A.
c1

,JL
.. .‘ “

«Elsa -
, . . .
'll

 

.§‘~’2x., "

-.i'

v "‘.
a; H

. .a
f

t ;
.
'r.‘
.u'

‘ 'r ". a

- u in.

L “32%“
”Pg" “ ”3‘-

..;JJ‘.:‘ 1‘; v’w
., .

«nu-- "

 

has

VERSITY LIBRARIES

2\\\\\\l‘.‘l‘l Ill

MIIGCH

\ INHIBIT!“ Ill

319

 

 

 

 

 

 

 

 

 

 

 

 

 

I

 

This is to certify that the

dissertation entitled

SERUM GANGLIOSIDES: ANALYTICAL METHODS AND
ANALYSIS OF SERA FROM BREAST CANCER PATIENTS

presented by

DOUGLAS ANDREW WI ESNER

has been accepted towards fulfillment
of the requirements for

Ph. D. BIOCHEMISTRY

degree in

ms 6 Kata/£7

Majoy rofessor

 

 

Date 1/4/94

 

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

 

 

LIBRARY

University

 

Michigan State

 

 

PLACE Ill RETURN BOX to remove thie checkout from your record.

TO AVOID FINES Mum on or Moro date duo.

 

DATE DUE DATE DUE

DATE DUE

 

 

‘-

 

;_-:l|

 

 

 

 

 

 

 

 

 

 

 

_—_-_l

 

 

FT—

MSU le An Affirmative Action/Equal Opportunity lnetltuion

 

 

 

SERUM GANGLIOSIDES: ANALYTICAL NIETHODS AND ANALYSIS OF

SERA FROM BREAST CANCER PATIENTS

By

Douglas Andrew Wiesner

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1993

ABSTRACT

SERUM GANGLIOSIDES: ANALYTICAL METHODS AND ANALYSIS OF
SERA FROM BREAST CANCER PATIENTS

By

Douglas Andrew Wiesner

Gangliosides are sialic acid-containing glycosphingolipids that are ubiquitous to
the mammalian plasma membrane. Alterations in gangliosides in the tissue, and in
some cases, the sera, due to aberrant glycosylation, have been associated with
oncogenesis. These changes in the ganglioside profiles may be useful in the clinical
diagnosis, prognosis and monitoring of the disease. In this study, we examined the
profiles of sera from breast cancer patients and an appropriate control group of female
to determine the nature of these profiles.

In order to adequately analyze the ganglioside profiles, new protocols for the
isolation and detection of gangliosides needed to be developed. I found that
modification of a three phase partition system was useful the isolation of gangliosides
from as little as 1 ml of sera. For the analysis of the ganglioside mixtures, a system
of 2-dimensional high-performance thin-layer chromatography and densitometry with
an image analyzer was well suited for our analytical needs. A derivitization
procedure for the analysis of components of gangliosides by gas chromatography was
developed for component analysis of the ganglioside mixtures.

The gangliosides of sera of 19 women with breast cancer and 16 women of an
appropriate control group was analyzed by two-dimensional high-performance thin-

layer chromatography gas chromatography. We found a significant increase (P < .05)

Douglas Andrew Wiesner
in the level of the total circulating gangliosides of the cancer sera over that of control.
In addition, we noted an increase in the level of complex (highly glycosylated)
gangliosides as well as the ratio of b-series gangliosides over the a-series gangliosides
in the sera of breast cancer patients. These results suggest a change in the
biosynthetic pathways of gangliosides associated with cancer. There was no alteration
in a specific ganglioside that could be classified as a tumor-associated antigen.

Free sialic acid and the activity of sialidase determined in the sera or plasma of both
groups to determine any possible contribution either might have in breast cancer. The
free sialic acid levels were found to be unchanged when examined with a modified
thiobarbituric acid assay. No sialidase activity was detected in either group when

assayed with a coupled enzyme assay.

To my parents, who taught me to work hard to achieve my goals,
yet also to enjoy life.

To my fiancee, Alexandra, for reminding me everyday.

iv

ACKNOWLEDGMENTS

I would like to thank my mentor, Dr. Charles C. Sweeley; "Chuck" to his
friends. His passion for research and discovery is contagious and inspiring. For
instilling in me this passion, and showing me how cool research can be, I say,
"Thanks, Chuck!"

I would also like to thank the following people:

My advisory committee, Drs. Richard Anderson, Nicolay Dimitrov, Rawle
Hollingsworth, and William Wells for their assistance and discussion as well as their
friendship.

Dr. Lyla Melkerson-Watson, for taking me under her wing and showing me
the ropes and where the Peanut Barrel was.

Bev Chamberlin and Dr. Zhi-Heng Huang for their assistance and knowledge
in the realms of gas chromatography and mass spectrometry.

Dr. Tom Deits, for his computer and bio/chemical expertise and friendship.

To the fantastic secretarial and administrative support staff, Diane, Julie,
Mary, Carol 8., Vicki, Susan, Frank, Mindy and Patty, who always had a smile for
me and were willing to aid me in some mess I had gotten myself into. Special thanks
to Carol Mc. , for her editorial services and assistance (asked for and otherwise), her

computer and especially her comradery.

The Hilberts, for their love and support, Sunday dinners, editorial services,
but especially their daughter’s hand.

Drs. Lyla Melkerson-Watson, Zhi-Heng Huang, Musti Swamy, William Hare,
and Kiyoshi Ogura, as well as Beverly Chamberlin, Bao-Jen Shyong, Misa Ogura,
Rob Dodds, Amy Poyner, Amanda Poxon, and Barb Kuhns, who are the members of
the Sweeley Lab, past and present, for their assistance, antics, dinners, craziness,
excellence, professionalism, good ideas and good humor. A perfect research
environment!

To all the folks down at the Barrel (as well as those who couldn’t make it),
Beta and Chris, Mike W., Sue, Leslie, Rich, Doug, John K., Lee, Jamie, Eric,
Paula, Mike D., Peter, Rath, John 5., Jim, Marty, Cindy, Angie, Andrea, Dave
Lewis, Kevin, Carla, Al, Stacey, Beth, Dave Levere, Liz, Barb, Carol, Rob and
Kerry, Ben and Jean and ..., Chuck and Mary and Paul, Luc and Odette, Steve, and
everyone else in the department for making this a great experience. Next time, you
are all coming to my house! Doug VI?

Lastly, the night crew, George, Nick and the rest, and the Phantom, for

keeping me company in the wee hours.

THANK YOU

vi

TABLE OF CONTENTS

LIST OF TABLES ....................................... x
LIST OF FIGURES ...................................... xii
LIST OF ABBREVIATIONS ............................... xiv
Chapter 1. Literature Review .............................. 1
Introduction ...................................... 1
Ganglioside Structure and Metabolism ....................... 2
Isolation of Gangliosides ............................... 11
Characterization of Gangliosides .......................... 13
High-Performance Thin-Layer Chromatography ............. 13
High-performance Liquid Chromatography ................ 16
Gas Chromatography and Mass Spectrometry .............. 17
Gangliosides in Cellular Interaction, Signal transduction, and Oncogenesis . 19
Cellular Interaction .............................. l9
Signal Transduction and Cell Growth ................... 21
Oncogenic Transformation .......................... 22
Sialic Acid and Sialidase ............................... 26
Breast Cancer and Gangliosides ........................... 27
Statement of the Problem ............................... 29
References ........................................ 31
Chapter 2. Isolation of Gangliosides from Serum or Plasma .......... 42
Abstract ......................................... 43
Introduction. ....................................... 44
Materials and Methods ................................ 46
Materials .................................... 46
Isolation of Total-Lipid Extract ....................... 46
Isolation of Gangliosides by Column Chromatography ......... 47

Isolation of Gangliosides by Diisopropyl Etherzn-Butanol:
Water Partition ................................. 50
Results and Discussion ................................ 51
Modification of Total-Lipid Extraction .................. 52
Chromatographic Isolation of Gangliosides ................ 52

Isolation of Gangliosides by Diisopropyl Ether:n-Butanol Partition. . 56

vii

Conclusion ........................................ 58
References ........................................ 6 1

Chapter 3. Characterization of Gangliosides by Two-Dimensional High-

Performance Thin-Layer Chromatography (HPI‘LC) ........ 62
Introduction ....................................... 63
Materials and Methods ................................ 64

High-Performance Thin-Layer Chromatography ............. 65

Visualization of Gangliosides .................... 67

Two-Dimensional Mapping ......................... 67

Image Densitometry .............................. 68
Results and Discussion ................................ 68

Identification and Quantitation of Gangliosides from Human Serum . 79
Conclusion ........................................ 79
References ........................................ 81

Chapter 4. Microscale Analysis of Glycosphingolipids by Methanolysis,

Peracetylation and Gas Chromatography ............... 82
Abstract ......................................... 83
Introduction ....................................... 84
Materials and Methods ................................ 85

Materials .................................... 85

Acid-Catalyzed Methanolysis and Peracetylation. ............ 86

Gas Chromatography ............................. 87

Derivatization Time and Temperature Optimization .......... 87

Isolation of GM3 from Plasma ....................... 87
Results and Discussion ................................ 88

Gas Chromatography of Peracetylated Methanolysis Products

of GM, ..................................... 89

Time and Temperature Optimization of Methanolysis Conditions . . 95

Stability of Peracetylated Methanolysis Products. ............ 95

Glycoconjugates Containing Fucose and N-Acetylglucosamine . . . . 98
Component Analysis of OMB Ganglioside Isolated from Human

Plasma. .................................... 101
Conclusion ....................................... 102
Acknowledgements .................................. 104
Footnotes ........................................ 105
References ....................................... 106

Chapter 5. Potential Methods for Characterization of Gangliosides ..... 108
Oligosaccharide Analysis with HPAE HPLC-PAD ............... 108

Experimental ................................. 109

Discussion ................................... 110
Radioisotope labeling, I-IPTLC, and Autoradiography ............ 111

Experimental ................................ 112

Discussion ................................... 112

viii

References ....................................... 1 15

Chapter 6. Analysis of Circulating Gangliosides of Breast Cancer Patient

and an Appropriate Control Group ................... 116
Abstract ........................................ 117
Introduction ...................................... 118
Materials and Methods ............................... 119

Materials ................................... 119

Isolation of Gangliosides from Serum and Plasma ........... 120

Two-dimensional High-Performance Thin-Layer Chromatography

(2-D HPTLC) of Gangliosides. ..................... 121

Gas Chromatography of Carbohydrates and Lipid Components of

Gangliosides. ................................ 122
Results ......................................... 123

High-Performance Thin-Layer Chromatography. .......... 123

Gas Chromatography. ........................... 130
Discussion ....................................... 144
Perspectives ...................................... 152
References ....................................... 155

Chapter 7. Sialidase and Free Sialic Acid in the Serum of Patients with

Breast Cancer ................................ 161

Abstract ........................................ 162

Introduction ...................................... 163

Materials and Methods ............................... 165

Materials ................................... 165

Coupled Enzyme Assay of Sialidase. ................. 165

4-MU—NeuAc Assay of Sialidase ..................... 166

Thiobarbituric Acid Assay of Free Sialic Acid. ............ 167

Preparation and Analysis of Serum and Plasma ............ 167

Results ......................................... 169

Sialidase .................................... 169

Sialic Acid .................................. 170

Discussion ....................................... 175

References ....................................... 181

Chapter 8. Discussion and Perspectives ....................... 184

References ....................................... 191
Appendix

Individual Patient Data ............................... 192

ix

Chapter 1.

Table 1.

Table 2.

Chapter 2.

Table 1.

Chapter 4.

Table 1.

Table 2.

Table 3

Chapter 6.

Table 1.

Table 2

Table 3.

Table 4.

LIST OF TABLES

Summary of Ganglioside Nomenclature ...........

Tumor-Associated Ganglioside Antigens ...........

Isolation of Circulating Gangliosides .............

Gas Chromatography of Peracetylated Methanolysis
Reference Compounds and Products from GMl

Ganglioside ............................

Carbohydrate Analysis of Reference Oligosaccharides
and GMl

Carbohydrate Analysis of an Equimolar Mixture of 6C
and Sialyldi Y-2 Gangliosides Plasma GM; Ganglioside

Summary of Ganglioside Nomenclature and Structure.

Average Ganglioside Content of Cancer Serum and a

Control Group Plasma. .....................

Averages of GC Analysis Peracetylated Methanolysis
Products of Plasma Ganglioside Mixture:Carbohydrates

and Long-Chain Bases ......................

Averages of GC Analysis Peracetylated Methanolysis

Products of Plasma Ganglioside Mixture: Fatty acids . . . .

24

..54

..92

..99

.100

. 128

129

140

141

Chapter 7.

Table 1.

Appendix

Table 1.

Table 2.

Table 3.

Sialic Acid Concentrations Found in 9 Breast Cancer

Patient Plasma Samples and 8 Female Control Sera . . . . 177
Patient Bio Data and TLC Data ................. 193
Patient GC Data: Sugars and Long-Chain Bases ....... 195

Patient GC Data: Fatty Acids .................. 196

xi

Chapter 1.
Figure 1
Figure 2
Figure 3

Chapter 3.
Figure 1

Figure 2

Figure 3

Figure 4

Chapter 4.

Figure 1

Figure 2
Chapter 6.

Figure 1

Figure 2a

Figure 2b

LIST OF FIGURES

GM2 Ganglioside ........................... 6
Biosynthesis of Core Glycolipid Structures. ........... 8
Biosynthesis of Gangliosides ..................... 10
Two-Dimensional Map of Gangliosides ............. 71

Standard Curve of Gangliosides Analyzed by 2-D HPTLC,

Based on Lipid-Bound Sialic Acid Content. .......... 74
Standard Curve of Gangliosides Analyzed by ID HPTLC,
Based on Lipid-Bound Sialic Acid Content. .......... 76
Standard Curve of Gangliosides Analyzed by 1-D HPTLC,
Based on Lipid-Bound Sialic Acid Content ........... 78
Chromatogram of the Peracetylated Methanolysis Products

from GMl Ganglioside ....................... 91
Methanolysis Time Course of GMI. ............... 97
HPTLC of Gangliosides ..................... 126
Box plots of 6M3 and Total Ganglioside Isolated from

Cancer Sera and Control Plasma. ................ 132
Box plot of Minor Gangliosides Isolated from Cancer Sera

and Control Plasma. ........................ 134

xii

Figure 2c Box plots of Important Ratios of Gangliosides. ...... 136

Figure 3 Gas Chromatograms of Peracetylated Methanolysis
Products of Gangliosides Isolated from Plasma of Serum. 139

Figure 4 Mass Spectrum of a-Acetoxytetracosanoic Methyl Ester
Analyzed by GC/MS. ....................... 143
Figure 5 Biosynthetic Pathway for Ganglia-Series Gangliosides . . . 147
Chapter 7.
Figure 1 Elution Profile of 1 ml of Plasma from a Control Subject
on a Sephacryl 5-200 HR Column. ............... 172
Figure 2 Elution Profile of 1 ml of Serum from a Cancer Subject
on a Sephacryl 8-200 HR Column. ............... 174
Chapter 8.
Figure 1 Possible Role of Gangliosides in Metastasis. ........ 190

xiii

LIST OF ABBREVIATIONS
Gangliosides abreviations, nomenclature, and structure - See Table 1, Chapter 1
Da - Daltons
Globo - Galal-4Ga161o4GchI-1Cer, Ganglio - GalNAcBl-4GaIBI-4Glc6l-lCer,
Lacto - Gala1-3G1cNAcBl—3Galf31-AGICB 1-1Cer, Neolacto - Gala1-4GlcNAc61-
3GalBl-401cBl-1Cer
DEAE - Diethylaminoethyl, HPLC - High-Performance Liquid Chromatography,
HPAE - High pH Anion-Exchange, PAD - Pulsed Amperometric Detection, HPTLC -
High-Performance Thin-Layer Chromatography, GC - Gas Chromatography, MS -
Mass Spectometry, FAB - Fast Atom Bombardment, CID - Collision-Induced
Dissociation, EIMS - Electron-Impact Mass Spectrometry, NMR - Nuclear Magnetic
Resonance
TMS - Trimethylsilyl, DMSO - Dimethyl sulfoxide
NGF - Nerve Growth Factor, PDGF - Platelet-Derived Growth Factor, EGF -
Epidermal Growth Factor
CM — Chloroform-Methanol, DIPE/ButOH - Diisopropyl ether/n-butanol
l- , 2-D - One-, Two-Dimensional, aq. - Aqueous
GSL - Glycosphingolipid, FAME - Fatty Acid Methyl Ester, HDA - Methyl
Heptadecanoate, Glc - Glucose, Gal - Galactose, GalNAc — N-Acetylgalactosamin,
GlcNAc — N-Acetylglucosamin, Fuc - Fucose, LCB - Long-Chain Base, NeuAc - N-
Acetylneuraminic Acid and Sialic Acid, Cer - Ceramide

LBSA - Lipid-Bound Sialic Acid

xiv

ABBREVIATIONS (Cont’d)

BSA - Bovine Serum Albumen , PBS - Phosphate Buffer Solution, HRP-B -
Horseradish Peroxidase conjugated to Cholera Toxin B-Subunit, 4-MU-NANA - 4-
Methylumbelliferyl-a-D—N—Acetylneuraminate

EIA - Enzyme-linked Imunoadsorbance

TBA - Thiobarbituric Acidl

A230,“, - Absorbance at 280 and 490 nanometers, respectively

XV

CHAPTER 1

Literature Review

Intrfiuction

Breast cancer is by far the leading cause of cancer among women and the
second most prevalent among all cancers after lung cancer. Women have a one-in-
eight chance of developing breast cancer over their lifetime. Although the mortality
rate of breast cancer has remained stable over many decades at 27 deaths per 100,000
women, the incidence rate has shown an annual 1% increase over the last century
(1). Survival after five years is close to 90% for those women with cancers caught at
an early stage when the cancer remains localized in the breast tissue ( 1). It becomes
obvious that early diagnosis and accurate staging of the tumor is key to a good
prognosis. Much research is devoted to finding new and better ways to characterize
the disease such that earlier diagnosis is possible and better prognosis follows.
Moreover, many breast cancers elude monitoring strategies, making their prognosis
difficult. New methods of monitoring these elusive types of cancers as well as better
means of monitoring the more detectable cancers is an ongoing struggle in the battle
against breast cancer.

A recent trend in the fight against all kinds of cancers is the analysis of
A glycosphingolipids associated with the cancers. Oncogenic transformation has long

been associated with an alteration in the composition of glycosphingolipids, both of

2

the neutral type and gangliosides, in the plasma membranes of cancer cells (2).
Numerous studies have indicated both direct and indirect roles of gangliosides in
tumorigenesis (reviewed in 3, 4). Furthermore, gangliosides have been shown to be
shed from these cancer cells into the local environment and eventually into the blood
stream (5-7). Some researchers have tried to take advantage of this phenomenon by
identifying the presence of unusual or increased levels of gangliosides in the plasma
of cancer patients as a circulating tumor marker (6, 8). This type of circulating
tumor marker can be extremely useful because it can be an indicator of the progress
and severity of the disease. It can be particularly advantageous because it only would

require a small aliquot of blood.

n 11 i nd Meta '

Over 300 different glycosphingolipids have been characterized with 79
different gangliosides being among them (9). Gangliosides are, by definition, sialic
acid-containing glycosphingolipids with molecular weights from 1000 - 5000 Da.
Gangliosides are typically amphipathic molecules, consisting of a hydrophobic tail and
a polar head group. The polar head group is an Oligosaccharide that is glycosidically
linked to the hydrophobic tail. The Oligosaccharide backbone found in mammals may
consist of varying proportions of glucose, galactose, N-acetylglucosamine, N-
acetylgalactosamine, or fucose. Sulfate can also be found as an O-linked ester. One
or more sialic acid residues are bound to this carbohydrate backbone, giving it a net
negative charge. It is believed that this net negative charge may confer gangliosides

with much of their bioreactivity (10). The hydrophobic tail, called cerarnide, consists

3
of a 10ng-chain sphingoid base with a fatty acid of varying length (14-26 carbons)

bound in an amide linkage. These biomolecules are ubiquitous components of the
extracellular aspect of the plasma bi-layer membrane in mammalian cells; the highest
concentrations are found in the brain. The structural similarity of ceramide to
diacylglycerol allows it to orient itself in the membrane with the carbohydrate portion
facing out into the extracellular matrix, often perpendicular to the orientation of the
hydrophobic tail (11). The composition of both the ceramide and the Oligosaccharide
have influence on the activity of the molecule. Figure 1 shows the structure of GM2
ganglioside.

The synthesis of gangliosides (nomenclature and structures are summarized in
Table 1) occurs largely in the Golgi apparatus by the step—wise transfer of sugars
from sugar nucleotides to the non-reducing end of the glycan portion of a
glycosphingolipid (or ceramide) by specific glycosyltransferases. Most mammalian
glycosphingolipids are found with four main core structures: globo (Galal-4GalBl-
4Gchl-1Cer), ganglio (GalNAcBl-46alBl-4G1cBl-1Cer), lacto (GalBl-3GlcNAcBl-
3Galfll-4Gchl-1Cer), and neolacto (GalBl-4GlcNAcB1-3Ga131-4Gchl-1Cer). Most
gangliosides have the ganglio core but are not restricted to this type of structure. The
biosynthetic pathways of glycolipid core structures are shown in Fig 2 and the
biosynthetic pathways of most gangliosides are shown in Fig. 3. Specific
glycosidases are responsible for the step-wise degradation of glycosphingolipids in a
reverse manner in which they were synthesized and occurs in the lysosomes.
Deficiencies of any of the glycosidases can block the catabolism and cause an

accumulation of lipid. Disorders such as Tay-Sach’s disease (12) and Gaucher’s

Table 1.

Summary of Ganglioside Nomenclature

 

 

Ganglioside Structure
6M4 NeuAca2-3GalBI-1Cer
precursor

LacCer GalBl—4Gchl-1Cer
a-series

GM3 NeuAca2-3GalBI-4GlcCer
6M2 GalNAcBl-4(NeuAcaZ-3)GalBl-4GlcCer
GM] Galfi1~3GalNAcBl-4(NeuAca2-3)GalBl-4GlcCer
GD 1. NeuAca2-3GalB l -3GalNAcBl -4(NeuAcaZ-3)Galfi l -4GlcCer
GT1. NeuAca2-8N euAca2-3GalB l -BGalNAcB 1 -4(NeuAcaZ-3 )GalB l -4GlcCer
b-series

GD3 (NeuAca2-8NeuAcaZ-3)Galfil-4GlcCer
GD2 GalNAcBl-4(NeuAca2-8NeuAcaZ-3)Galfil-4GlcCer
GD", GalBl-3GalNAcB14(NeuAca2-8NeuAca2-3)GalBl-4GlcCer
GT1 b N euAca2-3GalB 1 -3GalN AcBl -4(N euAca2-8NeuAca2—3 )GalB 1 AGlcCer
GQ“, NeuAca2-8NeuAca2-3Galfi l -3GalNAcB l 4(NeuAcaZ-8NeuAcaZ-3)Galfi l -4GlcCer
c-series

GT3 N euAca2-8NeuAca2-8NeuAca2-3Galfil-4GlcCer
GT2 GlcNAcBl4(NeuAca2-8NeuAca2-8NeuAca2-3)GalB1-4GlcCer
GT1 c Gall?1-3GlcNAcBl4(NeuAca2-8NeuAca2-8NeuAcaZ-3)Galfll-4GlcCer
GQl c N euAcaZ-3Galfi l-BGlcNAcfil4(NeuAca2-8NeuAca2—8NeuAca2-3)Gal[31—4GlcCer
lacto series

SPG NeuAca2-3GalB1-3GlcNAcBl-3GalBl-4GlcCer
2-6SPG NeuAca2-66alfil-3GlcNAcBl-3Galfll-4GlcCer
SLe' NeuAca2-3GalBl-3(Fucal-4)GlcNAc61-3GalBl-4GlcCer
SLe" NeuAca2-3GalBl-3(Fuca1-3)GlcNAcBl-BGalBl-4GlcCer

neolacto series
SNHC

Sialyl I

6C

saw-2

NeuAca2-3GalB 1-4GlcNAcBl-3GalBl-4GlcNAcBl-3Galfil-4GlcCer

(N euAca2-3GalB l 4GlcNAcB l -3 ,6)G 113 l 4GlcNAcB l -3GalB l 46ch
NeuAca2-6GalB l -4GlcNAcB l -3GalB 1 -4(Fuca l -3)GlcNAcB l -3GalB l 461ch
NeuAca2-3(Galfi l-4(Fuca l-3)GlcNAcB 1-3)zGalBl-4GlcCer

Figure l. GMZ Ganglioside. GalNAcB1-4(NeuAca2-3)GalB1-4Gchl-Cer.

 

 

 

mEESmu 02.283095

 

 

_ ,-
_ _ L
.22 _ 2am

 

o n
u 2
32V.
:0 oz: ufoz

O:

 

«Emoemcmo ~36

.u 33E

Figure 2. Biosynthesis of core glycolipid structures.

 

amamw 085
8 29.01neonaézsooaao a... 3.8

a... 83.2 / 829644584132on
Roda:

.8729.9.580522204133 >
/ 38.883 85w 890
3.8 sow—MEESSAéEZOG 87.96.45.938

33w 262.83“
bmo7320....Fm.amxn.~§.<52 32.8.8:
3206.3: lode:

 

 

 

«meow 5.0 8:30 #883
873819858-” Alma! 87582.58 8:39.38
8863
3.5: 19.8
373.83....” A a... 52.3.8 .8330
.w/ \M.

m92<mwo

makoamhm wmoo wh<¢o>zom¢<o o_n__._00>._@ “—0 m_mm:.—.z>m0_m « 9:5...—

Figure 3. Biosynthesis of Gangliosides.

 

 

10

200 1112.8 111.5 11.5
2.3 «.50 33220 >

 

£06 Alla—F0 AIIInFOG TuOG Alan—O
3.5 7.3.. «.50 7.32.3 >

 

2.8 Algae 111.26 1.1.20 111.20
6.52 .15 3.5 52on +

32690

more:

Stowe $30350

.5003 A .5026 In mEESmO

><>)_.._._.<n_ O_._.mz._.z>m0_m m0_m0_._02<0

ll

disease (13) have been attributed to the accumulation of GMZ ganglioside and
glucosylceramide due to defects in B-N-acetylhexosaminidase and glucocerebrosidase,

respectively.

Isolatng of fiangliosidg

Gangliosides can be isolated by first extracting all of the lipophilic material from the
biological tissue or fluid. Partitioning or column chromatography is employed to isolate the
ganglioside mixtures from the undesired lipid material, i.e. phospholipids, neutral
glycosphingolipids, neutral lipids, etc. The removal of salts and small molecular weight
contaminants is required at least once, and often twice, based on the particular method
chosen for the above steps. Total lipid extracts of biological materials are typically obtained
by dispersing the material in 10 to 20 volumes of varying ratios of chloroform and methanol,
typically 2:1 or 1: 1. It may be necessary to add a small bit of water if polysialylgangliosides
are sought. After adequate mixing or homogenization, the mixture is centrifuged or filtered
and the pellet extracted a second time. The supernatant fractions are combined and dried,
leaving the total lipid extract. Another method reported to extract gangliosides
quantitatively employs 9 vol of tetrahydrofuran:0.01 M phosphate buffer pH 6.8 (14).

To remove unwanted lipid material, one of two strategies is employed, that of
partitioning or column chromatography. Most early purifications involving partition used a
mixture of chloroform, methanol, and saline (15, 16) that created a biphasic solution in
which thegangliosides were recovered in the aqueous phase and neutral lipids partitioned
into the organic phase. While this type of preparation was good for crude separations,

substantial losses of less polar gangliosides, such as GM3, were seen. Furthermore, the

12

more polar neutral glycosphingolipids would partition with the gangliosides in the aqueous
phase, thus requiring further purification. These problems made it unusable for small
samples. Another method based on partitioning, developed by Iadisch et al. (17), employed
diisopropyl ether:n-butanol:water (4/2/3). This method was more suitable for smaller
samples with almost quantitative removal of phospholipids in two extractions with minimal
loss of gangliosides. Removal of salts and low molecular weight contaminants was then
achieved either by dialysis, size exclusion chromatography, or reverse phase-chromatography
(l8). Partitioning of these types leaves a semipure ganglioside mixture that is suitable for
many applications.

When a much purer preparation of gangliosides is desired, column chromatography is
employed. There are many ion-exchange methods based on diethylaminoethyl functional
groups conjugated to a solid support (19-21). Gangliosides bind to the column while neutral
and positively charged lipids are washed off. The gangliosides are eluted by adding salt to
the eluant. This step is typically followed by mild alkaline methanolysis to destroy any lipids
with O-linked fatty acids. After desalting by the methods described above, a column of silica
gel is used to remove sulfatides (21). The result is a fairly pure mixture of gangliosides.

Isolation of individual ganglioside species can be achieved by additional
chromatography on DEAE-Sephadex and silica gel (19—21), high-performance liquid
chromatography (22, 23), or preparative high-performance thin-layer chromatography.
Ganglioside mixtures purified by the above methods often are homogeneous with respect to
the Oligosaccharide moiety but heterogenous with respect to the ceramide portion. Reverse-
phase chromatography can be used to separate different species of a particular ganglioside

based on the hydrophobic tail heterogeneity.

13

ha ' i n 1‘ an li id

Characterization of gangliosides depends on the nature of the gangliosides and the
intent of the study. Mixtures of gangliosides are best analyzed by high-performance thin-
layer chromatography (HPTLC) employing numerous sensitive methods of visualization.
When a more homogeneous preparation is at hand, several different techniques are currently
available including different types of mass spectrometry (MS), high-performance liquid
chromatography (HPLC), gas chromatography (GC), proton and l3C-nuclear magnetic
resonance spectroscopy (NMR), as well as immunological techniques. These latter methods
are capable of providing much information about the nature of the ganglioside composition,

carbohydrate sequence, and linkage and anomerity of carbohydrates.

Hi gh-Perfomrance Thin-Layer Chromatography

Chromatography on high-performance thin-layer chromatography (HPTLC) plates is
widely used to separate mixtures of gangliosides and other glycolipids. HPTLC plates are
glass plates precoated with a 0.25 mm layer of uniform superfine silica gel, which has a
mean particle size of 5 pm and a mean pore diameter of 80 nm. The test sample is applied
near the bottom of the plate and the plate placed in a preequilibrated chamber with an
appropriate solvent to resolve the gangliosides of interest. The most common solvents
consist of a mixture of chloroform, methanol and water, containing either ammonia,
potassiumor calcium chloride, in ratios varying from 60/40/9 to 50/40/ 10, respectively. The
solvent is allowed to run the length of the plate, the plate dried, and the gangliosides

visualized by a colormetric spray or, more recently, by an immunological technique.

l4

Resorcinol-HCl (24) and orcinol-HZSO4 (25) are common colormetric sprays for the
detection of sialic acid residues of gangliosides and glucose residues of glycosphingolipids,
respectively. After heating, they react with the specific sugar to form a bluish purple spot.
For preliminary identification of gangliosides, the mobility and reactivity can be compared
with those of known reference compounds. A difficulty with this method is that more than
one ganglioside may have the same number of carbohydrates and such mixtures may be
difficult to separate. In addition, gangliosides with identical Oligosaccharide structures may
produce two or more bands due to differences in the ceramide moiety. These problems can
be overcome by altering the solvent system such that the bands of interest are better
resolved. Other techniques involve HPTLC of the acetylated derivative or use of alternate
types of plates, such as borate-impregnated silica gel. Alternatively, newer methods allow
for overlay techniques using antibodies developed for specific gangliosides or epitopes found
on gangliosides (26-28) or using lectins of the same nature (29). An aluminum-backed
HPTLC plate for this type of analysis is fixed with polyisobutylmethacrylate, immersed in a
primary antibody solution or specific lectin solution, then soaked in the second radiolabeled
antibody solution (27, 28) followed by autoradiography, or an avidin-biotin enzyme system
(30).

