Critical Reviews in Oncology/Hematology, 1992; 12:151-166 0 1992 Elsevier Science Publishers B.V. All rights reserved 1040-8428/92/$15.00

ONCHEM

151

00020

The blood group antigens (BGA)-related glycoepitopes A key structural determinant

in immunogenesis

and cancer pathogenesis

Gennadi V. Glinsky Cancer Research Center and Analytical

Bio-Chemistry

(Accepted

Laboratories,

2 December

Inc., Columbia, MO, USA

1991)

Contents I.

Abstract..........................................................

151

II.

Introduction.....................................................................

152

III.

Tumor

IV.

cell aggregation

Modification

and metastasis:

of in vitro cell aggregation

serummacromolecules V.

BGA-related

a pathogenetic

and metastasis

.. ........... ..

glycoepitopes

link

as key structural

.....

152

in vivo by glycoamines . ..

determinants

for cell aggregation

Aberrant

158

VII.

BGA-related

VIII

Tumor

glycosylation,

phenotypic

glycoantigens

cell aggregation

IX.

properties

of tumor cells, and tumor progression organogenesis

and cell differentiation

and serum macromolecules

of a new family of antimetastatic

Tnantigens

Conclusion:

divergence

in embryogenesis,

drugs.

Gennadi V. Glinsky received

a prospect

of anticarbohydrate

160

Research

Center and a Re-

Laboratories,

Columbia,

MO. Correspondence: Cancer

Research

65201, USA.

Dr. G.V. Center,

Glinsky,

M.D.,

3501 Berrywood

Ph.D., Drive,

Senior

161

vaccine

therapy

for treatment

161 of AIDS

and 163 164

his M.D. degree from the State Medical

Bio-Chemistry

160

...................................................

Chernovtzy, and a Ph.D. from the Institute of Oncology Academy of Sciences of Ukraine, Kiev, USSR. Dr. Glinsky at Analytical

a model

.........................................................

. . . . . . . . . ..-......................

a Senior Scientist at the Cancer

metastasis;

...............................

cancer........................................................................... References

in cancer

159

.............................................................

C. Lewis family of glycoantigens

Fellow

to

VI.

B. Galcl(l-3)Galepitope

is presently

in relation

156

A. Tand

search

153

cancerandimmunogenesis.........................................................

for development

Institute. Problems

and other

. .._._._.__.__.__.._......

Scientist,

Columbia,

MO

I. Abstract

This overview has been focused on the functional and pathophysiological aspects of blood group antigens (BGA)-related glycodeterminants. It has been postulated that in a broad range of histogenetically different tissues and organs BGA-related glycoepitopes are expressed on the cell surface at definite stages of cell differ-

152

entiation during embryogenesis, organogenesis, tissue repair, regeneration, remodeling and maturation when ‘sorting-out’ behaviour of one homotypic cell population from heterotypic assemblage of cells occurs. In this event the BGA-related glycoepitopes, if being expressed on the cell surface, play a role of key structural determinants in cellcell recognition, association and aggregation. This mechanism has been discussed in relation to immunogenesis regarding of antigen presentation, selfnon-self discrimination, positive and negative selection during thymic education. It is postulated that the appearance of the BGA-related glycoepitopes on the cell membrane is a consequence of the association of MHC and peptides, with subsequent elimination of cells carring high density of BGA-related glycoepitopes on their surface. In cancer it has been considered as a key mechanism of phenotypic divergence of tumor cells, immunoselection, tumor progression and metastasis. II. Introduction

The blood-group ABH determinants are major allogenie antigens in erythrocytes and many other human tissues. Historically, the term ‘blood-group antigens’ has been applied to the ABH and related antigens. However, they constitute major allogenic antigens of most epithelial cell types, and are also found in primary sensory neurons [1,2]. These antigens appear earlier in ectodermal or endodermal tissue than mesenchymal hematopoietic tissues and cells [2]. Therefore, the term ‘histo-blood group’ has recently been proposed for these allogenic carbohydrate antigens [3] that generally represent peripheral parts of the oligosaccharide chains of glycoconjugates. The structural aspects of the ABH antigens, relationships between their expression, development, differentiation and maturation, as well as malignant transformation, have been recently discussed [3]. In this overview, we have focused on the functional aspect of blood group antigens (BGA)-related glycoepitopes mainly regarding their key role in cell-cell recognition, association, adhesion, as well as homotypic and heterotypic cell aggregation in relation to immunogenesis, tumor progression and metastasis. III. Tumor cell aggregation and metastasis: a pathogenetic link

The process of cell adhesion consists of multiple steps, and the last model of such a multistep process was proposed by S. Hakomori and co-workers in 1989 141.The initial step is specific recognition between two cells in which multivalent homo- and heterotypic carbohy-

drate-carbohydrate interactions mediated by bivalent cations may play a major role [5]. Initial cell recognition through carbohydrate-carbohydrate interaction is followed by nonspecific adhesion, primarily mediated by Ca2+-sensitive adhesion molecules or sugar-binding proteins, or by pericellular adhesive meshwork proteins consisting of fibronectin, laminin, and their receptor systems (integrin). The third step of cell adhesion is the establishment of intercellular junctions, e.g., ‘gap junctions’, in which a cell-cell communication channel is opened through a specific structural protein. It is important that tumor cells retain, to a great extent, their ability to reveal initial stages of cell adhesion: at the same time cancer is characterized by profound disturbances in the subsequent stages of cell adhesion occurring with the involvement of extracellular matrix proteins and completing with the formation of specific ‘gap junctions’. Cell adhesion is processed mainly through recognition between similar types of cells. Thus, the ‘sortingout’ behaviour of homotypic cell populations plays a major role in morphogenesis and organogenesis. Specific interaction between identical or similar carbohydrate chains highly expressed in the same type of cells may provide a better explanation for sorting out than the interaction of relatively nonspecific adhesive molecules [4]. We suggested [6-91 that the recognition between identical types of cells that provides the basis for sorting one homotypic population out of a heterotypic assemblage of cells plays a major role in the metastatic spreading of tumor cells. This mechanism provides a basis for the formation of multicellular aggregates by tumor cells in the circulatory channel of tumor host and their subsequent arrest in blood vessel of target organs. One of the characteristic features of the growth of malignant tumors cultivated in vitro is their ability to form three-dimensional, spheroid structures termed multicell spheroids [lO,l 11. Similar (according to biochemical, immunocytochemical, morphological, and cytological criteria) spheroid-like formations are observed in rodent and human malignant tumors in vivo. Therefore, the multicell spheroids in vitro are considered a cytofunctional analog of the avascular stage of in vivo tumor development [l 11. Cell association and aggregation occur in the first stage of multicell spheroid formation and can be measured in vitro using the cellaggregation assay. Many studies indicate that cell surface components play an important role in metastasis [12-141, in particular in the formation of tumor cell aggregates and emboli. Circulating metastatic tumor cells tend to adhere to other tumor cells (homotypic adhesion) or to host cells including platelets and lymphocytes (heterotypic adhesion) to form emboli, the formation of

153

which is correlated with metastasis [15-171. The lodgment, attachment, and growth of blood-borne neoplastic cells depends largely on embolization [ 18,191. In a B 16 melanoma experimental model of metastasis [20,2 l] it has been shown that tumor cell clumps produce more lung metastases after i.v. injection than do single cells [ 151and has been demonstrated that a correlation exists between the tendency of cells to undergo homotypic and heterotypic aggregation in vitro and their metastatic potential in vivo [19,22-251. Most of these studies have demonstrated differences in cell surface glycoconjugates between tumor cells exhibiting high and low metastatic capacity. It is quite well established that cell surface constituents are involved in cell-cell and cell-substratum recognition and adhesion [26] and that such interactions are relevant for tumor cell embolization, specific arrest, and implantation in secondary sites [12-141. The emboli may lodge in narrow capillaries, nonspecifically, or they may adhere preferentially in specific organs, presumably as a result of cognitive interactions between the tumor cells and capillary endothelial cells or exposed basement membrane components [12-141. Tumor cell surface lectins play a major role in cellular interactions in vivo that are important for the formation of emboli and for the arrest of circulating tumor cells [27-3 11.Antilectin antibodies localize lectins in the cytoplasmic compartment of various normal cells as well as in the extracellular compartment, where they seem to be associated with the extracellular matrix after being secreted by the cells [3236]. Various functions have been proposed for the lectins, including mediation of intercellular recognition and adhesion [3642], control of cellular proliferation [32,43,44], organization of the extracellular matrix [32,34,36], and elastic fiber formation [45]. Tumor cells and tissues contain lectins that are similar in their sugarbinding specificities, molecular weights, and sequences to those isolated from normal cells [4649, 31, 5&54]. It was found by using monoclonal and polyclonal antilectin antibodies that the lectins were expressed on the cell surface and that their level correlated with malignant transformation and metastatic propensity [50,5 1, 531. A key role for the tumor-surface lectins in anchorage-independent growth in vitro and in tumor embolization (homotypic and heterotypic aggregation) during the metastatic process in vivo has been suggested [27311. The homotypic aggregation of B16 melanoma cells depends on the presence of fetal bovine serum [19,25]. One possible explanation for this serum requirement could be that a serum glycomacromolecule(s) mediates intercellular adhesion similar to the action of cell-cell adhesion molecules. Thus, the homotypic and heterotypic aggregation properties of tumor cells represent important biological

features of malignancy because these properties of transformed cells could determine metastatic ability of neoplastic tumors [ 12-19,22,23,25,27-3 1,411. The key structural determinants of the tumor cells that participate in cell recognition, association and aggregation have been determined as carbohydrates and carbohydrate-binding proteins [12-14,26-32.55-59). Our studies have shown that the serum glycoamines and other macromolecules may play important roles in changing the adhesion and aggregation properties in vitro and metastasis in vivo of cancer cells and, thus, may be involved in the process of modification of cell recognition, association and aggregation [69]. Since carbohydrates and carbohydrate-recognition structures are involved in cell aggregation, we have initiated the investigation of structural-functional interrelations of glycoamines, their synthetic structural analogs, and high molecular weight carbohydrate-associated serum tumor markers as a modifiers (inhibitor or stimulator) in vitro of tumor cell aggregation and metastasis in vivo. The results of ongoing research programs have provided increasing evidence that support the hypothesis that: (i) different molecular isoforms of glycomacromolecules in serum contain cell-originated glycoepitopes which are involved in cell recognition and cell association processes (the family of cell recognition glycoepitopes containing extracellular biopolymers (biomacromolecules) ~ CRG-CEB); (ii) these different molecular isoforms of the CRG-CEB and serum carbohydrate-binding proteins (e.g., anti-carbohydrate antibodies, lectins, etc.) play an opposite role either as an inhibitor and/or stimulator of cell association and aggregation, depending on the nature and size of the carrier molecules, valency of glycoepitopes (for CRG-CEB) and binding properties (for carbohydrate-binding proteins); (iii) human cancer is accompanied by changes in quantitative and qualitative characteristics of the CRG-CEB and carbohydrate-binding proteins in blood serum, and (iv) benign and malignant human tumors have different in vitro cell aggregation properties. IV. Modification of in vitro cell aggregation and metastasis in vivo by glycoamines and other serum macromolecules

