Glycoimmunomics of human cancer: current ... - Future Medicine

For reprint orders, please contact: [email protected]

REVIEW

Glycoimmunomics of human cancer: current concepts and future perspectives Mepur H Ravindranath†, Paul Yesowitch, Cindy Sumobay & Donald L Morton †Author

for correspondence John Wayne Cancer Institute, 2200 Santa Monica Blvd, Santa Monica, CA 90404–2302, USA Tel.: +1 310 449 5263; Fax: +1 310 449 5259; [email protected]

Keywords: antigen presentation, CD1, colon cancer, gangliosides, glycoantigens, glycoimmunomics, glycolipids, glycomics, glycoproteins, immunoglobulin M antibodies, immunosuppression, melanoma, natural killer T cells, O-linked and Nlinked oligosaccharides, ovarian cancer, prostate cancer, renal carcinoma, sarcoma, Siglec part of

Future strategies for the treatment of human cancer require a full appreciation of the intracellular and extracellular changes that accompany neoplastic transformation. The changes may involve a variety of micro- and macro-molecules, including, but not restricted to, peptides, proteins (with sugar and/or lipid moieties), oligosaccharides, glycolipids (neutral or acidic, e.g., gangliosides), ceramides, fatty acids and other lipids. Although several therapeutic approaches have been well developed in recent years, most of the reported studies focus on proteins and peptides. Glycoantigens and lipoantigens have been neglected. Elucidation of the profiles and properties of all molecules associated with tumor progression is required to develop a successful strategy to treat human cancer. This review describes the unique immunomics of tumor-associated glycoantigens and explains why the field of glycoimmunomics may yield clinically important biomarkers and treatments for the management of human cancer.

Cancer progression is a result of interactions between the neoplasm and its microenvironment. There is a correlation between tumor burden and increased serum levels of shed tumor antigens, which may remain free or form immune complexes. One or more immunosuppressive molecules released from a growing neoplasm into the circulation may lead to the progressive depression of immune functions. Any therapy that eliminates shed immunosuppressive antigens from the circulation should help restore immune competence and enable the immune system to attack the growing neoplasm. Since immunosuppression is proportional to tumor load, immunotherapy is usually less successful in patients with a large tumor burden. However, cytoreductive surgery can reduce tumor burden to a level that can be effectively treated by systemic immunotherapy [1]. A tumor is a heterogeneous colony of cells defined by their antigenic profiles. These cytoplasmic and cell-surface tumor-antigens are targets for passive and active specific immunotherapies. Heterogeneity in the expression of tumor-associated antigens within a tumor mass means that any therapy targeting a single antigen, receptor or tumor-associated molecule may not be effective against all cells of the tumor. Tumor cells that express no or low target-molecules can escape chemotherapeutic or immunological attack. A typical example is tumor stem cells, which may not express the same antigens as the other tumor cells and even, if they do, the expression may be too low for immune recognition [2]. Stem cells from primary tumors escape and metastasize to different sites, as documented in

10.2217/14796694.3.2.201 © 2007 Future Medicine Ltd ISSN 1479-6694

breast cancer [3,4]. Therefore, antigen-targeting treatments would be beneficial if they could target more than one tumor-associated antigen. The interaction between the immune system and a target-antigen depends on the nature of the antigen (protein, carbohydrate or lipid), the cell site of antigen expression (cytoplasm or cell surface), the density of the cell-bound antigen and the serum level of circulating (shed) antigen. Cellular components of the immune system include antigen-presenting cells, such as macrophages and dendritic cells, and effector cells such as T-helper, T-suppressor, T-regulatory and natural killer (NK) cells. Humoral components include B lymphocytes (CD5+ and CD5-), antigen-presenting cells, antibodies and cytokines. When the immune system targets a protein antigen, an antigen-presenting cell may engulf a tumor cell, an aggregate or a membrane fragment containing the antigen, and present the peptide antigens to T cells via major histocompatibility complex (MHC) molecules, with endogenous (exogenous, by cross-priming) peptides presented by MHC class I and exogenous peptides presented by MHC class II. Costimulation is required at this juncture; otherwise antigen presentation may lead to T-cell anergy. When a peptide antigen is presented with costimulation, the T cell that recognizes the peptide epitope may be activated and proliferate as a clone. These clones subsequently recognize the target antigen in the context of MHC class I and facilitate ‘killing’. Although tumor-associated proteins/peptides are valuable target tumor-associated antigens, the immune system can also recognize glycoantigens, though not via the same components of the Future Oncol. (2007) 3(2), 201–214