When glycosphingolipids on HPTLC plates have been visualized by conventional
methods using colormetric sprays, densitometric scanning methods have been developed to
quantify the spots (31—33). The resorcinol—positive plate is scanned with a scanning
densitometer in the transmittance mode with incident light at 580 nm. This gave linear
response in the range of 0.5-10 nmol with as little as 100 pmol capable of routinely being

detected (32). One study reported as little as 10 pmol of ganglioside being reproducibly

15

detectable. A more recent technique of quantification of spots on Chromatograms is that of
image densitometry (34). Chromatograms are scanned into a computer as a digital image of
1024 pixels by 1024 pixels. Specific software then computes the area of the spot and its
optical density, and integrates this information. This method of quantification is applicable
directly to resorcinol-positive thin-layer plates or to autoradiograms, which previously were
difficult to quantify. The method is linear down to a sample size of 10 pmol. Quantification
of individual gangliosides can be done by the method of Suzuki (35) in which gangliosides
are visualized with iodine, scraped from the plate, and quantified by a modified resorcinol
assay (24). This is also applicable for radiolabeled gangliosides by scrapping and counting
with a scintillation counter.

Two-dimensional HPTLC is often used when added resolution is necessary. A
ganglioside mixture is spotted in one comer of the plate, 1 cm from each edge, run in the
first dimension in a suitable solvent, then allowed to dry. The plate is then turned 90° and
run in the second dimension with a different solvent. This method takes advantage of the
differential resolving properties of two different solvent systems. Many qualitative and semi-
quantitative studies have been conducted to analyze interspecies differences in various tissues
(36-38). One reason for the not-so—wide-spread use of 2-D HPTLC, despite its resolving
power, is the lack of a suitable method of detection and quantification of the spots. One
study (36) used sequential scanning densitometry across the entire plate, summing the areas
of peaks corresponding to the same spot. This requires a special sequential scanning
densitometer. Linearity and sensitivity has been reported to be from 0.1 to 6 nmol. Image
densitometry, as described above, makes the quantification of spots run in two—dimensions as

simple as in one—dimension. Furthermore, it is applicable to direct analysis of resorcinol-

16

positive plates as well as autoradiography from radiolabeled lipids or by the many overlay

methods as described above (34).

High-performance Liquid Chromatography

Most methods involving the use of HPLC in the analysis of the Oligosaccharide
structure of gangliosides require the release of the intact Oligosaccharide from the ceramide
portion. Classically, this has been done by oxidation of the olefinic double bound by
ozonolysis (39) or osmium tetroxide (40) followed by alkaline-induced degradation releasing
the intact Oligosaccharide. Another chemical method involves the use of trifluoroacetolysis
(41). Recent characterizations of endoglycoceramidase (42), or similarly ceramide-glycanase
(43, 44) have made enzymatic release of the Oligosaccharide possible. Once the
Oligosaccharide has been freed, many different protocols involving HPLC are available,
depending on the nature of the Oligosaccharide. These are reviewed in 45. A difficulty in
the analysis of carbohydrates, whether they be mono- or Oligosaccharides, is the ability to
detect them in a simple manner with some degree of sensitivity. They have no natural
absorption in the ultraviolet or visible regions, nor can they be detected by fluorescence.
While other physical types of detection are possible, they are neither specific nor sensitive.
To overcome this, reacting the reducing end with a fluorophore such as pyridylamino
derivatives (46, 47) or ultraviolet-absorbing 1-phenyl—3-methy1-5-pyrazolone derivatives are
often employed. Another, more current method for the separation and detection of
underivatized mono- and Oligosaccharides involves the separation on high-pH anion exchange
chromatography with electrochemical detection with a pulsed amperometric detector (23, 48-

52).

17

Gas Chromatography and Mass Spectrometry

Gas chromatography (GC) is a useful tool for the estimation of components in
glycolipids. GC can give information on the Oligosaccharide composition and structure as
well as sphingoid base and fatty acid moieties. Most methods require some type of
derivatization in order to make the compounds volatile enough for analysis with gas
chromatography. Small Oligosaccharide molecules (2-3 sugars) released from glycolipids can
be analyzed by GC as permethylated alditols (53). Larger Oligosaccharides are best resolved
by HPLC as described above. GC is most useful for the estimation of the components of
glycolipids. Monosaccharides are released from the Oligosaccharide either by acid-catalyzed
hydrolysis or methanolysis. Hydrolysis is achieved using sulfuric acid (54), hydrochloric
acid (55), or acetic acid with sulfuric acid (56), followed by conversion to alditol acetates
(57, 58) and analysis by GC. This method is useful in the analysis of mixtures of free
sugars because single peaks are observed for each saccharide. Problems with this type of
determination include the need for different conditions of hydrolysis for different
carbohydrates and, often, the complete destruction of sialic acid. Furthermore, hydrolysis
often requires complex, time-consuming work-up procedures not suitable for small samples.

Methanolysis in dry methanolic hydrochloric acid (HCl), followed by
trimethylsilylation and analysis by GC (59-66), is often used instead of hydrolysis.
Typically, a sample of glycolipid is methanolyzed with 0.75 N methanolic HCL for 16—24 h
at 80 °C followed by neutralization. Fatty acids ean be removed by partitioning with hexane
and can be run separately if desired. Trimethylsilylation of the methyl glycoside with a
mixture of hexylmethyldisilazane, trichlorosilane and pyridine, now conveniently found in

commercial preparations, produces a derivative suitable for GC. The derivatized

18

carbohydrates are run on a non-polar gas chromatography column, producing multiple peaks
due to different anomers of the methyl glycoside. Quantification is based on the area of all
or one of the resultant peaks, relative to an internal standard, often mannitol. Identification
is based on comparison of retention times of known standards and retention indices. A
recent microscale method has reduced the sample volume and methanolysis time to 1.5 h
(67). This method also makes use of peracetylation to produce a stable derivative with useful
chromatographic properties.

Many of the above methods can and have been used in conjunction with mass
spectrometry. Pure monosaccharides as well as mixtures can be analyzed with gas
chromatography and electron impact mass spectrometry (GC-EIMS) as TMS (60, 68-72),
permethylated (73-77) and peracetylated derivatives (57, 63, 73). Mass spectral data give
more substantial evidence of the presence of a particular saccharide over basic comparisons
of retention times with known standards, which is the case with gas chromatography alone.
Intact Oligosaccharides, released from the ceranride, can also be analyzed by GC-EIMS but
this approach often requires previous derivatization such as permethylation (78). Structural
information is possible with the use of methylation analysis followed by GC-EIMS.
Oligosaccharides are released from the ceramide portion by a method stated above and
subjected to methylation with methyl iodide and methylsulfinylmethanide in dimethyl
sulfoxide (DMSO) (79). This procedure methylates all free hydroxyl groups, acetamide
residues, and earboxyl groups. This permethylated product is hydrolyzed to form partially
methylated monosaccharides. This mixture is reduced with sodium borohydride, producing
partially methylated alditols, which are acetylated with acetic anhydride. The products

contain O—acetyl groups at positions of substitution in the Oligosaccharide and on carbons

19

attached to ring oxygens. These products are readily analyzed by comparison of their
retention times and fragmentation patterns obtained with GC-EIMS. Information on positions
of glycosidic linkages and ring size of sugar substituents is possible. Methylation analysis is
widely used in the analysis of all types of glycoconjugates (60, 76, 80). GC-EIMS of intact
Oligosaccharides following permethylation is also possible (78) and yields information on the
sequence as well as the linkages.

Both Sweeley and Dawson (81) and Samuelsson and Samuelsson (82) used GC-EIMS
to analyze TMS derivatives of intact glycolipids. Information was obtained on the sequence
of sugar and ceramide, yet no molecular ion was seen. The analysis of glycolipids as large
as tetrahexosylceramide (83) was possible. Permethylation of glycolipids analyzed by GC-
EIMS (77, 84) provided better information on the exact ratios of sugars and sequence up to
eight carbohydrates. Urdal and Hakomori (85) used direct probe EIMS instead of GC-EIMS
of the permethylated derivative of a tumor associated ganglioside to determine differences in
the fatty acid moiety. The development of soft ionization techniques such as fast-atom
bombardment mass spectrometry (FAB-MS) (86) with collision-induced dissociation (22, 87,
88) permitted analysis of underivatized gangliosides. Costello (88) has developed a

systematic nomenclature for the identification of fragment ions observed with CID-FAB—MS.

11.15-151'1‘9 n '_1 -1. '11 AIM .41-1I 11,1! u-vn 1n! 0: '1' '1'»-
Cellular Interaction
Gangliosides and other glycosphingolipids found on the surface of the plasma
membrane have been associated with several cell binding events. Several studies have found

cell-surface gangliosides to be integral components in cell-cell adhesion. The compaction of

20

morula, one of the first cell-binding events in ontogenesis, is modulated by Le" glycolipid
(89). The Le" glycolipid has been shown to strongly bind other Le" glycolipids and
glycoconjugates containing the Le" determinant. It weakly binds the Ley and nLc4
determinants. The ganglioside GM3 can bind asialo-GM3 (lactosyl ceramide) (90) as well as
Gg3Cer with strong affinity (91). While these interactions are not as strong as some protein-
protein bonds, the observed clustering of gangliosides on the cell surface, instead of being
homogeneously distributed on the surface, suggests that there is strength in numbers (90).
Hakomori has postulated that the interaction of these glycolipids is not the main type of
adhesion of cells, but precedes a stronger binding event. The patches of gangliosides or
other glycolipids on the surfaces allow cells to come in contact and remain oriented long
enough for the more stable binding takes place (90).

Interaction of GM, with lactosyl ceramide is stronger in the presence of integrins
(91), suggesting that gangliosides and other glycolipids have some interaction with various
extracellular proteins. Evidence of a role for gangliosides in cell adhesion to substrata was
found when the antibodies against GDz and GD; were able to inhibit the attachment and
spreading of human melanoma cells to a number of extracellular proteins including
fibronectin, vitronectin, collagen, laminin and Arg—Gly-Asp—containing synthetic peptides
(92). Gangliosides, GDla and Gle, were able to inhibit fibronectin-mediated adhesion to
fibronectin-eollagen complexes (93). Spiegal et a1. (93) reported that complex gangliosides
(the more highly glycosylated or sialylated) have an important role in fibronectin
organization. Furthermore, GDla has been shown to bind laminin, suggesting a role for
gangliosides in neural cell adhesion (95).

In addition to the cell-cell and cell-strata interactions, gangliosides, as well as other

21

glycolipids, can act as receptors for various hormones and bacterial and viral toxins (90, 96,
97). Gangliosides have been associated with the binding of the hormones such as thyroid-
stimulating hormone (98), chorionic gonadotropin (99), as well as serotonin (100). GMl

has been well characterized as the receptor for the B-subunit of Cholera toxin ( 12).

Signal Transduction and Cell Grthh

Despite the topological arrangement of gangliosides on the outer leaflet of cells, there
is a growing body of evidence suggesting that gangliosides act in transmembrane signal
transduction. Gangliosides appear to influence cell transmembrane signaling in a variety of
different mechanisms including 1) their interaction with plasma membrane ion channels, 2)
direct interaction with plasma membrane growth factor receptor-associated kinases, and 3)
through degradative products that inhibit protein kinase C (101). It is through these
mechanisms that gangliosides are postulated to be involved in cell-cycle control.

The addition of exogenous gangliosides has been shown to stimulate Na+-K+~ATPase
activity in rat brain mitochondrial membranes (102) and the binding of GM 1 to the B subunit
of cholera toxin causes proliferation in lymphocytes and fibroblasts by inducing a rapid and
sustained increase in intracellular calcium (103). Moreover, changes in ganglioside
composition and metabolism have been associated with differentiation (104, 105). Several
lines of evidence suggest that these compositional changes are more than just a phenotypic
trait and that there is a specific role for gangliosides in differentiation. Exogenously added
gangliosides can either stimulate or inhibit cell growth depending on the cell type and
ganglioside added. Specifically, the addition of GM1 or GT“, enhances the effect of nerve

growth factor (NGF) on PC-12 pheochromocytoma cells (106) and Gle appears to have

22

NGF-like activities by itself (58). Cells may exhibit a glycolipid contact response that is
closely related to contact inhibition of cell growth, suggesting that gangliosides may be
involved in the regulation of cell growth. Synthesis of specific gangliosides can often
accompany density-dependent growth inhibition. This has been seen with GM3 in NIL cells
(107), GDla in 3T3 cells (108), and GM3 and GM1 in human fibroblasts (109).
Gangliosides may modulate cell growth through various growth factor receptors.
Studies have found that exogenously added GM3 and GMl to 3T3 fibroblasts inhibited their
growth (110). The addition of these gangliosides had little or no effect on the growth
factors binding to their respective receptors. It was later demonstrated that GM3 interacts
directly with the EGF-r and GMl or GM3 with PDGF-r and inhibits the autophosphorylation
of the receptor (111, 112). Furthermore, various metabolites of GM3 have different
modulating effects on the receptor. De-N-acetyl-GM3 activates the tyrosine kinase of EGF-r
and is a strong promoter of cell growth (1 l l, 112). On the other hand, lyso-GM3 (dc-N-
acyl-GM3), is a potent inhibitor of EGF—r autophosphorylation and cell growth (113).
Parallels have been drawn between the growth modulating effects of GM3 metabolites and

that of the phosphatidylinositol cascade (90).

Oncogenic Transformation

Oncogenic transformation has been characterized by aberrant synthesis and
organization of gangliosides (3). Alterations in the compositions of gangliosides are a result
of incomplete synthesis due to down regulation or blockage of certain biosynthetic pathways.
Precursor or less complex gangliosides can accumulate as a direct result. Also, neo-

synthesis, the synthesis of novel gangliosides with a different core structure, and retrogenic

23
synthesis, the reappearance of fetal lipids through activation of quiescent fetal

glycosyltransferases (3, 75), of gangliosides is possible. A possible result is an increase in
the less complex gangliosides with the disappearance of the more complex. Similarly, the
emergence of novel gangliosides in the form of different core structures or biosynthetic
pathways may accompany oncogenic transformation. Table 2 is a list of gangliosides that
have been well characterized in particular cancers. It is not yet clear as to what the role of
aberrant ganglioside expression is in the biology of cancer, however, it is easy to see the
potential functions gangliosides may have when one considers their suspected and understood
functions in normal cells as stated above.

The aberrant expression of gangliosides on the surface of cancer cells has also been
suggested to have an immunosuppressive function, possibly allowing gangliosides to escape
immune surveillance Gangliosides of cancer tissues have been demonstrated to inhibit
natural killer cell activity (114), T-cell response (140) and block the effect of IL-2 (141).
The various postulated effects of gangliosides on the immune system have been reviewed in
142. Of particular interest when considering the immunosuppressive aspects of gangliosides
is the phenomenon of their shedding from the surfaces of tumor cells into the local
environment and eventually the blood stream. Neuroblastoma and melanoma have both been
shown to shed gangliosides from their surface at a high rate (5, 7, 129). In particular,
neuroblastoma Ins elevated levels of GDZ in the tumor tissue as well as elevated level in
circulation (6, 8). The levels of GD2 in the plasma or serum of neuroblastoma patients was

closely correlated with the progression of the disease and has been suggested as a circulating

tumor marker (6, 8, 130). Furthermore, shed gangliosides of neuroblastoma were

24

 

.2 .3 «68202
.0: a: 2.26
ms
NM~ EOEB Edam .550 o<=026wuN60<=oZBN
as as .2 2
232.26 59.2852 So 80309282.: 60-3204-Bacaaézao
NN MEOumEDOuozz
we 25:0 282.8
E 38202 2
as 68:8 we: ~20 50.3-2:qu so-_no_oaaaorao<zao
.5003 hon:
o: 382...: 50-3.5-0 -o<=oz-o<o-a-o<azn= -30:28m-~8<8§A3<¢A8<32
N2 .02 .552 5.5
m2 .h «893383:
vfl .mfi «E9383 0388ch
NE can 38.28
E .8}
a _ .2 a .83 magmas moo auofiaofiozvn: 58-39.-_n_~om-~eo<=ozm-~8<=oz
27m: .5 .85., :20 5005-282”: seizuriaomasgoz
5:355“?
moococoaom €083. E_o€occo>m £036.? mDTU<mD~ Baonbm 3262.0

 

N 25G.

25

hZ

.8ch 058.05.;

25.3 3502.3

 

«Eocao.moo=oc< 80.62: m o m e

t _ 38.28 3.5 -8...E282.>_ 802854- _ 2828222323
E _ .8ch 0:885; ”Sme— o<=oZeN
«80568852.... .0032: 3 m 2

2 _ 88.28 .25 -83528282 80286..-_qaom-.n2zo_oaaao
u<=oniNdo<=oZdN

m 2

a: 2.26 582-2% 89.52.5853. 80-E2032.328.228.3230
o<=oZBN

we 2.26 m.
E ..Eosoau 2.3 .282-..“ 8o-.3-282n>_ 80-320..-85222220328
«Eocmouwoocovm o<=oZGN

2: 2.888.. f
3 «228.882. 228 £828 8o-.3=-28z.>_ 80-820..-5.8822220..-_23
805.80% mcoEE. 8652.55 238.? 52702253— ouaosbm Roms—2.5

...8. .232

26

demonstrated to be immunosuppressive in a lymphoproliferative assay in vitro (131).

' Ii i nd i li

Sialic acids are acetylated derivatives of neuraminic acid (2-keto—5-amino—3,5-
dideoxy-D—nonulosonic acid) and are widely found in nature as mucins, glycoproteins,
gangliosides, and milk Oligosaccharides (10, 143-146). Gangliosides are defined as
sialic acid-containing glycosphingolipids and the N-acetylneuraminic acid is thought to
be the only homolog found in humans (147). The presence of sialic acid on
glycoconjugates has been demonstrated to have a wide range of biological functions.
These include structural conformation (10, 146, 148, 149), recognition and masking
of binding sites (10, 146), and to confer mitogenic properties as well as the binding of
various substrates (119, 148). Concentrations of sialic acid, defined as total serum
sialic acid (all sialic acid containing glycoconjugates and free sialic acid) or lipid-
associated sialic acid (gangliosides), have been suggested to be useful in the diagnosis,
prognosis, and the monitoring of various disease states including several cancer types
(150-163).

Sialidases (N-acetylneuraminosyl glycohydrolases (EC 3.2.1.18» are the
enzymes that cleave sialic acid residues from their Oligosaccharide bases. Sialidases
have been characterized in humans from the plasma membrane, cytosol, lysosome, as
well as in the medium of cultured human fibroblasts (164, 165). There have been
reports of trace sialidase activity in plasma (166), yets its presence and activity are
inconclusive in either healthy individuals or cancer patients. Activity of sialidase

towards gangliosides was found in transformed hamster cells that was not found in the

27

normal cells (167). It was postulated that this sialidase activity directed towards
extracellular gangliosides may be involved in disrupting cell-cell binding events

necessary for transformation.

B n er u an [1 sides

Breast cancer is one of the most common and most dreaded cancers in women.
It has an unpredictable course and the risk of metastases continues for 20 years or
more. It accounts for 18% of cancer deaths in females. It is typically detected by
self examination or mammography and staging of the cancer is based on tumor size,
presence in lymph nodes and the degree of metastasis. The earlier the tumor can be
discovered and determining the degree to which it has spread is important in the
prognosis (168). The need for better methods of detection as well as basic
understanding of the disease continues to be an ongoing struggle.

Several studies have examined the utility of sialic acid in the circulation for the
diagnosis and staging of breast cancer (152, 154-156, 169-171). The results of these
studies typically indicate that sialic acid can be a useful marker for breast cancer yet
there are problems associated with these studies and therefore the approach is not yet
fully accepted (158, 172). The little work done that examines the ganglioside content
in the breast tumor tissue shows abberrant glycosylation and an alteration in the
ganglioside profile (173). There is evidence for a correlation between sialic acid
concentration and progression of disease; however, a causal relationship has not been
established. If these alterations in the ganglioside profiles can possibly be understood

and made useful to the physcian, it may be a great boon to the many women fated to

suffer the disease.

28

29
Statement of the Problem

Breast cancer is a tragic, and often times fatal disease. There can never be too
much information available to the physician when attempting to diagnose and treat
patients with this disease. Circulating gangliosides have been demonstrated to be
useful in the diagnosis and prognosis of other cancers (130, 174). This study seeks to
determine the utility of circulating gangliosides as tumor markers for the diagnosis
and monitoring of breast cancer. Additionally, by defining the circulating ganglioside
profile associated with breast cancer, some insight may be gained in the workings of
the disease. In order to thoroughly assess this potential, the following goals must be

met:

1. Identify or develop protocols for the isolation and purification of gangliosides
from small amounts of serum.

2. Identify or develop protocols for the quantification of gangliosides at the picomole
level.

3. Identify or develop protocols for the separation of complex ganglioside mixtures
using one- or two-dimensional high-performance thin-layer chromatography.

4. Create a map of relevant gangliosides by one- or two-dimensional high-
performance thin-layer chromatography to aid in identification of unknown
gangliosides.

5. , Analyze a suitable amount of samples of plasma or serum from breast cancer
patients and appropriate control group women.

6. Analyze the free sialic acids levels and activity of sialidase in the blood of breast

30

cancer patients and appropriate control group women.
7. Compare the two test groups’ ganglioside profiles as well as the free sialic acid

concentration and sialidase activiy and discuss any anomalies.

10.

11.
12.

13.

14.

15.

W

Marshall, B. (1993) Science 259, 618-621
Hakomori, S. (1981) Ann. Rev. Biochem. 50, 733-764
Hakomori, S., and Kannagi, R. (1983) J. Natl. Cancer Inst. 71, 231-251

Yogeeswaran, G. (1983) in Advances in Cancer Research, Vol. 38, pp. 289-
350, Acedemic Press, New York

Valentino, L. A., and Ladisch, S. (1992) Cancer Res. 52, 810-814

Valentino, L., Moss, T., Olson, E., Wang, He-J., Elashoff, R., and Ladisch,
S. (1990) Blood 75(1), 1564-1567

Bernhard, H., Meyer zum Buschenfelde, Karl-H., and Dippold, W. G. (1989)
Int. J. Cancer 44, 155-160

Ladisch, S., Wu, Zi-L., Feig, S., Ulsh, L., Schwartz, E., Floutsis, G.,
Wiley, F., Lenarsky, C., and Seeger, R. (1987) Int. J. Cancer 39, 73—76

Stults, C. L. M., Sweeley, C. C., and Macher, B. A. (1989) Methods
Enzymol. 179, 167-207

Reutter, W., Kottgen, E., Bauer, C., and Gerok, W. (1982) in Sialic Acids:
Chemistry, Metabolism and Function, Vol. 10 (Schauer, R., ed), pp. 263-306,
Springer-Verlag, New York

Hakomori, S. (1986) Chem. Phys. Lipids 42, 209-233

Cuatrecasas, P. (1973) Biochem. 12, 3547-3558

Brady, R. 0., and Barranger, J. A. (1983) in The Metabolic Basis of Inherited
Disease, 5th ed. (Stanbury, J. B., Wyngaarden, J. B., Fredrickson, D. S.,
Goldstein, J. L., and Brown, M. S., eds), pp. 842-856, McGrall-Hill, New
York

Tettamanti, G., Bonali, F., Marchesini, S., and Zambotti, V. (1973) Biochim.
Biophys. Acta 296, 160

Folch, J., Lees, M., and Sloan Stanley, G. H. (1956) J. Biol. Chem. 226, 497

31

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

32
Suzuki, K. (1965) J. Biochem. 12, 629-638
Ladisch, S., and Gillard, B. (1985) Anal. Biochem. 146, 220-231
Williams, M. A., and McCluer, R. H. (1980) J. Neurochem. 35, 266-269
Kundu, S. K. (1981) Methods Enzymol. 72, 174-183
Ledeen, R. W., Yu, R. K., and Eng, L. F. (1973) J. Neurochem. 21, 829-839
Ando, S., Waki, H., Kon, K., and Kishimoto, Y. (1987) in Gangliosides and
Modulation of Neuronal Functions. NA T0 ASI series, Vol. H7 (Rahmann,
H., ed), pp. 167-171, Springer-Verlag, Berlin

Ladisch, S., Sweeley, C. C., Becker, H., and Gage, D. (1989) J. Biol. Chem.
264, 12097-12105

Wantanabe, N. (1985) J. Chromatog. 330, 333-338
Svennerholm, L. (1957) Biochim. Biophys. Acta 24, 604
Svennerholm, L. (1956) J. Neurochem. 1, 42

Higashi, H., Fukui, Y., Ueda, S., Kato, S., Hirabyashi, Y., Matsumoto, M.,
and Naiki, M. (1984) J. Biochem. 95, 1517-1520

Magnani, J. L., Smith, D. F., and Ginsburg, V. (1980) Anal. Biochem. 190,
399-402

Magnani, J. L., Brockhaus, M., Smith, D. L., Ginsburg, V., Blaszczyk, M.,
Mitchell, K. F., Steplewski, Z., and Koprowski, H. (1981) Science 212, 55-56

Smith, D. F. (1983) Biochem. Biophys. Res. Comm. 115, 360-367

Buehler, J., and Macher, B. A. (1986) Anal. Biochem. 158, 283-287

Ando, S., and Yamakawa, T. (1973) J. Biochem. 73, 387

Ando, S., Chang, Nan-C., and Yu, R. K. (1978) Anal. Biochem. 89, 437-450

Trbojevic-Cepe, M., and Kracun, I. (1990) J. Clin. Chem. Clin. Biochem. 28,
863-872

Wiesner, D. A., and Sweeley, C. C. (1993) In Preparation

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

33
Suzuki, K. (1964) Life Sci. 3, 1227

Chigomo, V., Sonnino, S., Ghidoni, R., and Tettamamti, G. (1982)
Neurochem. I. 4, 397-403

Ledeen, R. W., Haley, J. E., and Skrivanek, J. A. (1980) Anal. Biochem.
112, 135-142

Nakamura, K., Ariga, T., Yahagi, T., Miyatake, T., Suzuki, A., and
Yamakawa, T. (1983) J. Biochem. 94, 1359-1365

Weigandt, H., and Biicking, H. W. (1970) Eur. J. Biochem. 15, 287
Hakomori, S. (1966) J. Lipid Res. 21, 789

Nilsson, B., and Svensson, S. (1978) Carbohydrate Res. 62, 377
Ito, M., and Yamagata, T. (1986) J. Biol. Chem. 261, 14278-14282

Li, Yu-T., Ishikawa, Y., and Li, Su-C. (1987) Biochem. Biophys. Res. Comm.
149, 167-172

Li, Su-C., DeGasperi, R., Muldrey, J. E., and Li, Y. (1986) Biochem.
Biophys. Res. Comm. 141, 346-352

Honda, S. (1984) Anal. Biochem. 140, 1-47

Hase, S., Ikenaka, T., and Matsushima, Y. (1978) Biochem. Biophys. Res.
Comm. 85, 257-263

Takemoto, H., Hase, S., and Ikenaka, T. (1985) Anal. Biochem. 145, 863-874

Townsend, R. R., Hardy, M. R., Cumming, D. A., Carver, J. P., and
Bendiak, B. (1989) Anal. Biochem. 182, 1-8

Townsend, R. R., Hardy, M. R., Hindsgaul, 0., and Lee, Y. C. (1988) Anal.
Biochem. 174, 459-470

Hardy, M. R., and Townsend, R. R. (1988) Proc. Natl. Acad. Sci. USA 85,
3289-3293

Hardy, M. R., Townsend, R. R., and Lee, Y. C. (1988) Anal. Biochem. 170,
54-62

52.

53.

54.
55.

56.

57.

58.

59.

61.
62.

63.

65.
66.
67.
68.

69.

70.

34
Lee, Y. C. (1990) Anal. Biochem. 189, 151-162

Wang, Wei-T., Matsura, F., and Sweeley, C. C. (1983) Anal. Biochem. 134,
398-405

Yu, R. K., and Ledeen, R. W. (1970) J. Lipid Res. 11, 506-516
Niedermeier, W., and Tomana, M. (1974) Anal. Biochem. 57, 363-368

Honda, S., Suzuki, S., Kakehi, K., Honda, A., and Takai, T. (1981) J.
Chromatog. 226, 341-345

Sjostrom, E., Haglund, P., and Janson, J. (1966) Svensk Poppertidning 69, 11

Sawardeker, J. S., Sloneker, J. H., and Jeanes, A. (1965) Anal. Chem. 37,
1602

Sweeley, C. C., and Walker, B. (1964) Anal. Chem. 36, 1461-1466
Rickert, S. J., and Sweeley, C. C. (1978) J. Chromatog. 147, 317-326
Vance, D. E., and Sweeley, C. C. (1967) J. Lipid Res. 8, 621-630
Reinhold, V. N. (1987) Methods Enzymol. 138, 244-249

Sweeley, C. C., and Vance, D. E. (1967) in Lipid Chromatographic Analysis.
Vol. 1 (Marinetti, G., ed), pp. 465-495, Marcel Dekker, New york.

Shimamura, M., Oku, M., and Yamagata, T. (1991) FEBS Lett. 290, 213-215
Chambers, R. E., and Clamp, J. R. (1971) Biochem. J. 125, 1009-1018
Chaplin, M. F. (1982) Anal. Biochem. 123, 336-341

Wiesner, D. A., and Sweeley, C. ( 1993) Anal. Biochem. In Press

Polito, A. J., Naworal, J., and Sweeley, C. C. (1969) Biochem. 8, 1811-1815

Elwood, P. C., Reid, W. K., Marcel], P. D., Allen, R. H., and Kolhouse, J.
F. (1988) Anal. Biochem. 175, 202-211

Sweeley, C., Bentley, R., Makita, M., and Wells, W. W. (1963) J. Am.
Chem. Soc. 85, 2497-2507

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

35

DeJongh, D. C., Radford, T., Hribar, J. D., Hanessian, S., Bieber, M.,
Dawson, G., and Sweeley, C. C. (1969) J. Am. Chem. Soc. 91, 1728-1740

Wong, C. G., Sung, Sun-S. J., and Sweeley, C. (1980) in Methods in
Carbohydrate Chemistry, Vol. viii, pp. 55-64, Acedemic Press,

Yang, H.-J., and Hakomori, S. (1971) J. Biol. Chem. 246, 1192-1200

Morgan, A. C., Crane, M. M., and Rossen, R. D. (1984) J. Natl. Cancer
Inst. 72, 243-249

Hakomori, S. (1985) Cancer Res. 45, 2405-2414
Keranen, A. (1976) Biochim. Biophys. Acta 431, 96—104

Fredman, P., Mansson, J. E., Svennerholm, L., Karlsson, K. A., Pascher, 1.,
and Samuelsson, B. E. (1980) FEBS Lett. 110, 80-84

Tanaka, Y., Yu, R., Ando, S., Ariga, T., and Itoh, T. ( 1984) Carbohydrate
Res. 126, 1-14

Hakomori, S. (1964) J. Biochem. 55, 205-208

Geyer, R., Geyer, H., Kuhnhardt, S., Mink, W., and Stirm, S. (1982) Anal.
Biochem. 121, 263-274

Sweeley, C. C., and Dawson, G. (1969) Biochem. Biophys. Res. Comm. 37,
6-14

Sammuelsson, K., and Sammuelsson, B. ( 1969) Biochem. Biophys. Res.
Comm. 37, 15-21

Sweeley, C. C., and Dawson, G. (1971) J. Lipid Res. 12, 56-64

Karlsson, K. A., Leffler, H., and Samuelsson, B. E. (1974) J. Biol. Chem.
249, 4819-4823

Urdal, D. L., and Hakomori, S. (1983) J. Biol. Chem. 255, 6869-6874

Egge, H., Peter-Katalinic, J ., Reuter, G., Schauer, R., Ghidoni, R., Sonnino,
S., and Tettamanti, G. (1985) Chem. Phys. Lipids 37, 127-141

Domon, B., and Costello, C. E. (1988) Biochem. 27, 1534-1543

Costello, C. E., and Vath, J. E. (1990) Methods Enzymol. 193, 738-768

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

36

Eggens, I., Fenderson, B. A., Toyokuni, T., Dean, B., Stroud, M. R., and
Hakomori, S. (1989) J. Biol. Chem. 264, 9476-9484

Hakomori, S. (1992) Tohoku J. Exp. Med. 168, 211-222

Kojima, N., and Hakomori, S. (1991) J. Biol. Chem. 266, 17552-17558
Cheresh, D. A., Pierschbacher, M. D., Herzig, M. A., and Mugoo, K. (1986)
J. Cell Biol. 102, 688-696

Kleinman, H. K., Martin, G. R., and Fishman, P. H. (1979) Proc. Natl.
Acad. Sci. USA 76, 3367-3371

Spiegal, S. Yamada, K., Hom, B. E., Moss, J., and Fishman, P. H. (1986)
Cell Biol. , 102, 1898-1906.