Glycoamines are a newly recognized class of endogenous small molecular weight biopolymers consisting of amino acids and sugar units with covalent structure of ester, Schiff bases and Amadori bonding. They have been investigated during the last 15 years as physiological components of human and rodent blood serum that merit interest as potential human humoral tumor makers [6&65]. Structurally the glycoamines represent carbohy-

154

drate-amino acid conjugates containing from 5 to 29 amino acids and from 1 to 17 carbohydrate residues [I&60,63,6668]. The level of these substances is substantially increased in blood serum from humans and animals with different forms of malignant solid tumors and leukemias. The chemical structure of glycoamines reveal mono-, di- and trisaccharides bound by ester links to the amino acids and assembled into higher molecular weight compounds via the formation of Schiff base and Amadori product-type bonds with the involvement of the amino groups of amino acids and aldehyde or keto groups of the carbohydrates [8,60,63,6669]. Elevated level of glycoamines in the body fluids of cancer patients may provide one of the important factors that determine the malignant behavior of tumors in vivo. Biochemical and structural analysis of aminoglycoconjugates revealed their polyamine-like properties [60,66,67,70,71]. This can be attributed to the polyamine-like structural organization of free amino groups of the amino acids in aminoglycoconjugates and accounts for the high biological activity of glycoamines. In particular, glycoamines bring about modifications in peptide-protein binding in blood plasma and subsequent alterations in the bioactivity of a wide range of protein and peptide bioregulators, neuropeptides, hormones, and peptide growth factors in malignant growth [60,70,71]. One of the in vitro macromolecular targets for the modification by aminoglycoconjugates is crz-macroglobulin [8,60,71-731. The results of investigations on the structure and biogenesis of glycoamines reveal a unique capacity of these com-

3LOOD SERUM

BLOOD SERUM FRACTIONS(BSF) >lOOKD

300

pounds for nonenzymatic modification of the peripheral part of the molecule by way of incorporation of new amino acid and carbohydrate units [8,60,68,69,74,75]. The analysis of the structural-function interrelations using chemical, biochemical, spectroscopic, and biological methods indicate that amino acids and carbohydrates are the main structural-functional determinants of glycoamines. Thus, the structural studies on natural glycoamines and their synthetic analogs, along with the identification of the structural-functional determinants of amino-glycoconjugates, provide the chemical basis for the hypothesis of humoral molecular ‘imprints’ [b 9,60,74,75]. Investigation of the biological properties of glycoamines reveal the carbohydrate component as the key structural-functional determinant. One of the main reactions of the biogenesis of glycoamines is the nonenzymatic interaction between the carbohydrate aldehydes and a-amino groups of amino acids [60,74,75]. Investigation of the biological action of the glycoamines as a humoral factor of cancer pathogenesis revealed their possible role in metastasis. We suggest that glycoamines may be involved in this process through a modification of cell association and aggregation properties of tumor cells [6-91. We have observed that the addition of glycoamines of low molecular mass strongly inhibits tumor cell aggregation in vitro by over 95%, thus bringing about dramatic changes in the adhesion properties of tumor cells and their disaggregation from multicellular aggregates (Fig. 1). However, high molecular mass blood serum proteins (over 100 kDa) stimulates

100 kDa increases the number of tumor cell colonies in the lungs more than 2-fold; whereas the fractions with a molecular purified

glycoamines

n

CONTROL

Fig. 2. Effect of combined

n

administration

dine on the metastatic

GCrCOAMINES

+ SPERMIDINE

of glycoamines

and spermi-

spreading.

mass < 10 kDa as well as HPLCconsiderably

inhibit

the formation

of lung metastases (over 90%) and remove the stimulation of metastasizing by the high molecular mass fraction of blood serum [69]. We suggest that recognition among identical types of cell which provides a basis for sorting one homotypic population out of a heterotypic assemblage of cells plays a major role in the metastatic spreading of tumor cells. This mechanism could provide a basis for the formation of multicellular aggregates by tumor cells in the circulatory channels of the tumor host and their subsequent arrest in blood vessels of the target organs [69]. Further, in vivo and in vitro investigations on glycoamines have revealed suppression of immunological response to tumor antigens. Thus, it is possible that glycoamines may be a humoral factor contributing to tumor-associated immunosuppression. The stimulation of spheroid formation in suspension, induced by serum, proved to be caused by immunoglobulins and is probably associated with reactions of the antigen-antibody type, and as we mentioned earlier, the glycoamine fraction in blood serum with a molecular mass below 10 kDa inhibits the in vitro re-association of lymphocytemacrophage aggregates isolated from the spleen of tumor-bearing animals. These data, in conjunction with the analysis of the glycoamine structures, enabled us to formulate the hypothesis of a monovalent mechanism of anti-adhesive and immunosuppressive actions attributed to aminoglycoconjugates of < 10 kDa [69]. Amino acids and carbohydrates have been identified as the main structural-functional determinants of the glycoamines. The functional identification of carbohydrate determinants was performed using the hemagglutination inhibition test, the peptide-protein binding assay, lectin chromatography, and in a test of the inhibition of aggregate formation by tumor cells [71]. The inhibition of tumor cells aggregate formation in vitro by glycoamines could be prevented by influence of temperature and/or various types of hydrolysis (Fig. 3), may be mediated by

synthetic glycoesters (Fig. 3), and involved galactose as a terminal carbohydrate residue (Fig. 3). We have investigated the effect of combined administration of glycoamines and spermidine on the metastatic spreading of Lewis carcinoma under the ablation of primary tumor. Administration of glycoamines and spermidine stimulated lung metastases 3-fold in mice, mainly because of the increase in the number of metastases. We have shown the stimulation of metastatic spreading after the administration of glycoamines and spermidine starts in the presence of primary tumor. The effect disappears when the administration starts 1 day after the ablation of primary tumor. Thus, the glycoamine and spermidine-dependent stimulation of metastatic spreading is associated with the activation of tumor cell dissemination from the primary focus (Refs. 8, 60; and Fig. 2). We have noted (Glinsky, G. et al., unpublished data) that human malignant breast tumors have a significantly higher capacity for formation of multicellular aggregates in vitro as compared with benign breast tumors, and this observation has a good correlation with prognosis of the disease. These observations strongly support the concept that glycoamines, other serum glycomacromolecules and serum carbohydrate-binding proteins (e.g. lectins, anticarbohydrate antibodies) could modify cell recognition, association and aggregation, and consequently, metastasis and immunological response. Further in vivo and in vitro investigations on glycoamines have revealed a suppression of immunological response to tumor antigens. Thus, it is possible that glycoamines may be a humoral factor contributing to tumor-associated immunosuppression. Since the glycoamines are able to block the cell adhesion process at various - both specific and nonspecific - stages, they can evidently be used as a new instrument in the pathogenetic antimetastatic cancer therapy. In this context of unquestionable theoretical and

156

100

90 6 80

01

02

03

ml

70 60 50 -

40

2

!

30

3

1

20

I

10

INFLUENCE INHIBITION SERUM

I-10

5

7 -

I

AND

OF ONCOSPHEROID

EXPERIMENTAL 1

I

OF TEMPERATURE

FRACTION

9

4

VARIOUS

8

TYPES

FORMATION

6Y

IL

OF HYDROLYSIS

GLYCOAMINES

1

ON THE

(BLOOD

kDa)

3

CONDITIONS:

HYDROLYSIS

WITH 0.1 N HCI

2. HYDROLYSIS

WITH

2N

3

WITH 4 N HCI

AT 37’C

HCI ATlO5’C

FOR

16 HOURS;

FOR

16

HOURS;

FOR

16

HOURS;

4

I

HYDROLYSIS

4. HYDROLYSIS

WITH

loo/.

5

INCUBATION

AT 37’c

6

INCUBATION

AT 105’C

6. BLOOD

FRACTION

(FETAL

MODIFICATION

FOR 16HOURS;

IN VITRO 1. HPLC

FOR 16 HOURS,

FRACTIONI-

9. CONTROL

AT 37’C

FOR 16 HOURS;

7. BLOOD SERUM SERUM

AT105’C

NH40H

CALF

kDa BELOW

WITHOUT 1 kDo

INCUBATION;

WITHOUT

INCUBATION,

control

GLYCOAMINES;

(l-10

5. FRACTION

(OVER

4

J

FORMATION

BY: GRADE

2. FRACTION

3.1

SERUM)

OF ONCOAGGREGATE

kDa) OF BLOOD

SERUM;

100 kDa) OF BLOOD

SERUM;

+ 5 2f5

I

4

_

_

1

INFLUENCE STRUCTURAL

ON ONCOSPHEROID ANALOGS

OF

FORMATION

IN VITRO

INFLUENCE

GLYCOAMINES

FORMATION 1. 6-0-G-a-MET-glucopyranoslde

hydrobromide;

2. 6-0-A-2.14-o-triG-d-Met-D-glucopyronos~de 3

2.3-0-diA-d-Met-D-glucopyronoslde

4.

