Biological functions of melanoma-associated antigens

Abstract

To date, dozens of melanoma-associated antigens (MAGEs) have been identified and classified into 2 subgroups, I and II. Subgroup I consists of antigens which expression is generally restricted to tumor or germ cells, also named as cancer/testis (CT) antigen. Proteins and peptides derived from some of these antigens have been utilized in promising clinical trials of immunotherapies for gastrointestinal carcinoma, esophageal carcinoma, pulmonary carcinoma and so on. Various MAGE family members play important physiological and pathological roles during embryogenesis, germ cell genesis, apoptosis, etc. However, little is known regarding the role of MAGE family members in cell activities. It is reasonable to speculate that the genes for subgroup I MAGEs, which play important roles during embryogenesis, could be later deactivated by a genetic mechanism such as methylation. In the case of tumor formation, these genes are reactivated and the resultant proteins may be recognized and attacked by the immune system. Thus, the subgroup I MAGEs may play important roles in the immune surveillance of certain tumor types. Here, we review the classifications of MAGE family genes and what is known of their biological functions.

INTRODUCTION

In 1991, researchers first isolated a melanoma-associated antigen (MAGE) gene, MAGE-A1[1]. This antigen, isolated from an MZ-2 human melanoma cell line, could be recognized by cytotoxic T lymphocytes (CTLs). In the following years, dozens of new MAGE gene were identified. MAGE proteins act as anti-tumoral immune targets, making these antigens a popular focus of immunotherapy researches on gastrointestinal carcinoma and other cancers[2]. Although much work has investigated the expression of MAGE genes and HLA-restricted epitopes, relatively little is known about the functions of the MAGE proteins. In this paper, we review the classifications of MAGE family genes and what is known of their biological functions.

CLASSIFICATIONS OF THE HUMAN MAGE PROTEIN FAMILY

In 1991, Terry Boon’s laboratory established a method for identifying tumor antigens based on tumor-specific CTLs recognition[1]. Using this method, the group identified a new tumor antigen which they called MAGE-1. Subsequently, this and other screening methods were used to identify a large number of tumor-specific and tumor-related antigens, including dozens of MAGEs (Table 1)[3-7]. The MAGE family proteins share certain homologous regions, including the MAGE homology domain (MHD). Sequence comparison and analysis revealed 3 subgroups of acidic MAGEs, termed A, B, and C, and one basic subgroup, MAGE-D, which includes Necdin, Restin and others[8]. Based on expression patterns, the MAGEs were further classified as belonging to either subgroup I or II. Members of subgroup I, including MAGE-A, -B, and -C, are expressed in malignant tumors and testis, but not in other normal tissues. These members are also named as cancer/testis (CT) antigen and tumor-specific antigen. In contrast, subgroup II MAGEs are expressed in various normal adult human tissues[4,9]. Interestingly, testis germ cells do not express MHC I/II molecules and cannot present the MAGE proteins, so testis tissue is generally immune-exempt. Based on this, tumor-expressed MAGE-A, -B, and -C proteins have become important targets for cancer immunotherapy, and some clinical trials are ongoing for treating gastrointestinal carcinoma, esophageal carcinoma, pulmonary carcinoma and so on. Now researchers are continuing to identify additional MAGE genes in the hope of identifying better therapeutic targets.

Table 1.