Laitinen, J ., Lopponen, R., Merenmies, J ., and Rauvala, H. (1987) FEBS
Lett. 217, 94-100

Fishman, P. H. (1982) J. Membr. Biol. 69, 85-97

Bock, K., Briemer, M. E., Bignole, A., Hansson, G. C., Karlsson, K.-A.,
Larson, G., Leffler, H., Samuelsson, B. E., Stromberg, N ., Svanborg-Eden,
C., and Thurin, J. (1985) J. Biol. Chem. 260, 8545

Mullin, B. R., Fishman, P. H., Lee, G., Aloj, S. M., Ledley, F. D., Winand,
R. J., Kohn, L. D., and Brady, R. O. (1976) Proc. Natl. Acad. Sci. USA 73,
842

Fishman, P. H., and Brady, R. O. (1976) Science 194, 906-915

Woolley, D. W., and Gommi, B. W. (1965) Proc. Natl. Acad. Sci. USA 53,
959

Zeller, C. B., and Marchase, R. B. (1992) Am. J. Physiol. 262, c1341-c1355

Leon, S., Facci, L., Toffano, F., Sonnino, S., and Tettamanti, G. (1981) J.
Neurochem. 37, 350—357

Spiegel, S., and Panagiotopoulos, C. (1988) Exp. Cell Res. 177, 414-427

Glickman, R. M., and Bouhours, J. F. ( 1976) Biochim. Biophys. Acta 424, 17

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

37

Igarachi, Y., Nojiri, H., Hanai, N., and Hakomori, S. (1989) Methods
Enzymol. 179, 521-539

Ferrari, G., Fabris, M., and Gorio, A. (1983) Dev. Brain Res. 8, 215-221

Sakiyama, H., Gross, S. K., and Robbins, P. W. (1972) Proc. Natl. Acad.
Sci. USA 69, 872-876

Yogeeswaran, G., and Hakomori, S. (1975) Biochem. 14, 2151-2156
Hakomori, S. (1970) Proc. Natl. Acad. Sci. USA 67, 1741-1743

Bremer, E. G., Hakomori, S., Bowen-Pope, D. F., Raines, E., and Ross, R.
(1984) J. Biol. Chem. 259, 6818-6825

Hanai, N ., Nores, G., MacLeod, C., Torres-Mendez, Carmen-R., and
Hakomori, S. (1988) J. Biol. Chem. 263, 10915-10921

Hanai, N., Dohi, T., Nores, G. A., and Hakomori, S. (1988) J. Biol. Chem.
263, 6296-6301

Hanai, N., Nores, G., Torres-Mendez, C.-R., and Hakomori, S. ( 1987)
Biochem. Biophys. Res. Comm. 147, 127-134

Grayson, G., and Ladisch, S. (1992) Cell. Immun. 139(1), 18-29

Tsuchida, T., Saxton, R., Morton, D., and Irie, R. (1989) Cancer 63, 1166-
1174

Hamelton, W. B., Helling, F., LLoyd, K., and Livingston, K. O. (1993) Int.
J. Cancer 53, 566-573

Magnani, J. L., Nilsson, B., Brockhaus, M., Zopf, d, Steplewski, Z.,
Kaproski, H., and Ginsburg, V. (1982) J. Biol. Chem. 257, 14365-14369

Kyogshima, M., Uemuta, K., and Taketomi, T. ( 1987) Japan J. Cancer 78,
1229-1237

Yogeeswaran, G., and Salk, P. L. (1981) Science 212, 1514-1516

Portoukalian, J ., Zwingelstein, G., and Dore, J. F. (1979) Eur. J. Biochem.
94, 19-23

Hersey, P., and Jamal, P. (1989) Pathology 21, 51-58

122.
123.

124.

125.

126.

127.

128.

129.

130.
131.
132.

133.

134.

135.

38

Fredman, P., Nilsson, 0., Svennerholm, L., Myrvold, H., Persson, B.,
Pettersson, S., Holmgren, J ., and Lindholm, L. (1983) Med. Biol. 61, 45-48

Yates, A. J., Thompson, D. K., Boesel, C. P., Albrightson, c, and Hart, R.
W. ( 1979) J. Lipid Res. 20, 428-436

Traylor, T. D., and Hogan, E. L. (1980) J. Neurochem. 34, 126-131

Kannagi, R., Nudelman, E., and Hakomori, S. (1982) Proc. Natl. Acad. Sci.
USA 79, 3470-3474

Nilsson, 0., Masson, J. E., Brezicka, T., Holmgren, J ., Lindholm, L.,
S6renson, S., Yngvason, G., and Svennerholm, L. (1984) Glycocon. J. 1, 43-
49

Clamp, J. R., Dawson, G., and Hough, L. (1967) Biochim. Biophys. Acta
148, 342

Fredman, P. (1988) in New Trends in Ganglioside Research: neurochemical
and Neuroregenerative Aspects, Vol. 14 (Ledeen, R. W., Hogan, E. L.,
Tettamanti, G., Yates, A. J., and Yu, R. K., eds), pp. 151-162, Liviana,
Press, Padova

Li, R., and Ladisch, S. (1991) Biochim. Biophys. Acta 1083, 57-64

Schulz, G., Cheresh, D. A., Varki, M. M., Yu, A., Staffileno, L. K., and
Reisfeld, R. A. (1984) Cancer Res. 44, 5914-5920

Shaposhnikova, G. 1., Prokazova, G. A., Buzinkov, G. A., Zvezdiva, N. D.,
Teplitz, N. A., and Bergelson, L. D. (1984) Eur. J. Biochem. 140, 567-570

Schulz, G., Cheresh, D. A., Varki, N. M., Yu, A., Staffileno, L. K., and
Reisfeld, R. A. (1984) Cancer Res. 44, 5914-5920

Wantanabe, T., PUkel, C. S., Takeyama, H., Lloyd, K. 0., Shiku, H., Li, L.
T. C., Travassos, L. R., Oettgen, H. F., and Old, L. .t. (1982) J. Exp. Med.
156, 1884-1889

Tai, T., Cahan, L. D., Tsuchida, T., Saxton, R. E., Irie, R. F., and Morton,
D. L. (1985) Int. J. Cancer 35, 607-612

Hakomori, S., Nudelman, E., Levery, S. B., and Patterson, C. M. (1983)
Biochem. Biophys. Res. Comm. 113, 791-798

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

150.

151.

39

Masson, J .-E., Fredman, P., Nilsson, 0., Lindholm, L., Holmgren, J ., and
Svennerholm, L. (1985) Biochim. Biophys. Acta 834, 110-117

Nilsson, 0., M55son, J .-E., Lindholm, L., Holmgren, J ., and Svennerholm,
L. (1985) FEBS Lett. 182, 398-402

Davidsson, P., Fredman, P., Collins, V. P., von Holst, H., Mansson, J .-E.,
and Svennerholm, L. ( 1989) J. Neurochem. 53, 705-709

Masson, J.-E., Fredman, P., Bigner, D. D., Molin, K., Rosengren, B.,
Fredman, H. S., and Svennerholm, L. (1986) FEBS Lett. 201, 109-113

Ladisch, 8., Becker, H., and Ulsh, L. (1992) J. Immunol. 8, 180-188

Marcus, D. M. (1988) in New Trends in Ganglioside Research: Neurochemical
and Neuroregenerative Aspects, Vol. 14 (Ledeen, R. W., Hogan, E. L.,
Tettamanti, G., Yates, A. J., and Yu, R. K., eds), pp. 173-180, Liviana

Press, Padova

Dyatlovitskaya, E. V., and Bergelson, L. D. (1987) Biochim. Biophys. Acta
907, 125-143

Kakehi, K., Suzuki, S., Honda, S., and Lee, Y. C. (1991) Anal. Biochem.
199, 256-268

Aminoff, D. (1961) Biochem. J. 81, 384-392

Stack, M. S., Pizzo, S. V., and Gonzalez-Gronow, M. ( 1992) Biochem. J.
284, 81-86

Jourdian, G. W., Dean, L., and Roseman, S. (1971) J. Biol. Chem. 246, 430-
435

Corfield, A. P., and Schauer, R. (1982) in Sialic Acids: Chemistry,
Metabolism, and Function, Vol. 10 (Schauer, R., ed), pp. 5-50, Springer-
Verlag, New York

Schauer, R. (1985) Trends Biochem. Sci. 117, 357-360

Schauer, R. (1982) Adv. Carbohydr. Chem. Biochem. 40, 131-234

Klapan, I., Katic’, V., ’Culo, F., Sabolovic’, D., C’uk, V., Fumic’, K., and
Simovic’, S. (1993) Eur. J. Can. 29A, 839-845

Crook, M. (1993) Clin. Biochem. 26, 31-38

152.

153.

154.

155.

156.

157.

158.

159.

160.

161.

162.

163.

164.

165.

166.

167.

40

Dnistn'an, F. M., Schwartz, M. K., Katopodis, N., Franchia, A. A., and
Stock, C. C. (1982) Cancer 50, 1815-1819

Fu, W. Y., Neale, E., and Shun, F. W. W. (1993) Int. J. Gyn. C. 3(3), 143-
146

Patel, P., Baxi, B. R., Adhvaryu, S. G., and Balar, D. B. (1990) Anticancer
Res. 10, 1071-1074

Dwivedi, C., Dixit, M., and Hardy, R. E. (1990) Experientia 46, 91-94

Riley, M., Tautu, C., Verazin, G., Gregory, J., Josiah, S., Prorol, J., and
Alhadeff, J. (1990) Clin. Chem. 36, 161-162

Stringou, E., Chondros, K., Kouvaris, J ., Kakari, S., and Papavassiliou, K.
(1992) Anticancer Res. 12, 251-256

Voigtmann, R., Pokomy, J ., and Meinshausen, A. (1989) Cancer 64, 2279-
2283

Haq, M., Haq, S., Tutt, P., and Crook, M. (1993) Ann. Clin. Bi. 30, 383-386

Vivas, I., Spagnuolo, L., and Palacios, P. (1992) Gynecol. Oncol. 46, 157-
162

Polivkova, J ., Vosmikova, K., and Horak, L. ( 1992) Neoplasma 39, 223-226

Lindberg, G., Rastam, L., Gullberg, B., and Eklund, G. A. (1992) Int. J.
Epidemiol. 21, 253-257

Kokiglu, E., Sémnez, H., Uslu, E., and Uslu, l. (1992) Cancer Biochem.
Biophys. 13, 57-64

Miyagi, T., Hata, K., Konno, K., and Tsuiki, S. (1992) Tohoku J. Exp. Med.
168, 223-229

Sweeley, C. C. (1993) in Sphingolipids and their Metabolites, lst ed., Vol. 26
(Bell, R. M., Hannum, Y. A., and Merrill, A. J., eds), pp. 235-252,
Academic Press, New york

Schauer, R., Veh, R. W., Wember, M., and Bushcer, Hans-P. ( 1976) Hoppe-
Seyler’s Z. Physiol. Chem. 357, 559-566

Schengrund, Cara-L., Lausch, R. N., and Rosenburg, A. (1973) J. Biol.
Chem. 248, 4424-4428

168.

169.

170.

171.

172.

173.

174.

41

Keys, H. M., Bakemeier, R. F., and Savlov, E. D. (1983) in Clinical
Oncology for Medical Students and Physicians, 6th ed. (Rubin, P., ed), pp.
120-141, American Cancer Society, New York

Chondros, K., Dapolla, V., Stringou, E., Petiniki, M., Vakouftsi, M., and
Kakari, S. (1991) Anticancer Res. 11, 2103-2106

Shamberger, R. J. (1984) J..Clin. Chem. Clin. Biochem. 22, 647-651

Sallay, S. I., Prosise, W. W., Morré, D. J., and Matyas, G. R. (1986) Can.
J. 1, 124-129

Vegh, Z., Kremmer, T., Boldizsar, M., Gesztesi, K. A., and Szajani, B.
(1991) Clin. Chim. Act. 203, 259-268

Dyatlovitskaya, E. V., Tekieva, E. A., Lemenovskaya, A. F., Somova, O.
G., and Bergelson, L. D. (1991) Biokhimiya 56(3), 560-564

Saha, S., and Chattopadhyay, U. (1988) Int. J. Cancer 41, 432-435

CHAPTER 2

Isolation of Gangliosides from Serum or Plasma

D.A. Wiesner and C.C. Sweeley

November 16, 1993

Department of Biochemistry
Michigan State University
East Lansing, MI

42

43

Abram!

In order to determine the most appropriate method for the isolation of
gangliosides from small samples of plasma or sera (1 ml), several different techniques
were considered. Solvent partitioning, column chromatography, and solid-phase
extraction were carefully examined and each step evaluated for its potential usefulness
in an overall procedure. We determined that a modified partitioning method using
diisopropyl ether, butanol and water was the most useful when combined with gel
filtration. Gangliosides could be isolated from small volumes of plasma or sera in
good yield and were suitable for analysis by high-performance thin-layer
chromatography (HPTLC). It was, however, necessary to further purify gangliosides
mixtures from plasma or sera with a DEAE-Sephadex column and gel filtration when

analysis by gas chromatography was desired.

44

Introgugti ii

To determine the usefulness of serum gangliosides as relevant circulating
tumors markers, a suitable method for their isolation was required. As a primary
goal in this study, the method had to be suitable for the isolation of gangliosides from
relatively small amounts of blood serum. Currently available methods are useful for
many types of applications, yet fall short of the requirements outlined by this study.
Most methods yield a total-lipid extract from the biological tissue or fluid by thorough
mixing and/or homogenization in a mixture of chloroform and methanol. Lipids are
solubilized and precipitated proteins removed by centrifugation. The classic Folch (l)
partition and its many modifications partition the total-lipid extract with mixtures of
chloroform, methanol, and saline, with the gangliosides partitioning into the aqueous
phase and phospholipids and other undesired components partitioning into the organic
phase. After lyophilization, salts and other low molecular weight contaminants are
removed by dialysis or size-exclusion chromatography. This method provides a crude
mixture of gangliosides suitable for many uses or ready for further clean up with
additional column chromatographic procedures. Problems with this method are the
need for large sample sizes and losses of less polar gangliosides (GM3) through
partitioning into the organic phase. Other methods for the isolation of a ganglioside
fraction from the total lipid extract involve various types of weak-anion exchange,
affinity, and size exclusion column chromatography. The most popular method
involves the use of a DEAE-Sephadex A-25 to remove neutral and positively charged

contaminants from the ganglioside fraction. Mild alkaline methanolysis is then used

45

to destroy any remaining phospholipids, with dialysis, size-exclusion or reverse-phase
chromatography used to remove salts and low molecular weight contaminants.
Chromatography on silicic acid removes residual negatively charged contaminants to
produce a relatively pure mixture of gangliosides. While this method of isolation
produces mixtures of gangliosides pure enough for most applications, scaling down of
the method to accommodate the small samples required by this study would be
required. Ladisch and Gilliard (2, 3) found that gangliosides could be isolated from
the total lipid extract with good recovery by partitioning between diisopropyl ether, n-
butanol and distilled water (6:4:5). The gangliosides were recovered in the aqueous
phase and phospholipids in the organic phase. After desalting on a small column of
Sephadex G-50, the sample was ready for analysis. This method is simple and
suitable for smaller samples; however, losses of the less polar ganglioside into the
organic phase become critical as the sample size decreases. Furthermore, while the
degree of purity was acceptable, often contaminants would make one-dimensional
thin-layer chromatography (l-D HPTLC) and quantitation by densitometry difficult if
not impossible. All but the Folch method were examined thoroughly for use in this
study. The reason the Folch method was not considered was the admitted loss of
gangliosides which would most likely be exacerbated when scaled down. Criteria for
choosing a particular method were simplicity, ability to handle small (i.e. 1 ml serum)
and numerous samples, and potential for clinical usefulness. It was determined that
none of the current methodologies, as is, would be suitable without some degree of
modification. The following summaries are of studies conducted to evaluate various

methods and modifications in an attempt to optimize each potentially useful step in an

46

overall procedure.

Material and Methods

Materials

Fresh outdated plasma was obtained from the American Red Cross (Lansing,
MI) and was stored at -70 °C until used. All solvents were of analytical grade or
better. DEAE-Sephadex A-25 and Sephadex G-50 were obtained from Pharmacia
(Uppsala, Sweden) and SepPak C13 cartridges were obtained from Waters (Milford,
MA). The Bond Elut aminopropyl and silica gel cartridges were from Analytichem
(Harbor City, CA). High-performance thin-layer chromatography plates (HPTLC)
were obtained from Merck (Darmstedt, Germany). Reference gangliosides were

obtained from Sigma (St. Louis, MO) unless otherwise noted.

Isolation of Total-Lipid Extract

The isolation of gangliosides along with other lipid material from sera or other
biological fluid was easily achieved by extraction with 10 vol of chloroformzmethanol
in varying proportions. We found that two extractions with chloroformzmethanol
(CzM, 1:1), first for 4 hours and then for 4 to 20 hours, gave quantitative recovery of
gangliosides in the total-lipid mix (4). Proteins and other insoluble materials were
removed by centrifugation after extraction. The additional removal of partially
soluble proteins and salts was achieved by reducing the volume to one-quarter and the
temperature to -70 °C followed by centrifugation. Modifications included

lyophilization of serum prior to total-lipid extraction with the intent to speed up future

47

drying steps. Also, dialyzing the plasma against water was examined to remove small
molecular weight contaminants at an early stage. Lastly, rotary evaporation was
examined versus evaporation under a stream of nitrogen with heating for the initial

removal of solvent.

Isolation of Gangliosides by Column Chromatography.

The use of DEAE-Sephadex chromatography for the purification of
gangliosides was originally designed for 20 ml of plasma (5). A subsequent report
suggested modifications on how to scale down by a factor of three (4). We extended
this scaling down to accommodate 1 ml of serum with the use of pasteur pipets or
champagne columns (60 mm X 5 mm ID. columns with 10 m1 reservoir). The
DEAE-Sephadex resin was washed 3 times with chloroform:methanol:0.8 M sodium
acetate (aq.), 30:60:8 (Solvent A), then allowed to soak over night in the same
solvent. The resin was then washed 3 times with chloroform:methanolzwater, 30:60:8
(Solvent B). Columns of approximately 0.3 and 1 m1 of resin were poured into
pasteur pipets and 2 ml of resin into a champagne column. These columns were fitted
with glass wool plugs forced in rather tightly in order to restrict the flow to a fairly
slow flow rate. The resin was washed with 15 ml more of solvent A. Samples of the
total lipid extract were applied in very small volumes, i.e. < 1 ml, and were
reapplied twice (the application volume is collected and reapplied to the top of the
column) to ensure total adsorption of the gangliosides to the column. The columns
were washed with 5 to 15 ml more of solvent A to remove neutral and positively

charged materials, then the ganglioside fractions were eluted with 5 to 15 ml of

48

solvent B. After reducing the eluates to dryness under a stream of nitrogen, the
samples were subjected to mild-alkaline methanolysis, by taking up in small volumes
of 0.1 N sodium hydroxide in methanol and incubated at 37 °C for 2-3 hours. This
was to destroy any acidic phospholipids still present. The reaction mixtures were
neutralized with acetic acid and extracted twice with equal volumes of hexane to
remove any residual neutral lipids and fatty acid methyl esters (FAME). The solvent
was again removed with a stream of nitrogen and oil-pump vacuum to ensure removal
of all traces of solvent.

Several techniques were employed to remove salts and low molecular weight
contaminants, including dialysis, size-exclusion chromatography on small Sephadex
G-50 columns. and solid-phase extraction cartridges. Dialysis was done with
prewashed dialysis tubing (0.32 ml/cm, 12-14, kDa cutoff) cut to 2-3 ml length and
was carried out for 2 days against water with numerous water changes. The dialysate
was then lyophilized and the residue taken up in a small volume of chloroform:
methanol (1:1) and further clarified by centrifugation if necessary. Alternatively,
salts could be removed by size exclusion chromatography with a small 10 ml column
of Sephadex G-50. The samples of ganglioside ( < 30 nmol) were applied to the top
of the column in a small volume (0.3 m1), allowed to flow at a slow rate (<0.5
ml/ min) with the gangliosides collected in the void volume. The void volume was
determined by blue dextran and the total volume by sodium chloride detected with
silver nitrate. The ganglioside fraction was lyophilized and further clarified if
necessary. The use of SepPak C18 cartridges to remove salts was also considered.

The SepPaks were fitted with either glass or plastic 10 ml syringes and washed with

49

10 ml water, methanol:water, methanol, chloroform:methanol, methanol,
methanol:water and 0.1 M sodium chloride each. Samples were applied to the
cartridge in a small volume of 0.1 M sodium chloride (< 1 ml) and reapplied twice to
ensure complete adsorption to the cartridge. The cartridge was then washed with 2
m1 of 0.1 M sodium chloride then 40 ml of distilled deionized water. The
gangliosides were then eluted with 2 ml of methanol and 10 ml of
chloroform:methanol (1:1), dried under a stream of nitrogen, and clarified as above if
necessary.

Two methods were evaluated for the purification of serum gangliosides, based
on solid-phase extraction techniques on commercially available cartridges. The first
method utilized a weak anion exchanger, Bond Elut aminopropyl cartridge. The
cartridge was washed twice with 10 m1 of methanol and 10 ml of chloroform.

Sample was applied in 2 ml of chloroform:methanol (85: 15) and reapplied twice to
ensure adsorption. The flow was gravity fed. The cartridge was then washed with 10
ml of chloroform, dichloromethylenezmethanol (8:2) and methanol. Gangliosides
were eluted with 0.5 N ammonium bicarbonate in methanol. The sample was dried
and desalted on a SepPak C18 cartridge, as outlined above. Mild alkaline
methanolysis was carried out as before and the sample was again desalted on a
SepPak C18 cartridge.

Another solid-phase extraction method attempted was based on silicic acid
chromatography. A Bond Elut silica gel cartridge was washed twice with chloroform
and methanol and the sample applied to the cartridge in 1 ml of chloroform and

reapplied twice. The cartridge was washed with 10 ml of chloroform,

50

acetonezmethanol (9:1), methanol, with the gangliosides being eluted with
methanol:water (9:1). The eluant was dried and subjected to mild alkaline

methanolysis with subsequent desalting on a SepPak C18 cartridge.

Isolation of Gangliosides by Diisopropyl ether:n-Butanol: Water Partition.

A method of isolating gangliosides from small samples from biological tissues
and fluids, described by Ladisch et al. (2), used a three-phase partition mixture of
diisopropyl ether, n-butanol, and water. The dried lipid extract was dispersed in 2
vol of diisopropyl ether:n-butanol (6:4) with sonication and vortexing. Water (1 vol)
or 0.1 M sodium chloride(aq.) was added and the mixture was sonicated and vortexed
for 2 min. The phases were separated with centrifugation at 1000 x g for 10 min.
The upper, phospholipid-containing organic phase was carefully removed with a 9 in.
pasteur pipet. To quantitatively remove phospholipids from the sample, the lower
ganglioside-containing aqueous phase was reextracted with fresh organic solvent. The
mixture was sonicated and vortexed a second time for two minutes and centrifuged to
separate the phases. After the upper organic layer was removed, the lower aqueous
phase was lyophilized to concentrate and remove all traces of solvent. The purified
gangliosides were desalted on a 10 ml Sephadex G-50 column, as stated above, with
the gangliosides collected in the void volume. The ganglioside fraction was
lyophilized and taken up in a small volume of chloroform:methanol (1:1) with
clarification by centrifugation if necessary. The authors stated that the gangliosides
were now pure enough for analysis by HPTLC but we found frequent interfering

impurities and low yields when applied to samples of l and 2 ml of serum or plasma.

51

Several modifications were attempted in order to increase the purity and the
yield. Modifications of the partitioning step itself consisted of repeating the partition
step a third time and altering the ratio of diisopropyl ether:n-butanol to 7:3. A third
partition was used to remove any potential phospholipids not removed in the first two
steps. The increase of diisopropyl ether in the organic mixture was derived from the
authors’ original report that the ratio of 6:4 was an optimization of both recovery of
gangliosides and the extraction of phospholipids. The ratio of 7:3 reported a slightly
higher recovery of gangliosides with a decrease in the removal of phospholipids. The
ratio of 7:3 was used with two partitioning steps and mild alkaline methanolysis, as
described above, was used to destroy any phospholipids that still remained after the
partitioning. The use of both a 10 ml column of Sephadex G-50 and SepPak C18
cartridges, both as described above, were compared for their ability to remove salts.
Furthermore, a second Sephadex G-50 chromatography step was. examined for its
ability to increase the purity of the mixture. Lastly, a small column of DEAE-
Sephadex A-25 (0.3 ml), utilized as stated above, was employed to remove residual

neutral glycolipids, with a subsequent desalting by an above method.

8 II I D' .
The individual isolation procedures were analyzed and compared by one-
dimensional high-performance thin-layer chromatography (l-D HPTLC). The
purified gangliosides from serum were quantitatively applied to a HPTLC plate 1 cm
from the edge and run in pre-equilibrated chambers until the solvent just reached the

top. The plates were visualized with resorcinol spray and compared. Table 1 lists

52

the many different isolation procedures attempted. These include the main
ganglioside purification step as well as the desalting method. Any additional steps
have also been listed. They have been rated according to the yield and degree of

purity of the procedure.

Modification of Total-Lipid Extraction

The initial step in the purification of gangliosides from serum typically
involves extraction of all lipid material with chloroform:methanol in various ratios.
We found that the use of 10 vol of chloroform:methanol per m1 of serum was
adequate for quantitative extraction of gangliosides from the serum. The use of
lyophilization prior to the extraction was attempted so as to make subsequent drying
steps more rapid. Dialysis of the serum and lyophilization was also examined as an
additional purification step. We found that the recovery of gangliosides was
considerably reduced when lyophilization, with or without prior dialysis, was used in
conjunction with a particular isolation procedure. Based on this, the basic isolation of

total-lipid material by chloroform:methanol 1:1 was employed with no modifications.

Chromatographic Isolation of Gangliosides

Many researchers have successfully obtained substantially pure preparations of
gangliosides using procedures based on DEAE-Sephadex A-25 ion-exchange. DEAE-
Sephadex removes the neutral and positively charged species and alkaline
methanolysis is then used to destroy residual negatively charged phospholipids.

Iatrobeads (silicic acid) chromatography is often used to remove other negatively

53

Table 1. Isolation of Circulating Gangliosides. All isolation based on 1 ml of
plasma. Abbreviations: C:M -chloroform:methanol; Seph.- Sephadex; MAM - mild-
alkaline methanolysis; DIPE/ButOH - diisopropyl ether/n-butanol; DEAE Seph -
diethylaminoethyl Sephadex; (+ + + +) - Excellent and (---) Very Poor.

54

 

 

+ dm>m imam
- 8:20
- 8-0 88 we moss-"Ea x N 8.32.283
+ + + + + + + 8-0 + 88 985 8-0 88 232 + 3 mosemeo x N
+ + + + + + + 8-0 88 252 + 3 :osEmaa x N
+ + + + + + 8-0 88 $22 + .8 mosemaa x N
+ + + + + + 8-0 + 88 ”-2.8. 8-0 .38
+ + + + + 8-0 88 x N to mosemaa x N
+ + + + + + 8-0 2.8 E. mesa-"Ea x N
+ + + + + 20 288m
+ + + + + + 8-0 88 to mosemza x N
- - - - - - 20 2388 252+ _...0 8:8 .20 2.8
- - - - - - 20 2.884.522 20 288m £2 55 2.8
+ + + 28.25
+ + + 20 8.88
+ + + 8-0 88 :22 + gem-mac .2 20 .2, 2 ..N
3:?— Eo; 850 wet—amon— 8282350 .20 5:28— :28..be BET—Sop.

«aq.-m

55

charged lipids such as sulfatides and free fatty acids. This procedure was based on
larger samples of plasma (10-50 ml); however, the authors state that it can
successfully be scaled down to one-third the size (4). We attempted to repeat this
result using 1 m1 of plasma and column sizes of approximately one-fifth (2 ml), one-
tenth (1 ml), and one-thirtieth (0.3 ml) the original size. Furthermore, the last
Iatrobead chromatography was not used because the sample would be pure enough for
HPTLC with only the previous steps. The potential contaminants of free fatty acids
and sulfatides do not appear purple with resorcinol spray and can be removed from
the area of interest by 2—D HPTLC (See below). The results following these
modifications were not encouraging. We found that while the purity was acceptable,
the recoveries were less than what was necessary for our studies. The nature of the
losses was unknown. The first wash of the DEAE-Sephadex did not appear to have
significant amounts of gangliosides as determined by HPTLC, yet it was difficult to
be certain about this point, due to the amount of neutral and positively charged lipids.
The addition of higher salt-containing solvents did not increase the yield. While this
method had potential usefulness for the isolation of small concentrations of
gangliosides, it was abandoned for the sake of time and simplicity.

The use of solid-phase extraction for the isolation of biological materials is an
emerging science that has found many uses due to its intrinsic simplicity and facility.
We attempted to use this technology to isolate gangliosides from serum. The
procedure is based on the method of ion-exchange chromatography, as is the
procedure based on DEAE-Sephadex. The Bond Elut cartridges are made of

aminopropyl functional groups on a solid support of silica gel as the weak anion-

56

exchanger. The negatively charged gangliosides are bound to the solid-phase while
contaminants are eluted with a variety of solvents of increasing polarity and salt
concentration. We found the protocol to be very poor both in terms of recovery and
purity. Resorcinol-positive purple bands corresponding to gangliosides were barely
visible amidst all the non-resorcinol positive spots. It was determined that small
homemade columns of DEAE-Sephadex, albeit not perfect, were far superior in
terms of both purity and yield, than commercially available solid-phase extraction

cartridges.