6-0-A-d-Met-D-glucopyranoside

2

512

32

OF SYNTHETIC

Initlol

OFLACTOSE(l)ANDMAtTOSE(Z)ON

ONCOSPHEROID

IN VITRO

concentration

ot

sugars

- 23mM

hydrobromide; hydrobromlde hydrobromide

Fig. 3. Modification

of oncoaggregate

practical interest would be the elucidation of molecular determinants responsible for homotypic selection of tumor, cells during tumor progression and structural identification of glycoamines and other serum factors involved in this process.

formation

in vitro by glycoamines.

V. BGA-related glycoepitopes as key structural determinants for cell aggregation in relation to cancer and immunogenesis Two important

conclusions

can be drawn

from the

157

experiments on the investigation of the effect of glycoamines and fractions of blood serum on the cell aggregation in vitro and on metastasizing. First, the formation of oncoaggregates and reassociation of immunoaggregates occurs with the involvement of common or similar structural determinant, probably of carbohydrate nature. Second, the humoral (serum) constituents (glycoamines and high molecular mass factors of blood serum) that are able to recognize these structural determinants are present not only in the cancer serum, but also (to a lesser extent) in the normal blood serum. Such structural determinants participating in the homotypic aggregation of histogenetically different types of cells may be, in particular, the carbohydrate determinants of the blood group antigen (BGA)-related glycoantigens, since normal blood serum contains anticarbohydrate antibodies (including those against BGA-related glycoepitopes), and preliminary data indicate the involvement of BGA-related glycoepitopes in the inhibition of hemagglutination by glycoamines. At present, most of the clinically used humoral tumor markers are glycomacromolecules (CEA, mucins, Ca 19-9, Ca 15-3, CA 26); and many of them contain the BGA-related glycoepitopes (Lewis family, T, Tn, I) [76 8 I] and reveal immunosuppressive properties [82]. Blood serum from most healthy individuals contains natural anticarbohydrate antibodies, including those against the BGA-related glycoepitopes (anti-A, B, H, 0, I, and Lewis family; anti-Tn-antibodies; antiasialo GM2 ([GalNac/?( l-+Gal/3(14)Glc-Cer]) [83]. Blood serum from oncological patients reveals a decrease in the titers of anti-T, anti-Tn, anti-I, and anti-Forssman anticarbohydrate antibodies [58,76,84]. Finally, the most characteristic manifestation of aberrant glycosylation of cancer cells is neo-synthesis (or ectopic synthesis), the synthesis of incompatible antigens and incomplete synthesis (with or without the accumulation of precursors) of the BGA-related glycoepitopes [85-871. BGA-related glycoepitopes are directly involved in the homotypic (tumor cells, embryonal cells) and heterotypic (tumor cells-normal cells) formation of cellular aggregates (e.g., Lewis X antigens; H antigens, polylactosamine sequences; and T- and Tn antigens), which was demonstrated in different experimental systems [55-591. BGArelated alterations in the tissue glycosylation pattern are detected in benign (premalignant) tumors with high risk of malignant transformation, in primary malignant tumors, and in metastases [87-931, i.e., they were demonstrated as typical alterations in different stages of tumor progression. Lymphoid cells (lymphocytes, granulocytes, monocytes) are Lewis family glycoepitope-positive and contain T antigen [3,94,95] (demasked after treatment with neuraminidase).

Based on these facts, we have suggested [6-91 that the proceses of thymic education and antigen presentation are accompanied by certain changes in the glycosylation pattern of the cellular membranes. These changes include the presentation of the BGA-related glycoepitopes _ in particular, of their cryptic forms. Aberrant glycosylation may occur as a result of binding of foreign peptides to MHC and subsequent changes in glycosylation pattern of complexes peptide-MHC and/or as a result of presentation on the cell membrane of peptide-MHC complexes and exposition of new previously masked glycodeterminants. A number of experimental data are now available that confirm this suggestion for antigen presentation. Effective antigen presentation occurs in cells expressing an increased number of histocompatibility antigens with decreased sialylation level [96]. The cells infected with AIDS virus or cytomegalovirus start to express a high level of LeY and LeX glycoepitopes, respectively [97,98]. This relatively nonspecific mechanism of the presentation of BGA-related glycoepitopes may be involved in the process of homotypic sorting of the cells at subsequent stages of formation of immune response. In view of this, natural anticarbohydrate antibodies may fulfill a double function: to accomplish the complement-dependent lysis of cells containing highdensity BGA-related glycoepitopes (e.g., in the thymic education it may lead to negative selection of T cells whose major histocompatibility complexes have a high affinity for their own presented antigens) and to favor the homotypic selection of cells revealing a medium or low density level of the BGA-related glycoepitopes. The recent study [99] shows that the binding of TCR to thymic MHC molecules rescues CD4 + 8 + cells from programmed cell death and induces first up regulation of the TCR level and then differentiation into CD4 + 8 - or CD4 - 8 + cells in the absence of any cell division. The proposed sequence of positive selection of class I MHC-restricted CD4 - 8 + T cells in the thymus (Huesmann, M. et al., 1991) can be combined with our concept as follows. Immature CD4+ 8 + thymocytes normally express a large variety of TCRs and are BGArelated glycoepitopes positive. Some of the CD4 + 8 + cells TCRs of which can be bind to class I MHC ligands on thymic epithelium are rescued from cell death and selected to mature. Thus, nondividing CD4+8 + cells with high TCR levels and masked BGA-related glycodeterminants (possibly, by sialylation) are formed. They then become nondividing CD4- 8+ TCRh’sh mature thymocytes with no expression of BGA-related glycoepitopes on their cell surface. On the stage of negative selection, T cells with high affinity receptors to self peptides-self MHC molecules become BGA-related glycoepitopes positive, for example, as a result of

158

transfer of self peptides from cell surface of antigen presenting cells into T cells and subsequent association of self peptides with MHC molecules of corresponding T cells. Thus, the T cells with high, medium, and low level of affinity of self peptide-self MHC complexes will have a corresponding level of association of self peptides to their own MHC and correspondingly, high, medium, and low level of BGA-related glycoepitopes expression on their cell surface. Subsequently, T cells with high density of BGA-related glycoepitopes on cell surface, as an immature thymocytes, will be a target for elimination. This ‘self peptides pick-up mechanism’ by T cells provides one of the possible explanations of how antigen-presenting function could participate in negative selection during thymic education and, probably, in positive selection. For example, association of TCRs with ligands could protect the former from proteolytic degradation and lead to TCR up-regulation. Subsequent upregulation of MHC could be a second checking and selective signal for T cells with high affinity to MHC-self peptide complexes. Similar mechanisms may be realized for tumor cells, but these interactions should also involve humoral glycomacromomolecules that contain BGA-related glycoepitopes. It seems very intriguing to suggest that a similar mechanism is involved in the development of tumor progression and determines both its direction and the selection of the metastasizing cell in cancer. Indirect confirmation of this idea is provided by two groups of facts: the preservation at all stages of tumor progression (from premalignant benign tumors with high risk of malignant transformation to metastatic foci) of the characteristic of primary malignant tumor alterations in the glycosylation pattern of BGA-related glycoepitopes of cellular membranes [87-931 and the accumulation during tumor progression of tumor cells with a low density level of histocompatibility antigens on the cellular membrane [IOO]. VI. Aberrant glycosylation, pbenotypic divergence of tumor cells, and tumor progression

Aberrant glycosylation of cell-membrane macromolecules is one of the universal phenotypic attributes of malignant tumors. A rather limited number of molecular probes based on monoclonal anticarbohydrate antibodies now enables the detection of over 9046 of the most widespread forms of human cancer [86,87], whereas the use of most informative kits of molecular probes for oncogene detection enables the detection of no more than 21-30s of human cancer [101,102]. It can be assumed that the concept of the oncogene that has significantly contributed to the understanding of molecular mecha-

nisms of the control of cell proliferation and the autonomyzation of tumor-cell proliferation is, on the whole, of rather limited importance for the elucidation of mechanisms of the formation of other major biological properties of malignant cells. The concept of aberrant glycosylation undoubtedly is of more universal importance for the understanding of molecular mechanisms of the malignant behavior of transformed cells in vivo, not only in the aspect of a classical approach to the knowledge of cellular and biological mechanisms of the formation of invasive and implantation properties, cell-adhesion disturbances, and the like, but also for the elucidation of mechanisms of distant or systemic action of malignant tumor on the homeostasis of host organism. A number of theoretical and experimental preconditions implies that a malignant tumor is able to alter the glycosylation pattern of glycomacromolecules in the extracellular medium - in particular, in blood serum. At least partly, these alterations may be characterized as the formation of humoral molecular imprints of glycomacromolecular antigens expressed on the membranes of tumor cells. One of the mechanisms of this process is evidently ‘shedding’ of glycomacromolecules from the membrane of tumor cells [103]. In view of the wide molecular range of biomolecules (from low molecular mass glycoamines to high molecular mass glycoproteins and multimolecular complexes) involved in these alterations and of the key role of the involved determinants in the development and maintenance of homeostasis, one can suggest that this mechanism is responsible for various distant (systemic) metabolic and biochemical effects of malignant tumors on the organs and tissues of the host organism. This stage of the biochemical generalization of cancer evidently may be characterized as the early or first phase of the morphological generalization of the pathological process that is typical of cancer-metastasis development. When considering the causes of aberrant glycosylation of the glycomacromolecules of tumor-cell membranes, one usually mentions the alterations in the glycosyltransferase activity (e.g., the activation of glycosyltransferase V) and, more seldom, the elevation in the glycosidase activity. In view of the important role of the primary structure of the protein carrier of carbohydrates in the determination of glycosylation specificity, it can be suggested that alterations in the expression of such glycocarrier molecules (e.g., ectopic in site and/or time expression) are also one of the causes of the changes in the glycosylation pattern of the membrane of tumor cells. In functional and biological aspects, the alterations in the glycosylation of glycomacromolecules of tumor-cell membranes can be related to the changes that reflect the autonomyzation and enhancement of cell di-

