LIBRARY
Michigan State
University
This is to certify that the
dissertation entitled
INTERACTIONS AMONG INVADING TICKS, WILDLIFE, AND
ZOONOTIC PATHOGENS
presented by
Sarah Anne Hamer
has been accepted towards fulfillment
of the requirements for the
PhD degree in Fisheries and Wildlife
Ecology, Evolutionary Biology,
and Behavior
,n , ‘ . ff,
( Major Professoo‘é Signame
29 April 2010
Date
MSU is an Affinnative Action/Equal Opportunity Employer
- —< —.--.—----—---c-o-o-n-a—o--—
PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.
DATE DUE DATE DUE DATE DUE
5/08 K:/Proj/Aoc&PrelelRC/DateDue.mdd
INTERACTIONS AMONG INVADING TICKS, WILDLIFE,
AND ZOONOTIC PATHOGENS
By
Sarah Anne Hamer
A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Fisheries and Wildlife
Ecology, Evolutionary Biology, and Behavior
2010
ABSTRACT
INTERACTIONS AMONG INVADING TICKS, WILDLIFE,
AND ZOONOTIC PATHOGENS
By
Sarah Anne Hamer
Emerging vector-borne zoonotic diseases arise from complex interactions among
pathogens, bridge vectors, wildlife hosts, and hmnans. Ixodes scapularis - the
blacklegged tick - hosts a suite of zoonotic pathogens it in the Midwestern United States,
including the agents of Lyme disease, human granulocytic anaplasmosis, and babesiosis.
Risk of these diseases is increasing due in part to spread of I. scapularis. Over six years,
I have investigated hypothesized co-invasion of I. scapularis and pathogens in the
Midwest by tracking the spatial and temporal dynamics of invasion using of wild bird,
mammal, pet dog, and vegetation drag surveys, with subsequent genetic analyses of
pathogens. My studies operated not only where I. scapularis is readily found, but also in
zones beyond its detected distribution yet susceptible to establishment.
In Michigan, I have documented in real-time a northward invasion of blacklegged
ticks from a focal area of recent blacklegged tick detection on the west coast. White-
footed mouse surveillance was most sensitive method for detection of low-density tick
populations due to their importance for feeding immature ticks as well as the ease with
which large sample sizes are acquired. Invasion was not as apparent in the eastward,
inland direction, and future studies will address ecological parameters that may explain
the differential distribution. Compared to the wildlife studies, surveillance of pet dogs
was less sensitive for detection of blacklegged tick distribution, likely because of the
common practice of canine tick chemoprophylaxis.
While the detected distribution of blacklegged ticks was focal and expanding, l
detected a wider distribution of the Lyme disease pathogen, Borrelia burgdozfiri, in
alternative tick species and wildlife hosts. This pattern was most evident in bird-
associated ticks at a focal site 90 km to the east of I. scapularis invasion front, where l
dentatus ticks removed from birds harbored 3.5% infection prevalence with B.
burgdorferi, and no I. scapularis ticks were found. The B. burgdorfizrz' strains found in
this scenario of ‘cryptic’ transmission were comprised of many novel types not
previously described from Lyme disease endemic areas, and also at least three strains
previously associated with disseminated human Lyme disease. I hypothesize that cryptic
cycles reduce the time lag between I. scapularis invasion and the build-up of infection
prevalence, and may result in the introduction of novel strains to human and canines.
Across five states of the Midwest that represent a continuum of establishment of I.
scapularis, I hypothesized that patterns of diversity of pathogens within 1. scapularis may
be useful in elucidating the broad-scale tick invasion and subsequent disease emergence.
Analysis of 1565 adult I. scapularis ticks from 13 sites across five Midwestern states
revealed that tick density, infection prevalence with multiple microbial agents, co-
infections, and strain diversity of B. burgdorferi were positively correlated with the
duration of establishment of tick populations, though observed differences were subtle.
Cumulatively, these data suggest that the invasion of ticks and emergence of various tick-
bome diseases may be more complex than the traditional scenario whereby infected,
invading ticks are the only means of introduction of pathogens to naive communities.
To my mentors, colleagues, family and friends.
iv
ACKNOWLEDGEMENTS
It has been my honor to be mentored by my graduate co-advisors, Drs. Graham
Hickling and Jean Tsao, who have overseen my seven-year dissertation program and
taught me how to succeed in a life in academia. I aspire to the tireless enthusiasm for
biology that Jean expresses, and to the manner in which Graham can break down
complex ecological questions into testable hypotheses. I also thank my graduate
committee: Dr. Edward Walker for welcoming me into his laboratory and providing
diagnostic infrastructure; Dr. Linda Mansfield for serving as a wonderful DVM/PhD role
model, and Dr. Kelly Millenbah for pushing me to remember the big picture of the
ecological and epidemiological significance of this dissertation amidst the details of the
data I am also grateful for Dr. Steve Schmitt for his wildlife veterinary leadership and to
Dr. Kay Gross for welcoming our crew year after year to the W. K. Kellogg Biological
Station. I thank Erik Foster for the great foundation from which this dissertation was
sparked.
I’ve been fortunate to work alongside many wonderful undergraduate, graduate,
and veterinary students both in the field and laboratory. In particular, I am thankful to
Jennifer Sidge and Michelle Rosen for their exceptional work ethic and friendship amidst
endless hours of blood-sampling mice and manual checks of DNA sequence quality. I
am especially indebted to Rich and Brenda Keith for seven years of collaboration,
resulting in hundreds of thousands of bird tick samples from the Pitsfield Banding
Station. Additionally, I thank other members of the Tick Team including Todd Lickfett,
Natalie Ochmanek, Mandy Kauffman, Pam Roy, Isis Kuzcaj, Emily Johnston, Andy
Flies, Aubrey Rankin, Michelle Haring, Jen Stowe, Emily Koppel, Chris Niebuhr, Aaron
Rydecki, Katie Boatman, Megan Schiable, Dana Galbreath, Christina Barber, Lyssa
Alexander, and the undergraduate Fisheries and Wildlife Club for years of volunteering
at deer check stations. S. Bolin, A. Murphy and N. Grosjean provided technical
assistance for indirect fluorescent antibody testing used in the canine chapter. Alan
Barbour and Bridgit Travinsky assisted with genetic typing.
I thank my family and friends for support, especially my parents, Marko and
Diane Yaremych. My mother has equipped me with the highest quality, hand-sewn bird
holding bags and tick drag cloths, and her bird bags are now used by omithologists across
six states! I am very thankful for the inspiration provided by my favorite collaborator
and husband, Dr. Gabriel Hamer, with whom I cannot wait to share a laboratory, recruit
students, co-write grants, and investigate emerging zoonotic diseases in our life together.
I acknowledge the institutions and agencies that welcomed our research crew onto
their properties for tick and wildlife collections: Michigan, Illinois, Indiana, Minnesota,
and Wisconsin Departments of Natural Resources, the National Park Service, W. K.
Kellogg Biological Station, and Fort McCoy Military Instillation.
Parts of this dissertation are co-authored by Dr. Jean Tsao, Dr. Graham Hickling,
Dr. Edward Walker, Dr. Linda Mansfield, Erik Foster, Jennifer Sidge, Michelle Rosen,
Pamela Roy, and Christina Barber.
This work was supported by the CDC-Division of Vector-Bome Infectious
Diseases Cooperative Agreement No. CIOOl71-Ol , the Companion Animal Fund of
Michigan State University College of Veterinary Medicine, and the National Science
Foundation Doctoral Dissertation Improvement Grant. I appreciate additional funding
from a number of Michigan State University sources: W. K. Kellogg Biological Station
vi
(Porter Research Award, Graduate Fellowship Award, G. H. Laufl‘ Research Award); The
Gradate School and College of Agriculture and Natural Resources (Summer Support
Fellowship, Graduate Retention Fellowship, Travel Grants, and Dissertation Completion
Fellowship, and Alumni Association Fellowship); College of Veterinary Medicine (Tracy
Anne Hammer Memorial Fellowship for Duachegree DVM Students, Travel Grants);
Department of Zoology (G. H. Wallace Ornithological Scholarship); Department of
Animal Studies (G. R. Hartsough Endowed Scholarship in Fur Animal Studies);
Department of Fisheries and Wildlife (Glassen Foundation Fellowship, Graduate Student
Organization Travel Grants); and the Program in Ecology, Evolutionary Biology, and
Behavior (Summer Support Fellowship and Travel Grants). Extemal funding was
obtained from the Wildlife Disease Association (Graduate Research Recognition Award),
Safari Clubs International Michigan Involvement Committee (Joseph G. Schotthoefer
Memorial Student Award) and Rocky Mountain Goat Foundation (Bill Burtness Award).
vii
TABLE OF CONTENTS
LIST OF TABLES ................................................................................... xi
LIST OF FIGURES ................................................................................. xii
CHAPTER 1
INTRODUCTION ..................................................................................... 1
REVOLUTIONS IN THE STUDY OF EMERGING VECTOR-BORNE ZOONOTIC
DISEASES ............................................................................................. 5
Abstract ................................................................................................. 5
Synthesis of invasion biology and emerging disease .............................................. 9
The molecular revolution ........................................................................... 14
Bloodmeal analysis ........................................................................ 20
The geographic revolution .......................................................................... 22
Risk modeling ............................................................................... 25
Vector-borne zoonotic disease vignettes ......................................................... 28
Plague ........................................................................................ 28
Lyme disease in the American Midwest ................................................. 32
T ick-borne encephalitis in Europe ....................................................... 36
American cutaneous leishmaniasis ...................................................... 38
Conclusion ........................................................................................... 41
References ............................................................................................... 45
CHAPTER 2
ZOONOTIC PATHOGENS IN IXODES SCAPULARIS, MICHIGAN ...................... 54
Abstract ............................................................................................... 54
Introduction ........................................................................................... 54
The study .............................................................................................. 55
Conclusion ........................................................................................... 57
References ............................................................................................ 62
CHAPTER 3
INVASION OF THE LYME DISEASE VECTOR IXODES SCAPUIARIS:
IMPLICATIONS FOR BORRELM BURGDORFERI ENDEMICITY ...................... 64
Abstract ............................................................................................... 64
Introduction ........................................................................................... 65
Materials and methods .............................................................................. 68
Site selection and sampling regime. ..................................................... 68
Mammal trapping ....................................................................... i ..... 69
Bird mist netting ............................................................................. 7O
Questing tick sampling ..................................................................... 7O
Borrelia burgdorferi detection ............................................................. 71
Nucleotide sequencing ..................................................................... 72
Statistics ...................................................................................... 73
viii
Results ................................................................................................. 73
Wildlife captures ............................................................................ 73
I. scapularis on wildlife .................................................................... 74
Questing ticks ............................................................................... 76
B. burgdorferi infection in drag-sampled ticks ........................................ 77
Host infection with B. burgdorferi ....................................................... 77
B. burgdorferi infection in I. scapularis removed from hosts ......................... 78
B. burgdorferi infection in alternative tick species .................................... 78
Nucleotide sequencing ..................................................................... 79
Opportunistic detection of other Borrelia species ..................................... 80
Discussion ............................................................................................. 80
Supplementary material 1 .......................................................................... 92
Supplementary material 2 .......................................................................... 94
References ............................................................................................ 102
CHAPTER 4
CANINES AS SENTINELS FOR EMERGING IXODES SCAPULARIS-BORNE
ZOONOSES RISK ................................................................................ 108
Abstract ............................................................................................... 108
Introduction ......................................................................................... 109
Materials and methods ............................................................................. 113
Sample acquisition ........................................................................ 1 13
Indirect Fluorescent Antibody (IFA) .................................................... 1 14
Western blot (WB) ......................................................................... 116
C6 enzyme-linked immunosorbent assay (ELISA) .................................... 1 16
Tick processing and polymerase chain reactions (PC R) ............................ 1 17
Sequence analysis ........................................................................ 118
Spatial analysis .................... _ ....................................................... 1 19
Early invasion survey .................... ’ ................................................. 119
Results ................................................................................................ 119
[FA .......................................................................................... 120
WB ............................................................................................ 120
C6 ELISA ................................................................................... 121
Tick PCR ................................................................................... 121
Sequence analysis ........................................................................ 122
Spatial analysis ....... ‘ .................................................................... 123
Early invasion survey ..................................................................... 123
Discussion ............................................................................................ 124
References ........................................................................................... 136
CHAPTER 5
CRYPTIC TRANSMISSION OF BORRELIA BURGDORFERI, B. ANDERSONII, AND
B. MIYAMOTOI BY BIRD-DERIVED TICKS IN THE ABSENCE OF THE
BLACKLEGGED TICK .......................................................................... 139
Abstract ............................................................................................. 139
Introduction ......................................................................................... 140
Materials and methods ............................................................................ 143
Bird mist netting ........................................................................... 143
Mammal trapping ......................................................................... 144
Tick and Borrelia spp. detection ........................................................ 146
Nucleotide sequencing ................................................................... 147
Statistics .................................................................................... 149
Results ............................................................................................... 150
Trapping success and infestation prevalence .......................................... 150
Phenology of bird-associated ticks ..................................................... 153
Ticks on eastern cottontails .............................................................. 153
Ixodes scapularis on white-footed mice and chipmunks ............................. 153
Ticks on non-target captures ............................................................ 154
Borrelia infection in bird-associated ticks ............................................. 154
Borrelia infection in mammal-associated ticks ........................................ 157
Borrelia infection in mammal ear biopsies ............................................ 157
Genotypes of B. burgdorferi, B. andersonii, B. miyamotoi. . .............................. 157
Discussion .......................................................................................... 160
References ........................................................................................... 189
CHAPTER 6
DIVERSITY OF ZOONOTIC PATHOGENS IN IXODES SCAPULARIS AS A
MARKER OF ITS INVASION OF ACROSS THE MIDWESTERN UNITED
STATES ............................................................................................. 197
Abstract ............................................................................................. 197
Introduction ......................................................................................... 198
Materials and methods ............................................................................ 202
Tick collections ............................................................................. 202
Pathogen detection ....................................................................... 204
Nucleotide sequencing .................................................................... 205
Statistics .................................................................................... 207
Results ............................................................................................... 209
Index of abundance of I. scapularis ...................................................... 209
Infection of ticks with multiple microorganisms ...................................... 210
B. burgdorferi genotypes .................................................................. 212
B. burgdorferi RST groups ............................................................... 212
B. burgdorferi strain richness and diversity ........................................... 212
B. burgdorferi population structure ..................................................... 214
B. burgdorferi phylogeny and network analysis ....................................... 215
Discussion .......................................................................................... 216
References ........................................................................................ _. ..240
CHAPTER 7
SYNTHESIS AND FUTURE DIRECTIONS .................................................. 250
Synthesis of results ................................................................................ 250
Future research and implications for Michigan ................................................ 259
References ........................................................................................... 266
LIST OF TABLES
Table 1.1. Glossary of vector-home zoonotic disease related-terms ......................... 43
Table 2.1. Michigan site-specific prevalences of adult and nymphal I. scapularis
infection with three pathogens. Site names correspond to Figure 2.1. Life stage:
A== adult; N = nymph ............................................................................... 60
Table 3.1. Tick-host associations and infestation prevalence on the inland and coastal
transects in Lower Michigan, May-June, 2004-2008. L = larva; N = nymph; A = adult..96
Table 3.2. Sequence confirmation and accession numbers of B. burgdorferi infection in
various sample types from coastal and inland sites in lower Michigan, 2004-2008. I =
inland transect site; C = coastal transect site. RST = 168 - 23S rRNA spacer type of B.
burgdorferi. AF = adult female; AM = adult male; N = nymph; L = larvae; WFMO =
white-footed mouse; SFSQ = southern flying squirrel; EAGS = eastern gray squirrel;
VIOP = Virginia opossum; EACH - eastern chipmunk, NOCA = Northern cardinal;
AMRO = American robin, SOSP = song sparrow; COON = raccoon, MJMO = meadow
jumping mouse; EACO = eastern cottontail. ................................................... 97
Table 4.1. Results of IFA and WB serological assays of canine serum samples from 18
clinics in lower Michigan. IF A endpoint titers are expressed as reciprocals. IFA suspect
positives are those samples with endpoint titers of 1:640 or more dilute. All IF A suspect-
positives, plus randomly-selected IFA negatives (number indicated above), were subject
to confirmatory WB. Overall seroprevalence computed using denominator of (total
samples — number of vaccine positives). In no case did an IFA negative sample test
positive by WB. Zone assignment relates to proximity of clinic to established
populations of I. scapularis. Neg = negative test result; Pos = positive test result; Vax =
vaccine positive test result ........................................................................ 134
Table 4.2. Tick species and life stage-specific results of PCR testing for B. burgdorferi,
A. phagocytophilum, and Babesia species in ticks removed from canine patients at clinics
throughout lower Michigan, 2005 ............................................................... 135
Table 5.1a. Infestation prevalences of all parasitized bird species with three species of
tick, Pitsfield Banding Station, 2004-2007. Some ticks were observed on birds yet not
removed and are categorized as unknown identity. Tick species-specific infestation
prevalences therefore underestimate true infestation. Additionally, a single Song Sparrow
harbored 6 Dermacentor variabilis larvae. ..................................................... 177
Table 5.1b. Bird species investigated for ticks and found to be uninfested, Pistfield ’
Banding Station, 2004-2007.......... ............................................................................... 179
Table 5.2. Small mammals as sentinels for I. scapularis and B. burgdorferi presence at
Pitsfield Bird Banding Station and Van Buren State Park, a hotspot for recently-invaded
I. scapularis 90 km west of Pitsfield. Samme size of mammals captured and infestation
xi
prevalence with two species of ticks is followed by sample size of ear biopsies tested and
infection prevalence with B. burgdorferi. ...................................................... 180
Table 5.3. Tick species-specific infection prevalence with 3 different Borrelia pathogens,
2004-2007, Pitsfield Banding Station. In addition to samples listed, a single adult female
I. dentatus and a pool of 6 D. variabilis larvae were negative for all pathogens. 1 8 1
Table 6.1. Best estimates of the relative dates of Ixodes scapularis population
establishment at the sub-state level in selected Midwestern areas based on published
literature. Field sites sampled in the current study are categorized based on their best
estimate relative establishment .................................................................. 228
Table 6.2. Index of abundance and infection status of 1595 adult I. scapularis adults
across levels of I. scapularis establishment in the Midwestern United States, 2006-2007.
Significant differences among establishment groups are noted with superscript letters.
UP = Upper Peninsula; LP = Lower Peninsula ................................................. 229
Table 6.3. Matrices of co-infection for each I. scapularis establishment group. Infection
prevalence (%) with each individual microbe is on the diagonal in bold. Expected co-
infection prevalence is in the upper triangle, and observed co-infection prevalence is in
the lower triangle. Observed prevalences are not different than expected unless indicated
with an asterisk. Bb = B. burgdorferi; Bm = B. miyamotoi; Ap = A. phagocytophilum;
Barn = Ba. microti; Bao = Ba. odocoilei ......................................................... 230
Table 6.4. Mixed strain infections and strain richness of single strain B. burgdorferi
infection in 458 infected adult I. scapularis adults across a continuum of I. scapularis
establishment in the Midwestern united States, 2006-2007. UP = Upper Peninsula; LP =
Lower Peninsula ..................................................................................... 231
xii
LIST OF FIGURES
Figure 1.1. Two-by-two table of hypotheses regarding the invasion of the blacklegged
tick, I. scapularis, versus that of the Lyme disease pathogen, B. burgdorferi, based on
observations of the distribution of each species. Hypotheses are described in detail in
Chapter 3 ............................................................................................... 4
Figure 1.2. Methods for field collections of vectors. a) Drag sampling for questing ticks;
b) Checking drag cloth for tick presence. c) CDC light trap elevated in tree canopy for
mosquito collections. d) Gravid trap for collections of gravid female mosquitoes. ....... 44
Figure 2.1. Tick dragging field sites within Michigan. A= Menominee North; B=
Menominee South; C= Duck Lake State Park; D= Saugatuck Dunes State Park; E= Van
Buren State Park. Gray shaded counties are those in which endemic (Upper Peninsula)
and recently-invaded (Lower Peninsula) 1. scapularis are known to occur... .... . . . .59
Figure. 3.1. Locations of study sites in Lower Michigan, 2004-2008. The four sites along
Michigan’s west coast comprise the coastal transect (from south to north, C1-C4), and the
four inland sites comprise the inland transect (from southwest to northeast, 11-14).
Shading in the Lower Peninsula represents the three-county region where I. scapularis
were detected on small mammals in 2002-2003 (Foster 2004). The cross-hatched county
in the Upper Peninsula is Menominee County, Michigan’s longstanding endemic focus of
I. scapularis and B. burgdorferi. .................................................................. 99
Figure. 3.2. Infestation of white-footed mouse (left rectangle) and eastern chipmunk
(right rectangle) with I. scapularis at the four sites of the coastal transect, May-June,
2004-2008. Level of shading indicates the proportion of animals infested, with the
number of animals screened for ticks inside the symbol. Adjacent to each symbol, B.
burgdorferi infection prevalence (%) in (i) I. scapularis larvae (minimum infection
prevalence), (ii) nymphs removed from hosts, and (iii) host ear biopsies are listed from
top to bottom, each followed by sample size in parenthesis. At site C4, no traps were set
in 2004-2006. At site C1 in 2004 and C2 in 2006, no chipmunks were captured despite
traps being set. Information on the inland transect sites, where infestation and infection
rates were close to zero, appears in the texthO
Figure. 3.3. Average densities and B. burgdorferi infection prevalences of drag-sampled
I. scapularis at the three sites on the coastal transect in May and June, 2004-2008. No I.
scapularis were dragged at site C4. Regression lines and coefficients for nymphal
densities are shown. Infection prevalence for each life stage (minimum infection
prevalence for larvae) is expressed as the percent positive, with the total number tested in
parenthesis below each year. C = coastal transect srtelOl
Figure 4.1. Locations of the 18 veterinary clinics (labeled A - R) that participated in the
2005 canine tick and serosurvey of Lower Michigan. Shading indicates counties in
Which I. scapularis has recently invaded (Lower Peninsula) or is endemic (Upper
Peninsula); shaded counties are those within which all three life stages of I. scapularis
xiii
have been documented (Foster 2004). Circles indicate the estimated Lyme disease
vaccination rates (expressed as a proportion) for the dogs at each clinic. The locations of
the two dogs seropositive for antibodies to B. burgdorferi from natural exposure are
shown. . ........................................................................................................................... 130
Figure 4.2. Proportions of the serum samples collected (n == 353) from individuals of the
various American Kennel Club breed groups. Dogs reported as a specified breed mix are
classified under that breed group (i.e. German Shepard mix = herding group) whereas
reports of mixed breed with no breed specification are listed as ‘Mix’ ..................... 131
Figure 4.3. Frequency distribution of IFA endpoint titers for B. burgdorferi antibody
detection in canine sera (n = 353). Sera with titers s 1:320 were classified as negative for
infection with B. burgdorferi. The remaining high-titer sera, plus sera from dogs reported
as having been vaccinated, were classified as natural positive, vaccine positive, or
negative based on Western blot .................................................................... 132
Figure 4.4. Distribution of clinics with canines harboring I. scapularis, B. burgdorferi-
irrfected ticks of any species, and Babesia-infected ticks of any species. The shading
indicates counties in which I. scapularis has recently invaded and is now established with
documented presence of all three life stages (Foster 2004) .................................... 133
Figure 5.1. Pitsfield Bird Banding Station in Vicksburg, MI (blue triangle) and Van
Buren State Park, the site of comparative small mammal sampling (red circle). Shading
indicates counties with documented established populations of blacklegged ticks at the
time the study started in 2004 (dark gray) and at the time the study ended in 2007 (light
gray), and the endemic pOpulation (diagonal lines; Hamer et a1. 2010). H. lep. = H.
leporispalustris; I. dent. = I. dentatus. ................................................................ 182
Figure 5.2. Phenology of larval and nymphal bird—associated ticks (a, I. dentatus; b, H.
leporispalustris) depicted as weekly mean proportions of infested birds (error bars are
standard error of the mean using year as replicate) ............................................. 183
Figure 5.3. Temporal variation in infection prevalence of B. burgdorferi, B. andersonii,
and B. miyamotoi from May-November, 2004-2007. Monthly mean infection
prevalences across all years (E SE of mean) is plotted. All tick species are
aggregated ............................................................................................ 184
Figure 5.4. Neighbor-joining phylogenetic tree and frequency distribution of B.
burgdorferi IGS haplotypes collected from Pitsfield Banding Station, 2004-2007. The
percentages of replicate trees in which the associated taxa clustered together in the
bootstrap test (1000 replicates) are shown next to the branches. Indigenous strains (novel
IGS mutants not previously reported) are indicated with a blue triangle; ubiquitous strains
(haplotypes previously reported in Bunikis et al. (2004a) or detected in I. scapularis
across the Midwest) are indicated with a red circle. Sequences conforming to the 16S —
23S RST group 1, 2, and 3 designations by the criteria of Liveris et a1. (1995) are
demarcated by the labels at the top left of shaded or unshaded groups. Total samples size
xiv
of each strain is indicated at the end of each branch label, of which a majority is from
bird-derived ticks; the 7 mammal-associated samples (from rabbit ticks and ear and small
mammal ears) are denoted in parentheses. ..................................................... 185
Figure 5.5. Minimum spanning tree (MST) of B. burgdorferi IGS haplotypes collected
from Pitsfield Banding Station, 2004-2007. Each black circular node connecting
haplotypes represents one mutational change; a 7 base pair indel that occurs within the
IGS is considered as one change. The size of each haplotype is proportional to its
frequency within the sampled population. The rectangles represent the haplotypes with
the highest outgroup weight, which correlates with haplotype age. The two haplotpyes
within RST 1 (IGS 1A and Novel PP) are linked to each other by two mutational
changes, yet unlinked to the rest of the network. Indigenous strains (novel IGS mutants
not previously reported) are indicated with a blue triangle; ubiquitous strains (haplotypes
previously reported by Bunikis et al. (2004a) or detected in I. scapularis across the
Midwest) are indicated with a red circle ......................................................... 186
Figure 5.6. Rarefaction curve for B. burgdorferi strains derived from Pitsfield Banding
Station, 2004-2007. Mean and standard deviations of the number of strains found in
subsarnples are shown .............................................................................. 187
Figure 5.7. Neighbor-joining phylogenetic tree and frequency distribution of B.
andersonii IGS haplotypes collected from Pitsfield Banding Station, 2004-2007. The
percentages of replicate trees in which the associated taxa clustered together in the
bootstrap test (1000 replicates) are shown next to the branches. Total samples size of
each strain is indicated at the end of each branch label, of which a majority is fi'om bird-
derived ticks; the 7 rabbit-associated samples (from rabbit ticks and a rabbit ear) are
denoted in parentheses. ............................................................................ 188
Figure 6.1a. Location of field sites for collection of I. scapularis across the Midwestern
United States, spring 2006-2007. Sites are categorized by establishment status of
documented 1. scapularis populations: highly-established (red circle); intermediate-
established (blue square); recently-invaded (green triangle) ................................. 232
Figure 6.1b. Adult female I. scapularis questing on forest understory vegetation ........ 233
Figure 6.2. Infection of 1595 adult I. scapularis with Borrelia spp., Anaplasma
phagocytophilum, and Babesia spp. organisms, and co-infections thereof, from across the
Midwestern United States, 2006-2007. Error bars are the standard error of the mean
prevalence of all sites within each establishment group ...................................... 234
Figure 6.3. Variation in proportion of ribosomal spacer (RST) types 1, 2, and 3 of B.
burgdorferi across the Midwestern United States, 2006-2007 ............................... 235
Figure 6.4. Rarefaction analysis of the influence of sample size of detected strain richness
of B. burgdorferi within adult I. scapularis across the Midwestern United States, 2006-
2007. Predicted mean strain richness (intervals are standard deviation) is modeled at
incremental sampling for each tick establishment group until the observed datapoint is
reached. The observed datapoint for each group is plotted as the final datapoint on each
curve ................................................................................................. 236
Figure 6.5. Unrooted neighbor-joining phylogram and frequency distribution of B.
burgdorferi IGS haplotypes collected from across the Midwestern United States, 2006-
2007. The percentages of replicate trees in which the associated taxa clustered together
in the bootstrap test (1000 replicates) are shown next to the branches when 60 or above.
Haplotypes previously reported in Bunikis et al. 2004 begin with ‘IGS’; all strains
beginning with ‘Midwest’ were not previously reported. Strains conforming to the 16S —
23S rRNA RST group 1, 2, and 3 designations by the criteria of Liveris et al. (1995) are
demarcated by the labels at the top left of shaded or unshaded groups. The representation
of each strain is expressed as a percent within each establishment group .................. 237
Figure 6.6. Minimum spanning network (MSN) of B. burgdorferi IGS haplotypes
collected from across the Midwestern United States, 2006-2007. Each black circular
node connecting haplotypes represents one mutational change; a 7 base pair indel that
occurs within the IGS is considered as one change. The size of each haplotype is
proportional to its frequency within the sampled population. The rectangles represent the
haplotypes with the highest outgroup weight, which correlates positively with haplotype
age. The two haplotypes within RST l (IGS 1A and Midwest L) are linked to each other
by one mutational change and are unlinked to the rest of the network due to a high degree
of divergence. Strains are categorized as to their occurrence within each of the three I.
scapularis establishment groups: highly-established only (white); intermediate-
established only (black); recently-invaded only (striped); two groups (white with thick
border); all three groups (gray) ................................................................... 239
Figure 7.1. Incidence (number of reported cases per 100,000 population) of Lyme disease
in humans in southwest Michigan (light bars) and the whole state of Michigan (dark bars)
in 1996-2009. Case numbers shared by E. Foster, Michigan Department of Community
Health (pers. comm.) .............................................................................. 265
CHAPTER 1
Introduction
“Nowadays we live in a very explosive world, and while we may not know where or
when the next outburst will be, we might hope to find ways of stopping it or at any rate
damping down its force. It is not just nuclear bombs and wars that threaten us, though
these rank very high on the list at the moment: there are other sorts of explosions -
ecological explosions. An ecological explosion means the enormous increase in numbers
of some kind of living organism - it may be an infectious virus like influenza, or a
bacterium like bubonic plague, or a ftmgus like that of the potato disease, a green plant
like the prickly pear, or an animal like the grey squirrel they may develop slowly and
they may die down slowly; but they can be very impressive in their effects, and many
people have been ruined by them ...”
«Charles Elton, (1958)
My dissertation investigates interactions among invading blacklegged ticks, their
wildlife hosts, and a suite of zoonotic pathogens that are vectored by blacklegged ticks
across the Midwestern United States. I have studied the invasion and establishment of
blacklegged ticks within a zone of active invasion in Michigan, and I compare these
findings to the status of ticks in several endemic areas where they have been established
for various amounts of time. I also present data on diversity and incidence of multiple
zoonotic pathogens across the Midwest. Throughout this dissertation, my emphasis is on
understanding the changing status of blacklegged ticks and the Lyme disease pathogen,
which is the agent of the most significant vector-borne disease presently occurring in the
United States (Bacon et a1. 2008).
My dissertation begins with this introductory chapter (Chapter 1), which reviews
important recent developments in the study of emerging and resurging vector-borne
zoonotic diseases. This review emphasizes recent advances in molecular and ‘
geographical analysis technologies have advanced the field of vector-borne disease
ecology. Lyme disease is introduced as one example in a series of four vignettes that
each illustrates contemporary problems in zoonotic disease emergence. The next five
chapters (Chapters 2-6) each present original data collected in Michigan and the broader
Midwestern US that have either been published within the scientific literature (Chapters
2-4) or is preparation for submission to a journal (Chapters 5-6). Each of these data
chapters is formatted as a stand-alone manuscript. Chapter 7 provides a synthesis of key
research findings and presents my ideas on future research directions.
The original research studies presented herein were sparked by findings of
blacklegged ticks in previously uncolonized areas of southwestern Michigan (Foster
2004). Early in my program, I developed some competing hypotheses regarding the
invading disease system, and organized them into a two-by-two table (Figure 1.1); these
hypotheses are described further in Chapter 3. Research to address these hypotheses
aimed to determine whether an invasion was indeed occurring, and if so, the extent to
which the blacklegged tick and the Lyme disease pathogen were co-invading (the
classical notion) versus independently invading or already established. Testing these
hypotheses required me to sample for the Lyme disease pathogen in wildlife hosts and
alternative tick species in areas outside of the detected distributional range of blacklegged
ticks (Chapters 3 and 5). This is a contrast to many prior studies that only tested for the
presence of the pathogen in areas with sympatric blacklegged ticks. Chapter 4 presents
an assessment of the use of pet dogs as sentinels for tick invasion and disease emergence;
this surveillance tool is well-utilized in Lyme disease endemic areas, but has not
previously been studied in an area undergoing invasion by blacklegged ticks.
In order to more completely understand how and why the invasion of this disease
system is occurring in Lower Michigan, I was interested in also studying the system
where it has been established for longer durations of time. The upper Midwestern United
States (one of the two broad endemic foci for Lyme disease; the other being the
Northeast) encompasses gradients of establishment of blacklegged ticks. An ecological
principle is that the center of origin of a species is likely to also serve as the center of
diversity for that species and related taxa (Cain 1944). 1 extended this notion to the
blacklegged tick and its associated pathogens, for which I hypothesized that the areas
where blacklegged ticks have been documented as established in the Midwest for the
longest duration (Minnesota, Wisconsin, Michigan’s Upper Peninsula) will harbor the
greatest diversity of tick-home pathogens, both at the inter- and intra-species levels, in
comparison to areas where blacklegged ticks are currently invading (Chapters 2 and 6).
This dissertation is therefore not only a means for better understanding the complex
nature of biological invasions in general, but also for collecting distributional and
infection prevalence data that may be useful to medical communities and outdoor
recreationalists to reduce risk of tick-borne diseases.
Figure 1.1. Two-by-two table of hypotheses regarding the invasion of the blacklegged
tick, I. scapularis, versus that of the Lyme disease pathogen, B. burgdorferi, based on
observations of the distribution of each species. Hypotheses are described in detail in
Chapter 3.
No I. scapularis invasion 1. scapularis invasion
Pathogen
arrives with I. .
scapularis ‘Status Quo’ ‘Dual-Invasron’
Pathogen is ‘ . , ‘Invasron
already Cryptic Cycle 1,
MdeSpread Cryptic Cycle’
Revolutions in the study of emerging vector-home zoonotic diseases
Abstract
Many vector-borne zoonotic diseases (V BZDs) are emerging into new zones or
resurging from their past or current endemic ranges, and such spread is usually mediated
by anthropogenic factors. Emerging VBZD systems involve invasive species, in which
the invader is a vector, pathogen, reservoir host, human, or combination thereof, and
VBZD emergence can be understood using the paradigms of classic invasive species
biology. The advent of molecular and geographic analytical tools has revolutionized the
study of VBZD. Through highlighting selected emerging VBZDs in a series of vignettes,
I demonstrate the synergy of the molecular and geographic analytical tools that allow for
a better understanding of disease across a heterogeneous landscape. -- a prerequisite for
effective control.
Introduction
Vector-borne zoonotic diseases (V BZDs) are significant sources of human
morbidity and mortality. VBZDs are a subset of zoonotic diseases, i.e., those disease
systems in which pathogens of humans circulate in animal reservoirs. VBZDs arise from
the interactions among four key players — pathogens, vectors, vertebrate reservoir hosts,
and humans - all within a common environment that is conducive for disease
transmission. The pathogens include viruses, rickettsiae, bacteria, protozoan parasites,
flukes, and filarial nematodes, and encompass a diverse array of life cycles. The
arthropod vectors involved in these disease systems typically are blood feeders in at least
one, or sometimes all, life stages and include mites and ticks (Order Acari, Class
Chelicerata) and many groups within the Class Insecta. Vertebrate hosts include many
species of mammals and birds.
Emergence is the process that results in a measurable increase in the prevalence of
infection in humans, vectors, or reservoir hosts. Emergence is characterized by an
expansion in geographic distribution of infection, and may involve infection of new host
species. Zoonotic pathogens comprise over half (5 8%) of all human pathogen species (n
= 1407 total), and nearly three-quarters (73%) of human pathogens that are classified as
emerging or reemerging (Woolhouse and Gowtage-Sequeria 2005). Zoonotic pathogens
are therefore more likely to show emergent properties compared to non-zoonotic
pathogens, owing to the transmission probabilities inherent in the interfaces of humans,
domesticated animals, and wildlife.
Many emerging VBZDs are receiving renewed attention given the availability of
new tools and technologies which facilitate their detection and assessment of the health
risks they pose across wide geographic areas. The purpose of this chapter is to review the
status of emerging VBZDs with a particular perspective on two aspects: First, I identify
parallels between emergence of VBZDs and invasion biology. Secondly, I highlight how
major technological revolutions in biological sciences (specifically in molecular biology
and quantitative geographic analyses) have spurred our understanding of emergence of
VBZD, providing new conceptual frameworks to not only allow retrospective description
but also prediction. These themes will be highlighted in a series of five vignettes.
VBZDs as multi-factorial systems. VBZDs represent complex ecological systems that
involve multi-factorial abiotic and biotic components. Abiotic factors include rainfall,
temperature, soil, topography, hydrology, and climate. Biotic factors include the
pathogen, the vector, and the suite of vertebrate reservoir hosts, each of which can be
further subdivided. For example, pathogens typically circulate as various strains that
have differing virulence properties; vectors have different life stages and pathogen
infection may or may not be maintained during transition from one stage to the next;
vertebrate reservoir host populations have age and sex structures that influence the ability
of the host to be infected and to infect. F urthennore, population dynamic processes in
vector and reservoir host populations influence pathogen transmission. For example, the
annual phenological appearance of nymphal Ixodes scapularis ticks prior to larvae results
in infection in white-footed mice, which then become an infective reservoir to larvae,
permitting perpetuation of infection at foci in the absence of transovarial transmission.
Landscape epidemiology provides a useful conceptual framework for
understanding the distribution of VBZDs. The field of landscape epidemiology was
influenced by the founding studies of tick-bome viral infections conducted by Evgeny
Pavlovsky in the former Soviet Union, where concepts of endemicity, focality, and
nidality were formed (Pavlovsky 1966). Landscape epidemiology assesses the presence
of appropriate abiotic and biotic factors which permit a pathogen to occur in a particular
geographic area and to persist there over time. These factors comprise a biogeocenose, or
aggregate of living and non-living environmental properties. Importantly, this aggregate
can be quantified using modern tools of landscape ecology such as satellite imagery, and
detailed databases of land cover and land use. This allows classification of the
heterogeneous and patchy nature of the landscape, with only particular patches being
suitable to support the disease system. These patches are called foci or nidi (see Table
1.1 for glossary of VBZD-related terms). The disease system is endemic when it exists at
nidi with no need for external inputs. Landscape epidemiology permits development of
risk maps, based upon models of environmental receptivity and landscape heterogeneity,
that provide guidance for public health interventions to reduce disease risk (see below).
The basic reproductive rate of a pathogen, R0, is a central concept in disease
ecology, and it can be defined as the number of secondary infections that arise from a
single infected host in a population (Anderson and May 1982). In the case of VBZD, the
basic reproductive rate of the pathogen depends upon the infection prevalence of the
vectors, their rate of transmission of the pathogen to reservoir hosts, and in turn, the rate
at which reservoir hosts infect new vectors. When a pathogen that enters a new host or
vector population, one of three outcomes may occur: it may become locally extinct
(R0<1); it may persist in steady-state (R0=1); or it may increase in prevalence (R0>l ).
The true reproductive rate of an invasive pathogen is not easily defined, as the number of
unsuccessful invasion attempts is not possible to measure. Humans can influence R0
through many different environmental perturbations, including intrusion into a landscape
of endemic transmission and changes in land use and land cover to increase vector
density, wildlife reservoir host density, and their interactions.
An important component of the biogeocenose that influences R0 is the
biodiversity of the community of vertebrate hosts. In VBZD systems in which the vector
is a generalist and there exists a community of vertebrate host species that differ in their
reservoir competence, the dilution model may operate (Ostfeld and Keesing 2000).
According to this model, a vertebrate community with high species diversity contains
more reservoir-incompetent hosts than does a poorly-biodiverse community that may be
dominated by only the most competent reservoirs. Thus, in the biodiverse community,
the less-competent and incompetent hosts divert vectors from feeding on the competent
hosts and thus dilute the infection prevalence and disease risk. In the context of VBZD,
the dilution model has been demonstrated for Lyme disease (LoGiudice et al. 2003;
LoGiudice et al. 2008), tick-borne encephalitis (Perkins et al. 2006), and West Nile virus
(Swaddle and Calos 2008), but lack of effects have also been documented (Loss et al.
2009) and separation of density versus diversity effects is challenging (Begon 2008).
One of the fascinating aspects of emerging VBZDs is that a single vector and a
few vertebrate hosts may support many pathogens within the same system at the same
foci. For example, the blacklegged tick Ixodes scapularis and small rodent hosts
maintain transmission of the agents of Lyme disease (bacterium), human granulocytic
anaplasmosis (rickettsia), babesiosis (protist), and deer tick virus encephalitis (virus) in
Northeastern and Midwestern United States.
Synthesis of invasion biology and emerging disease
Invasion biology is a branch of ecology which deals with changes in distribution
of organisms with undesirable properties such as exotic species, pests and weeds. Good
examples of recent note are the zebra mussel and the emerald ash borer. The process of
biological invasion has four main components (With 2002): (1) a landscape exists that
includes a receptive enviromnent at some locations for the establishment of the invading
organism, were it to be introduced; (2) introduction of the invasive species occurs ; (3)
the invasive species establishes into an endemic status, then spreads locally and ofien
proliferates to the detriment of native species of similar biology; (4) the invasive species
spreads to new regions from its newly established sites, permitting regional spread.
Parallels and analogous processes to invasive species are easily recognized with
emerging VBZDs, in that both involve: (1) a process involving a receptive environment,
an organism with invasive properties, and a route of introduction followed by
establishment and spread; (2) retrospective analyses dominate; (3) an anthropocentric
point of view dominates; (4) common methods of landscape ecology. Outstanding
examples of importance currently include plague, Lyme disease, tick-borne encephalitis,
American cutaneous leishmaniasis; each of these is discussed in detail below. The tempo
of invasion-establishment-spread dynamics may be very fast (e.g., West Nile virus,
transcontinental in a few years), decadal (e. g., Lyme disease and tick-bome encephalitis,
regional emergence and spread over decades), or emergence over centuries (plague). The
appearance of human infection may be the first signal that an invasion and establishment
of a pathogen has already occurred, given that surveillance for invasive VBZD pathogens
is not typically conducted among vectors, wildlife, or domesticated animals in the
absence of human disease.
Mechanisms of invasion of pathogens associated with VBZDs are poorly known
but clearly a common element driving most is anthropogenic change to the landscape.
Emergence of many VBZD is supported by rapid movement of people, goods, and
animals via commerce and other physical means. For example, the earliest records of
introduction of yellow fever virus and its urban vector Aedes aegvpti were with colonial
and slave-trade shipping routes from the Old to the New Worlds. Modern transportation
via aircraft and on large transport ships facilitates movement of pathogens either in
10
infected vectors (likely the case for West Nile virus), infected animals (rats and their fleas
in the case of plague), or on living agents such as migrating birds or pet animals taken
from one place to another. Climate change has been proposed as an element driving the
emergence of some VBZDs, although strong tests of this relationship are lacking
(Randolph 2004).
Revolutions in the study of VBZDs
The ability to study VBZDs has advanced markedly in the past 30 years with the
advent of new tools for molecular and geographic analysis. Herein, I refer to these
advances as revolutions, because they continue to be so substantial and allow for
comprehensive analyses of VBZDs that were not before possible. These revolutions
arose during the era (which continues into present day) of unprecedented increases in the
rate of emergence of many VBZDs. Whereas historically most diseases are responded to
in a retrospective fashion, the biological revolutions collectively afford the ability to
predict when and where transmission will occur such that limited public health resources
may effectively be targeted to slow species invasions and reduce human disease.
VBZD research has progressed where molecular tools meet spatial analysis. The
distribution of infected bites of vectors across a landscape can be measured by molecular
tools which are used to detect prevalence of infection with pathogens in vectors, which
can then be mapped to show variation spatially. For example, Ruiz et al. (2010) used
spatial and statistical modeling to analyze and predict mosquito infection with WNV at
fine spatial and temporal scales around Chicago, IL. The authors found that increased air
temperature was the strongest temporal predictor of mosquito infection, and some terrain
ll
characteristics, including impervious surfaces and elevation in Chicago, also helped to
predict infection. The origin of WNV that invaded New York in 1999 was determined
through phylogenetic analysis with 33 other WNV strains and 8 related viruses from
across the world, in which the New York strain most closely related to a WNV strain
from a dead goose in Israel in the prior year (Lanciotti et al. 1999), suggesting a
Mediterranean origin. Through a mosquito population genetic analyses based on
microsatellite loci, the rapid westward spread of the virus across the country (in contrast
to the predominant north-south patterns of migratory birds) has recently been attributed to
dispersal mediated by the mosquito vector Culex tarsalis (V enkatesan and Rasgon 2010)
The analysis of the recent outbreak of chikungunya virus in Italy in 2007, in which
autochthonous transmission of a viral strain phylogenetically most similar to strains
identified in an earlier outbreak in India (Rezza et al. 2007), provides yet another
example in which molecular and spatial analytical tools provide a central foundation for
outbreak investigations.
Prediction of emergence of a VBZD is possible only after understanding disease
systems in their natural enzootic foci (Rogers and Randolph 2003). A first step in VBZD
research is often to define the density and geographic distribution of one or more
components of the disease system (the vectors, hosts, and pathogens). Typically, vectors
are collected and their infection status is determined which allows computation of the
‘entomological risk’ (Mather et al. 1996), which is directly related to human disease risk.
Quantifying infected vectors provides important information for public health agencies as
a form of surveillance. Beyond surveillance of the current status, however, the tools of
the biological revolutions allow the use of data to achieve the goal of prediction. To do
12
so, the following steps are generally followed (and will be expanded upon below): (1)
Field specimens are collected from a place and during a time that is biologically
meaningfirl. Field samples may be individuals of the vector species, or samples of tissue
or blood from hosts including wildlife, domestic animals (including sentinel species), or
hmnans. (2) Field specimens are assayed for infection with the pathogen or antibodies to
the pathogen in the laboratory. Presence/absence, point prevalence, or prevalence across
a time period may be determined. Further genetic characterization of the pathogen and
strain-level associations with hosts may be made. (3) Infection data are modeled to
predict landscape areas with high VBZD risk. A discussion of various aspects of VBZD
research and advances in light of the biological revolutions follows.
Field collections
While the processing of samples in the laboratory and spatial analyses of data
have become sophisticated with the tools of the biological revolutions, the methods of
collecting field specimens have remained simple, effective, and relatively inexpensive.
For example, the drag cloth is a tool used for collecting host-seeking ticks, and it is a one-
meter-square piece of corduroy cloth that is light in color and is dragged over the low-
lying vegetation behind the field worker as he/she hikes through the forest or field
(Sonenshine 1993) (Figure 1.2). Host-seeking (questing) ticks respond to the cues of
carbon dioxide, heat, and movement of the incoming host, and attach to the large surface
area of the cloth upon contact. Ticks can easily be observed against the light cloth, and
are plucked off and preserved for laboratory work. The most common mode of collecting
mosquitoes is using a light trap. The original light trap design, the New Jersey Light
13
Trap, was made in 1942; in 1962, the design was revised and marketed as the Centers for
Disease Control and Prevention miniature light trap and it has not changed substantially
since then. This trap uses a battery to power a small fan which created a drafi to blow
mosquitoes into a mesh bag or container, and so long as the trap is running, the
mosquitoes cannot escape. A small light bulb is used as an attractant to draw mosquitoes
into the trap, and collections can be increased by ‘baiting’ the trap with dry ice which
emits carbon dioxide to mimic host respiration (Newhouse 1966). The trap is hung in a
tree at a height that is dictated by the research or surveillance questions (exposure of
humans versus canopy-dwellers to mosquitoes; Figure 1.2).
Molecular Revolution
The gold standards for diagnosing infection in a sample have indeed changed
since the germ theory of disease was described in 18605 and culture of organisms was the
mainstay of pathogen testing. With the availability of new molecular detection tools,
culture-based assays are now often replaced by direct detection of pathogen genetic
sequences using polymerase chain reactions (PCR) and genetic characterization by
sequencing which affords strain-level analysis. The need for culture to confirm
laboratory findings is often still a prerequisite for meeting case definitions for states (i.e.
Lyme disease diagnosis in non-endemic states requires culture of spirochetes from the
expanding erythema migrans rash as well as documentation of an inerease in antibody
titer). Molecular techniques do not require that the pathogens be alive for detection and
characterization, and thus such techniques are useful in human diagnosis even after
prophylactic antibiotics have been given and organisms have been killed.
14
The molecular revolution in a broad sense began in 1953, when the structure of
DNA was described by Watson and Crick, after which the field of molecular biology
emerged. VBZD research is inextricably partnered to the molecular biology field for the
diagnostic assays that are needed to detect and characterize pathogens. Hi gh-throughput
and field-amenable assays have advanced VBZD surveillance and research, allowing for
a high sensitivity of pathogen detection without the time and labor associated with culture
and thus allowing for timely communication of epizootics and epidemics among public
health agencies. Uncultivable VBZD pathogens have been identified using molecular
tools (Relman 1990; Bhattacharya et al. 2002) Such molecular approaches in VBZD
research, however, must be used in conjunction with ecological data on vector and host
distribution (Chaves et al. 2007). The following review highlights selected tools of the
molecular revolution that have particularly advanced VBZD research.
In the 19705, the field of serology developed, providing a suite of assays that
utilize the specificity of the binding between antigen and antibody for detection of
antibodies to specific pathogens within blood. These irnmunoassays assess a patients
exposure to a pathogen as opposed to detecting the pathogen itself. Serological assays
include fluorescent antibody tests and enzyme linked immunosorbent assays (Engvall
1971) which involve the sequential layering of antigen, patient sera, and conjugated
secondary antibodies in wells of a slide or microtiter plate with an absorbent surface.
Presence of specific antibodies in an unknown sample is indicated by a color change
upon addition of a substrate to the well which reacts with the conjugated secondary
antibody, if bound to primary antibody. Serological assays are useful in VBZD
surveillance because direct pathogen detection is limited by the usually short duration of
15
viremia, parasiternia, or bacteremia within a host, and by the seasonal activity of vectors,
whereas antibodies against pathogens are generally long-lived and provide a more
comprehensive assessment of pathogen presence in an area.
Polymerase chain reaction (PCR) was developed in the 1980sas a method for
amplifying a particular segment of an organism’s genome (Mullis 1987). In VBZD
research, PCR is used to test a vector or host tissue or blood for the presence of pathogen
DNA. Extracted genomic DNA from the organism is mixed with primers specific for a
segment of the pathogen’s genome, a DNA polymerase which synthesizes the region to
be amplified, and deoxynucleotide triphosphates for building new DNA. PCR takes
place in a therrnocycler, in which fluctuating temperatures allow DNA denaturation,
primer annealing to the single-stranded DNA, and synthesis of new DNA that is
complementary to the template. A typical reaction of 40 amplification cycles can be
completed within hours and results in exponential amplification. Primers can be chosen
to be highly specific to a single strain or species of pathogen, to be general so as to
amplify all bacteria (16S ribosomal RNA primers), or to amplify highly variable regions
within a pathogens genome for further characterization. PCR products are visualized
using gel electrophoresis which separates the amplified DNA based on fragment size.
Successful amplification of the pathogen of interest will appear as a band at the expected
fragment size upon comparison to a DNA ladder. Multiplex PCR employs multiple sets
of primers within a single reaction for simultaneous detection or multiple pathogens.
Real-time/quantitative PCR (rt-PCR/qPCR) provides more information that
traditional PCR through a real-time assessment of the amplification status of each
sample. Developed in 1993 (Higuchi 1993), rt-PCR/qPCR involve the addition of a
16
fluorescent-labeled DNA probe to the template, primers, deoxynucleotide triphosphates,
and DNA polymerase that are used in traditional PCR. During the annealing stage of the
reaction, the probe and primers bind to the template, causing the probe to degrade and
fluorescence to be emitted. The amount of bound probe is quantifiable based on the
amount of fluorescence emitted after each amplification cycle. One can observe the
amplification after each reaction cycle using a video camera in real-time, and the cycle
number during which a sample amplifies above threshold to be considered positive can
be determined. Thus, rt-PCR/qPCR provide an assessment of how positive a sample is.