201

REVIEW – Ravindranath, Yesowitch, Sumobay & Morton

immune system. Glycoantigens are often considered T-cell independent as they do not require the help of T cells for antibody production by B cells. However, there is accruing evidence to suggest that specialized T cells may recognize glycoepitopes of glycoantigens, particularly neutral and acidic glycolipids (which include gangliosides). These specialized T cells are designated as NK T cells, which are innate lymphocytes that share receptor structures and functions with conventional T cells and NK cells. They are specific to glycolipid antigens bound by the MHC class Ilike glycoprotein CD1d. One striking property of NK T cells is their capacity to rapidly produce large amounts of cytokines in response to T-cell receptor engagement, suggesting that activated NK T cells can modulate immune responses. Recent preclinical studies have revealed a significant efficacy of NK T-cell ligands, such as the glycolipid α-galactosylceramide, in the treatment of metastatic cancers. These findings suggest that appropriate stimulation of NK T cells could be exploited for the prevention or treatment of human cancer [5]. The molecular mechanism for lipid antigen recognition involves the insertion of the lipid portion of antigens into a hydrophobic groove to form CD1–lipid complexes, which contact T-cell surface receptors (TCRs) on NK T cells. Recognition of carbohydrate epitopes is precise, and lipid-reactive T cells alter systemic immune responses in infectious and autoimmune disease. These findings provide a previously unrecognized mechanism by which the cellular immune system can recognize alterations in many types of carbohydrate structures [6]. The MHC class Ilike CD1 family of antigen-presenting molecules is responsible for the selection of NK T cells. For example, CD1a is expressed on antigen-presenting cells, such as Langerhans cells and dendritic cells, where it mediates T-cell recognition of glycolipid and lipopeptide antigens that contain either one or two alkyl chains. Structural, biochemical and biophysical studies support the view that CD1 proteins directly bind the hydrophobic alkyl portions of these antigens, and position the polar or hydrophilic head groups of bound lipids and glycolipids for highly specific interactions with T-cell antigen receptors. The ligands presented by CD1d to NK T cells or other CD1d-restricted T cells, may include glycolipids from a marine sponge, bacterial glycolipids, normal endogenous glycolipids, tumor-derived phospholipids, neutral glycolipids and gangliosides [7]. It is still not 202

Future Oncol. (2007) 3(2)

known whether the NK T cells are directly involved in the cytotoxic killing of tumor cells, as observed with peptide antigens. In this regard, the recent discovery of sialicacid binding, lectin-like receptors, termed Siglec (sialic acid-binding immunoglobulin (Ig) superfamily lectins), on immune cells, and particularly on NK cells, deserves attention, as Siglec on the surface of NK cells can recognize gangliosides on tumor cells, bind to them and mediate cytotoxicity. Siglecs are transmembrane polypeptides, with a transmembrane domain and a cytoplasmic tail. The first Ig-like domain is the most important in carbohydrate recognition and the second Ig-like domain may also contribute to the binding. Eight members of the family have been described so far in humans, and each displays highly cell-type specific expression: Siglec-1/sialoadhesin (expressed on macrophages); Siglec-2/CD22 (on B lymphocytes); Siglec-3/CD33 (on myeloid precursors and monocytes); Siglec-4a/myelin-associated glycoprotein (on oligodendroglia and Schwann cells); Siglec-5 (on neutrophils and monocytes); Siglec-6/OBBP-1 (on B lymphocytes and placental trophoblasts); Siglec-7/AIRM1 (on NK cells and monocytes); Siglec-8 (on eosinophils); Siglec-9 (on monocytes and granulocytes) and Siglec-10 (on mast cells and T- and B cells) [8,9]. Siglec-5 bound preferentially to GQ1b, but weakly to GT1b, whereas Siglec-10 interacted only with the GT1b ganglioside. Siglec-7 and Siglec-9 bound to gangliosides GD3, GQ1b and GT1b bearing a disialoside motif, although Siglec-7 was more potent. Siglec-9 also interacted with the monosialoganglioside, GM3. Siglec-8 demonstrated low affinity to the gangliosides tested compared with other Siglecs. The unusual binding preference of recombinant Siglec-7-Fc protein for α2,8-linked disialic acids of ganglioside GD3, and its specific binding to GD3 synthase-transfected P815 target cells expressing high levels of GD3, suggest that GD3 expression on target cells can modulate NK cell cytotoxicity via Siglec-7dependent and -independent mechanisms [10]. However, it remains to be seen whether the CD1 family presents gangliosides to Siglec receptors on NK cells. Siglec-9 is unique in that it recognizes the Neu5Acα2–3Galβ1–4[Fuca1–3]GlcNAc structure (i.e., recognition is not disturbed by the fructose residue on the GlcNAc) and distinguishes between Neu5Acα2–3-Galβ1–4GlcNAc (typically found in N-linked glycans) and Neu5Acα2–3Galβ1–3-GalNAc (typically found in O-linked glycans and glycosphingolipids) [8]. The functional relevance of future science group

