
  compared	
  to	
  
2/23	
  (9%)	
  of	
  papillary	
  urothelial	
  carcinomas	
  with	
  aberrant	
  expression.	
  
There	
  was	
  no	
  expression	
  of	
  N-­‐cadherin	
  in	
  urothelial	
  carcinomas,	
  within	
  tumor	
  
adjacent	
  well-­‐differentiated	
  urothelium,	
  or	
  within	
  the	
  urothelium	
  of	
  normal	
  urinary	
  
bladder	
  control	
  samples.	
  
Comparing	
  urothelial	
  carcinomas	
  and	
  normal	
  urothelium,	
  there	
  was	
  no	
  significant	
  
difference	
  between	
  the	
  overall	
  E-­‐cadherin	
  pattern	
  (p=0.26),	
  E-­‐cadherin	
  strength	
  (p=0.24),	
  
degree	
  of	
  E-­‐cadherin	
  expression	
  loss	
  (p=0.38),	
  P-­‐cadherin	
  pattern	
  (p=0.31),	
  or	
  aberrant	
  P-­‐
cadherin	
  expression	
  (p=0.139).	
  	
  There	
  were,	
  however,	
  significant	
  differences	
  between	
  
urothelial	
  carcinomas	
  and	
  urothelium	
  of	
  normal	
  controls	
  in	
  terms	
  of	
  P-­‐cadherin	
  expression	
  
strength	
  (p<0.001)	
  and	
  degree	
  of	
  P-­‐cadherin	
  expression	
  loss	
  (p=0.013).	
  	
  
Examining	
  only	
  urothelial	
  carcinomas,	
  there	
  were	
  no	
  significant	
  associations	
  
between	
  histologic	
  classification	
  of	
  urothelial	
  carcinomas	
  as	
  infiltrative	
  or	
  papillary	
  and	
  E-­‐

	
  

78	
  

	
  
cadherin	
  expression	
  pattern	
  (p=0.622),	
  strength	
  of	
  E-­‐cadherin	
  immunoreactivity	
  
(p=0.629),	
  or	
  degree	
  of	
  E-­‐cadherin	
  expression	
  loss	
  (p=0.829);	
  degree	
  of	
  anaplasia	
  and	
  E-­‐
cadherin	
  expression	
  pattern	
  (p=0.684),	
  strength	
  of	
  E-­‐cadherin	
  immunoreactivity	
  
(p=0.489),	
  or	
  degree	
  of	
  E-­‐cadherin	
  expression	
  loss	
  (p=0.563);	
  or	
  depth	
  of	
  invasion	
  into	
  the	
  
urinary	
  bladder	
  wall	
  and	
  E-­‐cadherin	
  expression	
  pattern	
  (p=0.431),	
  strength	
  of	
  E-­‐cadherin	
  
immunoreactivity	
  (p=0.060),	
  or	
  degree	
  of	
  E-­‐cadherin	
  expression	
  loss	
  (p=0.273).	
  	
  There	
  
were	
  statistically	
  significant	
  associations	
  between	
  histologic	
  classification	
  of	
  urothelial	
  
carcinomas	
  as	
  infiltrative	
  or	
  papillary	
  and	
  P-­‐cadherin	
  expression	
  pattern	
  (p=0.014),	
  degree	
  
of	
  P-­‐cadherin	
  expression	
  loss	
  (p=0.011),	
  and	
  aberrant	
  P-­‐cadherin	
  expression	
  pattern	
  
(p=0.012);	
  however,	
  there	
  were	
  no	
  significant	
  associations	
  between	
  histologic	
  
classification	
  of	
  urothelial	
  carcinomas	
  as	
  infiltrative	
  or	
  papillary	
  and	
  strength	
  of	
  P-­‐cadherin	
  
immunoreactivity	
  (p=0.280);	
  degree	
  of	
  anaplasia	
  and	
  P-­‐cadherin	
  expression	
  pattern	
  
(p=0.659),	
  strength	
  of	
  P-­‐cadherin	
  immunoreactivity	
  (p=0.333),	
  or	
  degree	
  of	
  P-­‐cadherin	
  
expression	
  loss	
  (p=0.334);	
  or	
  depth	
  of	
  invasion	
  into	
  the	
  urinary	
  bladder	
  wall	
  and	
  P-­‐
cadherin	
  expression	
  pattern	
  (p=0.148),	
  strength	
  of	
  P-­‐cadherin	
  immunoreactivity	
  
(p=0.416),	
  or	
  degree	
  of	
  P-­‐cadherin	
  expression	
  loss	
  (p=0.357).	
  	
  
Immunohistochemistry:	
  β-­‐catenin	
  
None	
  of	
  the	
  ten	
  normal	
  urinary	
  bladder	
  control	
  samples	
  had	
  aberrant	
  expression	
  of	
  
β-­‐catenin,	
  while	
  18/33	
  urothelial	
  carcinomas	
  had	
  aberrant	
  expression	
  (Table	
  8).	
  	
  This	
  
difference	
  in	
  aberrant	
  labeling	
  for	
  β-­‐catenin	
  between	
  normal	
  urinary	
  bladder	
  controls	
  and	
  
urothelial	
  carcinomas	
  was	
  statistically	
  significant	
  (p=0.002).	
  	
  Evaluating	
  only	
  urothelial	
  
carcinomas,	
  there	
  were	
  significant	
  differences	
  in	
  aberrant	
  β-­‐catenin	
  expression	
  between	
  

	
  

79	
  

	
  
carcinomas	
  classified	
  as	
  infiltrating	
  or	
  papillary	
  (p=0.032),	
  but	
  no	
  significant	
  correlations	
  
with	
  anaplasia	
  grade	
  (0.102)	
  or	
  invasion	
  into	
  the	
  bladder	
  wall	
  (p=0.084).	
  
Survival	
  analysis	
  
	
  

Correlations	
  between	
  survival	
  time	
  and	
  clinico-­‐demographic	
  data	
  or	
  expression	
  of	
  

immunohistochemically	
  examined	
  markers	
  in	
  urothelial	
  carcinomas	
  as	
  defined	
  above	
  were	
  
evaluated	
  using	
  Kaplan-­‐Meier	
  statistics,	
  and	
  univariate	
  and	
  multivariate	
  analyses.	
  	
  For	
  
clinico-­‐demographic	
  data,	
  treatment	
  groups	
  were	
  defined	
  as	
  none,	
  only	
  local	
  resection,	
  
only	
  piroxicam,	
  and	
  combination	
  of	
  piroxicam,	
  surgical	
  resection,	
  and	
  systemic	
  
chemotherapy;	
  breed	
  groups	
  were	
  defined	
  as	
  Scottish	
  Terriers,	
  West	
  Highland	
  Whites,	
  
Beagles,	
  Shetland	
  Sheepdogs,	
  and	
  other	
  breeds;	
  and	
  sex	
  was	
  defined	
  as	
  female,	
  spayed	
  
female,	
  and	
  castrated	
  male	
  (there	
  were	
  no	
  intact	
  males	
  in	
  the	
  data	
  set).	
  
Using	
  Kaplan-­‐Meier	
  estimators	
  and	
  log	
  rank	
  tests,	
  there	
  was	
  no	
  significance	
  
difference	
  in	
  survival	
  time	
  between	
  groups	
  receiving	
  different	
  treatment	
  (p=0.578),	
  breed	
  
(p=0.739),	
  sex	
  (p=0.497),	
  anaplasia	
  grade	
  (p=0.239),	
  degree	
  of	
  infiltration	
  into	
  the	
  bladder	
  
wall	
  (p=0.230),	
  percentage	
  of	
  COX-­‐2	
  expressing	
  cells	
  within	
  tumors	
  (p=0.231),	
  E-­‐cadherin	
  
pattern	
  (p=0.103),	
  E-­‐cadherin	
  strength	
  (p=0.726),	
  E-­‐cadherin	
  loss	
  (p=0.401),	
  P-­‐cadherin	
  
strength	
  (p=0.963);	
  P-­‐cadherin	
  loss	
  (p=0.830).	
  	
  There	
  were	
  significant	
  differences	
  in	
  
survival	
  time	
  comparing	
  histopathologic	
  classification	
  of	
  urothelial	
  carcinomas	
  as	
  
infiltrating	
  or	
  papillary	
  (p=0.017),	
  percentage	
  of	
  HPGD	
  expressing	
  cells	
  (p=0.010),	
  HPGD	
  
expression	
  pattern	
  (p=0.042),	
  P-­‐cadherin	
  pattern	
  (p<0.001),	
  aberrant	
  P-­‐cadherin	
  
expression	
  (p=0.010),	
  and	
  aberrant	
  β-­‐catenin	
  expression	
  (0.004).	
  	
  	
  Examining	
  those	
  factors	
  
that	
  had	
  significant	
  correlations	
  with	
  survival	
  time,	
  Kaplan-­‐Meier	
  estimators	
  showed	
  
estimated	
  50%	
  cumulative	
  survival	
  times	
  of	
  237	
  days	
  (58.2	
  SE)	
  for	
  papillary	
  urothelial	
  

	
  

80	
  

	
  
carcinomas	
  compared	
  to	
  23	
  days	
  (28.8	
  SE)	
  for	
  infiltrating	
  urothelial	
  carcinomas;	
  of	
  482	
  
days	
  (224.4	
  SE)	
  for	
  urothelial	
  tumors	
  that	
  had	
  any	
  cells	
  expressing	
  HPGD	
  compared	
  to	
  183	
  
days	
  (74.9	
  SE)	
  for	
  urothelial	
  carcinomas	
  that	
  had	
  complete	
  lack	
  of	
  HPGD	
  expression;	
  of	
  618	
  
days	
  for	
  urothelial	
  carcinomas	
  with	
  HPGD	
  Pattern	
  2	
  expression	
  and	
  539	
  days	
  (330.7	
  SE)	
  
for	
  HPGD	
  Pattern	
  1	
  compared	
  to	
  258	
  (205.0	
  SE)	
  for	
  HPGD	
  Pattern	
  4	
  and	
  183	
  days	
  (73.9	
  SE)	
  
for	
  urothelial	
  carcinomas	
  with	
  lack	
  of	
  HPGD	
  expression;	
  of	
  267	
  days	
  (104.4	
  SE)	
  for	
  
urothelial	
  tumors	
  with	
  P-­‐cadherin	
  Pattern	
  1	
  compared	
  to	
  1	
  day	
  (1.1	
  SE)	
  for	
  urothelial	
  
tumors	
  with	
  P-­‐cadherin	
  pattern	
  2	
  or	
  3;	
  of	
  258	
  days	
  (152.7	
  SE)	
  for	
  urothelial	
  carcinomas	
  
with	
  non-­‐aberrant	
  P-­‐cadherin	
  expression	
  compared	
  to	
  1	
  day	
  (6.7	
  SE)	
  for	
  urothelial	
  
carcinomas	
  with	
  aberrant	
  P-­‐cadherin	
  expression;	
  and	
  of	
  618	
  days	
  (292.2	
  SE)	
  for	
  urothelial	
  
carcinomas	
  with	
  non-­‐aberrant	
  β-­‐catenin	
  expression	
  compared	
  to	
  62	
  days	
  (120.9	
  SE)	
  for	
  
urothelial	
  carcinomas	
  with	
  aberrant	
  β-­‐catenin	
  expression.	
  
Using	
  univariate	
  analysis	
  to	
  evaluate	
  correlations	
  between	
  survival	
  time	
  and	
  clinic	
  
pathologic	
  information	
  or	
  expression	
  of	
  immunohistochemically	
  examined	
  markers	
  in	
  
urothelial	
  carcinomas,	
  there	
  were	
  significant	
  correlations	
  between	
  survival	
  time	
  and	
  
histologic	
  classification	
  as	
  infiltrating	
  or	
  papillary,	
  percentage	
  of	
  HPGD	
  expressing	
  cells,	
  
pattern	
  of	
  HPGD	
  expression	
  within	
  tumors,	
  percentage	
  of	
  COX-­‐2	
  expressing	
  cells,	
  pattern	
  of	
  
P-­‐cadherin	
  expression,	
  and	
  aberrant	
  expression	
  of	
  β-­‐catenin	
  (Table	
  9).	
  In	
  multivariate	
  
analysis	
  including	
  only	
  the	
  factors	
  listed	
  above	
  that	
  had	
  statistical	
  significance	
  in	
  univariate	
  
analysis,	
  only	
  percentage	
  of	
  HPGD	
  expressing	
  cells	
  and	
  pattern	
  of	
  P-­‐cadherin	
  expression	
  
retained	
  statistical	
  significance	
  (Table	
  9).	
  
Western	
  blot	
  analysis	
  of	
  human	
  and	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines	
  

	
  

81	
  

	
  
All	
  antibodies	
  yielded	
  bands	
  of	
  expected	
  molecular	
  weight	
  in	
  normal	
  canine	
  tissues	
  
demonstrating	
  the	
  applicability	
  of	
  these	
  antibodies	
  for	
  use	
  in	
  the	
  dog.	
  Representative	
  
results	
  of	
  Western	
  blots	
  are	
  presented	
  in	
  Figure	
  6.	
  	
  None	
  of	
  the	
  examined	
  canine	
  urothelial	
  
carcinoma	
  cell	
  lines	
  expressed	
  HPGD.	
  	
  While	
  expression	
  varied	
  mildly	
  between	
  canine	
  
urothelial	
  carcinoma	
  cell	
  lines,	
  all	
  examined	
  cell	
  lines	
  retained	
  E-­‐	
  and	
  P-­‐cadherin	
  
expression	
  and	
  failed	
  to	
  express	
  N-­‐cadherin	
  (Figure	
  15).	
  Human	
  urothelial	
  carcinoma	
  cell	
  
lines	
  that	
  expressed	
  HPGD	
  (RT4,	
  UC14)	
  also	
  expressed	
  E-­‐	
  and	
  P-­‐cadherin,	
  but	
  failed	
  to	
  
express	
  N-­‐cadherin.	
  Conversely,	
  human	
  urothelial	
  carcinoma	
  cell	
  lines	
  that	
  did	
  not	
  express	
  
HPGD	
  (UC2,	
  UC3,	
  UC6,	
  UC12,	
  5367)	
  had	
  no	
  expression	
  of	
  E-­‐	
  and	
  P–cadherin	
  expression,	
  but	
  
expressed	
  N-­‐cadherin.	
  All	
  canine	
  and	
  human	
  urothelial	
  carcinoma	
  cell	
  lines	
  expressed	
  COX-­‐
2	
  and	
  β-­‐catenin	
  with	
  mild	
  variations	
  in	
  expression	
  between	
  cell	
  lines.	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

82	
  

	
  
Discussion	
  
Overall,	
  normal	
  canine	
  urinary	
  bladders	
  and	
  low	
  grade,	
  noninvasive	
  canine	
  
urothelial	
  carcinomas	
  typically	
  expressed	
  HPGD	
  in	
  superficial	
  epithelial	
  cells,	
  expressed	
  no	
  
COX-­‐2,	
  and	
  predominately	
  expressed	
  E-­‐	
  and	
  P-­‐cadherin	
  and	
  β-­‐catenin	
  along	
  cell	
  
membranes.	
  In	
  contrast,	
  a	
  significant	
  number	
  of	
  high	
  grade	
  urothelial	
  carcinomas	
  had	
  loss	
  
of	
  HPGD	
  and	
  increased	
  COX-­‐2	
  expression.	
  In	
  such	
  tumors,	
  P-­‐cadherin	
  was	
  occasionally	
  lost	
  
or	
  was	
  expressed	
  aberrantly	
  in	
  the	
  cytoplasm	
  and	
  β-­‐catenin	
  was	
  aberrantly	
  localized	
  
within	
  the	
  cytoplasm	
  or	
  nucleus,	
  while	
  membrane-­‐associated	
  E-­‐cadherin	
  expression	
  
remained.	
  N-­‐cadherin	
  was	
  not	
  expressed	
  by	
  canine	
  normal	
  urothelium	
  or	
  urothelial	
  
carcinomas.	
  
In	
  humans,	
  HPGD	
  has	
  been	
  shown	
  to	
  be	
  important	
  in	
  urothelial	
  differentiation	
  and	
  
its	
  expression	
  is	
  decreased	
  in	
  urothelial	
  malignancies.12,	
  19	
  In	
  canine	
  urothelial	
  carcinomas	
  
of	
  the	
  current	
  study,	
  while	
  expression	
  of	
  HPGD	
  in	
  urothelial	
  carcinomas	
  was	
  not	
  correlated	
  
with	
  histologic	
  classification,	
  grading,	
  or	
  depth	
  of	
  invasion	
  into	
  the	
  bladder	
  wall,	
  there	
  was	
  
an	
  association	
  between	
  HPGD	
  expression	
  and	
  prognosis	
  in	
  terms	
  of	
  survival	
  time	
  post	
  
diagnosis.	
  	
  COX-­‐2	
  overexpression	
  has	
  been	
  described	
  in	
  a	
  many	
  urothelial	
  carcinomas	
  of	
  
humans	
  and	
  increased	
  expression	
  is	
  associated	
  with	
  invasiveness,	
  metastasis,	
  and	
  
increased	
  mortality.1,	
  7,	
  24,	
  25	
  Previous	
  studies	
  of	
  COX-­‐2	
  expression	
  have	
  yielded	
  similar	
  
results	
  in	
  dogs.4	
  Consistent	
  with	
  these	
  reports,	
  there	
  were	
  positive	
  correlations	
  between	
  
COX-­‐2	
  expression	
  and	
  degree	
  of	
  anaplasia	
  and	
  depth	
  of	
  invasion	
  into	
  the	
  urinary	
  bladder	
  
wall	
  in	
  urothelial	
  carcinomas;	
  however,	
  COX-­‐2	
  expression	
  was	
  not	
  correlated	
  with	
  survival	
  
time.	
  These	
  data	
  suggest	
  that	
  pathways	
  regulating	
  prostaglandin	
  E2	
  are	
  altered	
  in	
  canine	
  
urothelial	
  carcinomas,	
  and	
  that	
  such	
  alterations	
  likely	
  play	
  roles	
  in	
  progression.	
  

	
  

83	
  

	
  
Decreased	
  expression	
  of	
  E-­‐cadherin	
  and	
  increased	
  expression	
  of	
  other	
  cadherins,	
  or	
  
cadherin	
  switching,	
  result	
  in	
  decreased	
  strength	
  of	
  cell-­‐cell	
  adhesions	
  and	
  an	
  increased	
  
propensity	
  for	
  cellular	
  migration.14-­‐16	
  In	
  terms	
  of	
  carcinogenesis,	
  cadherin	
  switching	
  is	
  
observed	
  with	
  epithelial-­‐to-­‐mesenchymal	
  transition	
  and	
  has	
  been	
  associated	
  with	
  
increased	
  invasiveness	
  and	
  metastasis.13-­‐16	
  	
  Epithelial-­‐to-­‐mesenchymal	
  transition	
  has	
  been	
  
well	
  described	
  in	
  human	
  bladder	
  cancers	
  and	
  has	
  been	
  suggested	
  to	
  involve	
  decreased	
  E-­‐
cadherin	
  expression	
  and	
  gain	
  of	
  N-­‐cadherin	
  and	
  P-­‐cadherin	
  expression.26-­‐32	
  Increased	
  P-­‐
cadherin	
  expression,	
  in	
  particular,	
  has	
  been	
  associated	
  with	
  a	
  significantly	
  worse	
  bladder	
  
cancer-­‐specific	
  survival	
  and	
  a	
  more	
  malignant	
  and	
  invasive	
  cancer	
  phenotype	
  in	
  humans.29,	
  
33	
  

In	
  5/36	
  canine	
  urothelial	
  carcinomas	
  of	
  the	
  evaluated	
  tumor	
  set,	
  there	
  was	
  loss	
  of	
  
preimembranous	
  association	
  of	
  E-­‐cadherin	
  expression	
  with	
  translocation	
  of	
  E-­‐cadherin	
  
expression	
  into	
  the	
  cytoplasm,	
  and	
  in	
  6/36	
  cases	
  there	
  were	
  significant	
  areas	
  of	
  E-­‐cadherin	
  
expression;	
  however,	
  differences	
  in	
  E-­‐cadherin	
  expression	
  were	
  not	
  statistically	
  significant	
  
between	
  normal	
  urothelium	
  and	
  urothelial	
  carcinomas,	
  or	
  between	
  various	
  histologic	
  
classification,	
  grading,	
  or	
  depth	
  of	
  invasion	
  into	
  the	
  bladder	
  all.	
  	
  Further,	
  no	
  correlations	
  
between	
  E-­‐cadherin	
  expression	
  and	
  survival	
  time	
  were	
  found.	
  In	
  contrast	
  to	
  E-­‐cadherin,	
  
differences	
  in	
  P-­‐cadherin	
  expression	
  between	
  normal	
  urotheluim	
  and	
  urothelial	
  
carcinomas	
  and	
  between	
  infiltrative	
  and	
  papillary	
  urothelial	
  carcinomas	
  had	
  statistical	
  
significance.	
  	
  In	
  addition,	
  loss	
  of	
  P-­‐cadherin	
  expression	
  and	
  aberrant	
  P-­‐cadherin	
  were	
  
negatively	
  associated	
  with	
  survival	
  time.	
  	
  The	
  absence	
  of	
  N-­‐cadherin	
  expression	
  in	
  all	
  
examined	
  canine	
  urothelial	
  carcinomas	
  and	
  within	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines	
  
contrasts	
  with	
  our	
  findings	
  in	
  human	
  urothelial	
  cell	
  lines	
  and	
  previous	
  reports	
  in	
  human	
  

	
  

84	
  

	
  
urothelial	
  carcinomas.	
  	
  These	
  data	
  suggest	
  that	
  while	
  there	
  may	
  be	
  some	
  alteration	
  in	
  
cadherin	
  expression	
  in	
  canine	
  urothelial	
  carcinomas,	
  cadherin	
  switching	
  as	
  has	
  it	
  been	
  
described	
  in	
  human	
  urothelial	
  carcinomas,	
  does	
  not	
  occur.	
  Despite	
  this	
  difference	
  between	
  
humans	
  and	
  dogs,	
  aberrant	
  expression	
  or	
  loss	
  of	
  P-­‐cadherin	
  likely	
  plays	
  a	
  role	
  in	
  the	
  
clinical	
  outcome	
  of	
  canine	
  urothelial	
  carcinomas.	
  
In	
  human	
  bladder	
  cancer,	
  up-­‐regulation	
  or	
  more	
  specifically	
  altered	
  distribution	
  of	
  
expression	
  with	
  nuclear	
  localization	
  of	
  β-­‐catenin	
  has	
  been	
  suggested	
  to	
  be	
  a	
  prognostic	
  
indicator	
  which	
  has	
  been	
  variably	
  negatively	
  associated	
  with	
  grade,	
  stage,	
  survival,	
  and/or	
  
recurrence.10,	
  34-­‐40	
  Aberrant	
  expression	
  (increased	
  cytoplasmic	
  and/or	
  nuclear	
  expression)	
  
of	
  β-­‐catenin	
  was	
  common	
  in	
  canine	
  urothelial	
  carcinomas,	
  and	
  in	
  particular,	
  in	
  infiltrating	
  
urothelial	
  carcinomas.	
  Further,	
  aberrant	
  β-­‐catenin	
  expression	
  was	
  strongly	
  associated	
  with	
  
decreased	
  survival	
  time	
  of	
  dogs	
  with	
  urothelial	
  carcinomas	
  by	
  univariate	
  statistical	
  
analysis,	
  but	
  lost	
  significance	
  with	
  multivariate	
  analysis.	
  	
  
It	
  has	
  recently	
  been	
  shown	
  that	
  cadherin	
  expression	
  in	
  normal	
  urotheluim	
  and	
  
urothelial	
  carcinomas	
  is	
  integrally	
  linked	
  to	
  prostaglandin	
  E2	
  regulation	
  pathways.	
  In	
  rats,	
  
partial	
  bladder	
  outlet	
  obstruction	
  results	
  in	
  increased	
  expression	
  of	
  COX-­‐2	
  and	
  decreased	
  
expression	
  of	
  E-­‐cadherin,	
  which	
  was	
  subsequently	
  remediated	
  by	
  COX-­‐2	
  inhibitors.41	
  In	
  
human	
  urothelial	
  carcinomas,	
  a	
  negative	
  correlation	
  between	
  COX-­‐2	
  and	
  E-­‐cadherin	
  
expression	
  has	
  been	
  suggested	
  with	
  high	
  grade	
  tumors	
  often	
  having	
  both	
  high	
  levels	
  of	
  
COX-­‐2	
  expression	
  and	
  decreased	
  E-­‐cadherin	
  expression	
  or	
  reduced	
  membranous	
  
association	
  of	
  E-­‐cadherin	
  expression.10	
  	
  Further,	
  in	
  vitro	
  studies	
  have	
  shown	
  that	
  treatment	
  
with	
  COX-­‐2	
  promoters	
  can	
  result	
  in	
  reduced	
  E-­‐cadherin	
  expression	
  and	
  that	
  knockdown	
  of	
  
COX-­‐2	
  results	
  in	
  increased	
  E-­‐cadherin	
  expression.10	
  PGDH	
  expression,	
  on	
  the	
  other	
  hand,	
  

	
  

85	
  

	
  
was	
  shown	
  to	
  increase	
  in	
  expression	
  with	
  urothelial	
  differentiation,	
  and	
  inhibition	
  of	
  PGDH	
  
expression	
  resulted	
  in	
  disruption	
  of	
  E-­‐cadherin	
  expression	
  at	
  cell-­‐cell	
  junction	
  in	
  cell	
  
lines.12	
  	
  
In	
  addition,	
  because	
  cadherins	
  form	
  associations	
  with	
  other	
  proteins	
  within	
  cells,	
  
they	
  are	
  integrally	
  linked	
  to	
  intracellular	
  signaling	
  and	
  trafficking.	
  β-­‐catenin	
  binds	
  to	
  the	
  
cytoplasmic	
  tail	
  of	
  cadherin	
  molecules.	
  In	
  this	
  capacity	
  it	
  functions	
  in	
  cell-­‐cell	
  adhesion,	
  but	
  
it	
  is	
  also	
  involved	
  in	
  a	
  variety	
  of	
  signaling	
  pathways	
  including	
  some	
  that	
  are	
  involved	
  in	
  
carcinogenesis,	
  such	
  as	
  the	
  Wnt	
  pathways.14,	
  42,	
  43	
  While	
  loss	
  of	
  cadherin	
  expression	
  alone	
  
has	
  not	
  been	
  shown	
  to	
  directly	
  result	
  in	
  intracellular	
  signaling,	
  altered	
  expression	
  of	
  
cadherins	
  has	
  been	
  shown	
  to	
  amplify	
  or	
  buffer	
  the	
  effects	
  of	
  such	
  pathways.14,	
  42,	
  43	
  Given	
  
the	
  associations	
  between	
  prostaglandin	
  E2	
  regulation	
  pathways	
  and	
  cadherin	
  expression,	
  
there	
  are	
  likely	
  also	
  associations	
  between	
  such	
  pathways	
  and	
  β-­‐catenin	
  regulation.	
  In	
  high	
  
grade	
  human	
  urothelial	
  carcinomas,	
  expression	
  of	
  β-­‐catenin	
  has	
  been	
  positively	
  correlated	
  
with	
  COX-­‐2	
  expression.10	
  Rather	
  than	
  being	
  only	
  a	
  consequence	
  of	
  altered	
  prostaglandin	
  E2	
  
regulation,	
  altered	
  β-­‐catenin	
  expression	
  and	
  Wnt	
  signaling	
  may	
  also	
  drive	
  alterations	
  in	
  
prostaglandin	
  E2	
  regulation.	
  	
  Recent	
  work	
  has	
  suggested	
  that	
  β-­‐catenin	
  expression	
  is	
  
inversely	
  related	
  to	
  the	
  expression	
  of	
  HPGD	
  at	
  least	
  in	
  intestinal	
  epithelium	
  and	
  within	
  
colorectal	
  tumor	
  cell	
  lines.11	
  
In	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines,	
  expression	
  of	
  HPGD	
  and	
  COX-­‐2	
  was	
  not	
  
associated	
  with	
  cadherin	
  switching	
  or	
  obvious	
  differences	
  in	
  β-­‐catenin	
  expression.	
  In	
  
canine	
  urothelial	
  carcinoma	
  cell	
  lines,	
  there	
  was	
  no	
  appreciable	
  expression	
  of	
  HPGD	
  and	
  
variable,	
  but	
  often	
  strong	
  expression	
  of	
  COX-­‐2.	
  	
  E-­‐cadherin	
  and	
  P-­‐cadherin	
  were	
  expressed	
  
in	
  all	
  canine	
  cell	
  lines,	
  while	
  there	
  was	
  no	
  appreciable	
  expression	
  of	
  N-­‐cadherin.	
  In	
  contrast	
  

	
  

86	
  

	
  
to	
  canine	
  urothelial	
  carcinomas	
  and	
  urothelial	
  carcinoma	
  cell	
  lines,	
  cadherins	
  were	
  
differentially	
  expressed	
  in	
  human	
  urothelial	
  cell	
  lines	
  and	
  cadherin	
  expression	
  pattern	
  was	
  
correlated	
  with	
  HPGD	
  status.	
  Specifically,	
  in	
  examined	
  human	
  urothelial	
  carcinoma	
  cell	
  
lines,	
  those	
  cell	
  lines	
  that	
  expressed	
  HPGH	
  also	
  expressed	
  E-­‐cadherin	
  and	
  P-­‐cadherin.	
  	
  Lack	
  
of	
  HPGD	
  expression	
  in	
  human	
  urothelial	
  carcinoma	
  cell	
  lines	
  was	
  consistently	
  associated	
  
with	
  gain	
  of	
  N-­‐cadherin	
  expression.	
  	
  Thus,	
  while	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines	
  
appear	
  similar	
  to	
  human	
  urothelial	
  carcinomas	
  in	
  terms	
  of	
  the	
  common	
  loss	
  of	
  HPGD	
  
expression	
  and	
  expression	
  of	
  COX-­‐2,	
  they	
  are	
  dissimilar	
  in	
  the	
  lack	
  of	
  associated	
  overt	
  
cadherin	
  switching.	
  
While	
  the	
  results	
  of	
  this	
  study	
  are	
  intriguing	
  and	
  statistical	
  correlations	
  between	
  
many	
  factors	
  and	
  survival	
  time	
  appear	
  strong,	
  this	
  study	
  is	
  limited	
  by	
  the	
  relative	
  low	
  
number	
  of	
  cases	
  and,	
  in	
  particular,	
  the	
  low	
  number	
  of	
  low	
  grade	
  urothelial	
  carcinomas.	
  
High	
  grade	
  urothelial	
  carcinomas	
  are	
  far	
  more	
  commonly	
  diagnosed	
  in	
  dogs	
  than	
  low	
  grade	
  
tumors	
  accounting	
  for	
  the	
  low	
  number	
  of	
  low	
  grade	
  tumors	
  in	
  the	
  examined	
  set.	
  This	
  may	
  
be	
  due	
  to	
  an	
  actual	
  lower	
  rate	
  of	
  occurrence	
  in	
  canine	
  populations	
  or	
  diagnosis	
  of	
  urothelial	
  
neoplasms	
  only	
  late	
  in	
  the	
  course	
  of	
  disease	
  when	
  larger,	
  more	
  aggressive	
  tumors	
  affect	
  
micturition	
  or	
  cause	
  other	
  clinical	
  disease.	
  Further,	
  euthanasia	
  is	
  common	
  in	
  veterinary	
  
medicine	
  and	
  is	
  especially	
  common	
  in	
  animals	
  with	
  urothelial	
  carcinomas	
  due	
  to	
  the	
  
negative	
  prognosis	
  historically	
  associated	
  with	
  such	
  tumors,	
  the	
  effects	
  on	
  micturition,	
  the	
  
common	
  non-­‐respectable	
  nature	
  such	
  tumors	
  at	
  the	
  time	
  of	
  diagnosis,	
  and	
  client	
  choice	
  due	
  
to	
  financial	
  or	
  homecare	
  constraints.	
  
Despite	
  the	
  inherent	
  limitations,	
  this	
  work	
  shows	
  that	
  expression	
  of	
  HPGD,	
  COX-­‐2,	
  
cadherins,	
  and	
  β-­‐catenin	
  is	
  altered	
  in	
  at	
  least	
  subsets	
  of	
  canine	
  urothelial	
  carcinomas;	
  

	
  

87	
  

	
  
however,	
  the	
  patterns	
  of	
  alterations	
  is	
  different	
  than	
  that	
  previously	
  described	
  in	
  humans	
  
or	
  observed	
  in	
  human	
  cell	
  lines	
  in	
  the	
  current	
  study.	
  This	
  suggests	
  that	
  while	
  human	
  and	
  
canine	
  urothelial	
  carcinoma	
  may	
  be	
  similar	
  in	
  terms	
  of	
  clinico-­‐pathologic	
  features,	
  drivers	
  
of	
  carcinogenesis	
  and	
  progression	
  of	
  these	
  tumors	
  may	
  differ	
  between	
  humans	
  and	
  dogs.	
  	
  
While	
  COX-­‐2	
  and	
  HPGD	
  have	
  been	
  suggested	
  to	
  be	
  correlated	
  with	
  features	
  of	
  epithelial-­‐to-­‐
mesenchymal	
  transition	
  such	
  as	
  cadherin	
  switching	
  in	
  humans,	
  such	
  may	
  not	
  be	
  the	
  case	
  in	
  
dogs.	
  	
  As	
  the	
  molecular	
  pathways	
  which	
  were	
  evaluated	
  in	
  this	
  study	
  have	
  potential	
  as	
  
therapeutic	
  targets,	
  developing	
  a	
  better	
  understanding	
  why	
  such	
  there	
  are	
  differences	
  
between	
  dogs	
  and	
  humans	
  and	
  how	
  this	
  impacts	
  treatment	
  response	
  through	
  additional	
  
study	
  could	
  lead	
  to	
  the	
  development	
  of	
  better	
  clinical	
  strategies	
  to	
  address	
  canine	
  and	
  
human	
  urothelial	
  cancers.	
  
	
  
Acknowledgments	
  
Special	
  thanks	
  are	
  due	
  to	
  Dr.	
  Deborah	
  Knapp	
  of	
  Purdue	
  University	
  for	
  contributing	
  5	
  
of	
  the	
  8	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines	
  evaluated	
  in	
  this	
  study.	
  Dr.	
  Sledge’s	
  graduate	
  
program	
  was	
  funded	
  by	
  Bristol-­‐Meyers-­‐Squibb	
  through	
  the	
  American	
  College	
  of	
  Veterinary	
  
Pathologists/Society	
  of	
  Toxicologic	
  Pathologists	
  coalition.	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

88	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
APPENDIX	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

89	
  

	
  
Figure	
  10:	
  Urinary	
  bladder,	
  Dog,	
  15-­‐hydroxyprostaglandin	
  dehydrogenase	
  (HPGD)	
  
immunohistochemistry	
  (IHC).	
  Immunoreactivity	
  (brown	
  labeling)	
  was	
  defined	
  by	
  
cytoplasmic	
  and/or	
  nuclear	
  labeling	
  and	
  categorized	
  by	
  the	
  percentage	
  of	
  positively	
  
labeled	
  urothelial	
  cells	
  and	
  the	
  overall	
  pattern	
  of	
  immunoreactivity	
  within	
  the	
  urothelium.	
  
Immunoreactivity	
  is	
  distributed	
  diffusely	
  throughout	
  100%	
  of	
  cells	
  in	
  superficial	
  1-­‐2	
  cell	
  
layers	
  of	
  the	
  urothelium	
  consistent	
  with	
  HPGD	
  Pattern	
  1	
  in	
  the	
  urinary	
  bladder	
  of	
  a	
  healthy	
  
control	
  dog	
  consistent	
  with	
  Pattern	
  1	
  (A).	
  In	
  a	
  grade	
  I	
  (low	
  grade)	
  papillary	
  carcinoma,	
  
there	
  is	
  immunoreactivity	
  in	
  20%	
  of	
  the	
  total	
  neoplastic	
  urothelial	
  cells	
  and	
  
immunoreactivity	
  is	
  patchy	
  and	
  randomly	
  distributed	
  throughout	
  <80%	
  of	
  cells	
  in	
  the	
  
superficial	
  most	
  1-­‐2	
  cell	
  layers	
  consistent	
  with	
  HPGD	
  Pattern	
  2	
  (B).	
  In	
  a	
  grade	
  III	
  (high	
  
grade)	
  invasive	
  carcinoma	
  extending	
  into	
  the	
  muscularis,	
  there	
  is	
  lack	
  of	
  appreciable	
  HPGD	
  
immunoreactivity	
  (C).	
  3,3’-­‐Diaminobenzidine	
  (DAB)	
  chromogen,	
  hematoxylin	
  counterstain.	
  

	
  
	
  
	
  
	
  
	
  

	
  

	
  

90	
  

	
  
Figure	
  11:	
  Urinary	
  bladder,	
  Dog,	
  Cyclooxygenase-­‐2	
  (COX-­‐2)	
  IHC.	
  Immunoreactivity	
  (red)	
  
was	
  defined	
  by	
  cytoplasmic	
  labeling	
  and	
  categorized	
  by	
  the	
  percentage	
  of	
  positive	
  
urothelial	
  cells.	
  Urothelium	
  from	
  the	
  urinary	
  bladder	
  of	
  a	
  healthy	
  control	
  dog	
  (A)	
  and	
  
neoplastic	
  urothelium	
  of	
  a	
  grade	
  1	
  (low	
  grade)	
  papillary	
  carcinoma	
  (B)	
  lack	
  appreciable	
  
immunoreactivity.	
  	
  There	
  is	
  strong	
  immunoreactivity	
  for	
  COX-­‐2	
  in	
  >30%	
  of	
  neoplastic	
  cells	
  
of	
  a	
  grade	
  III	
  (high	
  grade)	
  invasive	
  carcinoma	
  with	
  infiltration	
  of	
  the	
  muscularis	
  (C).	
  
Indirect	
  biotin	
  streptavidin	
  and	
  Fast	
  Red	
  chromogen,	
  hematoxylin	
  counterstain.	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

91	
  

	
  
Figure	
  12:	
  Urinary	
  bladder,	
  Dog,	
  E-­‐cadherin	
  IHC.	
  Immunoreactivity	
  (brown)	
  was	
  
subjectively	
  categorized	
  as	
  weak,	
  moderate,	
  or	
  strong;	
  assessed	
  according	
  to	
  the	
  
predominant	
  pattern	
  within	
  urothelial	
  cells	
  as	
  membrane	
  associated,	
  intracytoplasmic,	
  or	
  
mixed	
  membranous	
  and	
  cytoplasmic;	
  and	
  categorized	
  by	
  degree	
  of	
  loss	
  as	
  no	
  loss,	
  patchy	
  
loss,	
  or	
  complete	
  loss.	
  	
  In	
  urothelium	
  from	
  the	
  urinary	
  bladder	
  of	
  a	
  healthy	
  control	
  dog,	
  
there	
  is	
  strong	
  membrane	
  associated	
  immunoreactivity	
  throughout	
  the	
  full	
  urothelial	
  
thickness	
  and	
  no	
  loss	
  (A).	
  While	
  there	
  is	
  strong	
  membrane	
  associated	
  immunoreactivity	
  in	
  
superficial	
  cell	
  layers	
  in	
  a	
  grade	
  II	
  (high	
  grade)	
  papillary	
  carcinoma	
  with	
  substantia	
  propria	
  
infiltration,	
  the	
  predominant	
  pattern	
  is	
  moderate	
  cytoplasmic	
  immunoreactivity	
  with	
  
patchy	
  areas	
  of	
  loss	
  (B).	
  	
  In	
  a	
  grade	
  III	
  (high	
  grade)	
  invasive	
  carcinoma	
  with	
  infiltration	
  into	
  
the	
  muscularis,	
  there	
  is	
  strong	
  membrane	
  associated	
  immunoreactivity	
  in	
  90%	
  of	
  
neoplastic	
  cells	
  and	
  patchy	
  loss	
  (C).	
  DAB	
  chromogen,	
  hematoxylin	
  counterstain.	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

92	
  

	
  
Figure	
  13:	
  Urinary	
  bladder,	
  Dog,	
  P-­‐cadherin	
  IHC.	
  Immunoreactivity	
  (brown)	
  was	
  
subjectively	
  categorized	
  as	
  weak,	
  moderate,	
  or	
  strong;	
  assessed	
  according	
  to	
  the	
  
predominant	
  pattern	
  within	
  urothelial	
  cells	
  as	
  membrane	
  associated,	
  intracytoplasmic,	
  or	
  
mixed	
  membranous	
  and	
  cytoplasmic	
  labeling;	
  and	
  categorized	
  by	
  degree	
  of	
  loss	
  as	
  no	
  loss,	
  
patchy	
  loss,	
  or	
  complete	
  loss.	
  	
  In	
  urothelium	
  from	
  the	
  urinary	
  bladder	
  of	
  a	
  healthy	
  control	
  
dog,	
  there	
  is	
  strong	
  membrane	
  associated	
  immunoreactivity	
  throughout	
  the	
  full	
  urothelial	
  
thickness	
  and	
  no	
  loss	
  (A).	
  	
  In	
  a	
  grade	
  I	
  (low	
  grade)	
  papillary	
  carcinoma,	
  there	
  is	
  moderate	
  
to	
  weak	
  membrane	
  associated	
  immunoreactivity	
  and	
  patchy	
  loss	
  in	
  the	
  basal	
  cell	
  layers	
  
(B).	
  	