With a carefully made standard curve of known pathogen load, one can determine
absolute quantities of pathogen within samples. The quantitative information provided in
qPCR can be used to identify key host and vectors with high pathogen loads, as well as to
explore effects of pathogen aggregation and ‘super-spreaders’.
While pathogen detection may be adequately accomplished through amplification
of one conserved gene target, research questions that require genetic characterization for
strain differentiation may require amplification and sequencing of a gene target that is
variable among strains (such as internally transcribed spacers; ITS). Better
characterization yet may be gained in sequencing multiple genes, and this forms the basis
of multilocus sequence typing (MLST), developed in the 19903 (Spratt 1999). MLST
creates allelic profiles to distinguish bacterial genotypes through sequencing of seven
house-keeping genes that have sufficient variation due to mutation or recombination to
provide many alleles per locus. Sequences at these loci are concatenated and provide a
unique allelic profile which can be compared to libraries of allelic profiles for pathogens
available on the intemet (www.pubrnlst.org; www.mlst.net).
17
Microarrays, also known as gene chips, are a method for characterizing testing for
the presence of multiple pathogens or pathogen strains simultaneously and can be
designed for DNA, RNA, proteins, and antibodies, and have recently been used to
characterize VBZD agents (Liang et al. 2002; Broekhuijsen et al. 2003). A DNA
mircoarray is a slide or other substrate to which a collection of up to thousands of short
segments of DNA (probes) is affixed. An amplified PCR product is applied to the array
and DNA hybridization occurs to the individual probes with complementary bases. This
specific binding can be observed through fluorophores. Similarly, an antibody
microarray uses a collection of antibodies for antigen detection, and arrays are being
developed for characterization of antibodies in reservoirs of VBZD agents.
F ield-formatted assays are available for detection of the agents of visceral
leishmaniasis (fluorogenic probe-based PCR assay; (Quispe-Tintaya et al. 2005), West
Nile virus and St. Louis encephalitis (dipstick antigen detection assay; (Ryan et al. 2003),
and malaria (dipstick antigen detection assay; (Sattabongkot et al. 2004). F ield-formatted
assays for detection of antibodies to VBZD agents are available for visceral leishmaniasis
(Sundar 1998), and for Lyme disease, anaplasmosis, and ehrlichiosis (Stone et al. 2005).
These tools often provide results within minutes and allow for public health or vector
abatement districts to perform surveillance and respond immediately to increases in
infection, without a time lag for sample processing.
While the tools of the molecular revolution allow for rapid sample processing and
test results, careful interpretation of results and relation back to the ecology of the disease
system allows advancement of the VBZD field. If a pathogen is detected in a vector or
host using PCR, this does not necessarily imply that the pathogen is present in a sufficient
l8
load to be infectious. Conversely, if a pathogen is not detected, it may be the case that
the pathogen is truly absent, or, that the pathogen is present but in a load that is below the
detection threshold of the assay. For proper interpretation, the sensitivity and specificity
of the assays as well as the epidemiologically meaningful pathogen loads must be known.
A parallel technological advance has occurred to allow for the organization,
analysis, and dissipation of results from the molecular advances. Tenned bioinforrnatics,
advances in computational and statistical methods have been made to allow for analysis
and interpretation of biological data. Software and hardware have been developed that
use mathematical techniques to manage the unprecedented amounts of molecular
information that is generated with the tools of the molecular revolution, such as genome
sequencing. Genome-level data offer the ability to better define vector-host-pathogen
relationships, and to identify potential targets for control of pathogen transmission. The
genomes of some vectors are now known, including Anopheles gambiae (Holt et al.
2002), the mosquito vector of the malaraia parasite, and Aedes aegypti (N ene et al. 2007),
the mosquito vector of the yellow fever, chikungunya, and dengue viruses, and efforts are
underway for the sequencing of the genome of Ixodes scapularis (Hill and Wikel 2005),
the tick vector of the agents of Lyme disease, anaplasmosis, and babesiosis, and Culex
pipiens (https://www.broadinstitute.org/annotation/genome/culex_pipiens.4/Home.html),
the mosquito vector of West Nile virus. Of particular interest to VBZD researchers,
bioinforrnatics tools allow one to search the published genomes of thousands of
organisms, identify genes, and align similar sequences that may contain mutations,
deletions, and insertions. Large databases of identified genetic sequences are available
for searching such as GenBank, an open-access, web-based database of nucleotide
19
sequences hosted by the National Center for Biotechnology Information as a forum for
deposition of annotated sequence information from laboratories across the world.
GenBank is able to be searched using BLAST (the Basic Local Alignment Search Tool)
to determine the similarity of unknown gene sequences to identified and published
sequences (Altschul et al. 1990). Libraries of MLST allelic profiles are now available in
similar open-access format (http://pubmlst.org).
In addition to the tools for managing sequence information, significant computer-
based advances have been made for surveillance and reporting of VBZD. Web portals
are increasingly used for consolidation of surveillance data from different abatement or
public health districts, and this internet-based sharing of data allows for efficient response
to changes in prevalence. ArboNET, hosted by the Centers for Disease Control, is a
national electronic surveillance system to assist states in monitoring trends in mosquito-
borne disease.
Tools of the molecular revolution at work: the bloodmeal analysis
Identification of key hosts in VBZD systems allows for targeting of management
to reduce transmission. The bloodmeal analysis provides an example of a process in
VBZD research that has become increasingly sophisticated and informative with the
advent of the tools of the molecular revolution (see (Kent 2009) for review). Many
vectors are catholic (ie generalists) in their feeding behavior; that is, they are able to
successfully blood-feed on a variety of vertebrate host species, either opportunistically
(based on host availability) or specific to host class. Bloodmeal analysis is the process by
which the blood in a vector’s abdomen (that has been consumed from a host during blood
20
feeding event) is identified to taxonomic group of the host. When used in conjunction
with data on host availability, the bloodmeal analysis allows for assessments of vector-
host relationships and host selectivity by vectors; when used in conjunction with infection
data, routes of pathogen transmission may be identified. Many VBZD agents are
transmitted from host to vector when a host is systemically infected and a vector
bloodfeeds on the host, thus in-taking a bloodmeal containing pathogens. If the vector
subsequently engages in a second bloodfeeding event on a different host, then the
pathogens may be transmitted from vector to host. After each bloodfeeding event, the
vector must rest to digest the meal; for example, mosquitoes rest and use blood nutrients
to brood eggs; ticks use blood resources to molt into the next life stage.
In a primitive bloodmeal analysis, a bloodmeal may be dissected from the vector
and viewed microscopically- the lack of nuclei in erythrocytes is indicative of a
mammalian-derived bloodmeal whereas presence of nuclei indicates avian or reptilian-
derived bloodmeal. This technique is quite limited in its utility and a high-quality blood
smear fiom a freshly-engorged vector is required for cell visualization. In the mid
1970’s, the science of bloodmeal analysis advanced to use serology-based tests (Tempelis
1975). Serology-based tests apply the vector bloodmeal against a library of polyclonal
antibodies to candidate host species. If the bloodmeal host species was represented
within the library, then antigen-antibody binding occurs and the host can be identified.
However, these serology-based tests were often useful in identifying only to the
taxonomic order of the host due to cross-reactivity; furthermore, one must create the
library of all candidate hosts which may not be comprehensive. In the late 1990s, PCR-
based bloodmeal analyses were developed which, used in conjunction with heteroduplex
21
analysis (PCR-HDA), allowed for genus-specific bloodmeal identification (Boakye et al.
1999). After DNA extraction of the bloodmeal, PCR is used to amplify a highly-
conserved region of the mitochrodrial genome and the amplicon is mixed with a driver of
known identity. The mixture is denatured and then cooled for heteroduplex formation.
Samples are electrophoresed and relative mobility is compared to a set of known-identity
standard samples. As with the serology-based bloodmeal analysis, the PCR-HDA
approach also requires a library of standards whose acquisition from the field and
preparation in the lab is labor-intensive, and HDA patterns that are not represented within
the standards are not identifiable.
Modern bloodmeal analyses often use PCR of a mitochondrial DNA (e.g. the
cytochrome b gene), which is more useful than genomic DNA due to the high copy
number and the high level of divergence between species seen in mitochondrial genes.
The high sensitivity of PCR-based assays allow for partially-digested mosquito
bloodmeals to be used. PCR amplicons are then run against a series of DNA probes of
known identity, or are sequenced and compared to published sequences of known
identity. This approach to bloodmeal analysis has been employed to identify the host of
ticks from a previous life stage, when only small amounts of the bloodmeal remain
(Kirstein and Gray 1999; Pichon et al. 2003).
The Geographic Revolution
While the heterogeneous distribution of human diseases and their association with
certain biotic and abiotic landscape features have long been noted, the ability to quantify
such relations has advanced only recently with the new tools of the geographic
22
revolution. The science that characterizes the relationships between disease and
geography is referred to as landscape epidemiology, spatial epidemiology, medical
geographies, or health geographies. Efforts to detect clusters of human infection in
relation to the environment date to the cholera outbreak at the Broad Street Pump in
1854, in which John Snow mapped the street address of infected humans to detect their
concentric distribution about a water pump; infection ceased upon closure of the pump.
Russian parasitologist Evangy Pavlovsky (1884-1965) is credited for bringing attention to
the natural nidality of VBZDs, and notes the ‘propensity of the infection for definite
geographic landscapes’ and the ‘seasonality of the disease’ (Pavlovsky 1966). Many
VBZD systems involve a natural, sylvatic transmission of a pathogen among its reservoir
hosts, and human disease results only after anthropogenic perturbation to the
biogeoscenose, or translocation of the vectors, hosts, or pathogens to novel areas. Tools
of the geographic revolution include global positioning systems (GPS), remote sensing
(RS), geographic information systems (GIS), and spatial statistics (see review by Kitron
(1998). These tools are used in a complementary fashion in VBZD research to collect,
describe, analyze, and predict disease across a landscape.
GPS is the technology that allows researchers to determine the precise spatial
location of an entity using a handheld unit that receives signals from multiple satellites of
known locations orbiting the earth. In VBZD research, a GPS may be used on the
ground, for example, to map the locations of abiotic environmental features associated
with vector production such as storm water sewers which may be breeding sites for
mosquitoes (Irwin et al. 2008).
23
RS utilizes satellites to measure the amounts of electromagnetic radiation
reflected from different objects on the earth, and the output is a series of spatially-linked
values. Different features on the landscape are associated with a unique radiation signal
and so they may be identified based on their patterns after the satellite signals are
calibrated in relation to ground-based variables. Remote sensing has made possible the
creation of widely-available spatially-referenced datasets on landuse/landcover, and the
most common application of RS to VBZD research is in the identification of particular
landscape features and climactic features (Rogers and Randolph 2003). In many
developing countries where epidemiological or survey data lack, satellite sensors are
often used to gather environmental data that may be used to predict infection risk so that
control efforts may be efficiently targeted (Brooker et al. 2002). Fourier time-series
analysis is a technique for extracting cyclical temporal patterns from remotely-sensed
environmental data through creating a set of Fourier variables that describe the space-
time features of a landscape (Rogers and Randolph 2003).
GIS is the technology that forms the basis for the geographic information science
discipline, and consists of computer programs originated in the 19603 for displaying
layers of spatially-referenced data for map-making or more advanced statistical analyses
of spatial patterns. GIS is increasingly important in the epidemiology field (see review
by Clarke et al. (1996)). The overlay process is inherent to data processing in a GIS, in
which a series of individual sets of spatial data and attributes are matched based on
location, and these layers are then overlaid so they may be simultaneously visualized.
For example, range maps of reservoir hosts may be overlaid with those of a vector for a
given VBZD, and areas of overlap may be at risk of pathogen transmission. Whereas the
24
classic epiderrriologic focus is on the non-spatial characteristics of disease (age, sex,
time), GIS allows for investigation of the spatial characteristics of disease (Pfeilfer and
Hugh-Jones 2002). Within a GIS framework, one can identify the biotic and abiotic
environmental features that are associated with vector, host, or pathogen presence within
endemic areas. Modeling then allows for these same features to be identified in non-
endemic areas, and this allows assessment of the available suitable habitat in which
invasion of the disease system is likely.
Spatial statistics can be used in conjunction with the above tools to describe and
predict the distribution of vectors, hosts, infection, and suitable habitats. Statistical
analyses of spatial patterns often include measures of spatial autocorrelation.
Autocorrelation occurs when a value at a specific location is dependant upon the values
of neighboring locations such that neighboring locations are most similar to each other.
Kriging is a technique in which values for unsampled sites are generated based on
interpolation of values from the sampled sites. Two statistical methods that are
commonly employed to predict disease in relation to landscape features are logistic
regression and discriminant analysis, followed by an evaluation of the models
performance through comparison of the agreement between model predictions and actual
field observations (Brooker et al. 2002). These methods will be illustrated below in the
description of steps in the risk—mapping process.
Tools of the geographic revolution at work: risk modeling
The risk modeling process serves as an example in VBZD research of the
application of the tools of the geographic revolution to obtain a product that is useful to
25
understand disease risk. Risk models may be used to illustrate disease locations for
alerting medical communities and the public to risky areas, test hypotheses, identify
knowledge gaps, or prioritize and evaluate surveillance and control efforts (Kitron 2000).
The following steps are those that are generally followed for developing a VBZD risk
model, once the purpose of the model is defined. (1) Data acquisition. Datasets to be
modeled must be collected- usually through a combination of field surveys, compilation
of human case reports, and downloading of remotely-sensed images. These spatial data
are then overlaid within a GIS environment. (2) Modeling. Logistic regression modeling
is used to predict group membership; in the case of risk modeling, logistic regression is
used to evaluate the association between the probability that an area is at high risk for a
disease and various landscape features. The landscape features serve as the independent
variables, and the dependent variable of the model is discrete (such as high risk or low
risk). Multiple models are constructed that use different combinations of landscape
features to predict risk, and the most parsimonious model is selected using Akaike’s
Information Crietrion (AIC; (Akaike 1974). It is important to consider the next necessary
step- model validation- before making the model, validation efforts may require a subset
of data to be reserved (not used in model creation) for use in validation. For example,
Eisen et al. (2007) used 80% of available data to build their model of habitat suitablility
for plague, and reserved the remaining 20% for model evaluation; Brooker (2002) based
their model of schistosomiasis infection of a randomly-selected 50% subset of available
data, and used the remaining 50% for ROC analysis. (4) Model validation. Receiver
Operating characteristic curves (ROCs) can be used to assess the ability of the model to
discriminate between binary categories (ie high and low risk for human disease; high and
26
low habitat suitability for vector establishment) by plotting sensitivity (true positives;
cases in which observed disease falls within an area predicted by the model to by high
risk) against one minus specificity (false positives; cases in which the model predicts an
area to be high risk fiom which no disease is observed). The area under this curve
provides a measure of the accuracy of the model (see review by Brooker et al. (2002)).
Comparing the area under the ROC that results from different thresholds for assigning
predictions into the binary categories provides a basis for threshold selection and
improves the predictive accuracy of the final model. One can use the ROC analysis not
only to evaluate a model for the geographic region upon which it was developed, but also
to test how universal the model is when extrapolated to a new geographic area.
Some caution must be taken when interpreting risk maps. For example, areas
without data are not often apparent in a map and area boundaries are not often
biologically meaningful (Kitron 2000). Addressing temporal change in a risk map is
often difficult, and maps quickly become outdated when VBZD systems are expanding in
range. A fundamental problem exists in using the above steps to create and validate a
risk map for an emerging disease system (noted by (Eisen et al. 2007);(Ostfeld et al.
2005)): habitat patches without the disease system may be so because they are not
suitable for establishment of the vector, hosts, or pathogen; alternately, patches without
disease may be comprised of fully suitable habitat, but the disease system invasion has
simply not yet reached this area. While the former scenario represents predictive power
of the model, the later scenario may result in false positives. In the risk-mapping .
procedure, there is an inherent assumption that the vector or pathogen being modeled is
distributed everywhere that is suitable. However when a disease system is invading, this
27
assumption is violated. To circumvent this, one may restrict ground-truthing efforts to
the geographic range which is known to have already been invaded (see Foster (2004).
Furthermore, suitable habitats may be undergoing active control measures, and thus
seemingly suitable patches may not support the disease system due to the effect of the
control. A selective review of research studies in which VBZD systems are mapped to
explain current disease and predict firture disease distribution is presented by Ostfeld et
al. (2005).
VBZD Vignettes
Below, I present a review of four VBZD systems that are emerging over different
time scales, and whose study employs the tools of the molecular and geographic
revolutions to predict risk or elucidate new relationships which improve our
understanding of transmission and emergence. Each vignette is not intended as a
comprehensive review of disease etiology and transmission, but instead provides a brief
background to the system and highlights selected contemporary ecological research
programs that apply new molecular and geographic analytical tools yield novel insight.
Vignette #1: Plague
Plague is an ancient VBZD responsible for many past epidemics including three
great pandemics: the Justinian Plague in AD. 541-542, originating in Ethiopia or Egypt;
the Black Death in 1347-1351 in Eurasia; and the Modern Pandemic that began inlthe
China in 1855 (Perry and Fetherston 1997). Unique to the Modern Pandemic is the vast
geographic range expansion of disease across all continents, afforded by the movement of
28
infected rats and fleas on trade ships. Human plague continues into present day in Africa,
Asia, and the Americas, with an average of 2821 cases reported per year for the time
period of 1990-1999 of which over 80% are from Africa (Weekly Epidemiological
Record, WHO). While human plague is indeed an emerging VBZD, with emergence on
a slower time scale compared to most emerging VBZDs, enzootic foci of plague in
rodents and fleas have long been established and transmission occurs continuously at
many foci that does not regularly spill over and cause human disease (Duplantier 2005).
The disease is able to spread rapidly during epizootic periods- when fleas leave reservoir
hosts and feed on incidental, amplifying hosts that are susceptible to disease (review by
(Gage and Kosoy 2005).
The etiologic agent of plague is the gram-negative bacterium Yersinis pestis that
is maintained in nature in transmission cycles among rodents, lagomorphs, and their
fleas. There are three Y. pestis biotypes that correspond with the sites of origin of the
three pandemics (Guiyoule et al. 1994). Y. pestis was introduced to North America in
1900, and the strains of Y. pestis orientalis in North America are a subset of those found
in Asia- the putative source of invasion (Gage and Kosoy 2005). Y. pestis is now under
watch as a bioterrorism agent, and has been used as such in the past through the
deliberate exposure of infected human or animal carcasses, release of infected laboratory-
reared fleas, or weaponized pneumonic plague.
Two mechanisms of Y. pestis transmission have been identified: flea-borne and
aerosol droplet. In the former, infection of a susceptible host may result after a
mechanical transmission from contaminated flea mouthparts or feces, direct ingestion of
infected fleas, or, most commonly, regurgitation of bacteria by blocked fleas. Blocked
29
flea transmission occurs after a flea feeds on an infectious host and Y. pestis organisms
amplify in the fleas midgut and form a colony sufficient in size to form a blockage in the
mouthparts of infected fleas (Gage and Kosoy 2005). Subsequent bloodfeeding attempts
by the flea are unsuccessful because the bacterial mass prevents the passage of blood
from the proventriculus to the rnidgut, and the flea starves after repeated attempts. The
increase in bloodfeeding attempts allow for Y. pestis transmission when the flea
regurgitates the blood and Y. pestis organisms from the mass into the feeding site of the
host. Transmission via blockage is possible only after a long extrinsic incubation time (5
days to weeks depending on flea species), which contrasts with the relatively rapid rate of
epizootic spread observed in plague outbreaks. Early-phase transmission from fleas in
the absence of block formation has thus recently been identified as a transmission
mechanism (Eisen et al. 2006).
While many mammals may become infected with Y. pestis upon the bite of an
infected flea, the contribution of each host to the maintenance and spread of the pathogen
differs. Reservoir hosts are those hosts that maintain sylvatic foci of transmission at a
baseline level with their fleas and are able to do so in the absence of other hosts.
Reservoir hosts perpetuate the pathogen in inter-epizootic periods. Specific
characteristics qualify a host as a reservoir, including the ability to circulate high bacterial
titers in their blood; heavy infestations by flea vectors; and residence in burrows or nests
with heave flea infestations (Gage and Kosoy 2005). In contrast to reservoir hosts,
epizootic hosts are usually naive to the pathogen, and suffer high mortality upon infection
which may initially result from spill-over from the sylvatic cycles (shared fleas) and
subsequently develop into an epizootic. In North America, reservoir hosts for Y. pestis
30
include Peromyscus and Microtis species, and epizootic hosts include prairie dogs,
ground squirrels, chipmunks, and woodrats (Gage and Kosoy 2005). Epidemics result
when humans contract infection from fleas after encroaching in sylvatic or epizootic
transmission cycles— in the latter case, fleas may seek humans as a host when their
epizootic hosts die. Y. pestis infection in humans manifests in three ways in humans:
bubonic plague is characterized by swollen lymph nodes called buboes at the site of flea
bite, and is the most common form of disease in humans. Pneumonic plague is
characterized by Y. pestis infection of the lungs and person-to-person transmission is
possible via respiratory droplet aerosols— pneumonic plague is likely to result if plague is
weaponized and used in bioterrorism. Septicemic plague is characterized by Y. pestis
presence in the blood and may be accompanied by gangrene of the extreminities. Both
pneumonic and septicemic plague may develop from bubonic plague.
Reasons for the resurgence of human plague include the discontinued surveillance
for enzootic transmission once human cases cease; political unrest and other socio-
economic factors that result in poverty and unhygienic living; deforestation that puts
humans in contact with sylvatic cycles or creates new habitats for enzootic hosts and their
fleas; introduction of new reservoir hosts to enzootic cycles; and evolution of the
pathogen to increase host range or transmissibility (see review by Duplantier et al.
(2005)).
Tools of the molecular and geographic revolutions have together provided a better
understanding of plague transmission cycles, leading to prediction of zones of increased
risk of transmission. For example, Woods et al. (2009) developed a rapid bloodmeal
analysis using multiplex PCR to identify the host species for fleas that is sensitive up to
31
72 hours after a flea has fed. Using this technique on fleas collected in a pilot study in
peridomestic settings in eastern Afiica, where a majority of human cases occur in present
day, the authors identified humans as the most common host across five species of flea in
the Congo, followed by chickens and cats, and many fleas contained mixed bloodmeals
with two of these host species. Importantly, this technique allowed for identification of a
rat bloodmeal in a flea species that is typically associated only with humans, which
suggests a possible mechanism of zoonotic transmission that does not require the
presence of infectious rat fleas for bridging the pathogen from rats to humans. Eisen et
al. (2007) used the logistic regression model approach to predict that 14.4% of the four-
comer region of the southwestern United States has elevated risk of human plague.
Using the locations of human plague cases in from 195 7-2004 within an endemic focus of
disease as the dependent data in the model, high-risk plague habitat was identified based
on the presence of key habitat types at specific elevations (southern Rocky Mountain
pinon-juniper, Colorado plateau pinon-j uniper, Rocky Mountain ponderosa pine, and
southern Rocky Mountain juniper at elevations up to 2300m).
Vignette #2: Lyme disease in the American Midwest
Ixodes scapularis, the blacklegged or deer tick, is distributed discontinuously
throughout the eastern half of the United States, and is the vector of a suite of zoonotic
pathogens across its range, of which Borrelia burgdorferi, agent of Lyme disease, is the
leading cause of reported vector-borne disease in the United States (23,305 cases reported
in 2005; MMWR 2007). A majority of cases of human Lyme disease are reported from
two endemic foci in the northeastern and upper Midwestern United States, but new
32
evidence from the field suggests that these ticks are invading new areas from endemic
foci. The initial epidemiological investigation of Lyme disease occurred in the 19705
when a cluster of children were misdiagnosed with juvenile rheumatoid arthritis in Old
Lyme, CT, later to be associated with tick exposure and the presence of the B.
burgdorferi pathogen. This pathogen is now known to have occurred in the eastern
United States since at least the late 19th century (Steere et al. 2004). A series of changes
in the landscape occurred that were favorable to the Lyme disease emergence, and
continue today, such that Lyme disease remains the most conunon VBZD in the United
States. Lyme disease is a significant source of human morbidity throughout Europe as
well, where the tick species and pathogen species, host associations, and disease
manifestation are different.
I. scapularis exhibits a two-year life cycle across most of its range and develops
through larval, nymphal, and adult life stages (Spielman et al. 1985). During each life
stage, the tick engages in one bloodfeeding event on a host that may last from 3-8 days
after which the tick undergoes an inactive period for weeks to months and molts into the
next life stage. Different host individuals are used during each blood meal. During
bloodfeeding, transmission of B. burgdorferi may occur from tick to host or from host to
tick, with the exception that larval ticks are usually hatched uninfected, and thus are
unable to infect during the larval bloodfeeding event.
Multiple hosts are involved in Lyme disease ecology, as I. scapularis is a
generalist feeder, and B. burgdorferi is a generalist pathogen. White-footed mice,
chipmunks, birds, and lizards commonly host larval and nymphal ticks, whereas white-
tailed deer are the most important host for adult ticks. Each host species differs in
33
reservoir competency for B. burgdorferi. White-footed mice considered the most
efficient reservoir host for the pathogen, and in xenodiagnostic studies in the laboratory
demonstrate that infectious white-footed mice are able to infect 70-100% of susceptible
ticks (Donahue et al. 198 7). Conversely, white—tailed deer are unable to maintain
infection by B. burgdorferi and contain proteins of their complement system that kill the
bacterium in situ and can clear a feeding tick of its infection (Telford et al. 1988).
During European settlement of North America, deforestation and conversion of
the eastern United States to agriculture occurred, and deer were nearly extirpated due to
over-hunting. These two factors are likely responsible for removing ticks and Lyme
disease risk from the landscape, where maintenance of B. burgdorferi is hypothesized to
have occurred for millennia ((Barbour and Fish 1993; Steere et al. 2004)). Reforestation
occurred in the early 20th century, and deer, blacklegged ticks, and B. burgdorferi
expanded from refugia. As deer numbers continue to rise and support tick populations,
human land use patterns of building houses in the woods support increased contact
between humans and ticks. As ticks do not move large distances during off-host periods,
tick invasion to new habitats is a function of host movement. Mechanisms of tick
invasion may include movement of ticks by birds in their migrations from endemic sites,
movement by deer with relatively large home ranges, and a slower movement of ticks by
small mammals with adjacent home ranges.
Tools of the molecular and geographic analytical revolutions have lead to some
changes in our understanding of Lyme disease invasion, distribution, and maintenance.
For example, multilocus sequence typing has been used to infer a European origin of the
Lyme disease pathogen Borrelia burgdorferi (Margos et al. 2008) whereas earlier
34
research suggested a North American origin (Foretz et al. 1997; Marti Ras et al. 1997). A
national survey for nymphal I. scapularis was recently conducted across the range of I.
scapularis in the eastern half of the United States (Diuk-Wasser et al. 2006). Data on the
density and infection prevalence of nymphal ticks has been modeled to predict nymphal
density using remotely-sensed data and interpolated weather station data, which
generated a continuous probability surface of the risk of encountering nymphs in the US.
This model will be useful in identifying high risk areas for human exposure to infected
nymphs, which can aid in decisions about personnel protection measures, residential or
community intervention, and for vaccination programs (Diuk-Wasser et al. 2010).
Furthermore, new relationships in this VBZD system have been revealed by Gatewood et
al. (2009), who analyzed the relationships between climate, seasonal activity of I.
scapularis, and B. burgdorferi genotype frequency across the northeastern and
Midwestern United States. The authors found that the degree of seasonal synchrony of
the larvae and nymphs was predicted by the magnitude of the difference between summer
and winter daily temperature maximums. This finding has potential consequences for
human disease, because particular human-invasive pathogen strains were more common
among tick populations characterized by low seasonal synchrony.
Guerra et al. (2002) modeled the habitat suitability for the Lyme disease vector
Ixodes scapularis in the upper Midwestern United States. The authors used both
discriminant fimction analysis and logistic regression to identify the environmental
features that best characterized the difference between tick positive and tick negative
field sites. Soil order and landcover were identified as the main contributors to tick
presence; in particular, sandy to loamy sand soils with deciduous, dry-mesic forests were
35
significant. The authors then mapped where these significant habitat types occurred to
create a map of habitat suitability for the tick, which was 86% correct in classifying tick
status at sites. This model has been projected to new areas (lower Michigan) and has
been shown to maintain high predictive capacity in an independent test (Foster 2004).
Bloodmeal analyses have been used to identify the hosts of the Lyme disease
vector tick in Europe. Bloodmeal analysis of flat ticks to determine the host feed upon
during the previous bloodmeal is particularly challenging because the duration of time
since bloodfeeding is on the order of months, even up to a year, and the remaining host
DNA is therefore degraded (Randolph 2009). Using PCR of the 128 rDNA
mitochondrial gene (Humair et al. 2007), Cadenas et al. (2007) successfully identified the
host of 43.6% of questing ticks, with a range in the identification rate of 20-93%
depending on month of tick collection. Combining bloodmeal analysis with Borrelia
detection, the authors confirm the associations of B. afzelii and B. burgdorferi 3.3. with
rodents, and B. valaisiana and B. garinii with birds.
Vignette 3: Tick-home encephalitis in Europe
Tick-bome encephalitis (TBE) is caused by a flavivirus like dengue, West Nile,
and Japanese encephalitis viruses. Ixodes persulcatus and I. ricinus are the zoonotic
vectors for the more pathogenic Eastern and Western subtypes, respectively (Dumpis et
al. 1999). Climates and landscapes that promote greater seasonal overlap of nymphal and
larval host-seeking support more intense enzootic transmission and greater TBE risk
(Randolph et al. 1999). This is because the maintenance of TBE virus requires non-
systemic transmission from infected nymphs to larvae of the next cohort (Labuda et al.
1993a). There are no vertebrate species that can maintain TBE virus systemically.
36
Rodents (mainly Apodemus spp.), however, can support non-systemic TBE virus
transmission, and because they feed both stages, they are critical for TBE virus
maintenance in nature (Labuda et al. 1993b; Randolph et al. 1999). Thus, TBE virus
prevalence in ticks is much lower (0.5-5%, (Dumpis et al. 1999) compared with that of
B. burgdorferi sensu lato (the agent of Lyme borreliosis, 25-50% (Hubalek and Halouzka
1998), which systemically infects many host species throughout its range.
The incidence of TBE over the last two decades generally has increased in Europe
both within its endemic regions in Central and Eastern Europe (Sumilo et al. 2007) as
well as in Western Europe (Broker and Gniel 2003). In a series of papers, Randolph and
colleagues have scrutinized the causes for this increased incidence as putative factors
include i) warmer climate, ii) improved surveillance and diagnostic assays, and iii)
political changes. Using satellite imagery and field data from seven countries in Europe,
Randolph et al. (2000) examined the activity of immature stages of I. ricinus in relation
to TBE incidence. The authors found that larvae consistently started feeding and questing
earlier in the year at sites within TBEv foci than elsewhere, and the larval activity in the
spring was coincident with nymphs. Analysis of satellite-derived indices of land surface
temperature revealed that a rapid fall in temperature in the fall predicted this activity may
cause unfed larvae to overwinter with a spring emergence that is synchronous with
nymphs. However, while overall climate trends may have influenced TBE incidence in
predictable ways, the spatial heterogeneity in TBE dynamics within similar climatic
zones and ecological landscapes indicate that changing climate cannot solely explain the
dynamics of TBE (Randolph 2004; Rogers and Randolph 2006; Sumilo et al. 2007).
37
Instead, flow-on effects resulting from the breakdown of Communist rule created
landscapes more favorable for TBE enzootic transmission as well as changes in human
behavior that increased the risk of contact with infected ticks. After the transition from
Communist rule, at least two changes took place that affected the landscape to increase
tick population. Collective farming collapsed in the Baltic countries of Lithuania,
Latvia, and Estonia (Sumilo et al. 2006). Secondary succession replaced farmland and
pasture and led to a proliferation of wildlife and subsequent increase in vector ticks.
Concomitant industrial collapse resulted in warmer climates - 'brightening' - due to the
reduction of air pollutant emissions. This regional warming improved climate both for
tick survivorship and for human outdoors activities. Thus, these socio-economic changes
created more a suitable landscape for vector p0pulations. Education and the availability
of an effective vaccine allowed people to modify their risk of TBE and seemed to explain
the reduction of incidence in certain areas. Available vaccination and epidemiological
data in Lithuania and Latvia show that decreased TBE incidence is significantly
associated with increased vaccination rates two years prior and high TBE incidence three
Years prior (Sumilo et al. 2009).
Vignette 4: American cutaneous leishmaniasis
American cutaneous leishmaniasis (ACL) is a VBZD involving human infection
with protozoan parasites in the genus Leishmaniasis, in particular L. braziliensis, L.
amazonensis, L. panamensis, L. mexicana, and L. guyanesis. Infection in humans “is
multifarious, but is characterized by focal cutaneous lesions, progression to
mucocutaneous infection, and in some cases development of diffuse cutaneous lesions
38
(Weigle and Saravia 1996). The course of infection is distinct from those Leishmania
species that cause visceral leishmaniasis, although geographic distributions of the two
types of infections may overlap. Infected individuals often self-cure without therapy and
show a strong adaptive immune response involving T-cell recognition of Leishmania
antigens (Carvalho et al. 1995), leading to long term immunity but typically with
disfiguring scars at lesion sites. Enzootic transmission of these parasites occurs by bites
of a systematically diverse and geographically widespread complex of species of blood
feeding sand flies (subfamily Phlebotominae, family Psychodidae) in the New World
genus Lutzomyia. Vertebrate hosts of the parasites are wild rodents and marsupials
(opposums), edentates (slotlrs), and commensal rats. Some of these transmission cycles
are vertically stratified by forest floor and forest canopy. Transmission solely between
sand flies and humans has been suggested for ACL, but not definitively documented,
although a clinical case investigation reported large numbers of amastigotes (the parasite
stage infectious to sand flies) in skin lesions in a man with AIDS in Brazil, suggesting
that it might be possible under conditions of immunosuppression and consequent parasite
proliferation (Souza et a1. 1998).
The pristine zoonotic cycles for the parasites, vectors, and reservoir hosts occur in
widely distributed, undisturbed forests of Central and South America. Risk of infection
in humans was primarily occupational amongst forest workers and others entering the
forest environment, and was strongly skewed towards adult males (Rawlins et al. 2001).
Deforestation was accordingly predicted to decrease incidence of infections, but the
opposite has occurred; incidence of cases has been increasing steadily in Colombia,
Venezuela, Brazil, French Guyana, and northwest Argentina, the ratio of infections in
39
human males and females has equalized, and infections in children are now common
(Campbell-Lendrurn et al. 2001). This changing epidemiological pattern (not one merely
of human intrusion into pristine habitat, but environmental modification leading to habitat
disturbance and destruction) is correlated with adaptation of the entire transmission cycle
from the forest setting to a peridomestic setting localized in rural settlements. The
environmental changes are due to human population expansion and development of
agriculture; establishment of roads; gold mining; and military activities. The sand fly
vectors of some species have shifted larval habitats from forest litter associated with
trees, and sylvan rodent burrows; to human garbage and rodent burrows near houses; in
these peridomestic environments, vertebrate hosts competent for Leishmania infection are
available. Adult sand flies have adopted human dwellings for resting sites in lieu of
humid, shady forest sites. Whereas in forested settings in Brazil, sand fly feeding was
either primarily on sloths or rodents (Christensen et al. 1982), a bloodmeal analysis of
sand flies in deforested areas in Manaus, Brazil, an area with very high ACL incidence,
indicated that sand flies fed upon humans, rodents, sloths, dogs, and domesticated fowl
(N ery et al. 2004). Layered over the long term environmental changes and correlated
epidemiological consequences are strong space-time clusters of human infection, with
epidemics postulated to occur at intervals related to waxing and waning of immunity in
the human population even while overall incidence has increased; and geographic
localizations of infection regionally. The anthropogenic environmental changes
associated with emergence of ACL from pristine forest environments are consistent with
predictions of “drivers” of emergence for vector-borne diseases in general identified by
Harms and Baneth (2005), as well as with processes of emergence associated with habitat
40
modification and enhanced transmission or “parasite flow” (Daszak et al. 2001; Polley
2005).
The revolutions reviewed above have contributed substantially to elucidation of
the peridomestic emergence of ACL in South America. Parasite infection in sand flies, in
sylvatic vertebrate hosts, and in humans has been facilitated by application of PCR-based
methods for parasite detection (de Bruijn and Barker 1992; Rodrigues et al. 2002;
Mendonca et a1. 2004), and molecular-based methods have been widely applied to
elucidate systematic relationships and parasite phylogenies, trace evolutionary histories,
examine species and strain associations with disease manifestations (Cupolillo et a1.
1994). While New World sand fly identification and systematic relationships have
remained largely morphological (Young and Duncan 1994), new methods of ribosomal
spacer region sequence similarity have clarified species boundaries within the context of
geographic distributions (Beati et al. 2004).
Conclusion
VBZD research is united in its underlying motivation of reducing human
morbidity and mortality. Since the 19705, the VBZD research field has benefited from
revolutions in advent of tools to allow for diagnostics and assessment of infection across
landscapes. Molecular-based tools, especially PCR and gene sequencing technologies,
allow for determination of infection prevalence and strain-typing of pathogens in vectors
and hosts. Many such tools are now formatted for use in the field and are complemented
by web-based portals for data—sharing to allow for rapid response to epidemics and
epizootics. Geographic analytical tools, especially remote-sensing and GIS, provide
41
access to, and a platform for analysis of, spatial information across a spectrum of scales
to analyze global as well as local patterns of disease.
With these tools, it is possible to not only describe disease retrospectively, but to
predict disease across space and time. This has been done successfully for a number of
VBZD systems, a subset of which has been outlined above. The final link in achieving
the goal of reducing prevalence of human disease is to use such models to inform control
practices.
42
Table 1.1. Glossary of vector-borne zoonotic disease related-terms
Biogeocenose
an association of animals and plants in an area of the earth’s surface
together with its climate and microclimate, geologic structure, soil, and
water supply (Pavlovsky 1966).
Emergence
the process by which a pathogen or disease system becomes present in an
area, which may be due to a range expansion of the pathogen or disease
system from an endemic area to a new geographic area to which it is
novel (invasion) or through evolution of novel pathogens and host
associations within an endemic area.
Endemic
the quality of a pathogen or disease system that is native to a given area
(indigenous), or has existed in an area for a relatively long time.
Endemic is alternatively used to describe an organism that is uniquely
found in a given area, and not found naturally anywhere else.
Enzootic
the quality of a disease system being maintained within animal reservoirs
at a baseline level; disease of wildlife may or may not result.
Epidemic
an increase in incidence of disease of hmnans
Epizootic
an increase in incidence of disease of animals
Generalist
with regard to vector feeding, a generalist vector is a species that is able
to successfully obtain a bloodmeal from a variety of host species. With
regard to pathogens, a generalist pathogen is one that is able to infect a
variety of host species.
Incidental
host
An animal that may become infected with a pathogen but is not required
for maintenance of the pathogen in nature. Incidental hosts are often
referred to as ‘dead-end’ with respect to the pathogg.
Invasive
the quality of a pathogen, or disease system that is expanding in range
into areas where it previously did not exist; invasive species typically
cause damage to their new habitat.
Nidus
a geggraphic area in which a disease system is present.
Receptive
the quality of a habitat patch that contains the appropriate biotic and
abiotic factors to support the establishment of an invading species;
invasible.
Reservoir
competence
the importance of a host species to pathogen transmission and the
dynamics of infection. In VBZD studies, reservoir competence is often
measured as the proportion of vectors that become infected after feeding
on an infected host individual.
Reservoir
host
An animal that maintains a pathogen in nature and can infect vectors
upon blood-feeding.
Resurgence
the process by which a pathogen or disease system reappears in an area
from which it has historically been transmitted followed by a period of
local extirpation, or an increase in rate of transmission or virulence of a
pathogen within an endemic area.
Sylvatic
descriptor of pathogen or disease system that is transmitted among wild
animals
Zoonotic
__
a type of pathogen or disease system that is maintained in animal
reservoirs and is transmissible to humans
43
Figure 1.2. Methods for field collections of vectors. a) Drag sampling for questing ticks;
b) Checking drag cloth for tick presence. c) CDC light trap elevated in tree canopy for
mos uito collections. d) Gravid tra- for collections of ravid female mos uitoes. .
~
.., ..
7;? (N "‘7 ‘-~‘
. o4"§‘~1$‘ .
)1‘,’
n. 4
l
, .
.
\
O
._ 3 .
’\.’
‘\-L‘ j'
ihl'v:
I
References
Akaike H (1974). New Look at Statistical-Model Identification. Ieee Transactions on
Automatic Control AC19:71 6-723.
Altschul SF, Gish W, Miller W, Myers EW, and Lipman DJ (1990). Basic local
alignment search tool. J Mol Biol 215:403-410.
Anderson RM, and May RM (1982). Coevolution of Hosts and Parasites. Parasitology
85:411-426.
Bacon RM, Kugeler KJ, and Mead PS (2008). Surveillance for Lyme Disease --- United
States, 1992--2006. Morbidity and Mortality Weekly Reports 57:1 -9.
Barbour AG, and Fish D (1993). The Biological and Social Phenomenon of Lyme-
Disease. Science 260: 161 0-1616.
Beati L, Caceres AG, Lee JA, and Munstermann LE (2004). Systematic relationships
among Lutzomyia sand flies (Diptera: Psychodidae) of Peru and Colombia based
on the analysis of 128 and 288 ribosomal DNA sequences. Int J Parasitol 34:225-
234.
Begon M, editor. (2008). Efi’ects of host diversity on disease dynamics. Princeton
University Press, Princeton.
Bhattacharya S, Vijayalakshmi N, and Parija S (2002). Uncultivable bacteria:
Implications and recent trends towards identification. Indian Journal of Medical
Microbiology 20:174-177.
Boakye DA, Tang J, Truc P, Merriweather A, and Unnasch TR (1999). Identification of
bloodmeals in haematophagous Diptera by cytochrome B heteroduplex analysis.
Medical and Veterinary Entomology 13:282-287.
Broekhuijsen M, Larsson N, Johansson A, Bystrom M, Eriksson U, Larsson E, et al.
(2003). Genome-wide DNA microarray analysis of F rancisella tularensis strains
demonstrates extensive genetic conservation within the species but identifies
regions that are unique to the highly virulent F-tularensis subsp tularensis. Journal
of Clinical Microbiology 41 :2924-293 1.
Broker M, and Gniel D (2003). New foci of tick-bome encephalitis virus in Europe:
consequences for travellers fi'om abroad. Travel Med Infect Dis 1:181-184.
Brooker S, Hay SI, and Bundy DAP (2002). Tools from ecology: useful for evaluating
infection risk models? Trends in Parasitology 18:70-74.
Cadenas FM, Rais O, Humair PF, Douet V, Moret J, and Gem L (2007). Identification of
host bloodmeal source and Borrelia burgdorferi sensu lato in field-collected
Ixodes ricinus ticks in Chaumont (Switzerland). Journal of Medical Entomology
44:1109-1117.
45
Cain S (1944). Criteria for the indication of center of origin in plant geographical studies.
Torreya 43: 142-154.
Campbell-Lendrum D, Dujardin J-P, Martinez E, F eliciangeli MD, Perez JE, Silans
LNMPd, et al. (2001). Domestic and peridomestic transmission of American
cutaneous leishmaniasis: changing epidemiological patterns present new control
opportunities. Memérias do Instituto Oswaldo Cruz 96: 159-162.
Carvalho EM, Correia D, Bacellar O, Almeida RP, Lessa H, and Rocha H (1995).
Characterization of the Immune-Response in Subjects with Self-Healing
Cutaneous Leishmaniasis. American Journal of Tropical Medicine and Hygiene
53:273-277.
Chaves LF, Hernandez M—J, Dobson AP, and Pascual M (2007). Sources and sinks:
revisiting the criteria for identifying reservoirs for American cutaneous
leishmaniasis. Trends in Parasitology 23:31 1-316.
Christensen HA, Arias JR, de Vasquez AM, and de Freitas RA (1982). Hosts of Sandfly
Vectors of Leishmania Braziliensis Guyanensis in the Central Amazon of Brazil.
Am J T rop Med Hyg 31:239-242.
Clarke KC, McLafferty SL, and Tempalski BJ (1996). On epidemiology and geographic
information systems: A review and discussion of future directions. Emerging
Infectious Diseases 2:85—92.
Cupolillo E, Grimaldi G, and Momen H (1994). A General Classification of New-World
Leishmania Using Numerical Zymotaxonomy. American Journal of Tropical
Medicine and Hygiene 50:296-311.
Daszak P, Cunningham AA, and Hyatt AD (2001). Anthropogenic environmental change
and the emergence of infectious diseases in wildlife. Acta Tropica 78:103-116.
de Bruijn MHL, and Barker DC (1992). Diagnosis of New World leishmaniasis: Specific
detection of species of the Leishmania braziliensis complex by amplification of
kinetoplast DNA. Acta Tropica 52:45-58.
Diuk-Wasser MA, Gatewood AG, Cortinas MR, Yaremych-Hamer S, Tsao J, Kitron U,
et al. (2006). Spatiotemporal patterns of host-seeking Ixodes scapularis nymphs
(Acari : Iodidae) in the United States. Journal of Medical Entomology 43: 166-
176.
Diuk-Wasser MA, Vourc'h G, Cislo P, Hoen AG, Melton F, Hamer SA, et al. (2010).
Field and clirnate-based model for predicting the density of host-seeking nymphal
Ixodes scapularis, an important vector of tick-bome disease agents in the
eastern United States. Global Ecology and Biogeography in press.
46
Donahue JG, Piesman J, and Spielman A (1987). Reservoir Competence of White-Footed
Mice for Lyme-Disease Spirochetes. American Journal of Tropical Medicine and
Hygiene 36:92-96.
Dumpis U, Crook D, and Oksi J (1999). Tick-bome encephalitis. Clinical Infectious
Diseases 28:882-890.
Duplantier JD, J ., Chanteauc, S., and Carnield, E. (2005). From the recent lessons of the
Malagasy foci towards a global understanding of the factors involved in plague
reemergence. Vet. Res. 36:437-453.
Eisen RJ, Bearden SW, Wilder AP, Montenieri JA, Antolin MF, and Gage KL (2006).
Early-phase transmission of Yersinia pestis by unblocked fleas as a mechanism
explaining rapidly spreading plague epizootics. Proceedings of the National
Academy of Sciences of the United States of America 103: 15380-15385.
Eisen RJ, Enscore RE, Biggerstaff BJ, Reynolds PJ, Ettestad P, Brown T, et al. (2007).
Human plague in the southwestern United States, 195 7-2004: Spatial models of
elevated risk of human exposure to Yersinia pestis. Journal of Medical
Entomology 44:530-537.
Elton CS (1958). The ecology of invasions by animals and plants. University of Chicago
Press, Chicago
Engvall E, Perhnan, P. (1971). Enzyme-linked immunosorbent assay (ELISA).
Quantitative assay of irnmunoglobulin G. Immunochemistry 8:871 -874.
Foretz M, Postic D, and Baranton G (1997). Phylogenetic Analysis of Borrelia
burgdorferi Sensu Stricto by Arbitrarily Primed PCR and Pulsed-Field Gel
Electrophoresis. Int J Syst Bacteriol 47:1 1-18.
Foster ES (2004). Ixodes scapularis (Acari: Ixodidae) and Borrelia burgdorferi in
southwest Michigan: population ecology and verification of a geographic risk
model. Masters Thesis, Michigan State University, East Lansing.
Gage KL, and Kosoy MY (2005). Natural history of plague: perspectives from more than
a century of research. Annu Rev Entomol 50:505-528.
Gatewood AG, Liebman KA, Vourc'h G, Bunikis J, Hamer SA, Cortinas R, et al. (2009).
Climate and Tick Seasonality Are Predictors of Borrelia burgdorferi Genotype
Distribution. Applied and Environmental Microbiology 75 :2476-2483.
Guerra M, Walker E, Jones C, Paskewitz S, Cortinas MR, Stancil A, et al. (2002).
Predicting the risk of Lyme disease: Habitat suitability for Ixodes scapularis in the
north central United States. Emerging Infectious Diseases 8:289-297.
47
Guiyoule A, Grimont F, lteman I, Grimont PA, Lefevre M, and Carniel E (1994). Plague
pandemics investigated by ribotyping of Yersinia pestis strains. J Clin Microbiol
32:634-641.
Harrus S, and Baneth G (2005). Drivers for the emergence and re-emergence of vector-
borne protozoa! and bacterial diseases. Int J Parasitol 35: 1 309-13 1 8.
Higuchi R, Fockler, C., Dollinger, G., Watson, R. (1993). Kinetic PCR analysis: real-time
monitoring of DNA amplification reactions. Biotechnology 11:1026.
Hill CA, and Wikel SK (2005). The Ixodes scapularis Genome Project: an opportunity
for advancing tick research. Trends in Parasitology 21:151-153.
Holt RA, Subrarnanian GM, Halpem A, Sutton GG, Charlab R, Nusskem DR, et al.
(2002). The Genome Sequence of the Malaria Mosquito Anopheles gambiae.
Science 298:129-149.
Hubalek Z, and Halouzka J ( l 998). Prevalence rates of Borrelia burgdorferi sensu lato in
host-seeking Ixodes ricinus ticks in Europe. Parasitology Research 84: 167-172.
Humair PF, Douet V, Cadenas FM, Schouls LM, Van de Pal I, and Gem L (2007).
Molecular identification of bloodmeal source in Ixodes ricinus ticks using 128
rDNA as a genetic marker. Journal of Medical Entomology 44:869-880.
Irwin P, Arcari C, Hausbeck J, and Paskewitz S (2008). Urban wet environment as
mosquito habitat in the upper midwest. EcaHealth 5:49-57.
Kent RH (2009). Molecular methods for arthropod bloodmeal identification and
applications to ecological and vector-borne disease studies. Molecular Ecology
Resources 9:4-18.
Kirstein F, and Gray J S (1999). Blood meal identification in ticks: a promising tool in
ecological research on tick-bome diseases. Zentralblatt Fur Bakterialagie-
International Journal of Medical Microbiology V irology Parasitology and
Infectious Diseases 289:760-764.
Kitron U (1998). Landscape ecology and epidemiology of vector-borne diseases: Tools
for spatial analysis. Journal of Medical Entomology 35:43 5-445 .
Kitron U (2000). Risk maps: Transmission and burden of vector borne diseases.
Parasitology Today 16:324-325.
Labuda M, Jones LD, Williams T, and Nuttall PA (1993a). Enhancement of Tick-Bome
Encephalitis-Virus Transmission by Tick Salivary-Gland Extracts. Medical and
Veterinary Entomology 7: 1 93-196.
48
Labuda M, Stunzner D, Kozuch O, Sixl W, Kocianova E, Schaffler R, et al. (1993b).
Tick-Bome Encephalitis-Virus Activity in Styria, Austria. Acta Virologica
37:187-190.
Lanciotti RS, Roehrig JT, Deubel V, Smith J, Parker M, Steele K, et al. (1999). Origin of
the West Nile virus responsible for an outbreak of encephalitis in the northeastern
United States. Science 286:2333-2337.
Liang FT, Nelson F K, and Fikrig E (2002). DNA Microarray Assessment of Putative
Borrelia burgdorferi Lipoprotein Genes
10.1128/LAI.70.6.3300-3303.2002. Infect. Immun. 70:3300-3303.
LoGiudice K, Duerr ST, Newhouse MJ, Schmidt KA, Killilea ME, and Ostfeld RS
(2008). Impact of host community composition on Lyme disease risk. Ecology
89:2841-2849.
LoGiudice K, Ostfeld RS, Schmidt KA, and Keesing F (2003). The ecology of infectious
disease: Effects of host diversity and community composition on Lyme disease
risk. Proceedings of the National Academy of Sciences of the United States of
America 100:567-571.
Loss SR, Hamer GL, Walker ED, Ruiz MO, Goldberg TL, Kitron UD, et al. (2009).
Avian host community structure and prevalence of West Nile virus in Chicago,
Illinois. Oecalogia 159:415-424.
Margos G, Gatewood AG, Aanensen DM, Hanincova K, Terekhova D, Vollmer SA, et al.
(2008). MLST of housekeeping genes captures geographic population structure
and suggests a European origin of Borrelia burgdorferi. Proc Natl Acad Sci U S A
105:8730-8735.