159

vision and also to the changes associated with blocking the realization of the development and differentiation program. The alterations associated with the autonomyzation of cell division of transformed cells may be both the consequence of enhanced cell proliferation (since the entry of cells into mitotic cycle is accompanied by certain changes in the glycosylation of cellular membranes, and in tumor tissue the fraction of dividing cells is many orders of magnitude higher than that in the normal tissue) and the causal and/or supporting factors of the autonomous proliferation (e.g., the characteristic of transformation decrease in the level of GMs-glycolipids of cellular membranes facilitates the dimerization of the EGF-receptor molecules [104]. However, in the biological aspects it seems to us most important that a part of the alterations in the glycosylation of glycomacromolecules of cellular membranes of tumor cells is associated with the antigen-presentation function of nonself peptides - a universal common biological function of all nucleus-containing cells of higher mammals and humans. Mutations may create some variation in self peptides, but errors in gene expression, particularly translation, have been suggested as a possible source of major variability [105]. The rate of misincorporation of amino acids, as estimated in bacterial and in vitro systems, ranges from 10e4 to lop3 per amino acid [106]. Premature termination, a particular type of error, occurs in vivo in almost one-third of the 1000 amino acid long Escherichia coli /?-galactosidase molecules [ 1071.In the absence of any special mechanism, a cell displaying some lo6 MHC molecules on its surface would expose, at any time, a few hundred or a thousand distinct erroneous peptides. Their accumulation with time and with the large number of cells might develop a relatively large diversity: a total in the the order of lo7 variants could be generated [105]. The unavoidable noise of errors in gene expression, particularly translation, would be a major driving force in the generation of erroneous peptides in tumor cells. In association with antigen-presentation function this mechanism may provide one the major causes of phenotypic divergence of tumor cells because association of erroneous peptides with MHC-molecules would be accompanied by aberrant glycosylation of MHCglycoprotein-peptide complexes. In addition to the elevated (or even unchanged, but in combination with the increase in cell multiplication) genomic and mutation mutability of tumor cells, the antigen-presentation function and subsequent change in cellular surface with a definite type of the modification of the glycosylation pattern of cellular membrane glycoepitopes may provide the key cause of the phenotypic mutability and divergence of tumor cells. These phenotypic alterations

evidently may provide the immunoselective basis for tumor progression with a rather strict determination in tumoral disease of the selection direction, the latter largely depending on the predisposition of immunoselective factors of the host organism. Thus, the progression of tumor and the formation of complete malignant phenotype, including the metastatic ability, may be represented as the consequence of a monopathogenetic immunoselective process with the following stages: 1. The autonomyzation of the proliferation of tumor cells; 2. Genomic and mutation mutability, and alterations in the primary structure of proteins; 3. Aberrant glycosylation of membrane glycomacromolecules and phenotypic divergence of tumor cells; 4. The formation of extracellular humoral ‘molecular imprints’ of the glycoepitopes of glycomacromolecular antigens of cellular membranes of the tumor; 5. The immunoselection of tumor-cell clones with a low or medium density of membrane glycoepitopes with the involvement of the competition for common structural determinants of the immunocompetent cells and humoral factors (anticarbohydrate antibodies, soluble glycomacromolecules, and glycoamines); and 6. The formation of mechanisms of immunoresistance and homotypic association of tumor cells. It seems very promising to consider as the key role the function in these processes of natural anticarbohydrate antibodies, glycoamines and other serum glycomacromolecules, and BGA-related glycoepitopes. VII. BGA-related glycoantigens in embryogenesis, organogenesis and cell differentiation

Glycoantigens undergo dramatic changes during decell differentiation and maturation velopment, [85,108,109]. We will only consider one example that clearly shows appearance, disappearance and reappearance of BGA-related glycoantigens on the cell surface at definite stages of cell differentiation during embryogenesis, organogenesis, tissue formation and maturation. Cancer cells frequently express carbohydrate emantigens since they undergo a retrobryonic differentiation process during the course of malignant transformation [ 1lo]. SSEA- 1 (stage-specific embryonic antigen-l) first described by Solter and Knowles in 1978 [ill], is one of the embryonic antigens, which is specifically expressed on the murine preimplantation embryos at the morula stage. The SSEA-1 antigen is described chemically as a carbohydrate antigen carrying Lex-hapten and i-antigenic structures [ 112,113]. By using the specific monoclonal antibodies, it was shown that the SSEA-1 and its modified form of the antigen,

160

including the sialylated form (sialylated LeXor sialylated Le”-i) and fucosylated forms (Ley and poly Le’), are frequently accumulated in human cancer of various origins [114-l 181.Lung adenocarcinoma is one of the good sources of the SSEA-1 and its modified forms of the antigen [ 1191, and murine monoclonal antibodies directed to lung cancer cells frequently turned out to recognize SSEA-1 and related antigens [120,121]. Moreover, among the modified forms of the SSEA-1, sialylated SSEA-1 has been shown to be present in the sera of patients with adenocarcinoma of the lung, as well as in cancer cells [117,122,123]. Since SSEA-1 was first described as an embryonic antigen and subsequently found in various human cancers, the antigen has sometimes been regarded as an oncofetal or oncodevelopmental antigen [89,124-1271. The localization of three carbohydrate antigens, Lex, Ley, and sialylated Lex-i which are closely related to stage-specific embryonic antigen 1 in the lung of developing human embryos was investigated using specific monoclonal antibodies [ 1lo]. The three antigens all have a physiological significance as stage-specific developmental antigens of the human lung; those antigens were specifically present in the bud cells at each important step of the morphogenesis of the human lung, such as cells in the lung buds, bronchial buds, and terminal buds for the formation of the alveolus, and cells differentiating into bronchial gland cells. The three antigens gradually disappear in the later stage of development along with the maturation process of the lung. Stagespecific embryonic antigen 1 and related antigens are known to be associated with various human cancers, including lung cancer. Miyake et al. [l lo] suggest that the expression of these antigens in the lung cancer cells is the result of the retrodifferentiation of the cancer cells to the stages of immature embryonic lung cells. It was known that SSEA-1 antigens reappear in a certain lineage of cells after implantation during the course of morphogenesis of certain organs such as brain and urogenital organs [126,128], and it was shown the stagespecific expression of LeX and related antigens in developing lungs of human embryos [l lo]. One type of expression of the stage-specific antigen is characterized by their strict specificity and disappearance of the antigen in mature cells. This expression of SSEA-l-related antigen during the course of organogenesis has been regarded as the second wave of the expression of those antigens [l lo] assuming the appearance of these antigens in the preimplantation embryos as classically described [l 1l] as the first wave of the expression. The third type of the expression of stage-specific antigen is characterized by this persistency of the antigen after differentiation. This expression of the SSEA-l-related antigens has

been regarded as the third and the last wave of the expression of these antigens in intrauterine life [ 1lo]. VIII. Tumor cell aggregation properties and serum macromolecules in cancer metastasis: a model for development of a new family of antimetastatic VIII-A.

drugs

T and Tn antigens

Springer and his co-workers have found that the blood group precursor antigens T and Tn are carcinoma associated [58]. They have shown them to be unmasked on the external surface membranes of most primary carcinomas and their metastases. In all other tissues, T and Tn antigens are masked and are generally precursors in normal complex carbohydrate chains. T and Tn have not been detected in benign tumors nor in other diseased or healthy tissues [129-1411. Although T and Tn antigens are present in most tissues, they are, however, occluded by covering carbohydrate structures [ 142,143]. This is valid throughout life [144-1471; however, there is some evidence that T and Tn are stage-specific differentiation antigens expressed transiently on human fetuses [148]. All humans have pre-existing antibodies against T and Tn [144-1471. In the presence of complement human anti-T antibodies kill carcinoma cells in vitro [149]; obviously, they do not kill carcinomas in vivo, at least not those that become clinically manifested. Anti-T IgM constitutes 7-14s of total IgM and its concentration was about 2.5- and j-times higher than that of anti-T IgG and anti-T IgA, respectively [150, 1511. T and Tn antigens are expressed in immunoreactive form in about 90% of carcinoma tissues [58]. T antigen has been found in all colon and in 96% of breast carcinomas examined [ 13 l-l 351. Springer et al. found T antigen in 95% of 144 fresh signal samples of all types of primary carcinomas from all major organs and all seven metastatic lesions from four patients up to 6.5 years after removal of their T- and Tn-positive primary breast carcinoma strongly expressed these antigens [ 129,141,1521541. Tn antigen occurred in carcinomas about as frequently as T. Both of these antigens have been detected on the outer cell membranes of 25 of 26 human breast carcinoma-derived epithelial lines cell [134,135,152]. The authenticity of the T- and Tn-active structures have been demonstrated immunochemically in antibody inhibition studies with specific haptens [155-l 581; by a solid-phase immunoassay [ 150,15 11;and biochemically with glycosidases [ 1551and with glycosyltransferases [159-l 6 11. Thus, the unmasked T- and Tnspecific epitopes are unique carcinoma markers, particularly for breast and lung carcinomas. T, Tn and sialosyl-

161

Tn structures have been identified as human tumor-associated carbohydrate antigens [58,85,87,92,95,1621661. The expression of these antigens is highly specific to a variety of human cancers, and is essentially absent in normal tissues [85,87,92,166]. Elevated serum levels of T-antigen in patients with breast cancer has been observed [ 1671 and monoclonal anti-T antibody has been successfully used for in vivo radioimmunoimaging of primary and metastatic carcinoma both in animal models and human cancer, particularly to localize in vivo metastatic cancer during of an extended phase I clinical trials [ 1671. VIII-B.