  In	
  a	
  grade	
  III	
  (high	
  grade)	
  invasive	
  carcinoma	
  with	
  infiltration	
  into	
  the	
  muscularis,	
  
there	
  is	
  predominant	
  moderate	
  intracytoplasmic	
  immunoreactivity	
  with	
  patchy	
  loss	
  (C).	
  
DAB	
  chromogen,	
  hematoxylin	
  counterstain.	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

93	
  

	
  
Figure	
  14:	
  Urinary	
  bladder,	
  Dog,	
  β-­‐catenin	
  IHC.	
  Immunoreactivity,	
  the	
  percentage	
  of	
  cells	
  
with	
  membrane,	
  cytoplasmic,	
  mixed	
  membrane	
  and	
  cytoplasmic,	
  and	
  nuclear	
  
immunoreactivity	
  (brown)	
  was	
  assessed.	
  	
  	
  In	
  urothelium	
  from	
  the	
  urinary	
  bladder	
  of	
  a	
  
healthy	
  control	
  dog,	
  there	
  is	
  membrane	
  associated	
  immunoreactivity	
  in	
  100%	
  of	
  cells	
  and	
  
no	
  cytoplasmic,	
  mixed	
  membrane	
  and	
  cytoplasmic,	
  or	
  nuclear	
  immunoreactivity	
  (A).	
  	
  In	
  a	
  
grade	
  I	
  (low	
  grade)	
  papillary	
  carcinoma,	
  there	
  is	
  membrane	
  associated	
  immunoreactivity	
  
in	
  90%	
  of	
  cells,	
  cytoplasmic	
  immunoreactivity	
  in	
  10%	
  of	
  cells,	
  and	
  no	
  mixed	
  membrane	
  and	
  
cytoplasmic	
  or	
  nuclear	
  immunoreactivity	
  (B).	
  In	
  a	
  grade	
  III	
  (high	
  grade)	
  invasive	
  
carcinoma	
  with	
  infiltration	
  into	
  the	
  muscularis,	
  there	
  is	
  primary	
  cytoplasmic	
  
immunoreactivity	
  in	
  85%	
  of	
  cells	
  and	
  nuclear	
  immunoreactivity	
  in	
  15%	
  of	
  cells	
  (C).	
  DAB	
  
chromogen,	
  hematoxylin	
  counterstain.	
  	
  
	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  
94	
  

	
  
Figure	
  15:	
  Western	
  blot	
  comparing	
  expression	
  of	
  HPGD,	
  COX-­‐2,	
  E-­‐cadherin,	
  P-­‐cadherin,	
  N-­‐
cadherin,	
  and	
  β-­‐catenin	
  in	
  human	
  (RT4	
  and	
  UC3)	
  and	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines	
  
(AXA,	
  AXB,	
  AXC,	
  and	
  NK).	
  	
  None	
  of	
  the	
  examined	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines	
  
expressed	
  HPGD,	
  all	
  retained	
  E-­‐	
  and	
  P-­‐cadherin	
  expression,	
  and	
  failed	
  to	
  express	
  N-­‐
cadherin.	
  Human	
  urothelial	
  carcinoma	
  cell	
  lines	
  that	
  expressed	
  HPGD	
  (RT4)	
  also	
  expressed	
  
E-­‐	
  and	
  P-­‐cadherin,	
  but	
  failed	
  to	
  express	
  N-­‐cadherin.	
  Conversely,	
  human	
  urothelial	
  
carcinoma	
  cell	
  lines	
  that	
  did	
  not	
  express	
  HPGD	
  (UC3)	
  had	
  no	
  expression	
  of	
  E-­‐	
  and	
  P–
cadherin	
  expression,	
  but	
  expressed	
  N-­‐cadherin.	
  While	
  expression	
  varied,	
  COX-­‐2	
  and	
  β-­‐
catenin	
  were	
  expressed	
  by	
  all	
  canine	
  and	
  human	
  urothelial	
  carcinoma	
  cell	
  lines.	
  	
  

	
  

95	
  

	
  
Table	
  7:	
  Expression	
  of	
  HPGD	
  in	
  canine	
  urothelial	
  carcinomas	
  and	
  normal	
  urothelium	
  with	
  respect	
  to	
  lesion	
  classification	
  and	
  
degree	
  of	
  invasion	
  
	
  	
  
	
  	
  
Classification	
  
Normal	
  urothelium	
  
Infiltrative	
  
carcinoma	
  
Papillary	
  carcinoma	
  

Percentage	
  of	
  urothelial	
  cells	
  expressing	
  HPGD	
  
None	
  
	
  	
  
1/10	
  
(10%)	
  
5/8	
  (63%)	
  
9/28	
  
(32%)	
  

	
  	
  	
  	
  	
  Grade	
  I	
  

1/4	
  (25%)	
  

	
  	
  	
  	
  	
  Grade	
  II	
  

6/16	
  
(38%)	
  

	
  	
  	
  	
  	
  Grade	
  III	
  

2/8	
  (25%)	
  

	
  	
  
Degree	
  of	
  Invasion	
  
No	
  invasion	
  
Substantia	
  propria	
  
Muscularis	
  

	
  

	
  	
  
	
  	
  
3/12	
  
(25%)	
  
6/12	
  
(50%)	
  
5/12	
  
(42%)	
  

<5%	
  
	
  	
  
2/10	
  
(20%)	
  
3/8	
  
(38%)	
  
5/28	
  
(18%)	
  

5-­‐15%	
  
	
  	
  
1/10	
  
(10%)	
  

15-­‐30%	
  
	
  	
  
3/10	
  
(30%)	
  

>30%	
  
	
  	
  
3/10	
  
(30%)	
  

-­‐	
  
2/16	
  
(13%)	
  
3/8	
  
(38%)	
  

-­‐	
  
6/28	
  
(21%)	
  
1/4	
  
(25%)	
  
3/16	
  
(19%)	
  
2/8	
  
(25%)	
  

-­‐	
  
4/28	
  
(14%)	
  
1/4	
  
(25%)	
  
2/16	
  
(13%)	
  
	
  1/8	
  
(13%)	
  

-­‐	
  
4/28	
  
(14%)	
  
1/4	
  
(25%)	
  
3/16	
  
(19%)	
  

	
  
3/12	
  
	
  
(25%)	
  
1/12	
  
(8%)	
  
4/12	
  
(33%)	
  

	
  
2/12	
  
	
  
(17%)	
  
2/12	
  
(17%)	
  
3/12	
  
(25%)	
  

	
  
2/12	
  
	
  
(17%)	
  
2/12	
  
(17%)	
  

-­‐	
  
	
  	
  
	
  	
  
2/12	
  
(17%)	
  
2/12	
  
(17%)	
  

-­‐	
  

-­‐	
  

96	
  

Pattern	
  of	
  HPGD	
  expression	
  
Pattern	
   Pattern	
  
Pattern	
  1	
  
2	
  
3	
  
Pattern	
  4	
  
	
  	
  
	
  	
  
	
  	
  
	
  	
  
8/10	
  
1/10	
  
(80%)	
  
(10%)	
  
-­‐	
  
-­‐	
  
1/8	
  
2/8	
  
-­‐	
  
(13%)	
  
-­‐	
  
(25%)	
  
3/28	
  
1/28	
  
15/28	
  
(11%)	
  
(4%)	
  
-­‐	
  
(54%)	
  
1/4	
  
2/4	
  
(25%)	
  
-­‐	
  
-­‐	
  
(50%)	
  
2/16	
  
1/16	
  
7/16	
  
(13%)	
  
(6%)	
  
-­‐	
  
(44%)	
  
6/8	
  
	
  	
  
-­‐	
  
-­‐	
  
(75%)	
  
	
  	
  
	
  	
  
	
  	
  
	
  
	
  
	
  	
  
3/12	
  
3/12	
  
	
  
2/12	
  
	
  
4/12	
  
(25%)	
  
(25%)	
  
(17%)	
  
(33%)	
  
6/12	
  
6/12	
  
(50%)	
  
-­‐	
  
-­‐	
  
(50%)	
  
5/12	
  
7/12	
  
(42%)	
  
-­‐	
  
-­‐	
  
(58%)	
  

	
  
Table	
  8:	
  Expression	
  of	
  P-­‐cadherin	
  and	
  β-­‐catenin	
  in	
  canine	
  urothelial	
  carcinomas	
  and	
  normal	
  urothelium	
  with	
  respect	
  to	
  lesion	
  
classification	
  and	
  degree	
  of	
  invasion	
  
	
  	
  
	
  	
  
Classification	
  
Normal	
  urothelium	
  
Infiltrative	
  carcinoma	
  
Papillary	
  carcinoma	
  
	
  	
  	
  	
  	
  Grade	
  I	
  
	
  	
  	
  	
  	
  Grade	
  II	
  
	
  	
  	
  	
  	
  Grade	
  III	
  
	
  	
  
Degree	
  of	
  Invasion	
  
No	
  invasion	
  
Substantia	
  propria	
  
Muscularis	
  

P-­‐cadherin	
  
Nonaberrant	
  
Aberrant	
  
	
  	
  
	
  	
  
10/10	
  (100%)	
  
3/7	
  (43%)	
  
23/25	
  (92%)	
  
4/4	
  (100%)	
  
13/14	
  (93%)	
  
6/7	
  (86%)	
  
	
  	
  
	
  	
  
12/12	
  (100%)	
  
8/10	
  (67%)	
  
6/10	
  (60%	
  

	
  β-­‐catenin	
  
Nonaberrant	
  
Aberrant	
  
	
  	
  
	
  	
  

-­‐	
  
4/7	
  (57%)	
  
2/25	
  (8%)	
  
-­‐	
  
1/14	
  (7%)	
  
1/7	
  (14%)	
  
	
  	
  
	
  	
  
-­‐	
  
2/10	
  (17%)	
  
4/10	
  (40%)	
  

	
  
	
  
	
  
	
  
	
  

	
  

97	
  

10/10	
  (100%)	
  
1/8	
  (12%)	
  
14/25	
  (56%)	
  
3/4	
  (75%)	
  
8/14	
  (57%)	
  
3/7	
  (43%)	
  
	
  	
  
	
  	
  
7/12	
  (58%)	
  
6/10	
  (60%)	
  
2/11	
  (18%)	
  

-­‐	
  
7/8	
  (88%)	
  
11/25	
  (44%)	
  
1/4	
  (25%)	
  
6/14	
  (43%)	
  
4/7	
  (57%)	
  
	
  	
  
	
  	
  
5/12	
  (42%)	
  
4/10	
  (40%)	
  
9/11	
  (82%)	
  

	
  
Table	
  9:	
  Significant	
  Correlative	
  Results	
  of	
  Univariate	
  and	
  Multivariate	
  Analysis	
  of	
  Examined	
  Variables	
  with	
  Respect	
  to	
  Survival	
  
Time	
  as	
  Determined	
  by	
  p-­‐values	
  less	
  than	
  0.05	
  
Univariate	
  Analysis	
  
Variable	
  	
  
Histologic	
  Classification	
  
HPGD:	
  %	
  of	
  positive	
  cells	
  
HPGD:	
  tissue	
  localization	
  
COX-­‐2:	
  %	
  of	
  positive	
  cells	
  
P-­‐cadherin:	
  cellular	
  localization	
  
β-­‐catenin:	
  aberrant	
  expression	
  

Multivariate	
  Analysis	
  
p-­‐value	
  
0.018	
  
0.006	
  
0.044	
  
0.042	
  
<0.001	
  
0.008	
  

Variable	
  	
  
HPGD:	
  %	
  of	
  positive	
  cells	
  
P-­‐cadherin:	
  cellular	
  localization	
  
	
  	
  
	
  	
  
	
  	
  
	
  	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

98	
  

p-­‐value	
  
0.026	
  
0.003	
  
	
  	
  
	
  	
  
	
  	
  
	
  	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
REFERENCES	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

99	
  

	
  
REFERENCES	
  
	
  
1.	
  Gee	
  J,	
  Lee	
  IL,	
  Grossman	
  HB,	
  Sabichi	
  AL.	
  Forced	
  COX-­‐2	
  expression	
  induces	
  PGE(2)	
  and	
  
invasion	
  in	
  immortalized	
  urothelial	
  cells.	
  Urol	
  Oncol.	
  2008;26:	
  641-­‐645.	
  
2.	
  Greenhough	
  A,	
  Smartt	
  HJ,	
  Moore	
  AE,	
  et	
  al.	
  The	
  COX-­‐2/PGE2	
  pathway:	
  key	
  roles	
  in	
  the	
  
hallmarks	
  of	
  cancer	
  and	
  adaptation	
  to	
  the	
  tumour	
  microenvironment.	
  Carcinogenesis.	
  
2009;30:	
  377-­‐386.	
  
3.	
  Muller-­‐Decker	
  K,	
  Furstenberger	
  G.	
  The	
  cyclooxygenase-­‐2-­‐mediated	
  prostaglandin	
  
signaling	
  is	
  causally	
  related	
  to	
  epithelial	
  carcinogenesis.	
  Mol	
  Carcinog.	
  2007;46:	
  705-­‐710.	
  
4.	
  Ono	
  M.	
  Molecular	
  links	
  between	
  tumor	
  angiogenesis	
  and	
  inflammation:	
  inflammatory	
  
stimuli	
  of	
  macrophages	
  and	
  cancer	
  cells	
  as	
  targets	
  for	
  therapeutic	
  strategy.	
  Cancer	
  Sci.	
  
2008;99:	
  1501-­‐1506.	
  
5.	
  Taylor	
  JA,	
  3rd,	
  Pilbeam	
  C,	
  Nisbet	
  A.	
  Role	
  of	
  the	
  prostaglandin	
  pathway	
  and	
  the	
  use	
  of	
  
NSAIDs	
  in	
  genitourinary	
  malignancies.	
  Expert	
  Rev	
  Anticancer	
  Ther.	
  2008;8:	
  1125-­‐1134.	
  
6.	
  Taylor	
  JA,	
  3rd,	
  Ristau	
  B,	
  Bonnemaison	
  M,	
  et	
  al.	
  Regulation	
  of	
  the	
  prostaglandin	
  pathway	
  
during	
  development	
  of	
  invasive	
  bladder	
  cancer	
  in	
  mice.	
  Prostaglandins	
  Other	
  Lipid	
  Mediat.	
  
2009;88:	
  36-­‐41.	
  
7.	
  Shariat	
  SF,	
  Matsumoto	
  K,	
  Kim	
  J,	
  et	
  al.	
  Correlation	
  of	
  cyclooxygenase-­‐2	
  expression	
  with	
  
molecular	
  markers,	
  pathological	
  features	
  and	
  clinical	
  outcome	
  of	
  transitional	
  cell	
  
carcinoma	
  of	
  the	
  bladder.	
  J	
  Urol.	
  2003;170:	
  985-­‐989.	
  
8.	
  Celis	
  JE,	
  Ostergaard	
  M,	
  Basse	
  B,	
  et	
  al.	
  Loss	
  of	
  adipocyte-­‐type	
  fatty	
  acid	
  binding	
  protein	
  
and	
  other	
  protein	
  biomarkers	
  is	
  associated	
  with	
  progression	
  of	
  human	
  bladder	
  transitional	
  
cell	
  carcinomas.	
  Cancer	
  Res.	
  1996;56:	
  4782-­‐4790.	
  
9.	
  Gee	
  JR,	
  Montoya	
  RG,	
  Khaled	
  HM,	
  Sabichi	
  AL,	
  Grossman	
  HB.	
  Cytokeratin	
  20,	
  AN43,	
  PGDH,	
  
and	
  COX-­‐2	
  expression	
  in	
  transitional	
  and	
  squamous	
  cell	
  carcinoma	
  of	
  the	
  bladder.	
  Urol	
  
Oncol.	
  2003;21:	
  266-­‐270.	
  
10.	
  Jang	
  TJ,	
  Cha	
  WH,	
  Lee	
  KS.	
  Reciprocal	
  correlation	
  between	
  the	
  expression	
  of	
  
cyclooxygenase-­‐2	
  and	
  E-­‐cadherin	
  in	
  human	
  bladder	
  transitional	
  cell	
  carcinomas.	
  Virchows	
  
Arch.	
  2010;457:	
  319-­‐328.	
  
11.	
  Smartt	
  HJ,	
  Greenhough	
  A,	
  Ordonez-­‐Moran	
  P,	
  et	
  al.	
  beta-­‐catenin	
  represses	
  expression	
  of	
  
the	
  tumour	
  suppressor	
  15-­‐prostaglandin	
  dehydrogenase	
  in	
  the	
  normal	
  intestinal	
  
epithelium	
  and	
  colorectal	
  tumour	
  cells.	
  Gut.	
  2012;61:	
  1306-­‐1314.	
  
12.	
  Tseng-­‐Rogenski	
  S,	
  Lee	
  IL,	
  Gebhardt	
  D,	
  et	
  al.	
  Loss	
  of	
  15-­‐hydroxyprostaglandin	
  
dehydrogenase	
  expression	
  disrupts	
  urothelial	
  differentiation.	
  Urology.	
  2008;71:	
  346-­‐350.	
  

	
  

100	
  

	
  
13.	
  Gloushankova	
  NA.	
  Changes	
  in	
  regulation	
  of	
  cell-­‐cell	
  adhesion	
  during	
  tumor	
  
transformation.	
  Biochemistry	
  (Mosc).	
  2008;73:	
  742-­‐750.	
  
14.	
  Jeanes	
  A,	
  Gottardi	
  CJ,	
  Yap	
  AS.	
  Cadherins	
  and	
  cancer:	
  how	
  does	
  cadherin	
  dysfunction	
  
promote	
  tumor	
  progression?	
  Oncogene.	
  2008;27:	
  6920-­‐6929.	
  
15.	
  Stemmler	
  MP.	
  Cadherins	
  in	
  development	
  and	
  cancer.	
  Mol	
  Biosyst.	
  2008;4:	
  835-­‐850.	
  
16.	
  Wheelock	
  MJ,	
  Shintani	
  Y,	
  Maeda	
  M,	
  Fukumoto	
  Y,	
  Johnson	
  KR.	
  Cadherin	
  switching.	
  J	
  Cell	
  
Sci.	
  2008;121:	
  727-­‐735.	
  
17.	
  Brouxhon	
  S,	
  Kyrkanides	
  S,	
  O'Banion	
  MK,	
  et	
  al.	
  Sequential	
  down-­‐regulation	
  of	
  E-­‐cadherin	
  
with	
  squamous	
  cell	
  carcinoma	
  progression:	
  loss	
  of	
  E-­‐cadherin	
  via	
  a	
  prostaglandin	
  E2-­‐EP2	
  
dependent	
  posttranslational	
  mechanism.	
  Cancer	
  Res.	
  2007;67:	
  7654-­‐7664.	
  
18.	
  Dohadwala	
  M,	
  Yang	
  SC,	
  Luo	
  J,	
  et	
  al.	
  Cyclooxygenase-­‐2-­‐dependent	
  regulation	
  of	
  E-­‐
cadherin:	
  prostaglandin	
  E(2)	
  induces	
  transcriptional	
  repressors	
  ZEB1	
  and	
  snail	
  in	
  non-­‐
small	
  cell	
  lung	
  cancer.	
  Cancer	
  Res.	
  2006;66:	
  5338-­‐5345.	
  
19.	
  Tseng-­‐Rogenski	
  S,	
  Gee	
  J,	
  Ignatoski	
  KW,	
  et	
  al.	
  Loss	
  of	
  15-­‐hydroxyprostaglandin	
  
dehydrogenase	
  expression	
  contributes	
  to	
  bladder	
  cancer	
  progression.	
  Am	
  J	
  Pathol.	
  
2010;176:	
  1462-­‐1468.	
  
20.	
  Starkey	
  MP,	
  Scase	
  TJ,	
  Mellersh	
  CS,	
  Murphy	
  S.	
  Dogs	
  really	
  are	
  man's	
  best	
  friend-­‐-­‐canine	
  
genomics	
  has	
  applications	
  in	
  veterinary	
  and	
  human	
  medicine!	
  Brief	
  Funct	
  Genomic	
  
Proteomic.	
  2005;4:	
  112-­‐128.	
  
21.	
  Knapp	
  DW.	
  Animal	
  models:	
  naturally	
  occuring	
  canine	
  urinary	
  bladder	
  cancer.	
  In:	
  Lerner	
  
SP,	
  Schoenberg	
  MP,	
  Sternberg	
  CN,	
  editors.	
  Textbook	
  of	
  Bladder	
  Cancer.	
  Oxon,	
  UK:	
  Taylor	
  
and	
  Francis,	
  2006:171-­‐178.	
  
22.	
  Patrick	
  DJ,	
  Fitzgerald	
  SD,	
  Sesterhenn	
  IA,	
  Davis	
  CJ,	
  Kiupel	
  M.	
  Classification	
  of	
  canine	
  
urinary	
  bladder	
  urothelial	
  tumours	
  based	
  on	
  the	
  World	
  Health	
  Organization/International	
  
Society	
  of	
  Urological	
  Pathology	
  consensus	
  classification.	
  Journal	
  of	
  comparative	
  pathology.	
  
2006;135:	
  190-­‐199.	
  
23.	
  Knapp	
  DW,	
  Ramos-­‐Vara	
  JA,	
  Moore	
  GE,	
  Dhawan	
  D,	
  Bonney	
  PL,	
  Young	
  KE.	
  Urinary	
  
bladder	
  cancer	
  in	
  dogs,	
  a	
  naturally	
  occurring	
  model	
  for	
  cancer	
  biology	
  and	
  drug	
  
development.	
  ILAR	
  J.	
  2014;55:	
  100-­‐118.	
  
24.	
  Ke	
  HL,	
  Tu	
  HP,	
  Lin	
  HH,	
  et	
  al.	
  Cyclooxygenase-­‐2	
  (COX-­‐2)	
  up-­‐regulation	
  is	
  a	
  prognostic	
  
marker	
  for	
  poor	
  clinical	
  outcome	
  of	
  upper	
  tract	
  urothelial	
  cancer.	
  Anticancer	
  Res.	
  2012;32:	
  
4111-­‐4116.	
  
25.	
  Tabriz	
  HM,	
  Olfati	
  G,	
  Ahmadi	
  SA,	
  Yusefnia	
  S.	
  Cyclooxygenase-­‐2	
  expression	
  in	
  urinary	
  
bladder	
  transitional	
  cell	
  carcinoma	
  and	
  its	
  association	
  with	
  clinicopathological	
  
characteristics.	
  Asian	
  Pac	
  J	
  Cancer	
  Prev.	
  2013;14:	
  4539-­‐4543.	
  

	
  

101	
  

	
  
26.	
  Yun	
  SJ,	
  Kim	
  WJ.	
  Role	
  of	
  the	
  epithelial-­‐mesenchymal	
  transition	
  in	
  bladder	
  cancer:	
  from	
  
prognosis	
  to	
  therapeutic	
  target.	
  Korean	
  J	
  Urol.	
  2013;54:	
  645-­‐650.	
  
27.	
  Paliwal	
  P,	
  Arora	
  D,	
  Mishra	
  AK.	
  Epithelial	
  mesenchymal	
  transition	
  in	
  urothelial	
  
carcinoma:	
  twist	
  in	
  the	
  tale.	
  Indian	
  J	
  Pathol	
  Microbiol.	
  2012;55:	
  443-­‐449.	
  
28.	
  Omran	
  OM.	
  CD10	
  and	
  E-­‐cad	
  expression	
  in	
  urinary	
  bladder	
  urothelial	
  and	
  squamous	
  cell	
  
carcinoma.	
  J	
  Environ	
  Pathol	
  Toxicol	
  Oncol.	
  2012;31:	
  203-­‐212.	
  
29.	
  Bryan	
  RT,	
  Atherfold	
  PA,	
  Yeo	
  Y,	
  et	
  al.	
  Cadherin	
  switching	
  dictates	
  the	
  biology	
  of	
  
transitional	
  cell	
  carcinoma	
  of	
  the	
  bladder:	
  ex	
  vivo	
  and	
  in	
  vitro	
  studies.	
  J	
  Pathol.	
  2008;215:	
  
184-­‐194.	
  
30.	
  Bryan	
  RT,	
  Tselepis	
  C.	
  Cadherin	
  switching	
  and	
  bladder	
  cancer.	
  J	
  Urol.	
  2010;184:	
  423-­‐
431.	
  
31.	
  Khorrami	
  MH,	
  Hadi	
  M,	
  Gharaati	
  MR,	
  Izadpanahi	
  MH,	
  Javid	
  A,	
  Zargham	
  M.	
  E-­‐cadherin	
  
expression	
  as	
  a	
  prognostic	
  factor	
  in	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  bladder	
  after	
  
transurethral	
  resection.	
  Urol	
  J.	
  2012;9:	
  581-­‐585.	
  
32.	
  Jager	
  T,	
  Becker	
  M,	
  Eisenhardt	
  A,	
  et	
  al.	
  The	
  prognostic	
  value	
  of	
  cadherin	
  switch	
  in	
  
bladder	
  cancer.	
  Oncol	
  Rep.	
  2010;23:	
  1125-­‐1132.	
  
33.	
  Wang	
  P,	
  Lin	
  SL,	
  Zhang	
  LH,	
  et	
  al.	
  The	
  prognostic	
  value	
  of	
  P-­‐cadherin	
  in	
  non-­‐muscle-­‐
invasive	
  bladder	
  cancer.	
  Eur	
  J	
  Surg	
  Oncol.	
  2014;40:	
  255-­‐259.	
  
34.	
  Baumgart	
  E,	
  Cohen	
  MS,	
  Silva	
  Neto	
  B,	
  et	
  al.	
  Identification	
  and	
  prognostic	
  significance	
  of	
  
an	
  epithelial-­‐mesenchymal	
  transition	
  expression	
  profile	
  in	
  human	
  bladder	
  tumors.	
  Clin	
  
Cancer	
  Res.	
  2007;13:	
  1685-­‐1694.	
  
35.	
  Clairotte	
  A,	
  Lascombe	
  I,	
  Fauconnet	
  S,	
  et	
  al.	
  Expression	
  of	
  E-­‐cadherin	
  and	
  alpha-­‐,	
  beta-­‐,	
  
gamma-­‐catenins	
  in	
  patients	
  with	
  bladder	
  cancer:	
  identification	
  of	
  gamma-­‐catenin	
  as	
  a	
  new	
  
prognostic	
  marker	
  of	
  neoplastic	
  progression	
  in	
  T1	
  superficial	
  urothelial	
  tumors.	
  Am	
  J	
  Clin	
  
Pathol.	
  2006;125:	
  119-­‐126.	
  
36.	
  Hu	
  X,	
  Ruan	
  Y,	
  Cheng	
  F,	
  Yu	
  W,	
  Zhang	
  X,	
  Larre	
  S.	
  p130Cas,	
  E-­‐cadherin	
  and	
  beta-­‐catenin	
  in	
  
human	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  bladder:	
  expression	
  and	
  clinicopathological	
  
significance.	
  Int	
  J	
  Urol.	
  2011;18:	
  630-­‐637.	
  
37.	
  Kashibuchi	
  K,	
  Tomita	
  K,	
  Schalken	
  JA,	
  Kume	
  H,	
  Takeuchi	
  T,	
  Kitamura	
  T.	
  The	
  prognostic	
  
value	
  of	
  E-­‐cadherin,	
  alpha-­‐,	
  beta-­‐	
  and	
  gamma-­‐catenin	
  in	
  bladder	
  cancer	
  patients	
  who	
  
underwent	
  radical	
  cystectomy.	
  Int	
  J	
  Urol.	
  2007;14:	
  789-­‐794.	
  
38.	
  Kashibuchi	
  K,	
  Tomita	
  K,	
  Schalken	
  JA,	
  et	
  al.	
  The	
  prognostic	
  value	
  of	
  E-­‐cadherin,	
  alpha-­‐,	
  
beta-­‐,	
  and	
  gamma-­‐catenin	
  in	
  urothelial	
  cancer	
  of	
  the	
  upper	
  urinary	
  tract.	
  Eur	
  Urol.	
  2006;49:	
  
839-­‐845;	
  discussion	
  845.	
  

	
  

102	
  

	
  
39.	
  Moyano	
  Calvo	
  JL,	
  Blanco	
  Palenciano	
  E,	
  Beato	
  Moreno	
  A,	
  et	
  al.	
  [Prognostic	
  value	
  of	
  E-­‐
cadherina,	
  beta	
  catenin,	
  Ki-­‐67	
  antigen	
  and	
  p53	
  protein	
  in	
  the	
  superficial	
  bladder	
  tumors].	
  
Actas	
  Urol	
  Esp.	
  2006;30:	
  871-­‐878.	
  
40.	
  Urakami	
  S,	
  Shiina	
  H,	
  Enokida	
  H,	
  et	
  al.	
  Epigenetic	
  inactivation	
  of	
  Wnt	
  inhibitory	
  factor-­‐1	
  
plays	
  an	
  important	
  role	
  in	
  bladder	
  cancer	
  through	
  aberrant	
  canonical	
  Wnt/beta-­‐catenin	
  
signaling	
  pathway.	
  Clin	
  Cancer	
  Res.	
  2006;12:	
  383-­‐391.	
  
41.	
  Erdogru	
  T,	
  Celik-­‐Ozenci	
  C,	
  Seval	
  Y,	
  et	
  al.	
  The	
  restorative	
  effect	
  of	
  a	
  selective	
  
cyclooxygenase-­‐2	
  inhibitor	
  on	
  urothelial	
  cell-­‐cell	
  interactions	
  after	
  partial	
  bladder	
  outlet	
  
obstruction	
  in	
  rats.	
  BJU	
  Int.	
  2005;95:	
  664-­‐669.	
  
42.	
  Howard	
  EW,	
  Camm	
  KD,	
  Wong	
  YC,	
  Wang	
  XH.	
  E-­‐cadherin	
  upregulation	
  as	
  a	
  therapeutic	
  
goal	
  in	
  cancer	
  treatment.	
  Mini	
  Rev	
  Med	
  Chem.	
  2008;8:	
  496-­‐518.	
  
43.	
  Nowak	
  M,	
  Madej	
  JA,	
  Dziegiel	
  P.	
  Expression	
  of	
  E-­‐cadherin,	
  beta-­‐catenin	
  and	
  Ki-­‐67	
  
antigen	
  and	
  their	
  reciprocal	
  relationships	
  in	
  mammary	
  adenocarcinomas	
  in	
  bitches.	
  Folia	
  
Histochem	
  Cytobiol.	
  2007;45:	
  233-­‐238.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

103	
  

	
  
CHAPTER	
  4	
  
	
  
	
  
Evaluation	
  of	
  microsatellite	
  instability	
  and	
  DNA	
  mismatch	
  repair	
  protein	
  expression	
  
in	
  canine	
  urothelial	
  carcinomas	
  
	
  
Dodd	
  Sledge1,	
  Matti	
  Kiupel1,	
  James	
  H.	
  Resau2,	
  Kristin	
  Feenstra2,	
  Alecia	
  E.	
  Agner3,	
  Baolin	
  
Wu4,	
  Nicole	
  J.	
  Madrill3,	
  Heidi	
  Parker5,	
  Deborah	
  W.	
  Knapp6,	
  James	
  R.	
  Mickelson7,	
  Elaine	
  A.	
  	
  
Ostrander5,	
  Elizabeth	
  A.	
  McNiel8	
  
	
  
1Diagnostic	
  Center	
  for	
  Population	
  and	
  Animal	
  Health,	
  Michigan	
  State	
  University,	
  Lansing,	
  
MI	
  
2	
  Van	
  Andel	
  Institute,	
  Grand	
  Rapids,	
  Lansing,	
  MI	
  
3	
  College	
  of	
  Veterinary	
  Medicine,	
  Michigan	
  State	
  University,	
  East	
  Lansing,	
  MI	
  
4	
   Division	
   of	
   Biostatistics	
   and	
   5	
   Department	
   of	
   Veterinary	
   and	
   Biomedical	
   Sciences,	
  
University	
  of	
  Minnesota,	
  Twin	
  Cities,	
  MN;	
  	
  
6	
  Department	
  of	
  Veterinary	
  Clinical	
  Sciences,	
  Purdue	
  University,	
  West	
  Lafayette,	
  IN;	
  	
  
7	
  National	
  Human	
  Genome	
  Research	
  Institute,	
  National	
  Institutes	
  of	
  Health,	
  Bethesda,	
  MD.	
  
8	
  Tufts	
  Cummings	
  School	
  of	
  Veterinary	
  Medicine	
  and	
  Molecular	
  Oncology	
  Research	
  
Institute,	
  Boston,	
  MA,	
  
	
  

	
  

	
  

104	
  

	
  
Abstract	
  
	
  

Microsatellite	
   Instability	
   (MSI),	
   defined	
   as	
   the	
   accumulation	
   of	
   frameshift	
   mutations	
  

in	
   nucleotide	
   repeat	
   sequences	
   of	
   DNA	
   microsatellites,	
   is	
   considered	
   a	
   signature	
   for	
  
dysfunction	
   in	
   the	
   DNA	
   mismatch	
   repair	
   (MMR)	
   system.	
   We	
   investigated	
   MSI	
   in	
   canine	
  
tumors	
   with	
   breed	
   predispositions,	
   using	
   a	
   panel	
   of	
   22	
   microsatellites	
   from	
   15	
   different	
  
canine	
   chromosomes.	
   Genotypes	
   derived	
   from	
   tumor	
   DNA	
   and	
   normal	
   DNA	
   from	
   blood	
  
were	
  compared	
  and	
  the	
  percentage	
  of	
  unstable	
  microsatellites	
  were	
  calculated.	
  Our	
  results	
  
demonstrated	
  that	
  MSI	
  is	
  more	
  frequent	
  among	
  canine	
  urothelial	
  carcinomas	
  of	
  the	
  urinary	
  
bladder	
  compared	
  to	
  other	
  canine	
  malignancies	
  including	
  gastric	
  carcinoma	
  and	
  mammary	
  
tumors,	
   and	
   is	
   common	
   within	
   specific	
   breeds	
   and	
   phylogenic	
   clades	
   of	
   dog	
   suggesting	
  
likely	
  genetic	
  predispositions.	
  Further,	
  we	
  investigated	
  the	
  expression	
  of	
  the	
  MMR	
  proteins,	
  
MLH1,	
   MSH2,	
   MSH3,	
   and	
   MSH6,	
   in	
   a	
   set	
   of	
   36	
   canine	
   urothelial	
   carcinomas	
   of	
   the	
   lower	
  
urinary	
  tract	
  and	
  10	
  urinary	
  bladders	
  from	
  unaffected	
  dogs	
  using	
  morphometric	
  analysis	
  of	
  
immunoreactivity	
   as	
   determined	
   by	
   multispectral	
   imaging.	
   These	
   studies	
   confirmed	
  
consistent	
   expression	
   of	
   MMR	
   proteins	
   in	
   normal	
   canine	
   urothelium.	
   While	
   there	
   was	
   no	
  
significant	
   decrease	
   in	
   expression	
   of	
   other	
   MMR	
   proteins,	
   expression	
   of	
   MSH2	
   was	
  
significantly	
   decreased	
   in	
   urothelial	
   carcinomas	
   in	
   comparison	
   to	
   that	
   in	
   urothelium	
   of	
  
unaffected	
  urinary	
  bladders.	
  These	
  investigations	
  demonstrate	
  likely	
  MMR	
  dysfunction	
  in	
  
subsets	
  of	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas.	
  

	
  

105	
  

	
  
Introduction	
  
The	
   DNA	
   mismatch	
   repair	
   (MMR)	
   system	
   participates	
   in	
   a	
   variety	
   of	
   cellular	
  
processes.	
   Most	
   notably,	
   this	
   system	
   is	
   responsible	
   for	
   post	
   replication	
   recognition	
   and	
  
repair	
  of	
  base-­‐base	
  mismatches	
  and	
  the	
  resolution	
  of	
  insertion	
  and	
  deletion	
  loops	
  that	
  can	
  
occur	
   in	
   repetitive	
   regions	
   of	
   DNA.1-­‐5	
   It	
   has	
   been	
   estimated	
   that	
   the	
   fidelity	
   of	
   DNA	
  
replication	
   is	
   improved	
   by	
   the	
   MMR	
   system	
   50-­‐1000	
   fold	
   over	
   that	
   seen	
   with	
   the	
  
replication	
   machinery	
   alone.2,	
   3	
   Accordingly,	
   deficiencies	
   in	
   this	
   system	
   can	
   lead	
   to	
  
increased	
   rates	
   of	
   point	
   and	
   frame	
   shift	
   mutations	
   throughout	
   the	
   genome.	
   Such	
   mutations	
  
can	
   lead	
   to	
   loss	
   of	
   function	
   changes	
   in	
   tumor	
   suppressors	
   or	
   gain	
   of	
   function	
   changes	
   in	
  
tumor	
   oncogenes.6-­‐10	
   In	
   addition,	
   the	
   MMR	
   system	
   also	
   participates	
   in	
   a	
   variety	
   of	
   other	
  
processes	
   including	
   DNA	
   damage	
   recognition	
   signaling,	
   promoting	
   cell	
   cycle	
   arrest	
   and	
  
apoptosis,	
   homologous	
   recombination,	
   meiotic	
   recombination,	
   and	
   other	
   DNA	
   repair	
  
pathways.1,	
  2,	
  4,	
  11	
  Given	
  these	
  varied	
  roles	
  played	
  by	
  MMR,	
  it	
  is	
  not	
  surprising	
  that	
  defects	
  in	
  
MMR	
  facilitate	
  carcinogenesis.	
  It	
  has	
  long	
  been	
  known	
  in	
  humans	
  that	
  hereditary	
  defects	
  in	
  
MMR	
   promote	
   carcinogenesis	
   in	
   a	
   variety	
   of	
   tissues	
   including	
   the	
   colon,	
   endometrium,	
  
stomach,	
   skin,	
   prostate,	
   and	
   urinary	
   tract.11-­‐13	
   However,	
   defects	
   in	
   MMR	
   have	
   also	
   been	
  
reported	
   in	
   a	
   wide	
   number	
   and	
   variety	
   of	
   spontaneously	
   occurring	
   cancers	
   suggesting	
   a	
  
wide	
  role	
  in	
  cancer	
  development.1,	
  14-­‐16	
  
The	
   MMR	
   system	
   is	
   composed	
   of	
   a	
   set	
   of	
   highly	
   conserved	
   proteins.4,	
  

11	
  

Five	
  

proteins,	
   including	
   MSH2,	
   MSH3,	
   MSH6,	
   MLH1,	
   and	
   PMS2,	
   provide	
   the	
   backbone	
   of	
   the	
  
MMR	
   machinery.	
   Initial	
   DNA	
   damage	
   recognition	
   may	
   be	
   accomplished	
   by	
   either	
   MSH2-­‐
MSH6	
   heterodimers	
   (MutSα)	
   or	
   MSH2-­‐MSH3	
   heterodimers	
   (MutSβ).11	
   The	
   MutS	
  
heterodimers	
   are	
   then	
   capable	
   of	
   recruiting	
   the	
   MLH1-­‐PMS2	
   heterodimer,	
   MutL,	
   which	
  

	
  

106	
  

	
  
initiates	
  the	
  assembly	
  of	
  other	
  proteins	
  involved	
  in	
  excising	
  the	
  lesion,	
  re-­‐synthesizing	
  DNA	
  
and	
   ligating	
   the	
   gaps.	
   Failure	
   to	
   repair	
   DNA	
   errors	
   may	
   result	
   in	
   signaling	
   for	
   cell	
   cycle	
  
arrest	
   and/or	
   apoptosis.	
   Deficient	
   expression	
   or	
   abnormalities	
   in	
   any	
   of	
   these	
   proteins	
  
result	
   in	
   a	
   mutator	
   phenotype	
   in	
   which	
   DNA	
   mutations	
   can	
   accumulate	
   in	
   a	
   variety	
   of	
  
genes.6-­‐10	
  	
  
MMR	
   function	
   has	
   most	
   often	
   been	
   evaluated	
   through	
   analysis	
   of	
   microsatellites.	
  
Microsatellites	
   are	
   regions	
   of	
   nucleotide	
   repeats	
   located	
   throughout	
   the	
   genome.	
   These	
  
areas	
   are	
   prone	
   to	
   polymerase	
   slippage	
   during	
   replication	
   leading	
   to	
   formation	
   of	
   small	
  
insertion	
   and	
   deletion	
   loops.17	
   If	
   not	
   recognized	
   and	
   repaired	
   by	
   MMR,	
   buildup	
   of	
   frame	
  
shift	
   mutations	
   occurs	
   within	
   microsatellites.5,	
   11	
   In	
   cancers	
   with	
   defects	
   in	
   the	
   MMR	
  
system,	
  a	
  high	
  percentage	
  of	
  microsatellites	
  have	
  recognizable	
  mutations.	
  Such	
  buildup	
  of	
  
mutations	
   within	
   microsatellites	
   is	
   termed	
   microsatellite	
   instability	
   (MSI)	
   and	
   is	
  
considered	
  a	
  “signature”	
  for	
  MMR	
  dysfunction.11,	
  18	
  	
  
Cancer	
   is	
   common	
   in	
   pet	
   dogs.	
   	