Marti Ras N, Postic D, F oretz M, and Baranton G (1997). Borrelia burgdorferi Sensu
Stricto, a Bacterial Species "Made in the USA"? IntJ Syst Bacteriol 47:1 1 12-
1117.
Mather TN, Nicholson MC, Donnelly EF, and Matyas BT (1996). Entomologic index for
human risk of Lyme disease. American Journal of Epidemiology 144:1066-1069.
Mendonca MG, de Brito MEF, Rodrigues EHG, Bandeira V, Jardim ML, and Abath F GC
(2004). Persistence of Leishmania parasites in scars alter clinical cure of
American cutaneous leishmaniasis: Is there a sterile cure? Journal of Infectious
Diseases 189:1018-1023.
Mullis KBF, F. A. (1987). Specific synthesis of DNA in vitro via a polymerase-catalyzed
chain reaction. Methods in Enzymolagy 155:335-350.
Nene V, Wortrnan JR, Lawson D, Haas B, Kodira C, Tu Z, et al. (2007). Genome
Sequence of Aedes aegypti, a Major Arbovirus Vector. Science 316: 1 71 8-1723.
49
Nery LC, Lorosa NE, and Franco AM (2004). Feeding preference of the sand flies
Lutzomyia urnbratilis and L. spathotrichia (diptera: Psychodidae, Phlebotominae)
in an urban forest patch in the city of Manaus, Amazonas, Brazil. Mem Inst
Oswalda Cruz 99:571-574.
Newhouse VR, Chamberlain RW, Johnston JF, Sudia WD (1966). Use of dry ice to
increase mosquito catches of the CDC miniature light trap. Mosquito News 26:30-
35.
Ostfeld R, and Keesing F (2000). The function of biodiversity in the ecology of vector-
bome zoonotic diseases. Canadian Journal of Zoology-Revue Canadienne De
Zoologie 78:2061-2078.
Ostfeld RS, Glass GE, and Keesing F (2005). Spatial epidemiology: an emerging (or re-
emerging) discipline. Trends in Ecology & Evolution 20:328-336.
Pavlovsky EN (1966). Natural Nidality of T ransmissible Diseases With Special
Reference to the Landscape Epidemiology of Zooanthraponases. University of
Illinois Press, Urbana and London.
Perkins SE, Cattadori IM, Tagliapietra V, Rizzoli AP, and Hudson PJ (2006). Localized
deer absence leads to tick amplification. Ecology 87: 1981 -1986.
Perry RD, and F etherston JD (1997). Yersinia pestis--etiologic agent of plague. Clin
Microbial Rev 10:35-66.
Pfeiffer DU, and Hugh-Jones M (2002). Geographical information systems as a tool in
epidemiological assessment and wildlife disease management. Revue Scientifique
Et Technique De L Oflice International Des Epizaoties 21 :91-102.
Pichon B, Egan D, Rogers M, and Gray J (2003). Detection and identification of
pathogens and host DNA in unfed host—seeking Ixodes ricinus L. (Acari :
Ixodidae). Journal of Medical Entomology 40:723-731.
Polley L (2005). Navigating parasite webs and parasite flow: Emerging and re-emerging
parasitic zoonoses of wildlife origin. International Journal for Parasitology
35: 1279-1294.
Quispe-Tintaya KW, Laurent T, Decuypere S, Hide M, Banuls AL, De Doncker S, et al.
(2005). F luorogenic assay for molecular typing of the Leishmania donovani
complex: Taxonomic and clinical applications. Journal of Infectious Diseases
192:685-692.
Randolph SE (2004). Evidence that climate change has caused [‘]emergence' of tick-
borne diseases in Europe? International Journal of Medical Microbiology
Supplements 293:5 -1 5.
50
Randolph SE (2009). Tick-home disease systems emerge from the shadows: the beauty
lies in molecular detail, the message in epidemiology. Parasitology 136:1403-
1413.
Randolph SE, Green RM, Peacy MF, and Rogers DJ (2000). Seasonal synchrony: the key
to tick-bome encephalitis foci identified by satellite data. Parasitology 121:15-23.
Randolph SE, Miklisova D, Lysy J, Rogers DJ, and Labuda M (1999). Incidence from
coincidence: patterns of tick infestations on rodents facilitate transmission of tick-
bome encephalitis virus. Parasitology 118:177-186.
Rawlins SC, Tiwari T, Chadee DD, Validum L, Alexander H, Nazeer R, et al. (2001).
American cutaneous leishmaniasis in Guyana, South America. Annals of Tropical
Medicine and Parasitology 95:245 - 251.
Relman DA, J. S. Loutit, T. M. Schmidt, S. F alkow, L. S1 Tompkins (1990). The agent of
bacillary angiomatosis. An approach to the identification of uncultured pathogens.
N Engl J Med 323: 1 573-1 580.
Rezza G, Nicoletti L, Angelini R, Romi R, Finarelli AC, Panning M, et al. (2007).
Infection with chikungunya virus in Italy: an outbreak in a temperate region.
Lancet 370: 1 840-1 846.
Rodrigues EHG, de Brito MEF, Mendonca MG, Werkhauser RP, Coutinho EM, Souza
WV, et al. (2002). Evaluation of PCR for diagnosis of American cutaneous
leishmaniasis in an area of endemicity in northeastern Brazil. Journal of Clinical
Microbiology 40:3 572-35 76.
Rogers DJ, and Randolph SE (2003). Studying the global distribution of infectious
diseases using GIS and RS. Nature Reviews Microbiology 1:231-237.
Rogers DJ, and Randolph SE (2006). Climate change and vector-borne diseases. Pages
345-381 in Advances in Parasitology, Vol 62.
Ruiz MO, Chaves LF, Hamer GL, Sun T, Brown WM, Walker ED, et al. (2010). Local
impact of temperature and precipitation on West Nile virus infection in Culex
species mosquitoes in northeast Illinois, USA. Parasit Vectors 3:19.
Ryan J, Dave K, Emmerich E, Fernandez B, Turell M, Johnson J, et al. (2003). Wicking
assays for the rapid detection of West Nile and St. Louis encephalitis viral
antigens in mosquitoes (Diptera : Culicidae). Journal of Medical Entomology
40:95-99. -
Sattabongkot J, Kiattibut C, Kumpitak C, Ponlawat A, Ryan JR, Chan AST, et al. (2004).
Evaluation of the VecTest Malaria Antigen Panel assay for the detection of
Plasmodium falciparum and P-vivax circurnsporozoite protein in anopheline
mosquitoes in Thailand. Journal of Medical Entomology 41:209-214.
51
Sonenshine DE (1993). Biology af T icks. Oxford University Press, New York.
Souza ISe, Naiff RD, Guimaraes TC, Naiff MdF, Cupolillo E, Rodrigues WdA, et al.
(1998). American cutaneous leishmaniasis due to Leishmania (V iannia)
guyanensis as an initial clinical presentation of human immunodeficiency
virus infection. Journal of the European Academy of Dermatology & Venereology
10:214-217.
Spielman A, Wilson ML, Levine JF, and Piesman J (1985). Ecology of Ixodes-Darnmini-
Borne Human Babesiosis and Lyme Disease. Annual Review of Entomology
30:439-460.
Spratt BG (1999). Multilocus sequence typing: molecular typing of bacterial pathogens in
an era of rapid DNA sequencing and the intemet. Curr Opin Microbial 2:312-316.
Steere AC, Coburn J, and Glickstein L (2004). The emergence of Lyme disease. Journal
of Clinical Investigation 1 13:1093-1 101 .
Stone EG, Lacombe EH, and Rand PW (2005). Antibody testing and Lyme disease risk.
Emerging Infectious Diseases 11:722-724.
Sumilo D, Asokliene L, Bormane A, Vasilenko V, Golovljova I, and Randolph SE
(2007). Climate Change Cannot Explain the Upsurge of Tick-Borne Encephalitis
in the Baltics. Plos One 2:-.
Sumilo D, Bormane A, Asokliene L, Lucenko I, Vasilenko V, and Randolph S (2006).
Tick-bome encephalitis in the Baltic States: Identifying risk factors in space and
time. International Journal of Medical Microbiology 296:76-79.
Sumilo D, Bormane A, Vasilenko V, Golovljova I, Asokliene L, Zygutiene M, et al.
(2009). Upsurge of tick-borne encephalitis in the Baltic States at the time of
political transition, independent of changes in public health practices. Clinical
Microbiology and Infection 15:75-80.
Sundar S, Reed. SG, Singh. VP, Kumar. PC, Murray. HW. (1998). Rapid accurate field
diagnosis of Indian visceral leishmaniasis. Lancet 351:563-565.
Swaddle JP, and Calos SE (2008). Increased avian diversity is associated with lower
incidence of human West Nile infection: observation of the dilution effect. Plos
One 3:e2488.
Telford SR, III, Mather TN, Moore SI, Wilson ML, and Spiehnan A (1988).
Incompetence of Deer as Reservoirs of the Lyme Disease Spirochete. Am J Trap
Med Hyg 39: 1 05-109.
Tempelis CH (1975). Host-Feeding Patterns of Mosquitos, with a Review of Advances in
Analysis of Blood Meals by Serology. Journal of Medical Entomology 11:635-
653.
52
Venkatesan M, and Rasgon JL (2010). Population genetic data suggest a role for
mosquito-mediated dispersal of West Nile virus across the western United States.
Mol Ecol.
Weigle K, and Saravia NG (1996). Natural history, clinical evolution, and the host-
parasite interaction in New World cutaneous leishmaniasis. Clinics in
Dermatology 14:433-450.
With KA (2002). The landscape ecology of invasive spread. Conservation Biology
16:1192-1203.
Woods ME, Montenieri JA, Eisen RJ, Zeidner NS, Borchert JN, Laudisoit A, et al.
(2009). Identification of flea blood meals using multiplexed real-tirne polymerase
chain reaction targeting mitochondrial gene fiagments. Am J Trop Med Hyg
80:998-1003.
Woolhouse MEJ, and Gowtage-Sequeria S (2005). Host range and emerging and
reemerging pathogens. Emerging Infectious Diseases 11:1842-1847.
Young DG, and Duncan MA, editors. (1994). Guide to the identification and geographic
distribution of Lutzomyia sand flies in Mexico, the West Indies, central and South
America (Diptera: Psychodidae), Gainsville, FL.
53
CHAPTER 2
Zoonotic pathogens in ondes scapularis, Michigan
Hamer SA, Hickling GJ, Roy P, Walker ED, Foster ES, Barber CC, and Tsao JI (2007).
Zoonotic pathogens in Ixodes scapularis, Michigan. Emerging Infectious Diseases
7:1131-1133.
Abstract
Ixodes scapularis is endemic to a focal area of Michigan’s Upper Peninsula and is
currently invading the southwestern Lower Peninsula. Herein we report Borrelia
burgdorferi, A naplasma phagocytophilum and Babesia odocoilei infection within ticks
fiom both foci. Infection prevalence with these pathogens was greater in ticks collected
from endemic areas as compared to those collected from recently-invaded sites. These
findings emphasize the need for heightened awareness by human and veterinary health
professionals.
Introduction
Ixodes scapularis - the blacklegged tick - is currently the most infectious vector
species to humans in the United States, transmitting the agents of Lyme borreliosis,
human anaplasmosis (previously human granulocytic ehrlichiosis), and human
babesiosis. Since Lyme borreliosis and human anaplasmosis became notifiable diseases
in 1991 and 1999, respectively, annual incidence has been increasing with 19,804 and
537 reported cases in 2004 (Jajosky et al. 2006). The increase in I. scapularis-bome
disease is due, in part, to geographic range expansion of I. scapularis. Despite the broad
endemic areas to the west in Wisconsin and Minnesota and to the south in Indiana, active
and passive surveillance suggested that the only populations of 1. scapularis established
54
in Michigan prior to 2002 were in Menominee County in the Upper Peninsula (Walker et
al. 1998). In 2002-2003, however, active wildlife sampling and tick dragging in
southwestern Michigan indicated that invasion of I. scapularis was occurring (Foster
2004), with nearby tick populations in northwestern Indiana as the putative source. Here
we describe a recent assay of Borrelia spp. , Anaplasma phagocytophilum, and Babesia
spp. from a targeted sample of host-seeking ticks from an endemic area in Michigan’s
Upper Peninsula and from the Lower Michigan invasion zone, and conclude that
infection prevalence with pathogens was greater within endemic ticks.
The Study
Over a 1.5 week period during the adult I. scapularis spring questing peak in
April-May 2005, ticks were collected from two sites in Menominee County in
Michigan’s Upper Peninsula (endemic for I. scapularis) and three sites in southwestern
Michigan (recently-invaded by I. scapularis; Figure 2.1) by dragging a l-m2 corduroy
cloth over the forest floor for 500 m (Falco and Fish 1992). Drag cloths were checked at
20 m increments for ticks, which were removed and kept alive until identification. Tick
densities were greater at endemic sites than recently-invaded sites, and within invaded
sites, densities were greater at the sites closest to the putative source of invading ticks,
indicative of duration of establishment. Host-seeking adults have had two previous
bloodmeals, so we focused our efforts on adults to increase the probability of detecting
pathogens. We collected in spring, because the adult I. scapularis questing peak in
Michigan is greater then than in fall (Strand et al. 1992; Foster 2004). The endemic sites
are comprised of upland mixed hardwoods (Site A in Figure 2.1) and conifer/shrub cover
55
(Site B) along the Menominee River, while the southern sites contain mixed deciduous
forest overlying sand dunes along the eastern shores of Lake Michigan.
Ticks were aseptically dissected in half, with one half retained for a separate
study. DNA was extracted from the other half, following overnight lysis, using DNeasy
tissue kits (Qiagen, Valencia, CA) and used as template in three separate polymerase
chain reactions to detect the 163-233 rRNA intergenic spacer of B. burgdorferi or
B. miyamotoi (Bunikis et al. 2004), the p44 gene of A. phagocytophilum (Zeidner et al.
2000), and the 188 rRN A gene of Babesia-genus organisms (Armstrong et a1. 1998).
Babesia-positive amplicons were purified and sequenced for species identification.
From endemic sites A and B, a total of 31 ticks were collected: 28 adult and 1
nymphal I. scapularis, and 2 adult Dermacentor variabilis. Of the adult I. scapularis, 17
(60.7%) were positive for B. burgdorferi, 4 (14.3%) were positive for
A. phagocytophilum, and 2 (7.1%) were positive for Babesia spp. (later identified as Ba.
odocoilei through sequence analysis; Table 2.1). Two adults were co-infected with B.
burgdorferi and A. phagocytophilum and one adult was co-infected with B. burgdorferi
and Ba. odocoilei. Rates of co-infection did not deviate from random expectation
(Fisher’s exact tests, P = 0518—0640). The single nymph was infected with
B. burgdorferi only.
At newly-invaded sites C-E, a total of 105 ticks were collected: 91 adult and 10
nymphal I. scapularis, and 5 adult D. variabilis. Of the adult I. scapularis, 43 (47.3%)
were positive for B. burgdorferi, 1 (1.1%) was positive for A. phagocytophilum, and 4
(4.4%) were positive for Babesia spp. (later identified as Ba. odocoilei). All 4 Ba.
odocoilei-positive ticks were also infected with B. burgdorferi. This rate of co-infection
56
was significantly greater than random expectation (Fisher’s exact test, P = 0.046). Of 10
nymphal I. scapularis, 2 (20.0%) were positive for B. burgdorferi and 2 (20.0%) were
positive for A. phagocytophilum, including one co-infected tick (a nonsignificant co-
infection rate, Fisher’s exact test, P = 0.378).
Conclusions
These are the first records of A. phagocytophilum and Ba. odocoilei in ticks in
Michigan, and it is clear that they are present among both endemic and recently-invaded
I. scapularis. B. burgdorferi infection in 1. scapularis of Michigan has been reported
previously (Walker et al. 1994; Walker et a1. 1998; Foster 2004; Diuk-Wasser et al.
2006), and two culture-confirmed cases of human Lyme borreliosis in Michigan have
been reported, both contracted in Menominee County (Stobierski et al. 1994; Golde et al.
1998). Whereas B. burgdorferi and A. phagocytophilum are both zoonotic pathogens of
humans, Ba. odocoilei is an intraerythrocytic protozoan parasite maintained in
transmission cycles between I. scapularis and white-tailed deer and is not known to be
pathogenic to humans (Armstrong et al. 1998). We did not detect 1. scapularis infected
with either Ba. microti or the WAl piroplasm, the etiologic agents of human babesiosis in
North America (Homer et al. 2000), nor with B. miyamotoi.
A comparison of the B. burgdorferi-infection rate in adult ticks collected herein
from the endemic sites (60.7%) with equivalent records from the same county in 1992
(31.3%)(Walker et al. 1994) indicate that infection prevalence in the endemic focus has
increased over time (Fisher’s exact test, P < 0.001 ). B. burgdorferi infection prevalence
of adult I. scapularis has also increased within the invasion zone of southwest Michigan,
57
from 37.0% in 2002-2003 (F oster 2004)(collection site 5km south of the southernmost
site in 2006) to 47.3% in 2006, although this difference is only marginally statistically
significant due to small 2006 sample size (one-tailed Fisher’s exact test, P = 0.046).
In Indiana, the putative source of invading I. scapularis into lower Michigan,
I. scapularis and B. burgdorferi infected-I. scapularis were first documented in 1987 and
1991, respectively (Pinger et al. 1996). A. phagocytophilum and Ba. odocoilei have been
recently reported in Indiana (Steiner et al. 2006), with infection prevalences higher than
those reported herein for Michigan (11.8% and 10.3% in Indiana; 1.1 and 4.4% in
Michigan, respectively), which is suggestive of more recent establishment in Michigan.
Our findings of B. burgdorferi, A. phagocytophilum, and Ba. odocoilei infections
in adult and nymphal I. scapularis in endemic and recently-invaded areas of Michigan
provide evidence of the establishment of pathogen maintenance cycles within these
zones. Furthermore, infection prevalence was greater in ticks collected from endemic foci
as compared to those collected from the invasion front, though difference were not
always statistically significant. These data imply a risk of human Lyme borreliosis in
areas both endemic for and recently-invaded by I. scapularis. Indeed, since rising levels
of tick infection and co-infection can be expected with increasing time since tick
establishment, medical practitioners in Michigan should be including disease resulting
from these pathogens within their index of suspicion. Co-infections within the adult ticks
of our study were more common than expected based on individual pathogen infection
rates, suggesting that these adults were infected with different pathogens during their
larval and nymphal bloodmeals, that individual wildlife hosts may be infecting ticks with
58
multiple pathogens during a single bloodmeal, and/or that infection with one pathogen
facilitates subsequent infection with others.
59
"" I 7 I.“ M
I. _ (I
l ’ ;____:, J.
\ 1,
.-”“"“‘*“ v4-- .jw ~
2 M . ‘ 7 '
W X ...——I . q i x
.\ K \
n\ ‘
1.!
Figure 2.1. Tick dragging field sites within Michigan. A= Menominee North; B=
Menominee South; C= Duck Lake State Park; D= Saugatuck Dunes State Park; E= Van
Buren State Park. Gray shaded counties are those in which endemic (Upper Peninsula)
and recently-invaded (Lower Peninsula) 1. scapularis are known to occur.
60
.036 u n. .69 fimxm 9.55mi .cozomhcnoo *0 _m>m_ 58¢?me
o 35: F o 8de N 8de N 2 z
a. s «v o e s v c s F 6 3 9. 5 < Bone.
o o o o 62: P z
810.: :EN ctN 8.3: :5th N < 2585 __ao>o
8.4:. o 3va as: 233% am < Bums... m
o GNU P o AN: F 3.3 N a 2
as: c an: o 8.8; 8 < 882.. o
o o o 6.8: o N z
o o o o 3.8: v < 882.. o
o o o o 6.2:: P z
8.8. 8.»: 8.»: 6.8:. :8; NF < ascocm m
o 3m: 3m: 8d: 8.8; 2 < 258$ <
dd + an ad + an 3388 an 552.828qu .< 5:893 .m €338» .. as» 25.5 a.»
53.88.: ..8 .oz
32 .5er .oz
.02
.299? N 2 3:65 u< ”owfim £5 . _ .N aha—mi 05 9 gamete”.
magma 8mm .mcowofia 025 53, detach: utfizmeoa .N Ram—PE use :33 mo moose—“BE ammoommozm 53:22 ._.N 033.
61
References
Armstrong PM, Katavolos P, Caporale DA, Smith RP, Spiehnan A, and Telford SR
(1998). Diversity of Babesia infecting deer ticks (Ixodes dammini). American
Journal of Tropical Medicine and Hygiene 58:73 9-742.
150:1741-1755.
Diuk-Wasser MA, Gatewood AG, Cortinas MR, Yaremych-Hamer S, Tsao J, Kitron U,
et al. (2006). Spatiotemporal patterns of host-seeking Ixodes scapularis nymphs
(Acari : Iodidae) in the United States. Journal of Medical Entomology 43:166-
176.
Falco RC, and Fish D (1992). A Comparison of Methods for Sampling the Deer Tick,
Ixodes-Dammini, in a Lyme-Disease Endemic Area. Experimental & Applied
Acarolagy 14:165-173.
F aster ES (2004). Ixodes scapularis (Acari: Ixodidae) and Borrelia burgdorferi in
southwest Michigan: population ecology and verification of a geographic risk
model. Masters Thesis, Michigan State University, East Lansing.
Golde WT, Robinson-Dunn B, Stobierski MG, Dykhuizen D, Wang IN, Carlson V, et al.
(1998). Culture-confirmed reinfection of a person with different strains of
Borrelia burgdorferi sensu stricto. Journal of Clinical Microbiology 36: 1015-
1019.
Homer MJ, Aguilar-Delfin I, Telford SR, Krause PJ, and Persing DH (2000). Babesiosis.
Clinical Microbiology Reviews l3:451-+.
Jajosky RA, Hall PA, Adams DA, Dawkins FJ, Sharp P, Anderson WJ, et al. (2006).
Summary of nitifiable diseases- United States, 1991-2004. Morbidity and
Mortality Weekly Reports 53:1-79.
Pinger RR, Timmons L, and Karris K (1996). Spread of Ixodes scapularis (Acari:
Ixodidae) in Indiana: Collections of adults in 1991-1994 and description of a
Borrelia burgdorferi-infected population. Journal of Medical Entomology 33:852-
855.
Steiner FE, Pinger RR, Vann CN, Abley MJ, Sullivan B, Grindle N, et al. (2006).
Detection of Anaplasma phagocytophilum and Babesia odocorler DNA in Ixodes
scapularis (Acari : Ixodidae) collected in Indiana. Journal of Medical Entomology
43:437-442.
62
Stobierski MG, Hall WN, Robinsondunn B, Stiefel H, Shiflett S, and Carlson V (1994).
Isolation of Borrelia-Burgdorferi from 2 Patients in Michigan. Clinical Infectious
Diseases 19:944-946.
Strand MR, Walker ED, and Merritt RW (1992). Field studies on Ixodes dammini in the
Upper Peninsula of Michigan. Vector Control Bulletin of North Central States 1
1 1-1 8.
Walker ED, Smith TW, Dewitt J, Beaudo DC, and McLean RG (1994). Prevalence of
Borrelia-Burgdorferi in Host-Seeking Ticks (Acari, Ixodidae) from a Lyme-
Disease Endemic Area in Northern Michigan. Journal of Medical Entomology
31:524-528.
Walker ED, Stobierski MG, Poplar ML, Smith TW, Murphy Al, Smith PC, et al. (1998).
Geographic distribution of ticks (Acari : Ixodidae) in Michigan, with emphasis on
Ixodes scapularis and Borrelia burgdorferi. Journal of Medical Entomology
35:872-882.
Zeidner NS, Burkot TR, Massung R, Nicholson WL, Dolan MC, Rutherford J S, et al.
(2000). Transmission of the agent of human granulocytic ehrlichiosis by Ixodes
spinipalpis ticks: Evidence of an enzootic cycle of dual infection with Borrelia
burgdorferi in northern Colorado. Journal of Infectious Diseases 182:616-619.
63
CHAPTER 3
Invasion of the Lyme disease vector Ixodes scapularis: implications for Borrelia
burgdorferi endemicity
Hamer SA, Tsao 11, Walker ED, Hickling GJ. 2010. Invasion of the Lyme
disease vector Ixodes scapularis: implications for Borrelia burgdorferi
endemicity. EcoHealth. doi: 10.1007/310393-010-0287-0
Abstract
Lyme disease risk is increasing in the United States due in part to spread of
Ixodes scapularis, the principal vector of the spirochetal pathogen Borrelia burgdorferi.
A five-year study was undertaken to investigate a hypothesized co-invasion of I.
scapularis and B. burgdorferi in lower Michigan. We tracked the spatial and temporal
dynamics of the blacklegged tick and the spirochete using mammal, bird, and vegetation
drag sampling at eight field sites along coastal and inland transects originating in a zone
of recent 1. scapularis establishment. We document northward invasion of ticks along
Michigan’s west coast over the study period; this pattern was most evident in ticks
removed from rodents. B. burgdorferi infection prevalences in I. scapularis sampled from
vegetation in the invasion zone were 9.3 and 36.6% in nymphs and adults, respectively.
There was no evidence of I. scapularis invasion along the inland transect, however low-
prevalence B. burgdorferi infection was detected in other tick species and in wildlife at
inland sites, as well as at northern coastal sites in years prior to the arrival of
I. scapularis. These infections suggest that cryptic B. burgdorferi transmission by other
vector-competent tick species may be occurring in the absence of I. scapularis. Other
Borrelia spirochetes, including those that group with B. miyamotoi and B. andersonii,
were present at a low prevalence within invading ticks and local wildlife. This rapid
64
blacklegged tick invasion - measurable within five years — in combination with cryptic
pathogen maintenance suggests a complex ecology of Lyme disease emergence in which
wildlife sentinels can provide an early warning of disease emergence.
Introduction
Lyme disease accounts for over 90% of all reported vector-borne disease in the
United States with more than 20,000 cases reported annually; its current invasive spread
from endemic foci constitutes a major public health concern (Bacon et al. 2007). In the
United States, Lyme disease is caused by the bacterium Borrelia burgdorferi and the
blacklegged tick Ixodes scapularis serves as the vector for a majority of cases. Patterns of
human disease mirror the geographic distribution of B. burgdorferi-infected I. scapularis,
which is characterized by high-density endemic foci in the Northeast and upper
Midwestern United States. Increasing incidence is associated with both biological and
non-biological factors, including enhanced surveillance and awareness among medical
professionals, human encroachment into tick habitat, creation of peridomestic habitats
that attract wildlife hosts of ticks and pathogens, and increases in abundance and range
expansions of wildlife. Blacklegged ticks were likely widespread prior to the most recent
glaciation event, during which relict populations remained in refuges, from which they
and their vertebrate hosts later expanded (Steere et al. 2004). Over the past 50 years,
white-tailed deer (Odocoileus virginiaus) populations have undergone explosive growth
due to reversion of agricultural lands to forest and restrictions on hunting. This deer
expansion has facilitated the recent blacklegged tick expansion and Lyme disease
emergence throughout the Northeast and Midwest. I. scapularis continues to spread from
65
both the northeastern and Midwestern foci (Steere et al. 2004), but the mechanisms for
this spread are not established. 1. scapularis serves as the vector for multiple zoonotic
pathogens — including the agents of Lyme disease, human anaplasmosis, babesiosis, and
Powassan encephalitis — so its expansion constitutes increased risk of multiple diseases.
In 2002, a new population of I. scapularis was detected in the southwestern
corner of Michigan’s Lower Peninsula (F oster 2004). This represented a significant
change from the State’s previous reported distribution (Walker et al. 1998), which was
characterized by a probable lack of established 1. scapularis throughout the Lower
Peninsula. Established populations in the state were recognized at that time only in
Menominee County of the Upper Peninsula, adjacent to endemic foci in Wisconsin
(Walker et al. 1998). Nevertheless, a Midwestern model of habitat suitability (Guerra et
al. 2002) predicted that sandy oak upland forests provide highly suitable habitat for
I. scapularis, and I. scapularis indeed were discovered in such areas(Foster 2004).
Suitable habitat types are widespread throughout lower Michigan, so further invasion of
I. scapularis is thus predicted.
Tick versus pathogen invasion?
B. burgdorferi and A. phagocytophilum have been found within Michigan’s
recently-invaded tick populations (Hamer et al. 2007). The early detection of low-density
infected tick populations afforded us the rare opportunity to test, in real time, three non-
exclusive scenarios by which invasion of the Lyme disease vector and/or pathogen into
lower Michigan may be occurring: (i) In the ‘tick first’ scenario, new uninfected
I. scapularis populations become established as a result of long distance dispersal of adult
66
ticks by white-tailed deer. This deer-mediated invasion would introduce replete
uninfected adult ticks, but not B. burgdorferi (as deer are incompetent hosts for this
pathogen; (Telford et al. 1988)). In this scenario, we propose that B. burgdorferi enters
the system later, as result of a slower secondary invasion mediated by infected
mammalian or avian hosts. (ii) In the ‘dual-invasion’ scenario, mammalian or avian hosts
introduce infected I. scapularis to new areas in high enough numbers to allow both
I. scapularis and B. burgdorferi to establish. In this scenario, B. burgdorferi should be
detectable early in the invasion process, when I. scapularis densities are still low. (iii) In
the ‘spirochete-first’ scenario, enzootic transmission of the pathogen is maintained by
cryptic vectors and reservoir hosts. These vectors are generally wildlife host-specialists
that, unlike I. scapularis, do not bite humans (Telford and Spielman 1989b, a) and thus
these transmission cycles have no implications for human or canine disease risk. Invasion
of I. scapularis (either infected or uninfected) into zones of cryptic pathogen maintenance
will, however, create opportunities for bridging the pathogen to humans and canines.
We initiated comprehensive sampling to investigate the spatial and temporal
dynamics of the Lyme disease system in and beyond the zone of hypothesized
I. scapularis invasion in lower Michigan. To address invasion, we studied areas not only
where all three parasitic life stages of I. scapularis were endemic (presence of all three
stages and/or at least six individuals of a single stage constitute the Centers for Disease
Control and Prevention (CDC) definition of an ‘established’ papulation; (Dennis et al.
1998), but also in areas beyond its detected distribution. The objectives of this study were
to: (t) test whether the detected distribution of I. scapularis changed over the period
2004-2008; and (ii) assess the patterns of occurrence of B. burgdorferi in relation to that
67
of I. scapularis. To address the ‘tick-first’ hypothesis, each field site was sampled
regularly for the presence of I. scapularis to detect new colonization and/or trends in tick
abundance at each site. To address the ‘dual-invasion’ and ‘spirochete-first’ hypotheses, a
diverse community of hosts and their attached ticks were assayed to assess the
relationship between I. scapularis presence and B. burgdorferi infection. Thus, we
include results from a diverse assemblage of ticks, including those with low or no
vectorial capacity and no associated zoonotic risk (i.e., tick species that are incompetent
for spirochete transmission and/or species that do not regularly feed on human hosts).
Although unimportant for transmission, incompetent ticks may serve as bio-indicators of
B. burgdory‘izri presence in a given area if they contain an infectious blood meal or
transtadially passed spirochetes from a previous meal.
Materials and Methods
Site selection and sampling regime. From 2004-2008, we assessed on- and off-host ticks
and wildlife along two sampling transects established at a spatial scale that extended
beyond the known limit of the lower Michigan I. scapularis papulation (Figure 3.1). Both
transects originated near the southwestern Michigan zone where I. scapularis was first
detected in 2002(Foster 2004). From this origin, the coastal transect extended north along
Lake Michigan (sites labeled C1-C4; 81-145 km between sites: Van Buren State Park,
Van Buren Co.; Duck Lake State Park, Muskegon Co.; Orchard Beach State Park,
Manistee Co.; Sleeping Bear Dunes National Lakeshore, Benzie Co.; respectiVely). The
inland transect extended northeastward (sites labeled 11-14; 33-76 km between sites: F ort
Custer State Recreation Area, Kalamazoo Co.; Lux Arbor Reserve of Kellogg Biological
68
Station; Barry Co.; Ionia State Recreation Area, Ionia Co.; and Rose Lake Wildlife
Research Area, Clinton Co.; respectively). Both transects traversed areas where
established 1. scapularis had not been previously detected (Walker et al. 1998; F aster
2004), or were previously uninvesti gated. Sampling was conducted in oak-dominated,
closed-canopy deciduous forest when available, as this habitat type is positively
associated with I. scapularis presence in the endemic Midwest (Guerra et al. 2002).
Otherwise, sampling was conducted in the dominant forest type at the site. All data
presented here are for May and June, when I. scapularis larvae, nymphs, and adults are
simultaneously active. Each site was sampled in these months in all five years of the
study, except for site C4 (2007 and 2008 only) and site 11 (all years except 2006).
Mammal trapping. At each field site, small mammals were trapped along four to six
transects of 25 Sherman live traps (H. B. Sherman Traps, Tallahassee, FL) spaced 10 m
apart and baited with sunflower seed. Mediurn-sized mammals were trapped using 12-16
32x10x12-inch live traps (Tomahawk Live Traps, Tomahawk, WI) and wooden box traps
baited with peanut butter, jelly, and cat food on tortillas. Small mammals were
anesthetized using Isoflurane (IsoFlo, Abbot Laboratories, Abbott Park, IL); mediurn-
sized mammals were anesthetized using ketamine hydrochloride (Ketaset; F ort Dodge,
Overland Park, KS) and xylazine hydrochloride (Rompun; Bayer Health Care, Kansas
City, KS) followed by reversal with administration of yohimbine hydrochloride
(Antagonil; Wildlife Laboratories, Fort Collins, CO). Animals were identified 'to species
and sex by inspection. Each animal was examined for ticks, biopsied in both ears using a
2-mm (small mammals) or 4-mm (medium mammals) biopsy punch (Miltex Insturments,
69
York, PA), and ear-tagged (National Band and Tag, Newport, KY). Ticks and ear
biopsies were stored separately in 70% ethanol. Additional ear biopsies were obtained
from animals recaptured after an interval of 2 or more weeks; recaptures at a shorter
interval were simply checked again for ticks. All animals were released at the site of
capture. Small mammal trapping success rate was discounted by the number of empty,
tripped traps as follows: number mammal captures / (number traps set - 0.5*no. tripped
traps); this expression assumes that on average tripped traps were open for half the night.
Wildlife procedures were approved through Michigan State University’s Institutional
Animal Use and Care Committee permit #02-07-13-000.
Bird mist netting. At each site, six 12-m mist nets (Avinet, Dryden, NY) were used to
capture birds in the same areas where mammals were trapped. Nets were run from 0600 -
1200 h on fair weather days and checked every 45 min. Birds were weighed, identified to
species and sex, measured, searched for ticks, and leg-banded with federally-issued bands
before release. Mist netting was performed under federal permit #02640.
Questing tick sampling. Each site was sampled for questing ticks by dragging a l-m2
white corduroy cloth (F alco and Fish 1992) over the forest floor along the same six
transects that were used for mammal trapping. Drag-sampling was performed on rain-free
days in the late morning or late afternoon so as to avoid the hottest and least-humid times
of day (Schulze et al. 2001; Diuk-Wasser et al. 2006). At least 1000 m2 of vegetation
were sampled per visit. The cloth was inspected every 20 m, and attached ticks were
stored in 70% ethanol. Drag-sampling was not performed an excessively hot or wet days.
70
Borrelia burgdorferi detection. All ticks were identified to species and stage (Keirans
and Clifford 1978; Sonenshine 1979; Durden and Keirans 1996). Total DNA from ticks
and ear biopsies was extracted using the DNeasy Blood and Tissue Kit (Qiagen,
Valencia, CA) following the manufacturer’s animal tissue protocol with the following
modifications. Ticks were first bisected using a sterile scalpel or were pulverized in liquid
nitrogen, followed by an overnight incubation in lysis buffer. DNA was eluted using 50
ul elution solution warmed to 70° C. Ear bi0psies (one per animal), adult and nymphal
ticks were extracted individually, and conspecific larvae from the same individual animal
or drag transect were pooled for extraction. Given our specific questions, we
implemented a protocol for sub-sampling of ticks to assay for infection in the occasional
cases of heavily-parasitized hosts. In cases where more than three adult or nymphal ticks
of the same species, life stage, and sex were removed from an individual host, three were
randomly selected for testing (i.e., the maximum number of ticks tested from a host
would be, for each tick species present, six adults (three female and three male), three
nymphs, and a larval pool). By including a subset of ticks from many hosts, rather than
all ticks from a smaller ntunber of hosts (given limited resources), this sampling strategy
allowed for improved coverage of the host population, and also reduced statistical
concerns of non-independence that would arise if very large numbers of ticks were
collected from a small number of highly-infested hosts. B. burgdorferi strain B31-
infected nymphal I. scapularis acquired from the CDC and water served as the positive
and negative extraction controls, respectively.
B. burgdorferi was detected using either i) a nested polymerase chain reaction
(PCR) for the 16S — 23$ rRNA intergenic spacer region (IGS) of Borrelia spp. (Bunikis
71
et al. 2004) followed by visualization with gel electrophoresis or ii) a quantitative PCR
(qPCR) of a region of the 168 rRNA of B. burgdorferi (Tsao et al. 2004). PCR enzyme
kits were used throughout (nested PCR: PCR Supermix, lnvitrogen, Carlsbad, CA and
FailSafe PCR System, Epicentre, Madison, WI; qPCR: Universal PCR Master Mix,
Applied Biosystems, Foster City, CA). In addition to the positive and negative DNA
extraction controls for each batch of samples extracted, DNA from B. burgdorferi strain
B31-infected ticks from laboratory colonies at the CDC and water served as the positive
and negative PCR controls. In qPCR, a six-point standard dilution series of DNA
extracted fiom cultured spirochetes (104 '— 10'l organisms per 3 ul reaction volume)
served as positive controls (B. burgdorferi strain C336, a rRNA spacer type (RST) 11
strain). Reactions for qPCR were done with an ABI Prism 7900HT Sequence Detection
System (Applied Biosystems, Foster City, CA). Preliminary experiments showed that
both tests were able to detect positive samples containing a minimum 100 organisms.
Nucleotide sequencing. A subset of B. burgdorferi-positive ticks and ear biopsies was
subjected to DNA sequencing to confirm pathogen identity. Samples included: i) a
random subset of positive samples from sites where I. scapularis is common; ii) a
majority of test-positive samples from sites with an apparent absence of I. scapularis; and
iii) any IGS amplicons of approximately 500 bp, characteristic of the relapsing-fever
spirochete B. miyamotoi (cf. the 987 bp fragment characteristic of B. burgdorferi). If the
sample was determined to be positive using qPCR, then IGS PCR was performed to
generate the template for sequencing. The IGS product was purified (Qiagen PCR
Purification Kit; Qiagen, Valencia, CA) and sequences were determined in both
72
directions using an ABI Prism 3100 Genetic Analyzer (Applied Biosystems, F aster City,
CA). Sequences were identified as either B. burgdorferi or B. miyamotoi based on
comparisons to published sequences using the basic local alignment search tool in
GenBank (Altschul et al. 1990). B. burgdorferi strains were assigned to IGS ribosomal
spacer type (RST; (Liveris et al. 1995)) by alignment with the prototypical strains
published in Bunikis et al. (2004) using the program MEGA (Tamura et al. 2007).
Statistics. Logistic regression was used to assess trends in wildlife infestation and tick
infection over the 5-year sampling period. Fishers exact test was used to assess
differences in infestation among sites. Linear regression was used to assess trends in
nymphal densities within sites overtime. Within-year comparisons between the coastal
and inland transects were made by calculating the z-ratio and associated two-tail
probabilities for the difference between two independent proportions. A minimum
infection prevalence (MIP; i.e., assuming only one positive larva per pool) was used for
tests done on pooled larvae. Statistics were performed using Statistix 8 (Analytical
Software, Tallahassee, FL).
Results
Wildlife captures. We achieved a total of 1,667 mammal captures, comprised of 1,472
small and 195 medium mammal captures. Small mammals were captured in a total of
7,998 adjusted trap nights (18.4% overall capture rate). White-footed mouse was the most
commonly caught species (84.9% of all small mammal captures) — their capture rates did
not differ between coastal and inland transect sites (P = 0.95). Medium mammals were
73
captured in a total of 777 adjusted trap nights (25.1% overall capture rate), with similar
rates along both transects (P = 0.93). Raccoon was the most abundant species (65.6% of
all medium mammal captures). Despite escaped animals and the release of lethargic
shrews, 99.8% of captured mammals were processed, 82.5% of which were first captures
and 6.5% were recaptures after a minimum 2-week interval; these animals were
processed in full and used for computing the infestation and infection prevalences
reported below. The remaining 11% represent white-footed mice recaptured within 2
weeks; these animals were only checked for ticks and released. These latter mice are not
considered further in infestation analyses; however ticks removed from these animals are
considered in infection analyses. In total, 14 mammal species were parasitized by eight
species of ticks (Table 3.1).
A total of 747 bird captures, comprised of 55 species, occurred in 2180 net hours
for an average of 34.3 birds per 100 net hours. Due to escaped birds, 99.2% were
processed, of which 91.6% represent first captures and 6.7% were bird recaptured after a
minimum 2-week interval; these birds were processed in full and used in computing
infestation and infection prevalences below. The remaining 1.6% represent birds
recaptured within 2 weeks -— none harbored ticks. Birds were parasitized by three species
of ticks (Table 3.1; Supplement 1).
I. scapularis on wildlife. Along the inland transect, I. scapularis were rarely found on
wildlife, with only 0.5% of 574 white-footed mice parasitized. Two of these Were trapped
at site 13 in 2007 (3.0% infestation), with one harboring a single larva and nymph, and
one harboring four larvae. The other mouse was trapped at site 11 in 2008 (3.9%
74
infestation) and harbored a single nymph. None of the 73 eastern chipmunks, 133
individuals of other mammal species, or 504 birds on this transect harbored I. scapularis.
In contrast, a steep gradient of I. scapularis infestation of wildlife was evident
along the coastal transect. In all years the level of white-footed mouse infestation was
highest at the southernmost C1 site (75-100% with no significant difference among years;
P = .08), and larval and nymphal burdens averaged 4.5 and 0.6 per mouse, and 5.6 and
2.3 per infested mouse (maxima of 37 and 11, respectively). Infestation at the more
northern sites was much lower, but increased progressively during the study (Figure 3.2).
At site C2, infested mice were trapped in all five years, with infestation increasing
significantly overtime (P < 0.001). At site C3, no I. scapularis-infested mice were
observed until the third summer after which infestation increased (P = 0.76). Site C4
harbored a low proportion of infested mice in both 2007 and 2008 (9 and 10%,
respectively); these were the only two years when May-June sampling was conducted at
this site (in July 2005 none of 21 mice trapped at C4 harbored I. scapularis, whereas 60%
of 110, 13% of 54, and 0% of 16 mice were infested at sites C1, C2, C3 respectively).
Overall, 48.2% of 81 eastern chipmunks harbored I. scapularis along coastal sites.
Like mice, infestation was highest at site C1, with all chipmunks (n == 15) infested in all
years. Although fewer chipmunks were infested at the three more northern sites (Figure
3.2; P < 0.0001), infestation levels increased over time (P = 0.01 for sites C2 and C3
pooled). I. scapularis burdens on chipmunks were greatest at site C1; larval burdens there
were similar to those of mice (average 4.7 per chipmunk and 5.8 per larval-infested
chipmunk, maximum of 20), but nymphal burdens were much higher (average 7.2 per
chipmunk and 7.7 per nymphal-infested chipmunk, maximum of 36). Of 117 other
75
mammals caught on the coastal transect, 28.2% comprising ten species harbored
I. scapularis (Table 3.1; most were from CI (54.5%) or C2 (24.2%)). Of 237 bird
captures, I. scapularis was found on 14, of which 13 were from C1 and one was from C2.
These 14 birds comprised five species: American robin, chipping sparrow, Eastern
towhee, indigo bunting, and Northern cardinal (Supplement 1).
Questing ticks. A total of 58,459 m2 of drag sampling was conducted, with 26,800 and
31,650 m2 along the inland and coastal transects, respectively. A total of 1,984 ticks were
collected of which 87% were I. scapularis. Four other tick species were collected by drag
cloth: Dermacentor variabilis (n = 252), Amblyomma americanum (n = 4),
Haemaphysalis leporispalustris (n = 2), and I. dentatus (n = 5).
All but one I. scapularis were from coastal sites. Along the inland transect, a
single larva was dragged (at 13 in 2007). Conversely, along the coastal transect, 86 adults,
246 nymphs and 1,388 larvae were collected, of which 95.3, 91.5, and 93.3% respectively
were from site Cl. A significant gradient of I. scapularis density was detected, with the
greatest abundance at site Cl , where all life stages were present in all five summers of
sampling (peak May/June densities were 8.2 adults, 29.7 nymphs, and 90.8 larvae per
1000 m2) and nymphal abundance did not change across the five years (R2 = 0.14; P =
0.26). Many fewer, yet increasing, numbers of I. scapularis occurred at the two sites to
the north (Figure 3.3): at site C2, no I. scapularis were dragged at the start of the study
whereas all stages were present by the end with a significant increase in nymphal density
(R2 = 0.78; P = 0.02). At site C3, an increase in nymphal density occurred (R2 = 0.05, P =
76
0.09), and this was the only life stage present on drag cloths. No questing I. scapularis
were collected at site C4.
B. burgdorferi infection in drag-sampled ticks. Along the coastal transect, 36.6, 9.3,
and a minimum of 0.1% of adult, nymphal, and larval I. scapularis respectively were
infected with B. burgdorferi, with the majority of infected ticks (95.1%) collected at site
C1 (Figure 3.4). There was no significant change in infection prevalence among years at
site Cl (P = 0.38 — 0.71 for the three age classes); too few positive ticks were collected at
other coastal sites to test for equivalent trends. Along the inland transect, the single
I. scapularis larva that was dragged at site I3 in 2007 tested positive for B. burgdorferi
(Table 3.2). Of all D. variabilis collected on drag cloth, B. burgdorferi was found in 8 of
147 adults, 0 of 4 nymphs, and a minimum of 1 of 59 larvae processed in 15 pools, with
no apparent temporal trends. All infected drag-sampled D. variabilis came from site C1,
with the exception of two infected adults from sites 11 and 12. B. burgdorferi detection in
the additional drag-sampled tick species across the years was as follows: 0 of 4 A.
americanum, 0 of 1 A. maculatum, 0 of 1 I. caakei, 2 of 2 H. leporispalustris (both
nymphs from site C1), and 0 of 3 I. dentatus.
Host infection with B. burgdorferi. Of the 1056 white-footed mouse ear biopsies that
were collected from 2004-2008, only 2.1% of 57 6 tested positive from inland sites (11:
0/71; 12: 9/213; 13: 3/169; 14: 0/121, with no apparent trend overtime), whereas 11.9% of
480 coastal mice tested positive (P < 0.0002). Infection prevalence was high in all years
at site CI (21.6 - 80%), and low and inconsistent at sites further north (Figure 3.2). A
77
total of 148 chipmunk biopsies were assayed for infection, of which 4.2% of 72 were
positive along the inland transect (I1: 0/9; 12: 0/20; 13: 0/4; 14: 3/39, and 6.6% of 76 (all
at site C1) were positive along the coastal transect (Figure 3.2; no difference between
transects, P = 0.72). A total of 240 car biopsies from alternative mammal species were
assayed for infection: 6.5% of 124 and 4.3% of 116 were positive along the inland and
coastal transects, respectively (no difference between transects, P = 0.57). Positive
alternative mammals comprised four species (raccoon, red squirrel, meadow jumping
mouse, and Virginia opossum) at two inland sites (13 and I4) and three species (eastern
gray squirrel, raccoon, and meadow jumping mouse) at two coastal sites (C1 and C3).
B. burgdorferi infection in I. scapularis removed from hosts. Along the inland transect,
all five larvae removed fi'om two mice at site I3 in 2007 were negative, whereas the
single nymph from the same site and year was positive (Table 3.2), and the single nymph
from site 11 in 2008 was negative. Along the coastal transect, infection in I. scapularis
removed fi-om mice and chipmunks was highest at all years at site C1 and decreased at
sites to the north, with no infection in the ticks from site C4 (Figure 3.2). Alternative
mammalian hosts harboring I. scapularis were found at all coastal sites, but
B. burgdorferi infection in these ticks was found only at sites Cl and C2 (Supplement 2).
Birds harboring I. scapularis were found only at sites C1 and C2, and B. burgdorferi
infection in bird-derived I. scapularis occurred only at site C1 (Supplement 2).
B. burgdorferi infection in alternative tick species. Assaying samples of the seven
species of ticks other than I. scapularis removed from mammals and birds revealed
78
relatively high levels of B. burgdorferi infection in alternative ticks at sites with
sympatric I. scapularis (Supplement 2). Additionally, some alternative tick species from
inland sites in the apparent absence of I. scapularis were associated with low-level
infection prevalence. Infected D. variabilis removed from mice were found at sites 12 and
13, where a total of 7.1% of 98 nymphs (positives from 6 animals) and a minimum of
0.7% of 2006 larvae (positives from 14 animals) collected at these two sites together
were infected. Contrast these results with that at site C l , where sympatric I. scapularis
occur in high density: 17.6% of 74 nymphal (positives from 12 animals) and a minimum
of 2.5% of 672 larval (positives from 17 animals) D. variabilis removed from white-
footed mice were infected.
Of the alternative tick species collected from vertebrates, I. texanus was the most
abundant and was ubiquitous across our study sites, commonly found parasitizing
raccoons (Table 3.1). Infected attached I. texanus were found not only at all coastal
transect sites, but low-level infection was documented in all three life stages from inland
sites (2.4-2.6%; Supplement 2) as well. In summary, infected D. variabilis, I. texanus, I.
caakei, I. dentatus, I. marxi and H. leporispalustris were removed from coastal wildlife,
and infected D. variabilis, I. texanus, and I. dentatus were removed from inland wildlife
(Supplement 2).
Nucleotide sequencing. Pathogen identity was further confirmed by nucleotide
sequencing of the 168-238 rRNA IGS region of a subset of the PCR positive Samples
reported above (i.e., including infection in host ear tissue, 1. scapularis ticks, and other
tick species). We obtained sequence confirmation of 26 samples from coastal sites where
79
I. scapularis is abundant, and 28 samples from inland sites where I. scapularis is absent
or at a very low density. Representatives of all three B. burgdorferi RST groups were
found at sites characterized by both the presence and absence of I. scapularis (Table 3.2).
Opportunistic detection of other Borrelia species. Using IGS PCR followed by DNA
sequencing, we detected spirochetes that group with B. miyamotoi, an organism
genetically similar to relapsing fever group spirochetes, in ears from 12 white-footed
mice at sites C2-C4. Two tick samples (a single nymph and a pool of 20 larval
I. scapularis) removed from the same eastern chipmunk in 2005 at site C1 also tested
positive for B. miyamotoi-like spirochetes Representative B. miyamotoi-like sequences
from these ears and ticks are deposited in Genbank (accession numbers GQ856588-
GQ8565 89). We also detected spirochetes that group with B. andersonii, a Lyme
borreliosis group spirochete associated with I. dentatus ticks, birds, and rabbits (Marconi
et al. 1995), in one I. dentatus nymph removed from a song sparrow from site I3 in 2007
(GQ856590) and in ear tissue from one white-footed mouse from site C1 in 2006. As the
lGS region of B. andersonii was not previously characterized in Genbank, we identified
this species based on characterization of the 16S gene (data not shown).
Discussion
Lyme disease is emerging in the United States due largely to the spread of the
bridging vector, 1. scapularis. Both in the Northeast and the Upper Midwest, the spread
of this tick has been documented at a gross scale, but few have studied this spread in real
time (Falco et al. 1995; Schulze et al. 1998; Cortinas and Kitron 2006) are exceptions).
80
By monitoring ticks and hosts over five years, we assessed the temporal and spatial
dynamics of I. scapularis populations along coastal and inland transects in western lower
Michigan. Our aim was to investigate three non-mutually exclusive hypotheses to explain
how the disease system is emerging across the landscape; we refer to these as the ‘tick-
first’, ‘dual-invasion’, and ‘spirochete-first’ hypotheses. Our results suggest that there are
distinct but overlapping processes that determine I. scapularis/B. burgdorferi dynamics.
In some areas, observations of newly-arrived I. scapularis into a host/vector community
where B. burgdorferi is absent support the ‘tick-first’ and ‘dual-invasion’ hypotheses.
Conversely, in other areas, the detection of infected wildlife and alternative tick species
preceded the arrival of I. scapularis, consistent with the ‘spirochete-first’ hypothesis.