Isolation of Gangliosides by Diisopropyl Ether:n-Butanol Partition.

In 1983, Wong and Ladisch (6) found that a technique to delipidate serum by a
partitioning between diisopropyl ether, butanol and water (7) was not useful for the
removal of gangliosides with the other lipid material. In fact, they found that
gangliosides partitioned almost quantitatively into the aqueous phase while
phospholipids and less polar neutral lipids partitioned into the organic phase. Taking
advantage of this observation, they produced a protocol for the isolation of
gangliosides from plasma or serum (2, 3). The authors stated that this method was
suitable for qualitative work when used on 1 ml samples and qualitative and
quantitative if 2 ml of plasma or serum were used. Furthermore, the authors admitted
that there was approximately a 14% loss of 6M3, the least polar of plasma
gangliosides, into the organic phase. With GM3 being the most abundant ganglioside
in plasma, 50-70% of the total, this becomes important. Senn et al. (8) studied this

method extensively and found that the loss of GM3 into the organic phase increased

57

with decreasing sample size. They found that with samples of plasma as large as 5
ml, up to 24% of the GM3 was preferentially partitioned into the organic phase. This
suggests that samples of plasma or serum that were smaller than 5 m1 would suffer
substantial losses. In addition, they found the need for an additional step of DEAE-
Sephadex chromatography to remove residual neutral glycosphingolipids in order to
obtain HPTLC Chromatograms clean enough for densitometry. Our findings were
similar to those of Senn et al. (8). We experienced similar problems with purity and
recovery of small samples. Several attempts were made at modifying the Ladisch
procedure to increase recoveries from smaller amounts of serum because of the many
desirable characteristics of this method. These approaches have been summarized in
Table 1.

Initial modifications, additional partitions and/or additional chromatographic
steps on Sephadex G-50, were attempted to increase the degree of purity but met with
no success. Passing the semipure mixture obtained from the partition over a small
DEAE-Sephadex column and subsequent desalting on Sephadex G-50 was very useful
in removing many contaminants; however, they led to a slight decrease in the
recovery. Ladisch and Gilliard (3) observed that altering the ratio of DIPE and n-
ButOH to 7:3 from 6:4, respectively, increased slightly the partition of gangliosides
into the aqueous phase. This came at the cost of increased phospholipids partitioning
into the aqueous phase. We utilized this information and did the partition with
DIPEzn-ButOH ratio of 7:3, compensating for the increased phospholipid
contamination by utilizing mild-alkaline methanolysis to destroy the residual

phospholipids. The resulting FAMEs were removed by extraction with hexane. We

58

found that after desalting on a Sephadex G-50 column, our recoveries were much
greater and had higher proportions of GM3. The apparent decrease in the polarity of
the organic phase resulted in a more quantitative partition of the less polar
gangliosides, such as GM3, into the aqueous phase. Furthermore, the troublesome
interfacial fluff that makes quantitative removal of the organic phase difficult was

rarely encountered, aiding in the complete removal of the organic phase.

Conclusion

The goal for this particular part of this study was to identify or develop
method(s) for the isolation of gangliosides from serum. It was to be simple,
quantitative, and able to handle large numbers of small amounts in a reasonable
amount of time. After examining several different methods and their potential
modifications, it was determined that a modified method of Ladisch et al. (3) best
fulfilled the requirements. Briefly, 1 ml aliquots of plasma or serum were placed in
15 ml screw-capped test tubes. Methanol (5 ml) was added with vortexing and
sonication to produce a fine suspension. Chloroforrn (5 ml) was added with vortexing
and sonication, then the mixture was placed on a bed rocker in a cold room for at
least 6 h. The samples were centrifuged at 2600 rpm for 20 min to pellet the
insoluble material. The supernatants were removed to 100 ml round bottom flasks
and rotary evaporated untill only a small volume was left. These were then
transferred to 15 ml test tubes with small amount of chloroform:methanol (1:1). The
pellets were reextracted with 10 ml of chloroform:methanol for an additional 12 h.

Again, the supernatant was removed and rotary evaporated to a small volume, and

59

combined with the previous supematants. The volumes were then adjusted to
approximately 5 ml and the tubes placed in a -70 °C freezer for at least 4 hours to
precipitate any slightly soluble proteins or peptides. After centrifugation, the
supernatants were collected and dried under a stream of nitrogen and oil-pump
vacuum. The residues were taken up in 2 ml diisopropyl ether:n-butanol (7:3) with
vortexing and sonication untill a fine suspension was achieved. Distilled, deionized
water (1 ml) was added and the mixtures vortexed and sonicated for at least 2 min.
The mixtures were centrifuged for 10 min to separate the phases. A 9 in pasteur
pipet was used to remove the upper organic phase. The lower aqueous phase was
reextracted with 2 ml of fresh organic solvent, vortexed and sonicated, centrifuged,
and the organic phase removed. The lower, ganglioside-containing, aqueous phases
were lyophilized to concentrate and remove any traces of organic solvent. The
semipure mixtures of gangliosides were then subjected to mild-alkaline methanolysis
in 0.1 N sodium hydroxide in methanol for 2 h at 37 0C. After neutralizing with 1 N
acetic acid, the samples were extracted twice with hexane to remove any fatty acid
and other neutral contaminants. The samples were quickly dried under a stream of
nitrogen, taken up in 0.3 ml water, and applied to 10 ml Sephadex G-50 columns.
The columns were run at a flow rate of <0.5 ml/min and the gangliosides were
collected in the void volume, as determined by blue dextran. After lyophilization,
clarification and concentration, the samples were ready for analysis by HPTLC.
However, for analysis by gas chromatography, an additional step using DEAE-
Sephadex was necessary to remove the residual neutral glycolipids and other

contaminants. This step also required desalting on a 10 ml Sephadex G-50 column.

60

We found these methods to be quite suitable for the needs of this study.

REFERENCES

Folch, J., Lees, M., and Sloan Stanley, G. H. (1956) J. Biol. Chem. 226, 497
Ladisch, S., and Gillard, B. (1987) Methods Enzymol. 138, 300-306

Ladisch, S., and Gillard, B. (1985) Anal. Biochem. 146, 220-231

Ledeen, R. W., and Yu, R. K. (1982) Methods Enzymol. 83, 139-191

Yu, R. K., and Ledeen, R. W. (1972) J. Lipid Res. 680, 680-686

Wong, C. G., and Ladisch, S. (1983) J. Lipid Res. 24, 666-669

Cham, B. E., and Knowles, B. R. (1976) J. Lipid Res. 17, 176-181

Senn, Haas-J., Orth, M., Fitzke, E., Wieland, H., and Gerok, W. ( 1989) Eur.
J. Biochem. 181, 657-662

61

CHAPTER 3

Characterization of Gangliosides by Two-Dimensioal High-Performance Thin-

Layer Chromatography (HPTLC)

D.A. Wiesner and C.C. Sweeley

November 19, 1993

Department of Biochemistry
Michigan State University
East Lansing, MI 48824

J. Lipid. Res. (1993) In preparation.

Running Title: HPTLC of gangliosides.

62

63

Introduction

Thin-layer chromatography has become a universal tool for the separation and
identification of gangliosides. Technological advances in the sorbents and solvent
systems, as well as the means of detection and quantitation, have had a tremendous
role in the development of glycolipid science. Commercial high-performance thin-
layer plates allow for analysis of much smaller amounts of lipid with much better
resolution than earlier home-made thin-layer plates. Scanning densitometry has
simplified the quantitation of minute quantities of gangliosides that have been resolved
by one-dimensional high-performance thin-layer chromatography (l-D HPTLC) with
reproducible results. More recently, image densitometry has simplified HPTLC even
more. Gangliosides analyzed on 1-D HPTLC and visualized by a resorcinol-HCl
spray can easily be resolved and quantitated with image densitometry. The HPTLC
plate is scanned by a camera scanner and digitized into a computer image. Then,
specialized software computes the size and optical density of the bands and integrates
them to produce an integrated optical density. Image densitometry also makes the
analysis of two-dimensional high-performance thin-layer chromatography (2-D
HPTLC) much easier and useful. Resorcinol-positive spots can be detected and
quantitated in the same manner as l-D HPTLC. This becomes important when one
realizes the qualitative differential resolving power of 2-D HPTLC without a suitable
method for quantitative analysis.

In this portion of the study, both 1-D and 2-D HPTLC are examined for their

usefulness in analyzing gangliosides isolated from plasma. Various solvents and

64

running conditions are explored for use with both 1- and 2-D HPTLC. An image
densitometer, or image analyzer, is described and its usefulness explained. In
addition, a map of standard gangliosides as well as some rare gangliosides found in
serum and tumor tissue is created using 2-D HPTLC. This map will be useful in the
identification of gangliosides in complex mixtures such as those found in human

serum.

Materials and Mathggs

All gangliosides were obtained from Sigma (St Louis, MO) except as follows.
Ganglioside GM4 was isolated from chicken liver according to the method of Shiraishi
and Uda (1). Gangliosides GD3, GMlb, GTla, and GT2 were a gift from Dr. R.K.
Yu of the Medical College of Virginia, Virginia Commonwealth University.
Gangliosides 6C, sialyldi-Y-Z, sialyl Le", sialyl Lea, 2,6—sialylparagloboside,
sialylnorhexaosylceramide, and sialyl I were a gift from Dr. E. Nudelman of the
Biomembrane Institute. Ganglioside N-glycoyl-GD3 was a gift of Dr. S. Handa of
the Tokyo Medical and Dental University.

All solvents were of analytical grade or better. High-performance thin-layer
plates ( Kiesel gel 60, 10 x 10 cm) were obtained from Merck (Darmstadt, Ger). The
resorcinol spray was made according to Svennerholm et al (2, 3). Briefly, 5 ml of a
stock solution of resorcinol (2 g of resorcinol recrystallized from benzene in 100 ml
of water) is added to 40 ml of concentrated hydrochloric acid containing 0.125 ml of
0.1 M copper sulfate and the total volume brought to 50 ml with distilled, deionized

water. This reagent is stable at 4 °C until it turns green. The stock solution is stable

65

for many months at 4 °C.

High-Performance Thin-Layer Chromatography

All HPTLC plates were prerun in methanol-diethyl ether (1:1) and preactivated
by heating at 110 °C for 30 min prior to use. Plates were used immediately after
preactivation.

For l-D HPTLC, a solution of gangliosides was spotted in bands 5 mm wide,
1.5 cm from the bottom edge of the plate in microdroplets with a 10 pl Hamilton
syringe. By placing the bands 2 mm apart, up to 13 different samples could be
analyzed on the same plate. In many cases, 1500 pmol of GM4 ganglioside is run
with the test samples for quantitation purposes. Several different chromatography
solvents were tested in order to determine the most suitable solvent system for our
purposes, as follows: a) chloroform-methanol-0.2 % CaClz' 2H 20 (aq) (60:40:9),
b) chloroform-methanol-2.5 N NH3(aq) (60:40:9), c) chloroform-methanol-O.1 M KCl
(aq) (60:40:9) and d) chloroform-methanol-2.5 N NH3 and 0.1 M KCl (60:40:9). In
all cases, the solvent (35 ml) was placed in a wicked (Gel Blot Paper, Schleicher and
Schuell, Keene, NH) metal chromatography chamber (Aldrich, Milwaukee, WI.) and
allowed to equilibrate for at least 2 h. The plate was placed in the chamber and
allowed to develop until the solvent front reached a line 1.5 cm from the top of the
plate. The plate was removed and thoroughly dried in a 34 °C drying oven. After
the plate was fully dried, it was ready for visualization and analysis.

In the case of 2-D HPTLC, the plates were prepared in a similar way as

described above. Ganglioside mixtures were spotted in one comer, 1.5 cm from each

66

edge. The spotting was done conveniently with a 10 #1 Hamilton syringe.
Ganglioside GM, (1500 pmol) was applied directly on top of the spot containing the
ganglioside mixture. Once again, several different solvent systems were considered to
determine the one best-suited to the needs of this study. The different solvent systems
considered were a) chloroform-methanol-0.2% CaC12(aq) (50:40: 10) and chloroform-
methanol-2.5 N NH3 (aq) (50:40:10) (4), b) chloroform-methanol-O.25% KCl and
chloroform-methanol-2.5 N NH3 and 0.25% KCl(aq) (50:40:10) (4), c) n-propanol-
0.1% CaC12(aq) (80:20) and n-butanol-pyridine-O.l%Cac12(aq) (60:45:15) (5), and d)
chloroform-methanol-0.2% CaC12(aq) (50:40:10) and n-propanol-11.4 N NH3(aq)

(2: 1) (6). The solvent for the first dimension is listed first and the second dimension
listed second. In order to achieve the highest degree of reproducibility between plates
in different runs and within the same run, the running conditions were made as
similar as possible. Taking this idea further, up to 10 plates were run at one time.
This was done in multiple 600 m1 Pyrex beakers, wicked with Gel Blot paper, and
containing exactly 35 ml of solvent each. Watch glasses (6 in. diam.) were
conveniently used as lids and the spouts taped to insure a proper seal. The chambers
were allowed to equilibrate for at least two hours prior to use. The plates were
placed in the chambers and allowed to develop until the solvent front just reached a
line etched in the surface, 1.5 cm from the edge of the plate. At that point, the plates
were removed and allowed to dry thoroughly in a 34 °C drying oven. In the
meantime, the pyrex beakers were emptied and the wicks removed. After the beakers
had dried, new wicks were placed in the beakers along with 35 ml of the second

dimension solvent. The chambers were again allowed to equilibrate for at least 2 hr

67

before use. The dried plates were then rotated 90° from the first direction and placed
in the chambers and developed until the solvent front reached a second line etched 1.5

cm from the edge of the plate. The plates are removed and dried in the drying oven.

Visualization of Gangliosides

The visualization procedure for gangliosides on HPTLC plates was the same
for both 1- or 2-D HPTLC. A non-destructive method involved placing the plate in
an iodine chamber, which stains the gangliosides and other lipids yellow. After
marking the presence of ganglioside, the iodine can be removed by allowing the plate
to sit in the open air for the iodine to evaporate from the spots. Coomassie Brilliant
Blue staining of the gangliosides is also possible (7). Plates are submersed in a 20%
methanol solution containing 0.03% Coomassie Brilliant Blue R. Destaining is done
in a 20% methanol solution. Resorcinol-HCI spray, a destructive method which is
specific for sialic acid, is the most useful for the analysis of gangliosides. Plates are
sprayed lightly with the resorcinol spray in an air-powered atomizer, covered with a
clean glass plate, and placed in a 110 °C oven for 10 minutes. Gangliosides are

visualized as deep bluish-purple bands or spots.

Two-Dimensional Mapping
A two-dimensional map of standard gangliosides was created with HPTLC for
use in identifying gangliosides found in human serum. After a core framework was

developed with a mixture of standard gangliosides, the more unusual gangliosides

68

were included and their migrations, relative to the framework, recorded. When
necessary, the Biolmage Visage 110 image analyzer was used to map and compare

gangliosides on HPTLC plates using overlap and triangulation software.

Image Densitometry

Image densitometry was utilized to quantitate the resorcinol-positive spots from
both 1- and 2-D HPTLC plates. The plates were scanned by a Visage 110 Biolmage
image analyzer (Millipore, MA). The resulting computer image was 1024 pixels by
1024 pixels, with 155 different color shades. An integrated optical density value was
obtained by measuring the optical density and integrating it over the area of the spot.
The linearity of the relationship between the image densitometric response and the
ganglioside content was assessed using mixtures of gangliosides GM3, GMI, GD3,
and GDla' The densitometric response was relative to that of the internal standard

GM4.

R n D' i

Both one- and two-dimensional high-performance thin-layer chromatography
were considered for the characterization of gangliosides isolated from human serum
and plasma. One-dimensional HPTLC is inherently simpler and more able to handle
multiple samples but this can be at the cost of resolving power. Up to 13 different
samples can be assayed on a single plate in one dimension while 2-D HPTLC can be
run on only one sample per plate. The value of 2-D HPTLC comes as much greater

resolving power over l-D HPTLC. Figure 1 shows the two—dimensional map created

69

for the identification of gangliosides isolated from serum, along with a photograph of
an actual 2-D HPT LC plate in which gangliosides from 1 ml of serum have been run.
Table 1 lists all the gangliosides, their source, and means of identification. By
looking at the standard map, one can easily see the resolving power of two-
dimensional chromatography. Very complicated mixtures of gangliosides can be
resolved and identified using 2-D HPT LC. The solvent system of chloroform-
methanol-0.2% CaC12(aq) (50:40: 10) in the first dimension and n-propanol-11.3 N
NH3 (aq) (2:1) in the second could potentially separate 25 different species of
gangliosides, as shown in Figure 1. This solvent system was chosen over the others
mentioned in the materials section for its exceptional resolving power and its
reproducibility. Furthermore, when using this solvent system, gangliosides that were
identical with respect to the Oligosaccharide portion of the molecule were observed as
a single band. Many other solvents resolve gangliosides that are alike in the
Oligosaccharide into two or more bands based on their fatty acid chain length and
presence or absence of hydroxyl groups. This can help simplify an otherwise
complicated mixtures of gangliosides differing in their fatty acid composition. The
2-D pattern of spots was very consistent from run to run when certain precautions
stated above were followed. Of particular importance was the removal of the plate as
soon as the solvent front reached the mark. Allowing the solvent to "flow over the
top of the plate" could cause distortions that were difficult to resolve. For the one-
dimensional analysis, the solvent of chloroform-methanol-2.5 N NH3 and 0.25%

KCl (aq) (60:40:9) was found to be the most suitable, for reasons stated below, in

terms of linearity and reproducibility.

70

Figure 1. Two-Dimensional Map of Gangliosides. Standard gangliosides were
mapped out in two dimensions using chloroform-methanol-0.2% CaCl2 (aq)
(50:40:10) in the first dimension and n-propanol-l 1.3 N NH3(aq) (2:1) in the second.
Gangliosides found in serum or plasma, as shown in the photograph, have been
shaded, while other gangliosides were left clear. Unidentified non-ganglioside
material is represented as dark spots.

 

Figure la

 

 

 

 

9
(D
672 SN C I q M64
\0 MGG QSL x
3L; GM1
‘GT1b GT1a
.\
GD‘lb
.—GQ1b
O—ec
GSdiY-z Resorcinol Negative
Origin
Figurc 1b Figure 10

   

71

72

mixture of gangliosides ranging from 10 pmol to 5 nmol per component The
densitometric responses from the image analyzer were plotted and the results can be
seen in Figures 2 and 3 for 1- and 2-HPTLC, respectively. The lines represent the
results of a linear regression of the points corresponding to increased amounts of
ganglioside based on lipid-bound sialic acid content. In all cases the values of R2
(variance) were better than 0.99. However, when one compares Figure 2 with Figure
3, the superiority of 2-D HPTLC over l-D can be seen in terms of molar
densitometric response. The 2-D standard curve shows an equimolar response for
equal moles of lipid-bound sialic acid. The l-D curve shows a differential response
for different gangliosides containing equal amounts of sialic acid. This could be a
result of different band sizes or geometries following migrations. This conclusion
was reported by Mullin et al. (8), who found that different solvents and different
distances the solvent front travels can affect the densitometric response. They further
reported that this was not due to a difference in extinction coefficients of the
resorcinol reaction of sialic acids of different gangliosides on silica gel. Our results
were similar to theirs. We also found that equimolar amounts of ganglioside that
were spotted on an HPTLC plate in increasing size bands and not developed in a
solvent had increasing densitometric response, when visualized with resorcinol (Figure
4). Furthermore, the solvent chloroform-methanol-O.2% CaC12(aq) (60:40:9)
produced curves like Figure 5 similar to the report of Mullin (8) while chloroform-
methanol-2.5 N NH3 and 0.25% KCl (aq) (60:40:9) produced much more linear plots
(Figure 3.). These results suggest that solvent systems that differ in their salt content

can have a considerable effect on the migration patterns and densitometric response.

73

Figure 2. Standard curve of gangliosides analyzed by 2-D HPT LC, based on
lipid-bound sialic acid content. The solvent in the first dimension was chloroform-
methanol-0.2% CaC12(aq) (50:40:10) and n-prOpanol-ll.3 N NH3(aq) (2:1) in the
second. Gangliosides were visualized with resorcinol and quantitated by image
densitometry based on 1.5 nmol GM4 added as an internal standard. The inset box is
an enlargement of the region of the plot from 0 to 1 nmol. Lines represent linear
regressions of the points and are identified in the legend.

74

22 268 832.1225 .0 868202

 

 

 

 

 

 

-i b In

 

 

 

C\mv EcoEEa z QZLOGanoLm 5m
823.183 N8888-_o§5mE-E.o.o.2€ x: ”32.928

0.385 DIN Low 0?.50 0.80505

000.

000..

(009 itrwe/xo) asuodsaa

.N as»: ,

75

Figure 3. Standard curve of gangliosides analyzed by l-D HPTLC, based on
lipid-bound sialic acid content. The plate was developed in chloroform-methanol-
0.25% KCl and 2.5 N NH3(aq) (50:40:10). Gangliosides were visualized with
resorcinol and quantitated by image densitometry based on 1.5 nmol GM4 added as an
internal standard. The inset box is an enlargement of the region of the plot from 0 to
1 nmol. Lines represent linear regressions of the points and are identified in the
legend.

76

22 268 venom-.223 .0 8.2.262

 

 

r 0°C—
.50: .
........ nQUI 9
..... 3:01 .
1 0:01 0 ........ \\.s OOON

 
    
  

0.0 0.0 v.0 Nd 0.0

1 d a

      

  

(009 L-vwe/xo) asuodsea

 

 

 

 

 

1 x a coon
... - ooN
.- 00v
1 000
r .Q t r 000 .. 000v
. _ a. 000.
. t a; 4
. a com.
fix — p p p 080

 

Aofiovnomv ..Iz ZnN 0:0 8x Nmmdlocanuofi152000.830 J:o>_om

049.me or; 00 0.550 Upmvcmam
£95»...—

77

Figure 4. Standard curve of gangliosides analyzed by l-D HPTLC, based on
lipid-bound sialic acid content. The plate was developed in chloroform-methanol-
0.2% CaCl2 (60:40:9). Gangliosides were visualized with resorcinol and quantitated
by image densitometry.

78

Figure4.
Standard Curve of 1—D HPTLC

Solvent: chloroform- methanol-0.2%CaCl. (60/40/9)

 

 

 

 

0.7 - - r 1
U
0 -GM. >
0.6 - . _GM, 1
V -60“ . v
V "GTu
0.5 e -
3 0.4 - -
t:
8. o
m '
a 0.3 - " -
V
/ \v
' o
O
0.2 - . / -
. o
0.1 - //V\/ ..
“/
0.0 1 L l l
0.0 0.2 0.4 0.6 0.8 1.0

Nanomoles of Lipid-Bound Sialic Acid

79

Alkaline solvent systems have been suggested to improve the resolution of HPTLC
(4, 6). Our results show that the use of both potassium salt and ammonia in the
solvent system can improve the linearity of linear response when quantitated by
resorcinol and densitometry. Figure 2 shows that in the 2-D HPTLC with a solvent
system that contains salt in the first dimension and is alkaline in the second, somehow
the shape and geometry of the bands are controlled in a uniform way so as to give an
equimolar response of lipid-bound sialic response for equal moles of lipid-bound sialic

acid.

Identification and Quantitation of Gangliosides from Human Serum

Gangliosides isolated from 1 ml of serum were concentrated into the bottom of
a 6 ml conical centrifuge tube. GM4 (1500 pmol, 3 pl) was added and the entire
sample was placed on the corner of the plate in a quantitative manner. The plate was
run in two-dimensions as described above, stained with resorcinol and analyzed by
image densitometry. Figure 1 shows a photograph of a typical 2-D HPTLC plate of
gangliosides isolated from serum. Figure 1 also shows the map created in with 2-D
HPTLC and the location and identification of gangliosides found in serum. Further
information as to the quantitation of the gangliosides will discussed in detail in

Chapter 6.

ganglaaion

In this study we have presented a useful, reliable method for the thin-layer

chromatographic analysis of gangliosides isolated from serum or plasma. We suggest

80

that the methods described here would also be useful for the analysis of other
biological tissues or fluids. The solvent system of chloroform-methanol-0.2% CaCl2
(aq) (50:40: 10) in the first dimension and n-propanol-l 1.3 N NH3 (aq) (2:1)
adequately resolved the 13 or more different gangliosides found in serum. This,
combined with the use of image densitometry, provided linear responses from 30
pmol to 5 nmol lipid-bound sialic acid. As little as 10 pmol was sometimes detectable
but not reproducibly. Spots corresponding to gangliosides having lipid-bound sialic
acid concentrations lower than 100 pmol should be viewed with caution due to high
standard deviations. Image densitometry was also useful in developing the map from
which the identifications of unknown gangliosides were made. The images obtained
by scanning the 2-D HPTLC plates are easily saved indefinitely on magnetic tapes.
Actual plates must be covered and frozen, yet still fade after time. These images can
be compared to each other and even overlaid on the computer for help in locating the
many spots found on the map. When necessary, the computer can compensate for
minor distortions when two plates were being overlaid and compared. This method of
analysis was useful in characterizing gangliosides isolated from the sera of breast
cancer patients and an appropriate control group, as will be discussed in greater

detail in Chapter 6.

REFERENCES
Shiraishi, T., and Uda, Y. (1986) J. Biochem. 100, 553-561
Svennerholm, L. (1957) Biochim. Biophys. Acta 24, 604
Ledeen, R. W., and Yu, R. K. (1982) Methods Enzymol. 83, 139-191

Ledeen, R. W., Haley, J. E., and Skrivanek, J. A. (1980) Anal. Biochem.
112, 135-142

Nakamura, K., Ariga, T., Yahagi, T., Miyatake, T., Suzuki, A., and
Yamakawa, T. (1983) J. Biochem. 94, 1359-1365

Chigomo, V., Sonnino, S., Ghidoni, R., and Tettamamti, G. ( 1982)
Neurochem. I. 4, 397-403

Nakamura, K., and Handa, S. (1984) Anal. Biochem. 142, 406-410

Mullin, B. R., Poore, C. M. B., and Rupp, B. H. (1983) J. Chromatog. 278,
160-166

81

CHAPTER 4

Microscale Analysis of Glycosphingolipids by Methanolysis, Peracetylation and

Gas Chromatography.‘

D.A. Wiesner and C.C. Sweeley2

Department of Biochemistry
Michigan State University
East Lansing, MI 48823
(51 7) 353-5327
Fax (517) 353-9334

Anal. Biochem. (1993), In Press.

Running Title: Gas Chromatography of Glycosphingolipids.

Subject Category: Chromatographic Techniques, Carbohydrates.

82

83

ABSTRA T

A method is described for the analysis of pure samples of individual
glycosphingolipids by microscale methanolysis, peracetylation, and gas
chromatography. Solvolysis of glycosphingolipids in dry methanolic HCl and
peracetylation were conducted in a single 4.5 cm sealed capillary tube (2 mm id),
after which the products were directly injected into a gas chromatograph. Total-
component analysis (i.e. analysis of the sugar, fatty acid, and sphingosine moieties)
was possible after a 45 min chromatographic run. Time-course studies of the acid-
catalyzed methanolysis of GMl ganglioside at 80, 110, and 150 °C showed that
methanolysis was complete after 2 h at 110 °C. Rates of methanolysis of individual
components were compared and the release of the fatty acid moiety from the long-
chain base was shown to be the slowest reaction. The methanolysis of all glycosidic
bonds was complete in 0.5 h. Peracetylated methanolysis products were very stable
over time and provided for good gas chromatographic detection of subnanomolar
amounts of hexose, hexosamine, fatty acid, sialic acid, and long-chain sphingoid base
components. Recoveries of fucose and N-acetylglucosamine were determined with
reference samples of fucosylal-Zlactose and lacto-N-fucosylpentaose 11. Applications
of the method are presented for the component analysis of a gift mixture of 6C
ganglioside and sialyldi-Y-2 ganglioside and analysis of GM, isolated from human

mama

84

Intragagian

Glycosphingolipids (GSLs)3 are of interest because of the several biological
roles they have been proposed to have, such as in cellular interaction, differentiation,
and oncogenic transformation (1). More than two hundred GSLs have been
discovered and characterized (2). Analyses of their levels in physiological fluids and
tissues are difficult, especially because of their presence as mixtures in very low
amounts.

While it has been reported that the carbohydrate composition of
glycosphingolipids cannot be determined by a single method (3), there are a number
of different approaches for such analyses. Typical methods used for analysis of the
carbohydrate moiety of glycolipids involve either acid-catalyzed hydrolysis or
methanolysis. Hydrolysis is achieved using sulfuric acid (4), hydrochloric acid (5),
trifluoroaoetic acid (6), or acetic acid with sulfuric acid followed by conversion to
alditol acetates and analysis by gas chromatography (GC) (7). Problems with this
type of determination include the need for different conditions of hydrolysis for
different carbohydrates and, often, the complete destruction of sialic acid.
Furthermore, hydrolysis often requires complex, time-consuming work-up procedures.

Methanolysis in dry methanolic hydrochloric acid (HCl), followed by re-N-
acetylation followed by trimethylsilylation and analysis by GC, is often used instead
of hydrolysis and is the current method of choice in many laboratories. However,
complete methanolysis may require 12-24 h, depending on the methanolysis

conditions, and different strategies are often required for the complete recovery of

85

different components of both the carbohydrate and ceramide parts. To avoid losses of
the methyl glycosides during concentration, neutralization of the methanolysis
mixture is required. This involves either trituration with silver carbonate (8, 9) or
passing over an ion-exchange column (10-12), both of which are cumbersome and
unsuitable with small samples. The problem of neutralization has been addressed
somewhat for glycoproteins in a report by Chaplin (13) but other difficulties with this
procedure and its application to GSLs have not been adequately addressed.

In this paper, we describe a somewhat different approach for the methanolysis,
peracetylation, and gas chromatography of glycolipids, designed especially for use
with very small samples. The method requires only 0.5 h for complete release of all
sugar components and 2 h for total methanolysis of the amide linkage of fatty acid to
the sphingoid base, with no need for neutralization. The use of peracetylation
instead of trimethylsilylation for derivitization prior to GC results in stable derivatives

with useful chromatographic properties.

Matam‘ 5 and Mghadg
Materials

All gangliosides were obtained from Sigma (St. Louis, MO.) unless otherwise
stated. The gangliosides 6C (NeuAca2-6GalBl-4GlcNAcBl-3GalBl-4(Fucal-
3)G1cNAcBl-3GalBl-4Glc-Cer) and Sialyldi-Y-2 ( NeuAca2-6GalBI-4(Fucal-
3)GlcNAcB1-3GalB1-4(Fucor1-3)GlcNAcBl-3Ga161-4Glc-Cer) were a generous gift of
Dr. E. Nudelman of the Biomembrane Institute, Seattle WA. 2’-Fucosyllactose

(2’FL; Fucal-ZGalfll-4Glc) and lacto-N-fucopentaose II (LNFP; GalBl-3(Fucal-

86
4)G1cNAcBl-3GalBI-4Glc) were obtained from Oxford Glycosystems (Rosedale, NY).