   Increased	
   risks	
   for	
   development	
   of	
   specific	
  
carcinomas	
   within	
   certain	
   breeds	
   suggests	
   that	
   inheritable	
   defects	
   may	
   underlie	
  
carcinogenesis	
   in	
   affected	
   breeds.19-­‐21	
   Thus,	
   breed	
   predispositions	
   for	
   specific	
   cancers	
  
provide	
   an	
   opportunity	
   to	
   investigate	
   potential	
   differential	
   genetic	
   mechanisms	
   of	
  
carcinogenesis,	
   including	
   those	
   associated	
   with	
   MMR.	
   Given	
   the	
   diversity	
   of	
   cancer	
  
phenotypes	
   associated	
   with	
   MMR	
   dysfunction,	
   we	
   hypothesized	
   that	
   mutation	
   in	
   MMR	
  
could	
   underlie	
   some	
   of	
   the	
   cancer	
   susceptibility	
   in	
   dogs.	
   To	
   investigate	
   this	
   hypothesis,	
   we	
  
evaluated	
   for	
   MSI	
   in	
   canine	
   epithelial	
   tumors	
   that	
   may	
   have	
   a	
   genetic	
   basis	
   including	
  
gastric,	
   mammary,	
   and	
   urothelial	
   carcinomas.	
   Following	
   identification	
   of	
   a	
   high	
   rate	
   of	
   MSI	
  
in	
   canine	
   urothelial	
   carcinomas,	
   the	
   expression	
   of	
   MLH1,	
   MSH2,	
   MSH3,	
   and	
   MSH6	
   was	
  

	
  

107	
  

	
  
evaluated	
   in	
   such	
   tumors	
   using	
   immunohistochemistry	
   and	
   compared	
   to	
   that	
   of	
   normal	
  
urothelium,	
  rates	
  of	
  MSI,	
  histomorphologic	
  features	
  of	
  the	
  tumors,	
  and	
  survival	
  time.	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

108	
  

	
  
Materials	
  and	
  methods	
  
Evaluation	
  of	
  Microsatellite	
  Instability	
  in	
  Canine	
  Tumors	
  
This	
  study	
  was	
  approved	
  by	
  the	
  University	
  of	
  Minnesota	
  Institutional	
  Animal	
  Care	
  
and	
  Use	
  Committee.	
  Pet	
  dogs	
  with	
  diagnoses	
  of	
  gastric	
  carcinoma,	
  mammary	
  tumor,	
  or	
  
urothelial	
  (transitional	
  cell)	
  carcinoma	
  of	
  the	
  urinary	
  bladder	
  were	
  identified	
  through	
  the	
  
University	
  of	
  Minnesota	
  Veterinary	
  Medical	
  Center	
  and	
  Diagnostic	
  Laboratory,	
  and	
  through	
  
referring	
  veterinarians.	
  	
  
	
  

Upon	
   owner	
   consent,	
   blood	
   samples	
   from	
   each	
   dog	
   were	
   collected	
   in	
   7.5%	
   EDTA	
  

containing	
   tube	
   (Monoject®,	
   Tyco	
   Healthcare,	
   Mansfield,	
   MA)	
   and	
   refrigerated	
   at	
   4°C.	
  
Tumor	
  samples	
  were	
  collected	
  either	
  at	
  the	
  time	
  of	
  diagnostic	
  biopsy,	
  therapeutic	
  surgical	
  
excision,	
   or	
   necropsy.	
   DNA	
   was	
   isolated	
   from	
   blood	
   samples	
   and	
   tumor	
   tissue	
   using	
  
commercially	
   available	
   DNA	
   extraction	
   kits	
   (Purgene,	
   Gentra	
   Systems,	
   Inc,	
   Minneapolis,	
  
MN)	
   using	
   manufacturer	
   recommendations.	
   DNA	
   was	
   stored	
   at	
   -­‐80°C	
   pending	
  
microsatellite	
  PCR.	
  	
  
	
  

A	
   panel	
   of	
   22	
   MS	
   repeats	
   including	
   10	
   tetra-­‐,	
   11	
   di-­‐,	
   and	
   one	
   mononucleotide	
  

repeats,	
   which	
   were	
   located	
   on	
   15	
   different	
   canine	
   chromosomes,	
   was	
   used	
   to	
   evaluate	
  
MSI.	
   With	
   the	
   exception	
   of	
   the	
   mononucleotide	
   repeat,	
   which	
   we	
   identified,	
   the	
  
microsatellite	
  sequences	
  and	
  corresponding	
  PCR	
  primers	
  for	
  their	
  amplification	
  have	
  been	
  
reported	
   previously.22	
   An	
   18	
   nucleotide	
   tail	
   sequence	
   was	
   added	
   to	
   the	
   5’	
   end	
   of	
   each	
  
reverse	
   primer	
   using	
   previously	
   described	
   methods.23	
   The	
   microsatellite	
   sequences	
   were	
  
PCR	
  amplified	
  from	
  tumor	
  and	
  normal	
  (from	
  blood)	
  DNA	
  in	
  15	
  μl	
  reactions	
  including	
  12.5	
  
ng	
   of	
   DNA,	
   1.5	
   mM	
   MgCl2,	
   1	
   µM	
   of	
   the	
   forward	
   primer	
   and	
   0.3	
   µM	
   of	
   the	
   reverse	
   primer	
  
with	
  tail,	
  156	
  nM	
  of	
  a	
  fluorescent	
  dye-­‐labeled	
  primer	
  complimentary	
  to	
  the	
  tail	
  sequence,	
  

	
  

109	
  

	
  
300	
   µM	
   of	
   each	
   dNTP	
   (Fisher	
   Scientific,	
   Pittsburg,	
   PA)	
   and	
   1.5	
   U	
   Taq	
   Polymerase	
  
(HotStarTaq®,	
   Qiagen,	
   Valencia,	
   CA).	
   	
  	
   The	
   PCR	
   program	
   consisted	
   of	
   20	
   min	
   at	
   95°C,	
   40	
  
cycles	
  of	
  1	
  min	
  at	
  94°C,	
  1	
  min	
  at	
  56°C,	
  2	
  min	
  at	
  72°C,	
  and	
  a	
  final	
  extension	
  step	
  of	
  10	
  min	
  at	
  
72°C.	
   	
  PCR	
   products,	
   with	
   DNA	
   size	
   standards,	
   were	
   evaluated	
   using	
   capillary	
  
electrophoresis	
  on	
  the	
  CEQ	
  8000	
  Genetic	
  Analysis	
  System	
  (Beckman	
  Coulter,	
  Inc,	
  Fullerton,	
  
CA)	
  and	
  chromatograms	
  subsequently	
  were	
  analyzed	
  using	
  the	
  system	
  software.	
  Genotypes	
  
were	
  manually	
  verified.	
  Any	
  ambiguous	
  results	
  in	
  which	
  PCR	
  products	
  were	
  of	
  poor	
  quality	
  
or	
  when	
  genotypes	
  were	
  inconsistent	
  were	
  repeated	
  to	
  confirm	
  instability.	
  
	
  

For	
   each	
   tumor	
   MSI	
   was	
   evaluated	
   as	
   %MSI,	
   which	
   was	
   defined	
   for	
   each	
   case	
   as	
   the	
  

%	
   of	
   evaluated	
   MS	
   loci	
   in	
   which	
   there	
   was	
   detectable	
   variation	
   in	
   MS	
   length	
   in	
   at	
   least	
   one	
  
allele	
   between	
   DNA	
   isolated	
   from	
   tumors	
   and	
   DNA	
   isolated	
   from	
   blood.	
   Tumors	
   with	
   >	
  
25%	
   MSI	
   were	
   classified	
   as	
   high	
   level	
   MSI	
   (MSI-­‐H),	
   tumors	
   with	
   10%	
   >	
   MSI	
   <	
   25%	
   were	
  
classified	
   as	
   exhibiting	
   low	
   level	
   MSI.	
   Tumors	
   with	
   <10%	
   MSI	
   were	
   considered	
   stable	
  
(MSS).	
   	
   For	
   urothelial	
   carcinomas,	
   comparisons	
   of	
   %MSI	
   were	
   made	
   between	
   breeds	
   and	
  
phylogenic	
  clade.	
  	
  For	
  statistical	
  evaluation	
  with	
  respect	
  to	
  breed,	
  4	
  breeds	
  with	
  a	
  reported	
  
high	
   relative	
   risk	
   for	
   development	
   of	
   bladder	
   cancer	
   (Scottish	
   terrier,	
   West	
   Highland	
   white	
  
terrier,	
  Shetland	
  sheepdog,	
  and	
  beagle)	
  were	
  evaluated	
  as	
  independent	
  groups	
  while	
  dogs	
  
of	
  other	
  breed,	
  which	
  included	
  17	
  unique	
  breeds	
  and	
  mixed	
  breed	
  dogs,	
  were	
  categorized	
  
into	
  a	
  single	
  group.20,	
  24,	
  25	
  For	
  statistical	
  evaluation	
  with	
  respect	
  to	
  phylogenetic	
  clade,	
  dogs	
  
were	
  of	
  specific	
  breed	
  grouped	
  into	
  phylogenetic	
  clade	
  according	
  to	
  previous	
  description	
  of	
  
genetic	
   diversity	
   amongst	
   purebred	
   dogs	
   or	
   evaluated	
   as	
   a	
   separate	
   category	
   if	
   of	
   mixed	
  
breed.26	
  	
  
Evaluation	
  of	
  MMR	
  protein	
  expression	
  in	
  Canine	
  Urothelial	
  Carcinomas	
  

	
  

110	
  

	
  
A	
  series	
  of	
  36	
  dogs	
  that	
  were	
  diagnosed	
  with	
  urothelial	
  carcinomas	
  at	
  either	
  the	
  
Michigan	
  State	
  University	
  Veterinary	
  Teaching	
  Hospital	
  or	
  the	
  University	
  of	
  Minnesota	
  
Veterinary	
  Teaching	
  Hospital	
  through	
  biopsy	
  were	
  selected	
  for	
  inclusion	
  in	
  the	
  study	
  based	
  
on	
  owner	
  consent,	
  availability	
  of	
  clinical	
  history,	
  and	
  availability	
  of	
  paraffin-­‐embedded,	
  
formalin-­‐fixed	
  diagnostic	
  samples.	
  	
  For	
  each	
  urothelial	
  carcinoma	
  case,	
  descriptive	
  
information	
  was	
  obtained	
  from	
  medical	
  records	
  and	
  periodic	
  follow-­‐up	
  questionnaires	
  
including	
  age	
  at	
  diagnosis,	
  breed,	
  sex,	
  survival	
  time	
  from	
  the	
  date	
  of	
  diagnosis,	
  the	
  reported	
  
cause	
  of	
  death,	
  and	
  any	
  treatment	
  employed.	
  In	
  addition,	
  bladder	
  samples	
  from	
  10	
  healthy	
  
dogs	
  that	
  were	
  used	
  for	
  veterinary	
  student	
  teaching	
  purposes	
  were	
  harvested	
  in	
  10%	
  
buffered	
  formalin	
  and	
  routinely	
  processed	
  for	
  histologic	
  examination.	
  Five	
  µm	
  sections	
  of	
  
all	
  samples	
  were	
  routinely	
  processed	
  and	
  stained	
  with	
  hematoxylin	
  and	
  eosin.	
  Urothelial	
  
proliferative	
  lesions	
  were	
  classified,	
  and	
  papillary	
  carcinomas	
  were	
  graded	
  according	
  to	
  
the	
  WHO/ISUP	
  Consensus	
  Classification	
  System	
  as	
  previously	
  described	
  in	
  canine	
  
urothelial	
  tumors.27	
  In	
  addition,	
  the	
  degree	
  of	
  invasion	
  of	
  all	
  carcinomas	
  was	
  scored	
  as	
  no	
  
invasion,	
  invasion	
  into	
  the	
  substantia	
  propria,	
  or	
  invasion	
  into	
  the	
  muscularis.	
  	
  
Immunohistochemistry	
  (IHC)	
  was	
  preformed	
  to	
  evaluate	
  MSH2,	
  MSH3,	
  and	
  MSH6,	
  
and	
  MLH1	
  expression	
  in	
  normal	
  canine	
  bladder	
  and	
  canine	
  urothelial	
  carcinomas	
  using	
  
commercially	
  available	
  antibodies	
  which	
  were	
  shown	
  be	
  efficacious	
  for	
  use	
  in	
  dog	
  tissues	
  
by	
  Western	
  blot	
  (see	
  chapter	
  4).	
  Five	
  µm	
  sections	
  of	
  all	
  formalin-­‐fixed,	
  paraffin-­‐embedded	
  
tissues	
  were	
  processed	
  for	
  immunohistochemistry	
  and	
  labeled	
  with	
  a	
  mouse	
  monoclonal	
  
anti-­‐MLH1	
  antibody	
  (1:200,	
  BD	
  Biosciences,	
  San	
  Jose,	
  CA,	
  USA),	
  a	
  rabbit	
  polyclonal	
  anti-­‐
MSH2	
  antibody	
  (1:100.	
  Santa	
  Cruz	
  Biotechnology,	
  Dallas,	
  TX,	
  USA),	
  a	
  rabbit	
  polyclonal	
  anti-­‐
MSH3	
  antibody	
  (1:100,	
  Abcam,	
  Cambridge,	
  MA,	
  USA),	
  a	
  mouse	
  monoclonal	
  anti-­‐MSH6	
  

	
  

111	
  

	
  
antibody	
  (1:500,	
  BD	
  Biosciences,	
  San	
  Jose,	
  CA,	
  USA).	
  For	
  MLH1	
  and	
  MSH6,	
  
deparaffinization,	
  antigen	
  retrieval,	
  immunohistochemical	
  staining	
  and	
  counterstaining	
  
was	
  preformed	
  on	
  a	
  Benchmark	
  XT™	
  autostainer	
  (Ventana,	
  Tucson,	
  AZ,	
  USA)	
  using	
  an	
  
Enhanced	
  Alkaline	
  Phosphatase	
  Red	
  Detection	
  Kit	
  (Ventana)	
  that	
  uses	
  an	
  indirect	
  biotin	
  
streptavidin	
  and	
  Fast	
  Red	
  chromogen	
  detection	
  system.	
  	
  Retrieval	
  for	
  MLH1	
  and	
  MSH6	
  was	
  
accomplished	
  with	
  20	
  minutes	
  of	
  incubation	
  with	
  Cell	
  Conditioning	
  1	
  solution	
  (Ventana).	
  
For	
  MSH2,	
  deparaffinization,	
  antigen	
  retrieval,	
  immunohistochemical	
  staining	
  and	
  
counterstaining	
  was	
  preformed	
  on	
  a	
  Dako	
  Link	
  48	
  Autostainer	
  (Carpinteria,	
  CA,	
  USA)	
  using	
  
a	
  LSAB2	
  kit	
  (Dako),	
  which	
  employs	
  a	
  3,3’	
  diaminobenzidine	
  tetrahydrochloride	
  (DAB)	
  
chromogen	
  detection	
  system.	
  	
  For	
  MSH3,	
  deparaffinization,	
  antigen	
  retrieval,	
  
immunohistochemical	
  staining	
  and	
  counterstaining	
  was	
  performed	
  on	
  a	
  Bond	
  maX™	
  
Automated	
  Staining	
  System	
  (Leica	
  Biosystems,	
  Buffalo	
  Grove,	
  IL,	
  USA)	
  using	
  the	
  Bond™	
  
Polymer	
  Refine	
  Detection	
  System	
  (Leica	
  Biosystems),	
  which	
  employs	
  a	
  3,3’	
  
diaminobenzidine	
  tetrahydrochloride	
  (DAB)	
  chromogen	
  detection	
  system.	
  Retrieval	
  for	
  
MSH2	
  and	
  MSH3	
  was	
  accomplished	
  using	
  heat	
  induced	
  epitope	
  retrieval	
  and	
  incubation	
  
with	
  Bond™	
  Epitope	
  Retrieval	
  Solution	
  1	
  (Leica	
  Biosystems)	
  for	
  20	
  minutes.	
  Sections	
  of	
  
normal	
  urinary	
  bladder	
  used	
  as	
  positive	
  controls	
  were	
  run	
  in	
  parallel	
  to	
  cases	
  for	
  each	
  of	
  
the	
  IHC.	
  	
  For	
  negative	
  controls,	
  homologous	
  non-­‐immune	
  sera	
  or	
  buffer	
  replaced	
  primary	
  
antibodies.	
  
	
  

For	
  all	
  cases,	
  seven	
  randomly	
  picked	
  areas	
  from	
  each	
  slide	
  immunohistochemically	
  

labeled	
  for	
  the	
  respective	
  MMR	
  proteins	
  listed	
  above	
  were	
  imaged	
  using	
  a	
  CRi	
  Nuance	
  
multispectral	
  imaging	
  system	
  (Woburn,	
  MA,	
  USA).	
  The	
  resulting	
  image	
  cubes	
  were	
  
converted	
  to	
  optical	
  density	
  units,	
  then	
  mathematically	
  unmixed	
  to	
  separate	
  chromagen	
  

	
  

112	
  

	
  
labeling	
  from	
  counterstain	
  using	
  spectral	
  libraries	
  generated	
  from	
  imaging	
  of	
  only	
  
hematoxylin	
  stained	
  or	
  immunohistochemically	
  labeled	
  and	
  hematoxylin	
  counterstained	
  
control	
  specimens.	
  Colocalized	
  immunoreactivity	
  was	
  assessed	
  as	
  positive	
  pixels	
  per	
  image	
  
using	
  constant	
  thresholds	
  for	
  detection	
  using	
  the	
  multispectral	
  imaging	
  system	
  software.	
  
The	
  component	
  images	
  of	
  the	
  image	
  cubes	
  were	
  then	
  pseudocolored,	
  converted	
  to	
  
pseudofluorescent	
  format,	
  and	
  unmixed	
  for	
  counting	
  of	
  nuclei,	
  which	
  was	
  done	
  using	
  
Imagine	
  morphometric	
  analysis	
  software	
  (Van	
  Andel	
  Research	
  Institute	
  (Grand	
  Rapids,	
  MI,	
  
USA)	
  (Figure	
  16).	
  For	
  statistical	
  analysis	
  results	
  were	
  evaluated	
  for	
  each	
  case	
  as	
  mean	
  
number	
  of	
  positive	
  pixels	
  per	
  nucleus.	
  
Statistical	
  analysis	
  
All	
  statistical	
  analysis	
  was	
  preformed	
  using	
  SPSS	
  statistical	
  software	
  (Somers,	
  NY).	
  
Permutation	
  F	
  tests	
  were	
  used	
  to	
  evaluate	
  MSI	
  data	
  as	
  a	
  continuous	
  variable.	
  Fisher’s	
  exact	
  
tests	
  were	
  used	
  to	
  evaluate	
  categorical	
  treatment	
  of	
  MSI	
  with	
  tumors	
  defined	
  as	
  MSI-­‐H,	
  
MSI-­‐L	
  or	
  MSS	
  as	
  described	
  above.	
  Mann-­‐Whitney	
  U	
  tests	
  and	
  Kurskal-­‐Wallis	
  tests	
  were	
  
used	
  to	
  evaluate	
  correlations	
  between	
  breed,	
  sex,	
  histologic	
  classification,	
  degree	
  of	
  
anaplasia,	
  and	
  depth	
  of	
  infiltration,	
  and	
  immunohistochemical	
  expression	
  MMR	
  proteins	
  as	
  
morphometrically	
  evaluated	
  by	
  nuance	
  spectral	
  imaging.	
  Regression	
  was	
  used	
  for	
  pairwise	
  
comparisons	
  of	
  %MSI	
  and	
  immunohistochemical	
  expression	
  of	
  MMR	
  protein.	
  	
  Correlations	
  
between	
  survival	
  time	
  of	
  canine	
  urinary	
  carcinomas	
  and	
  clinico-­‐demographic	
  data,	
  
histologic	
  classification	
  and	
  grading,	
  and	
  immunohistochemical	
  expression	
  of	
  MMR	
  
proteins	
  were	
  evaluated	
  using	
  ANOVA.	
  For	
  all	
  statistical	
  analysis,	
  significance	
  was	
  defined	
  
as	
  p<0.05.	
  
	
  

	
  

	
  
113	
  

	
  
Results	
  
Evaluation	
  of	
  MSI	
  	
  
The	
  results	
  of	
  scoring	
  MS	
  aberrations	
  in	
  canine	
  epithelial	
  tumors	
  are	
  depicted	
  in	
  
Figure	
  17	
  and	
  Table	
  10.	
  While	
  MS	
  aberrations	
  were	
  uncommon	
  in	
  gastric	
  carcinoma	
  and	
  
most	
  mammary	
  tumors,	
  both	
  the	
  percentage	
  of	
  aberrant	
  MS	
  in	
  individual	
  tumors	
  and	
  the	
  
frequency	
  of	
  tumors	
  classified	
  as	
  MSI-­‐H	
  were	
  significantly	
  increased	
  in	
  canine	
  urothelial	
  
carcinomas	
  of	
  the	
  urinary	
  bladder.	
  Despite	
  a	
  higher	
  prevalence	
  of	
  MSI	
  in	
  canine	
  TCC,	
  the	
  
distribution	
  of	
  %MSI	
  was	
  wide,	
  ranging	
  from	
  0%	
  to	
  92%	
  of	
  MS	
  loci	
  being	
  affected	
  in	
  
individual	
  tumors.	
  	
  	
  
There	
  was	
  variation	
  in	
  the	
  magnitude	
  and	
  frequency	
  of	
  %MSI	
  in	
  canine	
  urothelial	
  
within	
  all	
  evaluated	
  breed	
  groups;	
  however,	
  there	
  were	
  differences	
  in	
  overall	
  %MSI	
  
between	
  breed	
  groups.	
  Specifically,	
  MSI	
  was	
  greater	
  in	
  urothelial	
  carcinomas	
  from	
  Scottish	
  
terriers	
  and	
  West	
  Highland	
  white	
  terriers,	
  less	
  pronounced	
  in	
  Shetland	
  sheepdogs,	
  and	
  
uncommon	
  in	
  Beagles	
  (Figure	
  18	
  and	
  Table	
  11).	
  In	
  permutation	
  F	
  tests	
  comparing	
  these	
  4	
  
breeds	
  and	
  excluding	
  all	
  other	
  breeds,	
  differences	
  in	
  MSI	
  were	
  statistically	
  significant	
  
(p=0.037);	
  however,	
  statistical	
  significance	
  was	
  lost	
  when	
  MSI	
  instability	
  was	
  compared	
  
between	
  the	
  4	
  breeds	
  along	
  with	
  a	
  category	
  including	
  all	
  other	
  breeds	
  (p=0.092).	
  Using	
  
Fisher’s	
  exact	
  test	
  to	
  compare	
  the	
  categorical	
  distribution	
  of	
  tumors	
  as	
  MSS,	
  MSI-­‐L,	
  and	
  
MSI-­‐H	
  according	
  to	
  breed,	
  there	
  were	
  statistically	
  significant	
  differences	
  whether	
  
comparing	
  only	
  the	
  four	
  high	
  risk	
  breeds	
  (p=0.016),	
  or	
  all	
  breed	
  categories	
  (p=0.017).	
  
Further,	
  there	
  were	
  significant	
  differences	
  in	
  MSI	
  expression	
  between	
  dog	
  phylogenetic	
  
clades,	
  (Figure	
  19).	
  Specifically	
  dogs	
  in	
  clade	
  5	
  had	
  high	
  levels	
  of	
  MSI	
  in	
  comparison	
  to	
  dogs	
  
from	
  other	
  clades.	
  

	
  

114	
  

	
  
MSI	
  was	
  observed	
  in	
  tetranucleotide	
  repeats	
  and	
  shorter	
  repeats	
  including	
  mono-­‐	
  
and	
  dinucleotide	
  repeats	
  in	
  carcinomas	
  urothelial	
  carcinomas.	
  (Table	
  12).	
  Permutation	
  F-­‐
tests	
  showed	
  significant	
  differences	
  in	
  %MSI	
  for	
  short	
  repeats	
  whether	
  comparing	
  only	
  the	
  
four	
  high	
  risk	
  breeds	
  (p=0.016),	
  or	
  all	
  breed	
  categories	
  (p=0.042);	
  however,	
  there	
  was	
  no	
  
significant	
  difference	
  between	
  %MSI	
  for	
  tetranucleotide	
  repeats	
  whether	
  comparing	
  only	
  
the	
  four	
  high	
  risk	
  breeds	
  (p=0.116),	
  or	
  all	
  breed	
  categories	
  (p=0.158).	
  Fisher	
  exact	
  tests	
  
revealed	
  significant	
  differences	
  in	
  category	
  of	
  MSI	
  in	
  the	
  short	
  repeats	
  among	
  breed	
  
categories	
  whether	
  comparing	
  only	
  the	
  four	
  high	
  risk	
  breeds	
  (p=0.017),	
  or	
  all	
  breed	
  
categories	
  (p=0.050).	
  	
  In	
  contrast,	
  no	
  statistical	
  association	
  could	
  be	
  identified	
  between	
  MSI	
  
category	
  in	
  tetranucleotide	
  repeats	
  and	
  breed	
  (p=0.390	
  for	
  only	
  high	
  risk	
  breeds,	
  p=0.455	
  
including	
  all	
  breed	
  categories).	
  	
  
Demographic	
  information,	
  histologic	
  classification,	
  and	
  grading	
  of	
  urothelial	
  carcinomas	
  
For	
  dogs	
  diagnosed	
  with	
  urothelial	
  carcinomas,	
  the	
  mean	
  age	
  at	
  initial	
  diagnosis	
  
was	
  10.5	
  years	
  of	
  age	
  (2.0	
  SD)	
  with	
  a	
  range	
  of	
  6.7	
  to	
  17.1	
  years	
  of	
  age.	
  Affected	
  dogs	
  
included	
  4	
  intact	
  females,	
  21	
  spayed	
  females,	
  and	
  11	
  castrated	
  males.	
  Represented	
  breeds	
  
included	
  5	
  Scottich	
  terriers,	
  2	
  West	
  Highland	
  white	
  terriers,	
  7	
  Beagles,	
  7	
  Shetland	
  
Sheepdogs,	
  and	
  14	
  dogs	
  of	
  other	
  purebred	
  or	
  mixed	
  breed.	
  	
  For	
  the	
  32	
  cases	
  in	
  which	
  
information	
  regarding	
  treatment	
  was	
  available,	
  7	
  had	
  no	
  treatment,	
  2	
  had	
  only	
  local	
  
resection	
  of	
  the	
  mass,	
  17	
  were	
  treated	
  with	
  only	
  piroxicam,	
  and	
  6	
  were	
  treated	
  with	
  
piroxicam,	
  surgical	
  resection,	
  and	
  systemic	
  chemotherapy	
  which	
  variably	
  included	
  
adriamycin,	
  cyclophosphamide,	
  and/or	
  mitoxantrone.	
  	
  Two	
  animals	
  were	
  alive	
  at	
  the	
  end	
  
of	
  study,	
  1	
  animal	
  died	
  from	
  causes	
  unrelated	
  to	
  the	
  urinary	
  carcinoma,	
  30	
  reportedly	
  died	
  

	
  

115	
  

	
  
or	
  were	
  euthanized	
  as	
  a	
  result	
  of	
  progressive	
  disease	
  related	
  to	
  the	
  urinary	
  bladder	
  tumors,	
  
and	
  3	
  were	
  lost	
  to	
  follow-­‐up.	
  	
  	
  
Of	
  the	
  36	
  urothelial	
  carcinomas,	
  8	
  were	
  diagnosed	
  as	
  infiltrating	
  carcinomas,	
  which	
  
had	
  no	
  appreciable	
  exophytic	
  papillary	
  component,	
  but	
  instead	
  primarily	
  infiltrated	
  into	
  
and	
  expanded	
  the	
  bladder	
  wall.	
  	
  The	
  remaining	
  28	
  cases	
  were	
  papillary	
  carcinomas,	
  the	
  
vast	
  majority	
  of	
  which	
  (24	
  of	
  28)	
  were	
  high	
  grade	
  (grade	
  II	
  or	
  III).	
  	
  Invasion	
  of	
  high-­‐grade	
  
papillary	
  carcinomas	
  into	
  the	
  bladder	
  wall	
  was	
  common	
  with	
  9	
  of	
  24	
  having	
  invasion	
  into	
  
the	
  substantia	
  propria	
  and	
  7	
  of	
  24	
  invading	
  into	
  the	
  muscularis.	
  	
  For	
  all	
  carcinomas,	
  the	
  
mean	
  survival	
  time	
  was	
  326	
  days	
  with	
  a	
  standard	
  deviation	
  of	
  341	
  and	
  range	
  of	
  0	
  to	
  1225	
  
days.	
  	
  For	
  infiltrative	
  carcinomas,	
  the	
  mean	
  survival	
  time	
  was	
  173	
  days	
  with	
  a	
  standard	
  
deviation	
  of	
  223	
  days	
  and	
  range	
  of	
  0	
  to	
  365	
  days.	
  	
  For	
  papillary	
  carcinomas,	
  the	
  mean	
  
survival	
  time	
  was	
  370	
  days	
  with	
  a	
  standard	
  deviation	
  of	
  359	
  days	
  and	
  range	
  of	
  0	
  to	
  1225	
  
days.	
  
MLH1,	
  MSH2,	
  MSH3,	
  and	
  MSH6	
  Immunohistochemistry	
  	
  
	
  

Nuclear	
  expression	
  was	
  observed	
  in	
  urothelium	
  for	
  all	
  evaluated	
  MMR	
  protein	
  

markers.	
  For	
  all	
  cases,	
  mean	
  with	
  standard	
  error,	
  median,	
  and	
  range	
  of	
  %MSI	
  and	
  
expression	
  of	
  MMR	
  proteins	
  MLH1,	
  MSH2,	
  MSH3,	
  and	
  MSH6	
  in	
  terms	
  of	
  positive	
  pixels	
  per	
  
nucleus	
  as	
  determined	
  by	
  IHC	
  and	
  Nuance	
  spectral	
  imaging	
  are	
  reported	
  in	
  Table	
  13.	
  	
  The	
  
expression	
  of	
  MLH1,	
  MSH2,	
  MSH3,	
  and	
  MSH6	
  and	
  %MSI	
  and	
  were	
  not	
  normally	
  distributed	
  
amongst	
  cases.	
  	
  	
  
Overall,	
  while	
  there	
  was	
  variation	
  in	
  the	
  expression	
  of	
  MMR	
  proteins	
  in	
  both	
  
urothelium	
  of	
  normal	
  urinary	
  bladders	
  and	
  urothelial	
  carcinomas	
  of	
  the	
  urinary	
  bladder,	
  
the	
  mean	
  expression	
  of	
  MSH2	
  was	
  lower	
  in	
  urothelial	
  carcinomas	
  than	
  in	
  urinary	
  bladders,	
  

	
  

116	
  

	
  
and	
  the	
  mean	
  expression	
  of	
  MSH6	
  was	
  higher	
  in	
  urothelial	
  carcinomas	
  than	
  in	
  normal	
  
urinary	
  bladders	
  (Figure	
  20).	
  	
  Also,	
  the	
  mean	
  ratio	
  of	
  MSH2	
  to	
  MSH6	
  was	
  lower	
  in	
  
urothelial	
  carcinomas	
  than	
  it	
  was	
  in	
  normal	
  urinary	
  bladders	
  (Figure	
  21).	
  Using	
  Mann-­‐
Whitney	
  U	
  tests,	
  there	
  were	
  significant	
  differences	
  between	
  urothelial	
  carcinomas	
  and	
  
normal	
  urinary	
  bladder	
  samples	
  in	
  terms	
  of	
  %MSI	
  (p=0.003),	
  the	
  expression	
  of	
  MSH2	
  
(p=0.001),	
  and	
  the	
  expression	
  of	
  MSH6	
  (p=0.003).	
  However,	
  there	
  was	
  no	
  significant	
  
difference	
  between	
  urothelial	
  carcinomas	
  and	
  normal	
  urinary	
  bladder	
  samples	
  in	
  terms	
  of	
  
the	
  expression	
  of	
  MLH1	
  (p=0.886)	
  or	
  the	
  expression	
  of	
  MSH3	
  (p=0.470).	
  	
  There	
  were	
  also	
  
no	
  significant	
  differences	
  between	
  infiltrating	
  carcinomas	
  and	
  papillary	
  carcinomas	
  in	
  
terms	
  of	
  %MSI	
  (p=0.921)	
  or	
  the	
  expression	
  of	
  MLH1	
  (p=0.116),	
  MSH2	
  (p=0.562),	
  MSH3	
  
(p=0.204)	
  or	
  MSH6	
  (0.751).	
  	
  Further,	
  using	
  Kreskas-­‐Wallis	
  tests	
  and	
  examining	
  only	
  
urothelial	
  carcinoma,	
  there	
  were	
  no	
  significant	
  differences	
  (p>0.05)	
  between	
  breed,	
  sex,	
  
degree	
  of	
  cellular	
  anaplasia,	
  and	
  degree	
  of	
  invasion	
  into	
  the	
  urinary	
  bladder	
  wall,	
  and	
  
expression	
  of	
  MLH1,	
  MSH2,	
  MSH3,	
  or	
  MSH6.	
  
In	
  regression	
  analysis	
  evaluating	
  all	
  cases	
  including	
  normal	
  urinary	
  bladders	
  and	
  
urothelial	
  carcinomas,	
  there	
  were	
  no	
  significant	
  correlations	
  between	
  %MSI	
  and	
  
expression	
  of	
  MLH1	
  (p=0.408),	
  MSH2	
  (p=0.263),	
  MSH3	
  (p=0.799),	
  or	
  MSH6	
  (p=0.463).	
  	
  
Excluding	
  normal	
  urinary	
  bladders,	
  there	
  remained	
  no	
  significant	
  correlations	
  between	
  
%MSI	
  and	
  expression	
  of	
  MLH1	
  (p=0.343),	
  MSH2	
  (p=0.518),	
  MSH3	
  (p=0.629),	
  or	
  MSH6	
  
(p=0.192).	
  
Survival	
  analysis	
  
	
  

Using	
  Cox	
  regression,	
  there	
  were	
  no	
  significant	
  correlations	
  in	
  urothelial	
  carcinoma	
  

between	
  survival	
  time	
  and	
  %MSI	
  (p=0.547),	
  MLH1	
  (p=0.446),	
  MSH2	
  (p=0.740),	
  MSH6	
  

	
  

117	
  

	
  
(p=0.262),	
  age	
  at	
  diagnosis	
  (p=0.290),	
  breed	
  (p=0.723),	
  sex	
  (p=0.485),	
  treatment	
  
(p=0.0633),	
  histologic	
  classification	
  of	
  urothelial	
  carcinomas	
  as	
  papillary	
  or	
  infiltrating	
  
(p=0.129),	
  degree	
  of	
  anaplasia	
  (p=0.054),	
  degree	
  of	
  invasion	
  into	
  the	
  bladder	
  wall	
  
(p=0.185).	
  	
  Expression	
  of	
  MSH3,	
  however,	
  was	
  associated	
  with	
  survival	
  time	
  (p=0.022).	
  	
  	
  
Using	
  Kaplan-­‐Meier	
  estimators	
  and	
  log	
  rank	
  tests,	
  there	
  was	
  no	
  significant	
  
correlation	
  between	
  survival	
  time	
  and	
  MSI	
  as	
  categorized	
  as	
  stable,	
  low,	
  or	
  high	
  (p=0.283);	
  
breeds	
  (p=0.704);	
  sex	
  (p=0.476);	
  histologic	
  classification	
  as	
  papillary	
  or	
  infiltrating	
  
(p=0.120);	
  cellular	
  anaplasia	
  (p=0.132);	
  degree	
  of	
  invasion	
  into	
  the	
  bladder	
  wall	
  (p=0.307);	
  
or	
  MSH3	
  expression	
  when	
  grouping	
  cases	
  by	
  having	
  >	
  or	
  <	
  the	
  mean	
  MSH3	
  expression.	
  
There	
  was,	
  however,	
  a	
  significant	
  correlation	
  between	
  survival	
  time	
  and	
  treatment	
  
(p=0.017).	
  	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

118	
  

	
  
Discussion	
  
These	
   investigations	
   demonstrated	
   common,	
   but	
   not	
   universal	
   occurrence	
   of	
   MSI	
   in	
  
canine	
  urothelial	
  carcinomas.	
  Variation	
  in	
  frequency	
  and	
  magnitude	
  of	
  MSI	
  in	
  these	
  cancers	
  
was	
   correlated	
   with	
   genetic	
   background	
   in	
   terms	
   of	
   breed	
   and	
   phylogenetic	
   clade	
  
suggesting	
   likely	
   divergent	
   pathways	
   of	
   urothelial	
   carcinogenesis.	
   There	
   were	
   significant	
  
differences	
   in	
   the	
   expression	
   of	
   the	
   MMR	
   proteins,	
   MSH2	
   and	
   MSH6,	
   between	
   canine	
  
urothelial	
  carcinomas	
  and	
  normal	
  urinary	
  bladder	
  as	
  evaluated	
  by	
  IHC	
  and	
  morphometric	
  
analysis	
  using	
  multispectral	
  imaging.	
  However,	
  variance	
  in	
  MMR	
  protein	
  expression	
  did	
  not	
  
correlate	
   with	
   presence	
   or	
   degree	
   of	
   MSI,	
   breed,	
   or	
   histomorphologic	
   features	
   of	
   urothelial	
  
carcinomas.	
   	
   Neither	
   MSI	
   nor	
   the	
   expression	
   of	
   MMR	
   proteins	
   excluding	
   MSH3	
   was	
  
associated	
  with	
  survival	
  time.	
  
Urinary	
  carcinomas	
  of	
  the	
  urinary	
  bladder	
  represent	
  one	
  of	
  the	
  many	
  cancer	
  types	
  
in	
   humans	
   in	
   which	
   a	
   significant	
   percentage	
   of	
   tumors	
   have	
   been	
   reported	
   to	
   show	
   defects	
  
in	
   MMR.	
   Development	
   of	
   such	
   tumors	
   has	
   been	
   associated	
   with	
   both	
   hereditary	
   and	
  
spontaneously	
   developing	
   MMR	
   deficiency.	
   While	
   often	
   occurring	
   spontaneously,	
  
development	
   of	
   bladder	
   cancers	
   associated	
   with	
   defective	
   MMR	
   also	
   has	
   been	
   well	
  
described	
   as	
   part	
   of	
   the	
   Lynch	
   syndrome.	
   The	
   Lynch	
   syndrome	
   is	
   a	
   hereditary	
  
predisposition	
   for	
   cancer	
   development	
   defined	
   by	
   somatic	
   mutation	
   in	
   MMR	
   genes.12	
   A	
  
wide	
   range	
   of	
   cancers	
   has	
   been	
   described	
   in	
   patients	
   with	
   Lynch	
   syndrome,	
   the	
   most	
  
extensively	
   studied	
   and	
   widely	
   reported	
   of	
   these	
   is	
   hereditary	
   nonpolyposis	
   colorectal	
  
carcinoma.1-­‐4,	
  15	
  A	
  recent	
  study	
  of	
  Lynch	
  syndrome	
  families	
  from	
  4	
  nations	
  including	
  the	
  US	
  
demonstrated	
   that	
   the	
   urinary	
   tract,	
   including	
   the	
   kidney,	
   renal	
   pelvis,	
   ureter,	
   and	
   bladder,	
  
had	
  the	
  highest	
  organ	
  system	
  specific	
  cancer	
  risk	
  after	
  the	
  colon	
  and	
  endometrium.28	
  

	
  

119	
  

	
  
	
  

The	
  importance	
  of	
  MMR	
  dysfunction	
  and	
  MSI	
  in	
  urothelial	
  carcinomas	
  of	
  the	
  upper	
  

urinary	
   tract	
   is	
   well	
   accepted	
   in	
   human	
   medicine.29-­‐39	
   Studies	
   of	
   the	
   MMR	
   in	
   urothelial	
  
carcinomas	
   of	
   the	
   urinary	
   bladder,	
   however,	
   have	
   yielded	
   variable	
   and	
   sometimes	
  
contradictory	
   data.	
   Cancers	
   developing	
   from	
   urothelium	
   of	
   the	
   upper	
   urinary	
   tract	
  
frequently	
   have	
   been	
   associated	
   with	
   MMR	
   defects	
   and	
   often	
   exhibit	
   MSI.29-­‐39	
   Although	
  
urinary	
   carcinomas	
   of	
   the	
   upper	
   urinary	
   tract	
   represent	
   the	
   most	
   common	
   urinary	
   tract	
  
cancers	
   reported	
   with	
   Lynch	
   syndrome,	
   bladder	
   cancers	
   also	
   occur	
   in	
   people	
   harboring	
  
heritable	
   MMR	
   defects.40-­‐43	
   Lynch	
   syndrome	
   associated	
   bladder	
   cancer	
   represents	
   a	
   very	
  
small	
   proportion	
   of	
   bladder	
   cancers;	
   however,	
   data	
   from	
   numerous	
   studies	
   over	
   the	
   last	
  
two	
   decades	
   indicate	
   that	
   a	
   substantial	
   subset	
   of	
   bladder	
   cancers	
   exhibit	
   MSI	
   suggesting	
  
that	
  spontaneous	
  development	
  of	
  MMR	
  defects	
  in	
  urothelial	
  carcinogenesis	
  is	
  common.	
   44,	
  
45	
  	
  

It	
  has	
  previously	
  been	
  suggested	
  that	
  due	
  to	
  the	
  similarities	
  of	
  urothelial	
  carcinomas	
  
of	
   the	
   urinary	
   bladder	
   in	
   dogs	
   and	
   humans,	
   that	
   these	
   canine	
   cancers	
   could	
   be	
   used	
   as	
   a	
  
research	
   model.20,	
   24	
   Similar	
   to	
   humans,	
   urothelial	
   carcinomas	
   of	
   the	
   urinary	
   bladder	
  
comprise	
   a	
   significant	
   proportion	
   of	
   canine	
   neoplasms.	
   Urothelial	
   carcinomas	
   have	
   a	
  
similar	
   histomorphologic	
   appearance	
   in	
   dogs	
   and	
   humans.	
   20,	
   24,	
   27	
   In	
   fact,	
   it	
   has	
   recently	
  
been	
   suggested	
   that	
   canine	
   urothelial	
   carcinomas	
   could	
   be	
   classified	
   and	
   graded	
   according	
  
to	
   the	
   World	
   Health	
   Organization/International	
   Society	
   of	
   Urologic	
   Pathology	
   consensus	
  
classification	
   system,	
   which	
   was	
   developed	
   for	
   use	
   in	
   human	
   pathology.27	
   Papillary	
   and	
  
invasive	
   carcinomas	
   are	
   prevalent	
   in	
   dogs	
   and	
   follow	
   a	
   similar	
   clinical	
   course	
   to	
   those	
  
reported	
   in	
   humans	
   suggesting	
   that	
   the	
   pathways	
   involved	
   in	
   the	
   development	
   and	
  
progression	
  of	
  these	
  tumors	
  may	
  be	
  homologous.20,	
   24	
  Further,	
  particular	
  breeds	
  including	
  

	
  

120	
  

	
  
the	
  Scottish	
  terrier,	
  West	
  Highland	
  white	
  terrier,	
  the	
  Shetland	
  Sheepdog,	
  and	
  Beagle,	
  have	
  
significantly	
   increased	
   risk	
   of	
   developing	
   urothelial	
   carcinomas,	
   making	
   easier	
   the	
   study	
   of	
  
the	
   genetic	
   basis	
   of	
   carcinogenesis	
   in	
   dogs.20,	
   24,	
   25	
   One	
   difference	
   between	
   human	
   and	
  
canine	
  urothelial	
  carcinomas	
  is	
  that	
  in	
  situ	
  carcinomas	
  only	
  rarely	
  identified	
  in	
  dogs	
  while	
  
they	
  are	
  by	
  far	
  the	
  most	
  commonly	
  recognized	
  form	
  of	
  urothelial	
  carcinoma	
  in	
  humans.24,	
  
25,	
  46	
   This,	
   however,	
   may	
   be	
   due	
   to	
   the	
   fact	
   that	
   bladder	
   tumors	
   are	
   generally	
   recognized	
   in	
  

dogs	
   only	
   in	
   late	
   stages	
   of	
   disease	
   when	
   they	
   are	
   large	
   enough	
   to	
   cause	
   significant	
  
functional	
  problems.	
  	