Here we summarize our key findings and consider their support for these hypotheses.
I. scapularis invasion. At a broad spatial scale, a continuum of endemicity of
I. scapularis exists in the Midwest, whereby Wisconsin, Minnesota, and a focal area in
Michigan’s Upper Peninsula harbor long-established tick populations (Jackson and
Defoliart 1970; Drew et al. 1988; Strand et al. 1992), Illinois and Indiana harbor ticks that
colonized later (Pinger and Glancy 1989; Bouseman et al. 1990), and Michigan’s Lower
Peninsula represents the most recent invasion front (Foster 2004; Hamer et al. 2007) . We
postulate that these Michigan ticks represent an expanding focus from recently-described
populations in northern Indiana (these are adjacent to the Michigan border and near Lake
Michigan; (Pinger et al. 1996). The density of questing nymphs at coastal site’Cl, where
I. scapularis was most common, averaged 16.8/1000 m2, which is higher than the
average density of 6.6/1000 m2 in I. scapularis-positive sites from throughout the eastern
81
U.S. (Diuk-Wasser et al. 2006). Densities of nymphs at the other positive sites in our
study were much lower.
As in many other states that have witnessed an emergence of Lyme disease in the
past four decades, white-tailed deer populations have increased sharply in Michigan’s
southern Lower Peninsula over the past 50 years, fiom near extirpation in the 19603 to
nearly one million animals presently as a consequence of harvest restrictions and
reforestation (MDNR 2002). Deer densities throughout the invasion zone have clearly
passed the threshold necessary to support tick populations. The reforestation that
supported the deer expansion also created an abundance of suitable habitat for other
woodland hosts and I. scapularis. Thus the observed recent invasion may be a
consequence of a sufficient number of introductions and large enough propagule size
deposited to achieve a sufficient density to establish and disperse.
The progressive increase in tick density along the coastal sites from south to north
among years is suggestive of rapid tick invasion and colonization northward along the
coastal dune forests of Lake Michigan. I. scapularis could only be detected at the
southernmost site at the start of the study, whereas all four sites harbored established
populations by the end of the study period. Numerous authors have proposed that climate
change may facilitate tick invasion of new areas due to a warming of the microclimate
experienced by the ticks and shorter periods of fi'eeze (e.g. (Githeko et al. 2000; Lindgren
and Gustafson 2001). The prospects for such shifts are debatable (Randolph 2004), and
I. scapularis in lower Michigan is well south of its distributional limits, so we presently
have no evidence to suggest climate change as a driver for this invasion.
82
The disparity in tick distribution between our two transects suggests that there are
important ecological differences between coastal and inland sites (Sonenshine et al.
1995), with the former allowing for more introductions of ticks, and/or higher success in
establishment after introduction. Hosts with high tick burdens and large home ranges,
such as deer and migratory birds, are critical for long distance dispersal of I. scapularis,
whereas small mammals with small home ranges may limit invasion by diverting ticks
from more vagile hosts (Madhav et al. 2004). Deer densities were not quantified in our
study, but densities of other mammalian and non-migratory birds during the study period,
as indexed by our trap and mist-net success, were similar between coastal and inland
sites. Migratory birds tend to concentrate along shorelines (Alerstam 1978) and birds fly
parallel to the long axis of Lake Michigan (Diehl et al. 2003), which may allow for more
drop-offs of ticks from birds at coastal compared to inland sites.
B. burgdorferi distribution, abundance, and diversity. Overall infection prevalences in
drag-sampled I. scapularis in the invasion zone were 8.6 and 36.6% in nymphs and
adults, respectively; these infection prevalences are generally lower than those that
characterize I. scapularis in the endemic Northeast (15% in nymphs (Gatewood et al.
2009); 49% in adults (Schulze et al. 2003)). The density of infected nymphs (DIN) has
been shown to be correlated with the incidence of human Lyme disease in a given area
(Stafford et al. 1998). Across our study, the DIN at site or was 1.5/1000 m2, which is
similar to DIN s in endemic sites of the northeast (1.3/1000 m2) and Midwest (1.0/1000
2 . . .
m ) (Gatewood et al. 2009). The overall prevalence of infection of white-footed mice at
83
site C1 was 32.2%, which is lower than that reported from endemic foci (for example,
76% by the end of summer in Connecticut (Barbour et al. 2009). In addition to the lower
nymphal infection prevalence as an explanation for lower prevalence of mouse infection,
the difference may be due in part to sampling prior to the end of the transmission period.
We detected B. burgdorferi not only in sites where I. scapularis is established or
apparently recently-invaded, but also within alternative tick species and wildlife hosts
occurring at sites in the apparent absence of I. scapularis. Cryptic transmission of
B. burgdorferi sensu lato has been documented in various host and tick species
(Anderson et al. 1989; Telford and Spielman 1989b; Brown and Lane 1992; Maupin et al.
1994; Oliver et al. 2003), but has not been investigated in an area of I. scapularis
invasion. The requirements for demonstration of a cryptic transmission cycle include the
presence and interactions of i) B. burgdorferi, ii) a vector-competent tick species, and iii)
reservoir-competent hosts. While our data are suggestive of a cryptic cycle and confirm
the presence of many reservoir competent hosts, and infection with several different
strains of B. burgdorferi in such hosts and in various tick species, it remains unclear
which tick species are responsible for maintaining transmission of B. burgdorferi in the
absence of I. scapularis. We documented infection in both questing and on-host D.
variabilis, as did Walker et al. (1998) at a highly endemic area in Michigan’s Upper
Peninsula, yet this tick species is not a competent vector (Piesman and Sinsky 1988). I.
cookei and I. marxi both had a low prevalence of infection in our study, but these infected
ticks were removed from animals (southern flying squirrel, raccoon) at site C l , where
I. scapularis and B. burgdorferi were at relatively high prevalence. Infections in these
alternative tick species thus may represent spillover from feeding on hosts shared by
84
I. scapularis. Furthermore, I. cookei is an incompetent vector (Ryder et a1. 1992). Two
infected tick species removed from wildlife at inland sites included I. texanus and
I. dentatus. Although the vector competency of I. texanus is unknown, its most important
host- raccoons- are known to be competent reservoirs, though only 35% as eflicient as
white-footed mice at infecting xenodiagnostic larvae (Fish and Daniels 1990). Both
1. dentatus and eastern cottontails are competent for transmission and maintenance of
B. burgdorferi sensu lato (Telford and Spielman 1989a), however parasitism of
I. dentatus on hosts other than birds and rabbits hosts was low during our study, and are
not likely to explain infections in these hosts. Lord et al. (1994) documented a site in
Pennsylvania at which isolations of B. burgdorferi from mice were common, yet
I. scapularis was not found, and suggested that some other mechanism besides
transmission by a tick vector was responsible.
We detected three different Borrelia species within ticks and hosts (B.
burgdorferi, B. miyamotoi, and B. andersonii), including a diverse assemblage of B.
burgdorferi strains. The diversity of B. burgdorferi RST types along the coast suggests
high rates of pathogen flow and/or multiple introduction events, rather than isolated
founder effects. Similarly, all three RST types were found in alternative tick vectors or
ear biopsies from sites beyond the invasion front. Studies of the population genetic
structure of B. burgdorflari will be useful for determining the dynamics of B. burgdorferi
in the presence and absence of I. scapularis, and how cryptic cycles affect B. burgdorferi
infection prevalence and strain types in the bridging vector 1. scapularis.
Intriguingly, a single I. scapularis larva dragged at an inland site was infected
with B. burgdorferi, and sequencing of the IGS region was used to identify this organism
85
as RST 2, which is the most common type in the Midwest. Evidence for transovarial
passage of B. burgdorferi, which may result in infected questing I. scapularis larvae, is
rare (but see Magnarelli et al. (1987) who found that transovarial transmission in I.
dammini occurs at a highly variable rate) While the single infected questing larva we
report herein is of little epidemiological significance, our finding was especially
surprising considering the extremely low density of I. scapularis in the area. One possible
explanation is this infection may represent spirochetes acquired during a partial larval
bloodmeal from an infected host that terminated in the larva being groomed off.
Implications for hypotheses. Our detections of B. burgdorferi shed light on our
hypotheses about the emergence of the lower Michigan I. scapularis/B. burgdorferi
system. At four sites (C3, C4, 11 , 13), we detected 1. scapularis on mice after previous
years of surveillance had failed to detect infested mice, and we interpret this as evidence
for tick invasion into these sites. At site C4, no B. burgdorferi has yet been found, despite
finding I. scapularis in both 2007 and 2008, supporting a tick-first process. At site C3,
the first detection of B. burgdorferi coincided with that of I. scapularis, supporting a dual
invasion process. In contrast, at sites 11 and 13, B. burgdorferi was found in other sample
types prior to detecting I. scapularis, supporting a spirochete-first process. As described
above, it may be that B. burgdorferi is being maintained in cryptic transmission cycles. It
is also possible that a very low density 1. scapularis p0pulation - below our detection
threshold - exists at such sites. Such low density 1. scapularis populations, hawever,
would not generally be considered capable of maintaining B. burgdorferi (Madhav et al.
2004; Ogden et al. 2008).
86
Implications for detection of I. scapularis/B. burgdorferi invasion. Studies of invasion
are most often initiated once the invader has reached densities that result in negative
impacts, including disease, as resources for active surveillance and research are not often
appropriated in the absence of a problem. Rarely are standardized investigations of
wildlife undertaken in areas of no or low tick density that can provide a sensitive real-
tirne warning system for invading ticks, spirochetes and impending Lyme disease risk.
One exception is Schulze et al. (1998), who reported a significant increase in human
Lyme disease cases in 1990-1995 in Hunterdon County, NJ, that occurred after a
geographic expansion in ticks on deer in the county between 1981 and 1987. F alco et al.
(1995) detected a 2.6-fold increase in questing I. scapularis nymphal density in endemic
Westchester Co., NY, in 1991 versus 1984 when nymphal densities increased from 13 to
34 nymphs/1000 m2, coincident with an increase in reported human cases of infection. In
contrast, at our study site C2, where nymphal drag data demonstrate invasion most
clearly, densities ranged from zero to 5.3 nymphs/1000m2 in 2004-2008 (Figure 3.2.)
Thus, we are assessing an invasion during its earliest stages, and human disease incidence
is still very low (see below).
We found that prevalence of infestation of I. scapularis on white-footed mice
provided a sensitive index of invasion (Lord et al. 1992), as detection of ticks on mice
generally preceded detection of ticks on drag cloths, alternative mammal species, and
birds. Mouse sampling was also more useful than two other methods of surveillance for
ticks and Lyme disease foci. Assessment of ticks on hunter-harvested deer in November
across the study area was hampered by low densities of adult ticks in this early-stage
invasion, and by a mismatch in the timing of the fall hunting season in relation to adult
87
questing phenology in the invasion zone (their major activity peak is in the spring; S.
Hamer, unpublished data). The efficacy of canine serosurveillance in the area was likely
reduced by widespread use of anti-parasite prophylaxis on pet dogs living in the invasion
zone (Hamer et al. 2009).
Similarly, alternative tick species and wildlife hosts enabled detection of B.
burgdorferi in the apparent absence of I. scapularis. Detection of cryptically cycling B.
burgdorferi decouples its invasion from that of I. scapularis, and may hasten B.
burgdorferi transmission dynamics among an invading I. scapularis ticks. Cryptically
cycling B. burgdorferi thus may decrease the time lag between first detection of
I. scapularis and first detection of B. burgdorferi-infected I. scapularis.
Study limitations. The results we present are based on sampling within the May and June
period only, which does not include all months of blacklegged tick activity. The most
commonly cited 1. scapularis phenology, based on studies in Westchester Co., NY (Fish
1993), is characterized by bimodal peaks in adult activity in the spring and fall, and
nymphal activity in the early summer which precedes peak larval activity in the late
summer. Gatewood et al. (2009) noted that while the Northeast is generally characterized
by distinctly separate peaks in nymphal and larval activity in June and August,
respectively, as indicated above, the Midwest is characterized by synchronous feeding of
nymphs and larvae in June through July, with a much smaller late-season peak in larval
activity. While we present five years of data from May and June only (early Summer), all
our sites were also sampled during July-August (late summer) in at least one year with
similar rodent trapping and drag sampling effort (unpublished data). At all sites where we
88
found I. scapularis in the late summer, this species was also found in the early summer
period reported here. Furthermore, in 2008-2009 we conducted a longitudinal study of
I. scapularis phenology at monthly intervals at site Cl (unpublished data). Larval
infestation of mice was greatest in June (83.8%), followed by a decline (44.1, 48, 24.5,
and 19.4% monthly infestation prevalence in July-October, respectively), and nymphal
infestation followed a similar trend (24.3, 8.5, 2.0, 2.0, and 0% in June-October,
respectively). These patterns support the synchronized feeding reported by Gatewood et
al. (2009) and are similar those reported by Godsey et al. (1987) in Wisconsin and Strand
et al. (1992) in Michigan’s Upper Peninsula. Thus, we conclude that our sampling
focused on the period during which we had the best chance of detecting infested mice
each year, and that this period was appropriate for answering our research questions.
A longer duration of sampling at all sites would have provided greater sensitivity
of detection of stable I. scapularis populations. While neither our data nor that of
Gatewood et al. (2009) show significant larval activity in the late summer, late summer
larval activity of greater (Northeast) or equivalent (Midwest) magnitude to early summer
activity has been reported (Kitron et al. 1991; Daniels et al. 1996; Jones and Kitron
2000). In the Northeast, the early peak in May-June represents the remainder of the
cohort of larvae that hatched the previous July and successfully overwintered, whereas
the second peak of higher magnitude in August represents the subsequent cohort of
newly-hatched larvae from recently laid eggs by adults that were active in spring as well
as the previous fall (after ovipositional diapause; (Daniels et al. 1996)). In the Midwest,
the separate peaks observed by some researchers at some sites are posited to result from
the separate oviposition periods of the Spring and fall adults (Kitron et al. 1991; Daniels
89
et al. 1996; Jones and Kitron 2000), whereby the early larvae may result from oviposition
by fall adults and the late summer larvae result from oviposition by spring adults.
Furthermore, there has been a suggestion that this bimodality may be more pronounced in
newly-establishing populations (Kitron et al. 1991; Daniels et al. 1996; Jones and Kitron
2000). Therefore, if blacklegged tick invasion in Michigan is driven by adult ticks
moving in during the spring (e. g., on deer), our sampling may miss the resulting larvae
emerging late that first summer.
Assaying ticks of all stages and species removed from hosts increased our
capability to detect the presence of B. burgdorferi at a site compared to assaying I.
scapularis or host tissue alone. Our protocol, however, did not maximally allow us to
detect B. burgdorferi infection as we did not assay all ticks removed from hosts. We
selected a random subset of three per species/stage/sex combination to assay for
infection, sometimes resulting in over ten ticks tested from some hosts that were infested
my multiple tick species and stages. Given limited resources, this protocol allows us to
gather infection data fiom ticks removed from a greater number and diversity of hosts as
well as to avoid over-representing heavily parasitized animals in the overall results of tick
infection. Testing of all ticks from such animals would have increased the odds of both
detecting a positive tick on a given animal and detecting different genospecies that co-
infect the same vertebrate host (although that latter was not a main study objective).
While comprehensive tick testing may have resulted in a finer temporal resolution of
invasion dynamics, we do not believe it would have altered our qualitative conclusions.
90
Implications for disease and prevention. Data collected from careful study of the
underlying ecology and transmission of zoonotic pathogens within animal reservoirs
before and during invasion can provide an early warning of increasing disease risk to
human and companion animals. Cryptic cycling of B. burgdorferi does not imply human
risk of Lyme disease in these areas. They, however, may hasten the establishment of
transmission dynamics of I. scapularis. They also may introduce different strains of
B. burgdorferi to I. scapularis, which may result in different clinical manifestations
(Wormser et al. 1999).
Indeed, a significant increase in confinned cases of human Lyme disease (R2 =
0.28; l-tailed P = 0.03) has occurred within the zone of active 1. scapularis invasion in
southwest/westem Michigan. From 1996 (when Lyme disease became reportable in
Michigan) through 2008, incidence in the 14-county region reached 0.63 cases per
100,000 people in 2008 (E. F aster, Michigan Department of Community Health, pers.
comm). Incidence remains substantially less than the average annual incidence in the 10
endemic states from which >93% of human Lyme disease across the US. is reported
(average of 29.2 cases per 100,000 population for 2003-2005; (Bacon et al. 2007). In
contrast with the significant peridomestic exposures that occur in the Northeastern United
States, recreational exposure is probably common in the Michigan invasion zone, which
is dominated by recreational areas, camm, vacation homes, and rural communities.
Recreational exposure may produce low reported disease incidence locally, as
recreational visitors will often return home before developing symptoms that lead them to
seek medical attention. Spatial epidemiology and trace-back of human Lyme disease
cases reported elsewhere in lower Michigan could help shed light on the this possibility.
91
Supplementary material I. Bird-tick associations and infestation prevalences on the
inland and coastal transects in Lower Michigan, May-June, 2004-2008. Number of birds
carrying larvae/nymphs/adults is indicated with the percent of birds infested with at least
one tick of any life stage in parenthesis. All ticks were tested for Borrelia burgdorferi;
those birds with infected ticks are indicated in Supplementary material 2.
92
Species
American goidtinch
American redstart
American robin
Baltimore oriole
black-billed cuckoo
black-capped Chickadee
black-throated green war
blue jay
blue-winged warbler
brown thrasher
brown-headed cowbird
cedar waxwing
chipping sparrow
common grackle
common yellowthroat
Connecticut warbler
downy woodpecker
Eastern bluebird
Eastern kingbird
Eastern phoebe
Eastern towhee
Eastem wood-pewee
Empidonax flycatchers
EurOpean starting
field sparrow
gray catbird
gray-checked thrush
great crested flycatcher
hairy woodpecker
house finch
house wren
indigo bunting
mourning dove
Northern cardinal
Northem waterthrush
orchard oriole
ovenbird
pileated woodpecker
red-eyed vireo
red-winged blackbird
rose-breasted grosbeak
song sparrow
swamp sparrow
Tennessee warbler
Trail's flycatcher
tree swallow
tufted titmouse
unknown
warbling vireo
white-crowned sparrow
Wilson's warbler
wood thrush
yellow warbler
yellow-breasted chat
yellow-sided flycatcher
fibw-throated vireo
flamers:
Q4
00!
7
adaawnnnnnwaahgw
&
Total
22
22
20
ooouooooomoo-socoro05:d-xoo-qoficow-so-sauoaaw-sw-sw-naro-sannoooa:oo
Coastal
no. I. scapularis I. dentatus H. lep.
418/0 (40)
1/1/0 (25)
0/2/0 (6.1)
01011 (11.1)
1/0/0 (33.3)
0/1/0 (100)
01110 (11.1)
01110 (4.8)
11110 (14.3) 01011 (14.3)
1/1/0 (11.1)
11110 (10) 212/0 (15)
Inland
I. dentatus H. lep.
33°38
911110 (53.8) 01210 (7.7)
21010 (10.5)
2/1/0 (66.7)
31010 (6.8)
11010 (16.7)
41310 (6.6)
21010 (50)
2/0/0 (22.2)
o
8 com-Ho--srsoo—noam-scoorAddwmofiawwmwaaoaao
13/0/0 (16.3) 0/1/0 (1.3)
1/0/0 (20)
13/6/0 (32.6) 1/2/0 (7.0)
1/0/0 (25)
2/1/0 (100)
1/1/0 (3.4)
1/0/0 (100)
Addgnwmmnowhoagm
1 237 ans/015.5) 4/4/1 (align/we) 504 57/23/0 (12.9) mm (1.2)
93
Supplementary material 2. Infection prevalence of each species of tick removed from
mammalian and avian hosts in lower Michigan, May-June, 2004-2008. Infection
prevalence is presented as the percent of ticks that tested positive with the sample size of
ticks tested is in parenthesis. Larval infection is reported as the minimrun infection
prevalence. LL = larvae; NN = nymph; AA = adult; misc. = miscellaneous tick species
indicated in column; A. amer. = A. americanum; H. lep. = H. leporispalustris; .
Regarding birds, only those species with positive ticks are listed below; a complete list of
all 56 bird species and associated ticks is provided in Supplementary material 1.
94
Host species Transact Stag variabilis egg/lads tax
Northern short-
tailed shrew Coastal LL 0 (1)
LL 0 (2)
N” 0 (1) 0 (1)
Inland AA 1.1 (91) o (1) 0 (1)
LL 0 (52)
NN 6.9 (29) 0(5)
Vlrginia opossum Coastal AA
LL 0(3) 100(1)
Southern flying MN 0 (2) I. marxl
squirrel Coastal AA 0 (3) 50 (2)
LL 0 (19)
"N 0(3) 0 (1)
Striped skunk Coastal AA 0(2)
Meadow vole Coastal LL 0 (1)
Long—tailed LL 0 (1) 0 (1)
weasel Inland NN 0(2)
LL 0.7 (2116) 0(5) 0 (1)
Inland NN 6.6 (106) 50 (2)
White-footed LL 2.4 (716) 5.4 (705) 0 (1)
mouse Coastal NN 17.6 (74) 37.8 (82) o (1)
LL 0(1) 2.4 (41) 0(1)
MN 0 (7) 2.6 (78) 0 (6)
Inland AA 1.4 (209) 2.4 (83) 0(2)
LL 0 (1) 0(5) 0 (158)
NM 0 (1) 9.1 (22) 0 (94) A. amer.
Raccoon Coastal AA 2.4 (164) 0 (2) 2.3 (88) 11.1 (9) 0 (2)
LL 0(2) 25(4)
Eastern gray NN 4O (5)
squirrel Coastal AA 0(6)
LL 0(2)
Fox squirrel Coastal NM 0 (1) H. I .
Inland AA 0(3) 12-9 (31) 0(2)
LL 66.7 (3)
Eastern NN 0 (1) 0 (3)
cottontail Coastal AA 0 (2) 0 (6)
LL 5 (20)
Inland NM 0 (2)
Eastern LL 0 (2) 9.8 (82) o (1)
chipmunk Coastal NN 26.2 (65) _I._marx_l
MN 0 (1) 0 (1)
Inland AA 0 (1) 0(3)
LL 0 (1)
Red squirrel Coastal AA 0 (3)
LL 0 (12)
Inland NN 0(1) 0(1)
LL 0 (1)
Meadow jumping NM 0 (2)
mouse Coastal AA 0 (1)
Inland LL 2-9 (35)
LL 7.7 (13) H-' -
American Robin Coastal NN 37.5 (16) 5°12)
LL 100(1)
Northern cardinal Coastal MN 100(1)
Brown-headed
cowbird Coastal LL 10° (1)
LL 5.1 (39)
Song sparrow Inland NN 12'5 (8)
Red-winged
blackbird Inland LL 2-1 ‘47)
D
I
I.
1
1
anus cookei dentatus misc.
95
.8588... c0833.... exam: £33
8:80.588 mEEoécE< 0.9..» a
5.3 “.23.... was cocoon. .8880 9.0
.—. 305039..” 5 8330.5 0.3 9.0330930 x0: Ucw means 23 oEogmlmflowamw
8.: 888 8.9 93.. 8.9 88:... Ba 8880 1
8.: 99. 88.: 8898 .8 28.... .828
8.9 :88 8.8.: em: 9 8880 88...
8.89 88. 89 9:8 N. 28.5 8.8.... 3885.
89 :88 89 88: m 8880
8.~: ::o 8.9 9:8 8.9 :88 8.9 9:8 8. 28.5 858» no...
8.: 88: .89 988. 8. : 88: .8 8880
8.9 8.98 E 28.5 .5556 Eosaom
.8: ::o .8: :88 a 8880
89 :88 .8: 3:8 89 :88 v 28.5 .583. 5888
.8 : 9:. a .288 8588 .6".
89 8.8 89 :8: 8 8880
.2538 .88 583w
8.: :88 8.: 9:8 8.9 998 8.8. .8899 8.9 N: :8 889 8:8 8 8880
88: :9. 889 8.898 889 88: so 28.5 8.8688”.
8.9 2:. 8.89 8898: 8.89 99.8. 888 .8880
8.9 98 8.89 8898... 88 28.5 385 882.353
89 9:. 89 88: N 28.5
.380; 3.8.9.3
.8: 88: a 8880 59. 388.2
.8: :5 89 9:8 m 8.88 5.3.8 85.8
8.9 :88 8.9 8: 88: 8.9.. . :89 9o: 8. 8880 8528 85.... 59.8w
— 28.5
. :89 888 8.9 88 8.8: 888 2 8880 5888 8.58.5
:9 :5 8.89 8:8 mm 28.5
n .3.»on 395m
1 2 28.5 8.838.... 52.82
erasaaetons‘ .t :22: .. select .8 82.09... .8 3:88.. .. statics» .. minutes .0 02:33 83: 8... also... «Qua
.oz
.88. 8. 5.3 3.25 8. <2... 5.3 238.5 525... 6:
.22. u < 888 u 2 at... u 8 .88
. . 8.8 28 28.8889. 82.5.29 ...m 28.8
96
Table 3.2. Sequence continuation and accession numbers of B. burgdorferi infection in
various sample types from coastal and inland sites in lower Michigan, 2004-2008. I =
inland transect site; C = coastal transect site. RST = 168 — 23S rRNA spacer type of
B. burgdorferi. AF = adult female; AM = adult male; N = nymph; L = larvae; WFMO =
white-footed mouse; SFSQ = southern flying squirrel; EAGS = eastern gray squirrel;
VIOP = Virginia opossum; EACH — eastern chipmunk, NOCA = Northern cardinal;
AMRO = American robin, SOSP = song sparrow; COON = raccoon, MJMO = meadow
jumping mouse; EACO = eastern cottontail.
97
8'10 Your Sample Type Host Species or bra RST 1,2, ,
C1 2004 D. variabilis AF Drag J (2 3) 63:56:31
ear biopsy WFMO 1 60856630
ear biopsy WFMO 2 60856631
I. manri AF SFSQ 1 60856615
I- scapularis L wmo 2 60856598
I. scapularis N EAGS 2 60856614
2005 D. variabilis L Drag 1 60856596
D. variabilis L WFMO 3 (50355500
I- scapularis N VIOP 2 60856594
I- scapularis N WFMO 3 60856595
I. texanus L SFSQ 2 60856601
2006 D.‘variabilis N WFMO 2 60856639
I. scapularis N Drag 2 60856592
I. scapularis N EACH 1 60856612
I. scapularis N WFMO 2 60856611
2007 D. variabilis N WFMO 3 60856643
ear biopsy EACH 3 60856632
. leporispalustris N Drag 3 60856609
I. scapularis L NOCA 3 60656619
I. scapularis N AMRO 3 60856616
62 2005 ear biopsy WFMO 2 60856620
2006 I. scapularis L WFMO 2 60856593
2008 ear biopsy WFMO 2 60856610
C3 2006 ear biopsy WFMO 3 60856623
ear biopsy WFMO 3 60856622
ear biopsy WFMO 3 60856621
12 2004 D. variabilis L WFMO 1 60856638
2005 D. variabilis L WFMO 1 60856602
2006 ear biopsy WFMO 2 60856628
ear biopsy WFMO 2 60856629
ear biopsy WFMO 2 60856627
ear biopsy WFMO 2 60856626
ear biopsy WFMO 2 60856625
ear biopsy WFMO 2 60856624
2007 D. variabilis L WFMO 3 60856603
I3 2004 D. variabilis L WFMO 2 60856599
D. variabilis L WFMO 2 60856597
I. texanus N COON 3 60856640
_ 2006 ear biopsy WFMO 2 60856642
ear biopsy MJMO 2 60856641
2007 I. scapularis N WFMO 2 60856606
I. scapularis L Drag 2 6U190359
D. variabilis L WFMO 2 60656608
D. variabilis L WFMO 2 60856607
D. variabilis L WFMO 2 60856605
D. variabilis L WFMO 2 60856604
ear biopsy VIOP 2 60856636
ear biopsy COON 2 60856634
ear biopsy COON 2 60856635
ear biopsy WFMO 2 60856633
2008 I. dentatus AF EACO 2 60856613
I. dentatus AM EACO 2 30856617
I4 2004 I. texanus N COON 1 30356313
2006 ear biopsy EACH 2 60856637
98
>
9
LL)?»
C2
Ii. '4
C1 ”
Figure. 3.1. Locations of study sites in Lower Michigan, 2004-2008. The four sites along
Michigan’s west coast comprise the coastal transect (from south to north, Cl-C4), and the
four inland sites comprise the inland transect (from southwest to northeast, 11 -I4).
Shading in the Lower Peninsula represents the three-county region where I. scapularis
were detected on small mammals in 2002-2003(Foster 2004). The cross-hatched county
in the Upper Peninsula is Menominee COunty, Michigan’s longstanding endemic focus of
I. scapularis and B. burgdorferi. .
99
880 05 5 0.30000
.800 9 08—0 0003 0008 003003 98 0208003: 00055 .0030. 000mg 05:: 05 :0 023800.05 .08 mafia 30.0 03080 00.50900 0003
35802.. 8 .88 a 8 as 38 a 6 as E 68638 5 am as; 88 8 .40 as a... $055.8 a 0% 238 3 35:8
£000 .8033 3 no. 80¢ 00%: 0.3 006005 .60 60: Q: can .38: 80¢ 00.580.— 308»: A5 Age—SSE 030002 ESE—ES BEE
0.083%00 .N 3 5 Eu 0000365 00300? {$03330 .m 438% £000 9 E000 :5. 4098.? 08 029: 30$ 8% 00000000 $0830
.«0 0058:: 05 £3, €300.05 2088.0 00 00300080 05 00000205 wagons.“ no _0>0.._ .wooméoom 6:57.32 000053 30000 05 mo 8%
So.“ 05 3 fiafiamsum N 53> 03:800.. Emma 0.5.8020 E00000 was A2mqfi00c «05 0958 000080023 90 0200303 .N.m .00:me
J
E ”.2. a $80.3 A823 .. $80.2 Q 2: 3:30 5 m 33.8
8:30. 6:3. 6:: ”68$... 60 8 § 0.8 $92. 6:60
882: 38:0 €08 6055:8000 893 $83 as;
S .35.» E F $8 :2 Q 2.83
5, . .8 6v 3 @8 8v 5 ”.8
§ , 88 n 2: as; 63 é a: 5.
EV . a: 3.: 2 88 av 33 3 Ag 5 60 as
3:92 ., 2: § 3 av Z: 6 ".8 § 3 § §
3 E § 5 E § é 8 § §
as .0 . c:
Q E 3
§ §
coca mocu «com
100
80000 n U .0000 000.0 3200 0000500000 8 000000 0008:: 88 05 :03 030000
800000 05 00 0000000000 .00 0028 00.0 0000:0500 00:00.08 88888v 0w80 00: £000 000 0000—8600 00:00.80 83000 0.00 0020000
00:08.? 00.0 080005000 .000 00:: 0.800080% .00 0:0 00 00308 0.003 0.2838000. .N 02 .wooméoom .082. 080 .082 8 000088
.8080 05 00 00:0 0005 08 .0 0.2808000. .N 00.088080 00 00000—05000 00:00.08 00.008033 .m 800 0030000 09803.. .m.m .0830
000 0.0 a 00. 0.: 00... 00 0.00 0000.00 3 0.00 :00 0.00 3.03
80 0.0 3.: 0.0 3 0.0 $00 0.0 600 0.00 A00 00 £00 0.0 $00 0.0 0005»:
,. 30 00 a: 0.0 00: 010 00 0.0 E 0.0 0000 0.0 as; 0.0 0020..
, no No 00
0000 Son moon moon voou moon 300 0000 meow voom moon .300 meow meow voou m
o '11 , w
m 00.0 n a flNHiI 00.0 n. 0 w
. 00.0 n m u .s
m 00. 0 00.0 u 00. 3 2
Ir 0.
m 000 8.0.28 00 m
w 20:52 I M
2000 00 m
0w>us WU . 1
0.0-00 m
3 Z
101
References
Alerstam T (1978). Re-Oriented Bird Migration in Coastal Areas - Dispersal to Suitable
Resting Grounds. Oikos 30:405-408.
Altschul SF, Gish W, Miller W, Myers EW, and Lipman DJ (1990). Basic Local
Alignment Search Tool. Journal of Molecular Biology 215:403—410.
Anderson JF, Magnarelli LA, Lefebvre RB, Andreadis TG, McAninch JB, Pemg GC, et
al. (1989). Antigenically Variable Borrelia-Burgdorferi Isolated from Cottontail
Rabbits and Ixodes-Dentatus in Rural and Urban Areas. Journal of Clinical
Microbiology 27: 13-20.
Bacon RM, Kugeler KJ, Griffith KS, and Mead PS (2007). Lyme disease - United States,
2003-2005 (Reprinted from MMWR, vol 56, pg 573-5 76, 2007). Jama-Journal of
the American Medical Association 298:278-279.
Barbour AG, Bunikis J, Travinsky B, Hoen AG, Diuk-Wasser MA, Fish D, et al. (2009).
Niche Partitioning of Borrelia burgdorferi and Borrelia miyamotoi in the Same
Tick Vector and Mammalian Reservoir Species. Am J T rop Med Hyg 81 :1 120-
1131.
Bouseman JK, Kitron U, Kirkpatrick CE, Siege] J, and Todd KS (1990). Status of Ixodes-
Dammini (Acari, Ixodidae) in Illinois. Journal of Medical Entomology 27:556-
560.
Brown RN, and Lane RS (1992). Lyme-Disease in California - a Novel Enzootic
Transmission Cycle of Borrelia-Burgdorferi. Science 256:1439-1442.
Bunikis J, Garpmo U, Tsao J, Berglund J, Fish D, and Barbour AG (2004). Sequence
typing reveals extensive strain diversity of the Lyme borreliosis agents Borrelia
burgdorferi in North America and Borrelia afzelii in EurOpe. Microbiology-ng
150:1741-1755.
Cortinas MR, and Kitron U (2006). County-level surveillance of white-tailed deer
infestation by Ixodes scapularis and Dermacentor albipictus (Acari: Ixodidae)
along the Illinois River. J Med Entomol 43:810-819.
Daniels TJ, Falco RC, Curran KL, and Fish D (1996). Timing of Ixodes scapularis
(Acari: Ixodidae) oviposition and larval activity in southern New York. Journal of
Medical Entomolog/ 33: 140-147.
Dennis DT, Nekomoto TS, Victor J C, Paul WS, and Piesman J (1998). Reported
distribution of Ixodes scapularis and in Ixodes paciflcus (Acari : Ixodidae) in the
United States. Journal of Medical Entomology 35:629-638.
Diehl RH, Larkin RP, and Black J E (2003). Radar observations of bird migration over the
Great Lakes. Auk 120:278-290.
102
Diuk-Wasser MA, Gatewood AG, Cortinas MR, Yaremych-Hamer S, Tsao J, Kitron U,
et al. (2006). Spatiotemporal patterns of host-seeking Ixodes scapularis nymphs
(Acari : Iodidae) in the United States. Journal of Medical Entomology 43:166-
176.
Drew ML, Loken KI, Bey RF, and Swiggum RD (1988). Ixodes-Dammini - Occurrence
and Prevalence of Infection with Borrelia Spp in Minnesota. Journal of Wildlife
Diseases 24:708-710.
Durden LA, and Keirans JE (1996). Nymphs of the genus Ixodes (Acari: Ixodidae) of the
United States: taxonomy, identification key, distribution, hosts, and
medical/veterinary importance. Thomas Say Publications in Entomology,
Entomological Society of America.
F alco RC, Daniels TJ, and Fish D (1995). Increase in abundance of immature Ixodes
scapularis (Acari: Ixodidae) in an emergent Lyme disease endemic area. J Med
Entomol 32:522-526.
F alco RC, and Fish D (1992). A Comparison of Methods for Sampling the Deer Tick,
Ixodes-Dammini, in a Lyme-Disease Endemic Area. Experimental & Applied
Acarolagy 14:165-173.
Fish D (1993). Population ecology of Ixodes dammini. in H. S. Ginsberg, editor. Ecology
and Environmental management of Lyme disease. Rutgers University Press New
Brunswick, NJ.
Fish D, and Daniels TJ (1990). The Role of Medium-Sized Mammals as Reservoirs of
Borrelia-Burgdorferi in Southern New-York. Journal of Wildlife Diseases 26:339-
345.
Foster ES (2004). Ixodes scapularis (Acari: Ixodidae) and Borrelia burgdorferi in
southwest Michigan: population ecology and verification of a geographic risk
model. Masters Thesis, Michigan State University, East Lansing.
Gatewood AG, Liebman KA, Vourc'h G, Bunikis J, Hamer SA, Cortinas R, et al. (2009).
Climate and Tick Seasonality Are Predictors of Borrelia burgdorferi Genotype
Distribution. Applied and Environmental Microbiology 75:2476-2483.
Githeko AK, Lindsay SW, Confalonieri UE, and Patz JA (2000). Climate change and
vector-borne diseases: a regional analysis. Bulletin of the World Health
Organization 78:1 136-1 147.
Godsey MS, Amundson TE, Burgess EC, Schell W, Davis JP, Kaslow R, et al. (1987).
Lyme-Disease Ecology in Wisconsin - Distribution and Host Preferences of
Ixodes—Dammini, and Prevalence of Antibody to Borrelia-Burgdorferi in Small
Mammals. American Journal of Tropical Medicine and Hygiene 37 :180-1 87.
103
Guerra M, Walker E, Jones C, Paskewitz S, Cortinas MR, Stancil A, et al. (2002).
Predicting the risk of Lyme disease: Habitat suitability for Ixodes scapularis in the
north central United States. Emerging Infectious Diseases 8:289-297.
Harner SA, Roy PL, Hickling GJ, Walker ED, Foster ES, Barber CC, et al. (2007).
Zoonotic pathogens in Ixodes scapularis, Michigan. Emerging Infectious Diseases
13:1131-1133.
Hamer SA, Tsao JI, Walker ED, Mansfield LS, Foster ES, and Hickling GJ (2009). Use
of tick surveys and serosurveys to evaluate pet dogs as a sentinel species for
emerging Lyme disease. American Journal of Veterinary Research 70:49-56.
Jackson J0, and Defoliart GR (1970). Ixodes-Scapularis Say in Northern Wisconsin.
Journal of Medical Entomology 7: 124-&.
Jones CJ, and Kitron UD (2000). Populations of Ixodes scapularis (Acari : Ixodidae) are
modulated by drought at a Lyme disease focus in Illinois. Journal of Medical
Entomology 37 :408-4 1 5.
Keirans JE, and Clifford CM (1978). The genus Ixodes in the United States: a scanning
electron microscope study and key to the adults. J Med Entomol Suppl 2:1-149.
Kitron U, Jones CJ, and Bouseman JK (1991). Spatial and Temporal Dispersion of
Immature Ixodes dammini on Peromyscus leucopus in Northwestern Illinois. The
Journal of Parasitology 77:945-949.
Lindgren E, and Gustafson R (2001). Tick-borne encephalitis in Sweden and climate
change. Lancet 358: 16-1 8.
Liveris D, Gazumyan A, and Schwartz I (1995). Molecular Typing of Borrelia-
Burgdorferi Sensu-Lato by Per-Restriction Fragment Length Polymorphism
Analysis. Journal of Clinical Microbiology 33:589-595.
Lord R, Humphreys J, Lord V, McLean R, and Garland C (1992). Borrelia burgdorferi
infection in white-footed mice (Peromyscus leucopus) in hemlock (Tsuga
canadensis) habitat in western Pennsylvania. J Wildl Dis 28:364-368.
Lord RD, Lord VR, Humphreys JG, and McLean RG (1994). Distribution of Borrelia
burgdorferi in host mice in Pennsylvania. J. Clin. Microbiol. 32:2501-2504.
Madhav NK, Brownstein J S, Tsao J1, and Fish D (2004). A dispersal model for the range
expansion of blacklegged tick (Acari : Ixodidae). Journal of Medical Entomology
41:842-852.
Magnarelli LA, Anderson J F , and Durland F (1987). Transovarial Transmission of
Borrelia burgdorferi in Ixodes dammini (AcarizIxodidae). The Journal of
Infectious Diseases 156:234-236.
104
Marconi RT, Liveris D, and Schwartz I (1995). Identification of Novel Insertion
Elements, Restriction-Fragment-Length-Polymorphism Patterns, and
Discontinuous 23s Ribosomal-Rna in Lyme-Disease Spirochetes - Phylogenetic
Analyses of Ribosomal-Rna Genes and Their Intergenic Spacers in Borrelia-
Japonica Sp-Nov and Genomic Group-21038 (Borrelia-Andersonii Sp-Nov)
Isolates. Journal of Clinical Microbiology 33:2427-2434.
Maupin GO, Gage KL, Piesman J, Montenieri J, Sviat SL, Vanderzanden L, et al. (1994).
Discovery of an Enzootic Cycle of Borrelia-Burgdorferi in Neotoma-Mexicana
and Ixodes-Spinipalpis from Northern Colorado, an Area Where Lyme-Disease Is
Nonendemic. Journal of Infectious Diseases 170:636-643.
MDNR (2002). Deer Management History in Michigan. Michigan Department of Natural
Resources. Available: http://www. michigan. gov/dnr/O, 1 60 7, 7-153-
10363_10856_1 0905 ~285 43--, 00. html [accessed December I, 2009].
Ogden NH, Lindsay RL, Hanincova K, Barker IK, Bigras-Poulin M, Charron DF, et al.
(2008). Role of migratory birds in introduction and range expansion of Ixodes
scapularis ticks and of Borrelia burgdorferi and Anaplasma phagocytophilum in
Canada (vol 74, pg 1780, 2008). Applied and Environmental Microbiology
74:3919-3919.
Oliver JH, Lin T, Gao L, Clark KL, Banks CW, Durden LA, et al. (2003). An enzootic
transmission cycle of Lyme borreliosis spirochetes in the southeastern United
States. Proceedings of the National Academy of Sciences of the United States of
America 100:11642-11645.
Piesman J, and Sinsky RJ (1988). Ability of Ixodes-Scapularis, Dermacentor-Variabilis,
and Amblyomma-Americanum (Acari, Ixodidae) to Acquire, Maintain, and
Transmit Lyme-Disease Spirochetes (Borrelia-Burgdorferi). Journal of Medical
Entomology 25:336-339.
Pinger RR, and Glancy T (1989). Ixodes-Dammini (Acari, Ixodidae) in Indiana. Journal
of Medical Entomology 26:130-131.
Pinger RR, Timmons L, and Karris K (1996). Spread of Ixodes scapularis (Acari:
Ixodidae) in Indiana: Collections of adults in 1991-1994 and description of a
Borrelia burgdorferi-infected population. Journal of Medical Entomology 33:852-
855.
Randolph SE (2004). Evidence that climate change has caused [‘]emergence' of tick-
borne diseases in Europe? International Journal of Medical Microbialogy
Supplements 293:5 - 1 5.
Ryder J W, Pinger RR, and Glancy T (1992). Inability of Ixodes-Cookei and
Amblyomma-Americanum Nymphs (Acari, Ixodidae) to Transmit Borrelia-
Burgdorferi. Journal of Medical Entomology 29:525-530.
105
Schulze TL, Jordan RA, and Hung RW (1998). Comparison of Ixodes scapularis (Acari:
Ixodidae) populations and their habitats in established and emerging Lyme
disease areas in New Jersey. J Med Entomol 35:64-70.
Schulze TL, Jordan RA, and Hung RW (2001). Effects of selected meteorological factors
on diurnal questing of Ixodes scapularis and Amblyomma americanum (Acari :
Ixodidae). Journal of Medical Entomology 38:3 1 8-3 24.
Schulze TL, Jordan RA, Hung RW, Puelle RS, Markowski D, and Chomsky MS (2003).
Prevalence of Borrelia burgdorferi (Spirochaetales : Spirochaetaceae) in Ixodes
scapularis (Acari : Ixodidae) adults in New Jersey, 2000-2001. Journal of Medical
Entomology 40:555-558.
Sonenshine DE (1979). Ticks of Virginia. Virginia Polytechnic Institute and State
University, College of Agriculture and Life Sciences, Blacksburg, VA.
Sonenshine DE, Ratzlaff RE, Troyer J, Demmerle S, Demmerle ER, Austin WE, et al.
(1995). Borrelia burgdorferi in Eastern Virginia: Comparison between a Coastal
and Inland Locality. Am J T rop Med Hyg 53:123-133.
Stafford KC, 3rd, Cartter ML, Magnarelli LA, Ertel SH, and Mshar PA (1998). Temporal
correlations between tick abundance and prevalence of ticks infected with
Borrelia burgdorferi and increasing incidence of Lyme disease. J Clin Microbiol
36:1240-1244.
Steere AC, Coburn J, and Glickstein L (2004). The emergence of Lyme disease. Journal
of Clinical Investigation 113:1093-1101.
Strand MR, Walker ED, and Merritt RW (1992). Field studies on Ixodes dammini in the
Upper Peninsula of Michigan. Vector Control Bulletin of North Central States 1
1 1-18.
Tamura K, Dudley J, Nei M, and Kumar S (2007). MEGA4: Molecular evolutionary
genetics analysis (MEGA) software version 4.0. Molecular Biology and Evolution
24:1596-1599.
Telford SR, Mather TN, Moore SI, Wilson ML, and Spielman A (1988). Incompetence of
Deer as Reservoirs of the Lyme-Disease Spirochete. American Journal of
Tropical Medicine and Hygiene 39:105-109.
Telford SR, and Spielman A (1989a). Competence of a Rabbit-Feeding Ixodes (Acari,
Ixodidae) as a Vector of the Lyme-Disease Spirochete. Journal of Medical
Entomology 26:1 18-121.
Telford SR, and Spielman A (1989b). Enzootic Transmission of the Agent of Lyme-
Disease in Rabbits. American Journal of T ropical Medicine and Hygiene 41:482-
490.
106
Tsao JI, Wootton JT, Bunikis J, Luna MG, Fish D, and Barbour AG (2004). An
ecological approach to preventing human infection: Vaccinating wild mouse
reservoirs intervenes in the Lyme disease cycle. Proceedings of the National
Academy of Sciences of the United States of America 101:18159-18164.
Walker ED, Stobierski MG, Poplar ML, Smith TW, Murphy AJ, Smith PC, et al. (1998).
Geographic distribution of ticks (Acari : Ixodidae) in Michigan, with emphasis on
Ixodes scapularis and Borrelia burgdorferi. Journal of Medical Entomology
35:872-882. *
Worrnser GaryA P, Liveris D, Nowakowski J, Nadelrnan RobertA B, Cavaliere LA F,
McKenna D, et al. (1999). Association of Specific Subtypes of Borrelia
burgdorferi with Hematogenous Dissemination in Early Lyme Disease. The
Journal of Infectious Diseases 180:720-725.
107
CHAPTER 4
Canines as sentinels for emerging Ixodes scapularis-home zoonoses risk
Hamer SA, Tsao JI, Walker ED, Mansfield LS, Foster ES, Hickling GJ. 2009. Canines as
sentinels for emerging Ixodes scapularis-borne zoonoses risk. American Journal of
Veterinary Research 70:49-5 6.
Abstract
To evaluate companion animal dogs as sentinels for emergence of Ixodes
scapularis and associated canine and zoonotic pathogens in a region undergoing invasion
by these ticks. Dogs brought to clinics located along a gradient of I. scapularis abundance
in Lower Michigan in 2001-2002 (pilot study), and July-August, 2005. Sera were
evaluated for Borrelia burgdorferi—specific antibodies using an indirect fluorescent
antibody test, western blot, and rapid enzyme-linked immunosorbent assay. Ticks from
dogs were subjected to PCR and DNA sequencing for identification of Borrelia and
Babesia species and Anaplasma phagocytophilum. Of 353 canine sera from 18 counties,
only two (0.6%) contained western blot-confirmed B. burgdorferi antibodies from
naturally-acquired infections; this low seroprevalence rate did not differ from the early
invasion study. Seventy-eight ticks of three species were collected Ten of 13 dogs
presenting with I. scapularis were from clinics within or immediately adjacent to the
known tick invasion zone. Six of 18 (33%) I. scapularis and 12 of 60 (20%) non-
competent vectors were positive for B. burgdorferi. No ticks were positive for A.
phagocytophilum, and three tested positive for Babesia spp. Seroprevalence, has
remained very low throughout five years of I. scapularis invasion of the study area, so a
canine serosurvey has been relatively ineffective in tracking early invasion dynamics.
Furthermore, the common practice of canine tick chemoprOphylaxis likely reduces
108
sensitivity of the canine serosurvey. Ticks testing positive for B. burgdorferi were more
common and widely dispersed than were the very small number of seropositive dogs.
We conclude therefore that in areas of low tick density, the use of canine patients as a
source of ticks for species identification and pathogen detection is preferable to canine
serosurvey as a method of surveillance. By retaining ticks from their canine patients for
later identification and pathogen testing, companion animal veterinarians can play an
important role in early detection in areas with increasing incidence of Lyme disease.
Introduction
Ixodes scapularis, the blacklegged tick, is abundant in parts of the northeastern
and upper Midwestern United States, and is expanding geographically into previously-
uninfested areas. Invasion of I. scapularis results in increased risk of disease to canines
and humans from pathogens vectored by this tick, including Borrelia burgdorferi,
Anaplasma phagocytophilum, and Babesia species, the agents of Lyme disease,
granulocytic anaplasmosis, and babesiosis, respectively. Lyme disease is the most
prevalent vector-home disease of humans in the United States and Europe (Steere et a1.
2004) and a common cause of acute arthritis and arthralgia in dogs (Appel et a1. 1993).
One area of recent 1. scapularis invasion is the Lower Peninsula of
Michigan(Foster 2004; Hamer et al. 2007), where I. scapularis presence and incidence of
B. burgdorferi infection in humans, canines, and wildlife have been increasing since the
early 20008. The putative sources of invasion are recently-established populations of I.
scapularis immediately to the south, in northern Indiana(Pinger et al. 1996) and initial
intensive surveys of B. burgdorferi-infected ticks and wildlife in Lower Michigan
109
demonstrate a gradient of I. scapularis density with the highest density in the
southwestern corner of the state. Along Michigan’s Lower Peninsula western lakeshore
in 2007, I. scapularis dragging densities ranged from 30 to 3 to O nymphs/1000 m2 at
sites in the south, central, and north, respectively, with B. burgdorferi found in all areas
where I. scapularis was found; drag sampling over 5600 m2 at inland sites over the same
time period revealed no I. scapularis nymphs (S. Hamer, unpublished data). The infested
zone is of considerable public health significance because the Lake Michigan coast is a
popular tourist site and harbors many state parks and recreation areas. Hence, it would be
useful to quantify and delineate geographically the current disease risk levels in order to
inform public health recommendations in the Lower Peninsula.
Dogs as sentinels for disease emergence. Assessing risk in zones of emerging Lyme
disease, as opposed to in endemic zones, is challenging. Lack of public and medical
awareness often hinders accurate diagnosis. Human case prevalence is initially low, so
the positive predictive value of any diagnostic test is similarly low. Field surveys of
ticks, mice, deer and other vertebrates can reveal trends in Lyme disease risk in these
areas, but such surveys are laborious and costly. Consequently, numerous researchers
have proposed that companion animal dogs are a useful sentinel animal that can assist in
assessing risk in areas endemic for Lyme disease (Falco et al. 1993; Merino et al. 2000;
Guerra et al. 2001; Bhide et al. 2004). Companion dogs typically have a close
association with their human owners, but are more active in habitats where they can come
into close contact with infected ticks. Eng et al. (1988) estimated that pet dogs were on
average six times more likely than their owners to be seropositive to B. burgdorferi, and
110
dog ownership is a significant risk factor for human tick-bome disease (Jones et al.
2002). Most canine serosurveys, however, have been conducted in areas endemic for I.
scapularis where human risk is already appreciated. Exceptions include Rand and
colleagues(Rand et al. 1991; Rand et al. 1996), who found in Maine that seropositive
canines identified areas of hmnan risk in advance of human cases, and Duncan and
colleagues (Duncan et al. 2004), who deemed the canine an apprOpriate sentinel in North
Carolina where canine seroprevalence was <1% and transmission to humans in that state
was rare. Nevertheless, the utility of canine serosurveys in areas where risk is low but
emerging remains unclear.