Glycoimmunomics of human cancer: current concepts and future perspectives – REVIEW

Tumor glycomics

the polypeptide by an O-glycosidic linkage from GalNAc to the hydroxyl (-OH) of serine or threonine are mucin-type sugar chains (Figure 1). Mucintype sugars often have a Galβ1,3GalNAc-disaccharide core. Sugar chains attached to the polypeptide by an N-glycosidic linkage from GlcNAc to the amide nitrogen of aspargine are aspargine-linked glycoproteins (Figure 1). Figure 1 shows the attachment of oligosaccharides by the O-glycosidic bond between GalNAc and the -OH group of serine or threonine (or lysine or proline; O-linked). The N-linked oligosaccharides are usually connected through a GlcNAc or GalNAc to the side chain of aspargine within the protein backbone. Structural variations in O-linked and N-linked heteroglycan moieties have been reviewed [13]. Heteroglycans associated with glycolipids have their own classification and terminology, as described in detail elsewhere [14]. Glycolipid antigens may be further categorized into neutral glycolipids, fucolipids and acidic glycolipids to include sialolipids and sulfolipids containing sulphate groups [15]. Sulfated glycolipids (sulfatides) are found in several normal and abnormal tissues. Glycolipids containing neutral and acid heteroglycans are associated with several genetic abnormalities. The most common glycolipids in normal and malignant human tissues are those that contain sialic acid, specifically, gangliosides (Figure 2). Chain elongation of gangliosides involves a series of gene-specific glycosyltransferases [16]. Gangliosides on the tumor cells and frozen sections can be identified using monospecific monoclonal antibodies, presented in brackets in Figure 2.

Glycoantigens are polymeric chains of diverse sugars such as glucose, galactose, mannose, fucose, N-acetylglucosamine, N-acetylgalactosamine, N-acetylneuraminic acid and N-glycolylneuraminic acid. The antigenic epitope may vary with the nature of the sugar and the glycosidic linkages (α and β) between sugars (e.g., Galα1,3Gal, Galα1,4Gal, Galβ1,3Gal, Galβ1,4Gal, Galβ1,3-GlcNAc, Galβ1,4GlcNAc, NeuAcα2,3Gal, NeuAcα-2,6Gal and NeuAcα2,8NeuAc). The linkages between two sugars are introduced by different gene-specific glycosyltransferases. The glycoantigens may occur as free polysaccharides (e.g., polysialic acids associated with neural cell adhesion molecules) or in association with peptides/proteins (proteoglycans and glycoproteins) or lipids (glycolipids) [12]. Glycoproteins usually contain multiple sugar chains with different structures. The sugar chain of glycoproteins is classified by its linkage to the polypeptide backbone. Sugar chains attached to

There is remarkable diversity in epitopes of the glycoantigen, which forms molecular contacts with the antibody. The complementary combining site (CCS) of the antibody is the paratope. The upper limit of the epitope size is determined by variable-region CCSs of the paratope. X-ray crystallographic studies of oligosaccharide–antibody complexes demonstrate how the CCS of antibodies against glycoantigens can bind to as many as six sugar residues, whereas the CCS of antibodies against peptide antigens can only accommodate four amino acids. The contact areas involve 255 A2 of the sugar residues and 304 A2 of the paratope. A total of 15 amino acids of the antibody establish 90 van der Waal forces and nine hydrogen bonds [17]. Antigen–antibody interactions may also involve salt bridges. The structures of some antibody–antigen complexes suggest the rigid binding of both.