  
	
  

In	
   the	
   current	
   study,	
   there	
   was	
   a	
   high	
   incidence	
   of	
   MSI	
   in	
   canine	
   urothelial	
  

carcinomas	
   in	
   comparison	
   to	
   other	
   tumors	
   with	
   strong	
   breed	
   predispositions	
   (gastric	
  
carcinomas	
   and	
   mammary	
   tumors).	
   Further,	
   MSI	
   in	
   canine	
   urothelial	
   carcinomas	
   was	
  
correlated	
   with	
   breed,	
   with	
   the	
   phylogenetically	
   related	
   breeds	
   Scottish	
   terrier	
   and	
   West	
  
Highland	
   white	
   terrier	
   often	
   having	
   high	
   MSI.26	
   Statistical	
   correlations	
   were	
   particularly	
  
strong	
  when	
  comparing	
  only	
  the	
  four	
  breed	
  groups	
  with	
  high	
  relative	
  risk	
  for	
  development	
  
of	
   urothelial	
   carcinomas.	
   The	
   relatively	
   higher	
   p-­‐value	
   obtained	
   in	
   permutation	
   tests	
   when	
  
breeds	
   not	
   predisposed	
   for	
   urothelial	
   carcinomas	
   was	
   included	
   in	
   analysis	
   as	
   an	
   additional	
  
category	
   can	
   be	
   explained	
   by	
   the	
   higher	
   variability	
   of	
   MSI	
   in	
   tumors	
   derived	
   from	
   this	
  
genetically	
  heterogeneous	
  group	
  composed	
  of	
  17	
  other	
  purebreds	
  and	
  mixed	
  breed	
  dogs.	
  
	
  

In	
   humans,	
   evidence	
   suggests	
   that	
   tumors	
   of	
   different	
   types	
   with	
   high	
   MSI	
   may	
  

differ	
  in	
  the	
  type	
  MS	
  repeat	
  affected.	
  In	
  human	
  upper	
  urinary	
  tract	
  urothelial	
  carcinomas,	
  
MSI	
   in	
   mono-­‐	
   and	
   dinucleotide	
   repeat	
   MS	
   is	
   common	
   and	
   usually	
   associated	
   with	
   defective	
  
DNA	
   mismatch	
   repair	
   (MMR).47,	
   48	
   While	
   the	
   underlying	
   mechanisms	
   of	
   development	
   and	
  
significance	
   are	
   poorly	
   understood,	
   MSI	
   in	
   tetranucleotide	
   repeats,	
   termed	
   EMAST	
  

	
  

121	
  

	
  
(elevated	
  microsatellite	
  alterations	
  at	
  select	
  tetranucloetides),	
  is	
  common	
  in	
  human	
  lower	
  
urinary	
  tract	
  urothelial	
  carcinomas,	
  while	
  MSI	
  in	
  mono-­‐	
  or	
  dinucleotide	
  repeats	
  is	
  rare.48,	
  49	
  
Differences	
  in	
  the	
  type	
  of	
  MS	
  affected	
  in	
  specific	
  tumors	
  may	
  indicate	
  that	
  while	
  MSI	
  occurs	
  
in	
   many	
   different	
   tumor	
   types,	
   the	
   mechanisms	
   underlying	
   MSI	
   development	
   within	
  
particular	
   types	
   of	
   tumors	
   varies.	
   In	
   the	
   currently	
   evaluated	
   canine	
   urothelial	
   carcinoma	
  
set,	
   MSI	
   was	
   observed	
   in	
   MS	
   composed	
   of	
   both	
   tertranucleotide	
   and	
   short	
   mono-­‐	
   and	
  
dinucleotide	
   repeats.	
   	
   While,	
   there	
   was	
   no	
   statistical	
   associations	
   between	
   MSI	
   in	
  
tetranucleotide	
   repeat	
   motifs	
   and	
   dog	
   breed,	
   there	
   were	
   associations	
   between	
   both	
  
magnitude	
   and	
   frequency	
   of	
   MSI	
   in	
   short	
   (mono-­‐	
   and	
   dinucleotide)	
   repeats	
   with	
   the	
  
phylogenetically	
   related	
   breeds	
   of	
   West	
   Highland	
   white	
   terrier	
   and	
   Scottish	
   terrier	
   being	
  
most	
   prominently	
   represented.	
   This	
   suggests	
   that	
   not	
   only	
   are	
   these	
   breeds	
   predisposed	
  
for	
  development	
  of	
  MSI,	
  but	
  that	
  the	
  mechanisms	
  behind	
  such	
  development	
  are	
  similar.	
  
	
  

We	
   also	
   hypothesized	
   that	
   the	
   degree	
   of	
   MSI	
   in	
   individual	
   canine	
   urothelial	
  

carcinomas	
  would	
  be	
  correlated	
  with	
  the	
  expression	
  of	
  MMR	
  proteins.	
  	
  Several	
  studies	
  have	
  
investigated	
  expression	
  of	
  MMR	
  proteins	
  in	
  bladder	
  cancer	
  using	
  immunohistochemistry	
  in	
  
humans.44,	
   50-­‐55	
   MMR	
   IHC	
   appears	
   to	
   correlate	
   fairly	
   well	
   with	
   MMR	
   functional	
   status	
   in	
  
human	
  colon	
  cancers	
  and	
  similar	
  findings	
  are	
  likely	
  in	
  other	
  tumors.56,	
  57	
  Multiple	
  studies	
  of	
  
MMR	
  IHC	
  in	
  human	
  bladder	
  cancer	
  report	
  common	
  loss	
  or	
  decreased	
  expression	
  of	
  MMR	
  
proteins	
   including	
   MLH1,	
   MSH2,	
   MSH6,	
   and	
   MSH3	
   in	
   subsets	
   of	
   tumors.40,	
   42,	
   50-­‐54,	
   58	
   The	
  
percentage	
   of	
   bladder	
   cancers	
   reported	
   to	
   exhibit	
   MMR	
   deficiency	
   according	
   to	
   IHC	
  
evaluation	
  of	
  MMR	
  proteins	
  varies	
  between	
  studies,	
  ranging	
  from	
  0-­‐69%.40,	
   42,	
   50-­‐54,	
   58	
  Such	
  
seemingly	
   variable	
   prevalence	
   may	
   be	
   attributable	
   to	
   many	
   factors	
   including	
   the	
   particular	
  

	
  

122	
  

	
  
proteins	
  evaluated	
  and	
  the	
  histopathologic	
  tumor	
  subtypes	
  included	
  in	
  the	
  study,	
  as	
  well	
  as	
  
technical	
  differences	
  in	
  the	
  performance	
  and	
  scoring	
  of	
  IHC.	
  	
  
	
  

In	
   the	
   canine	
   urothelial	
   carcinomas	
   of	
   the	
   urinary	
   bladder	
   of	
   the	
   current	
   study,	
  

lower	
   expression	
   of	
   MSH2	
   and	
   higher	
   expression	
   of	
   MSH6	
   were	
   in	
   comparison	
   to	
   that	
   of	
  
normal	
   urothelium	
   from	
   unaffected	
   urinary	
   bladders.	
   Such	
   relatively	
   low	
   expression	
   of	
  
MSH2	
   and	
   high	
   expression	
   of	
   MSH6	
   was	
   often	
   concurrent	
   within	
   individual	
   urothelial	
  
carcinomas,	
   and	
   there	
   were	
   strong	
   statistical	
   differences	
   between	
   the	
   ratios	
   of	
   MSH2	
   to	
  
MSH6	
  when	
  comparing	
  urothelial	
  carcinomas	
  to	
  normal	
  urinary	
  bladders.	
  The	
  expression	
  
of	
   MLH1	
   and	
   MSH3	
   was	
   not	
   significantly	
   different	
   between	
   canine	
   urothelial	
   carcinomas	
   in	
  
comparison	
  to	
  that	
  of	
  normal	
  urothelium	
  from	
  unaffected	
  control	
  bladders.	
  
	
  

Interestingly,	
   the	
   expression	
   of	
   neither	
   MSH2	
   nor	
   MSH6	
   in	
   canine	
   urothelial	
  

carcinomas	
   was	
   associated	
   with	
   the	
   frequency	
   or	
   degree	
   of	
   MSI.	
   In	
   fact,	
   decreased	
   MSH2	
  
and	
   increased	
   MSH6	
   expression	
   relative	
   to	
   that	
   in	
   normal	
   urinary	
   bladders	
   were	
   found	
  
more	
   frequently	
   in	
   canine	
   urothelial	
   carcinomas	
   than	
   was	
   high	
   MSI.	
   Given	
   the	
   lack	
   of	
  
correlation	
   between	
   MMR	
   protein	
   expression	
   and	
   MSI,	
   it	
   is	
   not	
   surprising	
   that	
   no	
  
association	
   was	
   found	
   between	
   MMR	
   protein	
   expression	
   and	
   breed,	
   which	
   was	
   strongly	
  
associated	
  with	
  MSI.	
  	
  The	
  exact	
  significance	
  of	
  this	
  disjunction	
  between	
  variations	
  in	
  MMR	
  
protein	
   expression	
   and	
   MSI	
   is	
   unclear.	
   	
   It	
   is	
   possible	
   that	
   IHC	
   for	
   MMR	
   proteins	
   is	
   more	
  
sensitive	
   in	
   identifying	
   variations	
   in	
   the	
   MMR	
   repair	
   pathway,	
   but	
   that	
   these	
   variations	
   are	
  
not	
  functionally	
  significant	
  and	
  therefore	
  do	
  not	
  result	
  in	
  MSI.	
  	
  It	
  is	
  also	
  possible	
  that	
  the	
  
relative	
  small	
  pool	
  of	
  MS	
  evaluated	
  in	
  this	
  study	
  compared	
  to	
  the	
  total	
  number	
  of	
  MS	
  in	
  the	
  
genome	
  was	
  not	
  completely	
  reflective	
  of	
  MMR	
  function.	
  In	
  human	
  colon	
  cancer,	
  there	
  has	
  

	
  

123	
  

	
  
been	
  considerable	
  controversy	
  in	
  deciding	
  how	
  best	
  to	
  evaluate	
  tumors	
  for	
  defective	
  MMR:	
  
microsatellite	
  analysis	
  or	
  IHC57,	
  59.	
  It	
  seems	
  that	
  both	
  methods	
  have	
  limitations.	
  	
  
	
  

With	
   the	
   rare	
   exception	
   of	
   MSH3	
   survival	
   correlating	
   with	
   survival	
   time,	
   there	
   not	
  

appreciable	
   correlations	
   between	
   MSI	
   or	
   MMR	
   protein	
   expression	
   and	
   survival	
   time	
   or	
  
histomorphologic	
   features	
   of	
   canine	
   urothelial	
   carcinomas	
   including	
   classification	
   as	
  
primary	
   infiltrative	
   or	
   papillary,	
   histologic	
   grade,	
   or	
   depth	
   of	
   invasion	
   into	
   the	
   urinary	
  
bladder	
   wall.	
   	
   The	
   significance	
   of	
   the	
   correlation	
   between	
   MSH3	
   expression	
   and	
   survival	
  
time	
  is	
  unclear,	
  especially	
  given	
  that	
  there	
  was	
  no	
  statistical	
  difference	
  in	
  MSH3	
  expression	
  
between	
  urothelium	
  of	
  normal	
  bladders	
  and	
  urothelial	
  carcinomas.	
  	
  Overall,	
  these	
  findings	
  
suggest	
   that	
   while	
   the	
   loss	
   of	
   MMR	
   in	
   certain	
   bladder	
   cancers	
   is	
   clear,	
   the	
   significance	
   of	
  
such	
   losses	
   is	
   less	
   certain.	
   In	
   humans,	
   an	
   assortment	
   of	
   correlations	
   between	
   clinical	
   and	
  
pathologic	
   variables	
   and	
   MMR	
   status	
   has	
   been	
   reported;	
   however,	
   there	
   is	
   not	
   universal	
  
agreement	
   as	
   to	
   which	
   variables	
   correlate	
   with	
   MMR	
   or	
   the	
   direction	
   of	
   the	
   correlation	
  
(positive	
  or	
  negative).	
  For	
  instance,	
  one	
  study	
  reports	
  the	
  loss	
  of	
  MLH1	
  and	
  MSH2	
  is	
  most	
  
common	
  in	
  high	
  grade	
  and	
  invasive	
  tumors50,	
  whereas	
  another	
  study	
  associates	
  MMR	
  loss	
  
with	
   noninvasive,	
   well-­‐differentiated	
   tumors53.	
   Further,	
   differential	
   expression	
   of	
   MMR	
  
proteins	
  including	
  MSH2,	
  MSH3,	
  and	
  MLH1	
  has	
  been	
  associated	
  with	
  urothelial	
  carcinoma	
  
grade	
   and	
   clinical	
   outcome.16,	
   48,	
   60	
   In	
   seeming	
   contradiction,	
   several	
   studies	
   indicate	
   that	
  
loss	
  of	
  MMR,	
  even	
  in	
  invasive	
  tumors	
  may	
  indicate	
  a	
  better	
  prognosis	
  similar	
  to	
  what	
  has	
  
been	
  reported	
  in	
  colon	
  cancers.50,	
  53,	
  54	
  
The	
  current	
  study	
  was	
  limited	
  by	
  the	
  relative	
  low	
  number	
  of	
  cases	
  and,	
  in	
  particular,	
  
the	
  low	
  number	
  of	
  cases	
  of	
  each	
  breed	
  and	
  the	
  low	
  number	
  of	
  low	
  grade	
  urothelial	
  
carcinomas.	
  High	
  grade	
  urothelial	
  carcinomas	
  are	
  far	
  more	
  commonly	
  diagnosed	
  in	
  dogs	
  

	
  

124	
  

	
  
than	
  low	
  grade	
  tumors	
  accounting	
  for	
  the	
  low	
  number	
  of	
  low	
  grade	
  tumors	
  in	
  the	
  examined	
  
set.	
  This	
  may	
  be	
  due	
  to	
  an	
  actual	
  lower	
  rate	
  of	
  occurrence	
  in	
  canine	
  populations	
  or	
  
diagnosis	
  of	
  urothelial	
  neoplasms	
  only	
  late	
  in	
  the	
  course	
  of	
  disease	
  when	
  larger,	
  more	
  
aggressive	
  tumors	
  affect	
  normal	
  micturition	
  or	
  cause	
  other	
  clinical	
  disease.	
  Further,	
  
euthanasia	
  is	
  common	
  in	
  veterinary	
  medicine	
  and	
  is	
  especially	
  common	
  in	
  animals	
  with	
  
urothelial	
  carcinomas	
  due	
  to	
  the	
  negative	
  prognosis	
  historically	
  associated	
  with	
  such	
  
tumors,	
  the	
  effects	
  on	
  micturition,	
  the	
  common	
  non-­‐resectable	
  nature	
  and	
  high	
  number	
  of	
  
cases	
  with	
  metastases	
  of	
  such	
  tumors	
  at	
  the	
  time	
  of	
  diagnosis,	
  and	
  client	
  choice	
  due	
  to	
  
financial	
  or	
  homecare	
  constraints.	
  Treatment	
  was	
  also	
  not	
  standardized	
  between	
  animals.	
  	
  
It	
  is	
  likely	
  that	
  variation	
  in	
  treatment	
  affected	
  survival	
  time,	
  especially	
  given	
  that	
  treatment	
  
was	
  one	
  of	
  the	
  few	
  variables	
  examined	
  in	
  this	
  study	
  that	
  correlated	
  with	
  survival	
  time.	
  
	
  

Despite	
   inherent	
   limitations,	
   this	
   study	
   does	
   suggest	
   that	
   canine	
   urothelial	
  

carcinomas	
   have	
   potential	
   as	
   a	
   model	
   for	
   further	
   evaluation	
   of	
   the	
   MMR	
   system	
   with	
  
respect	
   to	
   carcinogenesis.	
   	
   This	
   work	
   demonstrates	
   that	
   MSI	
   can	
   be	
   observed	
   in	
  
spontaneous	
   canine	
   tumors.	
   Further,	
   the	
   frequent	
   occurrence	
   of	
   MSI	
   and	
   the	
   variation	
   in	
  
expression	
   of	
   MMR	
   proteins	
   in	
   canine	
   urothelial	
   carcinomas	
   of	
   the	
   urinary	
   bladder	
  
compared	
  to	
  canine	
  epithelial	
  tumors	
  from	
  other	
  tissues	
  suggests	
  that	
  MSI	
  may	
  contribute	
  
to	
  urothelial	
  carcinogenesis	
  in	
  the	
  dog,	
  and	
  that	
  there	
  is	
  likely	
  an	
  inheritable	
  propensity	
  for	
  
MMR	
   dysfunction	
   in	
   specific	
   breeds.	
   The	
   clinical	
   implications	
   of	
   deficiencies	
   in	
   MMR	
   in	
  
canine	
   urothelial	
   carcinomas	
   are	
   less	
   clear	
   as	
   neither	
   MSI	
   nor	
   MMR	
   protein	
   expression	
  
correlated	
  with	
  survival	
  time	
  or	
  histomorphologic	
  features	
  often	
  associated	
  with	
  prognosis.	
  	
  	
  
	
  
Acknowledgements	
  

	
  

125	
  

	
  
These	
  studies	
  were	
  supported	
  in	
  part	
  by	
  grants	
  from	
  the	
  National	
  Cancer	
  Institute	
  
(K08-­‐CA89530,	
  R03_CA101030)	
  and	
  a	
  comparative	
  medicine	
  grant	
  from	
  the	
  University	
  of	
  
Minnesota	
  College	
  of	
  Veterinary	
  Medicine.	
  Portions	
  of	
  the	
  described	
  work	
  comprised	
  part	
  
of	
  Dr.	
  Sledge’s	
  PhD	
  training,	
  which	
  was	
  funded	
  by	
  Bristol-­‐Meyers-­‐Squibb	
  through	
  the	
  
American	
  College	
  of	
  Veterinary	
  Pathologists/Society	
  of	
  Toxicologic	
  Pathologists	
  coalition.	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

126	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
APPENDIX	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

127	
  

	
  
Figure	
   16:	
   Unmixed	
   composite	
   images	
   and	
   unmixed	
   pseudofluorecent	
   images	
   derived	
   from	
  
multispectral	
   imaging	
   of	
   MLH1,	
   MSH2,	
   MSH3,	
   and	
   MSH6	
   immunohistochemically	
   labeled	
  
urothelium	
   from	
   a	
   normal	
   urinary	
   bladder	
   control	
   and	
   a	
   grade	
   II	
   papillary	
   urothelial	
  
carcinoma.	
   	
   MLH1	
   and	
   MSH6	
   IHC:	
   Indirect	
   biotin	
   streptavidin	
   and	
   Fast	
   Red	
   chromogen,	
  
hematoxylin	
   counterstain.	
   MSH2	
   and	
   MSH3	
   IHC:	
   3,3’-­‐Diaminobenzidine	
   (DAB)	
   chromogen,	
  
hematoxylin	
  counterstain.	
  

	
  
	
  

	
  

128	
  

	
  
Figure	
   17:	
   Scatter	
   plot	
   depicting	
   the	
   frequency	
   of	
   microsatellite	
   aberrations	
   identified	
   in	
  
canine	
   gastric	
   carcinoma	
   tumor	
   (N=15),	
   mammary	
   tumors	
   (N=35),	
   and	
   urothelial	
  
carcinomas	
  of	
  the	
  urinary	
  bladder	
  (N=46).	
  Each	
  circle	
  represents	
  an	
  individual	
  tumor.	
  	
  The	
  
horizontal	
  bar	
  associated	
  with	
  each	
  dataset	
  indicates	
  the	
  mean	
  %MSI	
  for	
  the	
  group.	
  Many	
  
urothelial	
  carcinomas	
  have	
  high	
  %MSI	
  defined	
  by	
  25%	
  of	
  evaluated	
  MS	
  having	
  instability,	
  
while	
  only	
  few	
  mammary	
  tumors	
  and	
  no	
  gastric	
  carcinomas	
  have	
  high	
  %MSI.	
  	
  In	
  addition,	
  
the	
  mean	
  %MSI	
  for	
  urothelial	
  carcinomas	
  is	
  higher	
  than	
  that	
  of	
  other	
  evaluated	
  tumors.	
  

	
  
	
  
	
  
	
  
	
  

	
  

129	
  

	
  
Figure	
   18:	
   Scatter	
   plot	
   depicting	
   the	
   frequency	
   of	
   microsatellite	
   aberrations	
   identified	
   in	
  
canine	
   urothelial	
   carcinomas	
   stratified	
   by	
   breed.	
   Each	
   circle	
   represents	
   an	
   individual	
  
urothelial	
  carcinoma.	
  	
  The	
  horizontal	
  bar	
  associated	
  with	
  each	
  dataset	
  indicates	
  the	
  mean	
  
%MSI	
  for	
  the	
  group.	
  The	
  two	
  dashed	
  horizontal	
  lines	
  indicate	
  the	
  level	
  for	
  MSI-­‐L	
  and	
  MSI-­‐H	
  
classification.	
  Variation	
  in	
  the	
  fraction	
  of	
  MS	
  with	
  instability	
  within	
  individual	
  breed	
  groups	
  
is	
  common;	
  however,	
  in	
  comparison	
  to	
  other	
  breeds,	
  MSI-­‐H	
  is	
  common	
   in	
   West	
   Highland	
  
white	
  terriers	
  and	
  Scottish	
  terriers.	
  	
  

	
  
	
  
	
  
	
  
	
  

	
  

130	
  

	
  
Figure	
   19:	
   Scatter	
   plot	
   depicting	
   the	
   frequency	
   of	
   microsatellite	
   aberrations	
   identified	
   in	
  
canine	
   urothelial	
   carcinomas	
   stratified	
   by	
   phylogenetic	
   clade.	
   	
   Each	
   circle	
   represents	
   an	
  
individual	
  urothelial	
  carcinoma.	
  	
  The	
  horizontal	
  bar	
  associated	
  with	
  each	
  dataset	
  indicates	
  
the	
   mean	
   %MSI	
   for	
   the	
   group.	
   The	
   dashed	
   horizontal	
   lines	
   indicate	
   the	
   level	
   MSI-­‐H	
  
classification.	
  Variation	
  in	
  the	
  fraction	
  of	
  MS	
  with	
  instability	
  within	
  individual	
  phylogenetic	
  
groups	
  is	
  common;	
  however,	
  MSI-­‐H	
  is	
  common	
  phylogenetic	
  clade	
  5,	
  which	
  includes	
  West	
  
Highland	
  white	
  terriers	
  and	
  Scottish	
  terriers.	
  

	
  
	
  
	
  
	
  

	
  

131	
  

	
  
Figure	
   20:	
   Scatter	
   plots	
   depicting	
   immunoreactivity	
   of	
   MSH2	
   and	
   MSH6	
   in	
   canine	
   normal	
  
urinary	
   bladder	
   urothelium	
   and	
   urothelial	
   carcinomas	
   as	
   determined	
   by	
   morphometric	
  
analysis	
   of	
   multispectral	
   imaging.	
   Each	
   circle	
   represents	
   an	
   individual	
   case.	
   Horizontal	
   bars	
  
associated	
   with	
   each	
   dataset	
   indicate	
   the	
   mean	
   relative	
   expression	
   of	
   MSH2	
   or	
   MSH6	
   for	
  
each	
   group.	
   While	
   there	
   is	
   variation	
   in	
   relative	
   expression	
   of	
   MSH2	
   and	
   MSH6	
   in	
   both	
  
normal	
   urothelium	
   and	
   urothelial	
   carcinomas,	
   the	
   mean	
   relative	
   expression	
   of	
   MSH2	
   is	
  
lower	
   in	
   urothelial	
   carcinomas,	
   than	
   in	
   normal	
   urothelium	
   (A),	
   and	
   the	
   mean	
   relative	
  
expression	
  of	
  MSH6	
  is	
  higher	
  in	
  urothelial	
  carcinomas,	
  than	
  in	
  normal	
  urothelium	
  (B).	
  

	
  
	
  
	
  
	
  
	
  
	
  

	
  

132	
  

	
  
Figure	
   21:	
   Scatter	
   plot	
   depicting	
   ratio	
   of	
   MSH2	
   and	
   MSH6	
   immunoreactivity	
   in	
   individual	
  
cases	
  of	
  canine	
  normal	
  urinary	
  bladder	
  urothelium	
  and	
  urothelial	
  carcinomas	
  as	
  determined	
  
by	
   morphometric	
   analysis	
   of	
   multispectral	
   imaging.	
   Each	
   circle	
   represents	
   an	
   individual	
  
case.	
   Horizontal	
   bars	
   associated	
   with	
   each	
   dataset	
   indicate	
   the	
   mean	
   ratio	
   of	
   MSH2	
   and	
  
MSH6	
  for	
  each	
  group.	
  While	
  there	
  is	
  variation	
  in	
  the	
  ration	
  of	
  MSH2	
  to	
  MSH6	
  expression	
  for	
  
individual	
   cases	
   in	
   both	
   normal	
   urothelium	
   and	
   urothelial	
   carcinomas,	
   the	
   mean	
   ratio	
   of	
  
MSH2	
  to	
  MSH6	
  is	
  lower	
  in	
  urothelial	
  carcinomas,	
  than	
  in	
  normal	
  urothelium.	
  

	
  
	
  
	
  
	
  
	
  
	
  

	
  

133	
  

	
  
Table	
  10:	
  Distribution	
  of	
  microsatellite	
  instability	
  in	
  canine	
  epithelial	
  tumors	
  
MSS	
  

MSI-­‐L	
  

MSI-­‐H	
  	
  

Tumor	
  
type	
  

No.	
  
tumors	
  
(%)	
  

No.	
  tumors	
  
(%)	
  

No.	
  Tumors	
  (%)	
  

GC	
  

14	
  (93%)	
  

1	
  (7%)	
  

0	
  (0%)	
  

MT	
  

28	
  (80%)	
  

3	
  (9%)	
  

4	
  (11%)	
  

TCC	
  

16	
  (36%)	
  

16	
  (36%)	
  

13	
  (28%)	
  

GC	
  =	
  Gastric	
  carcinoma;	
  MT	
  =	
  Mammary	
  gland	
  tumor;	
  TCC	
  =	
  
Transitional	
  cell	
  carcinoma	
  of	
  the	
  bladder;	
  MSS	
  =	
  
microsatellite	
  stable;	
  MSI-­‐L	
  =	
  at	
  least	
  10%	
  MSI;	
  MSI-­‐H	
  =	
  at	
  
least	
  25%	
  MSI.	
  	
  
	
  
	
  
Table11:	
  Distribution	
  of	
  microsatellite	
  instability	
  in	
  canine	
  urothelial	
  carcinomas	
  of	
  the	
  
urinary	
  bladder	
  classified	
  by	
  breed	
  
#	
  of	
  
Tumors	
  

#	
  MSS	
  

#	
  MSI-­‐L	
  

#	
  MSI-­‐H	
  

	
  Mean	
  %MSI	
  

Westie	
  

3	
  

0	
  

0	
  

3	
  

43.7%	
  (±	
  28.2%)	
  

Scottie	
  

6	
  

1	
  

1	
  

4	
  

37.1%	
  (±	
  30.7%)	
  

Sheltie	
  

6	
  

3	
  

3	
  

0	
  

10.9%	
  (±	
  10.7%)	
  

Beagle	
  

4	
  

3	
  

1	
  

0	
  

4.8%	
  (±	
  6.7%)	
  

Other	
  	
  

26	
  

7	
  

13	
  

6	
  

22.8%	
  (±	
  25.1%)	
  

Breed	
  

MSS	
  =	
  microsatellite	
  stable;	
  MSI-­‐L	
  =	
  at	
  least	
  10%	
  MSI;	
  MSI-­‐H	
  =	
  at	
  least	
  25%	
  MSI.	
  	
  
	
  

	
  

134	
  

	
  
Table	
  12:	
  Distribution	
  of	
  microsatellite	
  instability	
  in	
  canine	
  urothelial	
  carcinomas	
  of	
  the	
  
urinary	
  bladder	
  by	
  breed	
  and	
  repeat	
  motif	
  

Breed	
  

Mean	
  %MSI	
  

Mean	
  %MSI	
  
Di/Mono	
  

Mean	
  %MSI	
  
Tetra	
  

Westie	
  

43.7%	
  (±	
  28.2%)	
  

52.8%	
  (±	
  21.0%)	
  

30.1%	
  (±	
  41.8%)	
  

Scottie	
  

37.1%	
  (±	
  30.7%)	
  

32.2%	
  (±	
  29.3%)	
  

45.7%	
  (±	
  36.1%)	
  

Sheltie	
  

10.9%	
  (±	
  10.7%)	
  

9.4%	
  (±	
  8.9%)	
  

13.0%	
  (±	
  17.8%)	
  

Beagle	
  

4.8%	
  (±	
  6.7%)	
  

4.2%%	
  (±	
  8.3%)	
  

5.6%	
  (±	
  6.4%)	
  

Other	
  	
  

22.8%	
  (±	
  25.1%)	
  

21.4%	
  (±	
  25.0%)	
  

25.6%	
  (±	
  26.4%)	
  

%	
  MSI	
  =	
  mean	
  percentage	
  of	
  MS	
  markers	
  demonstrating	
  instability	
  for	
  all	
  tumors	
  
and	
  all	
  MS	
  markers;	
  Mean	
  %MSI	
  Di/Mono	
  =	
  Mean	
  %MSI	
  for	
  di-­‐	
  or	
  mononucleotide	
  
repeats	
  only;	
  Mean	
  %MSI	
  Tetra	
  =	
  Mean	
  %	
  MSI	
  for	
  tetranucleotide	
  repeats	
  only.	
  	
  
Standard	
  deviation	
  is	
  given	
  in	
  parentheses.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

135	
  

	
  
Table	
  13:	
  Distribution	
  of	
  %MSI	
  and	
  immunohistochemically	
  evaluated	
  MMR	
  protein	
  
expression	
  in	
  canine	
  normal	
  urinary	
  bladders	
  and	
  urothelial	
  carcinomas	
  of	
  the	
  urinary	
  
bladder	
  

	
  	
  
Mean	
  
Median	
   Minimum	
  
Maximum	
  
%MSI	
  	
  
	
  	
  
	
  	
  
	
  
	
  
	
  	
  All	
  cases	
  
14.4%	
  (±3.2%)	
   9.5%	
  
0%	
  
89.0%	
  
	
  	
  Normal	
  urinary	
  bladder	
  
1.7%	
  (±1.7%)	
  
0%	
  
0%	
  
8.0%	
  
	
  	
  All	
  urothelial	
  carcinomas	
  
17.4%	
  (±3.7%)	
   13.6%	
  
0%	
  
89.0%	
  
	
  	
  	
  	
  Infiltrating	
  carcinomas	
  
27.2%	
  (±16.0%)	
   11.8%	
  
5.0%	
  
89.0%	
  
	
  	
  	
  	
  Papillary	
  carcinomas	
  
14.9%	
  (±2.5%)	
   14.0%	
  
0%	
  
38.0%	
  
MLH1	
  (postive	
  pixels	
  per	
  nucleus)	
  
	
  	
  
	
  	
  
	
  
	
  
	
  	
  All	
  cases	
  
740.3	
  (±113.6)	
   618.5	
  
7.1	
  
2268.0	
  
	
  	
  Normal	
  urinary	
  bladder	
  
775.5	
  (±225.2)	
   755.5	
  
26.6	
  
1245.0	
  
	
  	
  All	
  urothelial	
  carcinomas	
  
695.4	
  (±111.0)	
   492.9	
  
0.2	
  
2268.0	
  
	
  	
  	
  	
  Infiltrating	
  carcinomas	
  
571.1	
  (±229.2)	
   466.3	
  
15.0	
  
1353.9	
  
	
  	
  	
  	
  Papillary	
  carcinomas	
  
797.4	
  (±158.0)	
   653.1	
  
7.1	
  
2268.0	
  
MSH2	
  (postive	
  pixels	
  per	
  nucleus)	
  
	
  	
  
	
  	
  
	
  
	
  
	
  	
  All	
  cases	
  
351.0	
  (±75.9)	
  
209.7	
  
1.7	
  
1629.1	
  
	
  	
  Normal	
  urinary	
  bladder	
  
750.2	
  (±235.6)	
   768.2	
  
189.2	
  
1492.6	
  
	
  	
  All	
  urothelial	
  carcinomas	
  
328.8	
  (±67.6)	
  
175.8	
  
1.7	
  
1629.1	
  
	
  	
  	
  	
  Infiltrating	
  carcinomas	
  
164.8	
  (±52.1)	
  
134.8	
  
46.8	
  
343.2	
  
	
  	
  	
  	
  Papillary	
  carcinomas	
  
301.5	
  (±91.2)	
  
157.0	
  
1.7	
  
1629.1	
  
MSH3	
  (postive	
  pixels	
  per	
  nucleus)	
  
	
  	
  
	
  	
  
	
  
	
  
	
  	
  All	
  cases	
  
	
  	
  
	
  	
  
	
  
	
  
	
  	
  Normal	
  urinary	
  bladder	
  
1410.0	
  (±197.9)	
   1291.9	
  
684.0	
  
2566.0	
  
	
  	
  All	
  urothelial	
  carcinomas	
  
1179.6	
  (±125.9)	
   1107.8	
  
102.0	
  
2662.0	
  
	
  	
  	
  	
  Infiltrating	
  carcinomas	
  
915.8	
  (±248.7)	
   714.6	
  
200.0	
  
2133.0	
  
	
  	
  	
  	
  Papillary	
  carcinomas	
  
1264.0	
  (±144.5)	
   1143.9	
  
102.0	
  
2662.0	
  
MSH6	
  (postive	
  pixels	
  per	
  nucleus)	
  
	
  	
  
	
  	
  
	
  
	
  
	
  	
  All	
  cases	
  
570.7	
  (±96.2)	
  
399.3	
  
27.8	
  
2490.0	
  
	
  	
  Normal	
  urinary	
  bladder	
  
255.6	
  (±60.6)	
  
256.0	
  
112.4	
  
399.3	
  
	
  	
  All	
  urothelial	
  carcinomas	
  
603.2	
  (±94.9)	
  
417.1	
  
1.9	
  
2490.0	
  
	
  	
  	
  	
  Infiltrating	
  carcinomas	
  
746.8	
  (±243.7)	
   1054.0	
  
27.8	
  
1225.8	
  
	
  	
  	
  	
  Papillary	
  carcinomas	
  
619.8	
  (±131.4)	
   450.5	
  
41.0	
  
2490.0	
  
%	
  MSI	
  =	
  mean	
  percentage	
  of	
  MS	
  markers	
  demonstrating	
  instability	
  for	
  all	
  tumors	
  and	
  all	
  MS	
  
markers;	
  	
  Standard	
  deviation	
  for	
  means	
  	
  is	
  given	
  in	
  parentheses.	
  	
  	
  	
  

	
  

136	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
REFERENCES	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

137	
  

	
  
REFERENCES	
  
	
  
1.	
  Harfe	
  BD,	
  Jinks-­‐Robertson	
  S.	
  DNA	
  mismatch	
  repair	
  and	
  genetic	
  instability.	
  Annual	
  review	
  
of	
  genetics.	
  2000;34:	
  359-­‐399.	
  
2.	
  Hsieh	
  P,	
  Yamane	
  K.	
  DNA	
  mismatch	
  repair:	
  molecular	
  mechanism,	
  cancer,	
  and	
  ageing.	
  
Mechanisms	
  of	
  ageing	
  and	
  development.	
  2008;129:	
  391-­‐407.	
  
3.	
  Iyer	
  RR,	
  Pluciennik	
  A,	
  Burdett	
  V,	
  Modrich	
  PL.	
  DNA	
  mismatch	
  repair:	
  functions	
  and	
  
mechanisms.	
  Chemical	
  reviews.	
  2006;106:	
  302-­‐323.	
  
4.	
  Jiricny	
  J.	
  The	
  multifaceted	
  mismatch-­‐repair	
  system.	
  Nature	
  reviews.	
  Molecular	
  cell	
  
biology.	
  2006;7:	
  335-­‐346.	
  
5.	
  Preston	
  BD,	
  Albertson	
  TM,	
  Herr	
  AJ.	
  DNA	
  replication	
  fidelity	
  and	
  cancer.	
  Seminars	
  in	
  
cancer	
  biology.	
  2010;20:	
  281-­‐293.	
  
6.	
  Chung	
  H,	
  Young	
  DJ,	
  Lopez	
  CG,	
  et	
  al.	
  Mutation	
  rates	
  of	
  TGFBR2	
  and	
  ACVR2	
  coding	
  
microsatellites	
  in	
  human	
  cells	
  with	
  defective	
  DNA	
  mismatch	
  repair.	
  PloS	
  one.	
  2008;3:	
  
e3463.	
  
7.	
  Kim	
  CJ,	
  Lee	
  JH,	
  Song	
  JW,	
  et	
  al.	
  Chk1	
  frameshift	
  mutation	
  in	
  sporadic	
  and	
  hereditary	
  non-­‐
polyposis	
  colorectal	
  cancers	
  with	
  microsatellite	
  instability.	
  European	
  journal	
  of	
  surgical	
  
oncology	
  :	
  the	
  journal	
  of	
  the	
  European	
  Society	
  of	
  Surgical	
  Oncology	
  and	
  the	
  British	
  
Association	
  of	
  Surgical	
  Oncology.	
  2007;33:	
  580-­‐585.	
  
8.	
  Miquel	
  C,	
  Jacob	
  S,	
  Grandjouan	
  S,	
  et	
  al.	
  Frequent	
  alteration	
  of	
  DNA	
  damage	
  signalling	
  and	
  
repair	
  pathways	
  in	
  human	
  colorectal	
  cancers	
  with	
  microsatellite	
  instability.	
  Oncogene.	
  
2007;26:	
  5919-­‐5926.	
  
9.	
  Fernandez-­‐Peralta	
  AM,	
  Nejda	
  N,	
  Oliart	
  S,	
  Medina	
  V,	
  Azcoita	
  MM,	
  Gonzalez-­‐Aguilera	
  JJ.	
  
Significance	
  of	
  mutations	
  in	
  TGFBR2	
  and	
  BAX	
  in	
  neoplastic	
  progression	
  and	
  patient	
  
outcome	
  in	
  sporadic	
  colorectal	
  tumors	
  with	
  high-­‐frequency	
  microsatellite	
  instability.	
  
Cancer	
  genetics	
  and	
  cytogenetics.	
  2005;157:	
  18-­‐24.	
  
10.	
  Hampson	
  R.	
  Selection	
  for	
  genome	
  instability	
  by	
  DNA	
  damage	
  in	
  human	
  cells:	
  unstable	
  
microsatellites	
  and	
  their	
  consequences	
  for	
  tumourigenesis.	
  Radiation	
  oncology	
  
investigations.	
  1997;5:	
  111-­‐114.	
  
11.	
  Jascur	
  T,	
  Boland	
  CR.	
  Structure	
  and	
  function	
  of	
  the	
  components	
  of	
  the	
  human	
  DNA	
  
mismatch	
  repair	
  system.	
  International	
  journal	
  of	
  cancer.	
  Journal	
  international	
  du	
  cancer.	
  
2006;119:	
  2030-­‐2035.	
  