History of serodiagnosis of Lyme disease in canine patients. Diagnosis of Lyme disease
in canine patients is typically based on serology alone, and as in humans, has been
problematic since the emergence of the disease due to cross-reactivity to spirochetes
including T reponema denticola and Leptospira interrogans and lack of standardization of
laboratory protocols. Recommended practice has therefore been a two-step diagnostic
algorithm for determining infection of dogs with B. burgdorferi, whereby indirect
fluorescent antibody tests (IFA) or enzyme-linked immunosorbent assays (ELISA) is
used to screen sera and western blotting (WB)(Lindenmayer et al. 1990; Gauthier and
Mansfield 1999) is used to confirm suspect-positive samples. Recently, a B. burgdorferi
peptide antigen, the C6 region of the VlsE, has been deployed in a commercially-
available ELISA kit that provides sensitivity and specificity equaling that of the western
blot (Levy et al. 2003).
111
Canids as sentinels/or Lyme disease risk in Lower Michigan. In 1992-1993, prior to the
current I. scapularis invasion of southwestern Michigan, a canine serosurvey was
undertaken in six Michigan counties, including Menominee County in the Upper
Peninsula, the only Michigan county endemic for Lyme disease at that time (Walker et al.
1998). Using ELISA and IFA, 25 of 299 (8.4%) canids from Menominee County were
seropositive by one or both screening methods, whereas only one of 919 (0.1%) canids
from the Lower Peninsula tested positive by IF A (and was negative by ELISA). This
serosurvey thus establishes the almost complete absence of B. burgdorferi-exposed
canids in the Lower Peninsula in the early 1990s.
In response to initial findings of invading ticks on wildlife (Foster 2004), we
initiated a preliminary survey of canine exposure to Lyme disease ticks and pathogen in
the presumed focus of invasion of southwestern Michigan. In 2005, we undertook a new
serosurvey and tick survey of canine patients at veterinary clinics in Lower Michigan to
define the zones of disease risk in an area undergoing tick invasion beyond the focus.
Furthermore, we compare the results given the conventional two-tiered approach versus
the new C6 ELISA.
In designing the study, we hypothesized that B. burgdorferi seropositive canids
would be found throughout the known geographic distribution of I. scapularis, and
similarly, that I. scapularis-borne pathogens would be restricted to that same geographic
range. To test these hypotheses, our objectives were four-fold: i) to determine rates of
canid exposure to B. burgdorferi in an area undergoing active invasion by I. I scapularis;
ii) to determine rates of canid infestation by I. scapularis and other ticks; iii) to
determine the prevalence of B. burgdorferi, A. phagocytophilum and Babesia species in
112
ticks from canine patients, and iv) to assess patterns of seroprevalence, vaccination,
I. scapularis presence and tick infection in relation to the zone of I. scapularis invasion in
southwestern MI. Specifically, we predicted that all these measures would be highest at
clinics in the southwestern corner of the state — the putative site of earliest invading ticks
- and would decline with increasing distance from this focus.
Materials and Methods
Sample Acquisition. In spring 2005, 18 veterinary clinics consented to participate in our
canine tick and serosurvey; these clinics were selected along three transects that
originated in the recently-invaded southwestern corner of Michigan and extended radially
to the north, northeast, and east. These transects had been established previously for
ongoing wildlife and tick sampling, and traversed the known geographic boundary of the
invading ticks (S. Hamer unpublished data). An initial 133 clinics operating within the
area of interest were identified from online directories and geo-coded using ArcGIS 8.0
(ESRI, Redlands, CA). Six clinics distributed along each transect were then randomly
selected and contacted by phone to solicit participation in the study. For analysis, the
clinics were grouped into three zones based on their proximity to the recently-established
I. scapularis populations: Zone 1 included all clinics in counties where I. scapularis was
known to be established (n = 6; establishment defined as documented presence of all
three life stages of I. scapularis on drag cloth, wildlife, and/or humans); Zone 2 included
all clinics in counties that bordered Zone 1 (n = 4). Zone 3 included all clinics in
counties outside of Zones 1 and 2 (n = 8; Figure 4.1).
113
Clinics were asked to retain 30 serum samples from dogs acquired in the course of
routine blood-sampling from July 15 — August 15. We requested this number based on
the calculation that we would have a >95% chance of detecting at least one seropositive
dog, assuming a 50% response rate by the clinics and an invasion area prevalence of 1%
(cf. average canine seroprevalence for endemic sites in the northeastern US. is 30%
(Guerra et al. 2001). In recently-invaded southwest Michigan, I. scapularis adult and
nymphal host-seeking peaks in April and June, respectively (Foster 2004); the collection
period chosen was the time of year when dogs would have likely seroconverted (i.e., >4-6
weeks after exposure (Appel et al. 1993) from infection by either nymphs or adults.
Clinicians were instructed to store sera at -20 ° C until pick-up.
The clinic veterinarians were also asked to collect all ticks seen on canine patients
during the same July-August period. Canines presenting ticks were not necessarily the
same canines from which blood was collected. All ticks from each dog were stored in
vials of 70% ethanol, and the veterinarian completed a short questionnaire to indicate the
dog’s breed, sex, age, zip code, travel history, Lyme disease vaccination and diagnosis
status. After pick-up from clinics, sera were stored at —80 ° C before processing at the
Michigan State University Medical Entomology Laboratory.
Indirect Fluorescent Antibody (IFA). Sera were thawed and centrifuged for 10 min at
13,000 rpm, then serially diluted with sterile 1X phosphate buffered saline (PBS) to
1:320 and 1:640. Both dilutions were screened for antibody presence using I12-well IFA
slides fixed with whole cell B. burgdorferi organisms (low passage mixture of isolates
B31 and 297; Fuller Laboratories, Fullerton, CA). Slides were warmed to room
114
temperature and each well was loaded with 20 ul of sample or control. One positive and
one negative control (V MRD, Pullman, WA) were included on each slide. Slides were
incubated in a humidifying chamber for 30 min at 37 ° F. Samples were washed from
slides by rinsing with PBS three times after which excess PBS was tapped off onto a
paper towel. While wells were still moist, 20 ul of stained fluorescein isothiocyanate
(F ITC)-conjugated anti-dog IgG antibody was added to the wells. This conjugate was
prepared using sterile 0.1% Evans Blue solution in PBS to make a 1:100 dilution FITC-
conjugated anti-dog IgG (Kirkegaard and Perry Laboratories, Gaithersburg, MD). The
conjugated secondary antibody was incubated on the slide for 30 min at 37 ° F in the
humidifying chamber and washed as above.
For visualization, a cover slip was placed onto the moist slide, which was then
viewed under UV illumination at 4OOX (10X ocular, 40X objective) using fluorescent
microscopy (Eclipse E800, Nikon, Melville, New York). All slides were read by one
observer (SAH) who was blinded to the identity of the slides. Roughly 15 see were spent
scanning each of 10 fields within each well, with both the intensity and abundance of
fluorescing spirochetes compared to the negative control. Wells with well-resolved and
consistently bright yellow-green spirochetes were considered positive for anti-
B. burgdorferi antibodies at the tested dilution. Samples with no fluorescing spirochetes
at 1:640 were considered negative. All samples positive at 1:640 were tested in a series
of two-fold dilutions until an endpoint titer was found. Samples positive at or above
1:640 dilutions were considered suspect-positive and were subjected to confirmatory
testing using western blot.
115
Western blot (WB). Western blot analysis was used as a confirmatory test for all IFA
suspect-positive sera, all sera derived from canines that were reported as having been
vaccinated regardless of endpoint titer, and a random sample of the IFA-negative sera.
Western blots were performed using the QualiCode Canine Lyme Disease Kit
(Irnmunetics, Cambridge, MA); blots were prepared by the manufacturer and contained
18 separated antigens of B. burgdorferi strain B31. We followed kit instructions and
interpretive criteria were as follows: 2 or more bands in the p30 - p14 region without
both p31 (Outer surface protein (Osp) A) and p34 (Osp B) indicate a natural positive; 2 or
more bands in the p3 0 —— p14 region plus both p31 and p34 were indicative of vaccination;
and less than 2 bands in the p3 0 -~ pl 4 region were considered negative for the presence
of B. burgdorferi-specific antibodies. Overall seroprevalence was computed as: number
of WB-positive samples / (total number of samples assayed by IF A — number of WB-
confinned vaccine-positive samples).
C6 Enzyme-linked Immunosorbent Assay (ELISA). In addition to the two-step
diagnostic procedure outlined above, we subjected selected sera to a newly-available
commercial ELISA-based canine Lyme kit (SNAP 3Dx test, Idexx Laboratories,
Westbrook, ME) to evaluate its utility in an area of invading I. scapularis and to compare
it to the more laborious two-step process. The kit employs a synthetic peptide (C6) with
sequence homology to one of six invariable regions within the variable domain of VIsE
(V mp-like sequence, Expressed), a surface lipoprotein of B. burgdorferi. Antibodies to
VlsE are found in field-infected dogs, but not in vaccinated dogs (Levy et al. 2003). In
addition to assaying for the presence of antibodies to B. burgdorferi, this test
116
simultaneously assays for antibodies to Ehrlichia canis and antigen of Dirofilaria immitis
adult worms. Manufacturer’s instructions were followed to assay all sera with IF A
endpoint titers 2 1:2560 (n = 22) plus seven randomly-selected samples with lower
endpoint titers.
Tick Processing and Polymerase Chain Reactions (PCR). All ticks removed from dogs
were identified to life stage and species using standard taxonomic keys (Sonenshine
1979; Durden and Keirans 1996). Total DNA was extracted from each tick using the
animal tissue protocol of the DNeasy Tissue Prep extraction kit (Qiagen, Valencia, CA).
Known-infected nymphs from a CDC laboratory colony served as positive controls and
no-template wells served as negative controls for the extraction. Ticks were prepared for
extraction by slicing through the tick exoskeleton and midgut using a scalpel in a dry
microcentrifuge tube. Lysis buffers were added and the solution incubated overnight. In
the case that an engorged tick clogged the spin column, a sterile pipette tip was used to
dislodge the clog. DNA was eluted in a single 100111 elution. Five ul of DNA was used
in each of three separate PCRs of 50ul volumes to assay for B. burgdorferi, A.
phagocytophilum, and Babesia species. Gel electrophoresis was carried out using 10 ul
of PCR product in a pre-cast 4% agarose gel (E-gel system, Invitrogen, Carlsbad, CA).
The rrs—rrlA (16S - 238 rRNA) intergenic spacer (IGS) of B. burgdorferi was amplified
in a nested assay producing a 978 bp fragment (Bunikis et al. 2004). DNA extracted
from cultured spirochetes served as a positive control (spirochetes were cultured from
field-collected adult ticks from a different study in BSK-H complete media (Sigma-
Aldrich, St. Louis, MO) and incubated at 37° C). The p44 gene of A. phagocytophilum
117
(Zeidner et al. 2000) was amplified using a touchdown PCR program described in Steiner
et al. (2006) to produce a334-bp fragment. Babesia genus-specific PCR was performed
using primers for the 18S rRNA gene to produce a fragment of variable size, including a
408bp fragment for B. microti or a 437-bp fragment for B. odocoilei (Armstrong et al.
1998). Commercially-available B. microti organism (ATCC, Manassas, VA) was
extracted as above and used as a positive control. Ticks were removed from dogs at
different levels of engorgement, so their extracted DNA varied in concentration. PCR
optimization trials using serially-diluted templates suggested that overly-concentrated
template (from engorged ticks) inhibited detection of pathogen (data not shown), so we
tested all DNA samples at both full-strength and at a 1:10 dilution in the non-nested
PCRs. In all PCRs, negative controls consisted of wells in which all reagents except a
DNA template were added.
Sequence Analysis. We undertook DNA sequence analysis of PCR amplicons of B.
burgdorferi (to assess strain-level variation) and Babesia spp. (to identify to the species
level). The 40 ul of amplicon that remained afier electrophoresis was purified (PCR
Purification Kit, Qiagen, Valencia, CA) and used as a template for DNA sequencing at
the Research Technology Support Facility at Michigan State University on an ABIPrism
3100 Genetic Analyzer (Applied Biosystems, Foster City, CA). Sequences were
compared to those published in the National Center for Biotechnology Information
sequence database using the nucleotide-nucleotide basic local alignment search tool
(BLASTn) to identify the pathogen strains with the greatest sequence homology.
118
Spatial Analyses. Generalized linear models were used to assess the relationship
between tick/pathogen measures (i.e., seroprevalence, vaccination rate, IF A background
reactivity (IF A suspect-positives that tested negative on WB), presence of I. scapularis
on dogs, and presence of B. burgdorferi in ticks removed from dogs) and the geographic
zone of each clinic. All statistical analyses were performed using R-project (Team 2005).
Early Invasion Survey. We include data from a pilot serosurvey that we undertook in
2001-2002, which used similar methodology to collect samples from 17 clinics dispersed
throughout Zone 1. This survey was initiated in response to the first direct field
evidence, from drag and wildlife sampling, of invading I. scapularis in southwestern
Michigan (Foster 2004). Briefly, veterinarians were asked to store surplus blood samples
and place ticks from dogs in one common vial from June of 2001 through December of
2002. Sera were processed as above, except that suspect positives were defined as having
an IFA cutoff titer of 2: 1:320 in 2001 and 2 1:640 in 2002 on slides read by a single
observer (ESF). A subset of suspect positives was subjected to WB. Serology and tick
collection results are presented as a means for comparison with the 2005 survey data.
Results
In 2005, a total of 353 serum samples were provided by the enrolled veterinary
clinics (mean of 20 per clinic; range 3 -30) of which 174 (49.3%) were from female dogs
and 179 (50.7%) were from male dogs. All American Kennel Club breed groups were
represented (Figure 4.2). Twenty-one canids (5.9%) were reported as vaccinated, 113
(32.1%) were reported as not vaccinated, and 219 (62%) were of unknown vaccination
119
status. None were reported to have previously been diagnosed with Lyme disease nor to
have traveled to any other Lyme disease endemic area, although 232 (66%) reported
unknown travel history or did not answer this question.
IFA. Of 353 serum samples, 79 (22.4%) had endpoint titers _>_ 1:640 and thus were
considered suspect-positive for B. burgdorferi exposure (Figure 4.3). The maximum
endpoint titer detected was 1:81920 (Table 4.1).
WB. Western blot analysis was performed on 105 samples (79 suspect-positives, 3
reported vaccinated with IFA endpoint titers <1 :640, and 23 randomly-selected sera with
IFA endpoint titers < 1:640). Of the suspect-positives, 2 (2.5%) were WB positive due to
natural exposure; 19 (24.0%) were vaccine-positive (including 3 from canines reported as
not vaccinated and 4 from canines of unknown vaccine history), and 58 (73.4%) were
negative. The highest IF A endpoint titer in WB-negative samples was 1:5120. Two of
the low IF A-titer, reportedly-vaccinated dogs were confirmed as vaccine-positive
whereas the other was negative. WB-confirmed vaccine-positive sera had IFA endpoint
titers ranging from 1:320 — 1:20480 (median: 1:2560; Figure 4.3). All 23 of the
randomly-selected IFA-negative sera tested WB-negative.
The overall natural seroprevalence detected in the study was 2 of 332 (0.6%;
denominator discounted by number of vaccine-positives). The two natural-positive
samples had IF A endpoint titers of 1281920 and 1:10240 and were submitted from clinics
N and M, respectively (both within the I. scapularis invasion zone; Figure 4.1). The
overall WB-confirmed vaccine rate was 21 of 353 (5.9%; Table 4.1).
120
C6 ELISA. Twenty-nine sera were assayed for B. burgdorferi and E. canis antibodies
and D. immitis antigen using the SNAP 3Dx kit. Of the 22 samples with IF A endpoint
titers 2 1:2560, two tested positive for B. burgdorferi antibodies (these being the natural-
positives previously identified by WB). Of the 7 randomly-selected lower-titer sera, none
tested positive for B. burgdorferi antibodies. The Snap test was thus 100% sensitive and
100% specific when compared to our WB. No samples tested positive for the presence of
E. canis antibodies or D. immitis antigen.
Tick PCR. A total of 78 ticks from 55 individual dogs were submitted by the clinics
(mean = 4.9 per clinic; range 0-14). Three dogs had reported travel to I. scapularis-
endemic areas (two to Michigan’s UP and one to the northeastern US) and 28 did not
have responses to this question. Ticks comprised nymphs and adults of three species: I.
scapularis, I. cookei, and D. variabilis. A total of 18 I. scapularis were removed from 13
individual dogs (including one dog from clinic B which was reported to have traveled to
an endemic area; all other dogs with I. scapularis had no known travel to endemic areas).
Of the dogs with I. scapularis, 1 was from Zone 1 (clinic E), 9 were from Zone 2 (clinic
B), and 3 were from Zone 3 (clinics H and R; Figure 4.4). Fifty-two and eight D.
variabilis and I. cookei were removed from 35 and six dogs, respectively. Of the 78 ticks
submitted, 18 ticks removed from 15 dogs tested positive for B. burgdorferi including
ticks of all three species (Table 4.2). The prevalence of B. burgdorferi infection in I.
scapularis was 33% (6 of 18; nymphs and adults combined). The prevalence of
B. burgdorferi infection in D. variabilis was 17% (9 of 52; nymphs and adults
combined). The prevalence of B. burgdorferi infection in I. cookei was 38% (3 of 8;
121
nymphs and adults combined). Multiple infected ticks were removed from the same
individual dogs (three infected D. variabilis from one dog at clinic O, and two infected I.
scapularis from one dog at clinic N; both clinics are in Zone 1). B. burgdorferi-positive
ticks were identified at 9 of the 18 clinics (Figure 4.4). Results of positive and negative
extraction and PCR controls were as expected.
Three of 78 ticks (3.9%) were infected with Babesia spp; these comprised one
adult D. variabilis from clinic L and two nymphal I. cookei from clinics I and N (Figure
4.4). One Babesia-positive nymphal I. cookei from clinic N was co-infected with B.
burgdorferi. No ticks tested positive for A. phagocytophilum.
Sequence Analyses. All PCR amplicons for B. burgdorferi and Babesia spp.-positive
ticks were purified and sequenced. Sequences were obtained for 13 B. burgdorferi
amplicons with an average trimmed sequence length of 790 bases. These sequences were
found to have 2 98% sequence identity with previously reported B. burgdorferi l6S —
23S rDNA gene sequences of the following strains: B515 (11 = 4; accession number
AF467860.1); CS2 (11 = 2; accession number DQ437492.1); B356 (n = 2; accession
number AF467861.1); Lenz (n = 1; accession number EF537391.1); RSP2 (n = 1;
accession number EF649782.1); M1415 (n = 1; accession number EF537369.1); HBl (n
= 1; accession number EF537294.1); and 1P3 (n = 2; accession number EF537392.1).
Sequences were obtained for 2 Babesia amplicons (both from I. cookei nymphs) with an
average trimmed sequence length of 316 bases. The sequence from the nymph from
clinic N had 93% sequence identity with Babesia odocoilei Wisconsin 1 (accession
number AY237638.1), whereas the sequence from the nymph from clinic I had 91%
122
sequence identity with Babesia sp. BiCM002 (accession number A8053216.2). These
Babesia matches provide an indication of species similarity but are insufficient for
species identification.
Spatial Analyses. Canine vaccination rates were significantly higher in Zone I (I.
scapularis established) than in the other two zones (average rates for Zones 1, 2, 3 were
14.0, 1.7, 1.9%, respectively; P = 0.008). Seroprevalence, background IF A activity, 1.
scapularis presence and B. burgdorferi presence were all unrelated to clinic zone (P =
0.169, 0.728, 0.73 8, 0.616; respectively). The only clinics where multiple B. burgdorferi
positive ticks were found were those closest to the ‘endemic’ southwestern corner of the
state (i.e., clinics O and N) and just north of the invasion zone (clinic B). The only co-
infected tick (carrying both B. burgdorferi and Ba. odocoilei) was from the southwestern-
most clinic, N.
Early Invasion Survey. In 2001-02 a total of 2030 sera from 17 Zone 1 clinics were
screened by IFA, with 235 samples identified as suspect-positive. WB was performed on
20 suspects (selected randomly), resulting in 2, 12, and 6 samples classified as positive
due to natural exposure, vaccine positive, and negative; respectively. Extrapolating from
this albeit very small sample size of western blotted samples, these data suggest overall
seropositivity, vaccination, and background IF A activity rates of 1.2, 6.9, and 30%,
respectively. A total of 345 ticks were removed from dogs of which 26 (7.5%) were I.
scapularis; this species was recovered from 8 clinics and only adults were found. Ticks
of other species included: D. variabilis (n = 298; 86.4%); Rhipicephalus sanguineus (n =
123
7; 2.0%); I. cookei (n = 11; 3.2%) and Amblyomma americanum (n = 3; 0.9%). No
pathogen testing of these ticks was undertaken.
Discussion
Numerous researchers have proposed that companion animal dogs are a useful
sentinel animal that can assist in assessing risk in areas endemic for Lyme disease (F alco
et al. 1993; Merino et al. 2000; Guerra et al. 2001; Bhide et al. 2004). Most such
serosurveys, however, have been conducted in areas already endemic for I. scapularis.
Here we consider the efficacy of such surveys in areas where there is concern about low-
prevalence emerging disease.
Despite evidence from tick and other wildlife surveys that B. burgdorferi is
emerging in lower Michigan (Foster 2004), B. burgdorferi seroprevalence was extremely
low (2 in 334; 0.6%) in canine patients from lower Michigan. The two naturally-exposed
dogs detected in this study were from clinics within the I. scapularis invasion zone,
where local exposure can be expected. Canine seroprevalence in this zone during the
early invasion study in 2001-2002 was 1.2%, and neither the 2005 nor the 2001-2002
seropositivity rates reported herein are substantially greater than those found during the
1992-1993 Lower Peninsula serosurvey (0.1%; Walker et al. 1998). This leads us to
conclude that canine serosurvey is an insensitive approach to detection of the invasion of
I. scapularis — B. burgdorferi that has occurred over the timeframe bounded by these
studies (as detected by the wildlife sampling and vegetation dragging that began during
our 2001-02 pilot study).
124
In previous studies, canine seroprevalence has typically been reported to be 30-
50% in B. burgdorferi-endemic areas, and 3-5% in areas such as Alabama where Lyme
disease is considered largely absent (see (Bhide et al. 2004) for a review). Our low
seroprevalence estimate is explained, in part, by many of the clinics operating in the outer
zones where I. scapularis is apparently not yet established (Figure 4.1). Second,
seroconversion may not yet have occurred for all exposed dogs. In laboratory studies,
dogs exposed to infected adult ticks develop detectable antibodies 4-6 weeks after
exposure; titers increase for an additional 6-8 weeks and remain high for at least one year.
In contrast, 50% of dogs exposed to infected nymphs fail to seroconvert or convert only
after repeated exposure (Appel et al. 1993); therefore, some dogs infected in the summer
of 2005 may not have seroconverted by the July-August blood-sampling. Third, flea and
tick chemoprophylaxis, which is increasingly being recommended by veterinarians in
southwestern Michigan (see below), may be reducing levels of infection and subsequent
seroconversion.
In our data, background IFA activity was high, with 73.4% of IF A suspect-
positives (positive at 1:640 or higher dilution) not confirmed as natural or vaccine-
positive by either WB or C6 ELISA. The geographical distribution of these IF A high-
titer sera was unrelated to the location of the Lyme disease hotspot, so these IFA ‘false
positives’ may represent cross-reactivity with heterologous antibodies or subjectivity of
IFA slide interpretation. This high background reactivity even on negative samples
emphasizes the importance of using confirmatory testing such as the WB or rapid ELISA
kits as we have done in the current study. Given their widespread use of the Borrelia IF A
alone, considerable caution should be exercised in interpreting previous canine
125
serosurveys. We found the C6 rapid ELISA kit to be 100% sensitive and 100% specific
when compared to WB, confirming the value of this test for canine disease sentinel
studies (see also (Levy et al. 2003; Duncan et al. 2004; Stone et al. 2005).
Canine Tick Survey as an Alternative to Canine Serosurvey. The majority of the 13
dogs that harbored I. scapularis during our study period were from clinics within the
zone of established 1. scapularis (Zone 1; 1 dog) and just north of this zone (Zone 2; 9
dogs), demonstrating that the dog, in harboring I. scapularis, serves as a useful sentinel
for presence of this tick in areas undergoing invasion. Contrary to other studies
(Hinrichsen et al. 2001), however, the geographic distribution of I. scapularis was not
predictive of canine seropositivity. Six of 18 I. scapularis — from six different dogs —
plus 12 of 60 ticks of other species — from nine different dogs- were found to be infected
with B. burgdorferi. Thus it was much easier to find B. burgdorferi in ticks off the dogs
than to find antibodies to B. burgdorferi in the dog’s sera.
Of the three species of ticks that clinics submitted from canine patients, only
1. scapularis is regarded as a competent vector for B. burgdorferi and Babesia spp. Yet
in addition to the I. scapularis found infected with B. burgdorferi, we found 12 B.
burgdorferi PCR-positive and three Babesia sp. PCR-positive non-vector ticks, including
one co-infected I. cookei removed from a dog in the southwestern-most site in the study.
We posit that a bloodfed non-vector tick could become infected via three means (not
mutually exclusive). First, the non-vector tick may have been sampled during or soon
after feeding on a spirochetemic dog, so that PCR detects the spirochetemic bloodmeal
within the engorged tick. In our study, 39 individual dogs were parasitized by D.
126
variabilis or I. cookei, and B. burgdorferi-infected ticks of these species (n = 12) were
removed from nine individual dogs. That two individual dogs produced more than one
infected tick lends support to the spirochetemic dog scenario; however, no dogs in our
study were reported to have been diagnosed with Lyme disease (74.5% veterinarian
response rate) and canine seroprevalence was negligible. Second, we posit that
transstadial passage of pathogen acquired in a previous infectious bloodmeal (most likely
from a non-canine host) may result in an infected non-engorged tick, though the infected
non-vector is unable to transmit the pathogen. Walker et al.(Walker et al. 1994) cultured
B. burgdorferi from five questing D. variabilis, and found a 1.3% infection rate in this
species. Similarly, during our ongoing study of Borrelia infection in questing adult ticks
in the Midwestern states we found 6 of 59 (10.2%) D. variabilis to be PCR-positive for
B. burgdorferi (S. Hamer, unpublished data). Studies in the Lyme disease-endemic area
of Michigan's Upper Peninsula found six of 75 (8.0%) questing adult D. variabilis to be
infected, albeit with a spirochete load much lower than that of infected I. scapularis
collected simultaneously (P. Roy, pers. comm). These reports of low-moderate
prevalence, with very low spirochete loads, are consistent with transmission studies
showing D. variabilis to be an incompetent vector for B. burgdorferi (Piesman and
Sinsky 1988). Third, we note that non-vector ticks potentially could acquire spirochetes
by co-feeding with vector ticks, even in the absence of systemic infection of the canine
host. This mechanism, -- proposed for the maintenance of tick-borne encephalitis virus
in Europe(Randolph et al. 1996) -- has not been tested for non-vector ticks and B.
burgdorferi, and has only been demonstrated for I. scapularis under laboratory conditions
with artificially high tick burdens(Piesman and Happ 2001). Nevertheless, while the
127
mechanism remains unclear, our findings do suggest that assays of both the vector and
non-vector ticks found attached to canine patients can contribute to detection of pathogen
presence in a given area - irrespective of the tick species’ inability to transmit the
pathogen of interest.
Relative to the low apparent exposure of canids to B. burgdorferi based on our
serological findings, ticks testing positive for this pathogen were far more common than
expected. We hypothesize a posteriori that our tick and serum samples came from
different subpopulations of pet dogs: i.e., that serum samples came from a range of dogs
that included those prophylactically vaccinated and/or chemically protected against tick
bites; whereas tick samples mainly came from dogs not chemically protected from tick
exposure. In 2008, we conducted a short phone survey of veterinarians and licensed
veterinary technicians at nine of the participating veterinary clinics (3, 2, and 4 in Zones
1, 2, and 3; respectively) to assess whether anti-tick and anti-tick-borne pathogen
measures are recommended at the clinics. In responding, the clinicians were asked to
recall as best they could the situation in summer of 2005, when our sampling was
underway. Clinicians estimated that 20 — 65% (average 50%) of the pet dogs attending
their clinic were being actively protected against ticks during those summer months -
with topical agents being the most common form of prophylaxis - and that 5-25%
(average 12%) of dogs were being vaccinated against Lyme disease. These data suggest
that acquisition of serum from pet dogs randomly selected at veterinary clinics (which is
a common method of sampling in canine serosurveys; see current study, (Rand et al.
1996; Guerra et al. 2001; Hinrichsen et al. 2001)) will under sample the unprotected dogs
that are likely the most effective sentinels for pathogen transmission in an area.
128
Since this serosurvey, ongoing surveillance (through wildlife trapping, drag sampling and
human submissions) has documented continuing range expansion of I. scapularis in
southwestern Michigan (Hamer unpublished). We anticipate that I. scapularis and
associated pathogens will spread north and east from the present southwestern Michigan
hotspot in coming years, so that both veterinary and human clinicians will see an increase
in Lyme cases among their patients. Dogs are unlikely to be a sensitive sero-sentinel for
changes in tick and pathogen activity in these areas of low tick density. We conclude that
I. scapularis emergence, and consequent increase in canine and human disease risk, can
be predicted more effectively by surveys of I. scapularis and other non—vector ticks
removed from dogs than by serosurvey of the canines.
129
Figure 4.1. Locations of the 18 veterinary clinics (labeled A - R) that participated in the
2005 canine tick and serosurvey of Lower Michigan. Shading indicates counties in
which I. scapularis has recently invaded (Lower Peninsula) or is endemic (Upper
Peninsula); shaded counties are those within which all three life stages of I. scapularis
have been documented (Foster 2004). Circles indicate the estimated Lyme disease
vaccination rates (expressed as a proportion) for the dogs at each clinic. The locations of
the two dogs seropositive for antibodies to B. burgdorferi from natural exposure are
shown.
N
|:| County Boundary
"i Established I. scapularis
+ B. burgdorferi seropositive dog &
Vaccination Rate C
o 0
e >o.oo — 0.03 .3 .c ,D
e >0.03 — 0.04 ,,
fli‘ >0.04 - 0.07 l .H .I
>0.07—o.17 9F
‘ E G
>o.17 — 0.25 K,L .R
NMO .p .Q 1
130
Figure 4.2. Proportions of the serum samples collected (n = 353) from individuals
of the various American Kennel Club breed groups. Dogs reported as a specified
breed mix are classified under that breed group (i.e. German Shepard mix =
herding group) whereas reports of mixed breed with no breed specification are
listed as ‘Mix’.
10% 14%
El Herding
6% I Hound
a Mix
.E:?:§:EIE:£:?tEt:I ----- Eli-1 Non-Sporting
12% 323253335333333233333332323232553-1; :::: 8% DSporting
:}:{:}:}:§if:f:313;3' """" E] Terrier
4% I Toy
l Unknown
IIII Working
38%
131
Figure 4.3. Frequency distribution of IFA endpoint titers for B. burgdorferi antibody
detection in canine sera (n = 353). Sera with titers S 1:320 were classified as negative for
infection with B. burgdorferi. The remaining high-titer sera, plus sera from dogs reported
as having been vaccinated, were classified as natural positive, vaccine positive, or
negative based on western blot.
507
45 - El Vaccine Positive
.3 4o - 1 I Natural Positive
€35- EINegative
”304
E
3254
£20-
915i 3 :1
o
2104
:1
5“ I]
o ._ U L_,,-0_L flj_-__-_
O O O O O O O O O O
v ‘- .. z: :2 :2 2 a 8 a
IFA endpoint titer
132
Figure 4.4. Distribution of clinics with canines harboring I. scapularis, B. burgdorferi-
infected ticks of any species, and Babesia-infected ticks of any species. The shading
indicates counties in which I. scapularis has recently invaded and is now established with
documented presence of all three life stages (Foster 2004).
[:| County Boundary
u i Established I. scapularis
' Clinic (no I. scapularis; no infection)
I. scapularis on dog(s)
B. burgdorferi in tick(s) from dogs
0 Babesia sp. in tick(s) from dogs
133
.00. 0 .00. .0
N 0 p .3 N nu mm as 000 00 we» .80...
0 0 0 0 F 0 0 F F 08 v 0 m
0 .000 F 0 F 0 0 0 0 0 00000 00 0 0
0 0 0 0 0 0 F F 0 000 00 0 d
0 .000 F 0 P 0 0 v v 0 000. 00 0 o
.000 _. 9.0: 0 0 0 0 0 0 0 0 0020 00 . 0 z
.0000 F 0 F 0 0 0 0 0 F 0000. 0 F s.
0 0 0 0 0 0 F 0 0 000. 0. 0 ..
0 .0000 P 0 0 0 0 0 F 0 00.0 0 F v.
0 .0000 0 0 0 0 0 0 0 0. 0000. 00 F ..
0 0 0 0 F 0 F F P 0000 0. 0 _
0 0 0 0 F 0 . F F 00.0 0. 0 I
0 :00 0 0 F 0 0 P 0 0 0.00 E 0 0
0 E00 0 0 0 v 0 F F 0 0000 00 0 ...
0 .00: 0 0 0 0. 0 0 0 0. 0030 v0 0 m.
0 0 0 0 0 0 P P 0 0.0 0. 0 n.
0 0 0 0 0 0 0 0 0 000. 0. 0 o
0 0 0 0 v o 0 0 v 00.0 00 0 m
0 .000 F 0 F 0 0 F F v 000. 00 0 <
3 80 m; 08> m3 ac: m? 08> 95 we: 95 m! 05.80 x0... .8:
._ 08> .80 E. .80 E. .80 E. .08 E. .02. E. Show“... .0880 00000 80””: 800 0.5.0
33 .3532 8.38.. m; 0...... .0: 8.38m <0.
5000.. 0000 055000 05000> u 08> 5:000 0000. 025000
. 0050530000 2 05.0 .00 b55080 8 080—0.. 5085.000 0
5 0055000 05000.» .00 00a .
9 000.300 0003 A0
u 000 5000.0 0000 0.503: n w0Z 0.2500000 N 00 005003000
aoN .95 .3 055000 0000 20500 05009: <0: :0 5.0 0000 0:
8:0 I 00.050 .0005 00 080550000 w50= 0050500 0000—0000800 =000>O .mB 008085.500
>30 00000505 00555 00>:0w0: <0: 00003000500000 050 00250000000000 <00 5.. .0050
finch“; .00 00050 0.500000 553 00.0500 0005 0.0 0055009 0000000 <0: 0.0085000 00 0000000000 000 0.0050 0500000
50:2 0026. 5 055.0 m. 505 00.08000 5200 0.5.00 00 000000 .05w5800 m3 000 <0: .00 050.00% . _ .0 030...
134
Table 4.2. Tick species and life stage-specific results of PCR testing for B. burgdorferi,
A. phagocytophilum, and Babesia species in ticks removed from canine patients at clinics
throughout lower Michigan, 2005.
Number PCR positive (%)
Tick Species Life Stage Number Tested B. burgdorferi A. phagocytophilum Babesia spp.’
D. variabilis N’fi‘; 51‘ 9 (107-5) 3 1 (%)-0)
" °°°"°‘ N333; g :12 85°33 3 2 (303.3)
.000. $ng 126 58515? g g
* One I. cookei nymph was infected with Ba. odocoilei; identities of the remaining two
sequences remain general to Babesia genus.
135
References
Appel MJG, Allan S, Jacobson RH, Lauderdale TL, Chang YF, Shin SJ, et al. (1993).
Experimental Lyme-Disease in Dogs Produces Arthritis and Persistent Infection.
Journal of Infectious Diseases 167:651-664.
Armstrong PM, Katavolos P, Caporale DA, Smith RP, Spielman A, and Telford SR
(1998). Diversity of Babesia infecting deer ticks (Ixodes dammini). American
Journal of Tropical Medicine and Hygiene 58:73 9-742.
Bhide M, Travnicek M, Curlik J, and Stefancikova A (2004). The importance of dogs in
eco-epidemiology of Lyme borreliosis: a review. Veterinarni Medicina 49: 135-
I42.
Bunikis J, Garpmo U, Tsao J, Berglund J, Fish D, and Barbour AG (2004). Sequence
typing reveals extensive strain diversity of the Lyme borreliosis agents Borrelia
burgdorferi in North America and Borrelia afzelii in Europe. Microbiology-ng
150:1741-1755.
Duncan AW, Correa MT, Levine JF, and Breitschwerdt EB (2004). The dog as a sentinel
for human infection: Prevalence of Borrelia burgdorferi C6 antibodies in dogs
from southeastern and mid-Atlantic states. Vector-Borne and Zoonotic Diseases
4:221-229.
Durden LA, and Keirans J E (1996). Nymphs of the genus Ixodes (Acari: Ixodidae) of the
United States: taxonomy, identification key, distribution, hosts, and
medical/veterinary importance. Thomas Say Publications in Entomology,
Entomological Society of America.
Eng TR, Wilson ML, Spielman A, and Lastavica CC (1988). Greater Risk of Borrelia-
Burgdorferi Infection in Dogs Than in People. Journal of Infectious Diseases
158:1410-1411.
Falco RC, Smith HA, Fish D, Mojica BA, Bellinger MA, Harris HL, et al. (1993). The
Distribution of Canine Exposure to Borrelia-Burgdorferi in a Lyme-Disease
Endemic Area. American Journal of Public Health 83:1305-1310.
Foster E (2004). Ixodes scapularis (Acari: Ixodidae) and Borrelia burgdorferi in
Southwest Michigan: Population Ecology and Verification of a Geographic Risk
Model Thesis. Michigan State University, East Lansing.
Gauthier DT, and Mansfield LS (1999). Western immunoblot analysis for. distinguishing
vaccination and infection status with Borrelia burgdorferi (Lyme disease) in dogs.
Journal of Veterinary Diagnostic Investigation 11:259-265.
Guerra MA, Walker ED, and Kitron U (2001). Canine surveillance system for Lyme
borreliosis in Wisconsin and northern Illinois: Geographic distribution and risk
factor analysis. American Journal of Tropical Medicine and Hygiene 65:546-552.
136
Hamer SA, Roy PL, Hickling GJ, Walker ED, Foster ES, Barber CC, et al. (2007).
Zoonotic pathogens in Ixodes scapularis, Michigan. Emerging Infectious Diseases
13:1131-1133.
Hinrichsen VL, Whitworth UG, Breitschwerdt EB, Hegarty BC, and Mather TN (2001).
Assessing the association between the geographic distribution of deer ticks and
seropositivity rates to various tick-transmitted disease organisms in dogs. Journal
of the American Veterinary Medical Association 218: 1092-1097.
Jones TF, Garman RL, LaFleur B, Stephan SJ, and Schaffner W (2002). Risk factors for
tick exposure and suboptimal adherence to preventive recommendations.
American Journal of Preventive Medicine 23:47-50.
Levy SA, O'Connor TP, Hanscom JL, and Shields P (2003). Evaluation of a canine C-6
ELISA lyme disease test for the determination of the infection status of cats
naturally exposed to Borrelia burgdorferi. Veterinary Therapeutics 4: 172-177.
Lindenmayer J, Weber M, Bryant J, Marquez E, and Onderdonk A (1990). Comparison
of Indirect Immunofluorescent-Antibody Assay, Enzyme-Linked Immunosorbent-
Assay, and Western Immunoblot for the Diagnosis of Lyme-Disease in Dogs.
Journal of Clinical Microbiology 28:92-96.
Merino FJ, Serrano JL, Saz JV, Nebreda T, Gegundez M, and Beltran M (2000).
Epidemiological characteristics of dogs with Lyme borreliosis in the province of
Soria (Spain). European Journal of Epidemiology 16297-100.
Piesman J, and Happ CM (2001). The efficacy of co-feeding as a means of maintaining
Borrelia burgdorferi: a North American model system. Journal of Vector Ecology
26:216-220.
Piesman J, and Sinsky RJ (1988). Ability of Ixodes-Scapularis, Dermacentor-Variabilis,
and Amblyomma-Americanum (Acari, Ixodidae) to Acquire, Maintain, and
Transmit Lyme-Disease Spirochetes (Borrelia—Burgdorferi). Journal of Medical
Entomology 25:336-339.
Pinger RR, Timmons L, and Karris K (1996). Spread of Ixodes scapularis (Acari:
Ixodidae) in Indiana: Collections of adults in 1991-1994 and description of a
Borrelia burgdorferi-infected population. Journal of Medical Entomology 33:852-
855.
Rand PW, Lacombe EH, Smith RP, Gensheimer K, and Dennis DT (1996). Low
seroprevalence of human Lyme disease near a focus of high entomologic risk.
American Journal of Tropical Medicine and Hygiene 55:160-164.
Rand PW, Smith RP, and Lacombe EH (1991). Canine Seroprevalence and the
Distribution of Ixodes-Dammini in an Area of Emerging Lyme-Disease.
American Journal of Public Health 81:1331-1334.
137
Randolph SE, Gem L, and Nuttall PA (1996). Co-feeding ticks: Epidemiological
significance for tick-bome pathogen transmission. Parasitology Today 12 :472-
479.
Sonenshine DE (1979). Ticks of Virginia. Virginia Polytechnic Institute and State
University, College of Agriculture and Life Sciences, Blacksburg, VA.
Steere AC, Coburn J, and Glickstein L (2004). The emergence of Lyme disease. Journal
of Clinical Investigation 113:1093-1101.
Steiner FE, Pinger RR, Vann CN, Abley MJ, Sullivan B, Grindle N, et al. (2006).
Detection of Anaplasma phagocytophilum and Babesia odocoilei DNA in Ixodes
scapularis (Acari : Ixodidae) collected in Indiana. Journal of Medical Entomology
43:43 7442.
Stone EG, Lacombe EH, and Rand PW (2005). Antibody testing and Lyme disease risk.
Emerging Infectious Diseases 11:722-724.
Team RDC (2005). R: A language and environment
for statistical computing. in. R Foundation for Statistical
Computing, Vienna, Austria.
Walker ED, Smith TW, Dewitt J, Beaudo DC, and McLean RG (1994). Prevalence of
Borrelia-Burgdorferi- in Host-Seeking Ticks (Acari, Ixodidae) from a Lyme-
Disease Endemic Area in Northern Michigan. Journal of Medical Entomology
31 :524-528.
Walker ED, Stobierski MG, Poplar ML, Smith TW, Murphy AJ, Smith PC, et a1. (1998).
Geographic distribution of ticks (Acari : Ixodidae) in Michigan, with emphasis on
Ixodes scapularis and Borrelia burgdorferi. Journal of Medical Entomology
35:872-882.
Zeidner NS, Burkot TR, Massung R, Nicholson WL, Dolan MC, Rutherford J S, et al.
(2000). Transmission of the agent of human granulocytic ehrlichiosis by Ixodes
spinipalpis ticks: Evidence of an enzootic cycle of dual infection with Borrelia
burgdorferi in northern Colorado. Journal of Infectious Diseases 182:616-619.
138
CHAPTER 5
Cryptic transmission of Borrelia burgdorferi, B. andersonii, and B. miyamotoi by bird-
derived ticks in the absence of the blacklegged tick
Abstract
Like other areas in the eastern United States, blacklegged ticks, Ixodes scapularis,
and the Lyme disease pathogen, Borrelia burgdorferi, are invading lower Michigan. We
asked the question whether B. burgdorferi may be maintained in cryptic enzootic cycles
prior to the arrival of I. scapularis. In particular, we hypothesized that B. burgdorferi
may be maintained within this landscape through Ixodes dentatus-mediated cryptic
cycles. During the summer breeding and fall migratory seasons of 2004 —2008, we tested
this hypothesis by mist-netting 19,631 wild birds of 105 species at a focal site in Lower
Michigan, 90 km east of the zone of blacklegged tick invasion. Eleven percent of birds,
comprising 56 species, were infested with ticks, and a total of 12,301 ticks were removed
(86% were I. dentatus). No resident birds, rabbits or other small mammals sampled at
our site harbored I. scapularis, whereas 72.8% of small mammals sampled in the tick
invasion zone over the same period were infested. In bird-derived ticks, we identified
low-prevalence infection of three Borrelia species, including a minimum prevalence of
3.5% with B. burgdorferi, 1.6% with B. andersonii, and 0.7% with B. miyamotoi (this
constitutes the first report of B. miyamotoi in ticks from wild birds). At least 35% of
rabbits had ear tissue or ticks infected with either B. burgdorferi or B. andersonii.
We found a total of 25 B. burgdorferi rDNA intergenic spacer strains at this
cryptic site, the majority of which have not been previously reported from the Lyme
139
disease endemic northeastern or Midwestern US. Also present within cryptic
transmission cycles were at least three strains of B. burgdorferi previously associated
with disseminated Lyme disease in humans. This study represents the first time that
significant levels of B. burgdorferi have been found concurrently in birds, rabbits, and
their specialist ticks in the apparent absence of I. scapularis. Such cryptic cycles may
reduce the time lag between I. scapularis invasion and the build-up of infection
prevalence. Mobile infected birds and cryptic cycles together have the potential to
introduce novel B. burgdorferi strains and accelerate the increase in risk to human and
canines within 1. scapularis invasion zones.
Introduction
The predominant maintenance cycle of the Lyme disease pathogen Borrelia
burgdorferi in the eastern United States involves the blacklegged tick Ixodes scapularis -
the main enzootic and zoonotic vector - and the white-footed mouse Peromyscus
leucopus - the main reservoir host for the pathogen and preferred host for juvenile ticks
(James and Oliver 1990). The generalist feeding capacity of I. scapulairs allows this
vector to bridge the pathogen fiom wildlife reservoirs to humans and dogs- both which
may become diseased upon infection. Lyme disease is the most frequent vector-borne
disease in the northern hemisphere, with over 20,000 cases reported annually in the
United States, and incidence is increasing annually (Bacon et al. 2007).
In 2002, a new population of I. scapularis was detected in the southwestern
comer of Michigan’s Lower Peninsula (Foster 2004), where these ticks were historically
absent (Walker et al. 1998). A subsequent invasion process ensued as ticks colonized to
140
the north through coastal dune forests along the Lake Michigan lakeshore (Hamer et al.
2007; Hamer et al. 2010). In tracking this invasion, we documented the presence of B.
burgdorferi not only in recently-invaded I. scapularis and their small rodent hosts, but
also in alternative tick and wildlife species prior to the arrival of I. scapularis. Cryptic
pathogen maintenance — that is, enzootic maintenance of B. burgdorferi in ticks that
specialize on wildlife and do not commonly bite humans - was considered as one
explanation for the incongruent distributions of I. scapularis and B. burgdorferi.
Cryptic transmission cycles of B. burgdorferi have been described previously,
with the zoonotic potential of such cycles considered low because the vector species are
host specialists that do not regularly bite humans or dogs. For example, I. neotome/I.
spinipalpis (now synonymized) is a woodrat-specialist tick that rarely bites humans and
yet is more important that I. pacificus (the human-biting equivalent of I. scapularis in the
western United States) in maintaining B. burgdorferi (Brown and Lane 1992; Maupin et
al. 1994). Similarly, in the southeastern United States, I. affinis and 1. major — both
rodent-feeding enzootic vectors — are locally more important that I. scapularis in
maintaining B. burgdorferi (Oliver et al. 2003). I. dentatus feeds almost exclusively on
birds and eastern cottontail rabbits and has been shown to maintain transmission of B.
burgdorferi between birds and eastern cottontails (Sylvilagusfloridanus) in Nantucket,
MA, in a zone of sympatric I. scapularis and endemic Lyme disease (Telford and
Spielman 1989b), where the authors concluded that infection may occasionally be
exchanged between the I. dentatus- and I. scapularis cycles. In laboratory studies, I.
dentatus has been confirmed as a competent vector of B. burgdorferi, although less so
than I. scapularis (Telford and Spielman 1989a). A majority of North American bird
141
species (Brinkerhotf et al. 2010) as well as eastern cottontails (Telford and Spielman
1989b) are competent Borrelia reservoirs, although competence is generally less than that
of the white-footed mouse. Additional evidence of I. dentatus-associated B. burgdorferi
was found in Connecticut where sympatric I. scapularis also occur, when a spirochete
indistinguishable from B. burgdorferi strain B31 was isolated (Anderson et al. 1990).
Antigenic variants of B. burgdorferi have also been found in I. dentatus removed from
rabbits (Anderson 1989). These variants were later characterized as being sufficiently
distant genetically as to represent a new genospecies, B. andersonii (Marconi et al. 1995).
Consequently, I. dentatus is associated with both B. burgdorferi and B. andersonii in
some areas where it is sympatric with I. scapularis. The level of interaction between the
I. scapularis- and I. dentatus-driven cycles and their respective spirochetes remains
unclear, and such cycles have never before been investigated in an area of I. scapularis
absence or an area undergoing invasion by I. scapularis.
The diversity of B. burgdorferi can be assessed at a number of loci, of which the
16S-23S rRNA intergenic spacer (IGS) is an ecologically, epidemiologically, and
pathologically-informative target (Bunikis et al. 2004a). At least 24 IGS strains have
been described, which can more broadly be delimited into three ribosomal spacer types
(RST 1, 2, and 3; (Liveris et al. 1995). RST 1 strains are associated with a higher
frequency of disseminated infection in humans and more invasive disease in experimental
animals (Seinost et al. 1999; Wormser et a1. 2008). Gatewood et al. (2009) found that
infection in I. scapularis nymphs due to RST 1 strains was greater in the Northeastern as
compared to Midwestern United States, where RST 2 and 3 strains were relatively more
dominant. This pattern was associated with the degree of synchrony of feeding of the
142
immature stages of I. scapularis, in which there is a longer duration between nymphal
and larval activity in Northeast, which may favor longer-persisting RST 1 strains.
Similarly, within recently-invaded ticks in lower Michigan, Hamer et al. (2010) found
that only 13% of strains were within the RST 1 group. In addition to geographic
structure to B. burgdorferi genotypes, the influence of host species of pathogen genotype
is also the subject of study. Whereas Brisson and Dykhuizen (2004) found host
associations of different B. burgdorferi strains, Hanincova et al. (2006) found that most
of the known B. burgbrgeri genotypes can infect a range of hosts, and that cross-species
transmission among various mammalian hosts is common.
We hypothesized that B. burgdorferi can be maintained by Ixodes dentatus and
bird and eastern cottontail hosts in the absence of I. scapularis. Further, in consideration
of the previous evidence for host association of strains, and the host specificity of I.
dentatus on birds and rabbits (i.e., hosts not hosts used heavily in I. scapularis-driven
scenarios of B. burgdorferi maintenance), we hypothesized that the pathogen strain
diversity in bird-associated ticks differs from that of I. scapularis-mediated transmission
at endemic sites. We addressed these hypotheses by i) characterizing the tick parasites of
birds, their seasonal phenology, and their infection status with Borrelia spirochetes at a
focal site in lower Michigan; ii) searching for I. scapularis on mice and chipmunks at the
site; and iii) characterizing strain diversity of the Borrelia spp. spirochetes we detected.
Materials and methods
Bird mist netting. The Pitsfield Banding Station in Vicksburg, MI (Figure 5.1) is a late-
succession forest on old agricultural fields and gravel pits that were farmed and mined
143
through the 1970s. Bird banding has occurred at Pitsfield since 1985. The banding area
encompasses approximately 10 hectares and is used for collection of data on avian
productivity and survival. In 2004-2007 we collaborated with bird banders to mist-net
birds and collect tick samples for an average of 3 days per week during the bird breeding
season of May-August and for an average of 5 days/week during the fall migratory
season of September—November. We aimed to investigate birds in both the breeding and
fall migratory seasons so as to assess both local and bird-imported ticks and Borrelia
strains. Between 22 and 36, 12 m mist nets (Avinet, Dryden, NY) were used to capture
birds. Nets were run from sunrise for approximately 6 hours on fair weather days and
checked every hour. Birds were weighed, identified to species, sex, and age class (hatch
year or afier hatch year), and leg-banded with federally-issued bands before release.
Recaptures of previously-banded birds were noted. All birds were checked for ticks using
a magnifying head loupe and straws to blow feathers from cars. Searching was restricted
to the neck, head, ears, and face. Ticks were plucked and preserved in 70% ethanol. Due
to time constraints and bird safety, for a small number of birds, ticks were observed but
not removed and these birds are considered only in the overall infestation estimates with
ticks categorized as unidentified. For a different small number of birds, primarily those
that were very heavily infested, not all ticks were removed- in these cases; we assume the
removed ticks are representative of the tick species/ stages that were present. Michigan
State University Institutional Animal Use and Care Committee approved all research.