this selectivity of the Siglecs is unclear, but it appears that engagement of the Siglecs by polyvalent ligands may be able to elicit negative intracellular signals, thus silencing the cells that are inappropriately activated [8]. Siglecs may function in a manner similar to that of killer cell Ig-like receptors expressed on NK cells that send negative intracellular signals if engaged by MHC class I expressed on target cells [8]. The inhibitory effect of engaging Siglecs is evident from the works of Crocker’s [9,10] and Paulson’s [11] teams of investigators. Recently, Crocker and colleagues demonstrated that NK cells were less able to kill target cells bearing the ganglioside GD3, a high-affinity Siglec-7 ligand, suggesting the inhibition of NK activity by Siglec-7 [10]. Both Siglec-7 (p70/AIRM) and Siglec-9 are CD33-related Siglecs, expressed on NK cells and subsets of peripheral T cells. Similar to other inhibitory NK-cell receptors, they contain Ig receptor family tyrosine-based inhibitory motifs in their cytoplasmic domains. Sialic acid binding of Siglecs-7 and -9 is abrogated by mutations at the conserved Arg in the ligand-binding site, Arg124 for Siglec-7 and Arg120 for Siglec-9, which results in reduced inhibitory function in the nuclear factor of activated T cells (NFAT)/luciferase transcription assay, suggesting that ligand binding is required for the optimal inhibition of TCR signaling [11]. All these observations favor the view that the binding of glycoconjugates with Siglecs may lower the activation and activities of immune cells.

future science group

www.futuremedicine.com

Glycoimmunomics

203


Figure 1. N-linked and O-linked oligosaccharides observed in glycoproteins. CH2OH

OH

O

O

Serine threonine O-linked

GalNAc

Gal

O

E1–3

H2C

NH

O

NHCOCH3

Protein backbone

X

O

O Man D1–6

NH

O O CH2

E1–4

CH2OH

E1–4

O

O O

O

GlcNAc

O

HN

C

Aspargine N-linked

H2C

GlcNAc

O X

Man

CH2OH

Oligosaccharides

O Man

D1–3

NHCOCH3

NHCOCH3

O

Gal: Galactose; GalNAc: N-acetyl galactosamine; GlcNAc: N-acetyl glucosamine; Man: Mannose.

Glycoantigens stimulate primary, but not secondary B-cell responses and do not require T-cell help; hence they are T-independent antigens. There are two classes of T-independent antigens [18]. T-independent-1 glycoantigens at high concentrations stimulate the polyclonal proliferation of B cells and at low doses stimulate the division of B cells that secrete heteroglycan-specific antibodies. A rigorous response to these glycans occurs in cultured cells depleted of T cells, but can be modified by other cell types, such as monocytes or NK cells. Secondary exposure to these antigens does not result in accelerated kinetics and isotype switching typical of memory responses mediated by T cells. T-independent-2 glycoantigens stimulate antigen-specific responses in athymic nude mice, but do not stimulate polyclonal proliferation at high doses. In vitro responses may require interferon (IFN)-γ, which can also be provided by NK cells. IFN-γ may induce these T-independent-2 antigen-stimulated B cells to switch isotype production to IgG3 in animal models. The observation that glycoantigens can trigger B cells to produce IgM antibodies in T-cell-deficient mice indicates that glycoantigens are T-cell independent. Glycoantigens also fail to induce a memory response.

204


IgM antibodies directed against glycoantigens may be hexameric or polymeric [19], in contrast to conventional pentameric IgM, and may be without a J-chain [19–22]. Polymeric IgM antibodies without a J-chain appear to fix complement 20-fold more efficiently than conventional pentameric IgM [20]. Persistent IgM antibodies in patients with motor-neuron diseases are directed against carbohydrate residues of glycolipids [21–23]. The B cells sharing pan-T-cell antigen CD5 may augment the production of antiglycolipid antibodies [23]. Immunogenicity of tumor-associated carbohydrate antigens is confirmed by immortalizing B cells with Epstein–Barr virus and developing human monoclonal antibodies. Monospecificity of human, as well as murine, monoclonal antibodies against specific carbohydrate epitopes is confirmed using sensitive enzyme-linked immunosorbent assay (ELISA) [24–25]. Examining sera of healthy volunteers between the ages of 18 and 90, it was noted that antiganglioside IgM antibodies occur naturally at low levels in healthy individuals (Figure 3) [23,26]. Most interestingly, the antibody levels decline after the age of 50 years [23,26], a finding that is of tremendous significance for the majority of cancer patients, particularly prostate cancer



Figure 2. Gangliosides overexpressed in human cancers.