12.	
  Lynch	
  HT,	
  de	
  la	
  Chapelle	
  A.	
  Hereditary	
  colorectal	
  cancer.	
  The	
  New	
  England	
  journal	
  of	
  
medicine.	
  2003;348:	
  919-­‐932.	
  

	
  

138	
  

	
  
13.	
  Thibodeau	
  SN,	
  Bren	
  G,	
  Schaid	
  D.	
  Microsatellite	
  instability	
  in	
  cancer	
  of	
  the	
  proximal	
  
colon.	
  Science.	
  1993;260:	
  816-­‐819.	
  
14.	
  Hsu	
  HS,	
  Wen	
  CK,	
  Tang	
  YA,	
  et	
  al.	
  Promoter	
  hypermethylation	
  is	
  the	
  predominant	
  
mechanism	
  in	
  hMLH1	
  and	
  hMSH2	
  deregulation	
  and	
  is	
  a	
  poor	
  prognostic	
  factor	
  in	
  
nonsmoking	
  lung	
  cancer.	
  Clinical	
  cancer	
  research	
  :	
  an	
  official	
  journal	
  of	
  the	
  American	
  
Association	
  for	
  Cancer	
  Research.	
  2005;11:	
  5410-­‐5416.	
  
15.	
  Jacob	
  S,	
  Praz	
  F.	
  DNA	
  mismatch	
  repair	
  defects:	
  role	
  in	
  colorectal	
  carcinogenesis.	
  
Biochimie.	
  2002;84:	
  27-­‐47.	
  
16.	
  Kawakami	
  T,	
  Shiina	
  H,	
  Igawa	
  M,	
  et	
  al.	
  Inactivation	
  of	
  the	
  hMSH3	
  mismatch	
  repair	
  gene	
  
in	
  bladder	
  cancer.	
  Biochemical	
  and	
  biophysical	
  research	
  communications.	
  2004;325:	
  934-­‐
942.	
  
17.	
  Eckert	
  KA,	
  Hile	
  SE.	
  Every	
  microsatellite	
  is	
  different:	
  Intrinsic	
  DNA	
  features	
  dictate	
  
mutagenesis	
  of	
  common	
  microsatellites	
  present	
  in	
  the	
  human	
  genome.	
  Molecular	
  
carcinogenesis.	
  2009;48:	
  379-­‐388.	
  
18.	
  Umar	
  A,	
  Boland	
  CR,	
  Terdiman	
  JP,	
  et	
  al.	
  Revised	
  Bethesda	
  Guidelines	
  for	
  hereditary	
  
nonpolyposis	
  colorectal	
  cancer	
  (Lynch	
  syndrome)	
  and	
  microsatellite	
  instability.	
  Journal	
  of	
  
the	
  National	
  Cancer	
  Institute.	
  2004;96:	
  261-­‐268.	
  
19.	
  Dobson	
  JM.	
  Breed-­‐predispositions	
  to	
  cancer	
  in	
  pedigree	
  dogs.	
  ISRN	
  Vet	
  Sci.	
  2013;2013:	
  
941275.	
  
20.	
  Knapp	
  DW,	
  Ramos-­‐Vara	
  JA,	
  Moore	
  GE,	
  Dhawan	
  D,	
  Bonney	
  PL,	
  Young	
  KE.	
  Urinary	
  
bladder	
  cancer	
  in	
  dogs,	
  a	
  naturally	
  occurring	
  model	
  for	
  cancer	
  biology	
  and	
  drug	
  
development.	
  ILAR	
  J.	
  2014;55:	
  100-­‐118.	
  
21.	
  Seim-­‐Wikse	
  T,	
  Jorundsson	
  E,	
  Nodtvedt	
  A,	
  et	
  al.	
  Breed	
  predisposition	
  to	
  canine	
  gastric	
  
carcinoma-­‐-­‐a	
  study	
  based	
  on	
  the	
  Norwegian	
  canine	
  cancer	
  register.	
  Acta	
  Vet	
  Scand.	
  
2013;55:	
  25.	
  
22.	
  Guyon	
  R,	
  Kirkness	
  EF,	
  Lorentzen	
  TD,	
  et	
  al.	
  Building	
  comparative	
  maps	
  using	
  1.5x	
  
sequence	
  coverage:	
  human	
  chromosome	
  1p	
  and	
  the	
  canine	
  genome.	
  Cold	
  Spring	
  Harb	
  
Symp	
  Quant	
  Biol.	
  2003;68:	
  171-­‐177.	
  
23.	
  Oetting	
  WS,	
  Fryer	
  JP,	
  Oofuji	
  Y,	
  et	
  al.	
  Analysis	
  of	
  tyrosinase	
  gene	
  mutations	
  using	
  direct	
  
automated	
  infrared	
  fluorescence	
  DNA	
  sequencing	
  of	
  amplified	
  exons.	
  Electrophoresis.	
  
1994;15:	
  159-­‐164.	
  
24.	
  Knapp	
  DW,	
  Glickman	
  NW,	
  Denicola	
  DB,	
  Bonney	
  PL,	
  Lin	
  TL,	
  Glickman	
  LT.	
  Naturally-­‐
occurring	
  canine	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder	
  A	
  relevant	
  model	
  of	
  
human	
  invasive	
  bladder	
  cancer.	
  Urologic	
  oncology.	
  2000;5:	
  47-­‐59.	
  

	
  

139	
  

	
  
25.	
  Mutsaers	
  AJ,	
  Widmer	
  WR,	
  Knapp	
  DW.	
  Canine	
  transitional	
  cell	
  carcinoma.	
  Journal	
  of	
  
veterinary	
  internal	
  medicine	
  /	
  American	
  College	
  of	
  Veterinary	
  Internal	
  Medicine.	
  2003;17:	
  
136-­‐144.	
  
26.	
  Parker	
  HG,	
  Kim	
  LV,	
  Sutter	
  NB,	
  et	
  al.	
  Genetic	
  structure	
  of	
  the	
  purebred	
  domestic	
  dog.	
  
Science.	
  2004;304:	
  1160-­‐1164.	
  
27.	
  Patrick	
  DJ,	
  Fitzgerald	
  SD,	
  Sesterhenn	
  IA,	
  Davis	
  CJ,	
  Kiupel	
  M.	
  Classification	
  of	
  canine	
  
urinary	
  bladder	
  urothelial	
  tumours	
  based	
  on	
  the	
  World	
  Health	
  Organization/International	
  
Society	
  of	
  Urological	
  Pathology	
  consensus	
  classification.	
  Journal	
  of	
  comparative	
  pathology.	
  
2006;135:	
  190-­‐199.	
  
28.	
  Watson	
  P,	
  Vasen	
  HF,	
  Mecklin	
  JP,	
  et	
  al.	
  The	
  risk	
  of	
  extra-­‐colonic,	
  extra-­‐endometrial	
  
cancer	
  in	
  the	
  Lynch	
  syndrome.	
  Int	
  J	
  Cancer.	
  2008;123:	
  444-­‐449.	
  
29.	
  Bai	
  S,	
  Nunez	
  AL,	
  Wei	
  S,	
  et	
  al.	
  Microsatellite	
  instability	
  and	
  TARBP2	
  mutation	
  study	
  in	
  
upper	
  urinary	
  tract	
  urothelial	
  carcinoma.	
  Am	
  J	
  Clin	
  Pathol.	
  2013;139:	
  765-­‐770.	
  
30.	
  Blaszyk	
  H,	
  Wang	
  L,	
  Dietmaier	
  W,	
  et	
  al.	
  Upper	
  tract	
  urothelial	
  carcinoma:	
  a	
  
clinicopathologic	
  study	
  including	
  microsatellite	
  instability	
  analysis.	
  Mod	
  Pathol.	
  2002;15:	
  
790-­‐797.	
  
31.	
  Ehsani	
  L,	
  Osunkoya	
  AO.	
  Expression	
  of	
  MLH1	
  and	
  MSH2	
  in	
  urothelial	
  carcinoma	
  of	
  the	
  
renal	
  pelvis.	
  Tumour	
  Biol.	
  2014;35:	
  8743-­‐8747.	
  
32.	
  Eltz	
  S,	
  Comperat	
  E,	
  Cussenot	
  O,	
  Roupret	
  M.	
  Molecular	
  and	
  histological	
  markers	
  in	
  
urothelial	
  carcinomas	
  of	
  the	
  upper	
  urinary	
  tract.	
  BJU	
  Int.	
  2008;102:	
  532-­‐535.	
  
33.	
  Hartmann	
  A,	
  Zanardo	
  L,	
  Bocker-­‐Edmonston	
  T,	
  et	
  al.	
  Frequent	
  microsatellite	
  instability	
  
in	
  sporadic	
  tumors	
  of	
  the	
  upper	
  urinary	
  tract.	
  Cancer	
  Res.	
  2002;62:	
  6796-­‐6802.	
  
34.	
  Ho	
  CL,	
  Tzai	
  TS,	
  Chen	
  JC,	
  et	
  al.	
  The	
  molecular	
  signature	
  for	
  urothelial	
  carcinoma	
  of	
  the	
  
upper	
  urinary	
  tract.	
  J	
  Urol.	
  2008;179:	
  1155-­‐1159.	
  
35.	
  Mongiat-­‐Artus	
  P,	
  Miquel	
  C,	
  Van	
  der	
  Aa	
  M,	
  et	
  al.	
  Microsatellite	
  instability	
  and	
  mutation	
  
analysis	
  of	
  candidate	
  genes	
  in	
  urothelial	
  cell	
  carcinomas	
  of	
  upper	
  urinary	
  tract.	
  Oncogene.	
  
2006;25:	
  2113-­‐2118.	
  
36.	
  Roupret	
  M,	
  Azzouzi	
  AR,	
  Cussenot	
  O.	
  Microsatellite	
  instability	
  and	
  transitional	
  cell	
  
carcinoma	
  of	
  the	
  upper	
  urinary	
  tract.	
  BJU	
  Int.	
  2005;96:	
  489-­‐492.	
  
37.	
  Roupret	
  M,	
  Catto	
  J,	
  Coulet	
  F,	
  et	
  al.	
  Microsatellite	
  instability	
  as	
  indicator	
  of	
  MSH2	
  gene	
  
mutation	
  in	
  patients	
  with	
  upper	
  urinary	
  tract	
  transitional	
  cell	
  carcinoma.	
  J	
  Med	
  Genet.	
  
2004;41:	
  e91.	
  

	
  

140	
  

	
  
38.	
  Roupret	
  M,	
  Fromont	
  G,	
  Azzouzi	
  AR,	
  et	
  al.	
  Microsatellite	
  instability	
  as	
  predictor	
  of	
  
survival	
  in	
  patients	
  with	
  invasive	
  upper	
  urinary	
  tract	
  transitional	
  cell	
  carcinoma.	
  Urology.	
  
2005;65:	
  1233-­‐1237.	
  
39.	
  Roupret	
  M,	
  Hupertan	
  V,	
  Seisen	
  T,	
  et	
  al.	
  Prediction	
  of	
  cancer	
  specific	
  survival	
  after	
  
radical	
  nephroureterectomy	
  for	
  upper	
  tract	
  urothelial	
  carcinoma:	
  development	
  of	
  an	
  
optimized	
  postoperative	
  nomogram	
  using	
  decision	
  curve	
  analysis.	
  J	
  Urol.	
  2013;189:	
  1662-­‐
1669.	
  
40.	
  Catto	
  JW,	
  Meuth	
  M,	
  Hamdy	
  FC.	
  Genetic	
  instability	
  and	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  
bladder.	
  BJU	
  Int.	
  2004;93:	
  19-­‐24.	
  
41.	
  Skeldon	
  SC,	
  Semotiuk	
  K,	
  Aronson	
  M,	
  et	
  al.	
  Patients	
  with	
  Lynch	
  syndrome	
  mismatch	
  
repair	
  gene	
  mutations	
  are	
  at	
  higher	
  risk	
  for	
  not	
  only	
  upper	
  tract	
  urothelial	
  cancer	
  but	
  also	
  
bladder	
  cancer.	
  Eur	
  Urol.	
  2013;63:	
  379-­‐385.	
  
42.	
  Volanis	
  D,	
  Papadopoulos	
  G,	
  Doumas	
  K,	
  Gkialas	
  I,	
  Delakas	
  D.	
  Molecular	
  mechanisms	
  in	
  
urinary	
  bladder	
  carcinogenesis.	
  J	
  BUON.	
  2011;16:	
  589-­‐601.	
  
43.	
  Wadhwa	
  N,	
  Mathew	
  BB,	
  Jatawa	
  SK,	
  Tiwari	
  A.	
  Genetic	
  instability	
  in	
  urinary	
  bladder	
  
cancer:	
  An	
  evolving	
  hallmark.	
  J	
  Postgrad	
  Med.	
  2013;59:	
  284-­‐288.	
  
44.	
  Saetta	
  AA,	
  Goudopoulou	
  A,	
  Korkolopoulou	
  P,	
  et	
  al.	
  Mononucleotide	
  markers	
  of	
  
microsatellite	
  instability	
  in	
  carcinomas	
  of	
  the	
  urinary	
  bladder.	
  Eur	
  J	
  Surg	
  Oncol.	
  2004;30:	
  
796-­‐803.	
  
45.	
  Vaish	
  M,	
  Mandhani	
  A,	
  Mittal	
  RD,	
  Mittal	
  B.	
  Microsatellite	
  instability	
  as	
  prognostic	
  marker	
  
in	
  bladder	
  tumors:	
  a	
  clinical	
  significance.	
  BMC	
  Urol.	
  2005;5:	
  2.	
  
46.	
  van	
  der	
  Meijden	
  AP.	
  Bladder	
  cancer.	
  BMJ.	
  1998;317:	
  1366-­‐1369.	
  
47.	
  Amira	
  N,	
  Rivet	
  J,	
  Soliman	
  H,	
  et	
  al.	
  Microsatellite	
  instability	
  in	
  urothelial	
  carcinoma	
  of	
  the	
  
upper	
  urinary	
  tract.	
  J	
  Urol.	
  2003;170:	
  1151-­‐1154.	
  
48.	
  Catto	
  JW,	
  Azzouzi	
  AR,	
  Amira	
  N,	
  et	
  al.	
  Distinct	
  patterns	
  of	
  microsatellite	
  instability	
  are	
  
seen	
  in	
  tumours	
  of	
  the	
  urinary	
  tract.	
  Oncogene.	
  2003;22:	
  8699-­‐8706.	
  
49.	
  Watson	
  MM,	
  Berg	
  M,	
  Soreide	
  K.	
  Prevalence	
  and	
  implications	
  of	
  elevated	
  microsatellite	
  
alterations	
  at	
  selected	
  tetranucleotides	
  in	
  cancer.	
  Br	
  J	
  Cancer.	
  2014;111:	
  823-­‐827.	
  
50.	
  Catto	
  JW,	
  Xinarianos	
  G,	
  Burton	
  JL,	
  Meuth	
  M,	
  Hamdy	
  FC.	
  Differential	
  expression	
  of	
  
hMLH1	
  and	
  hMSH2	
  is	
  related	
  to	
  bladder	
  cancer	
  grade,	
  stage	
  and	
  prognosis	
  but	
  not	
  
microsatellite	
  instability.	
  Int	
  J	
  Cancer.	
  2003;105:	
  484-­‐490.	
  
51.	
  Kassem	
  HS,	
  Varley	
  JM,	
  Hamam	
  SM,	
  Margison	
  GP.	
  Immunohistochemical	
  analysis	
  of	
  
expression	
  and	
  allelotype	
  of	
  mismatch	
  repair	
  genes	
  (hMLH1	
  and	
  hMSH2)	
  in	
  bladder	
  
cancer.	
  Br	
  J	
  Cancer.	
  2001;84:	
  321-­‐328.	
  

	
  

141	
  

	
  
52.	
  Kawakami	
  T,	
  Shiina	
  H,	
  Igawa	
  M,	
  et	
  al.	
  Inactivation	
  of	
  the	
  hMSH3	
  mismatch	
  repair	
  gene	
  
in	
  bladder	
  cancer.	
  Biochem	
  Biophys	
  Res	
  Commun.	
  2004;325:	
  934-­‐942.	
  
53.	
  Mylona	
  E,	
  Zarogiannos	
  A,	
  Nomikos	
  A,	
  et	
  al.	
  Prognostic	
  value	
  of	
  microsatellite	
  instability	
  
determined	
  by	
  immunohistochemical	
  staining	
  of	
  hMSH2	
  and	
  hMSH6	
  in	
  urothelial	
  
carcinoma	
  of	
  the	
  bladder.	
  APMIS.	
  2008;116:	
  59-­‐65.	
  
54.	
  Rubio	
  J,	
  Blanes	
  A,	
  Sanchez-­‐Carrillo	
  JJ,	
  Diaz-­‐Cano	
  SJ.	
  Microsatellite	
  abnormalities	
  and	
  
somatic	
  down-­‐regulation	
  of	
  mismatch	
  repair	
  characterize	
  nodular-­‐trabecular	
  muscle-­‐
invasive	
  urothelial	
  carcinoma	
  of	
  the	
  bladder.	
  Histopathology.	
  2007;51:	
  458-­‐467.	
  
55.	
  Yamamoto	
  Y,	
  Matsuyama	
  H,	
  Kawauchi	
  S,	
  et	
  al.	
  Biological	
  characteristics	
  in	
  bladder	
  
cancer	
  depend	
  on	
  the	
  type	
  of	
  genetic	
  instability.	
  Clin	
  Cancer	
  Res.	
  2006;12:	
  2752-­‐2758.	
  
56.	
  Dietmaier	
  W,	
  Wallinger	
  S,	
  Bocker	
  T,	
  Kullmann	
  F,	
  Fishel	
  R,	
  Ruschoff	
  J.	
  Diagnostic	
  
microsatellite	
  instability:	
  definition	
  and	
  correlation	
  with	
  mismatch	
  repair	
  protein	
  
expression.	
  Cancer	
  Res.	
  1997;57:	
  4749-­‐4756.	
  
57.	
  Shia	
  J.	
  Immunohistochemistry	
  versus	
  microsatellite	
  instability	
  testing	
  for	
  screening	
  
colorectal	
  cancer	
  patients	
  at	
  risk	
  for	
  hereditary	
  nonpolyposis	
  colorectal	
  cancer	
  syndrome.	
  
Part	
  I.	
  The	
  utility	
  of	
  immunohistochemistry.	
  J	
  Mol	
  Diagn.	
  2008;10:	
  293-­‐300.	
  
58.	
  Bonnal	
  C,	
  Ravery	
  V,	
  Toublanc	
  M,	
  et	
  al.	
  Absence	
  of	
  microsatellite	
  instability	
  in	
  
transitional	
  cell	
  carcinoma	
  of	
  the	
  bladder.	
  Urology.	
  2000;55:	
  287-­‐291.	
  
59.	
  Zhang	
  L.	
  Immunohistochemistry	
  versus	
  microsatellite	
  instability	
  testing	
  for	
  screening	
  
colorectal	
  cancer	
  patients	
  at	
  risk	
  for	
  hereditary	
  nonpolyposis	
  colorectal	
  cancer	
  syndrome.	
  
Part	
  II.	
  The	
  utility	
  of	
  microsatellite	
  instability	
  testing.	
  J	
  Mol	
  Diagn.	
  2008;10:	
  301-­‐307.	
  
60.	
  Mylona	
  E,	
  Zarogiannos	
  A,	
  Nomikos	
  A,	
  et	
  al.	
  Prognostic	
  value	
  of	
  microsatellite	
  instability	
  
determined	
  by	
  immunohistochemical	
  staining	
  of	
  hMSH2	
  and	
  hMSH6	
  in	
  urothelial	
  
carcinoma	
  of	
  the	
  bladder.	
  Acta	
  pathologica,	
  microbiologica	
  et	
  immunologica	
  Scandinavica.	
  
2008;116:	
  59-­‐65.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

142	
  

	
  
CHAPTER	
  5	
  
	
  
	
  
Evaluation	
  of	
  DNA	
  mismatch	
  repair	
  in	
  novel	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  
carcinoma	
  cell	
  lines	
  
	
  
Dodd	
  Sledge1,	
  Elizabeth	
  A.	
  McNiel2,	
  Nicole	
  J.	
  Madrill3,	
  James	
  H.	
  Resau4,	
  Matti	
  Kiupel1	
  
	
  
1	
  Department	
  of	
  Pathobiology	
  and	
  Diagnostic	
  Investigation,	
  Diagnostic	
  Center	
  for	
  
Population	
  and	
  Animal	
  Health,	
  Michigan	
  State	
  University,	
  Lansing	
  MI	
  
	
  University,	
  Lansing,	
  MI;	
  	
  
2	
  Tufts	
  Cummings	
  School	
  of	
  Veterinary	
  Medicine	
  and	
  Molecular	
  Oncology	
  Research	
  
Institute,	
  Boston,	
  MA,	
  
3	
  College	
  of	
  Veterinary	
  Medicine,	
  Michigan	
  State	
  University,	
  East	
  Lansing,	
  MI	
  
4	
  Van	
  Andel	
  Institute,	
  Grand	
  Rapids,	
  Lansing,	
  MI	
  
	
  

	
  

	
  

143	
  

	
  
Abstract	
  
The	
  DNA	
  mismatch	
  repair	
  system	
  plays	
  important	
  roles	
  in	
  DNA	
  post	
  replication	
  repair,	
  
damage	
  recognition	
  signaling,	
  apoptosis,	
  and	
  meiotic	
  recombination.	
  Consequently,	
  defects	
  
in	
  MMR	
  are	
  associated	
  with	
  carcinogenesis	
  in	
  a	
  variety	
  of	
  tissues	
  and	
  have	
  been	
  shown	
  to	
  
affect	
  response	
  of	
  cancers	
  to	
  chemotherapy.	
  Recently,	
  we	
  have	
  shown	
  that	
  many	
  canine	
  
lower	
  urinary	
  tract	
  urothelial	
  carcinomas	
  have	
  deficiencies	
  in	
  MMR.	
  In	
  order	
  to	
  further	
  
investigate	
  the	
  role	
  of	
  MMR	
  in	
  carcinogenesis	
  and	
  the	
  significance	
  of	
  MMR	
  deficiency	
  in	
  the	
  
dog,	
  4	
  novel	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines	
  from	
  3	
  dogs	
  established	
  
from	
  primary	
  cultured	
  cells.	
  Characterization	
  of	
  all	
  cell	
  lines	
  including	
  immunophenotyping	
  
for	
  markers	
  of	
  urothelial	
  differentiation	
  was	
  consistent	
  with	
  an	
  urothelial	
  origin.	
  MMR	
  gene	
  
expression	
  in	
  each	
  cell	
  line	
  was	
  evaluated	
  using	
  qPCR	
  and	
  protein	
  expression	
  was	
  
evaluated	
  using	
  Western	
  blots.	
  Further,	
  MMR	
  gene	
  expression	
  was	
  evaluated	
  in	
  mice	
  
xenografts	
  derived	
  from	
  the	
  canine	
  cell	
  lines	
  using	
  qPCR	
  on	
  laser	
  capture	
  microdissection	
  
samples	
  of	
  tumor	
  and	
  MMR	
  protein	
  expression	
  was	
  evaluated	
  using	
  
immunohistochemistry.	
  	
  One	
  of	
  the	
  four	
  cell	
  lines,	
  TYLER2,	
  had	
  decreased	
  gene	
  and	
  protein	
  
expression	
  of	
  MSH2	
  and	
  MSH6,	
  which	
  form	
  portions	
  of	
  the	
  MMR	
  pathway.	
  	
  XTT	
  survival	
  
assays	
  were	
  subsequently	
  used	
  to	
  evaluate	
  differences	
  between	
  the	
  cell	
  lines	
  in	
  sensitivity	
  
to	
  a	
  panel	
  of	
  chemotherapeutics.	
  	
  In	
  comparison	
  to	
  other	
  cell	
  lines	
  including	
  one	
  derived	
  
from	
  the	
  same	
  primary	
  tumor,	
  TYLER2	
  had	
  relative	
  resistance	
  to	
  carboplatin	
  and	
  
sensitivity	
  to	
  oxaliplatin,	
  thiotepa,	
  and	
  methotrexate.	
  
	
  

	
  

	
  

144	
  

	
  
Introduction	
  
Urothelial	
  carcinomas	
  of	
  the	
  urinary	
  bladder	
  comprise	
  a	
  significant	
  proportion	
  of	
  
canine	
  neoplasms	
  and	
  have	
  historically	
  been	
  associated	
  with	
  a	
  poor	
  prognosis.	
  Because	
  
complete	
  surgical	
  resection	
  is	
  rarely	
  achievable	
  and	
  30-­‐50%	
  of	
  these	
  cancers	
  metastasize,	
  
most	
  animals	
  die	
  of	
  their	
  disease.1,	
  2	
  Although	
  frequently	
  treated	
  with	
  chemotherapy,	
  these	
  
cancers	
  respond	
  inconsistently;	
  however,	
  the	
  choice	
  of	
  chemotherapeutic	
  for	
  urothelial	
  
carcinomas	
  has	
  traditionally	
  been	
  empirical	
  rather	
  than	
  based	
  on	
  scientific	
  evidence.2,	
  3	
  	
  
It	
  is	
  clear	
  that	
  the	
  molecular	
  constitution	
  of	
  cancer	
  in	
  one	
  individual	
  can	
  differ	
  
considerably	
  from	
  that	
  of	
  cancers	
  in	
  others	
  despite	
  the	
  fact	
  that	
  the	
  cancers	
  may	
  have	
  a	
  
similar	
  morphology	
  and	
  anatomic	
  localization.	
  In	
  other	
  words,	
  cancers	
  even	
  of	
  the	
  same	
  
histologic	
  origin	
  can	
  be	
  molecularly	
  unique.	
  Thus,	
  in	
  therapeutic	
  terms,	
  a	
  standardized	
  
treatment	
  prescription	
  fails	
  to	
  address	
  this	
  diversity.	
  In	
  recent	
  years,	
  there	
  has	
  been	
  
considerable	
  interest	
  in	
  determining	
  ways	
  to	
  exploit	
  the	
  particular	
  molecular	
  constitution	
  
of	
  a	
  given	
  tumor	
  to	
  improve	
  therapeutic	
  outcomes.	
  One	
  feature	
  that	
  differs	
  between	
  tumors	
  
that	
  may	
  be	
  exploited	
  through	
  targeted	
  therapy	
  is	
  DNA	
  repair	
  capability,	
  including	
  the	
  
functionality	
  of	
  their	
  DNA	
  mismatch	
  repair	
  system	
  (MMR).	
  	
  
	
  

The	
  MMR	
  machinery	
  participates	
  in	
  a	
  variety	
  of	
  cellular	
  processes.	
  Most	
  notably,	
  this	
  

system	
  is	
  responsible	
  for	
  post	
  replication	
  repair	
  of	
  base-­‐base	
  mismatches	
  and	
  the	
  
resolution	
  of	
  insertion	
  and	
  deletion	
  loops	
  that	
  can	
  occur	
  in	
  repetitive	
  regions	
  of	
  DNA.4	
  The	
  
MMR	
  system	
  also	
  participates	
  in	
  a	
  variety	
  of	
  other	
  processes	
  including	
  DNA	
  damage	
  
recognition	
  signaling,	
  apoptosis,	
  meiotic	
  recombination,	
  and	
  other	
  DNA	
  repair	
  pathways.4	
  
Given	
  these	
  varied	
  roles	
  played	
  by	
  MMR,	
  it	
  is	
  not	
  surprising	
  that	
  defects	
  in	
  MMR	
  facilitate	
  
carcinogenesis.	
  Similarly,	
  it	
  is	
  not	
  surprising	
  that	
  defects	
  in	
  MMR	
  have	
  significant	
  effects	
  on	
  

	
  

145	
  

	
  
response	
  of	
  cancers	
  to	
  therapy.	
  The	
  reported	
  effects	
  of	
  MMR	
  deficiency	
  on	
  
chemotherapeutic	
  response	
  are	
  complex.	
  	
  Depending	
  on	
  the	
  mechanism	
  of	
  action	
  for	
  a	
  
given	
  chemotherapeutic,	
  MMR	
  deficiency	
  may	
  confer	
  either	
  drug	
  resistance	
  or	
  sensitivity.	
  	
  
	
  

Microsatellite	
  Instability	
  (MSI),	
  which	
  is	
  defined	
  as	
  the	
  accumulation	
  of	
  frame	
  shift	
  

mutations	
  in	
  nucleotide	
  repeat	
  regions	
  of	
  DNA	
  (called	
  microsatellites)	
  is	
  considered	
  a	
  
“signature”	
  for	
  MMR	
  dysfunction.4,	
  5	
  In	
  previous	
  investigations,	
  we	
  demonstrated	
  that	
  MSI	
  
is	
  frequent	
  in	
  canine	
  urothelial	
  carcinomas	
  of	
  the	
  urinary	
  bladder	
  and	
  has	
  strong	
  
associations	
  with	
  breed	
  and	
  phylogenetic	
  clade	
  (data	
  presented	
  in	
  Chapter	
  3).	
  	
  In	
  addition,	
  
we	
  showed	
  that	
  expression	
  levels	
  of	
  MSH2	
  and	
  MSH6	
  as	
  evaluated	
  by	
  
immunohistochemistry	
  differed	
  in	
  canine	
  urothelial	
  carcinoma	
  from	
  that	
  observed	
  in	
  
normal	
  canine	
  urotheluim	
  (data	
  presented	
  in	
  Chapter	
  3).	
  	
  Such	
  frequent	
  MSI	
  and	
  
alterations	
  MMR	
  protein	
  expression	
  suggests	
  MMR	
  dysfunction	
  is	
  common	
  in	
  canine	
  
urothelial	
  carcinomas	
  and	
  may	
  be	
  heritable.	
  The	
  exact	
  cause	
  and	
  significance	
  of	
  MMR	
  
dysfunction	
  in	
  canine	
  urothelial	
  carcinomas	
  remains	
  unclear.	
  	
  However,	
  based	
  on	
  reports	
  
of	
  differential	
  response	
  of	
  MMR	
  proficient	
  and	
  MMR	
  deficient	
  cancers	
  to	
  treatment	
  in	
  
humans,	
  it	
  is	
  likely	
  that	
  MMR	
  status	
  could	
  also	
  affect	
  treatment	
  response	
  in	
  dogs	
  with	
  lower	
  
urinary	
  tract	
  urothelial	
  carcinomas.	
  	
  	
  
	
  

For	
  the	
  current	
  study,	
  we	
  aimed	
  to	
  establish	
  in	
  vitro	
  and	
  xenograft	
  models	
  for	
  further	
  

study	
  of	
  MMR	
  dysfunction	
  in	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas.	
  	
  More	
  
specifically,	
  we	
  aimed	
  to	
  evaluate	
  the	
  expression	
  of	
  components	
  of	
  the	
  MMR	
  pathway	
  at	
  the	
  
gene	
  and	
  protein	
  level	
  in	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines	
  and	
  xenografts,	
  
and	
  to	
  correlate	
  differences	
  in	
  such	
  expression	
  to	
  treatment	
  response.	
  
	
  

	
  

	
  
146	
  

	
  
Materials	
  and	
  methods	
  
	
  

All	
  work	
  done	
  with	
  canine	
  tissues	
  or	
  with	
  xenografts	
  was	
  preformed	
  with	
  approval	
  

from	
  the	
  Michigan	
  State	
  University	
  Institutional	
  Animal	
  Care	
  and	
  Use	
  Committee.	
  
Characterization	
  of	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines	
  
With	
   owner	
   consent	
   for	
   inclusion	
   into	
   the	
   study,	
   four	
   canine	
   cells	
   lines	
   (ANGUS,	
  
KINSEY,	
   TYLER1	
   and	
   TYLER2)	
   were	
   generated	
   from	
   primary	
   cell	
   culture	
   of	
   urothelial	
  
carcinomas	
   of	
   the	
   lower	
   urinary	
   tract.	
   	
   ANGUS	
   was	
   derived	
   from	
   a	
   papillary	
   grade	
   III	
  
urothelial	
   carcinoma	
   of	
   the	
   urinary	
   bladder	
   with	
   invasion	
   into	
   the	
   muscularis	
   of	
   a	
   10-­‐year-­‐
old	
   neutered	
   male	
   Scottish	
   terrier,	
   KINSEY	
   was	
   derived	
   from	
   an	
   infiltrating	
   urothelial	
  
carcinoma	
  of	
  the	
  urinary	
  bladder	
  with	
  invasion	
  into	
  the	
  substantia	
  propria	
  of	
  a	
  12-­‐year-­‐old	
  
spayed	
   female	
   Australian	
   Shepherd,	
   and	
   TYLER	
   cell	
   lines	
   were	
   derived	
   from	
   a	
   single	
  
papillary	
   grade	
   II	
   urothelial	
   carcinoma	
   with	
   invasion	
   into	
   the	
   muscularis	
   of	
   the	
   prostatic	
  
urethra	
  of	
  an	
  8-­‐year-­‐old	
  neutered	
  male	
  Beagle.	
  All	
  animals	
  were	
  referred	
  to	
  the	
  Michigan	
  
State	
   University,	
   Veterinary	
   Medical	
   Center,	
   East	
   Lansing,	
   MI	
   for	
   treatment	
   of	
   primary	
  
bladder	
   disease.	
   Diagnoses	
   of	
   urothelial	
   carcinomas	
   were	
   made	
   based	
   on	
   results	
   of	
  
histopathologic	
   examination	
   and	
   evaluation	
   of	
   immunohistochemical	
   markers	
   at	
   the	
  
Diagnostic	
  Center	
  for	
  Population	
  and	
  Animal	
  Health,	
  Lansing,	
  MI.	
  	
  
For	
  ANGUS	
  and	
  TYLER	
  cell	
  lines,	
  neoplastic	
  cells	
  were	
  initially	
  isolated	
  from	
  macerated	
  
fresh	
   samples	
   of	
   each	
   tumor	
   taken	
   from	
   routine	
   diagnostic	
   biopsy.	
   For	
   KINSEY,	
   cell	
   lines	
  
were	
   derived	
   from	
   cell	
   pellets	
   derived	
   from	
   centrifugation	
   of	
   7mL	
   of	
   urine	
   taken	
   by	
  
cystocentesis.	
  	
  The	
  resulting	
  cell	
  pellet	
  was	
  suspended	
  in	
  media	
  and	
  plated	
  on	
  cell	
  culture	
  
plates.	
  	
  For	
  all	
  cell	
  lines,	
  neoplastic	
  cells	
  were	
  grown	
  on	
  uncoated	
  cell	
  culture	
  plates	
  in	
  1:1	
  
DMEM/F12	
  media	
  (Life	
  Technologies,	
  Grand	
  Island,	
  NY,	
  USA)	
  supplemented	
  with	
  10%	
  fetal	
  

	
  

147	
  

	
  
bovine	
  serum,	
  l-­‐glutamine,	
  and	
  penicillin/streptomycin	
  in	
  a	
  37°C	
  humidified	
  CO2	
  incubator.	
  
For	
  subculturing,	
  cells	
  were	
  harvested	
  from	
  culture	
  plates	
  using	
  0.05%	
  or	
  0.25%	
  Trypsin-­‐
EDTA	
  (Life	
  Technologies).	
  
Over	
   multiple	
   passages,	
   two	
   unique	
   cellular	
   morphologies	
   were	
   identified	
   in	
   the	
  
original	
   TYLER	
   cell	
   line,	
   which	
   were	
   separated	
   by	
   selective	
   trypsinization	
   over	
   several	
  
passages.	
   For	
   selective	
   trypsinization,	
   mixed	
   cell	
   populations	
   were	
   incubated	
   with	
   0.05%	
  
Trypsin-­‐EDTA	
   at	
   37°C	
   for	
   1-­‐3	
   minutes.	
   	
   Cells	
   easily	
   released	
   from	
   attachments	
   on	
   cell	
  
plates	
   by	
   such	
   light	
   trypsinization	
   were	
   removed	
   by	
   light	
   rinsing	
   of	
   plates	
   with	
   culture	
  
media.	
   	
   Easily	
   trypsinized	
   cells	
   were	
   subsequently	
   subcultured	
   separately	
   from	
   more	
  
adherent	
  cells	
  resulting	
  in	
  establishment	
  of	
  two	
  morphologically	
  distinct	
  cell	
  lines,	
  TYLER1	
  
and	
  TYLER2.	
  
In	
   addition	
   subclones	
   of	
   TYLER1	
   and	
   TYLER2	
   were	
   established	
   through	
   serial	
   dilution.	
  	
  
Single	
   colonies	
   derived	
   from	
   dilution	
   of	
   cell	
   suspensions	
   plated	
   on	
   96	
   well	
   plates	
   were	
  
identified	
   at	
   day	
   3	
   post	
   plating.	
   	
   Subclones	
   were	
   established	
   from	
   these	
   single	
   cell	
   colonies	
  
through	
  subsequent	
  expansion.	
  
Development	
  of	
  xenografts	
  
	
  

Xenografts	
  were	
  established	
  in	
  athymic	
  nude	
  mice	
  by	
  subcutaneous	
  injection	
  of	
  106	
  

cells	
  in	
  200ul	
  of	
  sterile	
  media	
  lacking	
  FBS	
  and	
  antibiotics	
  over	
  the	
  flank.	
  	
  Tumor	
  volume	
  
was	
  evaluated	
  twice	
  per	
  week	
  and	
  animals	
  were	
  euthanized	
  using	
  CO2	
  when	
  tumor	
  volume	
  
was	
  >0.5cm3	
  or	
  when	
  there	
  grossly	
  obvious	
  ulceration	
  of	
  the	
  skin	
  over	
  tumors.	
  Fresh	
  
samples	
  of	
  xenograft	
  tumors	
  were	
  flash	
  frozen	
  in	
  liquid	
  nitrogen	
  and	
  saved	
  at	
  -­‐80°C.	
  	
  	
  
Additional	
  fresh	
  tissue	
  samples	
  were	
  frozen	
  over	
  dry	
  ice	
  in	
  optimum	
  cutting	
  temperature	
  
(OCT)	
  media	
  and	
  samples	
  were	
  stored	
  at	
  -­‐80°C.	
  	
  Remaining	
  tumor	
  samples	
  were	
  fixed	
  for	
  

	
  

148	
  

	
  
24-­‐48	
  hours	
  in	
  10%	
  buffered	
  formalin	
  and	
  routinely	
  processed	
  for	
  histopathologic	
  
examination.	
  	
  
Immunophenotyping	
  of	
  primary	
  tumors,	
  cell	
  lines,	
  and	
  xenografts	
  
	
  

For	
  each	
  of	
  the	
  four	
  cell	
  lines,	
  cells	
  were	
  removed	
  from	
  75%	
  confluent	
  cell	
  culture	
  

plates	
  with	
  a	
  cell	
  scraper	
  into	
  culture	
  media.	
  	
  After	
  pelleting	
  of	
  cells	
  under	
  centrifugation,	
  
cells	
  were	
  suspended	
  in	
  3ml	
  of	
  10%	
  buffered	
  formalin.	
  	
  Following	
  6	
  hours	
  of	
  fixation,	
  cells	
  
were	
  pelleted	
  under	
  centrifugation,	
  and	
  the	
  resulting	
  cell	
  pellet	
  was	
  routinely	
  processed	
  
into	
  paraffin	
  blocks.	
  	
  
Markers	
  of	
  differentiation	
  were	
  evaluated	
  in	
  paraffin-­‐embedded	
  primary	
  tumors,	
  
cell	
  lines,	
  and	
  xenograft	
  samples	
  using	
  immunohistochemistry	
  for	
  MNF116,	
  E-­‐cadherin,	
  P-­‐
cadherin,	
  N-­‐cadherin,	
  Cytokeratin	
  7,	
  uroplakin	
  III,	
  Prostatic	
  acid	
  phosphatase	
  (PAP),	
  and	
  
vimentin.	
  Previously	
  described	
  protocols	
  were	
  used	
  for	
  IHC,	
  with	
  the	
  exception	
  that	
  for	
  
mouse	
  derived	
  primary	
  antibodies,	
  blocking	
  of	
  nonspecific	
  binding	
  to	
  the	
  mouse	
  tissues	
  
was	
  accomplished	
  by	
  preincubating	
  primary	
  antibody	
  with	
  secondary	
  antibody	
  and	
  mouse	
  
serum	
  prior	
  to	
  use	
  (see	
  chapters	
  1	
  and	
  2).	
  
Evaluation	
  of	
  MMR	
  protein	
  expression	
  by	
  Western	
  blots	
  
Western	
  blotting	
  was	
  used	
  to	
  evaluate	
  expression	
  of	
  MLH1,	
  MSH2,	
  MSH3,	
  and	
  MSH6	
  
in	
  a	
  cell	
  lines	
  along	
  with	
  a	
  variety	
  of	
  normal	
  canine	
  tissues.	
  Cell	
  lines	
  were	
  harvested	
  at	
  
75%	
  confluence	
  using	
  a	
  cell	
  scraper.	
  Protein	
  was	
  subsequently	
  isolated	
  through	
  incubation	
  
with	
  RIPA	
  buffer.	
  	
  Total	
  protein	
  was	
  quantitated	
  using	
  Pierce™	
  BCA	
  Protein	
  Assay	
  Kit	
  
(Thermo	
  Scientific,	
  Rockford,	
  IL,	
  USA),	
  a	
  Victor	
  X3	
  microplate	
  reader	
  (PerkinElmer,	
  Shelton,	
  
CT	
  USA)	
  reading	
  absorbance	
  at	
  562nm,	
  and	
  comparison	
  to	
  a	
  dilution	
  series	
  of	
  bovine	
  
serum	
  albumin.	
  	