Mammal trapping. To increase our search intensity for I. scapularis, we sampled the
small mammal community with a particular focus on white-footed mice and eastern
144
chipmunks, as these are the most sensitive indicators of low-density blacklegged ticks in
lower Michigan (Hamer et al. 2010). In addition to trapping small mammals at Pitsfield,
we also trapped at Van Buren State Park, 90 km due west of Pitsfield, where l scapularis
is known to have invaded within the past decade. This latter trapping was done to control
for I. scapularis phenology; both areas were sampled in May and June, when I.
scapularis larvae and nymphs are simultaneously active. Additionally, at Pitsfield we
targeted the capture of eastern cottontails, the preferred host for adult I. dentatus. In
2004-2009, small mammals were trapped for an average of three nights per summer using
an average of 100 Sherman live traps (H. B. Sherman Traps, Tallahassee, FL) spaced 10
m apart and baited with sunflower seed. Rabbits were trapped with wooden box traps
baited with apples; these traps on occasion also captured other medium-sized mammals
including raccoon (Procyon lotor), Virginia Opossum (Didelphis virginiaina), and
woodchuck (Marmota monax). Small mammals were anesthetized using Isoflurane
(lsoFlo, Abbot Laboratories, Abbott Park, IL) and medium mammals were anesthetized
using ketamine hydrochloride (Ketaset; Fort Dodge, Overland Park, KS) and xylazine
hydrochloride (Rompun; Bayer Health Care, Kansas City, KS), with yohimbine
hydrochloride (Antagonil; Wildlife Laboratories, Fort Collins, CO) used to antagonize
the xylazine. Vital data were obtained by inspection. Each animal then was examined for
ticks, biopsied from both ears using a 2-mm (small mammals) or 4-mm (medium
mammals) biopsy punch (Miltex Insturments, York, PA), and finally marked with a
uniquely numbered ear tag (National Band and Tag, Newport, KY). Ticks and ear
biopsies were stored separately in 70% ethanol. Animals recaptured within the same site
145
visit (i.e. recaptured the day after initial processing) were simply checked again for ticks.
All animals were released at the site of capture.
Tick and Borrelia spp. detection. Ticks were identified to species and stage (Keirans
and Clifford 1978; Sonenshine 1979; Durden and Keirans 1996). Representative tick
specimens are vouchered in the National Tick Collection at Georgia Southern University.
Species identity for a subset of ticks was confirmed molecularly by amplifying and
sequencing the tick 5.8S — 28S rDNA internally transcribed spacer (ITS-2) (Poucher et al.
(1999). Total DNA from all ticks and ear biopsies was extracted using the DNeasy Blood
and Tissue Kit (Qiagen, Valencia, CA) following the manufacturer’s animal tissue
protocol, but with the following modifications. Ticks were first bisected using a sterile
scalpel or were pulverized in liquid nitrogen, followed by an overnight incubation in lysis
buffer. DNA was eluted using 50 pl elution solution warmed to 70° C. Ear biopsies (one
per animal), and adult and nymphal ticks, were extracted individually; same-species
larvae from the same individual animal were pooled for extraction. B. burgdorferi strain
B31-infected nymphal I. scapularis generously provided by the Centers for Disease
Control and Prevention (CDC) served as the positive extraction control, and water as a
negative extraction control.
All tick and ear biopsy DNA extracts were tested for the presence of Borrelia
species using a nested polymerase chain reaction (PCR) for the 16S — 23S rRNA
intergenic spacer region (IGS) of Borrelia spp. (Bunikis et al. 2004a) followed by
visualization with gel electrophoresis. The outer forward primer is located at the 3’ end of
the 16S gene and the outer reverse primer is located in the coding sequence for ileT tRN A
146
in the spacer. Included in the amplicon is the alaT gene for tRNA and approximately 805
nucleotides of a total of 3052 nucleotides of the spacer, by B. burgdorferi B31
coordinates. PCR enzyme kits were used throughout (PCR Supermix, Invitrogen,
Carlsbad, CA and FailSafe PCR System, Epicentre, Madison, WI). DNA from B.
burgdorferi strain B31-infected ticks from the CDC served as a positive PCR control, and
water served as a negative PCR control. Preliminary experiments showed that this nested
assay remains sensitive in the range of 10° and 10'1 organisms. A subset of IGS-positive
DNA samples was subjected to amplification of the 16S rDNA gene using both a real-
time, quantitiative PCR (qPCR) with a probe specific to B. burgdorferi (Tsao et al. 2004)
or an assay using primers R811 and S5 as published in Rudenko et al. (2009).
Nucleotide sequencing. Species identification and strain typing of Borrelia-positive tick
and ear tissue samples was attained through DNA sequencing. All IGS products were
purified (Qiagen PCR Purification Kit; Qiagen, Valencia, CA) and sequences were
determined using the inner primers on an ABI Prism 3100 Genetic Analyzer (Applied
Biosystems, Foster City, CA). Sequences were identified as either B. burgdorferi, B.
andersonii, or B. miyamotoi based on comparisons to published sequences using the basic
local alignment search tool in GenBank (Altschul et al. 1990). For B. burgdorferi
sequences, a 500 nucleotide segment of the IGS was aligned with the prototypical strains
published in Bunikis et al. (2004a) using the ClustalW algorithims within the program
Mega4 (Tamura et al. 2007). Analysis of this fragment size allowed for identification of
the 10 main IGS groups (groups 1-10; the minimal matrix for differentiation of these 10
groups includes mutations which all occur within the first 309 nucleotides of the
147
fragment), and of 20 IGS subtypes (1A, 18, 2A/C, ZB, 2D, 3A, 3B/C, 3D, 4A, 4B, 5, 6A,
6B, 6C, 7A/B, 8A/C, 88, 8D, 9, 10), as presented in Bunikis et al. (2004a). Alignments
were manually checked. Additionally, sequences were identified to broad ribosomal
spacer type (RST 1, 2, or 3; (Liveris et al. 1995) based on clustering topology of the IGS
phylogenetic trees. In the case that a sequence we derived did not identically match any
published strain or any strain we previously found in I. scapularis across the Midwest, we
classified it as novel IGS mutant. Sequence chromatographs were manually scrutinized
for confidence in nucleotide assignments and evidence of mixed strain infections which
were noted and removed from further analysis. For at least one representative of each
detected strain, we also determined the sequence in the reverse direction and/or
determined the sequence in the forward direction twice to validate the occurrence of
unique mutations. A similar protocol was followed for B. miyamotoi sequences, for
which prototypical IGS strains for use in alignments were obtained from Bunikis et al.
(2004b). For B. andersonii, the IGS region was not previously published to Genbank and
therefore not available for comparisons to our strains. We therefore sequenced the 16S
gene of a subset of suspect—B. andersonii samples in both directions to allow for species
identification using the Rudenko et al. (2009) primers above, as this gene target has
previously been described for B. andersonii We found 100% 168 sequence homology to
four published sequences, including those found in the kidney of an eastern cottontail in
New York (Marconi et al. 1995), in I. dentatus larvae and nymphs from rabbits in
Missouri, and in a drag-sampled nymph in Georgia (Lin et al. 2003), thereby facilitating
future identification of B. andersonii based on our IGS sequences. A representative 168
sequence from our Michigan B. andersonii has been deposited in Genbank, accession no.
148
GU993310. Double-stranded sequencing of the ITS-2 region of the tick rDNA (to
validate tick species identity) was performed using the primers used in the PCR assays.
Resulting sequences were compared to published sequences in Genbank.
To evaluate phylogenetic relationships among B. burgdorferi haplotypes, we
constructed unrooted neighbor-joining phylogenies and unrooted minimum spanning
trees (MST) using Mega4 and TCS 1.21, (Clement et al. 2000; Tamura et al. 2007)
respectively. Evolutionary distances were computed using the Kirnura 2-parameter
method and are in the units of number of base substitutions per site. Percentage support
values for clades within the neighbor-joining tree were obtained from 1000 bootstrap
iterations. The MST method determines the gene network in which the total length of the
branches that connect haplotypes is minimized, and discrimination among equal-length
MSTs required the assumption that older alleles are more common than recently derived
alleles, and new mutations are more likely found in the same population as their ancestor.
Statistics. To determine mist netting efficiency, one net hour (NH) is defined as the
equivalent of one 12m net run for one hour. Chi-squared tests for independence were
used to assess coinfestations. Logistic regression was used to assess trends in tick
infestation and tick/host infection over the 4-year sampling period. Comparisons between
birds captured in the migratory versus breeding seasons were made by calculating the 2-
ratio and associated probabilities for the difference between two independent proportions.
A minimum infection prevalence (MIP; i.e., assuming only one positive larva per pool)
was used for tests done on pooled larvae. Statistics were performed using Statistix 8
(Analytical Software, Tallahassee, FL). The effect of sample size on strain richness was
149
assessed using a web-based ’rarefaction calculator’ (University of Alberta, Edmonton,
Canada; available at http://www.biology.ualberta.ca/jbrzusto/rarefact.php). Strain
richness was estimated by using the nonparametric model of Chao, which considers the
number of operational taxonomic units observed, and the frequency with which each was
observed, to estimate total population strain richness (Chao and Tsung-Jen 2003).
Evidence for gene conversion was examined using Sawyer’s Test in GENECONV
version 1.81 (http://www.math.wustl.edu/~sawyer/geneconv/), which tests the null
hypothesis that nucleotide substitutions observed in a set of aligned sequences are
randomly distributed (Sawyer 1989).
Results
Trapping success and infestation prevalence. A total 19,631 captures of birds,
representing 105 species, occurred in 48,030 net hours for an overall netting efficiency of
40.9 captures per 100NH. Efficiency was significantly higher in October and November
(53.3 birds/IOONH) versus May through August (33.4 birds/ 100NH), reflecting the influx
of migrants. Of all capture events, 19.5% were recaptures. Gray Catbird, Myrtle
Warbler, American Goldfinch, and White-throated Sparrow were the most abundant
species with a sample size of over 1000 captures of each species, together comprising
35.6% of all captures.
A total of 2074 bird captures (10.6% of all captures) were infested with ticks, of
which 74.2% were hatch year, 22.5% were after hatch year, and 3.3% were of unknown
age class. A total of 12,301 ticks were removed from birds, with a mean infestation
burden of 5.9 ticks per infested bird, and hatch year birds had a significantly higher
burden of ticks (6.2 per infested bird) versus after hatch year (5.2 per infested bird; P =
150
0.0003). Similar numbers of male and female birds were infested (366 males and 350
females), although the majority of birds fiom which ticks were removed were of
unknown sex (11 = 1,350 birds). Of all ticks, 86.4% were I. dentatus (10,363 larvae and
265 nymphs), 13.4% were Haemaphysalis leporispalustris (1535 larvae and 118
nymphs), 0.1% were I. scapularis (7 larvae and 6 nymphs), and 0.06% were
Dermacentor variabilis (6 larvae). Voucher specimens of ticks were deposited with the
National Tick Registry. Molecular identification of the tick ITS-2 region resulted in
visible differences in the fragment size of I. scapularis, I. dentatus, and H.
leporispalustris on agarose gel. Our 1. scapularis sequences matched with 100%
homology to published I. scapularis sequences. Because the ITS-2 regions of I. dentatus
and H. leporispalustris were not previously deposited in Genbank, our samples for these
two species matched poorly to published species (I. dentatus matched with 85%
homology to 1. pacificus; H. leporispalustris matched with 90% homology to H.
Iongicornis). Two different sequences for H. leporispalustris were found.
Representative ITS-2 sequences for all three tick species were deposited in Genbank
(GU993304- GU993307).
The most commonly parasitized bird species, in which at least one-third of all
individuals were infested, included Brown Thrasher, Carolina Wren, Eastern Towhee,
White-throated Sparrow, Song Sparrow, Lincoln's Sparrow, Hermit Thrush, American
Robin, and Yellow-breasted Chat (Table 5.1). In total, 285 birds were noted to harbor
ticks and were released with no ticks removed. Of the remaining birds, 84.6, 23.5, 0.2 and
0.06% of parasitized birds harbored I. dentatus, H. leporispalustris, I. scapularis and D.
variabilis, comprising 7.8, 2.2, 0.02 and 0.005% of all birds, respectively. Five of the t0p
151
. .w" “—2—
ten hosts most commonly parasitized by I. dentatus were also among the top ten most
parasitized by H. leporispalustris, including Brown Thrasher, Lincoln's Sparrow, Song
Sparrow, Eastern Towhee, and Carolina Wren. Co-infestation of birds with both
I. dentatus and H. leporispalustris simultaneously was found for 8.2% of all parasitized
birds (0.8% of all birds); this rate is approximately 4.5-times higher than what would be
expected by chance ()8 = 439.1; P < 0.001). All ticks were immature stages with the
exception of a single adult female 1. dentatus collected from a House Wren. The average
burden of I. dentatus larvae and nymphs on infested birds was 7.0 d: 0.5 (standard error)
and 1.7 i 0.2, respectively. Peak 1. dentatus larval burden occurred on an Eastern
Towhee which harbored 162 larvae that were removed and an estimated 80 ticks that
were left on the bird. Peak 1. dentatus nymphal burden was 11 on a Brown Thrasher that
also harbored 14 other ticks. The average burden of H. leporispalustris larvae and
nymphs on infested birds was 4.2 :1: 0.5 and 1.2 :t 0.1, respectively, with a peak larval
burden of 56 on a Swainson’s Thrush that harbored 2 additional ticks, and a peak
nymphal burden of 4 on a White-throated Sparrow that harbored 23 additional ticks.
Across all sampling, only three individual birds harbored I. scapularis: a
Swainson’s Thrush, a Hermit Thrush, and a Connecticut Warbler. All three were hatch
year birds upon their first capture during in late August and September of 2004-2005 and
were not captured subsequently. The Hermit Thrush was the most heavily infested, with
6 larvae and 5 nymphal I. scapularis, in addition to 2 H. leporispalustris larvae. The
Connecticut Warbler harbored a single I. scapularis larva. The Swainson’s Thrush
harbored a single I. scapularis nymph, in addition to 10 and 3 H. leporispalustris larvae
and nymphs, respectively. Communication with local omithologists and review of
152
species-specific breeding and migratory data (Brewer et al. 1991) indicate that these three
captures represent fall migrants from further north on a migratory stopover.
Phenology of bird-associated ticks. I. dentatus larvae exhibited bimodal peaks of
activity, with the earlier peak in June and the second peak in October-November, whereas
I. dentatus nymphs were mostly active in May-July, with smaller numbers throughout the
fall (Figure 5.2a). Conversely, H. leporispalustris larvae were most active in August-
September, when both stages of I. dentatus were least active. H. leporispalustris nymphs
had a low level of activity throughout the sampling period (Figure 5.2b).
Ticks on eastern cottontails. Nineteen captures of eastern cottontails represented 15
individual rabbits; all recaptures occurred at least 7 days after initial trapping and
therefore are considered as independent in the calculations of infestation prevalences. Of
these captures, 73.7% were infested by ticks of three species. The two most common tick
species on cottontails were I. dentatus and H. leporispalustris, which parasitized 63.2 and
26.3% of captures, respectively. Nymphs and adults of each species were found, with
maximal burdens of 1 and 24 for I. dentatus and 1 and 3 for H. leporispalustris,
respectively. Three cottontails were infested by both species simultaneously, and the
maximum overall tick burden was 27. One rabbit was infested by two adult male
Dermacentor variabilis in addition to nine 1. dentatus adults.
Ixodes scapularis on white-footed mice and chipmunks. At Pitsfield, no I. scapularis
were found on 65 mice and 59 chipmunks trapped in May-June of 2005-2009. Fourteen
153
of these mammals were trapped again the day after initial processing, and none harbored
ticks. Conversely, at Van Buren, 69.3% of 179 mice and 100% of 23 chipmunks were
infested with I. scapularis (Table 5.2); these mammals were all trapped within two weeks
of the dates of mammal trapping at Pitsfield so as to control for I. scapularis phenology.
Ticks on non-target captures. Non-target captures at Pitsfield included 3 raccoons, l
fox squirrel (Sciurus niger), 3 meadow jumping mice (Zapus hudsonius), 1 red squirrel
(Tamiasciurus hudsonicus), 3 northern short-tailed shrews (Blarina brevaucada), 4
Virginia opossum, and 3 woodchuck. Prevalence of ticks on these hosts included 33.3%
of raccoons, 25% of opossums, and 33.3% of woodchuck with D. variabilis, 66.6% of
woodchuck with I. cookei, and 66.6% of meadow jumping mice with I. dentatus.
Borrelia infection in bird-associated ticks. A total of 2202 ticks/larval pools were
assayed for infection with Borrelia species. Of these, 146 (6.6%) tested positive for
Borrelia species, and DNA sequencing of the rDNA IGS and/or 168 gene was used to
identify the pathogen in 128 samples: 78 (3.5%) were positive for B. burgdorferi, 35
(1 .6%) were positive for B. andersonii, and 15 (0.7%) were positive for B. miyamotoi
spirochetes. There were 18 additional samples which produced IGS bands at
approximately 980 base pairs in size (indicative of either B. burgdorferi or B. andersonii)
and an additional 22 samples that produced faint IGS bands at approximately 500 base
pair size (indicative of B. miyamotoi) that were unsuccessfully sequenced. Accordingly,
the infection prevalences for all species should be considered as minimum. B. burgdorferi
and B. andersonii infection prevalence in nymphs was significantly greater than that of
154
larval pools (B. burgdorferi: 5.2 and 3.2%, respectively; P = 0.02; B. andersonii: 6.2 and
0.5%, respectively; P < 0.0001). Conversely, 14 of the 15 sequence-confirmed B.
miyamotoi-positive samples were pools of larval I. dentatus (0.8% larval MIP), and a
single sample was a nymphal I. dentatus (0.3% nymphal infection prevalence; Table 5.3).
Infection of B. burgdorferi and B. andersonii was found in I. dentatus, H.
leporispalustris, and I. scapularis, whereas all B. miyamotoi-infected ticks were I.
dentatus (Table 5.3). Of the small number of I. scapularis found in our study, one of the
5 nymphal I. scapularis removed from the Hermit Thrush was infected with B.
burgdorferi, and the single larva removed from the Connecticut Warbler was infected
with B. andersonii.
All three Borrelia species were found in ticks removed from birds of both age
classes, and the prevalence of infection with each pathogen in each age class was not
different that what would be expected based on the age structure and overall tick
infection rates (hatch year, if = 0.73. df= 3, P = 0.87; after hatch year, x2 = 2.88, df= 3, P
= 0.41). Similarly, all three species infected ticks removed from both sexes of bird in
proportion to the sex structure and overall tick infection rates (female, 752 = 3.35. df= 3, P
= 0.34; male, x2 = 1.98, df= 3, P = 0.58). Annual variation in tick infection prevalence
was not significant B. burgdorferi (0.5, 6.3, 1.0, and 5.0%, in 2004-2007, respectively; R2
= 0.16; P = 1) or B. andersonii (0.3, 2.7, 1.0, and 1.8% in 2004-2007, respectively; R2 =
0.23; P = 1). All 15 B. miyamotoi-infected ticks were collected in 2007 (additional
samples suspect-positive for B. miyamotoi were collected in earlier years, but were not
sequence-confirmed; see above). Tick infection prevalence varied seasonally within
years (Figure 5.3). Aggregating all four years of the study, B. burgdorferi was present in
155
all months of May-November with the highest monthly prevalence of 1 1-12% in May
and July (Figure 5.3). B. andersonii was present in the months of May - October with
the highest monthly prevalence of 4.9-6% in May-July. B. miyamotoi was only found in
October and November with monthly prevalence of 1 and 5%, respectively. In the
months of June-August only, a time period which largely excludes the spring and fall
migrations, we detected 27 B. burgdorferi-positive ticks/pools (5.4% of ticks/pools tested
in those months) and 21 B. andersoni-positive ticks (4.2% of ticks/pools tested in those
months), comprising 34.6 and 60% of all positives, respectively. Of these June-August
positive samples, 59.3 and 61.9% were from hatch year birds, respectively.
Assuming no significant transovarial or co-feeding transmission, our data on
production of infected larval ticks (i.e natural xenodiagnosis; Table 5.1) implicate 19
species of bird as reservoir-competent for B. burgdorferi, and 7 species of bird as
reservoir-competent for B. andersonii. As further evidence of local infectious hosts for
both pathogens, 9 birds were associated with multiple infected ticks/tick pools, removed
during one or more separate capture events (B. burgdorferi: American Robin, Song
Sparrow, Swainson’s Thrush, and Whiteethroated Sparrow; B. andersonii: Brown-headed
Cowbird, Hermit Thrush, Northern Cardinal, and Song Sparrow). In contrast, B.
miyamotoi is a transovarially-transmitted spirochete (Scoles et al. 2004), and our findings
of infected ticks on birds do not necessarily implicate host reservoir competence. We
removed B. miyamotoi -positive ticks/tick pools from three host species, of which a
majority (73.3%) were from Northern Cardinal, and a minority (13.3% each) was from
American Robin and Hermit Thrush.
156
Borrelia infection in mammal-associated ticks. A total of 150 ticks removed from
white-footed mice, chipmunks, eastern cottontails, and non-target captures at Pitsfield
were assayed for infection with Borrelia species. The only host species associated with
positive ticks was the eastern cottontail, in which 15 of 111 (13.5%) I. dentatus adults
tested positive, with a minimum of four (3.6%) B. burgdorferi-positive ticks, a minimum
of 7 (6.3%) B. andersonii-positive ticks, and 4 samples of unknown pathogen identity
(because sequencing was unsuccessful).
Borrelia infection in mammal ear biopsies. At Pitsfield, one of 61 (1.6%) and two of
55 (3.6%) white-footed mouse and chipmunk ears tested positive for B. burgdorferi,
whereas the equivalent prevalences at Van Buren were 30.4% of 171 and 34.8% of 23,
respectively (Table 5.2). At Pitsfield, seven of 20 (3 5%) cottontail ear biopsies were
positive for Borrelia species, which included a minimum of four (20%) B. burgdorferi-
positive ears, a minimum of 1 (5%) B. andersonii—positive ears, and 2 ears of unknown
pathogen identity (no sequence). None of 13 car biopsies from the non-target captures
tested positive.
B. burgdorferi genotypes. B. burgdorferi IGS PCR products were successfully
sequenced from 80 samples, including 71 ticks/larval pools removed from birds
(including 39 larval pools and 11 nymphs of I. dentatus, 14 larval pools and six nymphs
of H. leporispalustris, and one nymph of I. scapularis) and nine mammal-associated
samples (ear tissue from one white-footed mouse, two eastern chipmunks, and one
eastern cottontail, and from five adult I. dentatus removed from five different rabbits).
157
Of the 80 sequences, 10% were interpreted as mixed strain infections due to the presence
of double-nucleotide peaks at polymorphic sites. These mixed strain infections comprised
six tick sammes removed from birds, one chipmunk ear, and one tick removed from a
rabbit, and are not considered in further analyses. Among all IGS strains, we found no
evidence for recombination using Sawyer’s test. Within the 500 nucleotide IGS
fragment, 58 sites were found to be polymorphic, including one. indel block of 7
nucleotides (treated as a single polymorphism).
In total, 25 IGS strains were found among 72 samples, including strains within all
three RST groups (GU993279- GU993303). Of all strains, 32% are ‘ubiquitous’ (i.e.,
also found in association with classic transmission cycles in the Northeast (Bunikis et al.
2004a) and/or Midwest (S. Hamer, unpublished)) and 68% are ‘indigenous’(i.e., novel
IGS mutants found only at Pitsfield in cryptic transmission; Figure 5.4). The only strains
found more than once were ubiquitous strains; all indigenous strains were singletons,
many being single or double nucleotide polymorphisms of ubiquitous strains. A MST of
Pitsfield strains shows, for example, that there were many indigenous strains very similar
to the ubiquitous strain IGS 2D, including 6 single nucleotide polymorphism mutants and
3 double nucleotide polymorphism mutants (Figure 5.5). A rarefaction curve to assess
strain richness based on the frequency and diversity of strains we sampled suggests that
true B. burgdorferi strain richness at Pitsfield is vast, as the rate at which new strains
were found per unit of individuals sequenced showed no sign of approaching an
asymptote (Figure 5.6). While we detected a total of 25 IGS strains within 72 samples
derived from investigations of over 19,000 birds and a small number of mammals, the
Chao-l non-parametric estimator of true species richness is 245.5 :t 100.7 strains.
158
The single bird-bome B. burgdorferi-positive from a I. scapularis nymph was
type ‘Midwest A’ of RST group 2 (Figures 5.4 and 5.5), a ubiquitous strain with a wide
Midwestern distribution. There was no difference in the proportion of indigenous strains ,
in bird-associated ticks during the migratory (May; September-November) versus non-
migratory (June-August) seasons (23.8 and 26.1%, P = 0.84). The avian species that
harbored ticks infected with indigenous strains during the migratory seasons included
species known to both breed and winter on the property (e. g., Northern Cardinal) as well
as species that do not breed on site that could have picked up these indigenous strains
elsewhere, or on site during a migratory stopover (e. g., Hermit Thrush and White-
throated Sparrow).
Of the seven B. burgdorferi sequences identified in mammal tissues or mammal—
associated ticks, all but one were ubiquitous strains, with all three RST groups
represented (Figure 5.4). The exception was an ear tissue sample from a rabbit that was
infected with strain ‘Novel 11’, an indigenous strain within RST 3. The single infected
white-footed mouse ear was infected with ‘Midwest A’, the chipmunk ear was infected
with ‘Midwest B’ and the I. dentatus adults removed from rabbits were infected with IGS
2D (11 == 2), IGS 1A (11 = l), and ‘Midwest K’ (n = l).
B. andersonii genotypes. B. andersonii IGS PCR products were successfully sequenced
from 33 samples (including 27 nymphs/larval pools from birds, 5 ticks from rabbits, and
1 rabbit ear). We found a total of 12 unique strains (‘Michigan Bird A -— L’, accession
nos. HM015226-237) of which all but two were represented by more than one sequence
in our populations. Within the 500 nucleotide IGS fragment, 57 sites were found to be
159
polymorphic, including three indels (two single nucleotide indels, and one 10 nucleotides
indel which was treated as a single polymorphism). The amount of evolutionary change
separating B. andersonii IGS strains was greater than that for B. burgdorfieri IGS strains
(Figure 5.7).
B. miyamotoi genotypes. B. miyamotoi IGS PCR products were successfully sequenced
from 12 samples (including 11 larval pools and 1 nymph of I. dentatus). A majority (11 =
11) were identical to the published North American strain of B. miyamotoi (Bunikis et al.
2004b). A single sample from a larval pool had a single nucleotide polymorphism within
the spacer, confirmed with reverse strand sequencing (GU993309).
Discussion
Our four-year dataset provides evidence of a unique cryptic transmission system
involving three Borrelia species- B. burgdorferi, B. andersonii, and B. miyamotoi -
maintained by bird-associated ticks in the apparent absence of the bridge vector
1. scapularis. The focal site from which we trapped birds is outside the detected range of
established 1. scapularis populations, based on both historic records and our findings over
the 2004-2007 sampling period (Hamer et a1. 2010). Whereas birds in general have
previously been considered to dilute the force of infection in endemic areas, by diverting
tick bites away from white-footed mice which are generally able to infect more ticks than
do birds (Giardina et al. 2000; LoGiudice et al. 2003), we present an ecological scenario
in which birds may accelerate increasing Lyme disease risk through their maintenance of
B. burgdorferi in the absence of classic I. scapularis/white-footed mouse transmission.
160
Absence of I. scapularis. Given our evidence based on small mammal and avian
investigations, we conclude that the distribution of B. burgdorferi is not congruent with
that of I. scapularis. I. scapularis was not encountered on any mammals or any resident
birds at our study site, and was rarely encountered on fall migrants, with three birds
found to carry 1. scapularis during late August through October. Given the species-
specific breeding and migratory distribution and timing of these species(Brewer et al.
1991), combined with the endemicity of I. scapularis in the southern Upper Peninsula
(Walker et al. 1994) and other zones in the upper Midwest, we posit that these migratory
birds picked up I. scapularis prior to arrival at Pitsfield. Furthermore, all three infested
birds were captured during the first two years of this four-year study, and no additional 1.
scapularis infestations on wildlife were subsequently found, suggesting that the
infestation of these birds did not represent the beginning of an appreciable process of
successful establishment.
Migratory birds are known to be an important source of adventitious I. scapularis
in areas beyond the range of established 1. scapularis (Klich et al. 1996), and may be
responsible for extending the geographic range of]. scapularis (Smith et al. 1996; Ogden
et al. 2008) and B. burgdorferi sensu lato (Weisbrod and Johnson 1989; Dubska et al.
2009). During the September — November fall migratory period of four years, we
searched a total of 14,667 birds, and found 3 (0.02%) to harbor a total of 13 I. scapularis
ticks. Given this average of 0.0009 1. scapularis per migratory bird, and given that not all
I. scapularis are likely to drop off while their host is on site, it seems unlikely that south-
bound migrants alone would seed a new population of I. scapularis at this site. This is
161
likely influenced by the mismatch between peak juvenile 1. scapularis activity in northern
endemic areas and the timing of fall migration. In contrast, Ogden et al. (2008) found
that north-bound migratory birds in the spring are very likely to be important in
expanding the northern geographic range of I. scapularis across a wide front in northern
Canada, with 0.35% of birds infested and an average of 0.007 I. scapularis per migratory
bird. It is likely only a matter of time before a critical mass of occasional bird-derived
engOrged 1. scapularis larvae and nymphs, combined with ticks moving in as a
continuation of the invasion process occurring to the west (Hamer et al. 2010), will seed a
reproducing population.
Bird-associated ticks. The vast majority of ticks removed from birds at our site were I.
dentatus (86.4%), followed by H. leporispalustris (13.4%) and I. scapularis (0.1%). In a
review of tick species assemblages on avain hosts across both sites endemic and non-
endemic for I. scapularis, a similar preponderance of I. dentatus has not been found. For
example, no I. dentatus were identified among 1643 ticks removed from 4317 birds from
north-central Wisconsin, eastern and central Minnesota, and Michigan’s northern Lower
Peninsula, where 92-98% of ticks were H. leporispalustris, followed by small numbers of
I. scapularis and D. variabilis (N icholls and Callister 1996). Similarly, no I. dentatus
were identified among 883 ticks removed from 1693 birds in the central Maryland
Piedmont, where 97.4% of ticks were I. scapularis (Scharf 2004). In two Lyme disease-
endemic foci, I. dentatus comprised a small proportion of ticks removed from birds
relative to I. scapularis (Westchester Co., NY: 1067 ticks; 3.2% I. dentatus, 96.5%
1- scapularis (Battaly and Fish 1993); Lyme, CT: 4,065 ticks; 1.6% I. dentatus, 3.7% H.
162
leporispalustris; 94.4% I. scapularis; (Stafford et al. 1995). In a recent study of tick
parasites of migratory birds in Canada, I. dentatus and I. scapularis were nearly equal in
abundance (34 and 29.3% of ticks, respectively), and H. leporispalustris comprised
22.5% of the sample (Ogden et a1. 2008). The preponderance of I. dentatus among the
birds we sampled may reflect a unique area of I. dentatus abundance, reflecting the tick
community composition prior to I. scapularis invasion, and the extent to which I.
dentatus and I. scapularis compete is unknown.
1. dentatus has a widespread distribution throughout the eastern United States,
with recoveries from at least 26 states and Ontario (Durden and Keirans 1996);
immatures parasitize several species of Passerine birds and all stages parasitize cottontail
rabbits. 1. dentatus may itself be undergoing range expansion, having been detected in
Michigan only since 1992, despite an established passive surveillance program since
1968 (Walker et al. 1998). Alternatively, it may be an endemic species that is not readily
detected due to its degree of wildlife host specificity. I. dentatus has been associated with
Rickettsia rickettsia, agent of Rocky Mountain spotted fever (Clifford et al. 1969),
Connecticut virus (Main and Carey 1980), B. andersonii (Anderson et a1. 1989; Marconi
et a1. 1995), and B. burgdorferi (Telford and Spielman 1989b) including antigenically
variable strains of B. burgdorferi (Oliver et al. 1996; Oliver et al. 1998). This tick species
has been posited as a ‘transition vector’ of B. burgdorferi, facilitating the movement of
the spirochete into new areas beyond the distributional limit of I. scapularis (Levine et a1.
1991)
Comparatively less is known about H. leporispalustris. The cottontail rabbit is
the dominant host for all stages of H. leporispalustris, and immatures can also be found
163
on wild birds (Sonenshine and Stout 1970). H. leporispalustris is recognized as a vector
of the agents of Rocky Mountain spotted fever (Sonenshine and Stout 1970), tularemia
(Cooney and Burgdorfer 1974), and Sawgrass vius (Sather et al. 1970). In Lyme disease-
endemic areas, few infected H. leporispalustris ticks are generally found relative to other
species (Stafford et al. 1995), yet infected H. leporispalustris have been documented
(Anderson and Magnarelli 1984; Levine et al. 1991), and vector competency is unknown.
In the current study, H. leporispalustris infection rates were slightly higher than that of I.
dentatus, though H. leporispalustris was relatively less abundant. In our study, H.
leporispalustris generally parasitized the same species as did I. dentatus, and a significant
rate of simultaneous co-infestation with both species was found, despite the differences in
peak activity periods of the ticks. Therefore, infection in H. leporispalustris may reflect
co-feeding transmission from I. dentatus, spirochetemic bloodmeals from infectious
hosts, transstadial transmission, or perhaps vector competency.
Ecology of tick-borne pathogens at Pitsfleld. Our documentation of infection of bird-
associated ticks with at least three species of Borrelia underscores the potential
importance of birds in Borrelia maintenance and dispersal. Local transmission of all
three pathogens is evident (see below) in that hatch year birds, known to have been born
on-site, harbored infected ticks at times that exclude the migratory period. Furthermore,
eastern cottontails -— important hosts for feeding adults of both species of bird-associated
ticks — were present on site, infected with both B. burgdorferi and B. andersonii.
164
B. burgdorferi. We detected a low prevalence of B. burgdorferi infection in bird-derived
(minimum of 3.5%) and rabbit-derived (minimum of 3.6%) I. dentatus, and suggest that
this pathogen is being actively transmitted at Pitsfreld. Our findings are similar to the
scenario reported by Telford and Spielman (1989b), except that this system is operating
in the apparent absence of I. scapularis. Ogden et al. (2008) reported that no larval and
15.4% of nymphal I. scapularis removed from birds were infected with B. burgdorferi,
whereas 0.3% of I. dentatus larvae and no I. dentatus nymphs were infected. Stafford et
al. (1995) found 4.3% of all I. dentatus in a Lyme endemic area to be infected with B.
burgdorferi, but 0% of H. leporispalustris. Because of the narrow host range of I.
dentatus on birds and rabbits, its involvement in B. burgdorferi maintenance is cryptic
with respect to disease (Telford and Spielman 1989b).
Vector competency trials have not been conducted for H. leporispalustris. As a
non-Ixodes genus tick, H. leporispalustris would not be predicted to be a competent
vector for B. burgdorferi. Nicholls and Callister (1996), however, report that 5% of
1,184 ticks removed from birds in a Lyme disease-endemic area were B. burgdorferi-
infected H. leporispalustris immatures. Herein we report a 4.2% overall infection
prevalence in H. leporispalustris, resulting largely from infected larval pools, indicating
acquisition of spirochetes from infected avian hosts, or transovarial transmission.
Transovarial transmission is considered negligible for B. burgdorferi (Piesman et
al. 1986; Magnarelli et al. 1987; Patrican 1997) but cannot be ruled out as a mechanism
giving rise to infected larvae in our study. In the absence of local maintenance and
transmission at our site, however, the rate of transovarial transmission from adult ticks
(infected elsewhere) to larval ticks picked up by local birds would have to be
165
extraordinarily common to explain the frequent detection of infected larvae in our study
(58 larval pools removed from 58 different birds). The diversity of avian hosts able to
maintain B. burgdorferi infection and infect ticks underscores the complex transmission
dynamics of these pathogens. Many species of birds have been proposed as competent B.
burgdorferi reservoir hosts using laboratory studies and natural xenodiagnostic
investigations (Anderson et al. 1990; Rand et al. 1998; Richter et al. 2000; Ginsberg et al.
2005). Using natural xenodiagnosis, we implicate 18 avian species as competent
reservoirs for B. burgdorferi, including 10 species which have not before been reported:
Brown-headed Cowbird, Blue Jay, Eastern Towhee, Fox Sparrow, Magnolia Warbler,
Slate-colored Junco, Swainson’s Thrush, Tufted Titrnouse, White—throated Sparrow, and
Yellow-breasted Chat. Given that infective birds were found both during the summer
breeding months and fall migratory months, we posit that migration of hosts from sites of
cryptic transmission may provide a mechanism for dispersal of cryptic B. burgdorferi to
new sites. Conversely, Ogden et al. (2008) detected little evidence of infectious reservoir
hosts among the Spring migratory birds arriving in Canada, as no larval I. scapularis
ticks in their study were found to be infected with B. burgdorfiari. This may result from a
waned infection that was contracted during the previous year’s nymphal season, or, lack
of exposure of these birds to infectious ticks.
In this cryptic cycle, the prevalence of B. burgdorferi in bird-associated ticks is
quite low, with an overall infection prevalence of 3.5% of ticks/larval pools. This raises
questions as to how the cycle can be maintained given this low prevalence, compared to
the relatively higher rates of infection in classic transmission cycles involving 1.
scapularis. Adjusting the data to reflect infection in only I. dentatus nymphs removed
166
from only those avian hosts defined by our study as being reservoir competent (see
above), the minimum infection prevalence is 5.1%. Infection of adult I. dentatus feeding
on rabbits makes maintenance of this pathogen more plausible; in our study, a minimum
of 20% of rabbits were infected based on the conservative index of infected ear tissue,
and a minimum of 28.6% of rabbits were associated with B. burgdorferi-infected ticks.
Presence of other Borrelia spp. spirochetes. The presence of B. andersonii and B.
miyamotoi at the cryptic site enriches our understanding of the reservoir hosts and vector
species that may be important for maintenance and transmission of these spirochetes.
Neither B. andersonii nor B. miyamotoi have been associated with human disease (though
B. andersonii has been detected in an I. dentatus tick removed from a human in
Connecticut (Anderson et al. 1996). The presence of these pathogens at our study site is
of uncertain epidemiological importance. Twelve unique IGS strains of B. andersonii
were present among bird and rabbit-associated ticks and rabbit cars at the cryptic site.
The amount of evolutionary distance that separates B. andersonii strains is much greater
than that which separates most cryptic B. burgdorferi strains, suggesting that the duration
of establishment of the B. andersonii maintenance system is longer than that of B.
burgdorferi, and that I. dentatus (if required for pathogen maintenance) may be an
endemic tick species in this area.
We report on the first detections of B. miyamotoi in ticks removed from passerine
birds. B. miyamotoi was first detected in North America in I. scapularis ticks (Scoles et
al. 2004) and was originally described in I. persulcatus ticks in Japan (Fukunaga et al.
1995). A B. miyamotoi spirochete has also recently been detected in tissues and blood
167
from Wild Turkey in Tennessee (Scott et al., in review». A definitive wildlife reservoir,
however, has not yet been identified. We suggest the Northern Cardinal may be
important in the life history of this species due to the significant association of infected
ticks with this species (11 of 15 infected tick pools were derived from 11 separate
Northern Cardinals). Northern Cardinals are a permanent resident in Michigan, and most
cardinals stay within 8 km of where they were raised (Brewer et al. 1991). While all the
birds carrying B. miyamotoi -infected ticks were captured during the fall migratory
period, it is likely that all hatch year cardinals with infected ticks were born on site
(analysis of capture records indicates that 7 hatch year cardinals associated were each
captured two to six times (average 3.8) between 2 August and 13 November 2007 - the
last day of banding for the season). Conversely, two Hermit Thrushes were carrying
infected ticks. As Hermit Thurshes are not known to breed in the area, the timing of our
findings suggests that migratory birds from the north may have brought B. miyamotoi to
the study site. Similarly, we found one American Robin that carried infected ticks in the
fall. This species in Michigan is typically a migrant, though a small proportion of birds
will overwinter in the southern counties. We thus have evidence to support both local
and migratory contributions to B. miyamotoi maintenance at Pitsfield. The pattern in our
data in which a majority of infected ticks were larvae may parallel the finding of Barbour
et al. (2009), in which B. miyamotoi infection in mouse hosts rose toward the end of the
summer, coincident with the larval phenology (and in this vertically-transmitted system,
larvae may infect hosts). Similarly, Scott et al. (in review) found that B. miyamotoi-
positive turkeys were more likely to be infested with larval as opposed to nymphal ticks.
In comparison to the high strain diversity observed within B. burgdorferi (see below), the
168
intraspecific diversity of the B. miyamotoi spirochetes at the IGS locus was much less, as
we found only two strains. Only a single strain -— the same strain that infected 91.2% of
our B. miyamotoi samples - was present within ticks in Connecticut (Bunikis et al.
200413).
Bird tick phenology and implications for Borrelia species maintenance. Larval I.
dentatus exhibited bimodal activity peaks in Spring/early summer and again in the fall,
with nymphs active in the spring thought mid-summer, similar to the findings of
Sonenshine and Stout (1970), Battaly et al. (1987), and Kollars and Oliver (2003). From
the conventional perspective of B. burgdorferi maintenance by I. scapularis in the
northeastern US, we expect that nymphal activity should precede larval activity, such that
nymphs can infect the hosts prior to their being fed upon by the larvae. At our study site,
I. dentatus nymphal activity clearly precedes the fall peak in larval activity, but nymphs
and the early larvae are simultaneously active. This non-optimal phenology, from the
perspective of the pathogen, may partially explain its low prevalence. However, more
synchronous feeding nymphal and larval I. scapularis has been observed in the Midwest
(Gatewood et a1. 2009)(Hamer et a1. 2010), with similar infection prevalences as found in
the Northeast. The longevity of birds, in comparison to mice, may be advantageous to
the pathogen in a cryptic scenario (though the duration of infectiousness of B. burgdorferi
in infected birds is unknown), as a bird infected by a nymph in the previous year may
infect a larva in the spring in the absence of infectious nymphs. Given that peak infection
prevalence in ticks for both B. burgdorferi and B. andersonii was toward the beginning of
the season in May-July, coincident with the synchronous activity of larvae and nymphs,
169
we suggest that co-feeding of larval and nymphal I. dentatus on the same avian host may
be important in pathogen transmission. Co-feeding was observed in 116 cases, 73 of
which occurred during weeks 20-29 of the spring/early summer tick activity peak, thus
offering a mechanism for Borrelia spp. maintenance. There are three birds in our dataset
from which we simultaneously removed both B. burgdorferi—positive larvae and nymphs.
Furthermore, birds have been shown to reactivate latent B. burgdorferi infection in the
face of migratory stress (Gylfe et al. 2000), and this may contribute to the infection in
spring and early summer larvae that we observed.
Strain diversity of B. burgdorferi. We detected an unprecedented level of diversity at
the 16S-23S ribosomal intergenic spacer of cryptic B. burgorferi within host and tick
populations not classically expected to be infected. Rarefaction analysis suggests that our
sampled richness reflects only a small proportion of true strain richness. Intergenic
spacer regions generally accumulate higher degrees of sequence variation between related
species than do coding regions, because spacers do not produce a functional gene product
and are free of selective constraint (with the exception of a tRNA gene that occurs within
the IGS of Borrelia species). Furthermore, because no detectable recombination is found
within the IGS (which is chromosomally-located), in contrast to other B. burgdorferi
typing schemes (such as ospC , which is located on a plasmid), IGS is a sensitive marker
of evolutionary change (Liveris et al. 1995), and alone is an efficacious genetic marker to
differentiate among strains of B. burgdorferi (Bunikis et al. 2004a).
At least three evolutionary mechanisms may contribute to the pattern of genetic
diversity we observed within the cryptic Borrelia, which is characterized by a small
170
number of ubiquitous strains that are also found in classic transmission at Lyme disease-
endemic sites (present within more than one individual in our sampled population), and a
larger number of indigenous strains that are novel mutants (present as singletons in our
sampled population). (1) Ecological opportunity is present by the invasion of a new
vector (1. dentatus), new host species (birds/rabbits) or new geographical regions
(Pitsfield) by B. burgdorferi, after which the one or few invading genotypes become the
most common. Among others, strain IGS 2D was likely a founder at Pitsfield, as it is the
most common strain, comprising 40.3% of all sequences, and is present in ticks removed
from both birds and rabbits. This bottleneck in the B. burgdorferi population, represented
by the presence of only a few founders, is followed by clonal expansion and
diversification of founders. This diversification is evident as a star phylogeny in the
MST, in which each novel strain is derived independently from its common ancestor.
Because of the nature of cryptic transmission, these novel variants are permitted within
birds, rabbits, and I. dentatus, but are not necessarily exposed to the mammalian immune
system to the same extent as would be newly-evolved strains in classic transmission
scenarios. Thus, strains that may otherwise have been ‘purged’ by a host immune system
in classic transmission are instead present in cryptic transmission. (2) Multiple-niche
polymorphism, also known as diversifying selection, can maintain diversity within the
population when the environment is heterogenous and no single genotype has the highest
fitness in all environments (Brisson and Dykhuizen 2004). This selection does not occur
on the IGS region itself, but on other areas of the genome which are linked to the IGS
region such as ospC. The environment experienced by cryptic B. burgdorferi, including
1. dentatus, rabbits, and at least 12 species of birds in our study, may therefore influence
171
the frequency with which we observed genotypes. For example, that strain IGS 2D was
so common at Pitsfield may reflect selection for this strain at this site in avian hosts and
the I. dentatus vector, or, its selection within classic transmission cycles (it is among the
most common strains across the Midwest) and spillover into this cryptic cycle. Host
associations, facilitated by selective killing of strains by host-specific innate immune
factors (Kurtenbach et al. 2002), have been documented as a factor leading to the
diversifying selection of Lyme disease spirochetes, at both the inter-species level (e. g. in
Europe, B. burgdorferi and B. afzelii are mammal-associated, whereas B. garim'i and B.
valaisiana are bird-associated (Dubska et al. 2009)) and at the intraspecies level of B.
burgdorfieri (Brisson and Dykhuizen 2004; Hanincova et al. 2006). (3) Negative
frequency-dependant balancing selection has been described for B. burgdorferi, in which
antigenic types that have already infected a host cannot establish a subsequent infection
in the same host because of the immune response, and therefore rare antigenic types are
more likely to establish infection in hosts infected with multiple strains, leading to high
levels of variation (Wang et al. 1999). Balancing selection acts not only on the ospC
locus, but on closely linked loci (Brisson and Dykhuizen 2004).
Given the low overall prevalence of B. burgdorferi infection in this cryptic
system, the degree to which we identified mixed strain infections (10% of all infected
samples) was surprisingly high. We recently identified 4% of adult I. scapularis from
Van Buren State Park within the recently-invaded area in Michigan’s Lower Peninsula as
harboring mixed-strain infections (unpublished data). (Furthermore, adults are likely to
have higher levels of mixing as compared to nymphs, given the additional bloodmeal
they have consumed). Thus, our encountering of mixed infections at 2.5 the prevalence
172
in the cryptic zone suggests that the degree of interaction among infected ticks and hosts
is high, and that hosts are being exposed to multiple strains. Mixed strain infections are
generally expected to occur in highest frequency in sites with well-established
transmission as opposed to recently-colonized foci with many na'r‘ve hosts and a reduced
diversity of strains due to invasion by one or a few founding organisms strains.
What is the epidemiological significance of a cryptic cycle? The epidemiological
significance of a cryptic cycle is dictated by a combination of the following five
parameters: (1) the infectivity of cryptic strains to humans; (2) the frequency at which
infected cryptic vectors directly feed on humans; (3) the presence or imminent invasion
of a bridging vector; (4) the frequencies at which the cryptic vector and bridging vector
share hosts; and/or (5) the extent to which the cryptic B. burgdorferi introduced into the
bridging vector serves as additive infection, above the level of infection that invading I.
scapularis may bring to the site, or the extent to which the cryptic B. burgdorferi strains
replace strains that the bridging vector circulates
(1) Two main genetic typing schemes are useful in assessing virulence of different
B. burgdorferi strains in humans: RST (a tripartite classification of the 25+ IGS strains)
and outer surface protein C (ospC; an antigenic plasmid-borne gene); ospC is in positive
linkage disequilibrium with IGS (Bunikis et al. 2004a). Some RST 1 strains are
associated with a higher frequency of disseminated infection in humans and more
invasive disease in experimental animals (Seinost et al. 1999; Wang et al. 2002;
Derdakova et al. 2004; Hanincova et al. 2008; Worrnser et al. 2008), and a bias toward
the relatively less invasive RST 2 and 3 strains has been found among infected I.
173
scapularis in the Midwest (Gatewood et al. 2009). At our cryptic site, we document the
presence of two RST 1 strains (IGS 1A and ‘Novel PP’) which alone underscores the
potential epidemiological significance of this cryptic cycle.
The pathogenicity in humans of the large number of novel IGS mutant strains is
unknown, though we have no a priori reason to suggest that single or double nucleotide
polymorphisms within the non-coding intergenic spacer region would necessarily
correspond to a change in virulence. To begin to assess the epidemiological risk
associated with novel strains found within this cryptic cycle, we sent total DNA from 11
of the samples with novel IGS types (Novels NN, O, 00, P, PP, R, S, T, V, W and X; all
were derived from ticks removed from birds) to a laboratory at the University of
California-Irvine for direct ospC typing (B. Travinsky and A. Barbour, unpublished data).
Of these, ospC was successfully amplified and sequenced from 10 samples, resulting in
detection of ospC A, 13, and K, which represent three of the four ospC major groups
associated with disseminated human Lyme disease (Seinost et a1. 1999). F urthennore,
many mixed strain infections were found at the ospC locus. Given that no evidence of
mixing was apparent in the portion of IGS that we scrutinized from these samples, this
suggests that ospC may more readily be mixed, and further underscores the complexity of
the pathogen diversity maintained at this site.
(2) Documented human-biting in I. dentatus is rare, yet has been reported a total
of 13 people across Michigan (Walker et al. 1992), North Carolina (Harrison et al. 1997)
West Virginia (Hall et al. 1991), Washington D. C. (Sollers 1955), Connecticut
(Anderson et al. 1996) and Maine (Keirans and Lacombe 1998).
174
(3) I. scapularis is invading from the west (Hamer et al. 2010). Pitsfield Banding
Station is on the trajectory for continued invasion, and has habitat features predicted to
support I. scapularis populations should they become introduced (Guerra et al. 2002).
(4) We document I. dentatus feeding on two meadow jumping mice of a total of
three captured at Pitsfield. This represents an expansion of the documented host range for
this tick species, and may afford the opportunity for small mammals to become infected
with cryptic Borrelia spp. (though the two mouse-infesting I. dentatus were both
uninfected larvae in our study). Similarly, an I. dentatus larva was found on a white-
footed mouse in Tennessee (Kollars 1992). As mice are key hosts for the bridging vector
1. scapularis, such adventitious feeding events by the cryptic vector on atypical hosts may
offer a mechanism for transmission of cryptic strains to bridging vectors and
subsequently, humans. In addition to atypical host utilization by the cryptic vector as a
mechanism for merging cryptic and classic transmission cycles, the bridge vector may
feed directly on cryptically-infected hosts. The frequency at which I. scapularis feeds on
rabbits in endemic areas is not widely studied. Telford and Spielman (1989b) found that
41, 38, 3% of rabbits were infested with I. scapularis larvae, nymphs, and adults,
respectively, and that 85, 62, and 85% were infested with the equivalent life stages of I.
dentatus. Thus, there exists potential for introduction of cryptic Borrelia into I.
scapularis populations through direct feeding by the bridge vector on cryptically-infected
wildlife.
(5) It is currently unknown whether the infection prevalence in invading I.
scapularis will be elevated due to the existence of cryptic Borrelia in the region
undergoing invasion. In southwest Michigan, B. burgdorferi was found within recently
175
invaded I. scapularis populations, and it is plausible that both I. scapularis and B.
burgdorferi may invade simultaneously (Hamer et al. 2010). Upon invasion into a zone
of cryptic transmission, the frequency with which I. scapularis feeds on infectious hosts
will likely dictate the additive nature of cryptic infection. The degree to which invading
strains may replace cryptic strains, and how host associations of strains may affect their
persistence and the blending of the invading and cryptic cycles is unknown.
We hypothesize that the presence of B. burgdorferi in birds and rabbits, which
precedes the arrival of I. scapularis, may reduce the time lag between I. scapularis
invasion and build-up of infection prevalence in I. scapularis. Thus, mobile infected
birds and cryptic cycles may have the potential to accelerate the increase in human
disease risk within an invasion zone, or, facilitate the establishment phase of invasion.