Galcl (l-3)

Gal epitope

Anti-Gal antibody is a natural IgG antibody present in unusually large amounts in the serum of healthy individuals and constitutes 1% of circulating IgG [ 1681. Anti-Gal was found to be a polyclonal monospecific antibody interacting specifically with an oligosaccharide residue with the structure Galcc(l-3)Galj$l_4)GlcNacR [ 169,170]. Anti-Gal is produced throughout life in humans as a result of what seems to be a constant antigenic stimulation by gastrointestinal bacteria [171]. A striking reciprocal evolutionary pattern in the distribution of anti-Gal and the Gala( l-3)Galj?( l+)GlcNac-R residue was observed in mammals [172,173]. The Gal&( l3)Gal/?( 14)GlcNac-R epitope is abundant on various nucleated cells and red blood cells from nonprimate mammals, prosimians and New World monkeys, however, their expression is diminished on cells of Old World monkeys, apes, and humans. In contrast, antiGal antibody is present in the serum of Old World monkeys, apes and man, but this antibody is absent from the blood of New World monkeys and nonprimate animals [172,173]. Expression of Galcc( l-3)Gal cell surface epitopes has been correlated with the metastatic potential of murine tumor cells. It has been demonstrated that there are more Galc$l-3)Gal residues in metastatic murine tumor cells than in normal or nonmetastatic or low-metastatic tumor cells [174-1771. Gala(l-3)Gal residues are expressed at the cell surface of malignant human cancer cells, including four cell lines and 50% of the malignant breast specimens obtained by aspiration biopsy [ 1781. In contrast, all benign breast biopsies and normal cells were Galcl(l-3)Gal negative. Affinity-purified anti-Gal antibody significantly inhibited tumor cell attachment in two in vitro assays: attachment to perfused human umbilial vein endothelium, and attachment to isolated laminin [ 1781.Since anti-Gal antibody is a natural component of all human sera, it has been suggested that it may be part of the natural antitumor defense system in hu-

mans [178]. However, these antibodies obviously do not protect against clinically manifested breast carcinoma, since at least 50% of malignant breast specimens are expressed Galcc(l-3)Gal epitope. VIII-C.

Lewis,family

of glycoantigens

Stage-specific embryonic antigen 1 and related antigens are known to be associated with various human cancers, including lung cancer. Cancer cells frequently express carbohydrate embryonic antigens since they undergo a retrodifferentiation process during the course of malignant transformation [ 1lo]. SSEA-1 (stage-specific embryonic antigen-l) first described by Solter and Knowles in 1978 [l 111, is one of the embryonic antigens which is specifically expressed on the murine preimplantation embryos at the morula stage. The SSEA-1 antigen is described chemically as a carbohydrate antigen carrying Lex-hapten and i-antigenic structures [ 112,113]. By using the specific monoclonal antibodies, it was shown that the SSEA-1 and its modified form of the antigen, including the sialylated form (sialylated LeXor sialylated Le”-i) and fucosylated forms (LeY and poly Le”), are frequently accumulated in human cancer of various origins [114-l 181. Lung adenocarcinoma is one of the best sources of the SSEA- 1 and its modified forms of the antigen [ 1191,and murine monoclonal antibodies directed to lung cancer cells frequently turned out to recognize SSEA-1 and related antigens [120,121]. Moreover, among the modified forms of the SSEA-1, sialylated SSEA-1 has been shown to be present in the sera of patients with adenocarcinoma of the lung, as well as in cancer cells [117.122,123]. Since SSEA-1 was first described as an embryonic antigen and subsequently found in various human cancers, the antigen has sometimes been regarded as an oncofetal or oncodevelopmental antigen [ 1241271. Miyake et al. [l lo] suggest that the expression of these antigens in the lung cancer cells is the result of the retrodifferentiation of the cancer cells to the stages of immature embryonic lung cells. The widespread expression of LeY antigen in epithelial cancers, particularly of nonsecretor individuals, has a therapeutic potential [94] since these individuals do not synthesize LeB and LeY antigens in normal tissues or have only low levels in a few specific sites. Thus, the expression of these antigens is highly tumor-restricted in nonsecretor individuals. Thus, the homotypic and heterotypic aggregation properties of tumor cells represent important biological features of malignancy because these properties of transformed cells could determine metastatic ability of neoplastic tumors. The key structural determinants of the tumor cells that participate in cell recognition. asso-

162

ciation, and aggregation have been determined as carbohydrates and carbohydrate-binding proteins. Specifically, the BGA-related glycoepitopes are directly involved in the homotypic (tumor cells, embryonal cells) and heterotypic (tumor cells-normal cells) formation of cellular aggregates (e.g., Lewis X antigens; H antigens, polylactosamine sequences; and T and Tn antigens), which was demonstrated in different experimental systems. Tumor cell aggregation properties are particularly important for the formation of tumor cell aggregates in the blood stream and their subsequent arrest in the blood vessels of target organs, since the lodgment, attachment, and growth of blood-borne neoplastic cells depend largely on embolization. The tumor cell aggregation properties and consequently, BGA-related glycodeterminants may be also involved in specific cell-cell adhesions required for invasion and extravasation of cancer cells during metastasis. The high, specific inhibition activity of tumor cell adhesion to hepatocytes by T- and Tn-specific glycoconjugates supports this suggestion 1581. In summary, different molecular isoforms of glycomacromolecules in serum contain cell-originated glycodeterminants which are involved in cell recognition, association and aggregation. These glycomacromolecules and serum carbohydrate-binding proteins (e.g., lectins, anticarbohydrate antibodies, etc.) play an opposite role as inhibitors (low molecular weight glycoamines) and/or stimulators (high molecular weight multiglycoepitopecontaining serum macromolecules and naturally occurring anticarbohydrate antibodies) of cell aggregation. Human cancer is accompanied by changes in quantitative and qualitative characteristics of these serum macromolecules, generally in the direction of excessive accumulation of the activities that support tumor cell aggregation in blood vessels and metastasis development. These conditions specifically are being changed by surgical elimination of primary tumor because as we have shown [60,61] in the early period after surgery (2-3 days) that the level of glycoamines (the major cell aggregation and metastasis inhibitory component in serum) decrease dramatically in the serum of oncological patients. The changes in the levels of stimulatory macromolecules are supposedly less radical since part of them (anticarbohydrate antibodies) are present permanently in serum of all healthy individuals [83] and it is well known that high molecular weight proteins in serum have a relatively long half-life time. Thus, our primary objective will be development and preclinical evaluation of a new family of antimetastatic drugs with biospecificity for glycodeterminants of tumor cell aggregation. Natural and synthetic compounds will be studied with

emphasis on the application of this treatment strategy as an additional therapy after surgical elimination of primary tumor and preventive therapy for local and/or metastatic recurrence. This family of potential antimetastatic drugs will be composed of several synthetic low molecular weight compounds which contain different core structures (di- and polylysine, amino acids-glucose copolymers, etc.) and appropriate glycoepitopes as a peripheral structure of the effector molecules. Some of these compounds will have potential as a selective ‘neoplastic drug delivery on target carrier’ molecules since the expression of definite glycoepitopes have been shown as unique for malignant tumors [58]. The advantage of using these types of molecules in comparison with anticarbohydrate monoclonal antibodies and/or lectins is the absence of nonspecific sterical interference and less possibility for development of secondary immunological reaction. Another advantage could be their expected short half-lifes that provide a good basis for effective pharmacological control of the serum levels of effector molecules. This potential pharmacodynamic and other (see above) advantages will be especially important in relation to the expected therapeutic use of these compounds during the relatively long period of course of patient management. In conclusion, initial observations have shown the effectiveness of these types of synthetic and natural molecules as inhibitors of in vitro cell aggregation and potential therapeutic agents in vivo [69,55,56,60,71,74,179-1851. Effective inhibition of an experimental metastasis has been shown (see above) by monoclonal antibodies against peptide cell adhesion determinants as well as against lectins that participate in tumor cell aggregation. However, the disadvantages of secondary immunological reaction development and the nonspecific nature of these bioaddressing molecules for tumor cells have blocked their potential therapeutic use. Cell aggregation glycodeterminants have never been a target for antimetastatic drug development. However, at least some of them have a unique design for anticancer drug potential: they have been detected in from 75% to more that 90% of human breast and lung carcinomas, and never been detected in normal healthy tissues [58,95]. The use of monoclonal antibodies and/or active immunization against these glycodeterminants obviously has very limited potential, since, as we already have mentioned, serum of healthy individuals contain naturally occurring anticarbohydrate antibodies of the same specificity [83,58] and immunoglobulins, particularly IgM, do not readily permeate blood vessels and have no access to extravascular tumor cells. The synthetic compounds which we propose to design as tools in anticancer therapy are essentially free from these limitations. Thus, this con-

163

cept contains ments:

two very

(i) the tumor

important

practical

cell aggregation

new ele-

glycodeterminants

as a target structure for antimetastatic drug development; and (ii) a new family of low molecular weight synthetic compounds with biospecificity for glycodeterminants of tumor cell aggregation metastatic cancer therapy.

as a tool

for anti-

IX. Conclusion: a prospect of anticarbohydrate

to the presented

have common which involve eralization

concept,

vaccine

and represent

of pathological

domain

of CD4. This conclu-

sion has been supported by analysis of the effects of mutations in CD4 on the interactions of CD4 with gp 120 and MHC class II: while the binding sites are separable because some mutations affect the gp 120 binding site

tions [ 196.1971. A serological

similarity

between

gp 120

of HIV and class II MHC has been reported [198,199]. Due to these structural similarities, after HIV infection

cancer and AIDS

critical stages of disease BGA-related glycoepitopes

tural determinants

sites for gp 120 and class II MHC overlap in the

but not the class II MHC binding site and vice versa, there are also some mutations that affect both interac-

therapy for treatment of AIDS and cancer According

binding

first immunoglobulin-like

pathogenesis as key struc-

the mechanism

process throughout

of genthe body.