  Proteins	
  were	
  separated	
  using	
  SDS-­‐PAGE.	
  	
  For	
  MLH1	
  and	
  MSH2,	
  proteins	
  

	
  

149	
  

	
  
from	
  each	
  sample	
  were	
  separated	
  using	
  10%	
  gels	
  (BioRad,	
  Hercules,	
  CA,	
  USA).	
  	
  For	
  MSH3	
  
and	
  MSH6,	
  proteins	
  from	
  each	
  sample	
  were	
  separated	
  using	
  7%	
  gels	
  (BioRad).	
  Precision	
  
Plus	
  Protein™,	
  Dual	
  Color	
  (BioRad)	
  was	
  used	
  for	
  protein	
  standards.	
  	
  Proteins	
  were	
  
transferred	
  to	
  nitrocellulose	
  by	
  electroblotting.	
  	
  Following	
  washes	
  in	
  TBST,	
  blots	
  were	
  
blocked	
  using	
  3%	
  powdered	
  skim	
  milk	
  and	
  incubated	
  overnight	
  with	
  primary	
  antibodies	
  in	
  
3%	
  powdered	
  skim	
  milk	
  as	
  follows:	
  1:200	
  mouse	
  monoclonal	
  anti-­‐MLH1	
  antibody	
  (BD	
  
Biosciences,	
  San	
  Jose,	
  CA,	
  USA),	
  1:1500	
  rabbit	
  polyclonal	
  anti-­‐MSH2	
  antibody	
  (Santa	
  Cruz	
  
Biotechnology,	
  Dallas,	
  TX,	
  USA),	
  1:3000	
  rabbit	
  polyclonal	
  anti-­‐MSH3	
  antibody	
  (1:100,	
  
Abcam,	
  Cambridge,	
  MA,	
  USA),	
  1:500	
  mouse	
  monoclonal	
  anti-­‐MSH6	
  antibody	
  (1:500,	
  BD	
  
Biosciences).	
  Following	
  washes	
  in	
  TBST,	
  blots	
  were	
  accordingly	
  incubated	
  for	
  2	
  hours	
  with	
  
either	
  1:7500	
  goat	
  anti-­‐mouse	
  IgG	
  or	
  1:10,000	
  goat	
  anti-­‐rabbit	
  IgG	
  secondary	
  antibody	
  
(Santa	
  Cruz	
  Biotechnology)	
  as	
  appropriate.	
  Blots	
  were	
  developed	
  using	
  a	
  
chemiluminescence	
  detection	
  system	
  (Thermo	
  Scientific)	
  and	
  using	
  Amersham	
  Hyperfilm™	
  
MP	
  autoradiography	
  film	
  (GE	
  Healthcare,	
  Little	
  Chalfont,	
  BM,	
  UK).	
  	
  Blots	
  were	
  subsequently	
  
stripped	
  for	
  1	
  hour	
  using	
  Western	
  blot	
  stripping	
  buffer	
  (Thermo	
  Scientific).	
  	
  Following	
  
washes	
  in	
  TBST	
  and	
  blocking	
  in	
  3%	
  bovine	
  serum	
  albumin	
  (BSA),	
  blots	
  were	
  incubated	
  
overnight	
  with	
  1:2000	
  goat	
  anti-­‐β-­‐actin	
  primary	
  antibody	
  in	
  3%	
  BSA.	
  	
  	
  Following	
  washes	
  in	
  
TBST,	
  blots	
  were	
  incubated	
  for	
  2	
  hours	
  with	
  donkey	
  1:7500	
  anti-­‐goat	
  secondary	
  antibody	
  
in	
  3%	
  BSA	
  (Santa	
  Cruz	
  Biotechnology,	
  Dallas,	
  TX,	
  USA),	
  developed	
  as	
  described	
  above.	
  	
  
qPCR	
  for	
  MMR	
  genes	
  
Cell	
  lines	
  were	
  harvested	
  at	
  75%	
  confluence	
  using	
  .05%	
  or	
  .25%	
  Trypsin-­‐EDTA.	
  Cell	
  
lysis,	
  degradation	
  of	
  genomic	
  DNA,	
  and	
  reverse	
  transcription	
  of	
  RNA	
  to	
  cDNA	
  was	
  

	
  

150	
  

	
  
accomplished	
  using	
  TaqMan®	
  Gene	
  Expression	
  Cells-­‐to-­‐Ct	
  kits	
  (Life	
  Technologies,	
  Grand	
  
Island,	
  NY,	
  USA)	
  using	
  manufacturer’s	
  instructions	
  and	
  104	
  cells.	
  	
  	
  
	
  

For	
  xenografts,	
  laser	
  capture	
  microdissection	
  (LCM)	
  was	
  used	
  to	
  isolate	
  neoplastic	
  

canine	
  urothelial	
  carcinoma	
  cells	
  from	
  mouse	
  tissues.	
  In	
  TYLER1	
  and	
  TYLER2	
  xenografts,	
  
epithelioid	
  portions	
  of	
  tumors	
  were	
  harvested	
  separately	
  from	
  portions	
  with	
  more	
  discrete	
  
cell	
  morphology.	
  	
  Briefly,	
  serial	
  sections	
  of	
  OCT	
  embedded	
  frozen	
  xenografts	
  blocks	
  were	
  
made	
  with	
  a	
  cryostat	
  microtome	
  using	
  RNase-­‐free	
  techniques.	
  	
  Following	
  brief	
  fixation	
  in	
  
methanol,	
  staining	
  with	
  hematoxylin,	
  and	
  dehydration	
  in	
  alcohol,	
  tissue	
  sections	
  were	
  kept	
  
in	
  xylene	
  to	
  maintain	
  dehydration	
  until	
  LCM	
  was	
  performed.	
  	
  Laser	
  capture	
  
microdissection	
  (LCM)	
  using	
  an	
  Arcturus	
  PixCell	
  IIe	
  (Arcturus	
  Engineering,	
  Mountain	
  View,	
  
CA,	
  USA)	
  with	
  targeted	
  regions	
  of	
  tumors	
  being	
  captured	
  with	
  CapSure®	
  Macro	
  LCM	
  caps	
  
(Life	
  Technologies).	
  	
  RNA	
  was	
  isolated	
  from	
  captured	
  regions	
  using	
  an	
  Arcturus	
  Paradise	
  
Plus	
  RNA	
  Extraction	
  and	
  Isolation	
  Kit	
  (Life	
  Technologies)	
  following	
  the	
  manufacture’s	
  
instructions.	
  	
  Reverse	
  transcription	
  was	
  accomplished	
  using	
  High-­‐Capacity	
  cDNA	
  Reverse	
  
Transcription	
  Kits	
  (Life	
  Technologies)	
  using	
  manufacturer’s	
  instructions.	
  
Quantitative	
  real-­‐time	
  PCR	
  was	
  performed	
  using	
  a	
  StepOnePlus™	
  Real-­‐Time	
  PCR	
  
System	
  (Applied	
  Biosystems,	
  Grand	
  Island,	
  NY,	
  USA)	
  using	
  commercially	
  available	
  
TaqMan®	
  gene	
  expression	
  assays	
  for	
  MLH1,	
  MSH2,	
  MSH3,	
  MSH6,	
  PMS2,	
  β-­‐actin,	
  and	
  
HPRT1,	
  and	
  a	
  custom	
  designed	
  TaqMan®	
  gene	
  expression	
  assay	
  for	
  GAPDH	
  (Applied	
  
Biosystems)	
  (Table	
  14).	
  Studies	
  were	
  performed	
  with	
  technical	
  triplicates	
  and	
  were	
  
repeated	
  twice	
  with	
  independent	
  RNA	
  isolation	
  and	
  generation	
  of	
  cDNA	
  between	
  studies.	
  
The	
  baseline	
  and	
  threshold	
  for	
  detection	
  was	
  set	
  by	
  the	
  StepOnePlus	
  software,	
  with	
  
threshold	
  Ct	
  being	
  defined	
  as	
  the	
  number	
  of	
  cycles	
  at	
  which	
  the	
  detection	
  of	
  fluorescent	
  

	
  

151	
  

	
  
signal	
  exceeded	
  the	
  automatically	
  set	
  threshold.	
  For	
  cell	
  lines,	
  relative	
  quantitation	
  of	
  
target	
  gene	
  expression	
  was	
  evaluated	
  using	
  the	
  ∆Ct	
  method	
  using	
  comparison	
  of	
  a	
  
compilation	
  of	
  control	
  genes	
  including	
  GAPDH,	
  β-­‐actin,	
  and	
  HPRT1.	
  For	
  xenografts,	
  only	
  
expression	
  of	
  MSH2	
  and	
  MSH6	
  genes	
  was	
  evaluated,	
  and	
  analysis	
  of	
  relative	
  expression	
  of	
  
target	
  genes	
  was	
  made	
  to	
  only	
  GAPDH	
  expression.	
  Relative	
  quantitation	
  of	
  gene	
  expression	
  
between	
  different	
  cell	
  lines	
  or	
  xenografts	
  was	
  performed	
  using	
  the	
  delta-­‐delta	
  Ct	
  (∆∆Ct)	
  
method.	
  
Immunohistochemistry	
  and	
  morphometric	
  analysis	
  for	
  MSH2	
  and	
  MSH6	
  in	
  xenografts	
  
Immunohistochemistry	
  (IHC)	
  was	
  preformed	
  to	
  evaluate	
  MSH2	
  and	
  MSH6	
  
expression	
  in	
  xenografts	
  using	
  the	
  antibodies	
  described	
  above	
  for	
  Western	
  blots.	
  Five	
  µm	
  
sections	
  of	
  all	
  formalin-­‐fixed,	
  paraffin-­‐embedded	
  tissues	
  were	
  processed	
  for	
  
immunohistochemistry.	
  For	
  MSH2,	
  deparaffinization,	
  antigen	
  retrieval,	
  
immunohistochemical	
  staining	
  and	
  counterstaining	
  was	
  preformed	
  on	
  a	
  Dako	
  Link	
  48	
  
Autostainer	
  (Carpinteria,	
  CA,	
  USA)	
  using	
  a	
  LSAB2	
  kit	
  (Dako),	
  which	
  employs	
  a	
  3,3’	
  
diaminobenzidine	
  tetrahydrochloride	
  (DAB)	
  chromogen	
  detection	
  system.	
  	
  Retrieval	
  for	
  
MSH2	
  was	
  accomplished	
  using	
  heat	
  induced	
  epitope	
  retrieval	
  and	
  incubation	
  with	
  citric	
  
buffer	
  for	
  20	
  minutes.	
  For	
  MSH6,	
  deparaffinization,	
  antigen	
  retrieval,	
  
immunohistochemical	
  staining	
  and	
  counterstaining	
  was	
  preformed	
  on	
  a	
  Benchmark	
  XT™	
  
autostainer	
  (Ventana,	
  Tucson,	
  AZ,	
  USA)	
  using	
  an	
  Enhanced	
  Alkaline	
  Phosphatase	
  Red	
  
Detection	
  Kit	
  (Ventana)	
  that	
  uses	
  an	
  indirect	
  biotin	
  streptavidin	
  and	
  Fast	
  Red	
  chromogen	
  
detection	
  system.	
  	
  Blocking	
  of	
  nonspecific	
  binding	
  to	
  the	
  mouse	
  anti-­‐MSH6	
  antibody	
  to	
  
mouse	
  tissues	
  was	
  accomplished	
  by	
  preincubating	
  primary	
  antibody	
  with	
  secondary	
  
antibody	
  and	
  mouse	
  serum	
  prior	
  to	
  use.	
  Retrieval	
  for	
  MSH6	
  was	
  accomplished	
  with	
  20	
  

	
  

152	
  

	
  
minutes	
  of	
  incubation	
  with	
  Cell	
  Conditioning	
  1	
  solution	
  (Ventana).	
  Sections	
  of	
  normal	
  
urinary	
  bladder	
  were	
  used	
  as	
  positive	
  controls	
  and	
  run	
  in	
  parallel	
  to	
  cases	
  for	
  each	
  of	
  the	
  
IHC.	
  	
  For	
  negative	
  controls,	
  homologous	
  non-­‐immune	
  sera	
  or	
  buffer	
  replaced	
  primary	
  
antibodies.	
  
	
  

With	
  the	
  exception	
  of	
  TYLER1	
  and	
  TYLER2,	
  seven	
  randomly	
  picked	
  areas	
  from	
  each	
  

slide	
  immunohistochemically	
  labeled	
  for	
  the	
  respective	
  MMR	
  proteins	
  listed	
  above	
  were	
  
imaged	
  using	
  a	
  CRi	
  Nuance	
  multispectral	
  imaging	
  system	
  	
  (Woburn,	
  MA,	
  USA).	
  For	
  TYLER1	
  
and	
  TYLER2,	
  imaging	
  was	
  selectively	
  performed	
  to	
  isolate	
  regions	
  with	
  epithelial	
  and	
  
discrete	
  cell	
  morphology.	
  	
  The	
  resulting	
  image	
  cubes	
  were	
  converted	
  to	
  optical	
  density	
  
units,	
  then	
  mathematically	
  unmixed	
  to	
  separate	
  chromogen	
  labeling	
  from	
  counterstain	
  
using	
  spectral	
  libraries	
  generated	
  from	
  imaging	
  of	
  only	
  hematoxylin	
  stained	
  or	
  
immunohistochemically	
  labeled	
  and	
  hematoxylin	
  counterstained	
  control	
  specimens.	
  
Colocalized	
  immunoreactivity	
  was	
  assessed	
  as	
  positive	
  pixels	
  per	
  image	
  using	
  constant	
  
thresholds	
  for	
  detection	
  and	
  the	
  multispectral	
  imaging	
  system	
  software.	
  The	
  component	
  
images	
  of	
  the	
  image	
  cubes	
  were	
  then	
  pseudocolored,	
  converted	
  to	
  pseudofluorescent	
  
format,	
  and	
  unmixed	
  for	
  counting	
  of	
  nuclei.	
  	
  Such	
  counting	
  was	
  done	
  using	
  Imagine	
  
morphometric	
  analysis	
  software	
  developed	
  at	
  the	
  Van	
  Andel	
  Research	
  Institute	
  (Grand	
  
Rapids,	
  MI,	
  USA)	
  (Figures	
  22	
  and	
  23).	
  Final	
  results	
  of	
  morphometric	
  analysis	
  of	
  
immunoreactivity	
  were	
  reported	
  as	
  the	
  mean	
  number	
  of	
  positive	
  pixels	
  per	
  nucleus	
  for	
  
each	
  evaluated	
  sample.	
  
Effect	
  of	
  chemotherapeutics	
  on	
  cell	
  survival	
  
Survival	
  curves	
  were	
  generated	
  for	
  the	
  each	
  of	
  the	
  four	
  cell	
  lines	
  in	
  response	
  to	
  a	
  panel	
  
of	
   therapeutic	
   agents	
   using	
   XTT	
   assay	
   kits	
   (Trevigen,	
   Gaithersburg,	
   MD,	
   USA)	
   following	
   the	
  

	
  

153	
  

	
  
manufacturer’s	
   instructions.	
   Evaluated	
   chemotherapeutics	
   included	
   cisplatin,	
   carboplatin,	
  
oxaliplatin,	
   methotrexate,	
   thiotepa,	
   lomustine	
   (CCNU),	
   paclitaxel,	
   cytarabine,	
   and	
  
gemcitabine.	
   Briefly,	
   2x104	
   cells	
   of	
   each	
   cell	
   line	
   were	
   incubated	
   with	
   a	
   range	
   of	
  
concentrations	
   of	
   each	
   cytotoxic	
   agent.	
   Following	
   incubation	
   with	
   XTT,	
   a	
   terazolium	
  
compound	
  that	
  is	
  converted	
  to	
  a	
  formazan	
  dye	
  within	
  viable	
  cells,	
  absorbance	
  was	
  assessed	
  
using	
   a	
   Victor	
   X3	
   microplate	
   reader	
   (PerkinElmer,	
   Shelton,	
   CT	
   USA).	
   Corrected	
   absorbance	
  
was	
  calculated	
  by	
  subtracting	
  absorbance	
  read	
  at	
  630nm	
  from	
  absorbance	
  read	
  at	
  490nm.	
  
Surviving	
  cell	
  fraction	
  was	
  extrapolated	
  from	
  forecast	
  models	
  based	
  on	
  standards	
  derived	
  
from	
  incubating	
  known	
  numbers	
  of	
  concurrently	
  plated	
  cells	
  from	
  each	
  cell	
  line	
  with	
  XTT.	
  
Survival	
   assays	
   were	
   performed	
   with	
   three	
   technical	
   repeats	
   and	
   repeated	
   twice.	
   Means	
   of	
  
percent	
   cell	
   survival	
   for	
   each	
   tested	
   chemotherapeutic	
   concentration	
   were	
   used	
   to	
  
generate	
  survival	
  graphs.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

154	
  

	
  
Results	
  
Characterization	
  of	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines	
  and	
  xenografts	
  
Immunophenotyping	
  of	
  all	
  primary	
  tumors	
  was	
  consistent	
  with	
  urothelial	
  origin.	
  	
  
Specifically,	
  neoplastic	
  cells	
  throughout	
  all	
  tumors	
  diffusely	
  had	
  cytoplasmic	
  
immunoreactivity	
  for	
  MNF116,	
  cytoplasmic	
  and	
  perimembranous	
  immunoreactivity	
  for	
  
cytokeratin	
  7,	
  and	
  perimembranous	
  immunoreactivity	
  for	
  E-­‐cadherin	
  and	
  P-­‐cadherin	
  
(Figure	
  24).	
  	
  In	
  addition,	
  there	
  was	
  perimembranous	
  immunoreactivity	
  for	
  uroplakin	
  III	
  in	
  
apical	
  cell	
  layers.	
  	
  There	
  was	
  no	
  immunoreactivity	
  for	
  N-­‐cadherin,	
  vimentin,	
  or	
  PAP	
  in	
  any	
  
primary	
  tumor.	
  
After	
  >20	
  passages	
  ANGUS	
  and	
  KINSEY	
  had	
  a	
  uniform	
  epithelioid	
  morphology.	
  	
  
These	
  cells	
  were	
  plump	
  polygonal	
  and	
  grew	
  as	
  distinct,	
  dense	
  colonies	
  (Figure	
  25).	
  	
  The	
  
original	
  TYLER	
  cell	
  line	
  after	
  >10	
  passages	
  was	
  composed	
  of	
  a	
  biphasic	
  population	
  of	
  cells	
  
including	
  plump	
  polygonal	
  epithelioid	
  cells	
  that	
  grew	
  in	
  sheets	
  and	
  aggregates,	
  and	
  a	
  
population	
  of	
  more	
  spindle	
  to	
  stellate	
  cells.	
  	
  After	
  separation	
  of	
  these	
  distinct	
  populations	
  
by	
  selective	
  trypsinization	
  followed	
  by	
  >10	
  passages,	
  the	
  cell	
  lines	
  TYLER1	
  and	
  TYLER2	
  
were	
  morphologically	
  characterized.	
  	
  TYLER1	
  was	
  composed	
  of	
  plump	
  polygonal	
  
epithelioid	
  cells	
  that	
  grew	
  in	
  discrete	
  colonies.	
  TYLER2	
  was	
  composed	
  of	
  spindle	
  to	
  stellate	
  
cells	
  that	
  aggregated	
  into	
  sheets	
  and	
  bundles	
  when	
  near	
  confluence.	
  
All	
  cell	
  lines	
  were	
  confirmed	
  as	
  epithelial	
  in	
  origin	
  by	
  immunoreactivity	
  to	
  the	
  
pancytokeratin	
  marker	
  MNF116.	
  	
  Further,	
  all	
  cell	
  lines	
  expressed	
  cytokeratin	
  7	
  which	
  given	
  
the	
  location	
  and	
  histomorphology	
  of	
  the	
  primary	
  urothelial	
  carcinomas	
  from	
  which	
  they	
  
were	
  derived,	
  is	
  consistent	
  with	
  an	
  urothelial	
  origin.	
  	
  In	
  addition,	
  many	
  cells	
  from	
  the	
  
TYLER1	
  and	
  TYLER2	
  cell	
  lines	
  expressed	
  prostatic	
  acid	
  phosphatase	
  (PAP)	
  (Figure	
  26).	
  	
  All	
  

	
  

155	
  

	
  
cell	
  lines	
  were	
  immunoreactive	
  for	
  vimentin,	
  and	
  no	
  cell	
  line	
  expressed	
  uroplakin	
  3,	
  a	
  
marker	
  expressed	
  late	
  in	
  urothelial	
  differentiation	
  and	
  only	
  by	
  apical	
  umbrella	
  cells	
  of	
  the	
  
urothelium.	
  
Characterization	
  of	
  xenografts	
  
ANGUS	
  failed	
  to	
  produce	
  appreciable	
  tumors	
  in	
  xenograft	
  models.	
  	
  KINSEY,	
  TYLER1,	
  
and	
  TYLER2	
  grew	
  well	
  as	
  xenografts	
  forming	
  rapidly	
  growing	
  masses.	
  In	
  xenografts,	
  
tumors	
  derived	
  from	
  the	
  KINSEY	
  cell	
  line	
  were	
  composed	
  of	
  neoplastic	
  plump	
  polygonal	
  
epithelial	
  cells	
  arranged	
  in	
  dense	
  sheets	
  and	
  trabeculae.	
  	
  Three	
  of	
  6	
  TYLER1	
  xenografts	
  and	
  
3	
  of	
  4	
  TYLER2	
  xenografts	
  were	
  composed	
  of	
  biphasic	
  cell	
  populations	
  including	
  multifocal	
  
regions	
  of	
  epithelioid	
  differentiation	
  and	
  extensive	
  regions	
  of	
  discrete	
  round	
  cells	
  (Figure	
  
27).	
  	
  Areas	
  of	
  epithelioid	
  differentiation	
  were	
  composed	
  of	
  nests	
  of	
  polygonal	
  neoplastic	
  
cells	
  that	
  often	
  had	
  central	
  sharply	
  defined,	
  open	
  or	
  proteinaceous	
  fluid	
  filled	
  cavitations.	
  	
  
Discrete	
  cells	
  were	
  round,	
  loosely	
  associated	
  with	
  one	
  another,	
  and	
  arranged	
  in	
  dense	
  
sheets	
  supported	
  by	
  scant	
  fine	
  fibrovascular	
  stroma.	
  	
  The	
  remaining	
  3	
  TYLER1	
  xenografts	
  
were	
  composed	
  of	
  only	
  epithelioid	
  cells	
  and	
  the	
  remaining	
  TYLER2	
  cell	
  line	
  was	
  only	
  
composed	
  of	
  discrete	
  cells.	
  	
  Such	
  biphasic	
  cell	
  populations	
  were	
  also	
  seen	
  in	
  xenografts	
  
derived	
  from	
  subclones	
  of	
  TYLER1	
  and	
  TYLER2	
  generated	
  through	
  serial	
  dilution.	
  
Using	
  immunohistochemistry,	
  neoplastic	
  cells	
  of	
  KINSEY	
  xenografts	
  had	
  
immunoreactivity	
  for	
  the	
  epithelial	
  markers	
  MNF116,	
  E-­‐cadherin,	
  P-­‐cadherin;	
  
immunoreactivity	
  for	
  cytokeratin	
  7	
  and	
  uroplakin	
  III,	
  which	
  are	
  expressed	
  by	
  urotheluim;	
  
diffuse	
  immunoreactivity	
  for	
  vimentin,	
  an	
  intermediate	
  filament	
  expressed	
  in	
  
mesenchymal	
  cells;	
  and	
  no	
  immunoreactivity	
  for	
  PAP,	
  a	
  marker	
  of	
  prostatic	
  epithelium	
  
(Figure	
  28).	
  	
  Epithelioid	
  populations	
  of	
  TYLER1	
  and	
  TYLER2	
  xenografts	
  had	
  expression	
  of	
  

	
  

156	
  

	
  
MNF116,	
  E-­‐cadherin,	
  P-­‐cadherin,	
  N-­‐cadherin,	
  cytokeratin	
  7,	
  and	
  uroplakin	
  III	
  (Figure	
  29).	
  	
  
Discrete	
  cell	
  populations	
  of	
  TYLER1	
  and	
  TYLER2	
  had	
  no	
  expression	
  of	
  MNF-­‐116,	
  E-­‐
cadherin,	
  P-­‐cadherin,	
  cytokeratin	
  7,	
  or	
  uroplakin	
  III.	
  	
  Both	
  cell	
  populations	
  had	
  strong	
  
expression	
  of	
  vimentin	
  and	
  few	
  scattered	
  individual	
  cells	
  of	
  both	
  populations	
  had	
  
immunoreactivity	
  for	
  PAP.	
  
Evaluation	
  of	
  MMR	
  protein	
  expression	
  by	
  Western	
  blots	
  
	
  

Western	
  blotting	
  yielded	
  bands	
  corresponding	
  to	
  expected	
  molecular	
  weights	
  for	
  all	
  

evaluated	
  markers	
  in	
  all	
  cell	
  lines.	
  	
  The	
  expression	
  levels	
  of	
  MLH1	
  and	
  MSH3	
  were	
  
relatively	
  consistent	
  across	
  cell	
  lines;	
  however,	
  the	
  relative	
  intensity	
  of	
  the	
  MSH2	
  and	
  
MSH6	
  bands	
  were	
  lower	
  in	
  the	
  TYLER2	
  cell	
  line	
  in	
  comparison	
  to	
  that	
  of	
  other	
  cell	
  lines	
  
consistent	
  with	
  lesser	
  expression	
  of	
  these	
  proteins	
  (Figure	
  30).	
  	
  	
  
qPCR	
  for	
  MMR	
  gene	
  expression	
  
With	
  the	
  exception	
  of	
  MSH6	
  in	
  KINSEY	
  in	
  which	
  there	
  was	
  moderate	
  variance,	
  there	
  
was	
  only	
  mild	
  variance	
  in	
  relative	
  gene	
  expression	
  between	
  technical	
  and	
  biologic	
  repeats.	
  	
  
Examining	
  expression	
  of	
  the	
  MMR	
  genes	
  in	
  cell	
  lines	
  relative	
  to	
  a	
  compilation	
  of	
  the	
  
expression	
  of	
  control	
  genes,	
  GAPDH,	
  β-­‐actin,	
  and	
  HPRT1,	
  TYLER2	
  had	
  lower	
  levels	
  of	
  
expression	
  of	
  MSH2	
  and	
  MSH6	
  than	
  ANGUS,	
  KINSEY,	
  and	
  TYLER	
  (Figure	
  31).	
  There	
  were	
  
no	
  obvious	
  differences	
  in	
  expression	
  of	
  MLH1,	
  MSH3,	
  PMS2	
  or	
  MLH1	
  between	
  evaluated	
  
cell	
  lines.	
  	
  There	
  was	
  significant	
  variance	
  in	
  the	
  relative	
  expression	
  of	
  MSH2	
  and	
  MSH6	
  
relative	
  to	
  GAPDH	
  for	
  all	
  evaluated	
  xenografts;	
  however,	
  mean	
  relative	
  expression	
  of	
  MSH2	
  
and	
  MSH6	
  was	
  higher	
  in	
  KINSEY	
  than	
  in	
  either	
  TYLER1	
  or	
  TYLER2,	
  and	
  the	
  there	
  were	
  no	
  
or	
  only	
  mild	
  differences	
  in	
  mean	
  relative	
  expression	
  of	
  MSH2	
  and	
  MSH6	
  between	
  TYLER1	
  
and	
  TYLER2	
  (Figure	
  32).	
  	
  In	
  evaluations	
  of	
  relative	
  MSH2	
  gene	
  expression	
  in	
  LCM	
  

	
  

157	
  

	
  
separated	
  epithelioid	
  and	
  discrete	
  cell	
  populations	
  of	
  TYLER1	
  and	
  TYLER2,	
  there	
  was	
  
differential	
  expression	
  of	
  MSH2	
  between	
  epithelial	
  and	
  discrete	
  cell	
  populations	
  for	
  both	
  
TYLER1	
  and	
  TYLER2	
  (Figure	
  33).	
  	
  There	
  was,	
  however,	
  no	
  difference	
  in	
  MSH2	
  expression	
  
comparing	
  the	
  separated	
  cell	
  groups	
  of	
  TYLER1	
  to	
  those	
  of	
  TYLER2.	
  	
  
Immunohistochemistry	
  and	
  morphometric	
  analysis	
  for	
  MSH2	
  and	
  MSH6	
  in	
  xenografts	
  
In	
  general,	
  the	
  mean	
  expression	
  levels	
  of	
  MSH2	
  and	
  MSH6	
  in	
  terms	
  of	
  positive	
  pixels	
  
per	
  nucleus	
  was	
  lower	
  in	
  discrete	
  cell	
  populations	
  than	
  in	
  epithelioid	
  populations	
  for	
  both	
  
TYLER1	
  and	
  TYLER2	
  (Figure	
  34;	
  Tables	
  15	
  and	
  16).	
  	
  Further,	
  the	
  mean	
  expression	
  levels	
  of	
  
MSH2	
  and	
  MSH6,	
  for	
  both	
  epithelioid	
  and	
  discrete	
  cell	
  populations	
  were	
  respectively	
  
higher	
  in	
  TYLER1	
  populations	
  than	
  in	
  TYLER2	
  populations.	
  	
  The	
  mean	
  expression	
  of	
  MSH6	
  
was	
  higher	
  in	
  KINSEY	
  in	
  comparison	
  to	
  that	
  of	
  TYLER2	
  cell	
  lines	
  and	
  discrete	
  TYLER1	
  cell	
  
populations,	
  but	
  was	
  not	
  different	
  from	
  that	
  of	
  epithelioid	
  portions	
  of	
  TYLER1.	
  	
  Expression	
  
of	
  MSH2	
  in	
  KINSEY	
  was	
  low	
  relative	
  to	
  that	
  observed	
  in	
  epithelioid	
  populations.	
  	
  However,	
  
there	
  was	
  marked	
  variation	
  in	
  quantitated	
  immunoreactivity,	
  and	
  particularly	
  large	
  
differences	
  in	
  immunoreactivity	
  for	
  MSH2	
  between	
  the	
  TYLER	
  cell	
  line	
  subgroups.	
  	
  	
  
Using	
  one-­‐way	
  ANOVA	
  tests,	
  there	
  were	
  significant	
  differences	
  in	
  the	
  
immunoreactivity	
  for	
  MSH6	
  between	
  all	
  cell	
  groups	
  (p<0.001)	
  and	
  when	
  comparing	
  only	
  
TYLER	
  xenografts	
  (p=0.015).	
  	
  There	
  was,	
  however,	
  no	
  significant	
  difference	
  between	
  the	
  
immunoreactivity	
  for	
  MSH2	
  whether	
  comparing	
  all	
  groups	
  (p=0.125)	
  or	
  only	
  TYLER	
  
xenografts	
  (p=0.230).	
  
Survival	
  assays	
  
Graphic	
  results	
  of	
  XTT	
  survival	
  assays	
  are	
  presented	
  in	
  Figure	
  35.	
  	
  	
  With	
  regards	
  to	
  
differences	
  in	
  chemotherapeutic	
  sensitivity	
  for	
  TYLER1	
  and	
  TYLER2,	
  there	
  were	
  

	
  

158	
  

	
  
appreciable	
  differences	
  in	
  sensitivity	
  to	
  carboplatin,	
  oxaliplatin,	
  methotrexate,	
  and	
  
thiotepa.	
  TYLER2	
  was	
  more	
  sensitive	
  to	
  oxaliplatin,	
  moderate	
  concentrations	
  of	
  
methotrexate,	
  and	
  lower	
  concentrations	
  of	
  thiotepa	
  than	
  TYLER1,	
  but	
  less	
  sensitive	
  to	
  
carboplatin.	
  	
  ANGUS	
  often	
  was	
  less	
  sensitive	
  to	
  chemotherapeutics	
  relative	
  to	
  other	
  
evaluated	
  cell	
  lines	
  including	
  for	
  carboplatin,	
  cisplatin,	
  thiotepa,	
  and	
  gemcitabine.	
  	
  
Sensitivity	
  of	
  KINSEY	
  was	
  similar	
  to	
  that	
  of	
  TYLER1	
  for	
  carboplatin,	
  TYLER1	
  and	
  TYLER2	
  
for	
  cisplatin,	
  ANUGS	
  and	
  TYLER1	
  for	
  oxaliplatin,	
  TYLER2	
  for	
  methotrexate,	
  and	
  TYLER1	
  
and	
  TYLER2	
  for	
  gemcitabine.	
  	
  For	
  thiotepa,	
  there	
  were	
  differences	
  in	
  sensitivity	
  for	
  all	
  
tested	
  cell	
  lines,	
  especially	
  at	
  the	
  lowest	
  tested	
  concentrations,	
  with	
  TYLER2	
  being	
  most	
  
sensitive,	
  followed	
  by	
  TYLER2,	
  KINSEY	
  and	
  ANGUS.	
  	
  There	
  was	
  no	
  appreciable	
  difference	
  in	
  
sensitivity	
  of	
  TYLER1,	
  TYLER2,	
  or	
  KINSEY	
  for	
  gemcitabine.	
  	
  There	
  were	
  no	
  or	
  only	
  mild	
  
differences	
  in	
  response	
  of	
  all	
  evaluated	
  cell	
  lines	
  to	
  paclitaxel,	
  cytarabine,	
  lomustine.	
  	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

159	
  

	
  
Discussion	
  
In	
  humans,	
  the	
  potential	
  application	
  of	
  MMR	
  as	
  a	
  therapeutic	
  target	
  is	
  highlighted	
  by	
  
the	
   differences	
   observed	
   in	
   prognosis	
   and	
   the	
   response	
   to	
   treatment	
   between	
   cancers	
   of	
  
the	
  same	
  type	
  that	
  differ	
  in	
  MMR	
  capacity.	
  In	
  terms	
  of	
  prognosis,	
  MMR	
  deficient	
  cancers	
  are	
  
generally	
   associated	
   with	
   a	
   more	
   favorable	
   clinical	
   outcome	
   than	
   those	
   that	
   are	
   MMR	
  
proficient.6-­‐10	
  In	
  colorectal	
  carcinomas,	
  for	
  example,	
  MMR	
  deficient	
  tumors	
  are	
  associated	
  
with	
  longer	
  survival	
  times	
  and	
  a	
  decreased	
  risk	
  of	
  metastasis	
  compared	
  to	
  those	
  that	
  are	
  
proficient.9	
   The	
   reason	
   for	
   this	
   difference	
   in	
   prognosis	
   is	
   likely	
   multifactorial.	
   Inherently,	
  
cancers	
   with	
   defects	
   in	
   MMR	
   are	
   genetically	
   less	
   stable	
   that	
   those	
   with	
   intact	
   MMR	
   leading	
  
to	
  an	
  increase	
  in	
  DNA	
  lesions,	
  which	
  promotes	
  cell	
  cycle	
  arrest	
  and	
  apoptosis	
  signaling.11	
  
Also,	
  there	
  is	
  evidence	
  that	
  MMR	
  deficient	
  cancer	
  cells	
  are	
  often	
  highly	
  immunogenic.	
  This	
  
is	
  proposed	
  to	
  occur	
  due	
  to	
  production	
  of	
  atypical	
  proteins	
  generated	
  through	
  frame	
  shift	
  
mutations	
   resulting	
   in	
   T	
   cell	
   mediated	
   immune	
   responses	
   directed	
   against	
   the	
   cancer	
  
cells.12	
   Additionally,	
   it	
   has	
   been	
   shown	
   that	
   several	
   genes	
   associated	
   with	
   antitumor	
  
immune	
  responses	
  are	
  over	
  expressed	
  in	
  MMR	
  deficient	
  cancers	
  and	
  cell	
  lines.11	
  
Given	
  the	
  range	
  of	
  DNA	
  repair	
  and	
  damage	
  signaling	
  pathways	
  with	
  which	
  the	
  MMR	
  
system	
  is	
  associated,	
  it	
  is	
  not	
  surprising	
  that	
  defects	
  in	
  MMR	
  also	
  have	
  significant	
  effects	
  on	
  
response	
   of	
   cancers	
   to	
   therapy.	
   The	
   effects	
   of	
   MMR	
   deficiency	
   on	
   chemotherapeutic	
  
response	
   are	
   complex.	
   MMR	
   deficiency	
   is	
   capable	
   of	
   conferring	
   either	
   drug	
   resistance	
   or	
  
sensitivity	
  largely	
  based	
  on	
  a	
  given	
  drug’s	
  mechanism	
  of	
  action.8,	
  13-­‐21	
  The	
  differential	
  effect	
  
of	
  the	
  MMR	
  system	
  relative	
  to	
  chemosensitivity	
  likely	
  in	
  part	
  reflects	
  the	
  ability	
  of	
  the	
  MMR	
  
machinery	
   to	
   participate	
   in	
   a	
   number	
   of	
   alternative	
   DNA	
   damage	
   processing/signaling	
  
pathways.	
   In	
   the	
   case	
   of	
   simple	
   methylated	
   bases	
   in	
   the	
   DNA	
   molecule,	
   it	
   has	
   been	
  

	
  

160	
  

	
  
proposed	
   that	
   MMR	
   may	
   be	
   involved	
   in	
   triggering	
   apoptosis	
   through	
   either	
   futile	
   repair	
  
attempts	
   or	
   through	
   conversion	
   of	
   the	
   alkylated	
   base	
   to	
   a	
   lethal	
   lesion	
   such	
   as	
   a	
   double	
  
strand	
   break.21	
   Thus,	
   in	
   this	
   case,	
   deficiency	
   of	
   MMR	
   results	
   in	
   tolerance	
   of	
   alkylated	
   bases	
  
and	
   drug	
   resistance.	
   In	
   contrast,	
   an	
   increased	
   sensitivity	
   to	
   cytotoxins	
   can	
   be	
   observed	
   if	
  
defective	
  MMR	
  results	
  in	
  failure	
  to	
  repair	
  certain	
  types	
  of	
  DNA	
  lesions.	
  For	
  example,	
  certain	
  
interstrand	
  cross-­‐links	
   are	
   recognized	
  by	
  the	
  MMR	
  system	
  and	
  repair	
  is	
  thought	
  to	
  occur	
  
through	
   MMR	
   mediated	
   homologous	
   recombination.13	
   Cells	
   deficient	
   in	
   MMR	
   do	
   not	
  
effectively	
   repair	
   these	
   cross-­‐links	
   and	
   are	
   sensitive	
   to	
   agents	
   that	
   induce	
   such	
   these	
  
lesions.	
  	
  
Of	
  the	
  4	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines	
  evaluated	
  in	
  the	
  
current	
  study,	
  TYLER2	
  was	
  considered	
  to	
  be	
  MMR	
  deficient	
  relative	
  to	
  the	
  other	
  cell	
  lines	
  
in	
   terms	
   of	
   MSH2	
   and	
   MSH6	
   gene	
   and	
   protein	
   expression.	
   Such	
   differential	
   expression	
   of	
  
factors	
   in	
   the	
   MMR	
   pathway	
   in	
   cell	
   lines	
   offers	
   the	
   opportunity	
   to	
   study	
   the	
   effect	
   of	
  
differential	
  expression	
  of	
  MMR	
  on	
  chemotherapeutic	
  response	
  in	
  an	
  in	
  vitro	
  canine	
  model.	
  	
  
This	
   is	
   particularly	
   true	
   as	
   TYLER1	
   and	
   TYLER2	
   provide	
   the	
   opportunity	
   to	
   evaluate	
  
naturally	
   occurring	
   differenced	
   in	
   MMR	
   capacity	
   while	
   sharing	
   a	
   similar	
   genetic	
  
background	
   having	
   been	
   derived	
   from	
   the	
   same	
   primary	
   tumor.	
   The	
   differential	
  
sensitivities	
   observed	
   in	
   the	
   MMR	
   deficient	
   TYLER2	
   to	
   chemotherapeutics	
   are	
   consistent	
  
with	
  reports	
  in	
  similar	
  studies	
  in	
  humans.	
  The	
  fact	
  that	
  ANGUS	
  was	
  resistant	
  to	
  a	
  number	
  of	
  
chemotherapeutics	
  relative	
  to	
  other	
  cell	
  lines	
  is	
  not	
  surprising,	
  as	
  the	
  primary	
  tumor	
  was	
  
aggressively	
  treated	
  with	
  chemotherapy	
  prior	
  to	
  establishment	
  of	
  the	
  cell	
  line.	
  	
  	
  
Cisplatin,	
  carboplatin,	
  and	
  oxaliplatin	
  are	
  platinum-­‐containing	
  chemotherapeutics	
  that	
  
act	
  by	
  forming	
  DNA	
  adducts	
  and	
  interstrand	
  crosslinks.	
  	
  Recognition	
  of	
  such	
  DNA	
  damage	
  

	
  

161	
  

	
  
promotes	
   apoptosis	
   through	
   subsequent	
   activation	
   of	
   apoptotic	
   signaling	
   pathways.	
   In	
  
humans,	
   resistance	
   to	
   cisplatin	
   and	
   carboplatin	
   has	
   been	
   reported	
   in	
   MMR	
   deficient	
  
cancers,	
   while	
   no	
   such	
   resistance	
   has	
   been	
   reported	
   to	
   oxaliplatin.8,	
   18,	
   22-­‐26	
   While	
   these	
  
drugs	
   have	
   similar	
   mechanisms	
   of	
   action,	
   it	
   has	
   been	
   suggested	
   oxaliplatin-­‐related	
   DNA	
  
adducts	
   are	
   not	
   normally	
   recognized	
   by	
   MMR	
   and	
   thus	
   deficiency	
   has	
   little	
   effect	
   on	
  
effectiveness	
   of	
   oxaliplatin.8,	
  

26	
  

	
   Response	
   of	
   canine	
   lower	
   urinary	
   tract	
   urothelial	
  

carcinomas	
   to	
   carboplatin	
   parallel	
   these	
   findings,	
   where	
   the	
   MMR	
   deficient	
   cell	
   line	
  
TYLER2	
   was	
   relatively	
   resistant;	
   however,	
   there	
   was	
   no	
   difference	
   in	
   response	
   to	
   cisplatin,	
  
and	
  TYLER2	
  was	
  more	
  sensitive	
  to	
  oxaliplatin.	
  