Neutral glycolipids

[MB 3.6] GD3

GM3

LacCer

[KM696] Gg3

[14G2a]

GM2

GD2

[GMB16] Gg4

[14G2a] O-AcGD2

[GGR12]

GM1a

GD1b

Ceramide E1,1Glc

[GMR17]

[GD1a-1]

GM1b

[GMR5]

GD1a

GT1b

E1,4Gal E1,4GalNAc E1,3Gal

[GMR17]

O-AcGD1a

[GMR17]

[GMR11]

O-AcGM1b

D2,3Sia

GT1a a-series

GQ1b b-series

SiaD2,8Sia 2,30AcSia 2,80AcSia

cells may induce serum antiganglioside IgM antibodies, clearing them from the microenvironment. A number of earlier studies reported that gangliosides are incapable of inducing an antibody response, and that induction of antiganglioside antibodies may require exogenous adjuvants. The repeated injection of gangliosides

(CaP) patients, since the incidence of cancer increases above the age of 50. The levels of antiganglioside IgM increased with tumor burden and disease stage in patients with sarcoma [27] and colorectal carcinoma [28] (Figure 3). These observations led us to hypothesize that tumor gangliosides released or shed from tumor

Figure 3. Serum titers of naturally occurring antiganglioside IgM antibodies in (A) 33 healthy volunteers and in (B) 110 patients with early or advanced colorectal carcinoma. B Colon cancer patients (all stages) (n = 110)

18,000

18,000

16,000

16,000

14,000

14,000

Serum IgM titer

Serum IgM titer

A Healthy volunteers (n = 33)

12,000 10,000 8000 6000

12,000 10,000 8000 6000

4000

4000

2000

2000 0

0 GM3 GM2 GM1 GD1a GD1b GD2 GD3 GT1b Gangliosides (3 nmol/well)

GM3 GM2 GM1 GD1a GD1b GD2 GD3 GT1b Gangliosides (3 nmol/well)

IgM: Immunoglobulin M. Reprinted from Cryobiology 45, 10–21 (2002), with permission from Elsevier.



205


in rabbits did not elicit an immune response [29]; however, antiganglioside antibodies were augmented after admixing them with foreign carrier proteins, such as pig serum [30,31], serum albumin [32,33], human erythrocyte glycoprotein [34], foreign erythrocytes [35] or a mixture of meningococcal outer membrane proteins, cationized bovine serum albumin, multiple antigenic peptides, polylysine and keyhole limpet hemocyanin [36], or with bacterial carriers such as Salmonella minnesota or Mycobacterium bovis [37]. Importantly, the glycoantigens presented in the context of an intact membrane elicited a better immune response than soluble antigens [38–40]. While the need for adjuvants for inducing an immune response to gangliosides has been extensively reviewed [40,41], it is far from clear as to how gangliosides released or shed from tumor cells can induce antiganglioside IgM without exogenous adjuvants. It is probable that tumor necrosis in a proliferating population of tumor cells may act as a natural endogenous adjuvant. To test this possibility, we determined serum levels of total gangliosides and antiganglioside antibodies before and after inducing necrosis by cryoablation in hepatic metastases from colorectal cancers [28]. As predicted, the serum total ganglioside level increased significantly after cryoablation; concomitant with this increase and without any exogenous adjuvants, there was a significant increase in the titer of antiganglioside