In the case of AIDS, the effector part of this mechanism is presented by HIV glycoproteins; in the case of cancer, it is presented by cancer cells and serum glycomacromolecules. Since the postulated critical role of BGA-related glycoepitopes in cell differentiation during tissue remodeling, repair, regeneration and immunogenesis in both cases - AIDS and cancer - homeostasis will be affected and destroyed, immunosuppression and immunodeficiency will be developed. Like embryogenesis. tissue regeneration, repair and remodeling could be accompanied by expression of BGA-related glycoepitopes on definite stage of cell differentiation. In this particular stage of differentiation cells that are expressed of BGArelated glycoepitopes become a target of HIV-infection. During the course of HIV infection and AIDS development HIV could use a receptor-driving mechanism, e.g., change the predominant on the first stage of infection protein-protein type of ligand-receptor interaction on the carbohydrate-protein or charbohydrateecarbohydrate types of virus ligand-target cell receptor interaction on the late stage of infection. HIV could use this mechanism for dramatic increased of spectrum of target cells and tissues and for generalization of infection throughout the host body. Molecular basis for this hypothesis is provided by data that shows similarities between class II MHC and HIV proteins. Sequence homologies have been reported between a conserved region of gp 120 and MHC class II [186], and also between Nef of HIV and MHC class II [187]. Ziegler and Stites [188] and Andrieu et al. [189] have suggested that gp 120 could resemble class II MHC and the immune response to gp 120 could cross-react with class II MHC since the envelope glycoprotein gp 120 is complementary to CD4 [190] and CD4 is complementary to class II MHC [ 1911. Recombinant gp 120 can block the interaction of CD4 with MHC class II [192-1941 as can antibodies that recognize the gp 120 binding site of CD4 [195]. Thus, the

of primarily coproteins

target cells it is very possible become glycosylated

ilar) to the host cell glycoproteins, MHC

molecules.

Consequently,

identically

that HIV gly(or very sim-

in particular HIV

class II

glycoproteins

could be (after being glycosylated by host cell glycosyltransferases) identical to host cells unique pattern of BGA-related glycoepitopes and, subsequently, HIV could use these highly carbohydrate structural units for expansion of infection. The idea that MHC molecules are the carriers for BGA-related glycoepitopes provides a molecular basis for a structural link between the ABO and HLA systems. Obviously, the presented concept provides a theoretical basis for application of anticarbohydrate vaccine therapy for treatment of cancer and AIDS. The efficiency of an experimental cancer anticarbohydrate vaccine therapy has been reported [179,180] and possibility of its application for AIDS treatment has been suggested [200]. However, the strategy of anticarbohydrate vaccine therapy that can be developed on the basis of our concept is quite different. In the case of cancer it should be antimetastatic vaccine therapy only after surgical removal of primary tumor. In the case of AIDS it should be postexposure anticarbohydrate vaccine therapy. In both cases this kind of therapy supposes to be immunosuppressive because it will affect the immune cells recognition, association and communication. However, we believe that in both cases it will be appropriate since patients will need this therapy transitorily and it will prevent the generaiization of the disease. In the case of AIDS, this treatment will block HIV infection spreading from primarily infected cells throughout the body, e.g., brain. lung, liver, etc. In cancer, it will prevent a metastasis through blocking of tumor cell aggregation in the blood stream and it will affect immunoselection of the immunoresistant and highly metastatic clones of malignant cells during tumor progression. The paradoxical efficiency of immunosuppressive therapy for treatment of experimental immunodeficiency syndrome has been shown recently (see above). It seems more promising to us to use as a tool of anticarbohydrate vaccine therapy not antibodies and/or lectins, but glycoepitopes them-

164

selves linked to the appropriate low molecular weight carrier for clasterization and increasing of bioaffinity. The effectiveness of these types of synthetic and natural molecules as an inhibitor of cell aggregation in vitro and potential therapeutic agent in vivo has recently been suggested and shown [69,55,56,60,71,74,179-1851.

35 Nowak

TP, Kobiler

602&6030,

D, Roe1 LE, Barondes

36 Zalik SE, Thompson

LW, Ledsham

37 Cook GM, Zalik SE, Milos N, Scott VA. J Cell Sci 38: 293-304, 1979. 38 Drickamer

K. J Biol Chem, 263: 9557-9560,1988. FL, Chesterton

40 Monsigny

CJ. FEBS Lett 122: 157-165,

41 Raz A, Lotan

R. Cancer

USA 83: 2248-2252,

Metastasis

Oriol R, LePendu Clausen

18: 39-62,

J, Mollicone

H, Hakomori

R. VOX Sang 51: 161-171,

SI. VOX Sang 56: I-20,

Eggest I, Fenderson

44 Moutsatsos

1981.

B, Toyokuni

N, Hakomori

SI. J Biol Chem 264: 20159-20162,

Glinsky

GV. J Tumor

Marker

Glinsky

GV. In: Molecular

Oncol.

stract Book, Moscow, 8 Glinsky

Oncodevelopment

XVIIIth

tional Society for Oncodevelopmental

12. Clifton; 9 Glinsky

USSR, September

The Humana

Marker

Oncol.,

11 Sutherland

23-27, Cancer

Ab-

Chapter

W. J Cancer

Res Clin Oncol

RM. Science 240: 177-184.

13 Nicolson

GL. Exp Cell Res 150: 3322, 1984.

14 Roos E. Biochim

Biophys

15 Fidler IJ. Eur J Cancer LA, Kleinetman

1984.

1973. Res 36: 8899894,

PA. J Nat1 Cancer

Inst 64: 1173-l 178,

R, Feizi T. Biochem Biophys

Res

48 Gabius

H-J, Engelhardt

R, Cramer

F. Anticancer

Res 6: 573-

578,1986. SH. Proc Nat1 Acad Sci USA 83: 760337607,

1986. L, Carmi P, Karakash

EMBO J 3: 29792983,

R, Lotan

D, Lotan

R.

1984.

L, Lotan

52 Raz A, Avivi A, Pazarini

R. Cancer

Res 46: 3667-3672,

1986.

G, Garmi P. Exp Cell Res 173: 109-l 16,

L, Zvibel I, Lotan

R. Int J Cancer

39: 353-

54 Raz A, Carmi P, Pazarini

G. Cancer

55 Fenderson

PW, Nudelman

BA, Andrews

56 Fenderson 1990.

Res 48: 645649,

S, Sundberg

H, Hako-

1987.

BA, Eddy EM, Hakomori

57 Lindenberg

1988.

E, Clausen

SI. Rev Articles

12: 173376,

SJ, Lundblad

A. Reprod

K, Kimber

Fert 83: 1499158, 1988.

1980. 19 Winkelhake

JL, Nicolson

GL. J Natl Cancer

lnst 56: 285291,

1976. 20 Fidler IJ. Nat New Bio1242: 21 Nicolson

GL, Brunson

1488149, 1973.

KW, Fidler IJ. Cancer

Res 38: 4105-4111,

1978. 22 Fidler IJ, Bucana C. Cancer Res 37: 394553956, 1977. 23 Gasic GL, Gasic TB, Galanti N, Johnson T, Murphy Cancer

11: 706718,

24 Nicolson

GL,

JL. Nature

(Lond.)

58 Springer

GF. Science 224: 1198-1206,

59 Springer

GF, Cheinsong-Popov

255: 23&232,

62 Glinsky

26 Damsky

C, McLellan

W, Fidler IJ. Nature

(Lond.)

284:

YuL, Kukhar Russian).

KA, Buck, CA. In: Ivatt,

R. ed. Biology

of Glycoproteins. New York: Academic Press; l-64, 1983. 27 Lotan R, Raz A. Cancer Res 43: 2088-2093, 1983. 28 Lotan R. Lotan D, Raz A. Cancer Res 45: 4349-4353, 1985. 29 Lotan R, Cavvalero D, Lotan D, Raz A. Cancer Res 49: 1261-

33 Briles EB, Gregory 5288537,

P, Kornfeld

S. J Cell Biol 81:

1979.

RF,

1580-1589.

W, Fletcher

1986.

Haywood-Reid 1984.

PL, Barondes

SH. J Cell Biol 98:

1980.

VV, Surmilo

VP. Proc Acad

GV, Vinnitsky GV, Birkmayer

Osadchaya 65 Glinsky Tumor 66 Glinsky

NI, Shvachko

LP, Radavsky

Sci USSR 228: 495499,

VB. Exp Oncol

9: 78-80,

1986 (in

1987 (in Rus-

LP, Ivanova GV,

Marker

Surmilo Onco15:

GD, Vinnitskaya AB. Tumor NI,

AB, Koritskaya

Marker

Liventsov

Onco15:

LN,

249, 1990.

VV, Sidorenko

MV.

J

250, 1990.

GV. Proc Acad Sci Ukr SSR, B series 3: 69-72,

1988 (in

Russian). 67 Glinsky

1268, 1989. 30 Meromsky L, Lotan R, Raz A. Cancer Res 46: 527&5275, 31 Raz A, Lotan R. Cancer Res 41: 3642-3647, 1981. 32 Barondes SH. Science (Wash. D.C.), 23: 1259-1264, 1984.

V, Desoi PR,

sian). 64 Glinsky

CH, Knudsen

68-71,

GV, Glinsky

63 Glinsky

1975.