Gemcitabine	
  and	
  cytarabine,	
  which	
  were	
  evaluated	
  in	
  the	
  current	
  study,	
  are	
  pyrimidine	
  
nucleoside	
  analogs.	
  Radiosensitization	
  with	
  gemcitabine	
  in	
  the	
  face	
  of	
  MMR	
  deficiency	
  has	
  
been	
  suggested.27,	
   28	
  Human	
  carcinoma	
  MLH1	
  and	
  MSH2-­‐deficient	
  cell	
  lines	
  were	
  found	
  to	
  
be	
  sensitively	
  to	
  cytarabine.29	
  In	
  contrast	
  to	
  these	
  results,	
  there	
  was	
  no	
  apparent	
  difference	
  
in	
  the	
  sensitivities	
  of	
  MMR	
  proficient	
  and	
  deficient	
  cell	
  lines	
  to	
  gemcitabine	
  or	
  cytarabine.	
  
In	
   humans,	
   MMR	
   deficiency	
   results	
   in	
   drug	
   resistance	
   to	
   certain	
   alkylating	
   agents	
  
including	
   the	
   SN1	
   methylators,	
   temozolomide	
   and	
   dacarbazine.8,	
  

13,	
   18,	
   19,	
   21	
  

However,	
  

reports	
   suggest	
   that	
   MMR	
   deficient	
   cells	
   are	
   highly	
   sensitive	
   to	
   many	
   of	
   the	
   interstrand	
  
cross-­‐linking	
  alkylators,	
  including	
  lomustine	
  (CCNU)	
  and	
  mitomycin	
  C.13,	
   21	
  The	
  role	
  MMR	
  
directly	
   played	
   in	
   sensitivity	
   to	
   alkylators	
   in	
   our	
   canine	
   urothelial	
   carcinoma	
   cell	
   lines	
   is	
  
unclear.	
  	
  While	
  the	
  MMR	
  deficient	
  TYLER2	
  cell	
  was	
  most	
  sensitive	
  to	
  the	
  alkylator,	
  thiotepa,	
  
there	
   was	
   wide	
   variation	
   in	
   response	
   to	
   treatment	
   for	
   all	
   tested	
   cell	
   lines	
   regardless	
   of	
  
MMR	
   status,	
   and	
   there	
   were	
   no	
   apparent	
   differences	
   in	
   response	
   of	
   the	
   MMR	
   deficient	
  
TYLER2	
  to	
  lomustine.	
  	
  	
  

	
  

162	
  

	
  
Methotrexate	
   is	
   a	
   folic	
   acid	
   inhibitor.	
   	
   There	
   are	
   contradictory	
   reports	
   regarding	
   the	
  
effect	
  of	
  MMR	
  proficiency	
  on	
  sensitivity	
  to	
  methotrexate.	
  One	
  study	
  suggested	
  an	
  increased	
  
sensitivity	
  to	
  methotrexate	
  in	
  MMR	
  deficient	
  cell	
  lines.	
  	
  In	
  another	
  study,	
  comparison	
  of	
  the	
  
effect	
   of	
   methotrexate	
   on	
   a	
   cell	
   line	
   with	
   inactivation	
   of	
   MLH1	
   and	
   decreased	
   expression	
   of	
  
PMS2	
   to	
   a	
   MMR	
   proficient	
   cell	
   line	
   found	
   increased	
   sensitivity	
   in	
   the	
   MMR	
   proficient	
   cell	
  
line30	
   In	
   the	
   canine	
   urothelial	
   carcinoma	
   cell	
   lines,	
   the	
   MMR	
   deficient	
   TYLER2	
   was	
   more	
  
sensitive	
  to	
  methotrexate	
  than	
  the	
  proficient	
  TYLER1.	
  
Paclitaxel	
  (taxol)	
  is	
  an	
  inhibitor	
  of	
  mitosis	
  that	
  acts	
  by	
  interfering	
  with	
  breakdown	
  of	
  
microtubules.	
   Previous	
   studies	
   have	
   found	
   no	
   association	
   between	
   MMR	
   proficiency	
   and	
  
paclitaxel	
  sensitivity,	
  and	
  no	
  obvious	
  differences	
  were	
  observed	
  in	
  cell	
  lines	
  of	
  the	
  current	
  
study.24,	
  31	
  	
  
	
  

The	
   differential	
   expression	
   of	
   MSH2	
   and	
   MSH6	
   likely	
   influenced	
   the	
   differential	
  

sensitivities	
   of	
   TYLER2	
   to	
   select	
   chemotherapeutics	
   similar	
   to	
   that	
   described	
   in	
   the	
   human	
  
literature;	
   however,	
   differences	
   in	
   other	
   carcinogenesis	
   related	
   pathways	
   cannot	
   be	
  
excluded	
   as	
   contributing	
   to	
   relative	
   chemosensitivity	
   or	
   chemoresistance	
   in	
   the	
   current	
  
study.	
   	
   While	
   TYLER1	
   and	
   TYLER2	
   were	
   derived	
   from	
   the	
   same	
   tumor,	
   produced	
   similar	
  
tumors	
  in	
  xenografts,	
  and	
  maintained	
  similar	
  expression	
  profiles	
  of	
  immunohistochemical	
  
markers,	
   the	
   exact	
   similarity	
   of	
   the	
   molecular	
   constitution	
   of	
   these	
   cell	
   lines	
   is	
   unknown.	
  	
  
This	
   is	
   particularly	
   true	
   given	
   that	
   there	
   are	
   marked	
   differences	
   in	
   morphology	
   of	
   these	
  
cells	
   in	
   culture	
   likely	
   suggesting	
   differences	
   between	
   the	
   cell	
   lines	
   outside	
   the	
   MMR	
  
pathway.	
  Further	
  the	
  growth	
  of	
  tumors	
  in	
  xenografts	
  composed	
  of	
  biphasic	
  cell	
  populations	
  
suggests	
   that	
   each	
   cell	
   line	
   may	
   be	
   composed	
   of	
   a	
   mixed	
   cell	
   population,	
   components	
   of	
  
which	
   may	
   differ	
   in	
   MMR	
   proficiency;	
   however,	
   some	
   of	
   the	
   xenografts	
   of	
   TYLER1	
   and	
  

	
  

163	
  

	
  
TYLER2	
   cell	
   line	
   subclones	
   also	
   had	
   epithelioid	
   and	
   discrete	
   cell	
   populations	
   suggesting	
  
divergent	
  differentiation	
  within	
  the	
  mouse	
  rather	
  than	
  injection	
  of	
  multiple	
  cell	
  types.	
  
Overall,	
   these	
   studies	
   suggest	
   that	
   there	
   are	
   similarities	
   in	
   dogs	
   and	
   humans	
  
regarding	
   chemosensitivity	
   of	
   MMR	
   proficient	
   and	
   deficient	
   cells.	
   	
   In	
   addition,	
   the	
  
demonstrated	
   variance	
   in	
   chemosensitivity	
   suggests	
   likely	
   clinical	
   implications	
   for	
   dogs	
  
affected	
   with	
   lower	
   urinary	
   tract	
   urothelial	
   carcinoma	
   with	
   regards	
   to	
   selection	
   of	
  
treatment.	
  	
  	
  
	
  
Acknowledgments	
  
Dr.	
  Sledge’s	
  graduate	
  program	
  was	
  funded	
  by	
  Bristol-­‐Meyers-­‐Squibb	
  through	
  the	
  
American	
  College	
  of	
  Veterinary	
  Pathologists/Society	
  of	
  Toxicologic	
  Pathologists	
  coalition.	
  
Funding	
  for	
  portions	
  of	
  this	
  work	
  were	
  provided	
  by	
  a	
  companion	
  animal	
  fund	
  grant	
  
through	
  the	
  Michigan	
  State	
  University	
  College	
  of	
  Veterinary	
  Medicine.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

164	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
APPENDIX	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

165	
  

	
  
Figure	
  22.	
  Multispectral	
  imaging	
  of	
  MSH2	
  immunohistochemistry	
  in	
  canine	
  lower	
  urinary	
  
tract	
  urothelial	
  carcinoma	
  xenografts.	
  Immunoreactivity	
  for	
  MSH2	
  in	
  unmixed	
  images	
  was	
  
detected	
  by	
  3,3’-­‐Diaminobenzidine	
  (DAB)	
  chromogen	
  with	
  hematoxylin	
  counterstain.	
  
Colocalization	
  of	
  immunoreactivity	
  (yellow)	
  was	
  determined	
  using	
  set	
  thresholds	
  defined	
  
from	
  spectral	
  libraries	
  of	
  controls.	
  	
  Morphometric	
  analysis	
  used	
  to	
  count	
  nuclei	
  in	
  images	
  
converted	
  to	
  psedofluorescene.	
  
	
  

	
  

166	
  

	
  
Figure	
  22	
  (cont’d)

	
  
	
  
	
  

	
  

167	
  

	
  
	
  
Figure	
  23:	
  Multispectral	
  imaging	
  of	
  MSH6	
  immunohistochemistry	
  in	
  canine	
  lower	
  urinary	
  
tract	
  urothelial	
  carcinoma	
  xenografts.	
  Immunoreactivity	
  for	
  MSH2	
  in	
  unmixed	
  images	
  was	
  
detected	
  by	
  indirect	
  biotin	
  streptavidin	
  and	
  Fast	
  Red	
  chromogen,	
  hematoxylin	
  
counterstain.	
  Colocalization	
  of	
  immunoreactivity	
  (yellow)	
  was	
  determined	
  using	
  set	
  
thresholds	
  defined	
  from	
  spectral	
  libraries	
  of	
  controls	
  and	
  morphometric	
  analysis	
  used	
  to	
  
count	
  nuclei	
  in	
  images	
  converted	
  to	
  psedofluorescene.	
  

	
  

168	
  

	
  
Figure	
  23	
  (cont’d)

	
  
	
  
	
  

	
  

169	
  

	
  
	
  
Figure	
  24:	
  Photomicrographs	
  from	
  the	
  initial	
  diagnostic	
  biopsy	
  and	
  immunohistochemistry	
  
(IHC)	
  on	
  the	
  papillary	
  grade	
  II	
  urothelial	
  from	
  which	
  TYLER	
  cell	
  lines	
  were	
  derived.	
  	
  The	
  
histomorphology	
  of	
  the	
  neoplasm	
  and	
  pattern	
  or	
  immunoreactivity	
  for	
  the	
  tested	
  IHC	
  
markers	
  are	
  consistent	
  with	
  a	
  urothelial	
  carcinoma.	
  	
  Neoplastic	
  polygonal	
  urothelial	
  cells	
  
are	
  arranged	
  in	
  dense	
  nests	
  that	
  occasionally	
  have	
  open	
  or	
  necrotic	
  centers,	
  and	
  that	
  are	
  
supported	
  by	
  supported	
  by	
  scant	
  fine	
  fibrovascular	
  stroma,	
  hematoxylin	
  and	
  eosin	
  (A).	
  
Neoplastic	
  epithelial	
  cells	
  have	
  strong	
  cytoplasmic	
  immunoreactivity	
  for	
  the	
  
pancytokeratin	
  marker	
  MNF116	
  (B)	
  3,3’-­‐Diaminobenzidine	
  (DAB)	
  chromogen,	
  hematoxylin	
  
counterstain.	
  While	
  stromal	
  cells	
  are	
  immunoreactive,	
  neoplastic	
  urothelial	
  cells	
  are	
  not	
  
labeled	
  for	
  vimentin	
  (C),	
  indirect	
  biotin	
  streptavidin	
  and	
  Fast	
  Red	
  chromogen,	
  hematoxylin	
  
counterstain.	
  There	
  is	
  strong	
  perimembranous	
  immunoreactivity	
  for	
  cytokeratin	
  7	
  
throughout	
  the	
  neoplastic	
  cells	
  (D),	
  strong	
  apical	
  expression	
  Uroplakin	
  3	
  in	
  cells	
  at	
  the	
  
center	
  of	
  nests	
  (E),	
  and	
  no	
  immunoreactivity	
  for	
  prostatic	
  acid	
  phosphatase	
  (PAP),	
  DAB	
  
chromogen,	
  hematoxylin	
  counterstain.	
  	
  

	
  
	
  

170	
  

	
  
Figure	
  25:	
  	
  Phase	
  contrast	
  photomicrographs	
  of	
  ANGUS,	
  KINSEY,	
  original	
  TYLER	
  cell	
  lines.	
  	
  
ANGUS	
  (A)	
  and	
  KINSEY	
  (B)	
  cell	
  lines	
  are	
  composed	
  of	
  plump	
  polygonal	
  cells	
  that	
  grow	
  in	
  
intimately	
  associated	
  colonies	
  and	
  sheets.	
  	
  Cells	
  of	
  ANGUS	
  contain	
  occasional	
  refractile,	
  
sharply	
  demarcated,	
  vacuoles.	
  	
  The	
  original	
  TYLER	
  cell	
  line	
  is	
  composed	
  of	
  a	
  biphasic	
  
population	
  of	
  plump	
  polygonal	
  cells	
  that	
  generally	
  form	
  dense,	
  tightly	
  aggregated	
  colonies	
  
and	
  sheets,	
  and	
  an	
  intervening	
  population	
  of	
  loosely	
  arranged	
  spindle	
  to	
  stellate	
  cells	
  (C).	
  	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

171	
  

	
  
Figure	
  26:	
  Phase	
  contrast	
  photomicrographs	
  of	
  TYLER1	
  and	
  TYLER2	
  cell	
  lines	
  and	
  
photomicrographs	
  of	
  IHC	
  for	
  differentiation	
  markers	
  in	
  TYLER1	
  and	
  TYLER2	
  cell	
  lines.	
  In	
  cell	
  
culture,	
  TYLER1	
  is	
  comprised	
  of	
  plump	
  polygonal	
  cells	
  that	
  generally	
  form	
  dense,	
  tightly	
  
aggregated	
  colonies	
  and	
  sheets.	
  	
  TYLER2	
  is	
  composed	
  of	
  cells	
  that	
  have	
  a	
  spindle	
  to	
  stellate	
  
morphology	
  and	
  grow	
  in	
  loose,	
  haphazardly	
  arranged	
  patterns.	
  	
  TYLER1	
  and	
  TYLER2	
  have	
  
similar	
  expression	
  of	
  immunohistochemical	
  markers.	
  In	
  both	
  cell	
  lines,	
  individual	
  and	
  
scattered	
  aggregates	
  of	
  cells	
  have	
  cytoplasmic	
  immunoreactivity	
  for	
  cytokeratin	
  7	
  and	
  
prostatic	
  acid	
  phosphatase	
  (PAP),	
  3,3’-­‐Diaminobenzidine	
  (DAB)	
  chromogen,	
  hematoxylin	
  
counterstain.	
  	
  Cells	
  from	
  both	
  lines	
  have	
  strong	
  cytoplasmic	
  expression	
  of	
  vimentin,	
  
indirect	
  biotin	
  streptavidin	
  and	
  Fast	
  Red	
  chromogen,	
  hematoxylin	
  counterstain.	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

	
  

172	
  

	
  
Figure	
  27:	
  Photomicrographs	
  from	
  a	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  
xenograft	
  derived	
  from	
  the	
  TYLER2	
  cell	
  line	
  showing	
  two	
  morphologically	
  distinct	
  cell	
  
populations,	
  hematoxylin	
  and	
  eosin	
  stain.	
  Neoplastic	
  cell	
  populations	
  are	
  biphasic	
  with	
  
regions	
  showing	
  epithelioid	
  or	
  discrete	
  cell	
  morphology.	
  	
  In	
  the	
  lower	
  magnification	
  
photomicrograph	
  (A),	
  epithelioid	
  cells	
  are	
  arranged	
  in	
  nests	
  and	
  packets,	
  which	
  often	
  have	
  
central	
  cavitations	
  that	
  are	
  open	
  or	
  that	
  contain	
  flocculent	
  eosinophilic	
  fluid	
  and	
  sloughed	
  
degenerate	
  cells.	
  	
  Discrete	
  cells	
  are	
  arranged	
  in	
  dense	
  sheets	
  supported	
  by	
  scant	
  fine	
  
fibrovascular	
  stroma.	
  	
  In	
  the	
  higher	
  magnification	
  photomicrograph,	
  epithelioid	
  cells	
  are	
  
plump	
  polygonal,	
  and	
  have	
  distinct	
  cell	
  borders,	
  moderate	
  amounts	
  of	
  eosinophilic	
  
cytoplasm,	
  and	
  rare,	
  sharply	
  defined,	
  open	
  intracytoplasmic	
  vacuoles.	
  	
  Discrete	
  cells	
  are	
  
densely	
  packed,	
  but	
  loosely	
  arranged,	
  round,	
  and	
  have	
  scant	
  eosinophilic	
  cytoplasm	
  and	
  
variably	
  distinct	
  cell	
  borders.	
  	
  	
  

	
  
	
  
	
  
	
  
	
  

	
  

173	
  

	
  
Figure	
  28:	
  Photomicrographs	
  of	
  IHC	
  for	
  differentiation	
  markers	
  in	
  a	
  canine	
  lower	
  urinary	
  
tract	
  urothelial	
  carcinoma	
  xenograft	
  derived	
  from	
  the	
  KINSEY	
  cell	
  line.	
  Neoplastic	
  epithelial	
  
cells	
  have	
  strong	
  permembranous	
  to	
  cytoplasmic	
  immunoreactivity	
  E-­‐cadherin	
  (A),	
  
perimembranous	
  immunoreactivity	
  for	
  P-­‐cadherin	
  (B),	
  no	
  labeling	
  for	
  N-­‐cadherin	
  (C)	
  
confirming	
  an	
  epithelial	
  origin,	
  3,3’-­‐Diaminobenzidine	
  (DAB)	
  chromogen,	
  hematoxylin	
  
counterstain.	
  There	
  is	
  strong	
  perimembranous	
  immunoreactivity	
  for	
  cytokeratin	
  7	
  
throughout	
  the	
  neoplastic	
  cells	
  (D),	
  no	
  immunoreactivity	
  for	
  prostatic	
  acid	
  phosphatase	
  
(PAP)	
  (E),	
  which	
  is	
  consistent	
  with	
  an	
  urothelial	
  origin,	
  DAB	
  chromogen,	
  hematoxylin	
  
counterstain.	
  	
  Similar	
  to	
  what	
  was	
  observed	
  in	
  plated	
  cell	
  lines,	
  neoplastic	
  cells	
  have	
  
cytoplasmic	
  expression	
  of	
  vimentin	
  (F),	
  indirect	
  biotin	
  streptavidin	
  and	
  Fast	
  Red	
  
chromogen,	
  hematoxylin	
  counterstain.	
  

	
  
	
  
	
  
	
  

	
  

174	
  

	
  
Figure	
  29:	
  Photomicrographs	
  of	
  IHC	
  for	
  differentiation	
  markers	
  in	
  a	
  canine	
  lower	
  urinary	
  
tract	
  urothelial	
  carcinoma	
  xenograft	
  derived	
  from	
  the	
  TYLER2	
  cell	
  line.	
  	
  In	
  regions	
  of	
  
epithelioid	
  differentiation,	
  neoplastic	
  epithelial	
  cells	
  have	
  strong	
  permembranous	
  E-­‐
cadherin	
  (A),	
  P-­‐cadherin	
  (B),	
  N-­‐cadherin	
  (C),	
  and	
  Uroplakin	
  3	
  (D)	
  consistent	
  with	
  a	
  
urothelial	
  origin,	
  3,3’-­‐Diaminobenzidine	
  (DAB)	
  chromogen,	
  hematoxylin	
  counterstain.	
  
Discrete	
  cell	
  populations	
  were	
  not	
  immunoreactive	
  for	
  E-­‐cadherin	
  (A),	
  P-­‐cadherin	
  (B),	
  N-­‐
cadherin	
  (C),	
  or	
  Uroplakin	
  3	
  (D).	
  	
  There	
  was	
  rare	
  cytoplasmic	
  immunoractivity	
  for	
  prostatic	
  
acid	
  phosphatase	
  (E)	
  in	
  scattered	
  individual	
  cells	
  in	
  both	
  epitheliod	
  and	
  discrete	
  cell	
  
populations	
  (E),	
  suggesting	
  some	
  degree	
  of	
  prostatic	
  epithelial	
  differentiation,	
  DAB	
  
chromogen,	
  hematoxylin	
  counterstain.	
  	
  Similar	
  to	
  what	
  was	
  observed	
  in	
  plated	
  cell	
  lines,	
  
epitheloid	
  and	
  discrete	
  cells	
  have	
  cytoplasmic	
  expression	
  of	
  vimentin	
  (F),	
  indirect	
  biotin	
  
streptavidin	
  and	
  Fast	
  Red	
  chromogen,	
  hematoxylin	
  counterstain.	
  

	
  
	
  
	
  

	
  

175	
  

	
  
Figure	
  30:	
  Western	
  blots	
  show	
  relative	
  decreased	
  expression	
  of	
  MSH2	
  and	
  MSH6	
  in	
  the	
  
TYLER2	
  cell	
  line	
  in	
  comparison	
  to	
  that	
  of	
  ANGUS,	
  KINSEY,	
  and	
  TYLER1.	
  	
  Actin	
  expression	
  was	
  
evaluated	
  to	
  ensure	
  equal	
  protein	
  loading.	
  	
  	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

176	
  

	
  
Figure	
  31:	
  Graph	
  depicting	
  results	
  of	
  qPCR	
  for	
  MLH1,	
  MSH2,	
  MSH3,	
  MSH6,	
  and	
  PMS2	
  in	
  the	
  
canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines,	
  ANGUS,	
  KINSEY,	
  TYLER1,	
  and	
  
TYLER2	
  cell	
  lines.	
  	
  Columns	
  represent	
  relative	
  quantification	
  of	
  each	
  gene	
  as	
  determined	
  by	
  
the	
  ΔΔCt	
  method	
  with	
  evaluation	
  relative	
  to	
  a	
  compilation	
  of	
  β-­‐actin,	
  GAPDH,	
  and	
  HPRT1	
  
gene	
  expression	
  and	
  relative	
  to	
  expression	
  in	
  TYLER2.	
  Associated	
  bars	
  represent	
  standard	
  
deviation.	
  KINSEY,	
  ANGUS,	
  and	
  TYLER1	
  respectively	
  had	
  greater	
  than	
  10	
  fold,	
  5	
  fold,	
  and	
  4	
  
fold	
  mean	
  relative	
  expression	
  levels	
  of	
  MSH2	
  above	
  that	
  of	
  TYLER2.	
  	
  KINSEY	
  had	
  greater	
  
than	
  6	
  fold	
  mean	
  relative	
  expression	
  levels	
  and	
  TYLER2	
  and	
  ANGUS	
  had	
  greater	
  than	
  1	
  fold	
  
mean	
  relative	
  expression	
  of	
  MSH6	
  above	
  that	
  of	
  TYLER2.	
  There	
  were	
  less	
  than	
  a	
  fold	
  
differences	
  in	
  mean	
  relative	
  expression	
  of	
  MLH1,	
  MSH3,	
  and	
  PMS2	
  between	
  all	
  cell	
  lines.	
  	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

177	
  

	
  
Figure	
  31	
  (cont’d)	
  

	
  
	
  
	
  
	
  
	
  
	
  

	
  

178	
  

	
  
Figure	
  32:	
  Graph	
  depicting	
  results	
  of	
  qPCR	
  for	
  MSH2	
  and	
  MSH6	
  in	
  canine	
  lower	
  urinary	
  tract	
  
urothelial	
  carcinomas	
  xenografts	
  derived	
  from	
  TYLER1,	
  and	
  TYLER2	
  cell	
  lines	
  laser	
  capture	
  
microdissection	
  separated	
  from	
  mouse	
  tissues.	
  	
  Columns	
  represent	
  relative	
  quantification	
  of	
  
each	
  gene	
  as	
  determined	
  by	
  the	
  ΔΔCt	
  method	
  with	
  evaluation	
  relative	
  to	
  GAPDH	
  gene	
  
expression	
  and	
  relative	
  to	
  expression	
  in	
  TYLER2.	
  Associated	
  bars	
  represent	
  standard	
  
deviation.	
  Variance	
  in	
  the	
  mean	
  relative	
  quantification	
  of	
  gene	
  expression	
  is	
  depicted	
  by	
  
wide	
  standard	
  deviations.	
  The	
  mean	
  expression	
  of	
  MSH2	
  relative	
  to	
  GAPDH,	
  was	
  greater	
  
than	
  15	
  fold	
  higher	
  in	
  Kinsey	
  than	
  in	
  both	
  TYLER1	
  and	
  TYLER2.	
  	
  There	
  was	
  no	
  appreciable	
  
difference	
  in	
  the	
  mean	
  expression	
  of	
  MSH2	
  between	
  TYLER1	
  and	
  TYLER2.	
  	
  The	
  mean	
  
expression	
  of	
  MSH6	
  relative	
  to	
  GAPDH	
  was	
  greater	
  than	
  25	
  fold	
  higher	
  in	
  KINSEY	
  than	
  in	
  
TYLER2	
  and	
  greater	
  than	
  5	
  fold	
  greater	
  in	
  KINSEY	
  than	
  in	
  TYLER2.	
  	
  The	
  mean	
  relative	
  
expression	
  of	
  MSH6	
  in	
  TYLER2	
  was	
  greater	
  than	
  4	
  fold	
  higher	
  than	
  that	
  of	
  TYLER1.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

179	
  

	
  
Figure	
  32	
  (cont’d)	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

180	
  

	
  
	
  
Figure	
  33:	
  Graph	
  depicting	
  results	
  of	
  qPCR	
  for	
  MSH2	
  in	
  laser	
  capture	
  microdissection	
  
separated	
  epitheliod	
  and	
  discrete	
  cell	
  populations	
  of	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  
carcinomas	
  xenografts	
  derived	
  from	
  TYLER1,	
  and	
  TYLER2	
  cell	
  lines.	
  	
  Associated	
  bars	
  
represent	
  standard	
  deviation.	
  Columns	
  represent	
  mean	
  relative	
  quantification	
  of	
  MSH2	
  
expression	
  as	
  determined	
  using	
  the	
  ΔΔCt	
  method	
  with	
  evaluation	
  relative	
  to	
  GAPDH	
  gene	
  
expression	
  and	
  relative	
  to	
  expression	
  in	
  the	
  TYLER1	
  discrete	
  cell	
  population.	
  There	
  was	
  
over	
  a	
  fold	
  difference	
  in	
  the	
  expression	
  of	
  MSH2	
  in	
  TYLER1	
  epitheliod	
  and	
  TYLER2	
  
epitheliod	
  populations	
  in	
  comparison	
  to	
  that	
  of	
  the	
  TYLER1	
  and	
  TYLER2	
  discrete	
  cell	
  
populations,	
  respectively.	
  	
  	
  There	
  were	
  only	
  mild	
  differences	
  in	
  relative	
  MSH2	
  expression	
  
between	
  the	
  separately	
  isolated	
  epithelial	
  and	
  round	
  cell	
  populations	
  when	
  comparing	
  
TYLER1	
  and	
  TYLER2.	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

181	
  

	
  
Figure	
  33	
  (cont’d)	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

182	
  

	
  
Figure	
  34:	
  Graph	
  depicting	
  results	
  of	
  morphometric	
  analysis	
  of	
  multispectral	
  imaging	
  of	
  
immunoreactivity	
  of	
  MSH2	
  and	
  MSH6	
  in	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas	
  
xenografts	
  derived	
  from	
  KINSEY,	
  TYLER1,	
  and	
  TYLER2	
  cell	
  lines.	
  	
  For	
  TYLER1	
  and	
  TYLER2,	
  
areas	
  with	
  epithelioid	
  and	
  discrete	
  morphology	
  were	
  imaged	
  and	
  analyzed	
  separately.	
  
Quantitated	
  immunoreactivity	
  is	
  reported	
  as	
  positive	
  pixels	
  per	
  nucleus.	
  	
  Columns	
  
represent	
  mean	
  of	
  expression	
  of	
  MSH2	
  and	
  MSH6	
  between	
  multiple	
  xenografts	
  with	
  bars	
  
representing	
  standard	
  deviation	
  (KINSEY:	
  n=6;	
  TYLER1	
  epithelioid:	
  n-­‐4;	
  TYLER1	
  discrete:	
  
n=5;	
  TYLER2	
  epithelioid:	
  n=3;	
  TYLER2	
  discrete:	
  n=4).	
  Marked	
  variation	
  in	
  the	
  expression	
  
of	
  MSH2	
  and	
  moderate	
  variation	
  in	
  the	
  expression	
  of	
  MSH6	
  within	
  xenografts	
  derived	
  from	
  
the	
  same	
  cell	
  line	
  is	
  evidenced	
  by	
  large	
  standard	
  deviations.	
  The	
  mean	
  expression	
  levels	
  of	
  
MSH2	
  and	
  MSH6	
  in	
  terms	
  were	
  lower	
  in	
  discrete	
  cell	
  populations	
  than	
  in	
  epitheliod	
  
populations	
  for	
  both	
  TYLER1	
  and	
  TYLER2.	
  	
  Further,	
  the	
  mean	
  expression	
  levels	
  of	
  MSH2	
  
and	
  MSH6,	
  for	
  both	
  epithelioid	
  and	
  discrete	
  cell	
  populations	
  were	
  respectively	
  higher	
  in	
  
TYLER1	
  populations	
  than	
  in	
  TYLER2	
  populations.	
  	
  The	
  mean	
  expression	
  of	
  MSH6	
  was	
  
higher	
  in	
  KINSEY	
  in	
  comparison	
  to	
  that	
  of	
  TYLER2	
  cell	
  lines	
  and	
  discrete	
  TYLER1	
  cell	
  
populations,	
  but	
  was	
  not	
  different	
  from	
  that	
  of	
  epithelioid	
  portions	
  of	
  TYLER1.	
  	
  Expression	
  
of	
  MSH2	
  in	
  KINSEY	
  was	
  low	
  relative	
  to	
  that	
  observed	
  in	
  epithelioid	
  populations.	
  	
  	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

183	
  

	
  
Figure	
  34	
  (cont’d)	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

184	
  

	
  
	
  
Figure	
  35:	
  Results	
  of	
  XTT	
  survival	
  assays	
  comparing	
  survival	
  of	
  canine	
  urothelial	
  carcinoma	
  
cell	
  lines	
  upon	
  exposure	
  to	
  specific	
  chemotherapeutics.	
  With	
  regards	
  to	
  differences	
  in	
  
chemotherapeutic	
  sensitivity	
  for	
  TYLER1	
  and	
  TYLER2,	
  there	
  were	
  appreciable	
  differences	
  
in	
  sensitivity	
  to	
  carboplatin,	
  oxaliplatin,	
  methotrexate,	
  and	
  thiotepa.	
  	
  For	
  carboplatin,	
  there	
  
was	
  low	
  survival	
  of	
  TYLER1	
  relative	
  to	
  TYLER2	
  especially	
  at	
  higher	
  tested	
  concentrations;	
  
however,	
  TYLER2	
  was	
  more	
  sensitive	
  to	
  oxaliplatin	
  than	
  TYLER1.	
  	
  	
  TYLER1	
  was	
  less	
  
sensitive	
  than	
  TYLER1	
  to	
  moderate	
  concentraions	
  of	
  methotrexate	
  and	
  lower	
  
concentrations	
  of	
  thiotepa.	
  	
  ANGUS	
  often	
  was	
  less	
  sensitive	
  to	
  chemotherapeutics	
  relative	
  
to	
  other	
  evaluated	
  cell	
  lines	
  including	
  for	
  carboplatin,	
  cisplatin,	
  thiotepa,	
  and	
  gemcitabine.	
  	
  
Sensitivity	
  of	
  KINSEY	
  was	
  similar	
  to	
  that	
  of	
  TYLER1	
  for	
  carboplatin,	
  TYLER1	
  and	
  TYLER2	
  
for	
  cisplatin,	
  ANUGS	
  and	
  TYLER1	
  for	
  oxaliplatin,	
  TYLER2	
  for	
  methotrexate,	
  and	
  TYLER1	
  
and	
  TYLER2	
  for	
  gemcitabine.	
  	
  For	
  thiotepa,	
  there	
  were	
  differences	
  in	
  sensitivity	
  for	
  all	
  
tested	
  cell	
  lines,	
  especially	
  at	
  the	
  lowest	
  tested	
  concentrations,	
  with	
  TYLER2	
  being	
  most	
  
sensitive,	
  followed	
  by	
  TYLER2,	
  KINSEY	
  and	
  ANGUS.	
  	
  There	
  was	
  no	
  appreciable	
  difference	
  in	
  
sensitivity	
  of	
  TYLER1,	
  TYLER2,	
  or	
  KINSEY	
  for	
  Gemcitabine,	
  but	
  ANGUS	
  had	
  relative	
  
resistance.	
  	
  There	
  were	
  no	
  or	
  only	
  mild	
  differences	
  in	
  response	
  of	
  all	
  evaluated	
  cell	
  lines	
  to	
  
taxol,	
  cytarabine,	
  and	
  lomoustine.	
  	
  	
  
	
  
	
  
	
  
	
  
	
  

	
  

185	
  

	
  
Figure	
  35	
  (cont’d)	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

186	
  

	
  
	
  
Table	
  14:	
  TaqMan®	
  gene	
  expression	
  assays	
  used	
  for	
  qPCR	
  
Amplicon	
  
Gene	
  
length	
  
Species	
  
Number	
  
MLH1	
  
105	
  
Dog	
  
Cf02666410_m1	
  
Cf02626772_m1	
  
MSH2	
  
83	
  
Dog	
  
MSH3	
  
103	
  
Dog	
  
Cf02643735_m1	
  
MSH6	
  
77	
  
Dog	
  
Cf02641004_g1	
  
PMS2	
  
109	
  
Dog	
  
Cf02644577_m1	
  
HPRT1	
  
102	
  
Dog	
  
Cf02626256_m1	
  
β-­‐Actin	
  
139	
  
Human	
  
Hs03023880_g1	
  
	
  	
  
	
  	
  
	
  
	
  
Canine	
  GAPDH	
  custom	
  designed	
  TaqMan	
  gene	
  expression	
  
assay:	
  
Forward	
  Primer:	
  TCAACGGATTTGGCCGTATTGG	
  
Reverse	
  Primer:	
  TGAAGGGGTCATTGATGGCG	
  
FAM	
  labeled	
  probe:	
  CAGGGTGCTTTTAACTCTGGCAAAGTGGA	
  
	
  
	
  
	
  
	
  
	
  
Table	
  15:	
  Distribution	
  of	
  MSH2	
  expression	
  in	
  canine	
  urothelial	
  carcinoma	
  xenogafts	
  as	
  
evaluated	
  by	
  morphometric	
  analysis	
  of	
  immunohistochemistry	
  

Cell	
  line	
  xenograft	
  
KINSEY	
  
TYLER1	
  epithelioid	
  
TYLER1	
  discreet	
  
TYLER2	
  epithelioid	
  
TYLER2	
  discreet	
  

N	
  
6	
  
4	
  
5	
  
3	
  
4	
  

Mean	
  MSH2	
  expression	
  	
  	
  	
  	
  	
  	
  (±	
  
Standard	
  Deviation)	
  
267.1	
  (±111.9)	
  
1275.7	
  (±855.3)	
  
453.4	
  (±808.3)	
  
826.3	
  (±1318.3)	
  
128.0	
  (±212.1)	
  

	
  
	
  

	
  

187	
  

Minimum	
   Maximum	
  
127.3	
  
409.4	
  
564.6	
  
2544.7	
  
42.6	
  
1896.3	
  
24.2	
  
2347.7	
  
9.1	
  
445.7	
  

	
  
Table	
  16:	
  Distribution	
  of	
  MSH6	
  expression	
  in	
  canine	
  urothelial	
  carcinoma	
  xenogafts	
  as	
  
evaluated	
  by	
  morphometric	
  analysis	
  of	
  immunohistochemistry	
  
	
  

Cell	
  line	
  xenograft	
  
KINSEY	
  
TYLER1	
  epithelioid	
  
TYLER1	
  discreet	
  
TYLER2	
  epithelioid	
  
TYLER2	
  discreet	
  

N	
  
6	
  
4	
  
5	
  
3	
  
4	
  

Mean	
  MSH6	
  expression	
  	
  	
  	
  	
  	
  	
  	
  (±	
  
Standard	
  Deviation)	
  
796.2	
  (±	
  158.5)	
  
672.8	
  (±325.9)	
  
197.9	
  (±153.9)	
  
371.7	
  (±293.8)	
  
95.2	
  (±59.7)	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

188	
  

Minimum	
   Maximum	
  
643.7	
  
1128.5	
  
275.5	
  
1035.7	
  
77	
  
453.9	
  
54.3	
  
634.1	
  
38.9	
  
176.8	
  

	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
REFERENCES	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  
	
  

	
  

189	
  

	
  
REFERENCES	
  
	
  
1.	
  Catto	
  JW,	
  Azzouzi	
  AR,	
  Amira	
  N,	
  et	
  al.	
  Distinct	
  patterns	
  of	
  microsatellite	
  instability	
  are	
  
seen	
  in	
  tumours	
  of	
  the	
  urinary	
  tract.	
  Oncogene.	
  2003;22:	
  8699-­‐8706.	
  
2.	
  D'Errico	
  M,	
  de	
  Rinaldis	
  E,	
  Blasi	
  MF,	
  et	
  al.	
  Genome-­‐wide	
  expression	
  profile	
  of	
  sporadic	
  
gastric	
  cancers	
  with	
  microsatellite	
  instability.	
  European	
  journal	
  of	
  cancer.	
  2009;45:	
  461-­‐
469.	
  
3.	
  Hewish	
  M,	
  Lord	
  CJ,	
  Martin	
  SA,	
  Cunningham	
  D,	
  Ashworth	
  A.	
  Mismatch	
  repair	
  deficient	
  
colorectal	
  cancer	
  in	
  the	
  era	
  of	
  personalized	
  treatment.	
  Nature	
  reviews.	
  Clinical	
  oncology.	
  
2010;7:	
  197-­‐208.	
  
4.	
  Jacob	
  S,	
  Praz	
  F.	
  DNA	
  mismatch	
  repair	
  defects:	
  role	
  in	
  colorectal	
  carcinogenesis.	
  
Biochimie.	
  2002;84:	
  27-­‐47.	
  
5.	
  Mylona	
  E,	
  Zarogiannos	
  A,	
  Nomikos	
  A,	
  et	
  al.	
  Prognostic	
  value	
  of	
  microsatellite	
  instability	
  
determined	
  by	
  immunohistochemical	
  staining	
  of	
  hMSH2	
  and	
  hMSH6	
  in	
  urothelial	
  
carcinoma	
  of	
  the	
  bladder.	
  Acta	
  pathologica,	
  microbiologica	
  et	
  immunologica	
  Scandinavica.	
  
2008;116:	
  59-­‐65.	
  
6.	
  di	
  Pietro	
  M,	
  Sabates	
  Bellver	
  J,	
  Menigatti	
  M,	
  et	
  al.	
  Defective	
  DNA	
  mismatch	
  repair	
  
determines	
  a	
  characteristic	
  transcriptional	
  profile	
  in	
  proximal	
  colon	
  cancers.	
  
Gastroenterology.	
  2005;129:	
  1047-­‐1059.	
  
7.	
  Schwitalle	
  Y,	
  Kloor	
  M,	
  Eiermann	
  S,	
  et	
  al.	
  Immune	
  response	
  against	
  frameshift-­‐induced	
  
neopeptides	
  in	
  HNPCC	
  patients	
  and	
  healthy	
  HNPCC	
  mutation	
  carriers.	
  Gastroenterology.	
  
2008;134:	
  988-­‐997.	
  
8.	
  Casorelli	
  I,	
  Russo	
  MT,	
  Bignami	
  M.	
  Role	
  of	
  mismatch	
  repair	
  and	
  MGMT	
  in	
  response	
  to	
  
anticancer	
  therapies.	
  Anti-­‐cancer	
  agents	
  in	
  medicinal	
  chemistry.	
  2008;8:	
  368-­‐380.	
  
9.	
  Cejka	
  P,	
  Stojic	
  L,	
  Marra	
  G,	
  Jiricny	
  J.	
  Is	
  mismatch	
  repair	
  really	
  required	
  for	
  ionizing	
  
radiation-­‐induced	
  DNA	
  damage	
  signaling?	
  Nature	
  Genetics.	
  2004;36:	
  432-­‐433.	
  
10.	
  Flanagan	
  SA,	
  Robinson	
  BW,	
  Krokosky	
  CW,	
  Shewach	
  DS.	
  Mismatched	
  nucleotides	
  as	
  the	
  
lesions	
  responsible	
  for	
  radiosensitization	
  with	
  gemcitabine:	
  a	
  new	
  paradigm	
  for	
  
antimetabolite	
  radiosentizers.	
  Molecular	
  cancer	
  therapeutics.	
  2007;6:	
  1858-­‐1868.	
  