IgM antibodies (Figure 4). Control patients, whose hepatic metastases were resected or destroyed by radiofrequency ablation, did not exhibit significant changes in serum ganglioside or antiganglioside antibody titers. The antibody titers associated with cryoablation-induced necrosis indicate that an antiganglioside immune response does not require an exogenous adjuvant. It appears that tumor necrosis may cause endogenous adjuvants to elicit antibodies against gangliosides shed or released from the tumor. This could be a mechanism to eliminate gangliosides that are known to be immunosuppressive. Gangliosides have long been known to suppress the immune functions of T- and B cells (Figure 5) [42–53]. T cells from cancer patients are often functionally impaired, imposing a barrier to effective immunotherapy. Most pronounced are the alterations characterizing tumor-infiltrating T cells, which, in renal cell carcinomas (RCCs) include defective nuclear factor (NF)-κB activation and a heightened sensitivity to apoptosis [52,53]. Coculture of renal-tumor cell line (SK-RC-45) or RCC tissue-derived gangliosides with Jurkat T cells and peripheral blood T lymphocytes induced apoptosis of T cells. Jurkat T cells overexpress a protein, RelA, which protects the T cells against cell death. However, upon coincubation of SK-RC-45 with the T cells, there was a decrease in their RelA(p65) and p50 protein levels, which coincided with the

Figure 4. Serum ganglioside levels and IgM autoantibody titers before and after cryosurgical ablation in two representative patients (1 and 2). Patient 1

Patient 2 35

2000

25 20

1000

15 0 -30

0

10 30 60 90 120 150 180 210 240 Days post-cryoblation

Serum IgM titer

Serum IgM titer

30

6000

30

5000 4000

25

3000

20

2000

15

1000 10

0 -30

Serum ganglioside Anti-GM2 IgM Anti-GD1b IgM

0

30

60 90 120 150 Days post-cryoblation

Serum ganglioside (mg/dl)

35

Serum ganglioside (mg/dl)

3000

180

Anti-GT1b IgM Anti-GM1 IgM Anti-GM3 IgM

The abscissa is days post-cryoablation, the left ordinate is the serum IgM titers, and the right ordinate is the serum ganglioside level expressed in mg/l. Ig: Immunoglobulin. Reprinted from Cryobiology 45, 10–21 (2002), with permission from Elsevier.

206




onset of apoptosis. The disappearance of RelA/p50 protein was mediated by a caspasedependent pathway, since pretreatment of T lymphocytes with a pan caspase inhibitor prior to coculture with SK-RC-45 blocked RelA and p50 degradation. SK-RC-45-gangliosides also mediated this degradative pathway, since blocking ganglioside synthesis in SK-RC-45 cells with the glucosylceramide synthase inhibitor, PPPP, protected T cells from tumor-cell-induced RelA degradation and apoptosis. Ganglioside GM2, purified from RCC, promoted T-cell dysfunction [53]. Importantly, anti-GM2 antibody (DMF10.167.4) blocked 50–60% of the T-cell apoptosis. Furthermore, the RCC tissue-derived gangliosides also suppressed IFN-γ and interleukin-4 production by CD4+ T cells at concentrations (1 ng/ml–100 pg/ml) well below those that induce any detectable T-cell death (4–20 µg/ml). Anti-GM2 antibody also suppressed the production of IFN-γ, suggesting that RCC-derived GM2 can downregulate cytokine production by CD4+ T cells. The role of natural antibodies to gangliosides is likely the elimination of the immunosuppressive gangliosides from circulation. The natural induction of antibodies to gangliosides, as

observed after cryoablation of colon cancer metastasized to the liver [28], increase with stages of tumor progression in sarcoma [27], as well as after vaccine stage III melanoma [54]. Their findings suggest that the primary role of naturally occurring antiganglioside antibodies may be to eliminate gangliosides released or shed from tumors to prevent immunosuppression and restore immunocompetence. Glycomics & glycoimmunomics of prostate cancer

CaP progresses imperceptibly in its early stages and is potentially curable when organ confined. We hypothesized that the endogenous IgM response to CaP-associated gangliosides could serve as an early immunological event in tumorigenesis. Although GM3, GM2, GD3 and GD2 have been identified in human CaP tissue [55], the ganglioside profile of established human CaP cell lines is not known. Of the four American Type Culture Collection (ATCC) cell lines, two (LNCaP FGC and LNCaP FGC-10) are androgen receptor (AR)-positive (+) cells from lymph node metastases, and two (DU 145 and PC-3) are AR-negative (-) cells from brain and bone metastases. In addition, we have examined a

Figure 5. Hypothesis underlying tumor glycoimmunomics. Tumor-cell proliferation

Tumor cells undergo necrosis

Glycoantigens are released from necrotic tumors [27,28]