1984.

R, Schirrmacher

Tegtmeyer H. J Biol Chem 258: 5702-5706, 1983, 60 Glinsky GV. J Tumor Marker Onco14: 193-221, 1989, 61 Glinsky GV, Nikolaev VG, Ivanova AB, Shemchuk AS, Lisetsky VA. Exp Oncol2:

S. Int J

1973.

Winkelhake

25 Raz A, Bucana 3633364, 1980.

34 Cerra

SJ, Thorpe

mori SI. Dev Biol 122: 21-27,

1983.

J, Saidel GM. Cancer

J, Orr W, Ward

Bio18: 218-229,

127: 68&686,1983.

53 Raz A, Meromsky

1982.

1976. 18 Varani

SH. Science (Wash.

360, 1987.

Acta 738: 2633284,

16 Glaves D. Br J Cancer 48: 6655673,

RP, Barondes

1987.

Acta 695: 113-176,

9: 223-227,

1987.

1988.

GL. Biochim

Biophys

113: 101-122,

Nat1

1988. C, Piver MS et al. Tumor

SR, Thorpe

Commun

51 Raz A, Meromsky

5:231, 1990.

JL. Proc

1987.

50 Raz A, Meromsky NM. J Tumor

M, Wang

1987.

D, Mecham

239: 1539-1541,

49 Gitt MA, Barondes

1990; 7. Markers.

RA, Berezhnaya

12 Nicolson

17 Liotta

of the Interna-

Press, in press.

GV. Semyonova-Kobzar

IO Mueller-Kleiser

meeting

pro-

Biology and Medicine.

GV. In Sell S (ed.) Serological

DC.)

47 Carding

5: 206, 1990.

Oncology.

applications.

1989.

M, Schindler

Acad Sci USA 84: 645226456,

46 Allen HJ, Karakousis

1989.

Kojima

teins and clinical

M, Hako-

1987.

G et al. Gene 60: 1977204, 1987.

IK, Wade

45 Hinek A, Wren

1989.

T, Dean B, Stroud

mori SI. J Biol Chem 264: 94769484,

1986.

Rev 6: 433452,

SH, Jesse1 TM. Proc Nat1 Acad Sci

1986.

43 Jia S, Mee R, Morford SI. Semin Hematol

1980.

M, Kieda C, Roche AC. Biol Cell 47: 95%110,1983.

42 Regan LJ, Dodd J, Barondes

Hakomori

IM. J Cell Sci 88: 483493,

1987.

39 Harrison

References

SH. J Biol Chem 252:

1977.

GV. Proc Acad Sci Ukr SSR. B series 4: 4349,

1988.

68 Glinsky GV, Linetsky MD, Shemyakin VV, Markin VV, LiventSOYVV, Ivanova AB. Tumor Marker Oncol 5: 107-l 18, 1990. 69 Glinsky GV, Linetsky MD. Tumor Marker Oncol 5: 13760, 1990. 70 Glinsky

GV. Tumor

71 Glinsky

GV, Linetsky

Marker MD,

Oncol

1: 249-294,

Ivanova

1987.

AB, Osadchaya

LP. Se_

165

myonova-Kobzar

RA, Surmilo

NI. J Tumor

Marker

Oncol

5:

72 Glinsky Tumor

GV,

Surgova

Marker

73 Surgova

TM,

Oncol5:

TM, Glinsky GV,

Marker

VB. J

MD et al. Proc Acad Sci

1, 1988 (in Russian). AS, Shilin

VV et al. J Tumor

VV et al. J Tumor

Rev 6,541-565,

Marker

PS, Brown

glycosilation

in tumor

cells. New

14365-14369,

JL, Horn

B, Brockhaus

D. Cancer

Res 48:

2, Koprowski

V. Cancer

Hakomori

Y, Ueda I, Koba-

Res 49: 344553451, SI, Levine

86 Hakomori

ChL,

glycosilation

123, 92296.

SI. Marcus

DM

cells. New York: A.R. Liss.

1988.2077212. 87 Hakomori

88 Abe K, Hakomori

Res 52: 257-331,

SI, Ohshiba

S. Cancer

SH, Shi ZR, Kim YS. Histochem SH, Yuan M. Gukushi

Res 46: 2639-2644.

Res 46: 59765984,

Res 46: 26277

A, Kim YS. Cancer

M, Montgomery

CK, et al. Cancer

Res 49:

SH, et al. Cancer

Res 46: 5985-

Altered

glycosylation

ChL, Hakomori in tumor

cells. New York:

W, Goldenberg

LJ, Knowles

SL, Spitalnik

120 Cuttita

F, Rosen

Immunol

LJ, Boes J, Ploegh

19: 537-542, M, Hayami

DM. Cancer

Res 49:

HL, Melief CJM.

Eur J

E, Fenderson

BA. et al. J Exp Med 169:

M, Scott B, Kisielow P, von Boehmer

JL, Cuttitta

SRJD, Ginsburg

220: 3 18, 1983. N. Galton

J. Hirota

M. Katz D.

1985.

Y, Kitahara

A, et al. Int J Cancer 41: 344349.

PW,

Knowles

Goodfellow

PN.

BB. Int J Cancer SG, Marder

RJ, Minna

Rosen ST. J Histochem D. Dev Biol83:

Shevinsky

29: 523-531.

JD. Mulshine

Cytochem

Bronson

DL.

JL, Polovina

MR.

32: 982-988.

I. Martinez-Hernandez 391-398,

LH,

1982. 1984.

A, Knowles

BB, Soher

1981.

I. Knowles

BB, Solter D. Cancer

M, Boyer AM, Schwarting

Res 43: 669-

GA. Proc Nat1 Acad Sci

1985.

129 Springer

GF, Desai PR. Transplant

Proc 9: 1105, 1977.

130 Springer

GF, Desai PR, Banatwala

I. J Nat1 Cancer

DR, Taylor

37: 142. 1980.

Inst 54: 335,

1975. CR. Oncology

132 Boland CR, Montgomery

CK. Kim YS. Proc Nat1 Acad Sci USA

133 Orntoft

TE. Mors NPO.

H. Cell 66: 533-

134 Anglin

Eriksen

Res 45: 447452,

1985.

JH

MP,

Jr,

Lerner

G. Jacobsen

Nordquist

NO, Poulsen

RE.

Nature

HS.

(Land.)

2691254. 1977. K, Yoshioka

6: 3599380,

T, Bieberich

C, Jay G. Annu Rev Immunol

135 Laurent

JC, Noel P, Faucon

M. Carcinogenesis

7: 1037-1042,

ER, Hamilton

137 Coon

1986.

SR, et al. Nature

327: 293-297,

29: 260. 1978.

J, Weinstein

J, Watanabe

Lloyd KO. Cancer

T, Tokumaru

Res 49: 745-752,

T, Tagaki 1989.

H, Nakazato

H,

RS, Summers

J. Am J Clin Path01

77: 692.

1982. 138 Summers

1987.

deleted.

M. Biomedicine

136 Limas C, Lange P. Cancer 46: 1366. 1980.

1988.

102 Bos JL, Fearon

104 Reference

JD. Proc Nat1 Acad Sci

1988.

Cancer

540,1991.

103 Sakamoto

Res 46: 4751

79: 2051. 1982. PW, Gonczol

99 Huesmann

101 Barbacid

D. Kim YS. Cancer

C, et al. Cancer

AF, Minna

PI, Suyama

Res 45: 435437,

131 Howard N, et al. J Exp Med 167: 323%

1347-I 360, 1989.

100 Tanaka

BB. Solter

M, Magnani

Biophys

122 Chia D, Terasaki

128 Yamamoto

331, 1988. 98 Andrews

Res 46: 2619-

1981.

LC, Brockhaus

AR Liss, 1988;

1989.

M, Kashiwagi

BG,

R. J Biol Chem

T, et al. Cancer

PF, Dubois

S, Gazdar

USA 78: 45914595,

678, 1983.

1989.

96 Booy CJP, Needjes

E, Levery SB, Kannagi

Y, Tachikawa

USA 82: 304553049,

95 Stein R. Chen S, Grossman

T, Kulander

1985.

Res 44: 1142-I 147, 1984.

DM (eds.).

SI, Markus

2355243.

97 Adachi

R. Fukushi

127 Fox N, Damjanov

5992, 1986. 94 Lloyd KO. In: Reading

S, Shepard

1984.

126 Fox N, Damjanov

1977204, 1989. M, Itzkowitz

S. J Biol Chem

SI. J Biol Chem 258: I1793-

R, Hakomori

S. Nudelman

259: 46814685.

125 Combs

1986.

SH, Yuan

93 Kim YS, Yuan

KM, Hakomori

SingerJW,CancerRes45:3711-3717,

124 Andrews

SH, Yuan M, Ferrell LD, Palekar

92 Itzkowitz

E, Levery SB, Hakomori

1982.

123 Zenita K. Kirihata

J 18: 155-163,1986.

Y, et al. Cancer

2632, 1986. 91 Itzkowitz

BB, Solter D, Evans MJ.

11797, 1983.

Cancer

89 Itzkowitz

A, Knowles

1981.

R, Nudelman

V. Arch Biochem

1989.

1986. 90 Itzkowitz

1983.

R. Cancer

1988.

257: 14865-14874,

121 Huang

SI. Adv Cancer

Y, Kannagi

4755. 1986.

1985.

Hakomori

in tumor

Inst 71: 231-251,

0, Okada

BB. Proc Nat1 Acad Sci USA 75: 5565-5569,

118 Shi ZR. McIntyre

1989.

P. J Immunol

Res 45: 24052414,

SI. In: Reading

(eds.). Altered

1986,

1978.

R. J Nat1 Cancer

292: 156158,

119 Spitalnik SI. Cancer

& Hall,

2626, 1986.