Infection of birds from cryptic cycles may then allow bird dispersal of B. burgdorferi
strains such that the movement and maintenance of B. burgdorferi is un-linked from that
of the main bridging vector. Further studies are needed to understand the role of birds in
generating or maintaining genetic diversity of B. burgdorferi, as well as B. andersonii
and B. miyamotoi, and the consequences for infection of humans or canines with these
bird-associated strains.
176
Table 5.1a. Infestation prevalences of all parasitized bird species with three species of
tick, Pitsfield Banding Station, 2004-2007. Some ticks were observed on birds yet not
removed and are categorized as unknown identity. Tick species-specific infestation
prevalences therefore underestimate true infestation. Additionally, a single Song Sparrow
harbored 6 Dermacentor variabilis larvae.
177
’- I. Ctr-infestation Un-
Common name Total Overall dentatus H. lep. scapularis I. dent, H. Iep Identified
American Goldfinch 1398 0.9 0.6 0.7 0 1
American Redstart 222 1.0 1,0 '
American Robin' 566 34.9 26.5 1.9 1,6 8.0
Black and White Warbler 60 1.7 1,7
Black-capped Chickadee 716 3.9 3.6 0.1 Q1 03
Blue Jay' 109 28.4 2.2 33,
Brown Thrasher‘i' 104 46.2 44.2 11.5 1.6 1.0
Brown-headed Cowbird' 29 13.8 13.8
Canada Warbler 42 7.1 7,1
Carolina Wren‘ 21 42.9 28.6 19.5 4,3
Chestnut-sided Warbler'l' 114
Common Grackle 159 27.4 2.8 1.3 0.6 5]
Common Yellowthroat‘ 326 7.4 3.7 5.0 0.6
Connecticut Warbler 9 11.1 1 1.1
Downy Woodpecker 212 1.4 1.4
Eastem Bluebird 31 3.2 3.2
Eastern Towhee‘ 75 42.7 29.3 17.3 6.7 2.7
E. White-crowned Sparrow 53 22.6 17.0 13.3 7.5
Field Sparrow 196 9.7 7.7 1.5 0.5 1.2
Fox Sparrow" 94 27.7 25.5 3.2 2.1 1.6
Golden-crowned Kinglet 249 0.4 0.4
Gray Catbird'i' 2775 4.9 3.3 2.1 0.3 0.2
Gray-checked Thrush 58 19.0 13.8 1.3 5.2
Hairy Woodpecker 22 4.5 4.5
Hermit Thrush'i' 563 35.2 26.6 2.8 0.2 1.8 7.5
House Finch 252 1.6 1.6
House Wren't 135 31.1 14.8 23.0 6.7
Indigo Bunting 64 3.1 1.6 1.6 1.6 1.6
Lincoln’s Sparrow 24 37.5 37.5 12.5 12.5
Magnolia Warbler‘ 777 0.8 0.5 0.3
Mourning Warbler 54 5.6 5.6
Myrtle Warbler 1665 3.5 2.9 0.2 0.7 0.5
Nashvlle Warbler 428 1.6 0.9 0.8
Northem Cardinal‘ 768 27.9 22.9 4.0 1.8 2.9
Northern Waterthnrsh 27 3.7 3.7
Orange-crowned Warbler 57 1.8 1.8
Ovenbird' 384 3.1 0.8 2.3
Purple Finch 64 7.8 6.3 1.6
Red-breasted Nuthatch 25 4.0 4.0
Ruby-crowned Kinglet 660 1.7 1.4 0.3
Slate-colored Junoo‘ 466 7.8 5.2 2.1 0.9 06
Song Sparrow'rt 500 39.0 28.4 13.0 5.6 3.2
Swainson's Thrush’ 364 13.2 3.6 1.2 0.3 0.5
Swamp Sparrowt 37 18.9 1.8 2.7 5-5
Tennessee Warbler 790 0.4 0.1 0.3
Tufted Titmouse‘f 302 19.9 16.6 3-3
Veery 22 22.7 13.6 10.0
White-breasted Nuthatch 86 1.2 1.2
White-throated Sparrow“ 1159 4.8 29.9 4.6 2.2 9-6
Wilson's Warbler 165 0.7 0-7
Winter Wren 29 24.1 24.1
Wood Thrush 209 8.6 6.7 2.4 0-5
Yellow Warbler 187 0.5 0.5
Yellow-breasted Chat“ 6 33.3 33.3
‘ designated as competent reservoir for B. burgdorferi
T designated as competent reservoir for B. andersoni
1 a single Song Sparrow harbored 6 D. variabilis larvae
and a single House Wren harbored one D. variabilis adult
178
Table 5.1b. Bird species investigated for ticks and found to be uninfested, Pistfield
Banding Station, 2004-2007.
‘
—*
Common name Total captured
Acadian Flycatcher 5
American Woodcock 5
American Tree Sparrow 45
Baltimore Oriole 28
Black-billed Cuckoo 2
Bay-breasted Warbler 44
Blue-gray Gnatcatcher 4
Blue-headed Vireo 37
Blackbumian Warbler 9
Blackpoll Warbler 56
Brown Creeper 42
Brewster's Warbler 1
Black-throated Blue Warbler 46
Black-throated Green Warbler 54
Blue-winged Warbler 93
Cedar Waxwing 157
Cemlean Warbler 1
Chipping Sparrow 20
Cape May Warbler 9
Cooper's Hawk 2
Eastern Kingbird 1
Eastern Phoebe 48
Eastern Wood-Pewee 70
Great Crested Flycatcher 2
Golden-winged Warbler 9
House Sparrow 2
Hooded Warbler 5
Least Flycatcher 40
Mourning Dove 19
Northern Parula 1
Orchard Oriole 1
Philadelphia Wreo 3
Prairie Warbler 2
Prothonotary Warbler 1
Rose-breasted Grosbeak 141
Red-bellied Woodpecker 34
Red-eyed \fireo 73
Ruby-throated Hummingbird 319
Red-winged Blackbird 4
Scarlet Tanager 36
Sharp-shinned Hawk 15
Tree Swallow 2
Traill's Flycatcher 63
Virginia's Warbler 1
Warbling Vireo 38
White-eyed Vireo 12
Western Palm Warbler 13
Yellow-billed Cuckoo 9
Yellow-bellied Flycatcher 57
Yellow-shafted Flicker 36
Yellow-throated Vireo 5
179
3a 8 3 8. mm 4.8 E :8 0% 2. 55m co>
on on on o no 3 5 Na o 8 233K
at cocoa masons. d c.5388 .2 vegan: 3% 86.9 nascent. d 6.538» .2 8&8 new
85.655 .02 .
.02 85.99.”. .oz
#985 6m JOE Sesame.
oozes 8.8.722;
oz
x9665 ..mm axy c8982.
xSEQEo Ecummm
..EetbehMSQ .m 5? ooze—«>05 comaoomam use 68mm“ 86.35 .86 mo 3% 295m .3 336:8 mm 83% mo 866% 93 5S, coco—«>89
53833 38 623%... manage me on? cashew .Eommxm .«0 “$3 Ex ca uteEmeea. .N cougamtm—Eoooc com “cameo: m 41.3 88m
eon—m §> use aoufim magnum “Em Bowman 8 3:805 rebehwaae .m use 4.253653. .~ com £0538 me flea—EBB =eEm .N.m 635.
180
mud 5m mm.m Nam Rd and a; $9 $6QO :4.
52 m on N 6.562388 .2
mm? mmh 3 F mmd mad «on 65922632862 .1
and 22m 8.2V mom mad Etc mod var 22262ch .2
202053.28 .m .Eoflobcm .m 2.2.3.2893 .m B59 .02 2o2oEmA2E .m 2208695 .m totonmcno .m “22mm: .02
B2925 =26an 6908:. Enema.
e235»: goon Eta...
.mcowofimn =e com gamma: one? 0238— M252??? .Q 0 mo Boa 6. EB 6.582.623 .N 295m :3? 635m e .33: BEES 8 H2256?“
5 doufim wfiecwm Eocmzm Koomtvoom .mcowofimm 6.23.2er Eocotmc m 2.23 coco—«>65 eczema: owmoommtmomocmm x02. .m.m 636,—.
181
Figure 5.1. Pitsfield Bird Banding Station in Vicksburg, Ml (blue triangle) and Van
Buren State Park, the site of comparative small mammal sampling (red circle). Shading
indicates counties with documented established populations of blacklegged ticks at the
time the study started in 2004 (dark gray) and at the time the study ended in 2007 (light
gray), and the endemic population (diagonal lines; Hamer et al. 2010). H. lep. = H.
leporispalustris; I. dent. = I. dentatus.
0 30 60 kilometers
LLLLLLLIJ
182
62 23200 Eon—226.com , «6:92 2:32. 6:3... >22
2. -.2nal..-..-.-:.n...m:l ! 000 ow
.. S o «w m.
1. u
a
- 8.0 P M
_ J
. D.
., 0N0 S
mt2m32mqmtoqo2 .I 3 and
>02 conouoO ..onEoEom 33914 d
cod 1.
O
. I.d
1.. n.
u a I.
02. o S O
1. u
. a
9. o D. a.
“D
8.0 m.
caE>z I 3.226222% .2 Am 63
mam.‘ D
Agnew—QB me. So» mamm: SEE Br .20 note 6.235% one when 8:3 £59 688%: no 309895
:38 >23? mm @8033 Awmhnéemmtemfl E .9 522826226 .N .3 620: BHBOmmeE E23823 use Ete— ao awe—0:25 .N.m oSwE
183
Figure 5.3. Temporal variation in infection prevalence of B. burgdorferi, B. andersonii,
and B. miyamotoi from May-November, 2004-2007. Monthly mean infection
prevalences across all years (:t SE of mean) is plotted. All tick species are aggregated.
I B. burgdorferi
0.20 l I B. andersoni
" l ' B. m‘ amoto’
2
a
> 0.12 -
2 . n l
n.
l: 0.08 1
o
3 004’ ’ DE EL. like at.
c
- ..L
0.00 ‘1“ “—fit" -’—l'“ *‘rr '-‘-TT ”T "TH
May June July August September October November
184
Figure 5.4. Neighbor-joining phylogenetic tree and frequency distribution of B.
burgdorferi IGS haplotypes collected from Pitsfield Banding Station, 2004-2007. The
percentages of replicate trees in which the associated taxa clustered together in the
bootstrap test (1000 replicates) are shown next to the branches. Indigenous strains (novel
IGS mutants not previously reported) are indicated with a blue triangle; ubiquitous strains
(haplotypes previously reported in Bunikis et al. (2004a) or detected in I. scapularis
across the Midwest) are indicated with a red circle. Sequences conforming to the 16S —
23S RST group 1, 2, and 3 designations by the criteria of Liveris et al. (1995) are
demarcated by the labels at the top left of shaded or unshaded groups. Total samples size
of each strain is indicated at the end of each branch label, of which a majority is from
bird-derived ticks; the 7 mammal-associated samples (from rabbit ticks and ear and small
mammal ears) are denoted in parentheses.
RST 2
IGS 20 and 8 associated strains
. 29(2) 8 strainsAall n=1
Midwest A. 2(1)
118
‘51
Midwest r 4211
Midwest 001,}; 1
Midwest F 1
22 71 L IGS 5 .1.
Midwest v A 1
Midwest MQ1
54% Midwest “10(1)
Midwest HH A1
90—631 Midwest 8. 10(1)
IGS 71117301
Midwest ll 161(1)
Midwest w Q1
73 simmer
* i 7 Midwest PPA 1
RST 3
29
70
40
55
her 1
99L les1A01
0.005
185
Figure 5.5. Minimum spanning tree (MST) of B. burgdorferi IGS haplotypes collected
from Pitsfield Banding Station, 2004-2007. Each black circular node connecting
haplotypes represents one mutational change; a 7 base pair indel that occurs within the
IGS is considered as one change. The size of each haplotype is proportional to its
frequency within the sampled population. The rectangles represent the haplotypes with
the highest outgroup weight, which correlates with haplotype age. The two haplotpyes
within RST 1 (IGS 1A and Novel PP) are linked to each other by two mutational
changes, yet unlinked to the rest of the network. Indigenous strains (novel IGS mutants
not previously reported) are indicated with a blue triangle; ubiquitous strains (haplotypes
previously reported by Bunikis et al. (2004a) or detected in 1. scapularis across the
Midwest) are indicated with a red circle.
IGS 20
KK (£2) P '—*'
A A A A I A
O
a A
mo
A A
I amt: Mldweet B
186
Figure 5.6. Rarefaction curve for B. burgdorferi strains derived from Pitsfield Banding
Station, 2004-2007. Mean and standard deviations of the number of strains found in
subsamples are shown.
No. different strains
30
A
51 r
201 {iii
5; if}
101I {1’
l {l
5‘;
0 20 40 60 80
No. B. burgdorferi sequences determined
187
Figure 5.7. Neighbor-joining phylogenetic tree and frequency distribution of B.
andersoniii IGS haplotypes collected from Pitsfield Banding Station, 2004-2007. The
percentages of replicate trees in which the associated taxa clustered together in the
bootstrap test (1000 replicates) are shown next to the branches. Total samples size of
each strain is indicated at the end of each branch label, of which a majority is fiom bird-
derived ticks; the 7 rabbit-associated samples (from rabbit ticks and a rabbit car) are
denoted in parentheses.
99 {Michigan Bird A
47
I 9"Michigan Bird B
22 *
64 flMichigan Bird C
23 99
L—Michigan Bird 0
14_ #:Michigan Bird E
79 99
fiMichlgan Bird F
-— 57 —1Michigan Bird G
99
Michigan Bird H
9i.
54‘ .Michigan Bird I
62 fiMichigan Bird J
97 100
aMichigan Bird K
92 fiMichigan Bird L
100
'r 9.
0.005
188
References
Anderson J F (1989). Epizootiology of Borrelia in Ixodes Tick Vectors and Reservoir
Hosts. Reviews of Infectious Diseases 1 1:S1451-Sl459.
Anderson JF, and Magnarelli LA (1984). Avian and Mammalian Hosts for Spirochete-
Infected Ticks and Insects in a Lyme-Disease Focus in Connecticut. Yale Journal
of Biology and Medicine 57:627-641.
Anderson JF, Magnarelli LA, and Stafford KC (1990). Bird-Feeding Ticks Transstadially
Transmit Borrelia-Burgdorferi That Infect Syrian-Hamsters. Journal of Wildlife
Diseases 26: 1-10.
Bacon RM, Kugeler KJ, Griffith KS, and Mead PS (2007). Lyme disease - United States,
2003-2005 (Reprinted from MMWR, vol 56, pg 573-576, 2007). Jama-Journal of
the American Medical Association 298:2 78-279.
Barbour AG, Bunikis J, Travinsky B, Hoen AG, Diuk-Wasser MA, Fish D, et al. (2009).
Niche Partitioning of Borrelia burgdorferi and Borrelia miyamotoi in the Same
Tick Vector and Mammalian Reservoir Species. Am J T rap Med Hyg 81:1120-
1131.
Battaly GR, and Fish D (1993). Relative Importance of Bird Species as Hosts for
Immature Ixodes-Dammini (Acari, Ixodidae) in a Suburban Residential
Landscape of Southern New-York-State. Journal of Medical Entomology 30:740-
747.
Battaly GR, Fish D, and Dowler RC (1987). The Seasonal Occurrence of Ixodes-
Dammini and Ixodes-Dentatus (Acari, Ixodidae) on Birds in a Lyme-Disease
Endemic Area of Southeastern New-York-State. Journal of the New York
Entomological Society 95:461-468.
Brewer R, McPeek GA, and Adams J, Raymond A., editors. .(1991). The Atlas of
Breeding Birds of Michigan. Michigan State Umversrty Press, East Lamsng.
189
Brinkerhoff RJ, Folsom-O'Keefe CM, Tsao K, and Diuk-Wasser MA (2010). Do birds
affect Lyme disease risk? Range expansion of the vector-borne pathogen Borrelia
burgdorferi. Frontiers in Ecology and the Environment 0.
Brisson D, and Dykhuizen DE (2004). ospC diversity in Borrelia burgdorferi: Different
hosts are different niches. Genetics 168:713-722.
Brown RN, and Lane RS (1992). Lyme-Disease in California - a Novel Enzootic
Transmission Cycle of Borrelia-Burgdorferi. Science 256: 1439-1442.
Bunikis J, Garpmo U, Tsao J, Berglund J, Fish D, and Barbour AG (2004a). Sequence
typing reveals extensive strain diversity of the Lyme borreliosis agents Borrelia
burgdorferi in North America and Borrelia afzelii in Europe. Microbiology-ng
150:1741-1755.
Btmikis J, Tsao J, Garpmo U, Berglund J, Fish D, and Barbour AG (2004b). Typing of
Borrelia relapsing fever group strains. Emerging Infectious Diseases 10:1661-
1664.
Chao A, and Tsung-Jen S (2003). Nonparametric estimation of Shannon’s index of
diversity when there are unseen species in sample. Environmental and Ecological
Statistics 10:429-443.
Clement M, Posada D, and Crandall KA (2000). TCS: a computer program to estimate
gene genealogies. Molecular Ecology 9: 1657-1659.
Clifford CM, Sonenshine DE, Atwood EL, Robbins CS, and Hughes LE (1969). Tests on
Ticks from Wild Birds Collected in the Eastern United States for Rickettsiae and
Viruses. Am J Trop Med Hyg 18:1057-1061.
Cooney JC, and Burgdorfer W (1974). Zoonotic Potential (Rocky Mountain Spotted
Fever and Tularemia) in the Tennessee Valley Region: I. Ecologic Studies of
Ticks Infesting Mammals in Land Between the Lakes. Am J Trop Med Hyg 23:99-
108.
Derdakova M, Dudioak V, Brei B, Brownstein J S, Schwartz I, and Fish D (2004).
Interaction and transmission of two Borrelia burgdorferi sensu stricto strains in a
tick-rodent maintenance system. Applied and Environmental Microbiology
70:6783-6788.
Dubska L, Literak I, Kocianova E, Taragelova V, and Sychra O (2009). Differential Role
of Passerine Birds in Distribution of Borrelia Spirochetes, Based on Data from
Ticks Collected from Birds during the Postbreeding Migration Period in Central
Europe. Applied and Environmental Microbiology 75:596-602.
Durden LA, and Keirans JE (1996). Nymphs of the genus Ixodes (A cari: Ixodidae) of the
United States: taxonomy, identification key, distribution, hosts, and
190
medical/veterinary importance. Thomas Say Publications in Entomology,
Entomological Society of America
Foster ES (2004). Ixodes scapularis (Acari: Ixodidae) and Borrelia burgdorferi in
southwest Michigan: population ecology and verification of a geographic risk
model. Masters Thesis, Michigan State University, East Lansing.
F ukunaga M, Takahashi Y, Tsuruta Y, Matsushita 0, Ralph D, McClelland M, et al.
(1995). Genetic and phenotypic analysis of Borrelia miyamotoi sp. nov., isolated
from the ixodid tick Ixodes persulcatus, the vector for Lyme disease in Japan. Int
J Syst Bacteriol 45:804-810.
Gatewood AG, Liebman KA, Vourc'h G, Bunikis J, Hamer SA, Cortinas R, et al. (2009).
Climate and Tick Seasonality Are Predictors of Borrelia burgdorferi Genotype
Distribution. Applied and Environmental Microbiology 75:2476-2483.
Giardina AR, Schmidt KA, Schauber EM, and Ostfeld RS (2000). Modeling the role of
songbirds and rodents in the ecology of Lyme disease. Canadian Journal of
Zoology-Revue Canadienne De Zoologie 78 :2184-2197.
Ginsberg HS, Buckley PA, Balmforth MG, Zhioua E, Mitra S, and Buckley F G (2005).
Reservoir competence of native north American birds for the Lyme disease
spirochete, Borrelia burgdorferi. Journal of Medical Entomology 42:445-449.
Guerra M, Walker E, Jones C, Paskewitz S, Cortinas MR, Stancil A, et al. (2002).
Predicting the risk of Lyme disease: habitat suitability for Ixodes scapularis in the
north central United States. Emerg Infect Dis 8:289-297.
Gylfe A, Bergstrom S, Lunstrom J, and Olsen B (2000). Epidemiology - Reactivation of
Borrelia infection in birds. Nature 403:724-725.
Hall JE, Armine Jr. JW, Gais RD, Lp;anko VF, Hagenbuch BE, Gerencser VF, et al.
(1991). Parasitization of humans in West Virginia by Ixodes cookei (Acari:
Ixodidae), a potential vector of Lyme borreliosis. Journal of Medical Entomology
28: 186-189.
Hamer SA, Roy PL, Hickling GJ, Walker ED, Foster ES, Barber CC, et al. (2007).
Zoonotic pathogens in Ixodes scapularis, Michigan. Emerging Infectious Diseases
13:1131-1133.
Harrier SA, Tsao JI, Walker ED, and Hickling GJ (2010). Invasion of the Lyme disease
vector Ixodes scapularis: implications for Borrelia burgdorferi endenricity.
EcoHealth.
Hanincova K, Kurtenbach K, Diuk-Wasser M, Brei B, and Fish D (2006). Epidemic
spread of Lyme borreliosis, northeastern United States. Emerging Infectious
Diseases 12:604-611. '
191
Hanincova K, Ogden NH, Diuk-Wasser M, Pappas CJ, Iyer R, Fish D, et al. (2008).
Fitness variation of Borrelia burgdorferi sensu stricto strains in mice. Applied and
Environmental Microbiology 74: 153-157.
Harrison BA, Engber BR, and Apperson CS (1997). Ticks (Acari: Ixodidae) uncommonly
found biting humans in North Carolina. Journal of Vector Ecology 22:6-12.
James AM, and Oliver JH (1990). Feeding and Host Preference of Immature Ixodes-
Dammini, I-Scapularis, and I-Pacificus (Acari, Ixodidae). Journal of Medical
Entomology 27:324-330.
Keirans JE, and Clifford CM (1978). The genus Ixodes in the United States: a scanning
electron microsc0pe study and key to the adults. J Med Entomol Suppl 2:1-149.
Keirans JE, and Lacombe EH (1998). First records of Amblyomma americanum, Ixodes
(Ixodes) dentatus, and Ixodes (Ceratixodes) uriae (Acari : Ixodidae) from Maine.
Journal of Parasitology 84:629-631.
Klich M, Lankester MW, and Wu KW (1996). Spring migratory birds (Aves) extend the
northern occurrence of blacklegged tick (Acari: Ixodidae). Journal of Medical
Entomology 33:581-585.
Kollars TM (1992). New Host of Ixodes-Dentatus (Acari, Ixodidae), and lst Record in
Tennessee. Journal of Medical Entomology 29:715-716.
Kollars TM, and Oliver J H (2003). Host associations and seasonal occurrence of
Haemaphysalis leporispalustris, Ixodes brunneus, I.cookie, I-dentatus, and I-
texanus (Acari : Ixodidae) in Southeastern Missouri. Journal of Medical
Entomology 40:103-107.
Kurtenbach K, Schafer SM, Sewell HS, Peacey M, Hoodless A, Nuttall PA, et al. (2002).
Differential survival of Lyme borreliosis spirochetes in ticks that feed on birds.
Infection and Immunity 70:5893-5 895.
Levine JF, Sonenshine DE, Nicholson WL, and Turner RT (1991). Borrelia-Burgdorferi
in Ticks (Acari, Ixodidae) from Coastal Virginia. Journal of Medical Entomology
28:668-674.
Lin T, Oliver JH, and Gao LH (2003). Comparative analysis of Borrelia isolates fi'om
southeastern USA based on randomly amplified polymorphic DNA fingerprint
and 16S ribosomal gene sequence analyses. F ems Microbiology Letters 228:249-
257. '
Liveris D, Gazumyan A, and Schwartz I (1995). Molecular typing of Borrelia burgdorferi
sensu lato by PCR-restriction fragment length polymorphism analysis. J Clin
Morobiol 33:589-595.
192
LoGiudice K, Ostfeld RS, Schmidt KA, and Keesing F (2003). The ecology of infectious
disease: Effects of host diversity and community composition on Lyme disease
risk. Proceedings of the National Academy of Sciences of the United States of
America 100:567-571.
Magnarelli LA, Anderson JF, and Durland F (198 7). Transovarial Transmission of
Borrelia burgdorferi in Ixodes dammini (Acarizlxodidae). The Journal of
Infectious Diseases 156:234-23 6.
Main AJ, and Carey AB (1980). Connecticut virus: A new swgrass group virus from
Ixodes dentatus (Acari: Ixodidae). Journal of Medical Entomology 17:473-476.
Marconi RT, Liveris D, and Schwartz I (1995). Identification of Novel Insertion
Elements, Restriction-Fragment-Length-Polymorphism Patterns, and
Discontinuous 23s Ribosomal-Rna in Lyme-Disease Spirochetes - Phylogenetic
Analyses of Ribosomal-Rna Genes and Their Intergenic Spacers in Borrelia-
Japonica Sp-Nov and Genomic Group-21038 (Borrelia-Andersonii Sp-Nov)
Isolates. Journal of Clinical Microbiology 33:2427-2434.
Maupin GO, Gage KL, Piesman J, Montenieri J, Sviat SL, Vanderzanden L, et al. (1994).
Discovery of an Enzootic Cycle of Borrelia-Burgdorferi in Neotoma-Mexicana
and Ixodes-Spinipalpis from Northern Colorado, an Area Where Lyme-Disease Is
Nonendemic. Journal of Infectious Diseases 170:636-643.
Nicholls TH, and Callister SM (1996). Lyme disease spirochetes in ticks collected from
birds in midwestem United States. Journal of Medical Entomology 33:3 79-3 84.
Ogden NH, Lindsay LR, Hanincova K, Barker IK, Bigras-Poulin M, Charron DF, et al.
(2008). Role of migratory birds in introduction and range expansion of Ixodes
scapularis ticks and of Borrelia burgdorferi and Anaplasma phagocytophilum in
Canada. Applied and Environmental Microbiology 74:1780-1790.
Oliver JH, Jr.,Chandler FW, Jr., James AM, Huey LO, Vogel GN, and Sanders F H, Jr.
(1996). Unusual strain of Borrelia burgdorferi isolated from Ixodes dentatus in
central Georgia. J Parasitol 82:93 6-940.
Oliver JH, Jr., Kollars TM, Jr., Chandler F W, Jr., James AM, Masters EJ, Lane RS, et al.
(1998). First isolation and cultivation of Borrelia burgdorferi sensu lato from
Missouri. J Clin Microbiol 36:1-5.
Oliver JH, Lin T, Gao L, Clark KL, Banks CW, Durden LA, et a1. (2003). An enzootic
transmission cycle of Lyme borreliosis spirochetes in the southeastern United
States. Proceedings of the National Academy of Sciences of the United States of
America 100:1 1642-1 1645 .
Patrican LA (1997). Absence of Lyme disease spirochetes in larval progeny of naturally
infected Ixodes scapularis (Acarizlxodidae) fed on dogs. J Med Entomol 34:52-55.
193
Piesman J, Donahue JG, Mather TN, and Spielman A (1986). Transovarially acquired
Lyme disease spirochetes (Borrelia burgdorferi) in field-collected larval Ixodes
dammini (Acari: Ixodidae). J Med Entomol 23:219.
Poucher KL, Hutcheson HJ, Keirans JE, Durden LA, and Black WC (1999). Molecular
genetic key for the identification of 17 Ixodes species of the United States (Acari :
Ixodidae): A methods model. Journal of Parasitology 85:623-629.
Rand PW, Lacombe EH, Smith RP, and Ficker J (1998). Participation of birds (Aves) in
the emergence of Lyme disease in southern Maine. Journal of Medical
Entomology 35:270-276.
Richter D, Spielman A, Komar N, and Matuschka FR (2000). Competence of American
robins as reservoir hosts for Lyme disease spirochetes. Emerging Infectious
Diseases 62133-138.
Rudenko N, Golovchenko M, Grubhoffer L, and Oliver JH (2009). Borrelia carolinensis
sp nov., a New (14th) Member of the Borrelia burgdorferi Sensu Lato Complex
from the Southeastern Region of the United States. Journal of Clinical
Microbiology 47:134-141.
Sather GE, Lewis AL, Jennings W, Bond J0, and Hammon WM (1970). Sawgrass virus:
a newly described arbovirus in Florida. Am J Trop Med Hyg 19:319-326.
Sawyer S (1989). Statistical Tests for Detecting Gene Conversion. Molecular Biology
and Evolution 6:526-538.
Scharf WC (2004). Immature ticks on birds: Temporal abundance and reinfestation.
Northeastern Naturalist 11:143-150.
Scoles GA, Papero M, Beati L, and Fish D (2004). A Relapsing Fever Group Spirochete
Transmitted by Ixodes scapularis Ticks. Vector-Borne and Zoonotic Diseases
1:21-34.
Seinost G, Dykhuizen DE, Dattwyler RJ, Golde WT, Dunn J], Wang IN, et a1. (1999).
Four clones of Borrelia burgdorferi sensu stricto cause invasive infection in
humans. Infection and Immunity 67:3518-3524.
Smith RP, Rand PW, Lacombe EH, Morris SR, Holmes DW, and Caporale DA (1996).
Role of bird migration in the long-distance dispersal of Ixodes dammini, the
vector of Lyme disease. Journal of Infectious Diseases 174:221-224.
Sellers H (1955). Ixodes dentatus (Marx) collected from a human (Acarina, Ixodidae).
Proceedings of the Entomological Society of Washington 57:120.
Sonenshine DE (1979). Ticks of Virginia. Virginia Polytechnic Institute and State
University, College of Agriculture and Life Sciences, Blacksburg, VA.
194
Sonenshine DE, and Stout IJ (1970). Contribution to Ecology of Ticks Infesting Wild
Birds and Rabbits in Virginia-North-Carolina Piedmont (Acarina-Ixodidae).
Journal of Medical Entomology 7:645-&.
Stafford KC, Bladen VC, and Magnarelli LA (1995). Ticks (Acari, Ixodidae) Infesting
Wild Birds (Aves) and White-Footed Mice in Lyme, Ct. Journal of Medical
Entomology 32:453-466.
Tamura K, Dudley J, Nei M, and Kumar S (2007). MEGA4: Molecular evolutionary
genetics analysis (MEGA) software version 4.0. Molecular Biology and Evolution
24:1596-1599.
Telford SR, and Spielman A (1989a). Competence of a Rabbit-Feeding Ixodes (Acari,
Ixodidae) as a Vector of the Lyme-Disease Spirochete. Journal of Medical
Entomology 26:1 18-121.
Telford SR, and Spielman A (1989b). Enzootic Transmission of the Agent of Lyme-
Disease in Rabbits. American Journal of Tropical Medicine and Hygiene 41 :482-
490.
Tsao JI, Wootton JT, Bunikis J, Luna MG, Fish D, and Barbour AG (2004). An
ecological approach to preventing human infection: Vaccinating wild mouse
reservoirs intervenes in the Lyme disease cycle. Proceedings of the National
Academy of Sciences of the United States of A merica 101:18159-18164.
Walker ED, Poplar ML, and Russell HL (1992). Ixodes-Dentatus (Acari, Ixodidae) in
Michigan - lst State Records and Occurrence on a Human. Great Lakes
Entomologist 25:303-304.
Walker ED, Smith TW, Dewitt J, Beaudo DC, and McLean RG (1994). Prevalence of
Borrelia-Burgdorferi in Host-Seeking Ticks (Acari, Ixodidae) from a Lyme-
Disease Endemic Area in Northern Michigan. Journal of Medical Entomology
3 1 :524-528.
Walker ED, Stobierski MG, Poplar ML, Smith TW, Murphy AJ, Smith PC, et al. (1993).
Geographic distribution of ticks (Acari : Ixodidae) in Michigan, with emphasis on
Ixodes scapularis and Borrelia burgdorferi. Journal of Medical Entomology
35:872-882.
Wang GQ, Ojaimi C, Wu HY, Saksenberg V, Iyer R, Liveris D, et al. (2002). Disease
severity in a murine model of Lyme borreliosis is associated with the genotype of
the infecting Borrelia burgdorferi sensu stricto strain. Journal ofInfectious
Diseases 186:782-791.
Wang IN, Dykhuizen DE, Qin WG, Dunn JJ, Bosler EM, and Lufi BJ (1999). Genetic
diversity of ospC in a local population of Borrelia burgdorferi sensu stricto.
Genetics 151:15-30.
195
Weisbrod AR, and Johnson RC (1989). Lyme-Disease and Migrating Birds in the Saint
Croix River Valley. Applied and Environmental Microbiology 55:1921-1924.
Wormser GP, Brisson D, Liveris D, Hanincova K, Sandigursky S, Nowakowski J, et al.
(2008). Borrelia burgdorferi genotype predicts the capacity for hematogenous
dissemination during early Lyme disease. Journal of Infectious Diseases
198:1358-1364.
196
CHAPTER 6
Diversity of zoonotic pathogens in Ixodes scapularis as a marker of its invasion of across
the Midwestern United States
Abstract
The center of origin theory predicts that genetic diversity will be greatest at the
site of origin because of a long period of time available for evolution. Conversely,
diversity is predicted to decrease with distance from the origin, because invasion and
colonization are associated with founder effects that reduce genetic variation in newly-
arrived populations. Given the close association of Ixodes scapularis - the blacklegged
tick - and the suite of zoonotic pathogens it transmits, we posited that patterns of diversity
of I. scapularis-borne pathogens across its Midwestern range may be useful in better
understanding the broad-scale invasion of this tick and subsequent disease emergence.
Analysis of 1565 adult I. scapularis ticks from 13 sites across five Midwestern states
reveals that tick infection prevalence with multiple microbial agents, co-infections, and
strain diversity of Borrelia burgdorferi - agent of Lyme disease - were positively
correlated with the duration of establishment of tick populations. The observed
differences in these parameters across gradients of establishment were, however, subtle,
as recently-invaded ticks harbored diverse infections. Our data suggest that the invasion
of ticks and emergence of various tick-bome diseases may be more complex than the
traditional scenario whereby infected, invading ticks are the only means of introduction
of pathogens to native communities.
197
Introduction
The distribution and incidence of vector-borne pathogens and the diseases they
cause are inextricably linked to that of their invertebrate vectors (Gubler 1998; Otranto et
al. 2009). As such, patterns of invasion and establishment of vectors can be used to
predict disease, and similarly, pathogen diversity and prevalence may provide an indirect
understanding of the endemicity of the disease system. In disease ecology, a disease
system is considered endemic in an area when a pathogen persists in a given vector
and/or host population without the need for external inputs. Ixodes scapularis, the
blacklegged tick, is an epidemiologically important vector of multiple zoonotic pathogens
with an endemic distribution in the upper Midwestern and Northeastern United States.
Blacklegged ticks were widespread prior to the Pleistocene glaciation, which receded
10,000 years ago, throughout which time relict populations remained in refugia in the
Northeastern United States and in Northwestern Wisconsin (Spielman 1988). In the mid-
twentieth century, the reversion of agricultural lands to forest and implementation of deer
hunting regulations that allowed deer populations to increase in number supported
expansion of I. scapularis and its vertebrate hosts from refugia (Steere et al. 2004).
I. scapularis continues to spread from endemic areas both in the Northeastern and
Midwestern United States, but the mechanisms for this spread are not established.
The establishment of a new population of ticks requires the invasion of a new site
by a sufficient propagule, and the new site must be receptive with appropriate biotic and
abiotic features to allow for successful tick survival and reproduction. With regard to I.
scapularis, the Centers for Disease Control and Prevention (CDC) defines an established
population as one in which at least six individual ticks of a single life stage, or at least
198
two life stages, are found (Dennis et al. 1998). With regard to pathogens, establishment
is a description of the stability of the infection. Mathematically, establishment can be
demonstrated when the basic reproductive rate (R0) is greater than one; that is, the more
than one secondary infections result from a single infectious unit (Anderson and May
1978), and the extent to which Ro exceeds one may dictate the magnitude of zoonotic risk
(Telford et al. 1991). Across the Midwest, a non-monotonic continuum of establishment
of I. scapularis is apparent (Table 6.1), with tick populations first detected in Wisconsin,
Minnesota, Michigan’s Upper Peninsula, and Northwestern Illinois (Jackson and
Defoliart 1970; Drew et al. 1988; Bouseman et al. 1990; Strand et al. 1992), later
detected in northern Indiana (Pinger and Glancy 1989; Bouseman et al. 1990; Pinger et
al. 1991; Pinger et al. 1996; Cortinas and Kitron 2006), and most recently detected in the
Chicago-area and Michigan’s Lower Peninsula (Foster 2004; Jobe et al. 2006; Hamer et
al. 2007; Jobe et al. 2007). The current tick invasion in lower Michigan is particularly
well-docurnented, with detectable new wildlife infestations moving northward annually
(Hamer et al. 2010).
The three most significant 1. scapularis-borne zoonotic pathogens are Borrelia
burgdorferi, Anaplasma phagocytophilum, and Babesia microti. All three pathogens
have similar life histories, in which wild rodents serve as important (although not the
only) natural reservoirs (Donahue et al. 1987; Telford and Spielman 1993; Levin et al.
2002)and disease results when I. scapularis serves as a bridging vector through becoming
infected by infectious wildlife and subsequently delivering the pathogens to humans. The
epidemiology of each disease reflects the geographic distribution and phenology of I.
scapularis as well as the human behaviors that influence exposure to ticks. B.
199
burgdorferi is a spirochetal bacterium that causes Lyme disease, the most frequent human
vector-borne disease in the northern hemisphere (Bacon et al. 2007). First reported in
1975 in Old Lyme, Connecticut (Steere et al. 1977), over 20,000 human cases are now
reported annually in the United States, with increasing annual incidence (Bacon et al.
2007). Anaplasma phagocytophilum, a gram-negative, obligate intracellular rickettsial
bacterium of neutrophils, causes human anaplasmosis, formerly known as human
granulocytic ehrlichiosis. This disease was first reported among patients in Minnesota
and Wisconsin in 1994 (Bakken et al. 1994; Chen et a1. 1994). In the United States, a
130% increase in incidence of human anaplasmosis has been reported in 2007 over 2003
levels with a majority of cases reported from the I. scapularis-endemic foci in the
Northeast and upper Midwest (Hall-Baker et al. 2009). Babesia microti is one of four
apicomplexan parasites of the genus Babesia that infect erythrocytes and causes a
malaria-like disease called babesiosis in humans. The first significant cluster of disease
was reported among residents of Nantucket Island, Massachusetts, in 1977 (Ruebush et
al. 1977). Babesiosis is usually self-limiting, but can cause severe morbidity and
mortality in asplenic, elderly, and immunecompromised patients. Whereas both Lyme
disease and anaplasmosis are reportable diseases, babesiosis is not, and so incidence
estimates are not reliable -— nevertheless this disease is considered to be emerging
(V annier et al. 2008). In a review of co-infections of the above pathogens within Ixodes
ticks from California, Wisconsin, and the Northeast, the prevalence of co-infections was
found to be highest among ticks collected from regions of Lyme disease endemicity in
the Northeastern states, where infection with each individual pathogen is also highest
(Swanson et al. 2006).
200
In addition to variation in the species of pathogens infecting ticks at different
locations, within-pathogen variation occurs. This latter variation has implications both
for altering the risk of pathogen transmission and for efficient diagnosis and treatment
(Swanson et al. 2006). Strain-level variation has been best characterized for B.
burgdorferi and is maintained through diversifying selection, as the environment is
heterogenous and no single genotype has the highest fitness in all environments (Brisson
and Dykhuizen 2004).
A center of origin is a geographical area where a group of organisms, either
domesticated or wild, first developed its distinctive properties (Cain 1944). This concept
is typically applied to studies of crop domestication, and is a core foundation of dispersal
biogeography (Bremer 1992). The center of origin of a pathogen is generally also its
center of genetic diversity, and the former is hypothesized to correspond with the center
of origin of its host if host/pathogen coevolution has occurred. For example, the leaf
spotting cereal pathogen, Mycosphaerella graminicola, was found to have highest levels
of gene diversity and allele richness in Israel —- the center of origin of wheat (Zhan et al.
2003; Banke et al. 2004). The diversity of a pathogen, however, is not always highest at
the center of origin of its host. For example, while high genetic diversity of the scale
pathogen of barley, Rhynchosporium secalis, was expected in the Fertile Crescent of the
Middle East where barley has been cultivated for thousands of years, diversity was
instead greatest in Scandanavia and Switzerland, leading the authors to suggest that
cultivated barley was not the original host of the pathogen (Zaffarano et al. 2006).
Diversity is predicted to decrease with distance from the origin, because invasion and
201
colonization events are associated with founder effects that reduce genetic variation in
newly-arrived populations (Carlson and Templeton 1984).
We tested the center of origin theory using I. scapularis and associated pathogens
not only to apply this theory to a vector-borne disease system for the first time, but also to
learn more about the biology of these organisms. Given the close association of I.
scapularis and the pathogens it transmits, we posited that patterns of prevalence and
diversity of I. scapularis-bome pathogens across its Midwestern range may be useful in
better understanding the mechanisms underlying the broad-scale invasion of this tick and
subsequent disease emergence. Assuming a clonal evolution of B. burgdorferi
(Dykhuizen et al. 1993), a population with greater genetic diversity is likely to have had a
longer period of time available for evolution (Wang et al. 1999a). Herein we hypothesize
that infection prevalence, interspecific, and intraspecific diversity of pathogens within
ticks will positively correlate with the duration of establishment of tick populations, due
to the longer period of time for evolution in areas with established population. The
objectives of this study were to test the predictions for a center of origin hypothesis
through assessment of the following traits of the disease system across a continuum of l.
scapularis establishment in the Midwest: (1) the population abundance of I. scapularis;
(2) prevalence of Borrelia spp., Anaplasma phagocytophilum, and Babesia spp., and
mixed infections in I. scapularis; and (3) strain-level diversity of B. burgdorferi.
Materials and Methods
Tick collections. Ticks were collected by drag cloth during the spring adult questing
season (15 April - 15 May) of 2006—2007 at a total of 13 sites across the Midwest (Figure
202
6.13): Castle Rock State Park, Ogle Co., IL; Indiana Dunes National Lakeshore, Porter
Co., IN; Tippecanoe River State Park, Pulaski Co., IN; Van Buren State Park, Van Buren
Co., MI; Duck Lake State Park, Muskegon Co., MI; Fort McCoy Military Instillation,
Monroe Co., WI; Governor Dodge State Park, Iowa Co., WI; Saugatuck Dunes State
Park, Allegan Co., MI; Menominee North and South, Menominee Co., MI; Churchill
Woods Forest Preserve, DuPage Co., IL; Fort Sheridan, Lake Co., IL; and St. Croix State
Forest, Pine Co., MN. We targeted our sampling time for the peak period of activity of
the adult life stage of I. scapularis because adults have had two prior bloodmeals and
therefore have had a greater chance of being infected in comparison to nymphs. Due to
close geographic proximity (between 30-70 km between sites) and tick establishment
status, similar drag densities, and inadequate individual sample sizes, we grouped the two
Chicago area sites (Chicago North and South), the two Menominee Co. sites (Menominee
North and South), and two sites in Michigan’s Lower Peninsula (Saugatuck and Duck
Lake) for our analyses. Based on the time of first documented 1. scapularis within these
regions, we assigned sites to establishment groups as follows: Minnesota, Wisconsin,
western Illinois, and Michigan Upper Peninsula sites are ‘long-established’ tick
populations; northern Indiana sites are ‘intennediate-established’; and Chicago-area and
lower Michigan sites are ‘recently-invaded’.
Each site was sampled for questing ticks (Figure 6. lb) by dragging a 1-rn2 white
corduroy cloth (Falco and Fish 1992) on rain-free days in the late morning or late
afternoon to avoid the hottest and least-humid times of day (Schulze et a1. 2001; Diuk-
Wasser et al. 2006). The cloth was inspected every 20 m, and attached ticks were stored
in 70% ethanol. Drag-sampling was not performed on excessively hot or wet days. In
203
2006, ticks were removed from the cloth and placed into chambers with a piece of
vegetation for moisture and were kept alive for transport to the laboratory at Michigan
State University. In 2007, ticks were removed from the cloth and placed directly into
70% ethanol. At each site, we attempted to collect a quota of 100 adults to allow for an
adequate sample size of infected ticks for analysis, but this was not always possible in
low-density sties. In some cases, the distance that was sampled was not directly recorded,
and was estimated based on the duration of time spent sampling and number of ticks
collected. An index of tick abundance was calculated as the number of adult I. scapularis
collected per 1000m2 of drag sampling.
Pathogen detection. Ticks were identified to species and stage (Keirans and Clifford
1978; Sonenshine 1979; Durden and Keirans 1996). In 2006, live ticks were aseptically
bisected with one half used to culture B. burgdorferi (data not presented). Total DNA
from the remaining half tick in 2006 and the whole tick in 2007 was extracted using the
DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA) following the manufacturer’s
animal tissue protocol, with modifications as described in Harner et al. (2010). B.
burgdorferi strain B31-infected nymphal I. scapularis kindly provided by the Centers for
Disease Control and Prevention (CDC) served as the positive extraction control, and
water as a negative extraction control.
All ticks were tested for the presence of three classes of pathogens in three
separate PCR reactions -- for Borrelia species, Babesia species, and Anaplasma
phagocytophilum;-with subsequent sequencing of Borrelia-positive and Babesia-positive
samples to identify the species and/or strain. A PCR enzyme kit was used in all assays
204
(PCR Supermix, Invitrogen, Carlsbad, CA), and water was used as a negative control in
all assays. All assays were run in a 50 pl reaction volume. Borrelia species were detected
using a nested polymerase chain reaction (PCR) for the 16S - 23S rRNA intergenic
spacer region (IGS) as described by Bunikis et al. (2004a), resulting in a product size of
approximately 980bp for B. burgdorferi and 500bp for B. miyamotoi. DNA from B.
burgdorferi strain B31-infected ticks kindly provided by the CDC served as a positive
PCR control. Additionally, we screened 2007 samples for B. burgdorferi using a
quantitative PCR (qPCR) targeting the 168 gene following the protocol of Tsao et al.
(2004), and the reported infection prevalence includes samples positive on either assay.
Babesia genus-specific PCR was performed using primers for the 188 rRNA gene to
produce a fragment of variable size, including a 408bp fragment for B. microti or a 437-
bp fragment for B. odocoilei (Armstrong et al. 1998). Commercially-available B. microti
organism (ATCC, Manassas, VA) was extracted as above and used as a positive control.
The p44 gene of A. phagocytophilum (Zeidner et a1. 2000) was amplified using a
touchdown PCR program described in Steiner et a1. (2006) to produce a 334—bp fragment.
Infected laboratory colony I. scapularis nymphs provided by D. Fish at Yale University
were extracted as above and used as a positive control.
Nucleotide sequencing. Species identification and strain typing of Borrelia-positive and
Babesia-positive ticks was attained through DNA sequencing. Additionally, a subset of
randomly selected sammes that were PCR-positive for A. phagocytophilum was
sequenced to confirm the species identity. All products were purified (Qiagen PCR
Purification Kit; Qiagen, Valencia, CA) and sequences were determined on an ABI Prism
205
3100 Genetic Analyzer (Applied Biosystems, Foster City, CA) using the primers that
were also used in PCR. Borrelia sequences were identified as either B. burgdorferi or B.
miyamotoi and Babesia sequences were identified as either Ba. microti or Ba. odocoilei,
based on comparisons to published sequences using the basic local alignment search tool
in GenBank (Altschul et al. 1990).
Strain-level analyses were further conducted for approximately half of all B.
burgdorferi samples, and all B. miyamotoi samples For B. burgdorferi sequences, a 500
nucleotide segment of the IGS was aligned with the prototypical strains published in
Bunikis et al. (2004a) using the ClustalW algorithims within the program Mega4
(Tamura et al. 2007). Analysis of this fragment size allowed identification of the 10 main
IGS groups (groups 1-10; the minimal matrix for differentiation of these 10 groups
includes mutations which all occur within the first 309 nucleotides of the fragment), and
of 20 IGS subtypes within the main groups (1A, 13, 2A/C, 2B, 2D, 3A, 3B/C, 3D, 4A,
4B, 5, 6A, 6B, 6C, 7A/B, 8A/C, SB, 8D, 9, 10), as presented in Bunikis et al. (2004a).
Additionally, sequences were identified to broad ribosomal spacer type (RST 1, 2, or 3;
(Liveris et a1. 1995) based on clustering topology of the IGS phylogenetic trees. In the
case that a sequence we derived did not completely match any published strain, we
classified it as novel IGS mutant and assigned an alphabetical nomenclature beginning
with ‘Midwest’. Sequence chromatographs were manually scrutinized for confidence in
nucleotide assignments and evidence of mixed strain infections, which were excluded
from analyses. For at least one representative of each detected strain, we also determined
the sequence in the reverse direction and/or detemrined the sequence in the forward
direction twice to validate the occurrence of unique mutations. A similar protocol was
206
followed for B. miyamotoi sequences, for which prototypical IGS strains for use in
alignments were obtained from Bunikis et al. (2004b).
Statistics. Kruskal-Wallis non-parametric one-way analysis of variance was used to
assess differences in the index of population abundance of ticks and tick and infection
prevalence among establishment groups. Comparisons in infection between male and
female ticks were made by calculating the z-ratio and associated probability for the
difference between two independent proportions. Chi-squared test for independence was
used to assess co-infections. Statistics were performed using Statistix 8 (Analytical
Software, Tallahassee, FL). The effect of sample size on strain richness was assessed
using a web-based program called the rarefaction calculator (University of Alberta,
Edmonton, Canada; http://www.biology.ualberta.ca/jbrzustoharefactphp). Strain richness
was estimated by using the nonparametric model of Chao, which considers the number of
operational taxonomic units observed and the frequency with which each was observed to
estimate the total population strain richness including unsampled strains (Chao and
Tsung-Jen 2003).
A series of standard ecological diversity indices were computed to assess the B.
burgdorferi strain-level diversity within and among the three establishment groups.
Strain richness (alpha diversity) was tabulated as a simple count of the number of strain
types in each group. Evenness is a measure of the relative abundance of the different
species making up the richness of an area (Krebs 1978). This measure is constrained
between zero and one in which the most even (least variable) community has a value
close to one. Shannon’s Diversity Index considers both richness and evenness (Shannon
207
and Weaver 1949). Sarensen’s Similarity Index (a measure of Beta diversity) was
computed for each pairwise combination of establishment groups (Sarensen (1948); this
measure is constrained between zero and one in which a value of zero indicates no
overlap between communities, whereas a value of one indicates that exactly the same
strains are found in both communities.
Genetic differentiation within a pathogen among geographic populations was
tested using an analysis of molecular variance (AMOVA) (Excoffrer et al. 1992) using
Arlequin 3.1 (Excoffier et a1. 2005) to partition genetic variation into within- versus
among-population components. The fixation index (F ST) is a measure of the level of
population genetic differentiation that reflects the proportion of total genetic variability
that is due to the net differences between populations(Weir and Cockerham 1984). F ST
ranges from zero to one, where a value of one indicates that sub-populations are
completely different and values <0.1 indicate that subpopulations are not different. The
significance of the fixation index is tested using a nonparametric permutation approach
(Excoffier et al. 1992). Population pairwise F ST values were computed to test the null
hypothesis of no difference between populations by perrnutating haplotypes between
populations, in which the P-value of the test is the preportion of permutations leading to a
F ST value larger or equal to the observed one (Reynolds et al. 1983). The molecular
diversity index Theta__k (an estimate of Theta obtained from the observed number of
alleles (Ewens 1972)) and the mismatch distribution (Rogers and Harpending 1992) were
computed for each establishment group. Mismatch distribution assesses the distribution
of the number of differences (mismatches) between pairs of DNA sequences in a sample.
The shape of this distribution is affected by the past demography of a population, such
208
that the distribution is usually multimodal in samples from populations at demographic
equilibrium, reflecting the highly stochastic shape of gene trees. Conversely, a unimodal
mismatch distribution is characteristic of a population that has passed through a recent
demographic expansion. Goodness of fit of the observed mismatch distribution to the
sudden expansion model (unimodal distribution) allowed for assessment of significance
through the calculation of Harpending’s Raggedness Index (Harpending 1994).