N-acetylneuraminic acid (sialic acid, NeuAc), glucose, and heptadecanoic acid were
also obtained from Sigma. All solvents were of reagent grade or better, while
chloroform and methanol were HPLC grade. Pyridine was distilled before use.
Methanolic HCl, 0.75 N, was prepared by bubbling gaseous HCl through
methanol, then titrated with base to obtain the normality. This concentrated stock
solution was stored at -20 °C and diluted with methanol to 0.75 N as necessary. The
stock solution was titrated periodically to determine any loss of HCl. The stock

reagent was stable for up to one month.

Acid- Catalyzed Methanolysis and Peracetylation.

Samples (10 pmol to 10 nmol) of a neutral glycosphingolipid or ganglioside
were dissolved in chloroform:methanol (1: 1) and transferred into capillary tubes, 2.5
cm x 2 mm, with one end sealed. The solvent was quickly removed from the tubes
by placing them in a high vacuum (i.e. oil pump). Dry 0.75 N methanolic HCl (25
pl) and methyl acetate (5 pl) were added and the tops of the tubes were sealed in a
flame. The sealed tubes were placed in an oven at the stated temperature for an
appropriate time. Following heating, the tubes were allowed to cool briefly, one end
scored and carefully cracked open. The methanolic HCl was quickly removed with a
high vacuum. Methyl heptadecanoate (HDA, derived by solvolysis of heptadecanoic
acid with methanolic HCl), 2.1 nmol in 5 pl of methanol, was added as the internal
standard. Again, the methanol was removed in vacuo. Pyridinezacetic anhydride

(1:1, 5 pl), made fresh daily, was added, the tubes were rescaled in a flame, and the

87

peracetylation was allowed to proceed for at least 1.5 h at room temperature.

Gas Chromatography

Tubes were scored at one end and cracked open just prior to injection of
aliquots into the gas chromatograph. The sample, in 'pyridinezacetic anhydride, was
injected (1 to 3 pl) into an Hewlett Packard (Avondale, PA) 5890 gas chromatograph
fitted with a 60 m DB-l column (J & W Scientific, Folsom, CA) and a flame
ionization detector. After injection, the temperature was held at 150 °C for 15 min,
then raised to 300°C linearly at 4 °C per min. Hydrocarbons (12-20, 24-30, even)
were coinjected for use in the calculation of retention indices. Peak areas were

measured with a Hewlett Packard 3392A integrator.

Derivatization Time and Temperature Optimization

Methanolysis of GM, ganglioside (8 nmol) was carried out in duplicate at
temperatures of 80, 110, and 150°C. The reaction was stopped after periods of 0 to
24 h by removing the tubes from heat, brief cooling, opening, and solvent removal in
vacuo. The samples were analyzed by gas chromatography following peracetylation
for 2 h. Yields of glucose, galactose, N-acetylgalactosamine (GalNAc) , fatty acids,
sialic acid, and long-chain sphingoid base (LCB) were relative to that of the internal
standard. Absolute amounts of these products can be derived from the results,
provided accurate molar responses of each relative to internal standard is known.

Stability of the peracetylated derivative was determined by conducting the

methanolysis of GM, for 2 h, peracetylation, and injecting at increasing times

88

following derivitization.

Isolation of GM, from Plasma

Gangliosides were isolated from plasma using a modified method described by
Ledeen and Yu (14). Heparinized plasma was extracted twice with 20 vol of
chloroform:methanol (1:1), the supernatants were combined and solvent removed
under a stream of nitrogen. The total lipid extract was fractionated on DEAE-
Sephadex A-25 (Pharmacia Fine Chemicals, Piscataway, NJ.) and the ganglioside
fraction was subjected to mild-alkali-catalyzed methanolysis in methanolic 0.1 N
NaOH to destroy glycer0phospholipids. After dialysis, the semipure mixture was
applied to an Iatrobeads column (15 ml) to remove acidic contaminants. The pure
gangliosides were then applied to high-performance thin-layer chromatography
(HPTLC) plates (E. Merck, Darmstedt, Germany) which were developed in
chloroform:methanol:2.5 N NH, (50:40:10). Ganglioside bands were identified by
placing the dry plate in a chamber containing iodine crystals until the bands became
visible. After allowing the iodine to sublime off, the GM, band was then scrapped off
and eluted from the gel with chloroform:methanolzwater (5:5: 1). Dissolved silica and
other contaminants were then removed with a 10 cm Sephadex G-50 column, eluted

with water.

89
Rgalts and Discussion

Gas Chromatography of Peracetylated Methanolysis Products of GM ,

Figure 1 shows a typical Chromatogram of the peracetylated methanolysis
products of GM, ganglioside (GalBl-3GalNAcBl-4(NeuAca2-3)Ga161-4Gchl-lCer).
GM, was chosen because it contains most of the components (except fucose and N-
acetylglucosamine which have unique peaks) that one would likely find when studying
a mixture of gangliosides. The separation and response of the different peaks is
apparent. Table 1 shows the peaks and retention indices corresponding to the O-
acetylated methanolysis products of galactose, glucose, N-acetylgalactosamine, sialic
acid (N-acetylneuraminic acid), d18:1 and d20:1 sphingosines, and several fatty acids.
For comparison, Table 1 also reports the retention indices from peaks corresponding
to O-acetylated methyl fucoside, N-acetylglucosaminide, 0,0,N-triacetyl
dihydrosphingosine, and 1,3,4-tri-O-acetyl-N-acetyl phytosphingosine although they
do not appear in this sample of GM,. All peaks are easily detectable and
quantitatable. A mix of hydrocarbons, C14 to C30, except for C20, was coinjected
for use in the calculation of retention indices. None of the hydrocarbon peaks
interfered with any of the peaks of interest except for overlap of dihydrosphingosine
and C28 hydrocarbon. This can be overcome by the removal of C28 from the
hydrocarbon mix if dihydrosphingosine is suspected. The use of methyl
heptadeeanoate as the internal standard was chosen to avoid interference with any of
the glycolipid products. This internal standard had one centrally located peak in the
Chromatogram and was stable under the running conditions. Samples ranging from 10

pmol to 10 nmol were found to be reproducible and the peaks were linear in

90

Figure l. Chromatogram of the peracetylated methanolysis products from GM,
ganglioside. See Table l for peak identification and text for derivatization and
chromatography conditions.

91

Figure 1.

 

 

 

 

 

NH

 

2

2:22

21:31.1

 

 

 

 

 

 

 

 

 

TABLE 1.

 

 

Gas Chromatography of Peracetylated Methanolysis Reference

Compounds and Products from GM, Ganglioside.

 

 

Parent Structure Derivative Structure Peak“ Retention
Index”
Fucose Methyl at-L-fucopyranosidec 1613
Methyl (i-L-fucopyranosidec 1614
Galactose Methyl n?-D-galactofuranoside l 1886
Methyl a-D-galactopyranoside 2 1891
Methyl B-D-galactopyranoside 3 1999
Glucose Methyl a-D-glucopyranoside 4 1904
Methyl B-D-glucopyranoside 5 1910
N-Acetylgalactosamine Methyl 2-acetamido-2-deoxy-n?-D-galactofuranoside(?) 6 2067
Methyl 2-acetamido-2-deoxy-n?-D-galactofuranoside(?) 7 2073
Methyl 2-acetamido-2-deoxy-a-D-galactopyranoside 8 2088
Methyl 2-acetamido-2-deoxy-B-D-galactopyranoside 9 2094
N-Acetylglucosamine Methyl 2-acetamido-2-deoxy-oi-D-glucopyranosideF 2101
Palmitic Acid Methyl Palmitate (Cl8:0) 10 2127
Methyl 2-acetamido-2-deoxy-B-D-glucopyranoside‘ 2138
Arachidic Acid Methyl Eicosanate (C20:0) 11 2331
Behenic Acid Methyl Docosanate (C22:0) 12 2520
Sialic Acid Methyl (methyl 5-acetamido-3,S-dideoxy-D-glycero-D- 13 2565
galacto-a-nonulo-pyranosid) onate
Sphingosine 0,0,N-triacetylsphing-4-enine (d18: l) 14 2689
3-O-methyl-O,N-diacetylsphing-4-enine‘ 15, 16 2702,273 l
5-O-methyl-O,N-diacetylsphing-3-enine‘I
Sphinganine Dihydrosphingosine (d18:0)‘ 2800
4-Hydroxysphinganine Phytosphingosine (118:0)‘ 2879
Cm-Sphingosine 0,0,N-triacetyleicosasphing-4-enine (d20: l) 17 2891
O-methyl-O,N-diacetyleicosasphing-4-enine‘ 18 2902
O-methyl-O,N-diacetyleicosasphing-B-enine‘ 19 293 l
Hydrocarbons Cl8-20 A,B 1800,2000
Hydrocarbons C24-30 C .D, 2400,2600.
E.F 2800,3000

 

:Refers to Fig. 1
"Based on co-injected hydrocarbons;
cNot found in GM,, but provided for comparison;

4The two O-methyl sphingosines (C18) are 3-O-methyl-A‘-sphingosine and S-O-methyl-A3-sphingosine;
their order of elution from the GC column is unknown.The same two peaks were observed for

eicosasphingosine (d20: l)

92

 

93

response. Actual recovery of pure glycosphingolipid when compared to recoveries of
authentic methyl glycosides was found to be better than 95%.

The multiple peak profiles observed in Figure l are a result of the
methanolysis reaction of carbohydrates which produces anomeric mixtures of methyl
a- and B-glycopyranosides along with some glycofuranosides, as in the case of
galactose. Fucose and N-acetylglucosamine also have multiple peak profiles
representing different anomeric and ring forms resulting from methanolysis. With
this method, fucose had only two major peaks as opposed to other reports (13,15).
These peaks have all been identified in Table 1. Sphingosine ( trans-A4-sphingenine,
d18: 1) has three peaks resulting from methanolysis. The first two (see Table l) are a
result of O-methylation of the hydroxyl at the C3 carbon of the sphingosine molecule.
The third is the peracetylated derivative of he sphingoid base. This was also seen
with the 20 carbon homolog (4-eicosphingenine, d20:1). Both phytosphingosine and
dihydrosphingosine had only one peak, which gives evidence of the need for the
double bond adjacent to the hydroxy group for O-methylation. Other possible peaks
in all cases were minor and not considered.

The ratio of the combined areas of the three galactose peaks to the combined
areas of the two glucose peaks was 1.9. In the same way, the sialic acid peak and the
sum of the four GalNAc peaks relative to the glucose peaks were 1.0 and 0.8,
respectively. The total area of the fatty acid peaks was 1.36 relative to glucose and
the total of the sphingoid base peaks was 0.31 relative to glucose. Thus, there was
good agreement in the molar ratios of galactose, sialic acid, and glucose. Low

galactose/glucose ratios, described previously (9, 14), were believed to have been due

94

to side reactions that result when complex methods are used for HCl removal. The
use of peracetylation instead of trimethylsilylation allows galactose and N-
acetylgalactosamine to be run together without overlap (free amine groups are, by
nature of the reaction condition, converted to N-acetyl derivatives). The ratio of
GalNAc peaks to glucose was somewhat low (0.81) but still in the stoichiometric
range with glucose. The slightly low ratio could be a result of the multiplicity of
peaks. There was no evidence for partially methanolyzed disaccharides containing N-
acetylgalactosamine. Quantititation could be based on the sum of the peaks or a ratio
with just the largest GalNAc peak being used for calculation. The four-peak profile
has been shown to be useful in confirming the presence of GalNAc over other
aminosugars such as N-acetylglucosamine, which has only two major peaks.

The fatty acid ratio was slightly high, which we attribute at least partially to
the fact that the fatty acids are more reduced than sugars and consequently have a
higher molar response in the flame ionization detector. The sphingoid bases had the
lowest response of all, but the peaks were useful for qualitative and quantitative
purposes. We assume the low response is due to side reactions (16) and
decomposition.

Certain fatty acids interfere with certain sugars. That is, palmitate and
galactose, unsaturated C18 fatty acids and N-acetylgalactosamine, and C24 fatty acids
and the sphingoid bases overlap. This problem can be easily avoided by removal of
the fatty acid methyl esters with hexane after methanolysis and prior to the addition of
the internal standard. This modification has no effect on the recovery of the sugar

moieties. The fatty acid fraction can be analyzed separately, if necessary.

95

Time and Temperature Optimization of Methanolysis Conditions

Aliquots of GM, were placed in the reaction tube with the methanolic HCl and
methyl acetate, sealed, and subjected to temperatures of 80, 110, and 150 °C for
times ranging from 0 to 24 h. Figure 2 shows the time course for methanolysis at
110°C. It is apparent that the release of all sugars is complete after only 0.5 h while
methanolysis of the ceramide is complete after about 2 h. At 80°C, total release was
achieved after 5 h; at 150°C, release of the sugars was quite rapid but so was
degradation of the peracetylated methyl glycosides (data not shown).

Standard conditions in the literature for methanolysis of glycosphingolipids has
generally been 18-24 h at 80 °C. One report (17) described that the need for such
long methanolysis times was due to the difficulty in breaking the glucose-ceramide
bond. Our data show that in the sealed capillary tubes methanolysis of the glucose-to-
ceramide bond was complete after only 0.5 h and that the limiting reaction was the
release of the fatty acid from the sphingoid base. A possible explanation of the
decreased time may be that in the sealed tube, the relatively small volume of the tube
and head space after sealing maintains the HCl concentration in the methanol during

the reaction.

Stability of Peracetylated Methanolysis Products.

To assess the stability of the peracetylated methanolysis products, samples of
GM, were derivatized and run at increasing times following peracetylation. We found
that the GC areas of the representative peaks were consistent over a period of one

week. It was also determined that peracetylation was complete after only 2 h (data

96

Figure 2. Methanolysis time course of GM,. Gal: galactose; Glc: glucose; FA:
fatty acid; NeuAc: sialic acid; GalNAc: N-acetylgalactosamine; LCB: sphingoid base.

Figure 2.

nmoles

97

 

 

1.6

1.4-

1.2‘

 

Gal

 

 

 

 

LCB

 

0.5

dd

1.5

”—1

2:5
Time (11)

03-4

24

98

not shown)

Glycoconjugates Containing F ucose and N-Acetylglucosamine

The method was evaluated for use with compounds containing fucose and/or
N-acetylglucosamine. Because pure samples of appropriate glycosphingolipids were
not available to us, commercial samples of fucosylal-Zlactose (2’-FL) and lacto-N-
fucopentaose II (LNFP) were used for this study. Stock solutions of 2’-FL (2 mg/ml)
and LNFP (1 mg/ml) were prepared in water. Aliquots of 2’FL (2.5 pl; 10.3 nmol)
and LNFP ( 10p]; 11.7 nmol) were analyzed in triplicate after lyophilization in the
standard methanolysis tubes. Methanolysis was at 110°C for 2 h.

Although pure gangliosides were not available to us, a very small amount of
an equimolar mixture of ganglioside 6C (NeuAca2-6GalBl-4GlcNAcBl-4GalBl-
4(Fucal-3)GlcNAcB 1-3GalBl-4G10Cer) and Sialyldi-Y-2 (NeuAca2-6GalBl-4(Fuca1-
3)GlcNAcBl-4GalB 1-4(Fucal-3)GlcNAcBl-3GalB 1-4GlcCer) was subjected to the
standard procedure.

Table 2 shows the results of the carbohydrate analysis of 2’ FL and LNFP as
well as authentic GM, Ganglioside. Table 3 shows the carbohydrate analysis of the
sample containing equimolar amounts of 6C and sialyldi-Y-2 gangliosides. The
recovery 05-80%) of the fucose moiety was always below that expected. This may
be attributed in part to the increased volatility of fucose (one less OH group than
glucose) as well as, perhaps, to some degree of degradation. There recoveries and
ratios of glucose and galactose were very close to expected except for cases in which

an acetylated aminosugar was linked to a galactose residue. The de-N-acetylation of

99

TABLE 2.

 

Carbohydrate Analysis of Reference Oligosaccharides and GM,

 

 

Compound Component Actual Recovery % Yield Ratio to
(nmol) (nmol) Glucose
2’FL Fuc 10.3 8.3 81 0.82
Gal 10.3 10.2 99 1.0
Glc 10.3 10.1 98 --
LNFP Fuc 11.7 8.8 77 0.77
Gal 23.4 21.2 91 1.81
Glc 11.7 11.7 100 --
GlcNac 11.7 9.5 81 0.81
GM, Gal 1.62 1.50 93 1.90
Glc 0.81 0.79 98 --
GalNAc 0.81 0.62 75 78
NeuAc 0.81 0.79 98 1.00

100

TABLE 3.

Carbohydrate Analysis of an Equimolar Mixture of 6C and Sialyldi
Y-2 Gangliosides Plasma GM, Ganglioside

 

 

Ganglioside Component Ratio to Theoretical Relative
Glucose Ratio to Yield
Glucose

6C and Sialyldi-Y-2 Fuc 1.13 1.5 0.75
Gal 3.13 3 1.0
Glc l -- --
GlcNAc 2.06 2 1.0
NeuAc 1.0 l 1.0

GM, Gal 1.04 l 1.0
Glc 1.0 -- --
NeuAc 1.0 1 1.0
Fatty Acid 1.75 1 1.75
Sphingoid 0.43 1 0.43
Base

101

the aminosugar during methanolysis makes its glycosidic bond more resistant to
cleavage by methanolysis, hence a decrease in both the acetylated aminosugar and
adjoining galactose. Moreover, it appears form our limited data that a de-N-
acetylated galactosamine linkage may be more resistant to cleavage than a de-N-
acetylated glucosamine linkage. The detection of disaccharides containing hexosamine
and galactose were not found; however, such compounds may have retention times
that were greater than the time span of our GC run. This problem was not seen in
the analysis of the mixture of 6C and sialyldi-Y-2 gangliosides, which gave
stoichiometric ratios (Table 3). The problem of varying recoveries can be addressed
by the co-running of appropriate standards, using the same experimental conditions,
and determining a conversion factor for each component, as should be done in any
case. With these considerations, we believe that a wide variety of glycolipids and

other glycoconjugate can be analyzed by this procedure.

Component Analysis of GM, Ganglioside isolated from human plasma.

GM, ganglioside was isolated from human plasma as described. Following
methanolysis, the fatty acid methyl esters were removed by partitioning twice with
hexane in the reaction tube. This was achieved with the use of a 10 pl Hamilton
syringe. The upper phase was pulled into the barrel and expelled into the lower
phase. After brief centrifugation to separate the layers, the upper hexane phase was
removed. Methyl heptadecanoate (2.1 nmol in 5 pl) was added to the remaining
lower phase as the internal standard, followed by peracetylation (5 pl pyridine:acetic

anhydride, 1:1) for 2 h.

102

Gas chromatography and subsequent analysis of the peak areas from the
plasma GM, gave the component ratios shown in Table 3. The observed ratios of the
sugars were in good agreement with the composition of GM,. The fatty acid ratio to
glucose was higher than expected, beyond that explained by a higher detector
response. This was perhaps due to slight phospholipid contamination. The types and
relative abundance of fatty acids found in GM, in our sample agreed with reported
data (18, 19), with palmitate and stearate (Cl6:0 and C18:0) being the most abundant
and tetracosanoate and tetracosenoate (C24:0 and C24: 1) being the next most

abundant components.

Qanclasian

Gas chromatography of neutral glycosphingolipids and gangliosides is often the
first step in understanding the structures of these complex glycolipids. The
Oligosaccharide portion of the molecule has always drawn the most attention.
Alterations in the structures of the Oligosaccharide moieties with oncogenic
transformation has been known for some time, as well as the antigenic specificity of
this part of the molecule (1, 20). More recent studies have focused on the fatty acid
and sphingoid base moieties and their importance in transmembrane signaling events.
Spatial locations dictated by the ceramide moiety may determine how the course of
enzymatic glycosylations (21). This is suggested by the fact that certain
Oligosaccharides tend to be associated with specific fatty acid mixtures. Furthermore,
the presence of or-hydroxy fatty acids may also have an effect on the glycosylation

pathway (20) and thus define certain facets of the neoplastic state. The need for a

103

simple, rapid, and comprehensive method for the analysis of the many different
components of glycolipids is evident. We believe the method proposed herein will

greatly facilitate the analysis of these interesting biomolecules.

104

ACKNOWLEDGEMENTS

We thank Dr. Richard L. Anderson for his helpful suggestions.

105

FOOTNOTES

‘ This work supported by NIH Grant #DK12434.

2 To whom correspondence should be sent.

3 Abbreviations used: GSL, glycosphingolipid; GC, gas chromatography; GM,,
NeuAca2-3GalB 1-4Gchl-1Cer; HDA, heptadecanoic acid; GM,, GalBl-4GalNAcBl-
4(NeuAca2-3)Galfi1-4Gchl-1Cer; 6C, NeuAca2-6GalBl-4GlcNAcB1-3Ga161-
4(Fuca1-3)G1cNAcBl-3Gall31-4Glc-Cer; Sialyldi Y-2, NeuAcaZ-6GalBl-4(Fucal-
3)GlcNAcB 1-3GaIB l-4(Fuca 1-3)GlcNAcB 1-3GaIB 1-4Glc-Cer; 2 ’ -Fucosyllactose;
(2’FL) Fucal-2GalBl-4Glc, lacto-N-fucopentaose II, (LNFP) GalBl-3(Fucal-
4)GlcNAcBl-3Gal,81-4Glc, Glc, glucose; Gal, galactose; GalNAc, N-
acetylgalactosamine; GlcNAc, N-acetylglucosamine; Cer, ceramide; LCB, long-chain

base; HPTLC, high-performance thin-layer chromatography.

10.

11.

12.

13.

14.

REFERENCES.

Hakomori, S. ( 1981) Ann. Rev. Biochem. 50, 733-764.

Stults, C. L. M., Sweeley, C. C., and Macher, B. (1989) Methods Enzymol.
179, 167-214.

Hakomori, S. (1983) in Handbook of Lipid Research, 3: Sphingolipid
Biochemistry (Kanfer, J., and Hakomori, S., Eds.), Vol. 1, pp. 1—150, Plenum
Press, New York.

Yu, R. K., and Ledeen, R. W. (1970) J. Lipid Res. 11, 506-516.
Niedermeier, W., and Tomana, M. (1974) Anal. Biochem. 57, 363-368.
Honda, S., Suzuki, S., Kakehi, K., Honda, A., and Takai, T. (1981) J.
Chromatog. 226, 34-345.

Yang, H.-J., and Hakomori, S. (1971) J. Biol. Chem. 246, 1192-1200.
Chambers, R. E., and Clamp, J. R. (1971) Biochem. J. 125, 1009-1018.
Laine, R. A., Esselman, W. J., and Sweeley, C. C. (1972) Methods Enzymol.
28, 159-167.

Rickert, S. J ., and Sweeley, C. C. (1978) J. Chromatog. 147, 317-326.
Conchie, J. (1976) in Methods in Carbohydrate Chemistry (Whistler, R. L.,
and BeMiller, J. N., Eds), Vol. 7, pp. 195-199, Academic Press, New York.
Levvy, G. A., Hay, A. J ., Conchie, J., and Strachan, I. (1970) Biochim.
Biophys. Acta 222, 333-338.

Chaplin, M. F. (1982) Anal. Biochem. 123, 336—341.

Ledeen, R. W. and Yu, R. K. (1982) Methods Enzymol. 38, 139-191.

106

15.

16.

l7.

l8.

19.

20.

21.

107

Vance, D. E., and Sweeley, C. C. (1967) J. Lipid Res. 8, 621-630.

Gaver, R. C., and Sweeley, C. C. (1966) J. Am. Chem. Soc. 88, 3643-3647.
Shimamura, M., Oku, M., and Yamagata, T. (1991) FEBS Lett. 290, 213-
215.

Kundu, S. K., Diego, S. O., and Marcus, D. M. (1985) Arch. Biochem.
Biophys. 238, 388-400.

Tao, R. V. P., and Sweeley, C. C. (1970) Biochim. Biophys. Acta 218, 372-
375.

Ladisch, S., Sweeley, C. C., Becker, H., and Gage, D. (1989) J. Biol. Chem.
264, 12097-12105.

Kannagi, R., Nudelman, E., and Hakomori, S. (1982) Proc. Natl. Acad. Sci.

USA 79, 3470-3474.

CHAPTER 5
Potential Methods for Characterization of Gangliosides

The use of high-performance thin-layer chromatography and gas
chromatography are well suited for the analysis of gangliosides isolated from
biological tissues and fluids and they have been very useful in our studies of
gangliosides from sera. However, these methods were not the only methods
considered or tried. Other methods attempted were Oligosaccharide analysis using
high-pH anion-exchange high-performance liquid chromatography with pulsed
amperometric detection (HPAE HPLC-PAD) and radioisotope labeling and
autoradiography. In this chapter, these methods will be discussed briefly along with

the work done and the observations made.

Oligosaccharide Analysis with HPAE HPLC-PAD '

High pH anion-exchange high-performance liquid chromatography (HPAE
HPLC) was developed for the analysis of mono- and Oligosaccharides (1-6). It is an
aqueous based HPLC system run in buffers of very high pH. Carbohydrates form
oxyanion species at a pH of about 13, and are therefore separatable by anion-
exchange chromatography used in this system. Pulsed amperometric detection (PAD)
can successfully detect carbohydrates in the picomolar region. The alcohols of
carbohydrates are easily detected by the electrochemical oxidation, with free radical

intermediates adsorbing to the anode surface, reducing electrode current. Deposits

108

109

are removed by oxidative desorption with a pulse of high positive potential. Mono-
and Oligosaccharide analysis could be completed in under 2 h with identification done
by the comparison of retention behavior with that of known standards. Ganglioside
standards as well as complex mixtures of gangliosides could be analyzed in this way
following the release of the Oligosaccharides from their ceramide tails either
chemically or enzymatically. The Oligosaccharide mixtures could then be separated
by HPAE-HPLC, identified and quantitated with the PAD, giving information on the

Oligosaccharides of the ganglioside mixture.

Experimental

Oligosaccharides from authentic gangliosides or mixtures of gangliosides
isolated from plasma or serum were cleaved from their ceramide tails in one of
several ways. One method explored was an enzymatic procedure that used
endoglycoceramidase from Rhodococcus sp. (7, 8) or ceramide glycanase from
earthworm (9) or leech (10) to cleave the intact Oligosaccharide moiety from
ceranride. Another method made use of ozonolysis of the olefinic double bond of
sphingosine and alkali-induced degradation of the resulting aldehyde. If
monosaccharides were desired, hydrolysis in trifluoroacetic acid was performed.
After a brief clean-up procedure, the sample was injected into a Dionex Bio-LC fitted
with a CarboPac PA-l column and a Model PAD 2 detector. Samples were eluted in
a buffer of 0.1 N sodium hydroxide and a gradient of 0 to 150 mmol of sodium
acetate. Peak identifications were made by comparison with the retention times of

authentic standards as well as reported values. Quantitation was based on the internal

110

standard, mannose.

Discussion

High pH anion exchange high-performance liquid chromatography coupled
with pulsed amperometric detection has successfully been used to analyze mono- and
Oligosaccharides derived from glycoproteins (1-4, 6). The ability to resolve isomers
differing only in anomeric configuration and/or linkages is useful in the structural
analysis of many diverse Oligosaccharides. However, there were some difficulties
when this method was applied to the quantitative and qualitative analysis of
gangliosides mixtures isolated from plasma.

A necessary first step for analysis of gangliosides is the release of the
Oligosaccharide moiety from the ceramide tail. Aside from the expected problems
associated with clean-up, the enzymatic procedures suffered from a lack of specificity.
It has been shown that the specificity of these enzymes can vary towards different
ganglioside Oligosaccharide structures (7) . Gangliosides with larger Oligosaccharide
moieties were not as rapidly hydrolyzed as their less complex counterparts.
Therefore, it would be necessary to determine how the enzyme would behave towards
a diverse mixture of gangliosides such as found in plasma or serum. Chemical
release of the carbohydrate head group using ozonolysis would only be useful for
species containing sphingosine in their ceramide moiety as the double bond at carbon
4 is necessary for the reaction to proceed. The yield of the ozonolysis reaction would
have to be determined and considered.

The use of the HPAE HPLC-PAD system has great potential for structural

111

analysis of Oligosaccharides. However, the day to day fluctuations in the performance
of the instrument made quantitative and qualitative analysis difficult. Even with
extreme care, retention times would vary unpredictably from run to run and more so
day to day. Identification was based on retention indices, requiring daily running of
retention standards to determine the current retention behavior and detector response.
Quantitation on this system would be difficult due to each Oligosaccharide having its
own response factor, requiring several standard curves for estimations of
Oligosaccharide concentrations in a mixture. Analysis of the ceramide moieties would
either have to be omitted or determined by another means. This method is clearly
useful for semiqualitative and structural analysis of pure Oligosaccharides and
glycoconjugates, however, it becomes complicated when applied to the qualitative and

quantitative analysis of complex mixtures of gangliosides.

Radioisotope Labeling, HPTLC, and Autoradiography

Costello et al: (1 l, 12) found that the use of microscale reactions carried out
in the vapor phase for the analysis of glycosphingolipids increased the sensitivity
when analyzed with fast atom bombardment mass spectrometry. Specifically, the
reduction of amides to amines in the sphingosine moiety gave a much higher
molecular ion signal. The first step in the reduction was the glycosphingolipid being
reduced with borane vapor (3H,). Next, solvolysis of the borane complexes was
done in methanolic HCl vapor. Finally, oxidation of the organoborane was achieved
with ammonia/hydrogen peroxide vapor. These reactions reduce amides to amines as

well as reducing any olefinic bonds. It was observed that this reaction could be used

112

to label gangliosides with tritium by the reduction of amides, as well as any olefinic
bonds, with tritiated borane. The labeled gangliosides could then be analyzed with
HPTLC and autoradiography. This method had the potential of being extremely

sensitive and able to detect minute alterations in ganglioside profiles.

Experimental

Gangliosides disolved in chloroform-methanol (1:1) were coated onto the
inside of a capillary tube and dried in a vacuum. The ganglioside-containing capillary
tube was placed in a special apparatus designed for vapor phase reactions (13). The
apparatus was evacuated and the sample was reduced with the vapor of 1N borane
(3H3, BD3, or 8T3) in tetrahydrofuran under reduced pressure for 2 h at 45 OC. The
resulting borane complexes were solvolyzed in the apparatus in the same manner as
above in methanolic HCl (0.5 M) replacing the borane/tetrahydrofuran for 3 h at
room temperature. The methanolic HCl was evaporated under a vacuum and replaced
with of 30 % hydrogen peroxide and 3 N NH, to oxidize the samples for 1 hour at 45
°C. These reactions were all conducted in the same apparatus with the excess

reagents removed by evacuation before the next one is added.

Discussion

This method was under consideration for its ability to label minute mixtures of
gangliosides without dramatic alteration of the structure. This would allow for a
increase in the level of sensitivity of gangliosides detectable by conventional means

far below the 50 pmol level. This is desirable in order to detect very minor

113

gangliosides that might be of interest in biological tissues and fluids. While this
method could be potentially very useful in these studies, there were some important
issues to be addressed.

The first and the most serious problem was in the protocol. In order for the
reaction to proceed with a suitable yield, extremely high specific activities of tritiated
borane gas would have to be used. In order to safely use a highly radioactive gas,
already potentially harmful and reactive, would require an elaborate gas-train
assembly to contain the reaction and the reagents. Several different designs were
considered but none was found that satisfactorily dealt with the problem without
complicating it.