11.	
  Hart	
  JR,	
  Glebov	
  O,	
  Ernst	
  RJ,	
  Kirsch	
  IR,	
  Barton	
  JK.	
  DNA	
  mismatch-­‐specific	
  targeting	
  and	
  
hypersensitivity	
  of	
  mismatch-­‐repair-­‐deficient	
  cells	
  to	
  bulky	
  rhodium(III)	
  intercalators.	
  
Proceedings	
  of	
  the	
  National	
  Academy	
  of	
  Sciences	
  of	
  the	
  United	
  States	
  of	
  America.	
  2006;103:	
  
15359-­‐15363.	
  

	
  

190	
  

	
  
12.	
  Martin	
  SA,	
  Lord	
  CJ,	
  Ashworth	
  A.	
  Therapeutic	
  targeting	
  of	
  the	
  DNA	
  mismatch	
  repair	
  
pathway.	
  Clinical	
  cancer	
  research	
  :	
  an	
  official	
  journal	
  of	
  the	
  American	
  Association	
  for	
  
Cancer	
  Research.	
  2010;16:	
  5107-­‐5113.	
  
13.	
  Pors	
  K,	
  Patterson	
  LH.	
  DNA	
  mismatch	
  repair	
  deficiency,	
  resistance	
  to	
  cancer	
  
chemotherapy	
  and	
  the	
  development	
  of	
  hypersensitive	
  agents.	
  Current	
  topics	
  in	
  medicinal	
  
chemistry.	
  2005;5:	
  1133-­‐1149.	
  
14.	
  Sargent	
  DJ,	
  Marsoni	
  S,	
  Monges	
  G,	
  et	
  al.	
  Defective	
  mismatch	
  repair	
  as	
  a	
  predictive	
  marker	
  
for	
  lack	
  of	
  efficacy	
  of	
  fluorouracil-­‐based	
  adjuvant	
  therapy	
  in	
  colon	
  cancer.	
  Journal	
  of	
  clinical	
  
oncology	
  :	
  official	
  journal	
  of	
  the	
  American	
  Society	
  of	
  Clinical	
  Oncology.	
  2010;28:	
  3219-­‐
3226.	
  
15.	
  Takahashi	
  T,	
  Min	
  Z,	
  Uchida	
  I,	
  et	
  al.	
  Hypersensitivity	
  in	
  DNA	
  mismatch	
  repair-­‐deficient	
  
colon	
  carcinoma	
  cells	
  to	
  DNA	
  polymerase	
  reaction	
  inhibitors.	
  Cancer	
  letters.	
  2005;220:	
  85-­‐
93.	
  
16.	
  Valentini	
  AM,	
  Armentano	
  R,	
  Pirrelli	
  M,	
  Caruso	
  ML.	
  Chemotherapeutic	
  agents	
  for	
  
colorectal	
  cancer	
  with	
  a	
  defective	
  mismatch	
  repair	
  system:	
  the	
  state	
  of	
  the	
  art.	
  Cancer	
  
treatment	
  reviews.	
  2006;32:	
  607-­‐618.	
  
17.	
  Rader	
  KA.	
  Making	
  Mice:	
  Standardizing	
  Animals	
  for	
  American	
  Biomedical	
  Research	
  
1900-­‐1955.	
  Princeton,	
  NJ:	
  Princeton	
  University	
  Press,	
  2004.	
  
18.	
  Hahn	
  WC,	
  Counter	
  CM,	
  Lundberg	
  AS,	
  Beijersbergen	
  RL,	
  Brooks	
  MW,	
  Weinberg	
  RA.	
  
Creation	
  of	
  human	
  tumour	
  cells	
  with	
  defined	
  genetic	
  elements.	
  Nature.	
  1999;400:	
  464-­‐468.	
  
19.	
  Macleod	
  KF,	
  Jacks	
  T.	
  Insights	
  into	
  cancer	
  from	
  transgenic	
  mouse	
  models.	
  J	
  Pathol.	
  
1999;187:	
  43-­‐60.	
  
20.	
  Artandi	
  SE,	
  Chang	
  S,	
  Lee	
  SL,	
  et	
  al.	
  Telomere	
  dysfunction	
  promotes	
  non-­‐reciprocal	
  
translocations	
  and	
  epithelial	
  cancers	
  in	
  mice.	
  Nature.	
  2000;406:	
  641-­‐645.	
  
21.	
  Artandi	
  SE,	
  DePinho	
  RA.	
  Mice	
  without	
  telomerase:	
  what	
  can	
  they	
  teach	
  us	
  about	
  human	
  
cance?	
  Nat	
  Med.	
  2000;6:	
  852-­‐855.	
  
22.	
  Perkel	
  JM.	
  Telomeres	
  as	
  the	
  key	
  to	
  cancer:	
  could	
  hundreds	
  of	
  mouse	
  models	
  be	
  wrong?	
  
The	
  Scientist,	
  2002:38.	
  
23.	
  McKevitt	
  TP,	
  Nasir	
  L,	
  Devlin	
  P,	
  Argyle	
  DJ.	
  Telomere	
  lengths	
  in	
  dogs	
  decrease	
  with	
  
increasing	
  donor	
  age.	
  J	
  Nutr.	
  2002;132:	
  1604S-­‐1606S.	
  
24.	
  Nasir	
  L,	
  Devlin	
  P,	
  McKevitt	
  T,	
  Rutteman	
  G,	
  Argyle	
  DJ.	
  Telomere	
  lengths	
  and	
  telomerase	
  
activity	
  in	
  dog	
  tissues:	
  a	
  potential	
  model	
  system	
  to	
  study	
  human	
  telomere	
  and	
  telomerase	
  
biology.	
  Neoplasia.	
  2001;3:	
  351-­‐359.	
  

	
  

191	
  

	
  
25.	
  Bibby	
  MC.	
  Orthotopic	
  models	
  of	
  cancer	
  for	
  preclinical	
  drug	
  evaluation:	
  advantages	
  and	
  
disadvantages.	
  Eur	
  J	
  Cancer.	
  2004;40:	
  852-­‐857.	
  
26.	
  Hougton	
  PJ.	
  Human	
  tumor	
  xenografts	
  as	
  preclinical	
  models:	
  Value	
  and	
  limitations.	
  
American	
  Association	
  for	
  Cancer	
  Research	
  96th	
  Annual	
  Meeting.	
  Anaheim,	
  CA,	
  2005:33-­‐37.	
  
27.	
  Johnson	
  JI,	
  Decker	
  S,	
  Zaharevitz	
  D,	
  et	
  al.	
  Relationships	
  between	
  drug	
  activity	
  in	
  NCI	
  
preclinical	
  in	
  vitro	
  and	
  in	
  vivo	
  models	
  and	
  early	
  clinical	
  trials.	
  Br	
  J	
  Cancer.	
  2001;84:	
  1424-­‐
1431.	
  
28.	
  Izumi	
  Y,	
  di	
  Tomaso	
  E,	
  Hooper	
  A,	
  et	
  al.	
  Responses	
  to	
  antiangiogenesis	
  treatment	
  of	
  
spontaneous	
  autochthonous	
  tumors	
  and	
  their	
  isografts.	
  Cancer	
  Res.	
  2003;63:	
  747-­‐751.	
  
29.	
  Lindblad-­‐Toh	
  K,	
  Wade	
  CM,	
  Mikkelsen	
  TS,	
  et	
  al.	
  Genome	
  sequence,	
  comparative	
  analysis	
  
and	
  haplotype	
  structure	
  of	
  the	
  domestic	
  dog.	
  Nature.	
  2005;438:	
  803-­‐819.	
  
30.	
  Knapp	
  DW,	
  Glickman	
  NW,	
  Denicola	
  DB,	
  Bonney	
  PL,	
  Lin	
  TL,	
  Glickman	
  LT.	
  Naturally-­‐
occurring	
  canine	
  transitional	
  cell	
  carcinoma	
  of	
  the	
  urinary	
  bladder	
  A	
  relevant	
  model	
  of	
  
human	
  invasive	
  bladder	
  cancer.	
  Urologic	
  oncology.	
  2000;5:	
  47-­‐59.	
  
31.	
  Mutsaers	
  AJ,	
  Widmer	
  WR,	
  Knapp	
  DW.	
  Canine	
  transitional	
  cell	
  carcinoma.	
  Journal	
  of	
  
veterinary	
  internal	
  medicine	
  /	
  American	
  College	
  of	
  Veterinary	
  Internal	
  Medicine.	
  2003;17:	
  
136-­‐144.	
  
	
  
	
  

	
  

	
  

192	
  

	
  
CHAPTER	
  6	
  
	
  
	
  
Conclusions	
  
	
  
Dodd	
  Sledge	
  
	
  
Department	
  of	
  Pathobiology	
  and	
  Diagnostic	
  Investigation,	
  Diagnostic	
  Center	
  for	
  Population	
  
and	
  Animal	
  Health,	
  Michigan	
  State	
  University,	
  Lansing	
  MI	
  
	
  
	
  

	
  

	
  

193	
  

	
  
Summary	
  of	
  findings	
  
With	
  regards	
  to	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas,	
  these	
  studies	
  
highlight	
  the	
  potential	
  prognostic	
  significance	
  of	
  multiple	
  evaluated	
  markers,	
  emphasize	
  
potential	
  targets	
  for	
  directed	
  therapy,	
  and	
  highlight	
  similarities	
  between	
  dogs	
  and	
  humans	
  
in	
  multiple	
  carcinogenesis	
  pathways.	
  However,	
  these	
  studies	
  also	
  highlight	
  significant	
  
differences	
  between	
  urothelial	
  carcinomas	
  of	
  the	
  dog	
  and	
  human.	
  As	
  such,	
  these	
  findings	
  
contribute	
  to	
  the	
  field	
  of	
  canine	
  cancer	
  prognostics	
  and	
  give	
  hope	
  to	
  improved,	
  targeted	
  
treatment	
  for	
  bladder	
  cancer	
  in	
  dogs,	
  but	
  indicate	
  that	
  further	
  study	
  is	
  needed.	
  As	
  an	
  
animal	
  model,	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinomas	
  provide	
  an	
  interesting,	
  
although	
  not	
  perfect,	
  parallel	
  to	
  such	
  disease	
  in	
  humans.	
  	
  	
  
Histologic	
  classification	
  and	
  grading	
  form	
  cornerstones	
  of	
  routine	
  tumor	
  diagnostics	
  
and	
  prognostication;	
  however,	
  the	
  recently	
  proposed	
  classification	
  and	
  grading	
  scheme	
  for	
  
canine	
  proliferative	
  urothelial	
  lesions	
  based	
  on	
  the	
  World	
  Health	
  Organization	
  
(WHO)/International	
  Society	
  of	
  Urologic	
  Pathology	
  (ISUP)	
  consensus	
  system	
  accepted	
  in	
  
humans	
  had	
  not	
  previously	
  been	
  evaluated	
  with	
  respect	
  to	
  prognostic	
  relevance.	
  	
  This	
  
work	
  shows	
  that	
  there	
  are	
  biologic	
  differences	
  between	
  histologic	
  classifications	
  and	
  
grades	
  as	
  evidenced	
  by	
  differential	
  expression	
  of	
  differentiation	
  molecules	
  and	
  factors	
  
reflecting	
  specific	
  carcinogenesis	
  pathways.	
  	
  Specifically,	
  there	
  were	
  differences	
  in	
  
expression	
  of	
  uroplakin	
  III,	
  cytokeratin	
  7,	
  cyclooxygenase-­‐2	
  (COX-­‐2),	
  P-­‐cadherin,	
  and	
  β-­‐
catenin	
  between	
  classifications	
  of	
  proliferative	
  urothelial	
  lesions.	
  Differentiation	
  of	
  
urothelial	
  tumors	
  as	
  papillary	
  or	
  infiltrating	
  did	
  correlate	
  with	
  survival	
  time;	
  however,	
  
degree	
  of	
  invasion	
  and	
  histologic	
  grading	
  based	
  on	
  degree	
  of	
  anaplasia	
  did	
  not.	
  	
  

	
  

194	
  

	
  
Based	
  on	
  the	
  common	
  findings	
  of	
  invasion	
  of	
  carcinomas	
  into	
  the	
  bladder	
  wall	
  and	
  
metastasis,	
  it	
  is	
  likely	
  that	
  epithelial-­‐to-­‐mesenchymal	
  transition	
  occurs	
  in	
  canine	
  urothelial	
  
carcinomas.	
  Our	
  studies	
  suggest	
  that	
  the	
  mechanisms	
  underlying	
  such	
  transition	
  in	
  dogs	
  
differ	
  from	
  that	
  reported	
  in	
  humans.	
  In	
  humans,	
  correlations	
  exist	
  between	
  the	
  loss	
  of	
  
expression	
  of	
  15-­‐hydroxyprostaglandin	
  dehydrogenase	
  (HPGD)	
  and	
  cadherin	
  switching.	
  	
  
This	
  does	
  not	
  appear	
  to	
  be	
  exactly	
  the	
  case	
  in	
  dogs	
  as	
  decreased	
  expression	
  of	
  HPGD	
  
relative	
  to	
  that	
  of	
  normal	
  urothelium	
  was	
  common,	
  but	
  changes	
  in	
  cadherin	
  expression	
  
were	
  not	
  similar	
  to	
  that	
  described	
  in	
  humans.	
  	
  There	
  were	
  differences	
  in	
  the	
  expression	
  of	
  
P-­‐cadherin	
  in	
  canine	
  urothelial	
  carcinomas,	
  with	
  loss	
  or	
  aberrant	
  localization	
  of	
  P-­‐cadherin	
  
being	
  most	
  common	
  in	
  higher	
  grade,	
  invasive	
  carcinomas	
  and	
  being	
  associated	
  with	
  
survival	
  time.	
  	
  This	
  contrasts	
  with	
  the	
  increased	
  P-­‐cadherin	
  expression	
  described	
  with	
  
epithelial-­‐to-­‐mesenchymal	
  transition	
  in	
  humans.	
  	
  Further,	
  while	
  N-­‐cadherin	
  expression	
  is	
  
reported	
  with	
  epithelial-­‐to-­‐mesenchymal	
  transition	
  in	
  human	
  urothelial	
  carcinomas	
  and	
  
was	
  observed	
  in	
  human	
  urothelial	
  carcinoma	
  cell	
  lines	
  lacking	
  HPGD,	
  no	
  expression	
  of	
  N-­‐
cadherin	
  was	
  observed	
  in	
  any	
  of	
  the	
  examined	
  canine	
  urothelial	
  carcinomas.	
  This	
  suggests	
  
that	
  while	
  overall	
  histomorphology	
  and	
  clinical	
  progression	
  of	
  human	
  and	
  canine	
  are	
  
similar,	
  the	
  mechanisms	
  that	
  govern	
  invasion	
  and	
  metastasis	
  at	
  the	
  molecular	
  level	
  are	
  
different.	
  	
  
Microsatellite	
  instability	
  (MSI)	
  was	
  found	
  in	
  a	
  significant	
  proportion	
  of	
  canine	
  
urothelial	
  carcinomas	
  and	
  was	
  associated	
  with	
  genetic	
  background	
  in	
  terms	
  of	
  breed	
  and	
  
phylogenetic	
  clade	
  of	
  affected	
  dogs.	
  	
  This	
  finding	
  alone	
  makes	
  evaluation	
  of	
  MSI	
  and	
  the	
  
possible	
  heritable	
  basis	
  of	
  MSI	
  attractive	
  for	
  further	
  study.	
  	
  High	
  MSI,	
  however,	
  was	
  not	
  
associated	
  with	
  clinical	
  outcome	
  in	
  terms	
  of	
  survival	
  time	
  and	
  was	
  not	
  specifically	
  

	
  

195	
  

	
  
correlated	
  with	
  subsequent	
  evaluation	
  of	
  DNA	
  mismatch	
  repair	
  (MMR)	
  protein	
  expression.	
  	
  	
  
Differences	
  in	
  MMR	
  capacity	
  were,	
  however,	
  associated	
  with	
  variable	
  response	
  to	
  
chemotherapeutic	
  sensitivity	
  in	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines.	
  	
  Specifically,	
  the	
  
TYLER2	
  cell	
  line	
  had	
  decreased	
  expression	
  of	
  MSH2	
  and	
  MSH6,	
  resistance	
  to	
  carboplatin,	
  
and	
  increased	
  sensitivity	
  to	
  oxaliplatin,	
  methotrexate,	
  and	
  thiotepa	
  relative	
  to	
  TYLER1,	
  
which	
  was	
  derived	
  from	
  the	
  same	
  primary	
  tumor	
  and	
  considered	
  MMR	
  proficient.	
  	
  Such	
  
differences	
  in	
  response	
  to	
  treatment	
  in	
  vitro	
  suggest	
  that	
  there	
  are	
  likely	
  similar	
  differences	
  
between	
  canine	
  urothelial	
  carcinomas	
  in	
  vivo	
  and	
  that	
  MMR	
  capacity	
  should	
  be	
  taken	
  into	
  
account	
  with	
  regards	
  to	
  choice	
  of	
  therapeutic.	
  	
  
Given	
  the	
  preponderance	
  of	
  MMR	
  deficiencies	
  in	
  both	
  hereditary	
  and	
  spontaneous	
  
human	
  cancers	
  and	
  the	
  clear	
  differences	
  reported	
  in	
  therapeutic	
  response,	
  there	
  is	
  clearly	
  a	
  
need	
  to	
  optimize	
  treatment	
  plans	
  and	
  therapeutic	
  options	
  with	
  respect	
  to	
  MMR	
  capacity.	
  
Relevant	
  and	
  reproducible	
  models	
  for	
  study	
  of	
  MMR	
  and	
  the	
  effects	
  of	
  loss	
  of	
  MMR	
  function	
  
on	
  response	
  to	
  treatment	
  are	
  lacking.	
  Based	
  on	
  our	
  data,	
  MMR	
  is	
  frequently,	
  though	
  not	
  
exclusively,	
  defective	
  in	
  canine	
  urothelial	
  carcinomas.	
  As	
  such,	
  urothelial	
  carcinomas	
  in	
  
dogs	
  are	
  likely	
  to	
  provide	
  such	
  a	
  model	
  for	
  more	
  generalized	
  study	
  of	
  MMR.	
  
	
  
	
  

	
  

	
  

196	
  

	
  
Limitations	
  of	
  studies	
  and	
  unanswered	
  questions	
  
The	
  major	
  limitation	
  of	
  chapter	
  1	
  was	
  the	
  lack	
  of	
  clinical	
  follow-­‐up	
  data.	
  	
  This	
  was	
  a	
  
retrospective	
  study	
  and	
  cases	
  evaluated	
  were	
  retrieved	
  from	
  archives	
  of	
  routine	
  diagnostic	
  
specimens.	
  	
  As	
  such,	
  provided	
  clinico-­‐demographic	
  information	
  was	
  of	
  varying	
  quality	
  and	
  
completeness,	
  and	
  no	
  information	
  regarding	
  treatment,	
  clinical	
  progression,	
  or	
  clinical	
  
outcome	
  was	
  available.	
  Thus,	
  while	
  it	
  was	
  possible	
  to	
  compare	
  and	
  correlate	
  expression	
  of	
  
the	
  evaluated	
  immunohistochemical	
  markers	
  (uroplakin	
  III,	
  cytokeratin	
  7,	
  COX-­‐2	
  and	
  
caspase-­‐3)	
  to	
  histomorphologic	
  features	
  that	
  are	
  often	
  associated	
  with	
  prognosis	
  in	
  
humans	
  and	
  in	
  canine	
  other	
  cancers,	
  the	
  exact	
  prognostic	
  significance	
  of	
  differences	
  in	
  both	
  
histomorphologic	
  features	
  and	
  evaluated	
  immunohistochemical	
  markers	
  is	
  unclear.	
  COX-­‐2	
  
expression,	
  at	
  least,	
  was	
  further	
  examined	
  in	
  chapter	
  2	
  and	
  no	
  correlations	
  were	
  found	
  
between	
  expression	
  and	
  survival	
  time;	
  however,	
  the	
  grading	
  of	
  methods	
  and	
  means	
  of	
  
evaluation	
  of	
  COX-­‐2	
  expression	
  was	
  different	
  between	
  chapters	
  1	
  and	
  2.	
  
In	
  studies	
  in	
  which	
  survival	
  time	
  was	
  available	
  (chapters	
  2	
  and	
  3),	
  only	
  urothelial	
  
carcinomas	
  and	
  normal	
  urinary	
  bladders	
  were	
  evaluated,	
  and	
  there	
  were	
  particularly	
  low	
  
numbers	
  of	
  infiltrating	
  and	
  low	
  grade	
  papillary	
  urothelial	
  carcinomas.	
  	
  While	
  differences	
  
were	
  found	
  in	
  HPGD,	
  P-­‐cadherin,	
  and	
  β-­‐catenin	
  expression	
  and	
  in	
  microsatellite	
  instability	
  
between	
  urothelial	
  carcinomas	
  and	
  normal	
  urothelium,	
  it	
  is	
  unclear	
  if	
  such	
  findings	
  would	
  
be	
  discriminatory	
  between	
  urothelial	
  carcinomas	
  and	
  other	
  proliferative	
  urothelial	
  lesions	
  
or	
  inflammatory	
  disease	
  processes	
  affecting	
  the	
  lower	
  urinary	
  tract.	
  	
  However,	
  the	
  lack	
  of	
  
proliferative	
  urothelial	
  lesions	
  besides	
  urothelial	
  carcinomas	
  and	
  the	
  low	
  number	
  of	
  
infiltrating	
  and	
  low	
  grade	
  papillary	
  carcinomas	
  is	
  largely	
  reflective	
  of	
  the	
  types	
  of	
  cases	
  

	
  

197	
  

	
  
that	
  present	
  to	
  a	
  veterinary	
  referral	
  clinic	
  and	
  the	
  low	
  rate	
  of	
  low	
  grade	
  or	
  infiltrating	
  
carcinomas	
  in	
  the	
  overall	
  canine	
  population.	
  
Survival	
  time	
  from	
  the	
  date	
  of	
  diagnosis	
  was	
  the	
  major	
  indicator	
  of	
  prognosis	
  
evaluated	
  in	
  these	
  studies.	
  While	
  overall	
  survival	
  time	
  can	
  be	
  a	
  relevant	
  prognostic	
  
indicator,	
  it	
  can	
  be	
  skewed	
  by	
  many	
  factors.	
  	
  In	
  these	
  studies,	
  it	
  is	
  unclear	
  how	
  long	
  tumors	
  
may	
  have	
  been	
  present	
  before	
  diagnosis,	
  treatment	
  was	
  not	
  consistent,	
  information	
  
regarding	
  clinical	
  progression	
  was	
  not	
  available,	
  and	
  the	
  extent	
  of	
  disease	
  progression	
  at	
  
the	
  time	
  of	
  death	
  was	
  unclear.	
  	
  In	
  contrast	
  to	
  humans,	
  diagnosis	
  of	
  bladder	
  cancer	
  in	
  dogs	
  is	
  
often	
  not	
  made	
  until	
  late	
  into	
  disease	
  progression	
  when	
  tumors	
  cause	
  mechanical	
  problems	
  
or	
  when	
  systemic	
  disease	
  becomes	
  apparent.	
  Access	
  to	
  complete	
  clinical	
  follow-­‐up	
  was	
  
limited	
  as	
  most	
  cases	
  that	
  had	
  significant	
  survival	
  time	
  were	
  referred	
  back	
  to	
  primary	
  care	
  
veterinarians.	
  	
  Further,	
  monitoring	
  of	
  disease	
  progression	
  in	
  terms	
  of	
  changes	
  in	
  tumor	
  
size,	
  invasion,	
  or	
  development	
  of	
  metastasis	
  was	
  inconsistent	
  or	
  incomplete,	
  and	
  results	
  
were	
  often	
  not	
  available.	
  The	
  reason	
  for	
  death	
  in	
  most	
  of	
  these	
  animals	
  was	
  reported	
  to	
  be	
  
related	
  to	
  progressive	
  disease	
  related	
  to	
  the	
  urothelial	
  carcinomas.	
  In	
  most	
  cases,	
  however,	
  
necropsies	
  were	
  not	
  preformed	
  to	
  document	
  extent	
  of	
  tumor	
  growth	
  or	
  metastasis,	
  and	
  
death	
  was	
  not	
  by	
  natural	
  causes,	
  but	
  rather	
  by	
  elective	
  euthanasia.	
  The	
  choice	
  of	
  
euthanasia	
  is	
  a	
  complicated	
  issue	
  with	
  respect	
  to	
  statistical	
  evaluation	
  of	
  survival	
  in	
  canine	
  
cancer.	
  	
  This	
  is	
  especially	
  true	
  given	
  that	
  there	
  is	
  significant	
  variation	
  in	
  the	
  point	
  of	
  disease	
  
progression	
  at	
  which	
  the	
  decision	
  to	
  euthanize	
  is	
  made.	
  	
  In	
  lower	
  urinary	
  tract	
  diseases,	
  the	
  
choice	
  for	
  euthanasia	
  may	
  be	
  related	
  to	
  local	
  disease,	
  and	
  not	
  necessarily	
  metastasis	
  or	
  
other	
  systemic	
  progression.	
  Further,	
  in	
  veterinary	
  medicine,	
  home	
  care	
  considerations	
  and	
  
financial	
  impact	
  of	
  treatment	
  also	
  play	
  large	
  roles	
  in	
  the	
  decision	
  to	
  euthanize.	
  	
  Euthanasia,	
  

	
  

198	
  

	
  
however,	
  is	
  an	
  unavoidable	
  confounding	
  factor	
  in	
  analysis	
  of	
  cancer	
  related	
  statistics	
  in	
  
veterinary	
  medicine.	
  
The	
  exact	
  significance	
  of	
  deficiency	
  in	
  mismatch	
  repair	
  as	
  evaluated	
  in	
  chapter	
  3	
  
remains	
  unclear.	
  	
  While	
  microsatellite	
  instability	
  was	
  demonstrated	
  in	
  a	
  number	
  of	
  canine	
  
urothelial	
  carcinomas	
  and	
  was	
  associated	
  with	
  breed,	
  there	
  was	
  no	
  clear	
  association	
  of	
  MSI	
  
to	
  MMR	
  protein	
  expression	
  as	
  evaluated	
  by	
  immunohistochemistry.	
  	
  The	
  evaluation	
  of	
  
MMR	
  proficiency	
  is	
  complicated	
  even	
  in	
  human	
  medicine	
  where	
  it	
  has	
  been	
  extensively	
  
used.	
  Variable	
  recommendations	
  regarding	
  evaluation	
  immunohistochemcial	
  of	
  MMR	
  
protein	
  expression	
  and	
  determination	
  of	
  MSI	
  exist.	
  	
  The	
  reagents	
  used	
  for	
  evaluation	
  in	
  all	
  
of	
  these	
  studies	
  were	
  not	
  specifically	
  designed	
  for	
  use	
  in	
  dogs.	
  	
  This	
  is	
  particularly	
  true	
  of	
  
the	
  antibodies	
  used	
  to	
  assess	
  protein	
  expression	
  by	
  IHC.	
  	
  While	
  we	
  were	
  able	
  to	
  show	
  that	
  
the	
  antibodies	
  did	
  label	
  appropriately	
  sized	
  molecules	
  in	
  Western	
  blot,	
  aberrant	
  labeling	
  
cannot	
  be	
  completely	
  excluded	
  as	
  a	
  confounder.	
  	
  The	
  role	
  of	
  decreased	
  MMR	
  protein	
  
expression	
  rather	
  than	
  total	
  loss	
  is	
  unclear.	
  Further,	
  the	
  morphometric	
  methods	
  of	
  
evaluating	
  overall	
  MMR	
  protein	
  expression	
  differ	
  from	
  that	
  described	
  in	
  most	
  human	
  
studies.	
  	
  
In	
  canine	
  lower	
  urinary	
  tract	
  urothelial	
  carcinoma	
  cell	
  lines,	
  the	
  functional	
  significance	
  
of	
  decreased	
  relative	
  expression	
  of	
  MSH2	
  and	
  MSH6	
  is	
  unclear.	
  Evaluation	
  of	
  the	
  status	
  of	
  
the	
  MMR	
  system	
  of	
  evaluated	
  cell	
  lines	
  was	
  largely	
  inferred	
  from	
  evaluation	
  of	
  MMR	
  gene	
  
and	
  protein	
  expression	
  and	
  comparison	
  between	
  cell	
  lines.	
  	
  It	
  is	
  less	
  clear	
  to	
  what	
  degree	
  
the	
   relative	
   differences	
   in	
   MMR	
   gene	
   and	
   protein	
   expression	
   had	
   on	
   the	
   overall	
   function	
   of	
  
the	
   MMR	
   system.	
   Outside	
   of	
   the	
   studies	
   reported	
   in	
   this	
   dissertation,	
   MSI	
   was	
   evaluated	
   in	
  
ANGUS	
  and	
  the	
  initial	
  cultures	
  of	
  TYLER	
  using	
  the	
  methods	
  described	
  in	
  chapter	
  3.	
  ANGUS	
  

	
  

199	
  

	
  
did	
   not	
   exhibit	
   instability	
   in	
   any	
   of	
   the	
   microsatellites	
   evaluated.	
   In	
   contrast,	
   there	
   was	
  
variability	
   in	
   the	
   length	
   of	
   select	
   microsatellites	
   comparing	
   DNA	
   of	
   the	
   TYLER	
   cell	
   line	
   to	
  
that	
  of	
  DNA	
  of	
  blood	
  samples	
  collected	
  from	
  the	
  dog	
  from	
  which	
  TYLER	
  was	
  derived.	
  This	
  
does	
   suggest	
   some	
   functional	
   deficiency	
   microsatellite	
   instability	
   in	
   TYLER.	
   	
   However,	
  
differences	
   in	
   TYLER1	
   and	
   TYLER2	
   with	
   respect	
   MMR	
   functional	
   capacity	
   have	
   not	
   been	
  
specifically	
  assessed.	
  	
  	
  
While	
  the	
  TYLER	
  cell	
  lines	
  were	
  derived	
  from	
  the	
  same	
  primary	
  urothelial	
  
carcinoma,	
  it	
  is	
  likely	
  that	
  these	
  cell	
  lines	
  differ	
  not	
  only	
  in	
  MMR	
  gene	
  and	
  protein	
  
expression.	
  This	
  is	
  suggested	
  by	
  the	
  marked	
  phenotypic	
  variation	
  in	
  cell	
  culture	
  and	
  the	
  
production	
  of	
  biphasic	
  cell	
  populations	
  in	
  xenografts.	
  	
  These	
  cell	
  lines	
  were	
  derived	
  from	
  
primary	
  culture	
  and	
  as	
  such	
  are	
  likely	
  composed	
  of	
  a	
  heterogeneous	
  cell	
  population.	
  While	
  
such	
  a	
  heterogeneous	
  cell	
  population	
  in	
  cell	
  culture	
  may	
  more	
  closely	
  resemble	
  the	
  natural	
  
heterogeneity	
  of	
  cancer	
  in	
  vivo,	
  it	
  complicates	
  determination	
  of	
  the	
  effects	
  of	
  specific	
  
differences	
  in	
  in	
  vitro	
  studies.	
  	
  In	
  addition,	
  if	
  MMR	
  deficiencies	
  in	
  TYLER2	
  are	
  functional,	
  an	
  
increased	
  rate	
  of	
  mutation	
  and	
  an	
  associated	
  build	
  up	
  of	
  secondary	
  mutations	
  over	
  
numerous	
  passages	
  in	
  cell	
  culture	
  is	
  likely.	
  	
  The	
  reason	
  for	
  the	
  marked	
  variation	
  in	
  cell	
  
morphology	
  between	
  TYLER1	
  and	
  TYLER2	
  is	
  unknown,	
  but	
  is	
  likely	
  not	
  determined	
  only	
  
by	
  differences	
  in	
  MMR.	
  	
  As	
  such,	
  it	
  cannot	
  be	
  excluded	
  that	
  differences	
  in	
  sensitivity	
  to	
  
chemotherapeutics	
  observed	
  between	
  TYLER1	
  and	
  TYLER2	
  were	
  not	
  related	
  to	
  specifically	
  
to	
  differences	
  in	
  MMR	
  proficiency.	
  Further,	
  even	
  though	
  differences	
  in	
  response	
  to	
  
chemotherapeutics	
  were	
  observed	
  in	
  cell	
  culture,	
  this	
  in	
  vitro	
  response	
  may	
  not	
  be	
  
reflected	
  in	
  vivo.	
  
	
  

	
  

200	
  

	
  
Future	
  directions	
  
Larger	
   and	
   more	
   in	
   depth	
   prospective	
   studies	
   are	
   needed	
   to	
   more	
   fully	
   evaluate	
   the	
  
prognostic	
  significance	
  of	
  histologic	
  classification	
  and	
  grading.	
  Such	
  studies	
  should	
  ideally	
  
include	
  a	
  larger	
  representative	
  group	
  of	
  low	
  grade	
  and	
  infiltrating	
  carcinomas.	
  Such	
  studies	
  
should	
   also	
   include	
   careful	
   monitoring	
   to	
   evaluate	
   prognostic	
   significance	
   in	
   terms	
   of	
  
change	
   in	
   tumor	
   size,	
   change	
   in	
   stage,	
   and	
   time	
   to	
   metastasis.	
   Based	
   on	
   the	
   biologic	
  
differences	
   observed	
   in	
   tumor	
   classification	
   and	
   grade,	
   it	
   is	
   likely	
   that	
   prognostic	
  
significance	
  exists,	
  but	
  was	
  not	
  observed	
  in	
  the	
  studied	
  populations	
  of	
  this	
  dissertation	
  as	
  
only	
  survival	
  time	
  from	
  diagnosis	
  was	
  examined	
  and	
  there	
  were	
  low	
  sample	
  numbers,	
  with	
  
low	
  grade	
  and	
  infiltrating	
  carcinomas	
  being	
  particularly	
  underrepresented.	
  
The	
   mechanisms	
   that	
   underlie	
   epithelial-­‐to-­‐mesenchymal	
   transition	
   in	
   canine	
  
urothelial	
   carcinomas	
   remain	
   unclear	
   as	
   does	
   the	
   significance	
   of	
   loss	
   of	
   HPGD.	
   It	
   is	
   clear	
  
that	
   epithelial-­‐to-­‐mesenchymal	
   transition	
   in	
   dog	
   urothelial	
   carcinomas	
   is	
   different	
   from	
  
that	
   in	
   humans.	
   Comparative	
   studies	
   evaluating	
   other	
   features	
   of	
   epithelial-­‐to-­‐
mesenchymal	
   transition	
   are	
   needed	
   to	
   further	
   characterize	
   similarities	
   and	
   differences.	
  
Lack	
  of	
  HPGD	
  expression	
  was	
  seen	
  in	
  all	
  evaluated	
  canine	
  urothelial	
  carcinoma	
  cell	
  lines,	
  
and	
   while	
   there	
   was	
   some	
   degree	
   of	
   difference	
   in	
   P-­‐cadherin	
   expression,	
   loss	
   or	
   up	
  
regulation	
   was	
   not	
   often	
   observed.	
   	
   Knock	
   in	
   studies	
   for	
   HPGD	
   in	
   cell	
   lines	
   along	
   with	
  
evaluation	
  of	
  changes	
  in	
  cadherin	
  expression	
  and	
  comparative	
  invasion	
  assays	
  would	
  serve	
  
to	
  further	
  determine	
  the	
  functional	
  consequences	
  of	
  loss	
  of	
  HPGD.	
  	
  
It	
  is	
  unclear	
  if	
  MMR	
  deficiencies	
  play	
  a	
  role	
  in	
  development	
  of	
  other	
  types	
  of	
  cancers	
  
in	
  dogs.	
  Based	
  on	
  the	
  results	
  of	
  immunohistochemical	
  MMR	
  protein	
  evaluation	
  in	
  bladder	
  
tumors,	
   expression	
   levels	
   of	
   MMR	
   protein	
   may	
   not	
   be	
   well	
   correlated	
   with	
   MSI	
   given	
   the	
  

	
  

201	
  

	
  
methods	
  by	
  which	
  MSI	
  was	
  evaluated	
  in	
  these	
  studies.	
  	
  It	
  appears	
  that	
  lack	
  of	
  MSI,	
  at	
  least	
  
within	
   the	
   MS	
   evaluated	
   in	
   our	
   studies,	
   does	
   not	
   preclude	
   that	
   differences	
   or	
   at	
   least	
  
variation	
  in	
  MMR	
  may	
  be	
  present.	
  Given	
  the	
  fact	
  that	
  relative	
  differences	
  in	
  expression	
  of	
  
MMR	
   protein	
   expression	
   as	
   assessed	
   by	
   IHC	
   and	
   morphometric	
   analysis	
   of	
  
immunoreactivity	
   did	
   not	
   correlate	
   with	
   MSI,	
   it	
   would	
   be	
   of	
   interest	
   to	
   evaluate	
   MMR	
  
protein	
  expression	
  in	
  other	
  tumors	
  including	
  gastric	
  carcinomas	
  and	
  mammary	
  tumors,	
  for	
  
which	
   it	
   was	
   shown	
   that	
   MSI	
   is	
   uncommon.	
   It	
   is	
   likely	
   that	
   variance	
   in	
   MMR	
   protein	
  
expression	
  is	
  common	
  in	
  canine	
  cancer	
  and	
  possibly	
  normal	
  tissues.	
  	
  	
  
The long-term goal of establishing differences in response to chemotherapeutics with
respect to MMR status is to be able to make treatment selection for an individual patient based
on MMR system evaluation. The	
   results	
   of	
   chemotherapeutic	
   sensitivity	
   testing	
   in	
   canine	
  
urothelial	
   carcinomas	
   is	
   promising,	
   but	
   the	
   exact	
   role	
   of	
   MMR	
   proficiency	
   and	
   the	
   exact	
  
implication	
  these	
  data	
  have	
  in	
  animals	
  in	
  vivo	
  remains	
  unclear.	
  In	
  order	
  to	
  better	
  isolate	
  the	
  
effects	
  that	
  MMR	
  proficiency	
  have	
  on	
  treatment	
  response	
  in	
  vitro,	
  knockdown	
  and	
  knock-­‐in	
  
studies	
   would	
   be	
   required.	
   	
   TYLER1	
   and	
   TYLER2	
   offer	
   an	
   excellent	
   opportunity	
   for	
   such	
  
studies	
   due	
   to	
   their	
   common	
   origin,	
   but	
   differential	
   MMR	
   capacity	
   and	
   demonstrated	
  
differences	
   in	
   chemosensitivity.	
   In order to separate the effects of MMR deficiency form other
potential sources of treatment response differences, MSH2 could be selectively knocked down in
TYLER1 and other MMR proficient cell lines. Concurrent knock in studies could also be
preformed in TYLER2 to increase the expression of MSH2 to the comparative levels of the other
evaluated cell lines. Subsequent chemotherapeutic survival studies comparing manipulated cell
lines to wild types would better distinguish differences in treatment response specifically
associated with altered MMR expression.

	
  

202	
  

	
  
Further,	
  conditions	
  in cell culture vary greatly from in vivo conditions as such response
of cancerous cells to treatment may differ in each of these settings. Treatment dosages used in
cell culture systems may not be achievable or may be ineffective in live animals. Drug
concentrations that are applicable to cell cultures may cause toxicity in live animals or may not
achieve sufficient penetrance of the target tissues. Xenografts of derived from cell lines could be
evaluated in terms of in vivo response to treatment. The long-term goal for such studies would
be to support clinical studies in naturally occurring bladder cancer dogs evaluating the response
of tumors to standard chemotherapy with respect to the presence MSI and alterations in MMR
protein expression as evaluated by IHC in routine diagnostic biopsies.	
  
As a whole, the studies detailed in this dissertation suggest that many pathways are
involved in carcinogenesis of canine urothelial tumors. These include prostaglandin regulation
pathways, epithelial-to-mesenchymal transition, Wnt signaling, and DNA mismatch repair.
However, it is unclear if and how the observed differences relate to one another or what other
carcinogenesis related pathways may be altered in canine urothelial carcinomas. Carcinogenesis
is a multifactorial process involving defects in multiple pathways related to DNA repair, the cell
cycle, apoptosis, and a multitude of other processes. Once initiated, cancer progression involves
changes in the expression of a plethora of genes due to a wide variety of mechanisms. Such
individual changes and the sum of these changes act to determine clinical progression and
response to treatment.
MMR deficiencies are associated with not only an increased rate of mutation, but with
specific secondary mutations. In humans, whole genome gene expression profile of MMR
deficient cancers are distinct from those of cancers with MMR proficiency. Large scale genomic
analysis of MMR proficient and deficient canine urothelial carcinoma cell lines would allow for

	
  

203	
  

	
  
evaluation of expression profiles of other DNA repair pathways, for comparison of differential
expression of genes between MMR deficient and proficient cell lines, for comparison of changes
in gene expression of both MMR proficient and deficient cell lines that occur in response to
treatment, and for comparison of profiles in canine urothelial carcinomas to reported gene
profiles in human cancers. Similar to what has been documented in human cancers with variable
MMR capacity, differences in expression of genes related to apoptosis, antitumoral immunity,
and MMR would be expected. With evaluation of profiles with response to treatment, cell lines
with varying MMR capacity will undoubtedly have significant differences in expression of
multiple genes as MMR capacity is associated with varying sensitivity to cytotoxic agents. Such
differential gene profiles would serve to highlight additional potential targets for therapeutic
intervention. Comparisons of gene profiles in dogs to those of humans would also further the
supposition that dogs are accurate models of urothelial carcinogenesis and MMR deficiency for
human cancer.
	
  
	
  
	
  

	
  

204