To evaluate phylogenetic relationships among B. burgdorferi haplotypes, we
constructed an unrooted neighbor-joining phylogeny and minimum spanning network
(MSN) using Mega4 and TCS 1.21, (Clement et al. 2000; Tamura et al. 2007)
respectively. Evolutionary distances were computed using the Kimura 2-parameter
method and are in units of number of base substitutions per site. Percentage support
values for clades within the neighbor-joining tree were obtained from 1000 bootstrap
iterations. The MSN method determines the gene network in which the total length of the
branches that connect haplotypes is minimized; discrimination among equal-length
MSNs was achieved by assuming older alleles are more common than recently derived
alleles, and that new mutations are likely to be found in the same population as their
ancestor. Evidence for gene conversion was examined using Sawyer’s Test in
GENECONV version 1.81 (http://www.math.wustl.edu/~sawyer/geneconv/), which tests
the null hypothesis that nucleotide substitutions are randomly distributed (Sawyer 1989).
Results
Index of abundance of I. scapularis. Across all sites, a total of 1655 questing ticks were
collected, the majority (94.6%) of which were I. scapularis adults (816 females and 749
209
males). Also collected were 29 nymphal and one larval 1. scapularis, 59 adult
Dermacentor variabilis, and one adult Ambloymma americanum. The indices of
abundance of I. scapularis adults were highly variable within each establishment group,
and were not significantly different among groups (P = 0.95), although the trend was for
abundance to correlate positively with tick establishment status (26.6, 23.7, and 17.6
adults/ 1000m2, at long-established, intermediate-established, and recently-invaded sites,
respectively; Table 6.2).
Infection of ticks with multiple microorganisms. Five different microorganisms - B.
burgdorferi, B. miyamotoi, A. phagocytophilum, Ba. microti, and Ba. odocoilei - were
found among 1565 adult I. scapularis from across the Midwest, with overall percent
prevalences of 51.3, 2.2, 8.9, 0.3, and 4.5, respectively (Figure 6.2; Table 6.2). There
was no difference in infection between females and males for any microbe except B.
miyamotoi (females, 1.3%; males, 2.9%; P = 0.03) and A. phagocytophilum (females,
7.1%; males, 10.8%; P = 0.01). All microbes except Ba. microti were found in ticks of
all three establishment groups; Ba. microti was found in only two long-established
populations. Infection prevalence for all microbes and co-infections except Ba. odocoilei
trended highest in the long-established ticks, though prevalence was only significantly
higher in long-established ticks for A. phagocytophilum (11.4% versus 6.4 and 3.7% for
intermediate-established and recently-invaded ticks, respectively; P = 0.004; Table 6.2).
Ba. odocoilei infection prevalences trended highest in the recently-invaded ticks (7.0%;
versus 2.9 and 3.7% for long-established and intermediate-established ticks; P = 0.35).
210
The IGS of all 34 B. miyamotoi-positive samples was sequenced, and all samples were
100% homologous to ‘Type 4’ North American B. miyamotoi.
Overall, 8.9% of all ticks collected in the Midwest were co-infected with two or
more microbes (Table 6.2; Figure 6.2). Male ticks had a higher prevalence of co-
infections (11.2%) than females (6.9%; P = 0.003). Co-infections were found in all three
establishment groups and were most prevalent among long-established ticks (10.4%; P =
0.03) as compared to intermediate-established (6.7%) and recently-invaded ticks (6.4%;
Table 6.3). Across all data, the most common co-infection resulted from B. burgdorferi
and A. phagocytophilum which occurred in 90 ticks (5.8% of all ticks); this rate of co-
infection is 1.3 times higher than the expectation given the prevalence of each microbe
within the population (x2 = 11.03; P = 0.0009). Other co-infections that were observed in
a frequency that is significantly different than expected included B. burgdorferi and B.
miyamotoi (0.05% of all ticks; 2.1 times lower than expected; 78 = 9.89; P = 0.001) and
B. miyamotoi and Ba. odocoilei (0.3% of all ticks; 2.7 times higher than expected; )8 =
4.62; P = 0.03). The co-infections that were observed at an expected frequency included
B. burgdorferi and Ba. microti (0.1% of all ticks; x2 = 0.26; P = 0.61); B. burgdorferi and
Ba. odocoilei (2.4% of all ticks; )8 = 0.26; P = 0.61); B. miyamotoi and A.
phagocytophilum (0.3% of all ticks; x2 = 0.44; P = 0.51); A. phagocytophilum and Ba.
microti (0.06% of all ticks; x2 = 0.77; P = 0.38); and A. phagocytophilum and Ba.
odocoilei (0.3% of all ticks; 12 = 0.91; P = 0.34). Additionally, infection of ticks with
three microbes was observed 3 times, and a single tick was infected with four microbes.
All of these co-infections of 3 or more microbes occurred in ticks collected at long-
established sites.
211
B. burgdorferi genotypes. The IGS of 458 B. burgdorferi-positive samples (57.0% of all
positives) was sequenced, which included 62.0, 52.7, and 44.7% of PCR-positive samples
from the long-established, intermediate-established, and recently-invaded ticks,
reSpcctively. Across all sites, 21 (4.6%) had evidence of mixed-strain infections, based
on the presence of double-nucleotide peaks in the chromatograph at polymorphic sites as
defined by Bunikis et al. (2004a). The highest proportion of mixed strain infections was
found among long-established ticks (5.1%), followed by intennediate-established (3.9%)
and recently-invaded (2.9%; P = 0.35); mixed strain infections were present at all sites
harboring long-established ticks and intermediate-established ticks, but were only present
in one of the three sites harboring recently-invaded ticks (Table 6.4).
B. burgdorferi RST groups. Among the 437 single strain infections, all three RST
groups were represented, with RST 3 in highest abundance (58.3%) followed by RST 2
(35.5%) and RST l (6.2%). Strains of RST l were least abundant at all sites (Figure 6.3),
and were found at 3 of 5 long-established sites, both intermediate-established sites, and
two of three recently-invaded sites.
B. burgdorferi strain richness and diversity. We detected a total of 22 IGS strains
within 437 B. burgdorferi sequences across the Midwest. These 22 strain types included
8 prototypical IGS strains previously described by Bunikis et al. (2004a) from the
Northeastern United States (IGS 5, 1A, 2A/C, 2D, 4A, 6A, 7A/B, and 8A/C), 4 strains
that we have recently detected in bird-associated ticks (Midwest A, B, F, and K; Hamer et
al. unpublished) and 10 novel IGS variants that differed from published strains by at least
212
one nucleotide (Midwest C, D, E, G, H, I, J, L, M, and N; accession numbers
HN1015238-HM015247). Observed B. burgdorferi strain richness was highest within
ticks of the long-established group (19 strains), followed by intermediate-established (12
strains) and recently-invaded (10 strains), though this relationship is likely influenced by
sample size (297, 74, and 66 sequences determined in the three groups, respectively;
Table 6.4). Using a rarefaction analysis, we determined the rate at which new strains are
found per unit of individuals sequenced (Figure 6.4). At a common sample size of at
least 50, the long-established group is more rich than the intermediate-established group,
which is more rich than the recently-invaded group. In a separate analysis, the Chao-1
non-parametric estimator of true species richness using data across all sites in the
Midwest is 46.5 :1: 19.1. Richness is predicted to be greatest at long-established sites (27
+/- 8.3) followed by intermediate-established sites (14.3 +/- 2.8), and lowest at recently-
invaded sites (11 +/- 1.6). The percent of samples that comprised indigenous strains
(found in a single establishment group only) was greatest among long-established ticks
(11.8%) as compared to intermediate-established (2.7%) and recently-invaded (1.5%; x2
= 11.9, df= 2, P = 0.003; Figure 6.5).
Using a standard ecological index, we found that strain evenness among all three
establishment groups was comparable (0.83-0.85). Shannon diversity was greatest
among the long-established ticks (2.45), followed by intermediate-established (2.06) and
recently-invaded (1.95). The Sorenson’s Similarity Indices computed for pair wise
comparisons of establishment groups indicated that the intermediate-established versus
recently-invaded strain diversities were most similar (0.73), the long-established versus
213
recently-invaded was least similar (0.62), and long-established versus intermediate-
established was in-between (0.65).
Of the 500 nucleotides of the IGS that were assessed, the total number of
polymorphic sites across all sites was 42, including one indel block of 7 nucleotides that
was treated as a single polymorphism. Polymorphic sites were most common among
long-established samples (n = 40 polymorphic sites), followed by intermediate-
established (n = 36) and recently-invaded (n = 34). No significant fragments were found
using Sawyer’s test, which indicates that there is no evidence for recombination at this
locus. The molecular diversity index Theta_k was greatest in long-established sites and
smallest in recently-invaded sites, yet differences were not significant (Table 6.4). The
mismatch distribution was a significant fit to the model of sudden expansion for the
recently-invaded population (Harpending's raggedness index = 0.063; P = 0.02),
marginally significant for the intermediate-established population (Harpending's
raggedness index = 0.059; P = 0.05), and non-significant for the long-established
population (Harpending's raggedness index = 0.016; P = 0.41).
B. burgdorferi population structure. A total of 99.11% of IGS molecular variation
occurred within establishment groups, and 0.89% occurred among establishment groups.
Marginally significant differences were found in the identity and frequency of IGS
haplotypes among the three establishment groups (F St 0.0089; P = 0.06), indicating that
the majority of haplotypes are present in all three groups and that high genetic diversity is
maintained across the Midwest. Population pairwise FST analysis indicates that IGS
haplotype frequencies of long-established and intennediate-established populations were
214
not different (P = 0.32), whereas differences were significant between long-established
and recently-invaded populations (P = 0.07) and intennediate-established and recently-
invaded populations (P = 0.03). These differences in population structure are driven by
the differences in relative abundance of certain strains: most notably, IGS 8A/C
comprised 14.1 and 29.7% of long-established and intermediate-established strains,
respectively, yet was not present in the recently-invaded population. IGS 6A comprised
18.2% of the recently-invaded population, but only 8.1 and 2.7% of long-established and
intermediate-established populations (Figure 6.5).
B. burgdorferi phylogeny and network analysis. The topology of the IGS phylogenetic
tree (Figure 6.5) —— with broad grouping into three RST groups of which RST 3 is
paraphyletic, and firrther delineation of B. burgdorferi into about a dozen intraspecific
lineages — is well described (Bunikis et al. 2004a; Attie et al. 2007). Some of the tree
topology is star-shaped with polytornies that occurred only at sequences that differ by one
or two nucleotides. Midwest A was the most common strain across the Midwest and was
the only strain to be found at all sites; it constituted 19.5, 13.5, and 21.2% of samples
from long-established, intermediate-established, and recently-invaded ticks, respectively.
Seven strains were singletons within our sampled population (IGS 2A/C, Midwest E, I, J,
L, M, and N), and the percent of all strains that were singletons in each establishment
group did not differ significantly (21.1, 25.0, and 20% in long-established, intermediate-
established, and recently-invaded ticks; P = 0.95). Strains shared by more than one
establishment group were the most abundant within the sample, with eight ubiquitous
strains found in all groups and three strains found in two groups. Of the 11 strains that
215
were only found within one establishment group, the majority (72.7%) occurred in long-
established ticks, whereas two such strains occurred in intermediate-established ticks, and
a single such strain occurred in recently-invaded ticks. The unique IGS strains from each
establishment group are inter-digitated with those of the other establishment groups as
well as with ubiquitous strains, suggesting a recent shared history. Within our sampled
population, the maximal number of mutational steps separating a strain from the next
most homologous strain was three (Figure 6.6). There was no apparent geographical
structure in the relationships among the haplotypes; with the exception of the unlinked
RST 1 strains, the network exhibited some star-like topology with low levels of sequence
divergence and a high frequency of unique mutations.
Discussion
There have been many criteria proposed for establishing the center of origin of a
taxon, none of which independently is useful, but together may provide a framework for
assessing diversity within and among species in relation to establishment and invasion
(Cain 1944). These criteria were originally developed in the study of plant geography,
but some of the more frequently used criteria (as listed in (Crisci et al. 2003) and detailed
below) have direct application to better understanding the establishment of the I.
scapularis-borne disease systems. Below we recapitulate our data on relative tick
abundance, interspecific microbe diversity, and intraspecific microbe diversity across a
continuum of reported I. scapularis establishment in the Midwest and conclude that the
long-established populations of I. scapularis serve as a center of origin for the
Midwestern I. scapularis-home disease system. The observed gradients of relative
216
diversity are, however, quite shallow, and mechanisms accounting for these patterns are
speculated below.
1. The location of the area of greatest dominance and density of distribution. At the
tick-level, a gradient of was apparent in the indices of abundance of I. scapularis in
which the highest tick abundances occurred within the most long-established populations,
similar to the findings of Walk et al. (2009) in New Hampshire. Similarly, in a
standardized survey for I. scapularis nymphs across their distributional range, Duik-
Wasser et a1. (2006) found that within the Midwestern sampling sites, highest densities
were in Minnesota and Wisconsin, at densities comparable to those in Northeastern
endemic areas. While the overall trend in our dataset was for highest tick abundance in
longer-established populations, variation was apparent; most notably, the highest index of
abundance of adults found in the study was at Van Buren State Park, which was invaded
in the early 20005 (Foster 2004). We conducted a longitudinal study of questing times at
Van Buren, and found that adult activity is greater in the Spring than it is in the Fall
(unpublished data), which is consistent with other findings in the Midwest (Strand et al.
1992; Stancil 1999; Jones and Kitron 2000), but different than the characteristic
phenology in the Northeast in which adults are more active in the Fall (Fish 1993).
At the microbe-level, the prevalence of four of the five microbes we detected was
greatest in hi ghly-established ticks compared to intermediate-established and recently-
invaded ticks (though statistical difference was only detected for A. phagocytophil um).
At highly-established sites, the infection prevalences we report for B. burgdorferi
(52.1%), A. phagocytophilum (11.2%), and B. microti (0.5%) are similar to those reported
by Steiner et a1. (2008) from three sites across the Midwest, except the latter report no B.
217
microti. In two different Lyme disease-endemic areas in New Jersey in the Northeast,
Varde et al. (1998) report 43, 17, and 5% infection with these three agents, and Adelson
et al. (2004) report 33.6, 1.9, and 8.4% infection with these agents, indicating variability
in infection within endemic foci. While all three of these pathogens share wildlife
reservoirs (white-footed mouse) and a tick vector (Ixodes scapularis), the degree to
which other reservoirs and vectors are involved in transmission, and the efficiency of the
L scapularis transmission cycles, may influence the infection prevalence we observed.
Walk et al. (2009) and Hamer et al. (2007) found evidence that B. burgdorferi
infection prevalence is notably higher in Lyme disease endemic areas. Using a larger
dataset over a broader spatial and temporal scale, we found that newly-invaded
populations had nearly equivalent B. burgdorferi prevalence prevalences to long-
established populations, which suggests that there is not a long time lag between tick
invasion and build-up of high density pathogen prevalence within these populations.
This observation may result from a simultaneous dual-invasion process of both the tick
and the pathogen (Hamer et al. 2010), or from the existence of cryptic pathogen
maintenance cycles, in which certain hosts across the landscape are infected prior to the
arrival of the bridging vector 1. scapularis (Chapter 5). Also, it may be that very soon
after ticks invade an area, infection prevalence may be low, but once tick densities build
and more time has passed, prevalence builds quickly.
In addition to detecting the three zoonotic pathogens of main interest, our
sampling protocols also allowed for simultaneous detection of two tick-bome microbes
not known to be of epidemiological importance: B. miyamotoi and Ba. odocoilei. B.
miyamotoi was originally described in l. persulcatus ticks in Japan (FUkunaga et al. 1995)
218
and was first detected in North America in I. scapularis ticks (Scoles et al. 2004). These
spirochetes group genetically with the spirochetes responsible for relapsing fever diseases
in humans, are transovarially-transmitted, and rodents serve as a reservoir (Barbour et al.
2009). Across our samples from the Midwest, only one strain of B. miyamotoi was
present, which was the only strain reported among 22 samples from Connecticut (Bunikis
et al. 2004b)and this was the same strain that was identified among 22 field samples in
Connecticut (Ba. odocoilei is an intraerythrocytic protozoan parasite associated with
white-tailed deer and other cervids, and is genetically similar to Ba. divergans, the
causative agent of babesiosis in cattle, and a zoonotic Babesia species (Herwaldt et al.
2003). This parasite has a northcentral and northeastern distribution in the United States
(Steiner et al. 2006) as well as in Texas, Oklahoma, and Virginia (Waldrup et al. 1990).
Its widespread infection in human-biting ticks, combined with its genetic relatedness to
the human-infectious Babesia species, implies that it may cause disease in immune-
compromised people (Armstrong et al. 1998). The ubiquity with which we found these
microbes across the Midwestern I. scapularis range (B. miyamotoi, 6 of 10 sites; Ba.
odocoilei, 8 of 10 sites) underscores the importance of diagnostic assays that differentiate
them from the genetically and morphologically similar agents that pose public health risk
(i.e., B. burgdorferi and Ba. microti). The definitive diagnostic test for human babesiosis
is identification of parasites on Giemsa-stained thin blood smear (Wormser et al. 2006),
and the ability of Ba. microti to be specifically identified in the manner is unknown.
2. The location of the greatest variety of forms of the taxon. The number of microbial
taxa was greatest in the highly-established tick populations, where all five microbes were
219
present (B. burgdorferi, B. miyamotoi, A. phagocytophilum, Ba. microti, and Ba.
odocoilei). Ba. microti was not present in intermediate-established or recently-invaded
ticks. Similarly, as demonstrated by the population-specific rarefaction curves (Fig. 4),
B. burgdorferi strain richness was greatest in highly-established tick populations. Given
a common sample size of 60 infected ticks in all groups, for example, statistically greater
strain richness is found at highly-established sites (13.6i1.3) versus intermediate-
established (11.4:t0.7) and recently-invaded (9.4:t0.4). Nevertheless, the diversity of IGS
genotypes that we found within recently-invaded ticks was higher than we had
anticipated given the shorter time available for colonization of these diverse strains.
Similarly, Wang et al. (1999b) found that the ospC variation within a local population
was almost as great as the variation of a similar-sized sample of the entire species. It may
be that infected ticks are invading from areas of high pathogen prevalence, or that
uninfected, invading ticks readily encounter infectious hosts in the areas into which the
ticks are invaded — these hosts could be previously infected due to pathogen transmission
from other tick species.
Diversifying selection may influence patterns of B. burgdorferi strain diversity
through at least two mechanisms. First, host associations, facilitated by selective killing
of strains by host-specific innate immune factors (Kurtenbach et al. 2002), may maintain
diverse populations of Lyme disease spirochetes, at both the inter-species level (e.g. in
Europe, B. burgdorferi and B. afzelii are mammal-associated, whereas B. garinii and B.
valaisiana are bird-associated (Dubska et al. 2009)) and at the intraspecies level of B.
burgdorfiari (Brisson and Dykhuizen 2004). The magnitude of intraspecies host
association is, however, debated (Hanincova et a1. 2006). Secondly, the antigenic types
220
that have already infected a host cannot establish a subsequent infection in the same host
because of the immune response. Therefore, rare antigenic types are more likely to infect
previously-infected hosts, which leads to some hosts acquiring multiple strains and
maintenance of high levels of variation at the population level (Wang et al. 1999b).
3. The greatest number of overlapping distributions. Co-infections - acquired from
sequential feeding events on infected hosts or from a single feeding event of a co-infected
host - are generally expected to occur in highest frequency in sites with well-established
transmission as opposed to recently-colonized foci with many na'ive hosts and a reduced
diversity of strains due to invasion by one or a few founding ticks. Co-infections are
epidemiologically important because human infection with multiple pathogens may lead
to incomplete diagnosis and insufficient treatment (Swanson et al. 2006). At the inter-
species level, we found that ticks co-infected with multiple organisms were indeed most
common at long-established sites, where 10.4% of ticks harbored more than one microbe
(compared with 6.7 and 6.4% at intermediate-established and recently-invaded sites,
respectively). Across all our study groups, B. burgdorferi-A. phagocytophilum co-
infections occurred more frequently than would be expected given the prevalence of each
individual pathogen. Steiner et al. (2008) also found significant co-infection of these
pathogens in adult I. scapularis from Ft. McCoy, one of our long-established sites.
At the intra-species level, mixed strain B. burgdorferi infections were negligibly
most common in hi ghly-established ticks, where 5% of B. burgdorferi-infected ticks
harbored more than one IGS strain (compared with 4 and 3% in intermediate-established
and recently-invaded ticks, respectively). There was significant variability in the
221
pr0portion of mixed-strain infection within each establishment group, however, including
a maximum of 15% at Castle Rock, a highly-established site in Northwestern Illinois.
The comparable mixed strain infection prevalence between recently-invaded and highly-
established ticks is expected given the near-equivalent B. burgdorferi infection
prevalences, and suggests that the degree of interaction among infected ticks and hosts is
a newly invaded region is high. In comparison, Gatewood et al. (2009) report an overall
prevalence of mixed strain infections at the IGS locus of approximately 20% in nymphs
from across the range of I. scapularis establishment, including sites in both the Northeast
and Midwest, with no apparent geographic bias. The comparatively lower prevalence of
mixed strain infections in our study, especially in consideration that adults are likely to
have more mixed strains given the additional bloodmeal they have had in comparison to
nymphs, may be an artifact of the truncated region of the IGS that we analyzed (i.e.,
signals of mixing at polymorphic sites could have occurred in a region of the IGS that we
did not assess), or mixed strain infections may more commonly produce poor quality
sequences that we were not able to assess. Using the ospA and ospC gene targets, Wang
et al. (1999b) found 45 and 50% mixed-strain infections in adult I. scapularis from the
endemic area of Shelter Island, NY (the sites from which B. burgdorferi was first
isolated), and Guttrnan et al. (1996) found 60% of adults from the same areas harbored at
least two strains of B. burgdorferi. IGS appears not to be as sensitive as ospC in detecting
multiple strains, and therefore our report of mixed strain prevalence should be considered
a minimum, and as long as there is no facilitation or exclusion of particular strains, the
bias should be the same across sites.
222
4. The location of the most primitive form. In regard to B. burgdorferi genotypes, the
RST 1 group is ancestral to RST 2 and 3 (Bunikis et al. 2004a). While strain richness —-
one measure of genetic diversity - is less in RST 1 than in RST 2 and 3, the nucleotide
diversity (i.e., the number of nucleotide difference between strains) in RST l is greater
than RST 2 and 3 (Bunikis et al. 2004a). A low density of RST 1 strains was present
within all three establishment groups, the proportion of which did not relate to tick
establishment status (5.4, 8.1, and 7.6% in highly-established, intermediate-established,
and recently-invaded populations, respectively). The presence of more than one RST 1
strain, however, was found only within highly-established ticks. Across the Midwest,
RST 1 was only represented in 6.2% of infections. Furthermore, the RST 1 had less
strain richness (comprised of only 2 strains) as compared to RST 2 and 3 (with a total of
20 strains). Conversely, in a study of I. scapularis nymphal infection across its
distributional range including both the Midwestern and Northeastern United States,
Gatewood et al. (2009) found that RST 1 strains comprised 20.8% of all infected nymphs.
The apparent absence of RST 1 at two long-established sites and one recently-invaded
site combined with the overall low prevalence of this group in the Midwest may reflect
different mechanisms involved in maintaining Midwestern versus Northeastern B.
burgdorferi.
5. The identification of continuity and directness of individual variations or
modifications radiating from a 'centre of origin' along the highways of dispersal.
Assessment of the structure of the genetic data could allow for speculation as to the mode
or directionality of tick and pathogen dispersal, assuming a center of origin. For
223
example, if the long-established I. scapularis populations were the most dense, contained
the highest prevalence of infection with multiple pathogens and coinfections, and the
highest diversity of B. burgdorferi strains, and if the intermediate-established and
recently-invaded tick populations were characterized as nested subsets of these
parameters, this pattern may suggest a linear mode of invasion, with length of
establishment a useful predictor of the parameters. Conversely, if recently-invaded ticks
were equally or more abundant and infected with diverse pathogens, then a directionality
of dispersal from established foci would not be concluded. Across the Midwest, we
found marginally significant global population structure (P = 0.06). While the majority
of strain types found in any establishment group were shared among all three groups,
suggesting high rates of gene flow, the presence of indigenous strains within each
establishment group, however, suggests some isolation by distance and an emerging
evolutionary divergence. Over 99% of the genetic variation that was observed was due to
within group differences.
In our dataset, the IGS haplotype frequencies of the recently-established
pOpulation were significantly different than both the hi ghly-established and interrnediate-
established populations. The network analysis, exhibited by some star-like topology with
low levels of sequence divergence and a high frequency of unique mutations, is indicative
of a rapid population expansion, in the absence of selection. Similarly, analysis of
mismatch distributions suggests that a recent demographic expansion has occurred in the
recently-invaded population (P = 0.02) and intermediate-established p0pulation (P =
0.05), with no evidence for recent expansion in the long-established population (P =
0.41). The trend in the molecular diversity index Theta_k was that long-established sites
224
were the most diverse, and recently-invaded sites were least diverse, but this difference
was not significant.
The diversity gradient that we hypothesized was in fact quite shallow, as much of
our data show that recently-invaded sites harbor equally or only slightly less diverse
assemblages of B. burgdorferi than long-established sites. Speculation as to the
mechanisms that may account for the observed pattern may include the following: (i)
There is a high propagule pressure associated with invasion, not only of the tick, but also
of the pathogen, which may invade new sites in diverse wildlife reservoirs or in
alternative tick species. With many independent introduction events, of which each may
constitute introduction of a different strains, founder effects are not apparent. (ii) Even if
a limited number of strains are introduced to a new area, in this multi-host, complex
system, so many different reservoir species each present different selective pressures,
which leads to diversity. (iii) The time scale for build up of diverse populations may be
rapid (given our evidence for rapid demographic expansion) such that even if initial
invading diversity is low, our sampling occurs too late to appreciate this. (iv) A pre-
existing cryptic cycle of the pathogen may be maintained in advance of the invasion front
of I. scapularis, such that I. scapularis may encounter wildlife infected with diverse
strains upon its own invasion.
Epidemiological significance. Two main genetic typing schemes are useful in assessing
pathogenicity of different B. burgdorferi strains in humans: RST (a tripartite
classification of the 25+ IGS strains) and outer surface protein C (ospC; an antigenic
plasmid-borne gene); ospC is in positive linkage disequilibrium with IGS (Bunikis et al.
225
2004a). Some RST 1 strains are associated with a higher frequency of disseminated
infection in humans and more invasive disease in experimental animals (Seinost et al.
1999; Wang et al. 2002; Derdakova et al. 2004; Hanincova et al. 2008; Wormser et al.
2008a), and a bias toward the relatively less invasive RST 2 and 3 strains has been found
among infected I. scapularis in the Midwest (Gatewood et al. 2009). While our data
similarly demonstrate a strong bias toward RST 2 and 3 infections at our Midwestern
sites, RST 1 strains (IGS 1A and ‘Midwest L’) were found within ticks of all
establishment groups, including recently-invaded ticks, which underscores the rapidity
with which newly detected tick populations may pose a public health risk.
The pathogenicity in humans of the large number of novel IGS mutant strains is
unknown, though we have no a priori reason to believe that single or double nucleotide
polymorphisms within the non-coding intergenic spacer region would necessarily
correspond to a change in virulence. To begin to assess the epidemiological risk
associated with novel strains found within this cryptic cycle, we sent total DNA from 88
B. burgdorferi samples with comprising five IGS strains that were not previously
reported in the literature or in Genbank (Midwest A, E, J, K, and M) to a laboratory at the
University of Califomia-lrvine for direct ospC typing (B. Travinsky and A. Barbour,
unpublished data). Of these, ospC was successfully amplified and sequenced from 87
samples, resulting in detection of 48 single strain infections and 39 mixed-strain
infections. Present in single strain infections included 18 ospC types, including
representatives of the four ospC major groups that have been associated with
disseminated human Lyme disease (Seinost et al. 1999). The degree to which mixed
strain infections were present at the ospC locus within these samples (which had no
226
evidence of mixing at the IGS locus) further underscores the complexity of the
maintenance of pathogen diversity across the Midwest.
Data on genetic diversity of B. burgdorferi have practical implications for
interpretation of serodiagnoses. The IGS marker that we have used to quantify diversity
is linked to serodiagnostic antigens (including ospC), and variation has been found in the
ability of various diagnostic protocols for detection of diverse strains, with C6 ELISA
able to detect more strains than the standard two-tiered approach (Wormser et al. 2008b).
The confirmed cases of Lyme disease and human granulocytic anaplasmosis
reflect the patterns we observed based on tick data. In 2007, the national incidence of
Lyme disease was 9.2 cases/100,000 people (Hall-Baker et al. 2009), in which highly-
established Wisconsin and Minnesota reported incidences of 32.4 and 23.8 cases/100,000
people, respectively. Incidence in Illinois (1.2 cases/100,000 people), Indiana (0.9
cases/100,000 people) and Michigan (0.5 cases/100,000 people; on average, >55% of
reported cases in Michigan are from the Lyme-disease endemic hotspot in Menominee
County; E. Foster, pers. com). The incidence of human granulocytic anaplasmosis
reflects a similar trend, but of a lesser magnitude than B. burgdorferi. In 2007, the
national anaplasmosis incidence was 0.3 cases/100,000 people (Hall-Baker et al. 2009),
with Minnesota and Wisconsin reporting incidences of 6.2 and 1.2 cases/100,000 people,
respectively. The only other Midwestern state from which we collected ticks that also
reported cases was Illinois, with an incidence of 0.05 cases/100,000 people; no cases
were reported from Indiana or Michigan.
227
:58 can 5038.5 38 ..m a 32. ..88 ..a .0 82. 88 mz .92.... 882...
5.3m 5.5 2m .93.. .25 3.8...mb c an .88 ..m .0 use: .38 .28.. 88 23%.... .26.. am 58.5.... $.88”.
.02”. 0253de new 350 «can... 83 ..m «o ..oiqcil. 59 ..m .o 5ch com? >>z 6:39.. 35.383
sonnets...
x8”. 830 32 ..m .o :85. an: ..m .o 858:8 82 2,2 .m.o_._._.
58w .25 5.02 85528.... .38 ..m a 25.6 82 222.com .83 m 58.6.2
590 .6 82 ..m .o 26.0 .82 ..a a :33 .82 6.2.8 9.... m .3825: 85.58...»
.6022 to“. .330 .oEoBO 2.3 5:83 new 838.. 58.. .868 new 32 6.288.? .98..
.68» 22.3 2.. .o as.» 2.... 26.5.0 2253.358 88» was“
€333» 88x. Egg-.38»...
03..an he .2;
45852383 3332
058:3 flap :2: no woman contowofio 8w 3.5 House 05 5 BEES moi Eur.— .ofizfifl: 35:85 no women macaw 528322
vogue—om S .96. oufimbsm 05 «a “augmznfimo nous—25a $8338.. $305 mo 8:8 2:32 2: mo moaefiumo «mom . _ .w 2an
228
.8... no... 8... .00.." 3... 5.5 3.... 8m 80.: ..<
8.. 00..., 0 EN 0 3.3. and. 8 83 .__ 580 .2... 5.02 098.5
9.... 8.0 o v0... «0... 9...... .08 v0. 88 .... .s. ...8 0.0.0 5...... cm> 82......
02 Saw 0 o o 2.3 a... 0.. 8.... n: .s. 9...... .25 2... 838:8 -2283.
2.0 a...” 8... .30 3.. 2.2. 8...... SN 8...". ..<
3.0 8... o a. 8.. 3...... $.00 EN 8:. 2. ...m. 0.20 .02”. 0282...: 020.3%...
8d 0 o 8.... an. 0...... no.0 0. 83 2. 559.3 .2262 8...... 95...... 0.2.5....0...
.3...— 35 um... 08.: 8." 8."... «0.0.4. 2.0 .98 ..<
0.. : 8.0 o .00. o 8.8 8.0.. 8. 8mm .5 ..2 £80 9.... £52 masses...
8... 3.... 2.. 3. : .3 8a.. 8.3 :3 8.... .2. .822 ..o.
«0.2 3.0 o 8.0 SN «.3 00.0w .8 83 .2. ...m. 0.30 00.80 550.50
8... 3.... o 8.0 new $0.. 8.0. 8. 5.... .. ...m... 0.50 .8”. 0.80 85.32....
...m. o x... 0...: o 38 8a. .0 88 25. neon. 0.20 520 ..m -05..
2.88.... 5.80.8 020.... .002... .< .8085... 5800...... 58.: .02 88.5 850 8.0 3.50
.60 in in .m d 280!
.2. 8.0....
39 02.0.08... 8.82:.
8.300 0.8380» ..
638.80.. .039. u .3 8.1.8.8.. 8&3
n .5 .8050. 3.880%... a...» .008: 08 anew 6088.388 mace... 8088...... .80....:m.m .nooméoom 8.8m 00.83 60.8323.
0... 5 2088.358 0.33%.... N ..o m.0>0. 08.00 3.3... $8.388 .N :38 mom. ..o 8.08 8.80.... 9.0 008.08.... .0 x00... .m6 030...
229
Table 6.3. Matrices of co-infection for each I. scapularis establishment group. Infection
prevalence (%) with each individual microbe is on the diagonal in bold. Expected co-
infection prevalence is in the upper triangle, and observed co-infection prevalence is in
the lower triangle. Observed prevalences are not different than expected unless indicated
with an asterisk. Bb = B. burgdorferi; Bm = B. miyamotoi; Ap = A. phagocytophilum;
Barn = Ba. microti; Bao = Ba. odocoilei.
Long-established (N = 970)
Bb Bm Ap Barn Bao
Bb 52.1 1.6 5.9 0.3 2.0
Bm 0.9* 2.9 0.3 0.02 0.1
Ap 7.1* 0.3 11.4 0.06 0.4
Bam 0.2 O 0.1 0.5 0.02
Bao 2.2 4.1* 4.1 0 3.9
*P = 0.005 - 0.021
Intennediate-established (N = 297)
Bb Bm Ap Bam Bao
Bb 49.2 0.7 3.1 NA 1.8
Bm 0* 1 .3 0.1 NA 0.05
Ap 4.7* 0.3 6.4 NA 0.2
Bam NA NA NA 0 NA
Bao 1 .7 0 0 NA 3.7
*P = 0.027 - 0.048
Recently-invaded (N = 298)
Bb Bm Ap Barn 830
8b 51 0.2 1 .9 NA 3.6
Bm 0 0.3 0.01 NA 0.02
Ap 2.3 0 3.7 NA 0.3
Bam NA NA NA 0 NA
Bao 4 0 0 NA 7
230
$934.: 3.» 2 8 8... 8 __<
v : cod 3 .= 58m 25 £32 omflzo
m on 3.0 an n... =2 {E 29.0. 85m cm> 38>;
v m cod m n... =2 9.3 5.5 .25 628:3 358m
35.8.: 3.» 2 fl 3... t. __<
m 8 8.0 8 2. {ma 2me 52m ScSmaaF 35:38
5 2 mod 2 z. 855 manic. 9585.95
56.3.3 54. 2 Bu 3... 2n .2
m om. mod 5 n5 :2 snow can 582 85205.2
3 mm 8.0 3 .2, >80: to".
m. E 8d fl .3 {an 29m $80 55960
2 3. 2.0 3 .= fun. 93m fiom «:30 35:33
I 2 m... mod 2. 25. neon. 23m 590 .5 -95..
xtfiofi. 305.2. 32.0360» 523.355 30:362. 85» 85 «35$
£25 £83 :oanoE .oz
6.9:» .oz
2355“”
333 n ma 33833 .595 u as .nooméoom £83m coma: 888322 05 E “cogmznsmo “.2333“. .N we 8:35:80 a $88
3:6“ £33588 .N :53 cocoon“: wmv E cowoomfi .tmbohMSa .m Ema—m Bwfim mo mmofiot 38% Ba 32603 Exam v9.22 6.0 033.
231
- u u — u q ‘1‘
_
Ev. 03 9... as o
< .
Z
or 5:8 25
33¢ o 339.... 5.52
«950 «555-09350
9.00”. 330
5:5 cw> ,1. I
-950 xoamusmm a 0950 55300.
9.3 x25 ..
M Sous. to“. f
530w can 532 . W
35:55:: M590 um ,
M.
d.
/.
(3...... M
/ ill! Ill,
.3355 :8ka wongcfibfiooe ”€333 033v
conmmfifimoofiuogofim x286 cob conmznfimo$EwE ”macaw—smog miscaaom .N 380838 mo 33% inasmznfimo .3 woutowofio
Pa moxm .hooméoow macaw £88m 365 588322 05 888 mtfizmuou .N we 538:8 8.“ 8% Row mo :owaooq .26 833m
232
Flgure 6.1b. Adult female I. scapularis questing on forest understory vegetation
233
Figure 6.2. Infection of 1595 adult I. scapularis with Borrelia spp., Anaplasma
phagocytophilum, and Babesia spp. organisms, and co-infections thereof, from across the
Midwestern United States, 2006-2007. Error bars are the standard error of the mean
prevalence of all sites within each establishment group.
60 l
50 -1
401
09
O
N
O
B. burgdorferi prevalence
0+-
.3
O
l
Long-established Intermediate-
established
234
l .,
Recently-invaded
Other pathogen prevalence
D B. burgdorferi
I B. miyamotoi
I A. phagocytophilum
III Ba. microti
I Ba. odocoilei
El co-lnfecfions
Figure 6.3. Variation in proportion of ribosomal spacer (RST) types 1, 2, and 3 of B.
burgdorferi across the Midwestern United States, 2006-2007.
-RST1
CJRST 2
CHRST 3
235
Figure 6.4. Rarefaction analysis of the influence of sample size of detected strain richness
of B. burgdorferi within adult I. scapularis across the Midwestern United States, 2006-
2007. Predicted mean strain richness (intervals are standard deviation) is modeled at
incremental sampling for each tick establishment group until the observed datapoint is
reached. The observed datapoint for each group is plotted as the final datapoint on each
curve.
20 i
i
1
15 A;
w l
.E ’
1!; ‘ 5
m a
<5 1° 5‘
z I
y -I— Long-established
5 + -t—|ntennediate-established
—-— Rece ntIy-inv aded
0 V" "i" 7‘" "#T_'#‘#—""l ’F "l”* ' ‘fil ‘—
0 50 100 150 200 250 300
No. sequences determined
236
dnohw “cogmznfimo some 553» :5on a ma wommofixo E :65 :08 mo 228882qu 2: .328m 885::
S wowmnm mo a2 Q8 05 we. 23E 2: 3 Baggage 03 $3: .3 8 2834 mo £538 2: E mcozmcwmmow m 98 ,m ._ 98cm Ema
dime mg I m3 2: 8 wcmgflcg mcmwbm 60:82 Emsogofi Ho: 803 @8322. 53> wficfiwon mfigm =m ”.mOr 53» Emma voom
.130 3:55 5 3:88 £283on 89323: .03? no ow 5:3 3:0ch 2: 9 :6: :39? 8a @8823,” 25: «m2 mmbficon
05 E 550on 88330 85 83883 05 22:3 5 moo: 8823c mo 898583 “EH ..mooméoom £83m 3ch 528332 05
$88 Sow 860:8 $90293 mg twbohmxza .m mo monastic 5:033on cam Emcwoina wsmfiomiosnwmoc 388:: 6.0 oSmE
237
; a s l
«6 o o md
Pep New «.9 03
«.9 o 0 ad
0.0.. «.5 02 0.9
Q.” 0.0 Dd? N4.
oi 9m 0 rd
«6 o o no
.. w 3 3 3
3 N9 2.. we
r. p o o 5
We m6 Ym Na
«d o 3 o
a... o o 3
N6 o V... o
3 o o 3
06 o o .1.
N6 0 o m.o
a.» N9 3 5
«a 3 o o
0.: o Nam :1.
’06 o ...v v6
.2 3...»... 2.3388 33358
i $3.325 .8228... .93..
353.8 .. 6 2.55386
<6 $0. romp
._._ «0952.5 ll.
an we. 8
02... mo. Iljma
< 3032! L
(v we.
m .832: lllfi
.2 3.5.2:
o .832:
5' exonuclease for proofreading. Thus, there is a high error rate due to
misincorporation of bases during DNA replication (Strachan and Read 1999). Using this
enzyme, a 1 kb sequence that has gone through 20 cycles of duplication is estimated to
have an incorrect nucleotide resulting from a copying error in 40% of the newly
generated DNA strands. The final PCR amplicons will therefore be a mixture of very
similar, but not identical, DNA sequences (Strachan and Read 1999). Despite this PCR
error, however, DNA sequencing of the total amplicons is likely to yield the correct
sequence because the incorporation of incorrect bases in the final sequence is at random.
Thus, for each base position in the sequence, the contribution of one spurious base on one
or more strands will be overwhelmed by the contributions of the majority of strands
which have the proper base (Strachan and Read 1999). Furthermore, I have employed in
recent years a high-fidelity DNA polymerase with associated proofreading activity, such
that PCR amplicons have fewer spurious apparent mutations (Cline et al. 1996).
(ii) Regarding sequencing error, nearly all of the B. burgdorferi strains reported
herein have had a sequencing reaction run at least twice (forward and reverse strands, or
forward strand twice). The resulting sequences were aligned, and all of the unique
polymorphisms that set each novel strain apart from the common strains were apparent
on both sequences. If the polymorphism was introduced as a sequence error, the chance
of that same error occurring at the same site within independently-generated sequences
would be very unlikely.
258
Implications for the Lyme disease system in Michigan
The Vectors
Continued monitoring of 9 core sites along the coastal and inland transects will
allow for detailed description of the process of establishment. Our standardized and
repeated sampling at sites not only where I. scapularis was found, but also areas well
beyond its apparent distribution, has allowed for us to detect invasion and establishment
of ticks into new areas. This dataset provides a rare example of actual tracking of a
biological invasion as it occurs, as most medically important species invasions are
detected only once disease results. The 9 sites include a longstanding hotspot of invasion
(Van Buren), sites which have experienced measurable annual increases in tick density
(Duck Lake, Orchard Beach), sites which have been colonized but with no detectable
increase yet (Sleeping Bear Dunes, F ort Custer, Ionia), and sites which were not
colonized as of the 2008 field season (Lux Arbor, Rose Lake, and Proud Lake; although
additional 2009 sampling detected low-density I. scapularis at Lux Arbor). Though not
described in this dissertation, limited sampling at Proud Lake near Detroit in 2009
resulted in no I. scapularis findings. The inclusion of this site in future standardized
surveillance is suggested so as to stay ahead of the detected invasion front, as well as to
accumulate Detroit-area data that will be of use in addressing the concerns of a growing
number of people in that area who are interested in the local presence of ticks.
1 also suggest an assessment of blacklegged tick density at additional intermediary
sites within the coastal invasion zone as well as paired sites at intervals in the eastward,
inland direction from these sites. These coastal data will be useful in determination of a
rate of invasion of ticks. Such data can then be modeled across a series of different
259
scenarios of host-mediated invasion (i.e. migratory bird dispersal of ticks; deer dispersal;
small mammal dispersal) to identify a most likely mechanism (or combination thereof) to
explain the field observations. A series of biotic and abiotic environmental features
(including host densities and climactic conditions) at each site can be measured and
compared as a first step in addressing differences between coastal and inland landscapes
that may facilitate or hinder tick invasion and establishment.
Evidence of blacklegged ticks not associated with the southwestern invading
population or the Menominee County population has arisen in our checks of deer
harvested during the mid-November shotgun season (data not reported herein). In 2007,
were found three individual deer that each harbored a single adult I. scapulars from
unexpected areas: one infested deer was harvested in St. Charles of Saginaw County in
the east-central part of the state, and two infested deer were harvested in Gladwin County
in the northeastern Lower Peninsula. Epidemiologically, there is a difference between an
established population (defined by the Centers for Disease Control and Prevention by
detection of at least six individuals or at least two life stages) and a reported population
(one to a few ticks reported from a focal area, which alone does not indicate that the
environment is suitable for tick survival and reproduction). Whether the infested deer or
other such observations reflect ticks that have dropped off from mobile hosts into
unsuitable habitats (failed invasion), an early-invasion process in suitable habitats, or
established, yet previously undetected, populations is yet to be determined.
I therefore suggest that future research include a standardized survey for
established ticks at a level of resolution that is sub-county level to produce a very useful
tick distribution map. This project could be conducted with the assistance of groups
260
including the Michigan Department of Community Health (MDCH) and Michigan
Department of Natural Resources and Environment (MDNRE) in order to provide the-
scale field sampling over a broad area but within a short period of peak tick activity. This
exercise may also identify other independent foci from which invasion may be studied.
The Pathogen
A unique ecological scenario is apparent in a focal area in southwestern Michigan
field site, where I. scapularis has not yet been observed, but repeatable B. burgdorferi
infections of bird-associated ticks, rabbit-associated ticks, and rabbit tissues have been
detectable. This work has been made possible through our collaborations with bird
banders at Pitsfield Banding Station, who process tens of thousands of birds per season,
and provide us with thousands of ticks. The effort to obtain an equivalent sample size
without this collaboration would be near impossible. While this dissertation reports on
samples acquired in 2004-2007, we have already logged and begun identifying samples
collected in 2008-2009, and 2010 sample collections are underway. 1 suggest that future
research maintain this collaboration as a priority. This long-terrn data set will be quite
interesting, as this focal site is on the trajectory of invasion by I. scapularis. We have the
potential to determine the response of the current tick, wildlife, and pathogen
communities to the invasion and establishment by I. scapularis. Most interesting, we
may be able to understand how invading strains of the pathogen interact with cryptic
strains. In the meantime, while labor intensive, it may be quite valuable to attempt to
meet the gold standard of demonstrating pathogen presence in this area by obtaining the
cryptic pathogen in culture. To do this, 1 suggest that not only should we attempt to
261
culture from bird-associated ticks, but also ticks and ear biopsies obtained from rabbits,
as these sample types are associated with higher infection prevalence at the cryptic site.
Across our studies of invading, cryptic, and established B. burgdorferi in
Michigan and the Midwest, we detected unprecedented amounts of diversity at the IGS
locus, and rarefaction curve analysis suggests that true population diversity is greater than
what we detected. Characterization of additional genes of B. burgdorferi using an
approach such as multi-locus sequence typing will allow for better characterization of the
diversity of strains, and relationships of our strains to those reported from endemic foci
elsewhere, which may provide insight as to the origin or strains and mechanisms of
invasion. Furthermore, given the diversity of strains in the Midwest, future research
should address the efficacy of currently available human and canine diagnostic assays for
detection of the panel of Midwestern strains.
Implications for Public Health
There are direct human and canine health implications of this invading disease
- system. A significant increase in confirmed cases of human Lyme disease (R2 = 0.40; 1-
tailed P = 0.015) has occurred within the zone of active 1. scapularis invasion in
southwest Michigan, and this rate of increase is greater than that reported from the state
of Michigan as a whole (Figure 7.1). From 1996 (when Lyme disease became reportable
in Michigan) through 2009, incidence in the 14-county region peaked at 1.4 cases per
100,000 people (E. Foster, Michigan Department of Community Health, pers. comm.)
Though increasing, incidence in the invasion zone remains substantially less than the
average annual incidence in the 10 endemic states from which 93% of human Lyme
262
disease across the US. is reported (Connecticut, Delaware, Massachusetts, Maryland,
Minnesota, New Jersey, New York, Pennsylvania, Rhode Island, and Wisconsin). In
these states, an average of 29.2 cases per 100,000 population was reported for 2003-2005
(Bacon et al. 2007), and reported annual rates for a majority of these reference states
were relatively stable during 1992—2006 (Bacon et al. 2008). Given the low reported
incidence in the invasion zone, one must also be cognizant of the ‘iceberg phenomenon’
of human medicine, which is used to describe the common pattern whereby human
patients that report their clinical symptoms of disease represent only a small fraction of
the true extent of disease in the population (Lynch et al. 1987). Thus, the dynamic status
of blacklegged ticks and their infection reported in this dissertation from 2004-2008 in
lower Michigan thus represents an invasion in its early stages, and dissemination of these
results now can have a positive effect in reducing disease risk in the future.
I predict that blacklegged ticks are currently more widespread than we know. The
future may include not only a continued eastward expansion of ticks as a continuation of
the invasion we have monitored in western Michigan, but also expansion from other yet-
undocumented foci. I predict that human anaplasmosis will soon be diagnosed for the
first time in Michigan, as will babesiosis and Powassan encephalitis. We have detected
the pathogen responsible for anaplasmosis within the invading ticks, though its detection
within humans may be delayed due to a lack of specific testing when a human presents
with the generic symptoms it causes.
While the density of blacklegged ticks in the invasion hotspot at Van Buren is as
high as that which characterizes Lyme disease endemic areas of the Northeast, the
infection prevalence of ticks is lower, and nymphal infection prevalence (NIP; an index
263
of human risk) is therefore lower. In contrast with the significant peridomestic exposures
that occur in the Northeastern US, recreational exposure is probably common in the
Michigan invasion zone, which is dominated by recreational areas, camps, vacation
homes, and rural communities. Michigan is thus likely to export cases of B. burgdorferi-
infected visitors that were exposed recreationally in the invasion zone (as has been
documented at a children’s summer camp in summer 2009, E. Foster, pers. com.), and
there may be a time lag before the true public health impact of this invasion is realized.
Vigilance for ticks must be emphasized now so we may be prepared as risk
increases. Local or community-scale strategies to reduce disease risk should be
discussed, including educational outreach campaigns to teach personal protection
measures, and also tick management techniques that are currently being used on a limited
scale in the northeast, such as acaricide application to vegetation, manipulation of
vegetation to create tick—free transition zones around homes, and methods to reduce tick
abundance on deer or mice. Such strategies must be implemented in a manner to
preserve the strong interest in outdoor recreation in the state. In the end, Lyme disease is
preventable given adequate education of the public and medical community, and so
outreach campaigns should be a priority.
264
Figure 7.1. Incidence (number of reported cases per 100,000 population) of Lyme disease
in humans in southwest Michigan (light bars) and the whole state of Michigan (dark bars)
in 1996-2009. Case numbers shared by E. Foster, Michigan Department of Community
Ikmhh(penrconun).
.3
a:
I Southwest Michigan R2 = 0.403 P = 0.015
A
h
I All Michigan R2 = 0.452 P = 0.008
P P P r‘ r‘
h 03 00 O N
Wharf? m
Incidence per 100,000 people
it.
-=. 3.155331}!
9
o
1996
265
References
Cardo LJ (2006). Leishmania: risk to the blood supply. T ransfitsion 46:1641-1645.
Cline J, Braman J, and Hogrefe H (1996). PCR fidelity of pfu DNA polymerase and other
therrnostable DNA polymerases. Nucl. Acids Res. 24:3546—355 1.
Dink-Wasser MA, Gatewood AG, Cortinas MR, Yaremych-Hamer S, Tsao J, Kitron U,
et al. (2006). Spatiotemporal patterns of host-seeking Ixodes scapularis nymphs
(Acari : Iodidae) in the United States. Journal of Medical Entomology 43:166-
176.
Foster ES (2004). Ixodes scapularis (Acari: Ixodidae) and Borrelia burgdorferi in
southwest Michigan: population ecology and verification of a geographic risk
model. Masters Thesis, Michigan State University, East Lansing.
Hinckley AF, O'Leary DR, and Hayes EB (2007). Transmission of West Nile Virus
Through Human Breast Milk Seems to Be Rare. Pediatrics 119:e666—67l.
Iwarnoto M, Jemigan DB, Guasch A, Trepka MJ, Blackmore CG, Hellinger WC, et al.
(2003). Transmission of West Nile Virus from an Organ Donor to Four
Transplant Recipients. N Engl J Med 34822196-2203.
King LJ, Marano N, and Hughes J M (2004). New partnerships between animal health
services and public health agencies. Rev Sci Tech 23:717-725.
Pealer LN, Marfin AA, Petersen LR, Lanciotti RS, Page PL, Stramer SL, et al. (2003).
Transmission of West Nile virus through blood transfusion in the United States in
2002. N Engl J Med 349:1236-1245.
Schantz PM, Steurer FJ, Duprey ZH, Kurpel KP, Barr SC, Jackson JE, et al. (2005).
Autochthonous visceral leishmaniasis in dogs in North America. Journal of the
American Veterinary Medical Association 226:1316-1322.
Steere AC, Malawista SE, Snydman DR, Shope RE, Andiman WA, Ross MR, et al. .
(1977). Lyme artiritis- epidemic of oligoarticlar arthritis in children and adults in
3 Connecticut communities. Arthritis and Rheumatism 20:7-17.
Strachan T, and Read AP (1999). Human molecular genetics 2. BIOS Scientific
Publishers Ltd., Oxford, U. K.
266
S
E
R
A
R
B
L
Y
T
S
R
E
V
N
U
E
T
A
Y.
S
N
A
G
H
M