As was stated, hydroboration and oxidation will reduce amides to amine and
double bonds to single bonds. The original procedure was designed to be applied to
very simple, pure glycosphingolipids, such as galactosylceramide, which may have
one amine and possibly two double bonds, to be analyzed by mass spectrometry. The
degree of unsaturation was not a concern. Mixtures of gangliosides isolated from
biological sources most likely will have much more complex gangliosides containing
one or more N-acetylated hexosamines. Furthermore, there is the possibility of
heterogeneity in the ceramide moiety in both the sphingoid base and fatty acid, which
may have some degree of unsaturatation. Since any amide bond or olefinic bond is
capable of being reduced and therefore labeled, it is possible that GT“, could have
anywhere form 6 to 11 different sites for the reduction reaction while GM3 might
have only 2. Each tritium atom contributes to the detection signal and this could

make efforts at quantitation difficult. Moreover, each little change in the structure,

114

albeit small, will give the ganglioside different migration properties on thin-layer. In
essence, it would make an already highly complex mixture of gangliosides even more

complicated. It was for these reasons that this method was not pursued.

10.

11.

12.

13.

REFERENCES
Townsend, R. R., Hardy, M. R., Cumming, D. A., Carver, J. P., and
Bendiak, B. (1989) Anal. Biochem. 182, 1—8

Townsend, R. R., Hardy, M. R., Hindsgaul, O., and Lee, Y. C. (1988) Anal.
Biochem. 174, 459-470

Hardy, M. R., and Townsend, R. R. (1988) Proc. Natl. Acad. Sci. USA 85,
3289-3293

Hardy, M. R., Townsend, R. R., and Lee, Y. C. (1988) Anal. Biochem. 170,
54-62

Wantanabe, N. (1985) J. Chromatog. 330, 333-338
Lee, Y. C. (1990) Anal. Biochem. 189, 151-162
Ito, M., and Yamagata, T. (1986) J. Biol. Chem. 261, 14278-14282

Michalski, Jean-C., Corfield, A. P., and Schauer, R. (1986) Hoppe-Seyler’s
Z. Physiol. Chem. 367, 715

Li, Yu-T., Ishikawa, Y., and Li, Su-C. (1987) Biochem. Biophys. Res. Comm.
149, 167-172

Li, Su-C., DeGasperi, R., Muldrey, J. E., and Li, Y. (1986) Biochem.
Biophys. Res. Comm. 141, 346-352

Costello, C. E., and Vath, J. E. (1990) Methods Enzymol. 193, 738-768
Domon, B., Vath, J. E., and Costello, C. E. (1990) Anal. Biochem. 184, 191

Vath, J. E., Zollinger, M., and Beimann, K. (1988) Fresent'us Z. Anal. Chem.
331, 248-252

115

CHAPTER 6

Analysis of Circulating Gangliosides of Breast Cancer Patients and an
Appropriate Control Group.

D. A. Wiesner and C.C. Sweeley

November 15, 1993

Intl. J. Cancer, (1993) in preparation.

Department of Biochemistry
Michigan State University
East Lansing, MI 48824

Running Title: Serum Gangliosides of Breast Cancer.

116

117

Abstract

Gangliosides were isolated from the sera of recently diagnosed breast cancer
patients and from individuals of an appropriate control group of women. Quantitative
and qualitative analyses were by two-dimensional high-performance thin-layer
chromatography ( 2-D HPTLC) and gas chromatography (GC). The profiles of the
isolated gangliosides were determined by 2-D HPTLC, visualized with resorcinol, and
quantified by digital image densitometry. The profile of the cancer patients was
compared to that of the control, revealing an overall increase in total lipid-bound
sialic acid and a specific increase in polysialogangliosides. Furthermore, an increase
was noted in the b-series biosynthetic pathway over the a-series in the cancer sera, as
compared to the controls. Gas chromatographic analysis of the peracetylated
methanolysis mixtures derived from the total ganglioside fraction of cancer patients
supported the HPTLC data, with an increase in the total sialic acid and galactose
residues. No unusual gangliosides were found in either the normal or breast cancer
gangliosides profiles. The observed increase in the polysialogangliosides, specifically
the b-series gangliosides, is consistent with the idea that they may have a role in the

metastasis of breast cancer.

118

Introduction

A recent trend in the research of many cancer types is the analysis of
glycosphingolipids associated with the cancers. Oncogenic transformation has long
been associated with an alteration in the composition of glycosphingolipids, both of
the neutral type and gangliosides, in the plasma membranes of cancer cells (1-3).
Numerous studies have indicated both direct and indirect roles of gangliosides in
tumorigenesis (4-7). With the development of monoclonal antibodies against tumor-
associated antigens, the presence of complex, more highly glycosylated or sialylated
ganglioside tumor markers became more apparent. Magnani et al. (8) identified the
first monoclonal antibody-defined ganglioside of colon cancer and a large number of
antibodies to tumor-associated ganglioside antigens have subsequently been identified
(Reviewed in 9, 10). Furthermore, gangliosides have been shown to be shed from
these cancer cells into the local environment and eventually into the blood stream (11-
14). Some researchers have tried to take advantage of this phenomenon by
identifying the presence of unusual or increased levels of gangliosides in the plasma
of cancer patients as a circulating tumor marker (12, 15, 16). Circulating tumor-
associated antigens as well as tumor-associated tissue gangliosides have been
examined for their use in antibody therapy (17, 18). A circulating tumor marker of
this type may also be useful in the diagnosis and monitoring of breast cancer as well
as potentially in the treatment of cancer. In this study, we examined the ganglioside
content of several plasma and serum samples obtained from women suffering from

ductal cell carcinoma of the breast and an appr0priate control group to ascertain

119

whether such a circulating tumor marker may be present as well as to describe the

nature of the ganglioside profile from serum samples of breast cancer patients.

Materials and Methag

Materials

All solvents were of reagent grade or better, while chloroform and methanol
were HPLC grade and pyridine was distilled from a reagent stock before use. N-
acetylneuraminic acid (sialic acid, NeuAc), glucose, galactose, N-acetylglucosamine,
N-acetylgalactosamine and heptadecanoic acid were obtained from Sigma. DEAE-
Sephadex A-25 and Sephadex G-50 were obtained from Pharmacia (Uppsala,
Sweden). High-performance thin-layer chromatography plates (HPTLC) were
obtained from Merck (Darmstedt, Germany). Reference gangliosides were obtained
from Sigma (St. Louis, MO) unless otherwise noted. Ganglioside GM,, was isolated
from chicken liver according to the method of Shiraishi and Uda (5). Gangliosides
GD, GM,,), GT”, and GT2 were a gift from Dr. R.K. Yu of the Medical College of
Virginia, Virginia Commonwealth University. Gangliosides 6C, sialyldi-Y-2, sialyl-
Le" (SI.e"), sialyl-Lea (SLea), 2,6-sialylparagloboside (SPG),
sialylnorhexaosylceramide, and sialyl I were a gift from Dr. E. Nudelman of the
Biomembrane Institute.

Resorcinol spray was made according to Svennerholm et al. (19). Briefly, 5
ml of a stock solution of resorcinol (2 g of resorcinol recrystallized from benzene in
100 ml of water) is added to 40 ml of concentrated hydrochloric acid containing 0.125

ml of 0.1 M copper sulfate and the total volume brought to 50 ml with distilled,

120

deionized water. This reagent was stable at 4 °C until it turns green. The stock
solution was stable for many months at 4 °C.

Methanolic HCl, was prepared by bubbling gaseous HCl through methanol,
then titrating the acidic methanol with base to obtain the normality. This concentrated
stock solution was stored at -20 °C and diluted with methanol to 0.75 N as necessary.
The stock solution was titrated periodically to determine any loss of HCl. The stock
reagent was stable for up to one month.

Serum samples were obtained from women recently diagnosed with ductal cell
carcinoma of the breast. All patients were in the early stages and had not received
any treatment yet. Serum was obtained by allowing the patient blood to clot and
immediately separating the serum from the red cells by centrifugation . The serum
was stored at -70 °C prior to use. Plasma samples for the control group were
obtained from women 22 to 63 years of age who were not currently menstruating.
The blood was drawn after an overnight fast and treated with citrate to avoid clotting.
The samples were also immediately separated and the plasma stored at - 70 °C prior

10 1186.

Isolation of Gangliosides from Serum and Plasma.

A modified method of Ladisch et al. (20) was used for the extraction of
gangliosides from serum and plasma. Briefly, 1 ml of serum or plasma was extracted
twice with 10 ml of chloroform:methanol (CM, 1: l), with the gangliosides collected
in the supernatant and thoroughly dried. The residue was taken up in 2 ml of

diisopropyl ether:n-butanol (7:3) and 1 ml of double de-ionized water. After

121

sufficient vortexing and sonication, the sample was centrifuged and the phospholipid-
containing upper organic layer removed with a 9 in. pasteur pipet. The ganglioside-
containing aqueous lower layer was reextracted with fresh organic solvent. After
removal of the second organic layer, the aqueous layer was freeze-dried. Any
remaining phospholipids were destroyed by mild alkaline methanolysis in 0.1 N
NaOH in methanol for 2 h at 37 °C. The mixture was neutralized with acetic acid
and quickly dried under a stream of nitrogen. A small Sephadex G-50 column was
used to remove low molecular weight impurities and salts. Samples to be analyzed by
gas-chromatography (GC) were further purified by DEAE-Sephadex A-25 column

chromatography and again desalted on a Sephadex G-50 column.

Two-dimensional High-Perfonnance Thin-Layer Chromatography (2-D HPTLC) of
Gangliosides.

Reproducible two-dimensional thin-layer chromatography on silica gel was
achieved by running multiple plates in separate, identical, wicked chambers containing
like amounts of solvent, equilibrated for the same period, and run for the same
amount of time at the same time. This type of control over external variables, and
the solvent system of chloroform:methanol:0.2% CaC12 (50:40:10) in the first
direction and n-propanol:11.3 N NH3(aq.) in the second direction (21) provided very
useful and reproducible 2-D chromatography. A two dimensional reference map was
created of known standards and tumor-associated gangliosides for comparison to the
plasma and serum 2-D ganglioside profiles (Fig 1a). Gangliosides from 1 ml of

plasma or serum were spotted, along with GM4 (1.5 nmol) as an internal standard, on

122

a 10 cm x 10 cm HPTLC plate near the comer and run in two-dimensions.
Quantitation was performed on a Bio Image Visage 110 image densitometer and was
relative to the GM4 internal standard. Identification of gangliosides found in the
serum was based on co-migration with authentic standards as well as HPTLC data

found in the literature concerning serum gangliosides (22-27).

Gas Chromatography of Carbohydrate and Lipid Components of Gangliosides.
Ganglioside mixtures were derivatized for analysis by gas chromatography by
the method of Wiesner and Sweeley (28). Gangliosides isolated from serum or
plasma were placed in 2.5 mm x 4 cm capillary tubes and dried in vacuo.
Hydrochloric acid (25 pl, 0.75 N) in dry methanol and methyl acetate (5 pl) was
added, the tube sealed, and heated at 110 °C for 24 h. The tube was carefully
cracked open and the solvent quickly evaporated under a strong vacuum. Fatty acid
methyl esters were extracted with hexane and transferred to separate tubes for fatty
acid analysis. Methyl heptadecanate was added as an internal standard. The
methanolysis products were derivatized for GC with 5 pl pyridine:acetic anhydride.
(1:2) for 2 h. Gas chromatography was performed by a Hewlett Packard model
5890A gas chromatograph fitted with a 60 m DB—l column (0.25 mm ID. with a 0.1
mm film thickness). The sample (1-3p) was injected onto the column at 185 °C; after
15 min the column was heated at 4 0C per min to 290 °C and held for 5 min. The
sample was co-injected with a series of even hydrocarbons, C12 to C26 except C22,

for the use in calculating retention indices.

123
Results

For analysis of gangliosides by high-performance thin-layer chromatography,
sera were obtained from 17 women with ductal cell carcinoma of the breast and
plasma from 16 healthy women as a control group. For gas chromatography, 19 and
17 samples, respectively, were obtained. Samples of plasma and serum from the
same donor were obtained, isolated and analyzed for sialic acid content using the
thiobarbituric acid assay. The sialic acid content of the plasma and serum samples
was within 97% of each other. There was no difference found in the ganglioside
content between blood plasma and serum and henceforth will be considered

equivalent.

High-Performance Thin-Layer Chromatography

Fig. la shows a map created of known gangliosides used for identification of
ganglioside spots on the 2-D HPTLC. Figs. 1b and 1c show photographs of the 2-D
HPTLC plates of gangliosides isolated from cancer and control group samples,
respectively. Table 1 list the gangliosides found in serum and their structures. Table
2 shows the quantitation results of the 2-D HPTLC and image densitometry both in
terms of lipid-bound sialic acid and total lipid. We found and quantified 12
resorcinol-positive spots isolated from 1 ml of plasma or serum by 2-D HPTLC. The
spots were primarily identified by co-migration of authentic standards and comparison
with reported plasma gangliosides (22-25). While identification by co-migration is
not conclusive, it is suitable for preliminary analysis of the differences in profiles of

the control and cancer group.

124

The 16 human plasma samples comprising the control group contained a level
of recovery of 6.1 :t 1.5 nmol lipid-bound sialic acid/m1 of serum. These results are
in line with previous reports of between 7 and 20 nmol lipid-bound sialic acid per ml
from a single adult male (29, 30), pooled sera (31), or of unknown origin (22, 32), as
well as the reported average of 11 individuals (33). This report (33) examined 5
female and 6 male and found 9.9 i 3.8 and 11.1 i 2.8 nmol lipid-bound sialic acid
per ml of serum. In all of the reports as well as our findings, the total concentration
of lipid-bound sialic acid varied appreciably among individuals producing significant
standard deviations. In all the above studies, larger quantities of starting material
were used, and often correction factors were used to adjust the recoveries. In our
work, we were limited to only 1 ml of sera or plasma for starting material and made
no adjustment for losses during the isolation. To this difference we attribute our
lower recoveries. However, in all cases, great care was taken to make sure all
isolation steps were identical so that any losses would be reflected evenly in all
samples, including the work-up of both cancer patients and control samples in each
batch. The profiles of gangliosides determined from our control group were similar
to those of previous studies (22, 29, 33). GM, ganglioside accounted for 50-70% of
the total lipid-bound sialic acid, with the next most abundant ganglioside being GD3.

When the levels of gangliosides from sera of the control group are compared
to those of the cancer group, several observations can be made. Table 2 shows a
significant (P<0.05 using student’s t) increase in the total and all individual
ganglioside species from breast cancer serum over those of the control group. This

finding of increased sialic acid content has also been seen in other studies with other

125

Figure 1. HPTLC of Gangliosides. Figure 1a is a schematic diagram of several
gangliosides mapped with two-dimensional high-performance thin-layer
chromatography. Gangliosides found in plasma are shaded while other reference
gangliosides are white. Figure 1b and c are photographs of actual gangliosides
isolated from the serum of breast cancer patients (1b) and the plasma of a control
sample (1c).

 

 

 

 

Figure 1a

    

GT

‘0

 

 

GT1 b
.\
GD‘lb
.— GQ1b
O—ec .
C7-Sdiy-2 Resorcinol Negatlve
Origin
Figure lb Figure 1c

   

126

llll . 1"!

f.

(1. {tall ' ‘. .A

 

127

Table 1. Summary of Ganglioside Nomenclature and Structure. Gangliosides
found in plasma are marked with an asterisk (*).

 

 

 

 

 

128

 

 

Table 1 .
Ganglioside Structure
GM,, NeuAca2-3Galfll-1Cer
precursor
LacCer Gallil-4Gchl-1Cer
a-series
GM, * N euAca2-3GalB l -4GlcCer
GM2“ GalNAcBl-4(NeuAcaZ-3)GalBl-4GlcCer
GM, GalB1-3GalNAcB1-4(NeuAcaZ-3)GalBl-4GlcCer
GD ,.* NeuAca2-3Galfi 1 -3GalNAcB 1 -4(NeuAcc12-3)Gal[3 l -4GlcCer
GT, 1, NeuAca2-8NeuAc012-3Galfil-3Ga1NAcBl -4(NeuAcaZ-3)GalBl -4GlcCer
b-series
GD3* (NeuAcaZ-SNeuAca2-3)GalB l -4GlcCer
GDZ GalNAcfi l 4(NeuAca2-8NeuAcaZ-3)Galfi l -4GlcCer
GD , b" GalB1-3GalNAcfil-4(NeuAcorZ-8NeuAca2-3)Gal[3l -4GlcCer
GT , b "' NeuAca2-3GalBl-3GalNAcBl-4(NeuAca2-8NeuAcaZ-3)Galfi1-4GlcCer
GQ,,,"' NeuAca2-8N euAca2-3GalB 1 -BGa1N A03 1 -4(NeuAca2-8N euAca2-3)GalB 1 -4GlcCer
c-series
GT3 NeuAcaZ-SNeuAcaZ-8NeuAca2-3GalB1-4GlcCer
GT2 GlcNAcB 1 4(NeuAcaZ-8NeuAca2-8NeuAca2-3)GalB l -4GlcCer
GT, c GalBl-3GlcNAcB1-4(NeuAca2-8NeuAca2-8NeuAc012-3)GalB1-4GlcCer
GQ , c NeuAca2-3GalBl-3GlcNAcBl4(NeuAcaZ-8NeuAca2-8NeuAca2-3)GalB1-4GlcCer
lacto series
SPG * NeuAca2-3GalB l -3GlcNAcB l -3GalB 1 -4GlcCer
2-6SPG“ NeuAca2-6Galfi l -3GlcNAcB l -3GalB l-4GlcCer
SLe' NeuAca2-3GalB l -3(Fucal -4)GlcNAcB l -3GalB l -4GlcCer
SLe" NeuAca2-3GalB l -3(Fuca l -3)GlcNAcB l -3Gal[3 l -4GlcCer

neolacto series
SNHC*

Sialyl 1

6C

SdiY-2

NeuAca2-3GalB l 4G1cNAcBl-BGaIBl-4G10NAciil -3GalB l -4GlcCer
(NeuAca2-3GalBI-4GlcNAcfil-3 ,6)G 131 -4GlcNAcB l -3GalB 1-4GlcCer
NeuAca2-6GalB1-4GlcNAcBl-3GalB1-4(Fuca1-3)GlcNAcBl-BGaIB l-4GlcCer
NeuAca2-3(GalB 1 -4(Fuca l-3)G1cNAcB 1 -3)2GalB 1 -4GlcCer

129

.. 80:25.8 .3 82...; .0008 05 0. 02.8088 80.0 V m .1. J 8....» 38 00.8: 2...; 3:08 2 023800 8.0 V 8....

80280: sumo—28.200 0. 80282350038000... 00 2.3:.

.800 2 820 .o 2.3..

...00 ..5 820 2282.858 82.3-. 2 200 2.. 280.200 .800 8282.858 82.3.-.. .o 2... ..
8881.0 0: .00 .08 .

 

 

 

 

8.. 8.8 8.0 ...N8.N 89.0.2

8.. 8... ..N.8 .N... 0808:

...o 8... z... .8... .83

8. 8... 888 8N8. 8.8 8. 8.8N .288 . 88 3.8 .28
. N8 .8 N. 8N 8 :8 ...SN 88 88 £00

8 8... 88 8 N... N N84 ...888 H .8. 8: 280
N .N 88 88 a. 8 SN .38 3. No. 200
8 8. :8 8 8. 8 8NN 888 . 8.. 88. ..00
8. 8N8 4N8 N.N N2. .0 «N8 80. N8. 88 800
. a. N8 a. 8 N ..8. .8. ..8. ..8. 0:28
.v N8 8N N8 8N . 88 S 88 s. 802
. ..8 8 ..8 8 N 88 .N. 8N .N. 402

8 E o: N: o: 8 SN 88. SN 88. 0088-N
v 8: 8NN 8: 8NN .. SN :8 SN :8 008
8 N.N 88. 8 88. 8 8:. :8 8:. ...8 N20
8 ..No. 8N8 ..No. 8N8 N8 3N. .N88 ..NN. .N8 820

2883.8. .08 .82 .08 .82 <83... .08 M2 .08 8%..
.22 2.8.8 880-22.. .220 .23 .22 2.08 828.22.. . .22.. .23 8.82.880
O— ".6 _O.=:OU w— ".6 MQUCNU

 

 

 

80.820 0.8.0 .0008 a 0:8 05.58 .808 00 .0208 028029...» 00203.

N 03 ma.

130

cancers such as retinoblastoma (34) and melanoma (35). An increase in the more
complex gangliosides, i.e. more highly glycosylated and sialylated, was also
observed. Furthermore, the average ratio of the b-series gangliosides to the a-series
gangliosides is increased significantly in the cancer samples over the control. This
suggests an altered biosynthetic pathway of serum gangliosides associated with breast
cancer. Figures 2a, b and c show range plots of the HPTLC data. These figures
show the cancer samples having not only increased averages, but increased ranges and

medians over that of the control group.

Gas chromatography

Figures 3a and b show typical gas Chromatograms of peracetylated
methanolysis products of gangliosides isolated from serum or plasma. Figure 3a
shows a Chromatogram of the sugar and sphingoid base components while Figure 3b
shows the fatty acid species associated with them. Table 3 shows the concentrations
found of each peracetylated methanolysis product. The extra purification step was
necessary to remove the more polar, neutral glycosphingolipids not removed by the
isolation procedure. This extra step decreased the yield, yet much information was
still obtainable such as the conspicuous presence or absence of particular sugar
residues or ceramide components. Again, the high standard deviations reflect the
variability of the ganglioside concentrations between individuals.

When comparing the results from the control group to that of the cancer, the
data show a significant increase in the total lipid-bound sialic acid and galactose

residues. This supports the HPT LC data which show an increase in total gangliosides

131

Figure 2a. Box plots of GM3 and Total Ganglioside Isolated from Cancer Sera
and Control Plasma. Values are based on lipid-bound sialic acid concentrations.
Definition of boxes is as shown on the right side of the figure.

 

 

 

 

pmol: LBSK

Figure 22.

132

GMB and Total Gangliosides
from Serum of Cancer (C)
and Control (N)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ioooo —
‘i'
° 1
...,. , 1 l
O
C c N
GMS Total

 

95th
75th

 

500: (Median)
25th

 

 

. 5th Percentile

Key

133

Figure 2b. Box plot of Minor Gangliosides Isolated from Cancer Sera and
Control Plasma. Values are based on lipid-bound sialic acid concentrations. See
Fig. 2a for box definitions and Table 1 for abbreviations..

 

 

134

050

20

 

 

 

 

 

A—QU dam—U dfihn. 090 22m 005 «.02 0mm 0mm N20
@m
z o z o z o z 0 z 0 Z n. 2 U z o z o z
. — o _ _
.. H o
. o
o o
o In. 0
O O
H ..
0
G.

 

_

 

 

 

O

 

 

 

 

P _

 

 

 

 

bi P

L

 

 

 

 

.3

3.3000 090

Q.

pmonwo

50.5 800.82%:me Ednmm

.0" 2:»...—

com

08.

000—

OOON

v35“. selowd

135

Figure 2c. Box plots of Important Ratios of Gangliosides.

136

Figurelc.

Box Plot of Various Ratios of
Gangliosides in Cancer (C) and
Control (N).

 

10 l T o l T I I l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. ..-—
, .1.“- a
“LTl 8i
l ° ‘1‘ l
0 i . 2’ . i 1’
C N C N C N C N
I A B c D

A—A Series/B Series

' B—GMS/GDS

C—Mono/polysialyl gangliosides
D—GMZ/SPG 1

137

as well as an increase in the more complex gangliosides. The more complex
gangliosides have more sialic acid and galactose residues per molecule than the more
simple ones. It was interesting to note that sphingosine was the only long-chain base
found in either the cancer or the control group. Yu and Ledeen (29) found
sphingosine as the major component with traces of the C16 and C20 homologs in their
analysis of plasma gangliosides of a single male.

The fatty acid composition of the total ganglioside fraction (Table 4) was
similar to that reported by Yu and Ledeen (29) and Tao and Sweeley (36) with the
major fatty acids being palmitate (C16:O) and stearate (C1810) with the longer chain
fatty acids (C20:O—24:0) having a significant contribution. The fatty acid profile
determined by Kundu et al. (22) was somewhat different, identifying palmitate and
palmitoleate (C16:O and C1621), and stearate and oleate (C18:0 and C18:1) as the
major species with only trace amounts of the longer-chain species. Of particular
interest was our finding of a significant amount of (up to 7%) a-hydroxy fatty acids.
Kundu et al. (22) reported only traces and Yu and Ledeen (29) and Tao and Sweeley
(36) did not find any such compounds. The a—hydroxy fatty acids along with other
fatty acids were identified by gas chromatography-mass spectrometry as fatty acid
methyl esters. A mass spectrum of a-acetoxytetracosanoic methyl ester, from the
FAME fraction from derivatized gangliosides isolated from control plasma, is shown
in Figure 5.

By comparison of the fatty acid data of the control profile with the cancer, no
significant increase could be seen in any singular fatty acid species, yet there appears

to be a general increase in the ratio of long-chain fatty acid (22-24 carbon) to

138

Figure 3. Gas Chromatograms of Peracetylated Methanolysis Products of
Gangliosides Isolated from Plasma of Serum. Figure 3a is a typical gas
Chromatogram of the peracetylated methanolysis products of the sugar residues and
long-chain base of gangliosides isolated from plasma or serum. Figure 3b is the gas
Chromatogram of the fatty acid methyl esters of gangliosides isolated from plasma or
serum.

 

139

 

.3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. ..8 2...»...

 

~63

 

 

 

 

 

 

 

 

 

 

 

232

am 9:65

Table 3.

Averages of GC Analysis Peracetylated Methanolysis Products of Plasma Ganglioside
Mixture:Carbohydrates and long-chain bases.‘ll

 

Component Cancer n= 19 Control n= 17
Galactose 5164 :1; 1815* 3745 i 1477
Glucose 6676 i 2047 7143 i 3168
GalNAc 635 i 338 641 i» 352
GlcNAc 349 i 128 319 i 167
Unknown 630 j: 296 532 i 341
NeuAc 4650 j; 1862* 3899 i 1338
d18:1 4573 i 2367* 2803 i 1283
GlN/Gch 1.85 i086 2.17 i 0.94
Gal/Glcc 0.78 j; 0.20* 0.60 :9; 0.26
I-Iex/I-IexNAcd 7.30 2}; 1.86 8.88 i 6.03

 

‘pmole per m1 of serum.

bRatio of N-acetylgalactosamine to N-acetylglucosamine.
°Ratio of galactose to glucose.
dRatio of galactose and glucose to N-acetylgalactosamine and N-acetylglucosamine
*P < 0.05 compared to the control value using students t.

141

Table 4.

Averages of GC Analysis Peracetylated Methanolysis Products of Plasma
Ganglioside Mixture: Fatty acids

 

 

Component Cancer n= 193 % of Control n=17al % of

Total Total
14:0b 286 i 142 1.7 229 :t 129 1.8
15:0 182 i 85 1.1 165 i 95 1.3
16:1 263 i 198 1.6 266 i 246 2.0
16:0 4191 i 1919 26.0 3879 i 1325 30.0
18:2 395 i 317 2.4 419 i 352 3.3
18:1 609 i 401 3.7 530 j: 388 4.1
18:1 217i 103 1.3 171 i 99 1.3
18:0 2952 i 1494 18.1 2332 i 1091 18.1
20:0 465 i 275 2.8 353 i- 219 2.7
unk 1264 i 679 7.7 990 :t 361 7.7
22:0 1597 i 991 9.7 1094 i 566 8.5
23:0 616 j: 365 3.8 495 j: 277 3.8
24:1 1213 j: 856 7.4 940 j; 638 7.3
24:0 1242 j: 750 7.5 898 j; 435 7.0
h16:0c 162 i 113 1.0 121 i 131 0.9
h22:0c 192 j; 116 1.2 157 i 88 1.2
h23:0° 112 i 75 0.7 80 :t 46 0.6
h24:1° 186 1163 1.1 151 i 69 1.1
h24:0c 272 j; 200 1.7 247 j; 181 1.9
Total Fatty Acid 16415 j; 6693 12877 _-l_- 5674

 

$11101 per ml serum.

bNumber of carbon atomsznumber of double bands

cor-hydroxy fatty acids

J

Figure 4. Mass Spectrum of a-acetoxytetracosanoic methyl ester analyzed by
GC/MS. Taken from the fatty acid methyl ester of gangliosides isolated from
normal plasma.

 

 

143

Saw

:3. 59.2

mm:

 

144

short-chain fatty acid (14-20 carbon) moieties in cancer-associated gangliosides over
that of the control. Similarly, the ratio of stearate (18 carbon) to palmitate (16
carbon) has decreased in the cancer patients. It should be noted that these were
general trends and were not statistically significant. The results, however, along with
other results (37, 38), are consistent with the idea that the ceramide portion of the

ganglioside may have a greater role in carcinogenesis than previously thought.

Digg' ussion

Since the first observation of changes in glycosphingolipid composition in
tumor cells, in 1968-69 (39, 40), extensive studies have been undertaken to elucidate
the function of gangliosides in the biology of cancer. From these studies, several
tumor-associated ganglioside antigens (see Chapter 1, Table 1) have been identified,
primarily with the use of monoclonal antibody technology. After Black (41) noted
that several substances were shed from the surfaces of cancer cells into the local
environment and eventually in to the plasma, a logical step was to determine if these
tumor-associated ganglioside antigens were also found in the plasma and what their
possible function might be. This has been studied most extensively in neuroblastoma
and melanoma.

In the case of neuroblastoma, high levels of GDZ were found in the tissue and
in the plasma (11, 12, 15, 16). The levels of GD2 are typically below the detection
limit (<25 pmol) in serum of control samples. Patients with higher amounts of GD2
at the time of diagnosis had poorer prognosis than those who had lower levels (15).

Furthermore, levels of GDZ in the blood correlated well with the stage and progress

145

of the disease, the levels dropping off after successful treatment and increasing prior
to relapse (15). The authors have suggested the usefulness of GDZ as a diagnostic
tool as well as a tumor marker for monitoring of the disease. GD3, the precursor of
GDZ, has been found in relatively large amounts in the tissue and cells cultured from
human melanoma (17, 35 , 42, 43). While human melanoma cells in culture have
been shown to shed gangliosides, specifically GD3, into the medium, elevated levels
of the gangliosides could not be detected in the plasma or serum of patients (35). The
authors suggested that this was a local phenomenon masked by the already high levels
of 6D, found in control plasma.

There has been less extensive work done on the gangliosides in tissue and sera
of patients with breast cancer. Some of the first work on breast cancer was reported
about twenty years ago by Morré et al. (44). They found that the ganglioside profile
from rat mammary tumors induced by 7,l2-dimethylbenz[a]anthracene was depressed
in disialogangliosides with an increase in a ganglioside identified as GM]. They also
noted several alterations in the glycosyltransferases responsible for ganglioside
synthesis. Most notably, there was a 7-fold increase in GD3 synthase (SAT-2; 6M3
+ CMP-NAN -» GD3) and a 6ofold increase in GMZ synthase (GalNAcT-l; 6M3 +
UDP-GalNAc -’ GMZ) with a large decrease in GD1a synthase (SAT-4; GM1 +
UDP-Gal -» GDla) (See Fig 5.). In other studies, various sialyltransferases and
glycosyltransferases increased in tissues and cells in culture concomitant with
increased levels of glycolipids (31, 45 , 46).

In this study, we have examined the ganglioside profile in the serum of

patients with ductal cell carcinoma of the breast and the ganglioside profile of plasma

146

Figure 5. Biosynthetic pathway for ganglio—seriw gangliosides. See Table 1 for
structures.