Glycoantigens induce immunosuppression

GM2 inhibits Ig production in B cells and lowers NK activity [42–44] GM3 induces IL-10 in T cells [45] GD3 lowers NK cell activity [10,46,47] GD1a induces IL-6/IL-10 in T cells [45,48] GD1b binds to IL-2 [49] Suppresses Ig production [50] Suppresses IL-4/IL-5 production [51] GT1b reduces production of IgG/IgM in humans [50] and suppresses IL-4/IL-5 production [51] Degradation of NF-NB in T cells [52,53]

Augments production of pre-existing IgM [26,27]

To clear suppressive glycoantigens from the circulation and tumor microenvironment [26,27,54]

Serve as immunomarkers for the diagnosis of early phases of cancer [55,56]

Future Oncology

Ig: Immunoglobulin; IL: Interleukin; NF-κB: Nuclear factor κB; NK: Natural killer.



207


Figure 6. Ganglioside signature of prostate cancer demonstrating the evolution of GD2 (top) and presence of GD1a in prostate cancer cell lines DU145, PC3, HH-870, LNCaP FGC and LNCaP FGC-10 (bottom). Bone Met

Brain Met

Organ-confined

14G2a PC-3 A1

A1

GD2

After base treatment

HH870

A2

GD2

GD2

B GD2 only

DU145 Clone-lgG1

GD2 + O-AcGD2 upper (A1)

DU145 Clone-lgG2a

GD2 + O-AcGD2 upper (A1) + lower (A2)

DU145 Clone-lgG2b Immunostaining of DU145 GD1a with three different mAbs

GD1a

GD1a

GD1a Staining of four cell lines with GD1a-1 mAb

GD1a-1 PC-3

LNCaP FGC

LNCaP FGC-10

HH-870

Ig: Immunoglobulin; mAbs: Monoclonal antibody. Reprinted from Biochem. Biophys. Res. Commun. 324, 154–165 (2004), with permission from Elsevier; Int. J. Cancer 116, 368–377 (2005), with permission from Wiley-Liss, Inc.

recently developed AR- cell line (HH870) from an organ-confined CaP (stage T2b), which has provided an opportunity to compare the ganglioside signatures of CaP cell lines derived from organ-confined versus metastasized AR--CaP. We used antiganglioside monoclonal antibodies, resorcinol staining and mobility assessment to characterize the ganglioside signatures of human CaP cell lines. Our objective was to determine if specific antiganglioside IgM antibodies occur in the sera of patients with organconfined CaP (stage T1/2) after screening the profile for tumor-associated gangliosides in AR+ 208


(LNCaP FGC & LNCaP FGC-10) and AR(PC-3 and DU 145 from ATCC, and HH870 from Hoag Cancer Center) CaP cell lines. Analyses of ganglioside profiles of AR (+) and AR (-) CaP cell lines reveal the presence of GM2, GD2 and its O-acetylated derivatives, and GD1a. GD1a is present in both early and advanced CaP as well as in AR+ and AR- tumors (Figure 6). We examined serum antiganglioside IgM titers in 36 patients with early CaP (stage T1/T2, which is organ-confined), 27 patients with advanced CaP (stage T3/T4), 11 patients with benign prostatic hyperplasia and 11 healthy individuals [56]. We hypothesized future science group


Table 1. Antiganglioside immunoglobulin M profiles in sera from patients with prostate cancer, patients with benign prostatic hyperplasia and healthy volunteers. Group (N)

Target of antiganglioside antibody GM1

GM2

GM3

GD2

GD3

GD1a

GD1b

GT1b

Healthy (11)

4.9 + 0.61

5.7 + 0.84

5.3 + 0.98

4.82 + 0.51

5.9 + 1.00

5.0 + 0.60

6.0 + 1.02

5.9 + 0.96

BPH (11)

4.7 + 0.42

5.5 + 1.43

4.9+0.76

5.4 + 1.31

4.6 + 0.00

4.9 + 0.61

4.9 + 0.65

6.0 + 1.00

T1/T2 CaP (36)

5.3 + 1.01

5.4 + 1.05

5.2 + 1.03

5.7 + 1.13

4.9 + 0.79

5.8 + 1.11

5.4 + 1.11

6.2 + 1.40

T3/T4 CaP (27)

5.1 + 0.85

5.4 + 1.08

4.9 + 0.73

5.5 + 1.08

4.6 + 0.00

5.0 + 0.75

5.5 + 1.31

6.3 + 1.03

P value (ANOVA)

0.154

0.820

0.493

0.152