1984.

1979. 85 Hakomori

32-37,

Nature

117 Kannagi

1987.

3 1:385-397,

83 Lloyd KO, Old LJ. Cancer WW,

H. Ginsburg

M, Matsuoka

Res 47: 3451-3459,

TB et al. eds. Accu-

Chapman

1985.

K, Tanaka

Res 48: 7150-7158,

116 Hakomori

K, Kuroki

AA. Neoplasma

84 Young

257:

1983.

K, Totani

ta A. Cancer

SI, Kannagi M, Zenita

115 Fukushi Y, Kannagi

Res. 43: 548995492,

82 Hakim

M, et al. J Biol Chem

1982.

.I, Steplewski

81 Yamashita

109 Hakomori

314: 53-57,

114 Abe K, McKibbin

J. Nilsson

80 Magnani

JL. J Mol Biol 125: 407432.

113 Kannagi

JP, Magnani

213%2148,1988. 79 Magnani

107 Manley

112 Gooi HC, Feizi T, Kapadia

ChL, Hakomori

York: A.R. Liss, 1988; 279-294. 78 Linsley

London:

1978.

1987.

SI, Imura H. In: Reading

DM (eds.), Altered

1989.

H. In: Kirkwood

Processes.

111 Solter D, Knowles

Metastasis

R. Hakomori

racy in Molecular

110 Miyake

MD, Liventsov

RH, Grosjean

108 Feizi T. Nature

1990.

76 Dube VE. Cancer SI, Marcus

Vinnitsky

250, 1990.

119-136,

77 Kannagi

P, Claverie JM. Cell 56: 3277329,

83-126.

GV, Sidorenko

GV, Linetsky

Oncol5:

MV,

1990.

Ogorodniychuk

Oncol5:

75 Glinsky

Sidorenko

161-66,

Ukr SSR, B series 5: 79-8 74 Glinsky

105 Kurilsky

IO6 Buckingham

250, 1990.

139 Lehman

JL et al. Cancer TP. Cooper

140 Ghazizadeh

M. Kagawa

1127-l 130, 1984.

Res 43: 934, 1983.

HS. Mulholland

SG. Cancer

S. Izumi K. Kurokawa

53: 272, 1984. K. J Ural

132:

141 Springer

GF.

Murthy

MS, Desai

PR. Scanlon

EF. Cancer

45:

142 Springer

GF, Ansell NJ. Proc Nat1 Acad Sci USA 44: 182, 1958.

143 Lloyd KO, Kabat 144 Friedenreich

EA. Proc Nat1 Acad Sci USA 61: 1470, 1968.

V. The Thomsen

Copenhagen:

Hemagglutination

Levin & Munksgaard,

145 Burnet FM, Anderson J, Moullec

i48 Springer

J, Bernard

P, Waksman

150 Springer

MS, Tegtmeyer

Fluids. Oxford:

Pergamon.

EF.

K. eds. Pro-

69: 346, 1982.

G. In: Peeters H, ed. Protides

GF, Taylor

H, Scanlon

1979; Vol. 29,42-96.

GF, Desai PR. Naturwissenschaften

of the Biological

1984; Vol. 3 1.421424.

CR, Howard

DR, et al. Cancer

cology.

New York: Praeger,

154 Goodale

RL, Springer

GL, eds. Cellular

1983; Vol. 1,99-l

GF, Shearen

OnH.

Biophys

Res Commun

61: 470.

G. Immunitaetforsch

Exp Ther

130: 88,

1966. 158 Bray

GF, Desai PR. Carbohydr

J, Lemieux

RU,

TA. J Immunol

126: 1966,

1981. GF,

Desai

wissenschaften

PR, Schachter

H, Narasimhan

161 Desai PR, Springer

92:

II I, 1978.

H, Yamada

SI. Cancer

M. Hakomori

H, Metoki R, Hakomori

6: 403, 1979.

BM. Cancer

Res 51: 1406-1411,

1991. PO, Natoli

EJ, Calves MJ, Stockert

E, Oettgen

Old LJ. Proc Nat1 Acad Sci USA 84: 291 I-2915, PO, Ritter

7045-7050,

G, Srivastava

HF.

1987.

P, et al. Cancer

Res 49:

1989. T. Dean B, Hakomori

Lett 3 1: 2673-

SI. Tetrahedon

2676, 1990. MM, Bugianesi

RL. Robbins

JC, Doebber

TW, Schen

TY. J Med Chem 24: 1388895, 1981. BA, Zehavi

U. Hakomori

S. J Exp Med 160: 1591-

1596, 1984. 333: 215. 1988.

I88 Ziegler JL, Stites DP. Clin Immunol 189 Andrieu

345: 26, 1990.

Immunopathol41:

305-313,

JM, Even P, Venet A. AIDS Res 2: 1633174, 1986. LA, Nakamura

191 Gay D, Maddon

Y, Hakomori

S.

1985.

S. Cancer

192 Diamond

G, Smith

DH,

et al. Cell 50: 9755985,

P, Sekaly R, et al. Nature

DC, Sleckman

141: 3715-3717. Res 48: 43614367,

193 Rosenstein

1988.

328: 626629,

BP, Gregory

T, Lasky

SJ, Herrmann

SH. J Immunol

1987.

LA. J Immunol

1989.

YS, Burakoff

144: 52&

531.1990.

164 Kurosaka

A, Kitagawa

H, Fukui S, et al. J Biol Chem 263: 8724

8726, 1988. T, Clausen

H, Hirohashi

Res 48: 22142220,

VG, Schlom D. Cancer

167 Longenecker

J, Paterson

Res 46: 850-857,

BM, MacLean

ChL, Hakomori

SI, Markus

S, Ogawa

T, Iijima H, Hako-

1988.

KJ. Lyerly HK, Stanley SD, Austin AA, Matthews

AJ, Bennett

J, Magnani

JH,

1986.

GD, McEwan

195 Sattenau

AJ, et al. In: Reading

DM, eds. Altered

Glycosylation

in

196 Clayton 197 Lamarre Capon 198 Golding

EA, Peleg A, et al. J Exp Med 160: 1519-

1531, 1984.

170 Galili U, Buehler 1987.

J, Shohet

SB, et al. J Exp Med 165: 6933704,

RE, Hamadeh

AG, Weiss RA, Beverley

PCL. Science

LK. Sieh M, Pious DA, Reinherz

EL. Nature

339: 548-

RM et al. Infect

Immun

56:

D, Ashkenazi

A, Fleury

DJ. Science 245: 734746,

S, Smith

DH,

Sekaly

RP.

1989.

H, Robey FA, Gates FT, et al. J Exp Med 167: 914923,

1988. 199 Gelding

BA, Buehler J, et al. J Exp Med 162: 573-582,

1988.

QJ, Dalgleish

TJ.

142: 3091-3097,1989.

551, 1989.

169 Galili U, Macher 1985.

171 Galili U, Mandrel1

DP. J Immunol

234: 112&l 123. 1986.

tumor cells. New York: AR Liss, 1988; 3077320. 168 Galili U, Rachmilewitz

173G-1737.

194 Weinhold Bolognesi

mori S. Cancer Colcher

R, Ma-

1989.

1987.

T, Shimosato

Proc Nat1 Acad Sci USA 82: 703997043,

166 Johnson

Inst 812: 212-216,

180 Singhal

A, Fohn

1987.

JM, Lambotte

RR, Longenecker

190 Lasky

GF. J Immunogenet

S, Clausen

163 Takahashi

S. Natur-

63: 488, 1976.

JP, et al. Eur J Biochem

162 Hirohashi

165 Kjelden

B, Foidart

1986.

159 Springer 160 Carton

Inst 71:

JP. Exp Cell Res 155: 3455358,

187 Vega MA. Guigo R, Smith TF. Nature

Res 40: 183, 1975.

McPherson

Biol6: 99-l 14. 1985.

179 Fung PYS, Madej M, Koganty Res 50: 430843 14, 1990.

186 Young JAT. Nature

157 Springer

J. Tumor

M, et al. J Nat1 Cancer

J, McCoy

hieu Ph. J Nat1 Cancer

185 Fenderson

1974. Z, Uhlenbruck

IG, Varani

EJ III, Wicha

V, Colin C, Parent

184 Ponpipom

GF, Desai PR. Biochem

SB, et al. Proc Nat1 Acad Sci USA

IA, Bosnes V. Int J Cancer 40: 5055510,

183 Tokyokuni

30.

JG, Desai PR. Tegtmeyer

J Surg Res 35: 293, 1983. 155 Springer 156 Kim

177 Grimstad

182 Livingston GF, et al. In: Moloy PJ, Nicolson

MR, Shohet 1987.

IA, Varani

181 Livingston 55: 561-569.

1985. 153 Springer

263: 1755-

1984.

DW. Fed Proc Fed Am Sot

BH, de Week AL, Ishizaka

Base]: Karger,

151 Desai P, Springer 152 Springer

E, et al. J Biol Chem

1983.

178 Gastronovo

GF, Desai PR, Murthy

In: Kallos

SB, Kobrin

J, Lovet

176 Grimstad

110: 165, 1953.

6, 1984.

gress in Allergy.

84: 1369-1373,

1281-1287,

J. Blood 14: 1079, 1959.

H, Cromer

173 Galili U, Clark

175 Varani

SG. J Exp Biol Med 25: 213, 1947.

GF, Tegtmeyer

Exp Biol43: 149 Springer

U, Shohet

174 McCoy JP, Goldstein

Phenomenon.

1930.

146 Caselitz FH, Stein G. Immunitaetsforsch 147 Dausset

172 Galili

1762. 1988.

2949, 1980.

H, Shearer

1435, 1989. 200 Hansen JES, Clausen 1990.

GM, Hillman

K, et al. J Clin Invest 83: 143&

H. Nielsen C, et al. J Virol 64: 2833-2840.