 

147

28 1.125 1.5. 1.1.5 888..

 

 

1:3 310 8.5.2.3 >
200 A125 1200 A1 .00 1.1 .00 38.2.
35 1.83 23 5230 +
awn—.0 ‘10— CG ‘1 «Emu T N26 ‘11 ”:6 égfitfle
3.3 :6 «.50 8.5220 +

.0003 T 500.6 101 025800

><31h<n Ofigmoa mn=m03muz<0

148

from an appropriate control group. We determined that there is no significant
difference in the ganglioside content of serum versus plasma, nor is there any
significant difference in profiles of gangliosides of plasma with respect to age in our
sampling of plasma, which agrees with previous reports (33). There have been many
reports in the literature describing elevated levels of "lipid-associated sialic acid" in
the sera of breast cancer patients (47-52) using the method of Kapotodis et a1. (53) in
which gangliosides are precipitated with lipoproteins by phosphotungstic acid after
which the sialic acid is measured with resorcinol. These reports have been criticized
because of their apparent lack of specificity for "lipid-associated" sialic acid due to
the amount of acid-a-l-glycoprotein that coprecipitates (54, 55). Reported values of
"lipid-associated" sialic acid have ranged from 580 nmol/ ml for control to 1.9
pmol/ ml in cancer. When gangliosides have specifically been isolated and quantified,
the lipid-bound sialic acid content in control samples is around 8 to 15 nmol/m1 (29-
33). Our results were lower than previously reported values with around 6 nmol/ ml
LBSA in the control group and 7.5 nmol/m1 LBSA in the cancer group. We attribute
the low recovery to the small amount of starting material (1 ml). Samples of plasma
and serum for the control and cancer were run at the same time in mixed batches,
such that any losses incurred during isolation would be reflected similarly in all
samples. The results can therefore be compared and meaningful observations made.
When the gangliosides isolated from the cancer patients were compared to
those isolated from the control group by thin-layer chromatography, three major
observations were made. First, there was a significant increase in the overall content

of gangliosides; second, there was an increase in the more complex or

149

polysialogangliosides over the more common monosialogangliosides. lastly, and
most interestingly, there was an increase in the b-series gangliosides (GD3, Gle,
Gle, and Gle) relative to the a-series gangliosides (6M3, GMZ, and GDla)' These
have all been noted in Table 2. The increase in sialic acid and galactose, as
determined by gas chromatography, supported these observations (Table 3).
Dyatlovitskaya et al. (56) have examined the ganglioside profile of breast
cancer tissue and reported a large increase in what they called the " more polar
gangliosides", presumably the more complex gangliosides. Skipski et al. (57) found
increases in the polysialogangliosides in methylcholanthrene-induced mammary tumors
in rats, contrary to what Morré et al (44) observed in tumors induced by 7,12-
dimethylbenz[a]anthracene. In the Morré study (44), GMl was reported to be a major
ganglioside accumulating in the cancer cells. GMI was identified by comigration of
an authentic GM 1 standard run on a silica gel thin-layer plate in an ammonia-
containing solvent. GD3 will often comigrate with or surpass GM 1 in this type of
solvent. The increase in GD3 synthase (7 fold) suggests that their GM] band might
have contained some GD3. Factors in determining the profile of ganglioside
expression are relative ratios of glycosyltransferases as well as available substrates
(58-60). The ratio between 6M3 synthase (SAT-1) and GD3 synthase (SAT-2) has
been suggested to moderate the expression of the a-series gangliosides and b-series
gangliosides (58). Figure 5 shows the biosynthetic pathways to gangliosides in the a-,
b—,, and c-series (59-61) and Table 1 shows the structures. The enzymes 0M3
synthase (SAT-l) and GD3 synthase (SAT—2) are the initial step in each of the

respective series. In the Morré study (44), GD3 synthase activity increased almost 7-

150

fold more than 6M3 synthase, suggesting that there may be a shift from the a-series
biosynthetic pathway to the b-series and the increase in GMI may actually be an
increase in GD3. All the above experiments support our findings that the sera of
patients suffering from breast cancer have an increase in polysialogangliosides with a
significant increase in the b-series biosynthetic pathway over the a-series.

The apparent lack of fucose in the ganglioside fraction from either the cancer
or the control group was of interest. Fucosylated sialo-antigens have been found in
several different cancers. Sialylated Lewis antigens (SLea, SLe"; Table l) have been
associated with several adenocarcinomas, especially colorectal and pancreatic types
(62—65), including breast cancer tissue (63, 65). Interestingly, in the tissue, these
carbohydrate antigens were localized in sphingolipid fractions (62, 66), but are
associated with mucins in the plasma (65, 67, 68). Another fucosylated ganglioside
found in cancer is fucosyl-GMI, which was found in the tissue and plasma of small-
cell lung cancer. We did not detect any fucose in gangliosides isolated from either
serum of breast cancer or plasma of the control group by gas chromatography, nor
was a resorcinol-positive spot found that comigrated with either the SLea or SLe"
glycolipid. Our findings suggest that fucosylation of gangliosides is probably not
associated with breast cancer and the SLe" antigen (if present in sera) might be in the
form of mucins.

The question about what the impact of the shedding of gangliosides might be
in cancer is controversial at best. Is this phenomenon merely a secondary effect that
accompanies other more important events or does it play a more direct role?

Circulating gangliosides have been proposed to allow tumor cells to escape immune

151

system surveillance as well as allow tumor cells to metastasize through moderating
various cell-binding events (69-77). Of particular interest to this study is the effect of
the more highly sialylated gangliosides on the immune system and binding. Skipski er
al. (57) found that mouse mammary tumors were more metastatic when rich in
polysialogangliosides. Exogenous polysialogangliosides have been suggested to aid
tumor cells in metastasis through the inhibition of binding to subcellular matrix (9,
78-81). It is possible that breast tumor cells secrete significant amounts of
polysialogangliosides into the local environment surrounding the tumor cells. This
may be reflected in the slight but significant elevation in the gangliosides found in
circulation (35). Tumor cells may be able to disrupt the normal cell-substructure
binding by the release of matrix-binding polysialogangliosides which may block
binding to the substratum. By this means, tumor cells may become contact-
independent, allowing them to leave the local environment and metastasize elsewhere.
Whereupon the surface carbohydrates might again be altered in their structure to
facilitate binding and infiltration of a region remote to the original site.

While the exact mechanisms are not known, in general, a high content of lipid-
bound sialic acid either in the tissue or plasma seems to be linked to a poor prognosis
while lower lipid-bound sialic acid content indicates a better prognosis and the level is
characteristic of the disease state (9, 14-16, 31, 47, 48, 50, 51, 56, 68, 71, 72, 82).
Furthermore, alterations in the ceramide portion of the ganglioside molecule ( longer
or shorter chain length or a-hydroxylation of the fatty acid moiety) can be associated
with the shedding of gangliosides from the surface of cells (37, 38, 83, 84). It

becomes obvious that more information and additional studies are necessary to better

152

understand the importance of tissue and circulating gangliosides in the biology of

cancer.

Perspectives

It was the intent of this study to determine if a circulating ganglioside(s)
isolated from breast cancer patients could be useful as a tumor-associated antigen.
This was to be achieved by characterizing the profile of gangliosides found in breast
cancer sera and in a control or disease-free group. While much information was
obtained about the ganglioside components in serum samples, there was no single
ganglioside that could be considered a tumor-associated glycolipid antigen, i.e. a
novel or elevated concentration of a ganglioside not ordinary found in healthy
individuals’ blood. If one had been found, it could perhaps have been useful in
diagnosis, prognosis, and monitoring of the disease (15, 16, 85). Furthermore,
tumor—associated glycolipid antigens have been examined for use in various therapies.
Certain glycolipids, alone, have suspected antitumor properties (9, 86), while
antibodies raised against tumor-associated ganglioside antigens have been used in
trials and also appear to have anti-tumor activities (9, 87). Furthermore, gangliosides
isolated from tumors have been tested as immunogens (88-90), the goal being the
replacement of less specific therapies like chemotherapies with an anti-tumor immune
response in the host that would be present on a continual basis (91). The use of anti-
ganglioside antibodies in therapy can have several approaches. The antibodies
themselves can recruit the host immune system for defense against the cancer through

macrophage activation or "antibody—dependent cytotoxicity" (92), or antibodies could

153

have various toxins or radionucleotides attached to them (76). Labeled antibodies
could be used for tumor localization by scintillographic photoscanning (93).
Hakomori has suggested two new types of therapy he calls anti-adhesion therapy and
ortho—signaling therapy (76, 94). In anti-adhesion therapy, glycolipids or
antiglycolipid antibodies could be used to disturb the various carbohydrate-initiated
tumor cell binding events required for progression and metastasis. Ortho—signaling
therapy aims at disturbing the glycolipid-moderated signal transduction events.
Problems associated with these therapies are the heterogeneity of different tumor
antigens and the shedding of the antigens from the cells (76). Exact knowledge of the
antigens found in the tissue as well those shed into the plasma must be known in
order to appropriately design a useful therapy.

The results of this study are not of immediate clinical interest but are of
biochemical interest. The finding of an altered profile in the breast cancer sera,
especially in more complex gangliosides, is suggestive of a possible role of these
gangliosides in the biology of breast cancer, possibly in moderating important
carbohydrate-initiated binding events. In order to better characterize this possibility,
the profiles of gangliosides isolated from serum need to be examined over the course
of the disease to determine if the alterations are stage specific. Furthermore, breast
cancer tissue needs to be examined at all stages more closely to determine how and if
the gangliosides found in the tumors affect those in the circulation. Another
interesting experiment would be to determine what effect exogenous gangliosides have
on the various receptors found in breast cancer tissue. The loss of the estrogen or

progesterone receptor in breast cancer tissue results in a poor prognosis and its

154

disappearance is a common indicator for the diagnosis and prognosis of breast cancer
(95). The presence of the erbZ gene, which is thought to encode another growth
factor receptor, is an unfavorable prognostic factOr (95). The effect of several
gangliosides on the various growth factor receptors is well known (Reviewed in ref
10). Studies of cell growth in chemically defined media indicate that exogenously
added 6M3 and GMI alters the binding affinity of cells to platelet-derived growth
factor, epidermal growth factor, or fibroblast growth factor. To our knowledge, no
one has studied the interactions of gangliosides on growth factor receptors that are
important to breast cancer. These experiments would better help define the role of
gangliosides in the biology of breast cancer and aid in developing a useful strategies
of therapy based on ganglioside interactions with the cell-surface molecules and

mechanisms associated with tumorigenesis.

10.

ll.

12.

l3.

14.

15.

16.

17.

18.

REFERENCES

Hakomori, S. (1981) Ann. Rev. Biochem. 50, 733-764

Hakomori, S. ( 1990) J. Biol. Chem. 265, 18713-18716

Wong, C. G., and Ladisch, S. (1983) J. Lipid Res. 24, 666-669
Cham, B. E., and Knowles, B. R. (1976) J. Lipid Res. 17, 176-181
Shiraishi, T., and Uda, Y. (1986) J. Biochem. 100, 553-561
Nakamura, K., and Handa, S. (1984) Anal. Biochem. 142, 406-410

Usuki, S., Lyu, Shu-C., and Sweeley, C. C. (1987) J. Biol. Chem. 263, 6847-
6853

Magnani, J. L., Nilsson, B., Brockhaus, M., Zopf, d, Steplewski, Z.,
Kaproski, H., and Ginsburg, V. (1982) J. Biol. Chem. 257, 14365-14369

Hakomori, S. (1985) Cancer Res. 45, 2405-2414

Hakomori, S. (1986) Chem. Phys. Lipids 42, 209—233

Li, R., and Ladisch, S. (1991) Biochim. Biophys. Acta 1083, 57-64
Valentino, L. A., and Ladisch, S. (1992) Cancer Res. 52, 810-814
Ladisch, S. (1987) Adv. Pediatr. 34, 45-58

Ladisch, S., Wu, Zi-L., Feig, S., Ulsh, L., Schwartz, E., Floutsis, G.,
Wiley, F., Lenarsky, C., and Seeger, R. (1987) Int. J. Cancer 39, 73-76

Valentino, L., Moss, T., Olson, E., Wang, lie-J., Elashoff, R., and Ladisch,
S. (1990) Blood 75(1), 1564-1567

Schulz, G., Cheresh, D. A., Varki, M. M., Yu, A., Staffileno, L. K., and
Reisfeld, R. A. (1984) Cancer Res. 44, 5914-5920

Hamelton, W. B., Helling, F., LLoyd, K., and Livingston, K. O. (1993) Int.
J. Cancer 53, 566-573

Portoukalian, J., Carrel, S., Doré, Jean-F., and Riimke, P. (1991) Int. J.
Cancer 49, 893-899

155

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

156
Svennerholm, L. (1957) Biochim. Biophys. Acta 24, 604

Ladisch, S., and Gillard, B. ( 1987) Methods Enzymol. 138, 300-306

Chigomo, V., Sonnino, S., Ghidoni, R., and Tettamamti, G. (1982)
Neurochem. I. 4, 397-403

Kundu, S. K., Diego, 8. 0., and Marcus, D. M. (1985) Arch. Biochem.
Biophys. 238, 388-400

Mullin, B. R., Poore, C. M. B., and Rupp, B. H. (1983) J. Chromatog. 278,
160-166

Ledeen, R. W., Haley, J. E., and Skrivanek, J. A. (1980) Anal. Biochem.
112, 135-142

Nakamura, K., Ariga, T., Yahagi, T., Miyatake, T., Suzuki, A., and
Yamakawa, T. (1983) J. Biochem. 94, 1359-1365

Ando, S., Chang, Nan-C., and Yu, R. K. (1978) Anal. Biochem. 89, 437-450
Kundu, S. K. (1981) Methods Enzymol. 72, 185-204

Wiesner, D. A., and Sweeley, C. (1993) Anal. Biochem. Submitted

Yu, R. K., and Ledeen, R. W. (1972) J. Lipid Res. 680, 680-686

Ladisch, S., and Gillard, B. (1985) Anal. Biochem. 146, 220—231

Kloppel, T. M., Keenan, T. W., Freeman, M. J., and Morré, D. J. (1977)
Proc. Natl. Acad. Sci. USA 74, 3011-3013

Portoukalian, J ., Abdul-Malak, N., Zwingelstein, G., and Doré, J. F. (1978)
Biochem. Biophys. Res. Comm. 85, 916-920

Senn, Haas-J., Orth, M., Fitzke, E., Wieland, H., and Gerok, W. (1989) Eur.
J. Biochem. 181, 657-662

Portoukalian, J ., David, M. J ., Shen, X., Richard, M., and Dubreuil, C.
( 1989) Biochem. Int 18, 759-765

Bernhard, H., Meyer zum Buschenfelde, Karl-H., and Dippold, W. G. (1989)
Int. J. Cancer 44, 155-160

Tao, R. P., and Sweeley, C. C. (1970) Biochim. Biophys. Acta 218, 372-375

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

157

Kannagi, R., Nudelman, E., and Hakomori, S. (1982) Proc. Natl. Acad. Sci.
USA 79, 3470-3474

Ladisch, S., Sweeley, C. C., Becker, H., and Gage, D. ( 1989) J. Biol. Chem.
264, 12097-12105

Hakomori, s, and Murakami, W. T. (1968) Proc. Natl. Acad. Sci. USA 59,
254-261

Mora, P. T., Brady, R. 0., Bradley, R. M., and McFarland, V. W. (1969)
Proc. Natl. Acad. Sci. USA 63, 1290-1296

Black, P. H. (1980) N. Eng. J. Med. 303, 1415-1416
Hersey, P., and Jamal, P. (1989) Pathology 21, 51-58

Tsuchida, T., Saxton, R., Morton, D., and Irie, R. (1989) Cancer 63, 1166-
1174

Keenan, T. W., and Morré, D. J. (1973) Science 182, 935-937
Ganzinger, V., and Deutsch, E. (1980) Cancer Res. 40, 1300-1304

Bosmann, H. B., and Hall, T. C. (1974) Proc. Natl. Acad. Sci. USA 71,
1883-1887

Riley, M., Tautu, C., Verazin, G., Gregory, J., Josiah, S., Prorol, J., and
Alhadeff, J. (1990) Clin. Chem. 36, 161-162

Dwivedi, C., Dixit, M., and Hardy, R. E. (1990) Experientia 46, 91-94

Chondros, K., Dapolla, V., Stringou, E., Petiniki, M., Vakouftsi, M., and
Kakari, S. (1991) Anticancer Res. 11, 2103-2106

Patel, P., Baxi, B. R., Adhvaryu, S. G., and Balar, D. B. (1990) Anticancer
Res. 10, 1071-1074

Shamberger, R. J. (1984) J. Clin. Chem. Clin. Biochem. 22, 647-651

Michalski, Jean-C., Corfield, A. P., and Schauer, R. (1986) Hoppe-Seyler’s
Z. Physiol. Chem. 367, 715

Katopodis, N., Hirschaut, Y., Geller, N. L., and Stock, C. C. (1982) Cancer
Res. 42, 5270-5273

54.

55.

56.

57.

58.

59.

61.

62.

63.

65.

66.

67.

68.

69.

158

Voigtmann, R., Pokomy, J ., and Meinshausen, A. (1989) Cancer 64, 2279-
2283

Vegh, Z., Kremmer, T., Boldizsar, M., Gesztesi, K. A., and Szajani, B-
(1991) Clin. Chim. Act. 203, 259-268

Dyatlovitskaya, E. V., Tekieva, E. A., Lemenovskaya, A. F., Somova, O.
G., and Bergelson, L. D. (1991) Biokhimiya 56(3), 560—564

Skipski, V. P., Carter, S. P., Terbus-Kekish, O. I., Podlaski, F. J., Peterson,
R. H., and Stock, C. C. (1981) J. Natl. Cancer Inst. 67, 1251-1258

Ruan, S., and Lloyd, K. O. (1992) Cancer Res. 52, 5725-5731
Fishman, P. H., and Brady, R. O. (1976) Science 194, 906-915

Pohlentz, G., Klein, D., Schwartzmann, G., Schmitz, D., and Sandoff, K.
(1988) Proc. Natl. Acad. Sci. USA 85, 7044-7048

Caputto, R., Maccioni, H. J., Arce, A., and Cumar, F. A. (1971) Adv. Exp.
Med. Biol. 71, 27-44

Yang, H., and Hakomori, S. (1971) J. Biol. Chem. 246, 1192-1200
Fukushima, K., Hirota, M., Terasaki, P. I., Walcisaka, A., Togashi, H.,
Chia, D., Suyama, N., Fukushi, Y., Nudelman, E., and Hakomori, S. (1984)
Cancer Res. 44, 5279-5285

Macher, B. A., and Beckstead, J. H. (1990) Leukem. Res. 14, 119-130
Wang, WeiT., Lundgren, T., Lindh, F., Nilsson, B., Griinberg, G., Brown,
J. P., Mentzer-Dibert, holly, and Zopf, D. (1992) Arch. Biochem. Biophys.
292, 433-441

Magnani, J. L., Brockhaus, manfred, Smith, D. F., Ginsburg, V., Blaszczyk,
M., Mitchell, K. f, Steplewski, Z., and Koprowski, H. (1981) Science 212,
55-56

Koprowski, H., Herlyn, M., Steplewski, Z., and Sears, H. F. (1981) Science
212, 53-55

Yogeeswaran, G., and Salk, P. L. (1981) Science 212, 1514-1516

Niedermeier, W., and Tomana, M. (1974) Anal. Biochem. 57, 363-368

70.

71.

72.
73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

159
Grayson, G., and Ladisch, S. (1992) Cell. Immun. 139(1), 18-29

Dyatlovitskaya, E. V., and Bergelson, L. D. ( 1987) Biochim. Biophys. Acta
907, 125-143

Saha, S., and Chattopadhyay, U. (1988) Int. J. Cancer 41, 432-435

Ladisch, 8., Becker, H., and Ulsh, L. (1992) J. Immunol. 8, 180-188

Marcus, D. M. ( 1988) in New Trends in Ganglioside Research: Neurochemical
and Neuroregenerative Aspects, Vol. 14 (Ledeen, R. W., Hogan, E. L.,
Tettamanti, G., Yates, A. J., and Yu, R. K., eds), pp. 173-180, Liviana

Press, Padova

Portoukalian, J. (1989) in Gangliosides and Cancer (Oettgen, H. F., ed), pp.
207-216, VCH Publishers, New York

Hakomori, S. (1992) Tohoku J. Exp. Med. 168, 211-222

Shaposhnikova, G. 1., Prokazova, G. A., Buzinkov, G. A., Zvezdiva, N. D.,
Teplitz, N. A., and Bergelson, L. D. (1984) Eur. J. Biochem. 140, 567-570

Kleinman, H. K., Martin, G. R., and Fishman, P. H. (1979) Proc. Natl.
Acad. Sci. USA 76, 3367-3371

Spiegal, S., Yamada, K., Hom, B. E., Moss, J., and Fishman, P. H. (1986)
Cell Biol. 102, 1898-1906

Cheresh, D. A., Pierschbacher, M. D., Herzig, M. A., and Mugoo, K. (1986)
J. Cell Biol. 102, 688-696

Laitinen, J ., Lopponen, R., Merenmies, J ., and Rauvala, H. (1987) FEBS
Lett. 217, 94-100

Dnistrian, F. M., Schwartz, M. K., Katopodis, N., Franchia, A. A., and
Stock, C. C. (1982) Cancer 50, 1815-1819

Kannagi, R., Stroup, R., Cochran, N. A., Urdeal, D. L., Young, W. W., and
Hakomori, S. (1983) Cancer Res. 43, 4997-5005

Bassi, R., and Sonnino, S. (1992) Chem. Phys. Lipids 62, 1-9

Sasaki, H., Momoi, T., Yamanaka, C., Yorifuji, T., Kaji, M., and Mikawa,
H. (1991) Int. J. Cancer 47, 742-745

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

160

Maru, M., Haraguchi, M., Higashi, H., Kato, S., Kurimura, T., Naiki, M.,
and Wakamiya, N. (1993) Int. J. Cancer 53, 645-650

Houghton, A. N., Bajorin, D., Chapman, P., DiMaggio, J., Oettgen, H. F.,
and Old, L. J. ( 1989) in Gangliosides and Cancer (Oettgen, H. F., ed), pp.
232-238, VCH publishers, New York

Irie, R. F., and Chandler, P. J. ( 1989) in Gangliosides and Cancer (Oettgen,
H. F., ed), pp. 247-258, VCF Publishers, New York

Livingston, P. O., Ritter, G., Oettgen, H. F., and Old, L. J. (1989) in
Gangliosides and Cancer (Oettgen, H. F ., ed), pp. 293-300, VCF Publishers,
New York

Ritter, G., Livingston, P. O., Boosfeld, E., Wiegandt, H., Yu, R. K.,
Oettgen, H. F ., and Old, L. J. (1989) in Gangliosides and Cancer (Oettgen,
H. F., ed), pp. 301-314, VCF Publishers, New York

Stevenson, F. K. (1991) FASEB J. 5, 2250-2257

Adams, D. 0., Hall, T., Steplewski, Z., and Koprowski, H. (1984) Proc.
Natl. Acad. Sci. USA 81, 3506-3510

Goldberg, D. M., DeLand, F., Him, E., Bennett, S., Primus, F. J., van
Nagell, J. R., Estes, J. R., DeSimone, P., and Rayburn, P. (1981) N. Eng. J.
Med. 298, 1384-1388

Hakomori, S. ( 1991) Cancer Cells 3, 461-470

Marx, J. (1993) Science 259, 626-629

CHAPTER 7

Sialidase and Free Sialic Acid in the Serum of Patients with Breast Cancer

D.A. Wiesner and C.C. Sweeley

November 21 , 1993

Department of Biochemistry
Michigan State University
East Lansing, MI 48824

161

162

Abstract

Assays of free sialic acid and soluble sialidase activity were carried out in with
samples of serum from ductal cell breast cancer patients and plasma from a control
group of women was determined. Sialidase activity measured by the use of a coupled
enzyme assay, sensitive to 200 nU per ml of plasma or serum. The samples were
assayed either unmodified, or dialyzed against several buffers at different pH.
Fractions of serum or plasma fractionated over a gel filtration column were also
assayed. In none of the samples, whether from the cancer or the control group, was
any sialidase activity found. Free sialic acid was determined in several samples of
cancer sera and control plasma by a modified thiobarbituric acid assay mathematically
corrected for any contaminants. The concentrations of free sialic acid in the cancer
and control were 1.63 and 1.74 nmol per ml, respectively, and were not found to be
significantly different. There was no difference found in plasma versus serum in
either case. The activity of sialidase and concentration of free sialic acid were not

found to be useful in the diagnosis or prognosis of breast Cancer.

163

Intrfluction

Sialic acids are generally found on the non-reducing termini of carbohydrate
chains of glycoconjugates such as mucins, glycoproteins, gangliosides and milk
Oligosaccharides (1-5). They have been proposed to have a number of different
biological properties such as influencing the structural conformation of
Oligosaccharides (4-7) as well as involvement in recognition and the masking of
binding sites of macromolecules and cells (4, 5). Furthermore, the carboxylic acid of
sialic acids has been suggested to give many molecules their antigenicity (6). The
sialic acid moiety of gangliosides have been proposed to confer mitogenic properties
to the molecule as well as the binding site for various substrates (6, 8). The
hydrolysis of terminal sialic acids from their Oligosaccharide base is catalyzed by the
enzyme sialidase (neuraminidase). Knowing this, it is easy to see how the various
activities of different sialidases can be important to many biological processes. It is
conceivable that the action of a sialidase on gangliosides involved in cellular adhesion,
communication, and transmembrane signaling could have a profound effect on the
biology of the cancer state.

Sialidase activities have been associated with the plasma membrane, cytosol,
lysosome, and the extracellular aspect of the cell (9, 10). The source of the
extracellular sialidase is currently not known; it may be a product of proteolytic
cleavage of plasma membrane bound form, it may be exported from the cell as are
many of the other lysosomal glycosidases, or it may arise via a unique mechanism.

Sialidases typically have a slightly acidic pH Optimum of around 4.5-5.5. One of the

164

extracellular sialidases found in the medium of cultured fibroblasts had a more neutral
pH of around 6.5 (l 1). It is possible that sialidases found outside the cell, in the
local environment, could find themselves in the blood stream. The fact that all other
glycosidases have been found in the plasma in measurable amounts would suggest that
sialidase should also be found in circulation. While sialidases have been found on the
plasma membrane of erythrocytes and many lymphocytes (12), there has been only

one obscure report of a soluble plasma sialidase, reporting very low traces of activity.

This portion of my study was to determine if there is detectable sialidase
activity present in either plasma from the control group or that of patients with breast
cancer. The idea was that while apparently healthy individuals (control group) might
not have any measurable sialidase activity in their plasma, levels of sialidase in
plasma may be elevated in response to the cancerous state. Furthermore, alterations
in a plasma sialidase may result in the accumulation or depletion of free sialic acid.
Total sialic acid as well as lipid-bound sialic acid have been implicated in many
different diseased states and have been suggested as being useful for the diagnosis of
certain cancers (13-26). It is possible that the concentration of free sialic acid in
plasma may also be an indicator of the tumor burden and shed some light on some

mechanisms of tumor biology.

165

Materials and Methods
Materials

Cholera toxin B subunit-horseradish peroxidase conjugate was purchased form
LIST Biological Laboratory (Campbell, CA). All the following were purchased from
Sigma Chemicals (St.Louis, MO): free sialic acid, sialyllactose, all gangliosides, V.
claret sialidase (Type III), o-phenylenediamine tablets, bovine serum albumin, and
thiobarbituric acid. Hydrogen peroxide was purchased from J .T. Baker Inc.
(Phillipsburg, NJ) and Tween 20 as form Bio-Rad Laboratories (Richmond, CA).
Polystyrene EIA/RIA plates (96 well) were from GIBCO laboratories (Grand Island,

NY). All water used was distilled and deionized (ddi water).

Coupled Enzyme Assay of Sialidase.

A coupled enzyme assay (27) that measures very small activities of sialidase
was utilized for the determination of sialidase in serum or plasma. Ganglioside GDla
was coated onto the bottoms of 96 well polystyrene plates by placing 50 pl (25 pmol)
of the ganglioside in ethanol in the well and allowing to dry overnight. After washing
4x with 0.02% Tween 20 in Dulbecco’s modified phosphate-buffered saline solution
(Tween-PBS) and once with just PBS, the wells were blocked with 50 pl of a 1%
bovine serum albumin in PBS (BSA-PBS) for l h. The plate was washed as before
and 50 pl of an appropriate enzyme solution was added to each well and incubated at
37 °C for 45 min. The plate was again washed as before and 50 pl of the cholera
toxin B subunit-horseradish peroxidase conjugate solution (HRP-B, 0.64 mg B

subunit/ml of BSA-PBS) was added and allowed to bind to the GM 1 produced by the

166

action of sialidase for 1 h at 37 °. The substrate solution (10 mg of o-phylenediamine
and 30 pl of 30% hydrogen peroxide in 10 m1 of 0.1 M sodium acetate buffer, pH
5.5, 50 pl) was added after the wells were washed. After a timed interval, typically 2
min, in which the color was allowed to develop, the reaction was stopped with 50 p1
of 4 N H2804. The absorbance at 490 nm was measured with a Bio-Tek Instruments
Inc. EIA Reader EL-810 reader. Quantitation of the amount of GMI ganglioside
produced was based on a standard curve of pure GMl that was attached to the plate in
the same manner as GDla and treated with the HRP-B and substrate in the same
manner as above. This is easily achieved on the same plate by just omitting the

sialidase treatment on the region of the plate containing the GM].

4-MU-NeuAc Assay of Sialidase

As an additional method for the determination of sialidase, 4-
methylumbelliferyl-a-D-N-acetylneuraminate (4-MU-NeuAc) was used as a substrate
with the product 4-methylumbelliferyl (4-MU) measured fluorimetrically (28). The
assay mixture contained 50 pl of the test sialidase mixture, 25 pl of 32 mM CaClZ,
50p1 of the substrate solution, and 75 pl of an appropriate buffer. This mixture was
incubated for 2 h at 37 °C. The reaction was halted by the addition of 1 ml of the
"stop” solution (0.133 M glycine buffer, pH 10.7, containing 0.06 M sodium
chloride). A 100 pl aliquot of this was diluted to 2 ml with additional "stop" solution
and the fluorescence was measured in an Aminco fluoro-colorimeter at an excitation
wavelength of 365 nm and an emission wavelength of 448 nm. A 25 nmol/ml

solution of free 4-MU was determined to have a transmittance of 25% and a 50

167

nmol/ml a transmittance of 50%. Substrate solutions were either 4.0, 0.4, or 0.04

mM 4-MU-NeuAc.

Thiobarbituric Acid Assay of Free Sialic Acid.

Free sialic acid was measured by a modified version of the colorimetric
method of Aminoff (2, 29) and Warren (30). The samples to be analyzed and free
sialic acid (0.5 to 10 nmol) were taken up in 250 pl of dd water and thoroughly
mixed with 125 pl of 25 mM periodic acid in 0.125 N H2804. The tubes were
capped tightly and incubated for 30 min at 37 0C. Excess periodate was removed by
the addition of 125 pl of 1.6% sodium arsenite in 0.4 N HCl, shaking until the
yellow color disappears. A 1 ml solution of 0.9 M thiobarbituric acid (TBA),
adjusted to pH 9.0 with NaOH, was added, the tubes capped, and placed in a boiling
water bath for 7.5 min. The tubes were quickly cooled in an ice bath and 0.5 m1 of
n-butanol containing 5% HCl was added with vigorous mixing. After centrifugation,
the upper, color-containing organic phase was removed and the color measured at 549
nm and 532 nm. The absorbance at 532 nm was used to correct for interferences of

2-deoxy sugars that might be present.

Preparation and Analysis of Serum and Plasma
Heparinized plasma was collected from 8 adult women who were considered to
be in good health. The plasma was taken by venipuncture after an overnight fast and

the women were asked not to be menstruating. This was considered the control

168

group. Serum was taken from 9 women who had recently been diagnosed with
ductile cell carcinoma of the breast and had undergone no treatment as of that point.
Samples were used as soon as possible, and stored at -70 °C if necessary.

The coupled enzyme assay was used to determine the sialidase activity of
plasma prepared in several different ways. The different methods of preparation
included dialysis against buffers of various pHs and dd water. The activity of
unmodified plasma was also measured. In an attempt to isolate any potential sialidase
activity in both the cancer and control groups, 1 ml of each was run over a Sephacryl
S-200 HR gel filtration column (1 x 65 cm) with 1 ml fractions collected during
elution with 20 mM phosphate buffer, pH 6.5 containing 0.2 M sodium chloride.
Protein was determined for each fraction (absorbance at 280 nm) and sialidase activity
was determined by the coupled enzyme assay.

The TBA assay was used to determine the free sialic acid concentration of the
serum and plasma following a few brief clean-up steps. First, proteins were
precipitated from 1 ml aliquots of plasma or serum with 20 vol of cold ethanol. After
thorough mixing, the proteins were pelleted by centrifugation and the supernatants
removed and dried under a stream of nitrogen. The residues, after drying, were
taken up in a small volume of water (0.5 ml) and extracted twice with equal volumes

of hexane, lyophilized and analyzed by TBA analysis.

l 69
Results

 

Sialidase

Several different approaches were used to ascertain the presence of sialidase in
plasma or serum. Initially, 50 pl of heparinized plasma was assayed by the coupled
enzyme assay of Ogura et al. (27). The reported detection limit for this method was
3 nU per well with a well volume of 50 pl. Units were defined as pmols of product
formed (GMl) per minute. In our hands, 10 nU of commercial bacterial sialidase per
well was reproducibly detectable while 3 nmol was detectable, but was not reliable.
With this detection limit of 10 nU per 50 pl, as little as 200 nU of sialidase activity
per ml from plasma or serum was capable of being detected. Other glycosidase found
in plasma, such as B-galactosidase or B-glucosaminidase, have reported activities
around 20 and 600 mU per ml of plasma (31) and sialidase from sonicated human
fibroblasts was 1.2 mU/mg protein (31). While the ability to detect the activity of
sialidase with a sensitivity two orders of magnitude greater than necessary for other
plasma glycosidases, still no sialidase activity was found in the untreated plasma.
Recognizing that most sialidases have pH optima at a more acidic pH, aliquots of
plasma were dialyzed against 0.1 M sodium acetate buffers of either pH 4.2 or 5.5.
In addition, plasma was dialyzed against pure water for analysis of the possibility of
small molecular weight inhibitors. These were assayed with both the coupled enzyme
assay as well as using 4-MU-NeuAc as a substrate and measuring the fluorescence.
These assays showed little more than a trace of potential activity in plasma with either
no activity or results just barely over the detection limit. Where traces of sialidase

activity were found, these results were not reliably repeatable.

170

Another strategy employed in an attempt to find sialidase activity in plasma,
was fractionation of plasma or serum by size exclusion chromatography. Aliquots of
plasma or serum (1 ml) were applied to the top of a Sephacryl S-200 HR gel filtration
column. Each fraction was measured for protein (A280) and sialidase activity by the
coupled enzyme assay of a 50 p1 aliquot. Figures 1 and 2 show the elution profile
and sialidase activity from control plasma and cancer serum. Protein came off in a
large peak spanning 30 fractions. Sialidase was not detected in any of the fractions in
the profile of either the control or the cancer sample. It is possible that an inhibitor
may have interfered with the assay. To try to test for this, protein-containing
fractions were spiked with varying amounts of bacterial sialidase and assayed with the
coupled enzyme assay. The activity of the bacterial sialidase was highly inhibited in
these mixed fractions but this was later attributed to the absence of calcium ions in the
buffer or desorbtion of GDla from the plate through the action of plasma lipoproteins.
When the bacterial sialidase was assayed in the elution buffer (20 mM phosphate, pH
6.5, containing 0.2 M sodium chloride) with and without 4 mM CaClz, the enzyme
functioned as expected in the calcium containing buffer and was extremely low in the
calcium-free buffer. Due to multiple possibilities of inhibitory factors, this approach

was abandoned.

Sialic Acid
Plasma from 8 women comprising the control group and serum from 9 women
with ductal cell cancer of the breast were analyzed for their free sialic acid content.

The sialic acid concentration was computed using the equation derived by

171

Figure 1. Elution profile of 1 ml of plasma from a control subject on Sephacryl
S-200 HR column. Each fraction was assayed for protein (left axis, A280) and
sialidase activity (right axis, A490).

 

 

 

172

06W

 

 

0.28... 8.08 + 0.0.0... ...L

 

000.000.".

 

 

 

 

 

 

 

8%08 0.0 0.0 0.8 0.0 0.0 0... 0.0 0.0 0.8 08.0.
o .. - ..o
1...... -8
..- 4- >< 8.. 4... p» .1 >4 .8...
8.7 - .88 m
N- -m o
8.0- .80
0- -0
8.0- .88
v v

.n 8:»:—

173

Figure 2. Elution profile of 1 ml of serum from a cancer subject on Sephacryl 5"
200 HR column. Each fraction was assayed for protein (left axis, A280) and sialldase
activity (right axis, A490).

 

 

174

 

 

2.2.0.0. mw<w 1.1

0.905 Ill

 

 

80:52 00:00.“.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

om mm. on mo o0 mm om mv 0v mm om .mN 0N mw 0.. o
m.OI n — P — - —1 P — P P _ — p - moo.
0%
mdt .-
..
w 8.. N
6 8
O N 0
ma
m - m
m8” md
.8 V

175

Warren (30) which corrects for possible interference from 2-deoxyribose. The
amount of sialic acid present in a given sample can be determined from equation 1:
pmols Sialic acid =

63 E4
00D0549 - ———_— 0.0.532 X V (1)
6263-6154 6253-6164

where 61 and 62 are the molecular extinction coefficients X 10'3 cm/nmol of sialic
acid at 532 and 549 nm, respectively, and e3 and 64 are the molar extinction
coefficients X 10'3 cm/nmol of 2-deoxyribose at 532 and 549 nm. The values for 51,
52, 63, and 64 were calculated by measuring the absorbance of either free sialic acid or
2-deoxyribose at the specific wavelength and using Beer’s law. They were
determined to be 61 = 19, £2 = 48, 63 = 133, and 64 = 48 cm/nmol which is in
agreement of reported values (30).

No difference was found in sialic acid concentrations when serum and plasma
obtained from the same individual obtained at the same time were analyzed. Table 1
lists the values obtained for free sialic acid concentrations found in the 9 breast cancer
patients plasma and 8 control women samples. The averages were very similar and
showed no statistically significant difference. The cancer plasma was slightly lower

than that of the control group and had a greater range.

D' ion
Sialidase activity has been found in a wide range of mammalian tissues as well
as several subcellular locations. Activity has been found in the lysosome (9, 12, 32,

33), on the lysosomal membrane (9), on the plasma membrane (9, 12), in the cytosol

176

Table l. Sialic acid concentrations found in 9 breast cancer patient plasma sarI'IIDIeS
and 8 women control sera. Averages shown 1 SD. along with the median values-
concentrations were determined by the TBA method. See text for conditions and
calculations.

 

   

177

Table 1.
Free Sialic Acid Concentrations in Plasma and

Serum

 

Cancer (n = 9) Control (n = 8)

 

 

1.000 1.387
0.698 2.215
0.723 2.024
1.139 1.132
1.566 1.922
1.907 1.138
3.681 1.564
2.347 2.392
1.613

Average 1.63 j; 0.94 pmol/ml 1.72 i 0.48 pmol/ml

Median 1.56 1.74

 

178

(9, 34), and more recently, a secreted extracellular sialidase has been found (11, 35,
36). However, sialidase has rarely been reported to be present in measurable
amounts in plasma or serum (37—39), which is unexpected because most other
glycosidases can be found in measurable amounts. However, in 1976, Schauer et al.
(39) used radioactive sialic acid analogs in al-glycoprotein to measure sialidase
activity in plasma and human milk. They found minute amounts of sialidase activity
of 10 fU/ml. Their brief studies on this activity found a pH maximum of 5.5 and
norequirement for metal ions. They also found it to be stable in a frozen state for up
to three months. While the authors could not verify what the source of this sialidase
was, their experiments suggested that it did not come from autolytic blood cells or in
any particulate form. This paper is very intriguing yet suspicious at the same time.
The report of 10 fU/ ml of activity is extraordinary when considering the sensitivity of
the techniques of 1976. Even more strange is a total lack of citations of this report in
the bulk of the later literature. Even one of the authors fails to cite it in further
manuscripts in which it would be relevant.

When sialidase activity is reported to be found in plasma, it is typically
associated with some type of bacterial infection (37, 38). In this case, the sialidase
has been attributed to the bacteria, and not from the host. One of these reports
found no sialidase activity in plasma (37), while the other reported a value of 100
nU/ml based on a personal communication that was never published (38). This trace
amount would require detection methods more sensitive than the most current
methods. The method of Ogura et al. (27) is capable of detecting 200 nU/ml under

typical conditions and is considered to be one of the most sensitive.

179

The inability to find measurable sialidase activity in plasma or serum may be
due to one of several reasons. One possibility is the actual absence of sialidase in
plasma. This would be unusual because of the presence of the other glycosidases
found in plasma. The finding of sialidase activity in the medium of cultured
fibroblasts suggests that a mechanism, albeit unknown, is in place for the secretion of
sialidase into the extracellular space. Since the many other glycosidases are found in
plasma, a selective mechanism for the secretion of specific enzymes must be involved.

If a sialidase were secreted into the plasma or serum, it is highly likely that it
would be very tightly regulated. Studies have shown that sialidase in plasma can be
very harmful to the surrounding environment. Sialic acid acts as a biological mask in
many ways and the action of sialidase, i.e. the removal of sialic acid from
glycoproteins, Oligosaccharides, and sphingolipids, has many biological effects (6).
When free sialidase was injected into the blood of rats several effects were noted
including the removal of much of the serum glycoproteins and tissue dammage. The
removal of the masking effect of sialic acid can cause the removal of glycoproteins
(40) and erythrocytes (41) from the blood. Sialic acid can mask the antigenicity of
carbohydrates, thus when the sialic acid moiety is removed, a possible anti-immune
response can occur (7). One group of investigators found that desialylation of
plasminogen, the plasma zymogen of the fibrinolytic enzyme, fibrin, alters the
zymogen’s kinetics and binding without structure changes (1). Desialylation can also
activate the zymogen to 10 % of the fibrinolytic activity of the processed enzyme.
The fact that several bacterial infections are accompanied by the secreting of large

amounts of sialidase into the plasma and infected tissue suggests that sialidase aids in

180

the infection process by any of the above-mentioned processes (37, 38).
Extracellular sialidases have been shown to be involved in the control of the cell
cycle, acting through 6M3 and epidermal growth factor receptor (35, 36). It
becomes obvious that biological consequences of an unrestricted sialidase activity in
plasma could be hazardous and this strongly suggests that if there is indeed some
enzyme activity, it would be tightly regulated. Plasma may have some indigenous
inhibitor against sialidases such as salts or specific cations which have been shown to
be unnecessary or inhibitory to various plasma sialidases (34, 39, 42, 43). It is
possible that a plasma sialidase is associated with a protective protein such as B-
galactosidase as found in lysosomal sialidase in human placenta (33, 44-46). This
protective protein may closely regulate the activity of sialidase in a tightly controlled
manner.

The presence of sialidase in plasma or serum has yet to be proved in a
convincing way. We found no activity of sialidase in either breast cancer sera or
control plasma along with no alterations in the amount of free sialic acid, which tends
to agree with these findings. Our data show that a circulating sialidase may not have
a role in the progression of breast cancer, yet the possibility of an extracellular

sialidase working in the local environment of the tumor cells has not been determined.

10.

11.

12.

13.

14.

15.

REFERENCE
Stack, M. S., Pizzo, S. V., and Gonzalez-Gronow, M. (1992) Biochem. J.
284, 81-86
Aminoff, D. (1961) Biochem. J. 81, 384-392

Kakehi, K., Suzuki, S., Honda, S., and Lee, Y. C. (1991) Anal. Biochem.
199, 256—268

Jourdian, G. W., Dean, L., and Roseman, S. (1971) J. Biol. Chem. 246, 430-
435

Reutter, W., Kottgen, E., Bauer, C., and Gerok, W. (1982) in Sialic Acids:
Chemistry, Metabolism and Function, Vol. 10 (Schauer, R., ed), pp. 263-306,
Springer-Verlag, New York

Schauer, R. (1985) Trends Biochem. Sci. 117, 357-360

Schauer, R. (1982) Adv. Carbohydr. Chem. Biochem. 40, 131-234
Yogeeswaran, G., and Salk, P. L. (1981) Science 212, 1514-1516

Miyagi, T., Hata, K., Konno, K., and Tsuiki, S. (1992) Tohoku J. Exp. Med.
168, 223-229

Sweeley, C. C. (1993) in Sphingolipids and their Metabolites, lst ed., Vol. 26
(Bell, R. M., Hannum, Y. A., and Merrill, A. J., eds), pp. 235-252,
Academic Press, New york

Usuki, S., Lyu, Shu-C., and Sweeley, C. C. (1987) J. Biol. Chem. 263, 6847-
6853

Sagawa, J., Miyagi, T., and Tsuiki, S. (1990) J. Biochem. (Tokyo) 107, 452-
456

Klapan, 1., Katic’, V., ’Culo, F., Sabolovic’, D., C’uk, V., Fumic’, K., and
Simovic’, S. (1993) Eur. J. Can. 29A, 839-845

Crook, M. (1993) Clin. Biochem. 26, 31-38

Dnistrian, F. M., Schwartz, M. K., Katopodis, N., Franchia, A. A., and
Stock, C. C. (1982) Cancer 50, 1815—1819

181

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.
30.

31.

182

Pu, W. Y., Neale, E., and Shun, F. W. W. (1993) Int. J. Gyn. C. 3(3), 143-
146

Patel, P., Baxi, B. R., Adhvaryu, S. G., and Balar, D. B. (1990) Anticancer
Res. 10, 1071-1074

Dwivedi, C., Dixit, M., and Hardy, R. E. (1990) Experientia 46, 91-94

Riley, M., Tautu, C., Verazin, G., Gregory, J., Josiah, S., Prorol, J., and
Alhadeff, J. ( 1990) Clin. Chem. 36, 161-162

Stringou, E., Chondros, K., Kouvaris, J., Kakari, S., and Papavassiliou, K.
(1992) Anticancer Res. 12, 251-256

Voigtmann, R., Pokomy, J., and Meinshausen, A. (1989) Cancer 64, 2279-
2283

Haq, M., Haq, S., Tutt, P., and Crook, M. (1993) Ann. Clin. Bi. 30, 383-386

Vivas, 1., Spagnuolo, L., and Palacios, P. (1992) Gynecol. Oncol. 46, 157-

162
Polivkova, J., Vosmikova, K., and Horak, L. (1992) Neoplasma 39, 223-226

Lindberg, G., RAstam, L., Gullberg, B., and Eklund, G. A. (1992) Int. J.
Epidemiol. 21, 253-257

Kokaglu, E., Sbmnez, H., Uslu, E., and Uslu, I. (1992) Cancer Biochem.
Biophys. 13, 57-64

Ogura, K., Ogura, M., Anderson, R. L., and Sweeley, C. C. ( 1992) Anal.
Biochem. 200, 52-57

Potier, M., Mameli, L., Bélisle, M., Dallaire, L., and Melancon, S. B.
(1979) Anal. Biochem. 94, 287-296

Uchida, Y., Tsukada, Y., and Sugimori, T. (1977) J. Biochem. 82, 1425-1433
Warren, L. (1959) J. Biol. Chem. 234, 1971-1975
Wenger, D. A., and Williams, C. (1991) in Techniques in Diagnostic Human

Biochemical Genetics: A Laboratory Manuel (Hommes, F. A., ed), pp. 587-
617, Wiley-Liss, New York

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

183

Michalski, Jean-C., Corfield, A. P., and Schauer, R. (1986) Hoppe-Seyler’s
Z. Physiol. Chem. 367, 715

Hiraiwa, M., Nishizawa, M., Uda, Y., Nakajima, T., and Miyatake, T.
(1988) J. Biochem. 103, 86-90

Miyagi, T., and Tsuiki, S. (1985) J. Biol. Chem. 260, 6710-6716

Usuki, S., Hoops, P., and Sweeley, C. (1988) J. Biol. Chem. 263, 10595-
10599

Usuki, S., and Sweeley, C. C. (1988) Ind. J. Biochem. Biophys. 25, 102-105

Rodriguez-Iturbe, B., Katiyar, V. N., and Coello, J. (1981) N. Eng. J. Med.
304, 1506-1510

Schauer, R., Sander-Wewer, M., Gutschker-Gdaniec, G. H. M., Roggertin,
P., Tandow, E. A., and Hobrecht, R. (1985) Clin. Chim. Act. 146, 119-127

Schauer, R., Veh, R. W., Wember, M., and Bushcer, Hans-P. ( 1976) Hoppe-
Seyler's Z. Physiol. Chem. 357, 559-566

Ashwell, G., and Morell, A. G. (1974) Adv. Enzymol. 41, 99-128

Miiller, E., Schroder, C., Sharon, N ., and Schauer, R. (1983) Hoppe-Seyler’s
Z. Physiol. Chem. 364, 1410-1420

Corfield, A. P., Veh, R. W., Wember, M., Michalski, Jean-C., and Schauer,
R. (1981) Biochem. J. 197, 293-299

Caimi, L., Lombardo, A., Preti, A., Wiesmann, U., and Tettamanti, G.
( 1979) Biochim. Biophys. Acta 571, 137-146

Verheijen, F., Palmeri, S., Hoogeveen, A. T., and Galjaard, H. (1985) Eur.
J. Biochem. 149, 315-321

Hiraiwa, M., Uda, Y., Nishizawa, M., and Miyatake, T. (1987) J. Biochem.
101, 1273-1279

Verheijen, F. W., Palmeri, S., and Galjaard, H. (1987) Eur. J. Biochem. 162,
63-67

184

CHAPTER 8
Discussion and Perspectives

The goal of this study was to determine if the ganglioside profile in plasma or
serum from breast cancer patients was different than that which could normally be
found in plasma or serum of healthy individuals. A unique circulating ganglioside
profile could aid in the assessment of the onset of breast cancer and might be useful
for diagnostic and/or monitoring purposes. Furthermore, by determining the
components of circulating ganglioside profiles, more information would be at hand to
try to understand the mechanism of breast tumor growth and the propensity for these
tumors to metastasize. If a particular ganglioside were a tumor-associated antigen and
unique to breast cancer, a possible next step would be the development of antibodies
against it. Once these antibodies have been raised, assays could be developed for the
simplification of the detection of the ganglioside antigen with less complicated
isolation procedures. This ganglioside-antigen might also be a candidate for immune
therapy as discussed previously. The end result would be a quick and simple assay
that could provide the physician with potentially very useful information for
diagnosis, prognosis and the possibility for another choice in therapeutic strategies.

It was stated in the statement of the problem that in order to properly assess

this potential, the current technology and protocols would have to be reassessed and

185

optimized where possible. Limited amounts of plasma or serum obtainable from
patients, often less than 3 m1, requires that current isolation protocols be able to
maximize the recovery from such small volumes. We devised modifications to the
method of Ladisch et al. (1) that allow for both qualitative and quantitative analysis.
Our development of a system of 2-D HPTLC and image densitometry analysis has
aided in increasing the resolution and detection of gangliosides mixtures isolated from
plasma or serum. By employing these techniques and modifications, gangliosides can
be well-characterized from as little as 1 ml of plasma or serum.

Gas chromatography was originally employed to quantify ganglioside standards
for use on HPTLC. It was quickly recognized for its potential to characterize
mixtures of gangliosides by component analysis. In component analysis, unusual
sugars such as fucose could easily identify the presence of potentially interesting
gangliosides, as well as characterize other alterations in profiles by changes in the
ratio of certain sugars. The concomitant analysis of the fatty acids and long-chain
bases added to our information concerning the nature of the hydrophobic tail of
gangliosides found in the circulation. Our modifications to the current protocol for
analysis by GC made the method much faster, simpler to use, and capable of dealing
with smaller sample sizes.

After having developed a solid means of isolating and characterizing
gangliosides isolated from plasma or serum, we profiled the circulating gangliosides
of cancer patients and an appropriate control group. The methods and protocols we

developed were well suited for this type of analysis. We were able to successfully

characterize 12 different resorcinol-positive spots on 2-D HPTLC corresponding to

 

186

gangliosides isolated from 1 ml of plasma. Other studies required larger starting
volumes of plasma in order to achieve the degree of analysis we observed (2, 3). The
GC analysis was supportive of the HPTLC work and we were able to describe 5
different oz-hydroxy fatty acids associated with circulating gangliosides not previously
reported.

The concentration of free sialic acid was observed to be the same in the
circulation of both the control individuals or cancer patients. Also, we were unable to
find and sialidase activity in plasma or serum of either group. Consequently, neither
of these can be considered useful for diagnostic purposes. The apparent lack of
involvement of extracellular free sialic acid and sialidase in the alteration of
gangliosides and possibly other glycoconjugates suggests that any alterations in the
degree of sialylation of gangliosides most likely occurs within the cell and is
modulated by intracellular reactions.

We determined that the profiles of circulating gangliosides were altered in the
sera of breast cancer patients. There was not found a serum ganglioside that was
unique to breast cancer that could be characterized as a circulating tumor-associated
ganglioside antigen. We were, however, able to report some interesting biochemical
findings about the nature of circulating gangliosides associated with breast cancer. A
potential reason for the lack of finding a tumor-associated ganglioside in breast cancer
patients may be related to the nature of breast cancer. The two types of cancer that
have well-characterized tumor-associated ganglioside antigens are neuroblastoma and
melanoma. Both cancers are very fast growing, aggressive and metastatic. Breast

cancer, while highly metastatic, is not as aggressive and rapidly growing, and the

187

tumors are usually quite small at the time of diagnosis . The continuous cellular
division of more aggressive cancers such as neuroblastoma and melanoma may have
some causal effect on the shedding of gangliosides into the circulation. Furthermore,
this high degree of shedding may aid these aggressive cancers in their ability to grow
and metastasize by inhibiting the immune system.

Our findings of an increase in the relative proportions of the more complex
gangliosides, and in particular, b-series gangliosides, are consistent with the theory
that polysialogangliosides are involved in metastasis. The polysialogangliosides have
been shown to bind to various extracellular matrices, and one can speculate that free
gangliosides could compete for and occupy cellular binding sites, freeing tumor cells
from substrata anchoring and associated mechanisms. A possible model for the
involvement of gangliosides in metastasis is described in Figure 1: A) After a tumor
cell population becomes established in the breast tissue, alterations in the profile of
gangliosides found on the tumor cell surface reflect a move to the more complex
ganglioside. B) These gangliosides are subsequently shed from the cell surface. As
a result, current anchors may be released and potential binding cites of the
extracellular matrix are blocked. The cells are now freed of potential contact
inhibitory signals and continue to grow. The gangliosides shed from the now rapidly
growing tumor cells protect the cells from immune surveillance and allow the cells to
escape the local environment and enter the circulation. C) Once in circulation, the
cellular surface again alters its surface gangliosides and other glycoconjugates. D)
This new surface allows the tumor cells to bind to the blood cell walls and

subsequently invade new tissue and establish new regions of growth.

188

Further experiments will be necessary to support this hypothesis.
Ganglioside profiles from the circulation and the tissue would be monitored through
the progression of the disease, noting any profile changes and attempting to correlate
them with stage-specific tumorigenic events. This way, the role of circulating and
tissue gangliosides might be better understood in slow-growing yet metastatic tumors.
By better understanding of these tumorigenic processes, we become more able to

develop treatments which may eventually lead toa cure.

189

Figure 1. Possible role of gangliosides in metastasis. A)Tumor cell (T) in tissue
bound to matrix through ganglioside mediated interactions. B) Shedding of
gangliosides frees tumor cell from matrix and blocks matrix binding sites as well as
blocking detection by lymphocytes (L). C) Tumor cells translocate through the blood
cell wall epithelial cells (E), continually shedding gangliosides to block binding and
detection. D) Specific gangliosides reappear on tumor cell surface to initiate binding
to blood cell walls, and infiltration of new tissue.

190

 

 

 

 

 

 

 

 

 

 

REFERENCES

Ladisch, S., and Gillard, B. (1987) Methods Enzymol. 138, 300-306
Tao, R. P., and Sweeley, C. C. (1970) Biochim. Biophys. Acta 218, 372-375

Senn, Haas-J., Orth, M., Fitzke, E., Wieland, H., and Gerok, W. (1989) Eur.
J. Biochem. 181, 657-662

191

APPENDIX

 

Differentiation (Diff): PD—Poorlv differentiated, MWD- Medium

Patient Bio Data and HPTLC Data
All data ordered based on a/b ratio

Tabkel.
Exp# Age Diff.
3 60 WD
29 62 WD
27 59 WD
26 44 PD
31 60 MWD
23 56 MWD
2 52 MWD
28 54 MWD
9 53 PD
1 60 MWD
35 54 PD
4 56 MWD
21 44 PD
33 53 MWD
24 57 MWD
25 49 MWD
30.19 61 1MB
34 54 PD
20 70 MWD
22 50 PD
7 51 PD
11 52 WD
15 42 PD
36 55 PD
38 49 PD
39 60 MWD

192

 

ER. Stage 6M3 GM2 SP6

”0 2: 2: I! “U 23 I! hU I! ”U 2! ”U *U 25 “U 2! ”U I! I! "U “U "U 2! ”D "U “D

11
II
I
I
II
I
II
I
111 M
IV M
W M

111 M
IV
11
II

4218
3512
7678
4672
4023
4439
3626
4118
3289
2601
4087
4118
2598
3590
3897
4481
2522
2084
1981

pmol LBSA

164 118 26
711 281 289
1910 451 0
226 952 127
121 164 111
163 385 79
191 96 133
497 610 103
145 227 12
263 154 52
221 784 214
335 271 269
362 193 88
166 186 129
84 276 204
237 163 294
203 221 112
214 341 144
266 406 0

Estrogen Receptors (ER): P- Postitive, N- Negitive

52
71
157
117
28
57
129
149
14
121
215
229
53
22
69
879
11
O

0

2—631mg4 m26

0

13
57
73

33
16
158
23
65
16
109
77
41

well

Table ;.

AH data ordered based on a/b rafio

Exp#
3

29
27
26
31
23

2

28

(O

193

Patient Bio Data and HPTLC Data (Cont'd)

 

 

snhc GD3

45

GDla Gle GT1b ()le

229

667

nmol LBSA

Total

114 19 5897
241 3 6540
1579 609 13701
416 103 10060
146 27 6110
338 109 7235
231 95 6380
616 140 7967
322 64 5336
185 40 4492
448 220 8772
418 140 7259
319 61 5090
387 48 6629
481 148 7345
327 972 9148
323 118 5050
606 217 5201
1386 1086 7574

 

 

 

M/D M/DTan2/sp

4.73

39
a/b —- Ratio of a series/ b — series gangliosides
— Ratio of GM3 to GDS

M/D

M/DTQ - Rafio ofrnono-

to polysialogangliosides

M2/ spg — Ratio of GM2 to sialoparagloboside

363
3.12
304
399
2377

L39

194

Table 2 Gas Chromatography Data: Sugars and Long-Chain Base

Order is based on all: ratio [See Table 1]

Number Gal

30.1 9

E88885=uk385§ mxwaaatdwsawassaammw

4847
4307

[SIC
4975

9 838825328

84%

2 262%?

GalNAc GlcNAc Unk NeuAc d18:1

897.

538
480

§§§§§§S§

1022
710
1455
210

1283

285 880
358 558
318 381
238 1318
855 734
353 817
214 742
522 1014
108 380
418 849

4785
4824
4825
3408
8724
881 2
4887
7782
281 8

1 835
401 1
8255
8748
8832
8553
1578
4910
51 28
2888

3813
8843
4341
3882
1822
3611
1805

2501

195

Table 3. Patient Data — Gas Chromatography Data: Fatty Acids
Order is based on a/b ratio (See Table 1)

 

 

 

pmol
Exp# 14:0 15:0 16:1 16:0 18:2 18:1 18:1 18:0 20:0 u3 22:0]
3 236 131 188 1849 407 458 134 2097 282 604 544
29 493 389 577 4800 924 1099 365 4084 458 958 1405
27 355 202 292 4889 997 774 325 4730 605 740 1697
26 398 194 196 6826 107 582 172 2063 470 2537 1597
31 224 204 271 3515 593 570 233 3037 370 549 1286
23 421 263 234 6021 117 288 138 2518 511 1871 1394
2 188 122 184 1498 375 375 97 1417 205 830 527
28 436 360 949 8727 747 959 444 5855 981 950 2865
9 100 340 547 2518 756 823 136 2028 271 2079 611
1 368 180 283 343 44 647 193 3060 421 390 1441
35
4 487 247 235 4200 8'76 988 273 4392 461 2349 1333
21 65 115 150 5586 306 552 210 2729 946 2496 4084
33 102 156 473 5855 50 79 318 5288 1040 933 3477
24 147 80 73 2259 70 216 144 701 134 774 569
25 409 205 175 7908 122 1902 300 2047 0 1852 474
30.19 68 29 11 684 57 117 34 384 60 229 167
34 89 130 163 4551 654 577 212 5385 620 928 2049
20 457 203 214 4078 33 482 180 1876 354 2045 1186
22 296 121 81 1550 69 152 39 978 199 1059 590
7 -
11
15
36
38
39

196

Table 3. Patient Data — Gas Chromatography Data: Fatty Acids (cont‘d)
Order is based on a/b ratio (See Table 1)

 

 

 

Exp fl 23:0 24:1 24:0 111610 h22:0 h23:0 h24:1 112-Q]
3 189 513 395 44 91 0 60 75
29 687 1227 1302 106 259 158 333 511
27 930 1604 1176 152 325 186 377 505
26 580 1624 1100 473 117 97 176 195
31 611 1075 1093 94 201 125 228 428
23 432 1327 1057 164 177 113 113 190
2 288 430 511 46 120 102 57 94
28 1474 3852 2394 205 434 277 770 727
9 221 584 557 58 76 40 70 98
1 634 111 1046 69 198 99 9 297
35
4 530 670 1817 148 164 0 119 157
21 1330 2189 2848 210 309 240 159 410
33 1434 2298 2793 127 267 154 219 477
24 167 338 445 82 32 22 57 43
25 224 403 355 348 76 90 134 51
30.19 55 180 126 47 0 0 137 59
34 . 815 1036 1420 164 334 198 178 520
20 557 969 947 95 141 109 154 181
22 235 436 417 53 91 75 66 125
7
11
15
36
38
39

J_;