The MAGE family genes

Subtype	Gene	Access	Expression	Length of
	name	number	status	amino acids
MAGE-A	MAGE-A1	JC2358[27-29]	T	280
	MAGE-A2	AAH13098[30]	T	314
	MAGE-A3	AAH00340[30]	T	314
	MAGE-A4	BAA06841[31]	T	317
	MAGE-A5	NP_064402[32]	T	320
	MAGE-A6	AAA68875[5]	T	314
	MAGE-A7	NP_064677[33]	P	80
	MAGE-A8	AAH02455[30]	T	318
	MAGE-A9	P43362[5]	T	315
	MAGE-A10	AAA68869[5]	T	369
	MAGE-A11	P43364[5]	T	319
	MAGE-A12	AAF44789[34]	T	314
	MAGE-A13	U71148	P	-
	MAGE-A14	NT_011534.1	P	-
	MAGE-A15	NT_025309.1	P	-
MAGE-B	MAGE-B1	AAC23616[35]	T	347
	MAGE-B2	AAC23617[35]	T	319
	MAGE-B3	AAC23618[35]	T	346
	MAGE-B4	AAC23619[35]	T	346
	MAGE-B5	Q9BZ81[36]	T	111
	MAGE-B6	Q8N7X4[37]	T	407
	MAGE-B7	AC005297.1	P	-
	MAGE-B8	AC005297.1	P	-
	MAGE-B9	AC005297.1	P	-
	MAGE-B10	AC011693.5	T	348
	MAGE-B11	AC011693.5	P	-
	MAGE-B12	AC011693.5	P	-
	MAGE-B13	AC011693.5	P	-
	MAGE-B14	NT_025279.3	P	-
	MAGE-B15	NT_011752.1	P	-
	MAGE-B16	NT_025279.3	T	320
	MAGE-B17	NT_011766.3	T	203
MAGE-C	MAGE-C1	NP_005453[38-40]	T	1142
	MAGE-C2	NP_057333[41-42]	T	373
	MAGE-C3	AAK00358[36]	T	346
	MAGE-C4	AL023279.1	T	115
	MAGE-C5	NT_025337.2	P	-
	MAGE-C6	NT_025337.2	P	-
	MAGE-C7	NT_025337.2	P	-
MAGE-D	MAGE-D1	AAG09704[25,43]	N	778
	MAGE-D2	CAB10841[44-46]	N	606
	MAGE-D3	Q12816[47,48]	N	1387
	MAGE-D4	Q96JG8[49-52]	N	741
MAGE-E	MAGE-E1	Q96JG8[49-52]	N	741
	MAGE-E2	-	N	-
	MAGE-E3	NP_032879[53-55]	N	424
MAGE-F	MAGE-F1	NP_071432[56]	N	307
MAGE-G	MAGE-G1	AAG38607[57]	N	100
MAGE-H	MAGE-H1	AAG38608[58]	N	219
MAGE-I	MAGE-I1	NT_011638.3	P	-
	MAGE-I2	NT_011638.3	P	-
MAGE-J	MAGE-J1	NT_011577	P	-
MAGE-K	MAGE-K1	Z81311	P	-
MAGE-L2	MAGE-L2	NP_061939[59,60]	N	529
Necdin	Necdin	XM_007686[61-65]	N	321
Restin	Restin (Apr-1)	NP_054780[66-70]	N	219

T: Only expressed in tumor cells or germ cells. P: Pseudogene; N: Expressed in normal cells. -: Unknown.

CELLULAR FUNCTIONS OF MAGE PROTEINS

Regulation of subgroup I MAGE gene expression in normal somatic cells

In normal mature somatic cells, subgroup I MAGE genes are static. When cells become neoplastic, MAGE genes are activated and the corresponding proteins are expressed. Interestingly, some non-MAGE-A1-expressing tumor cells have been found to contain transcription factors capable of activating the MAGE-A1 promoter[10]. This suggests that the presence of the relevant transcription factor is not sufficient to trigger expression, indicating that alternative regulatory mechanisms may exist in terms of MAGE-1 gene activation.

Recent studies have shown that MAGE gene activation may be related to promoter demethylation[11]. The promoter of the MAGE-A1 gene contains several cis-regulatory sequences located from nt -792 to +47. Among them, the B and B’ domains are critical for MAGE-1 gene expression. Both domains contain Ets transcription factor binding sites, including a critical CpG bi-nucleotide site. When this CpG is methylated, the Ets transcription factor cannot bind to the B and B’ domains, and expression of MAGE-1 is inhibited. Therefore, promoter methylation inhibits the expression of MAGE-1[10]. Indeed, when non-MAGE-A1-expressing tumor cells were treated with the demethylating reagent, 5-aza-2-deoxycytidine (5DC), MAGE-A1 expression was induced[11]. When normally non-MAGE-expressing cells (i.e. fibroblasts) were treated with 5DC, some cells expressed MAGE genes[11], while others did not[12]. In explanation of this, the authors proposed that most normal cells possess strong methylating actions, and that demethylating reagents such as 5DC are insufficient to demethylate MAGE promoters to a degree that would allow for MAGE gene expression. Taken together, these studies suggest that both normal and tumor cells contain the transcription factors that activate subgroup I MAGEs, and that expression of these genes is regulated by promoter demethylation.

The functions of MAGE genes in germ cells

Most MAGE genes are expressed in germ cells under physiological conditions, but their functions remain unclear. Early work has focused on the functions of the mouse MAGE-b4 gene during embryonic development[13]. The mouse MAGE-b4 gene is expressed in adult and fetal reproductive gland cells and shares high homology with members of the human MAGE-B sub-type. The MAGE-b4 protein is located in the cytoplasm but not in the nucleus. In male testes, germ cells (gonocytes) continuously proliferate until they arrest at the G0/G1 stage, differentiate into foot cells and form the spermatic cords. After birth, the gonocytes differentiate into spermatogonia, which in turn undergo meiosis during adolescence and differentiate into sperm cells. The MAGE-b4 gene is highly expressed in germ cells suspended in G0/G1; in contrast, the gene is barely expressed in meiotic spermatogonia. Thus, MAGE-b4 likely plays an important role in male germ cells, perhaps by maintaining gonocytes in a non-proliferative state.

In female germ cells, MAGE-b4 is expressed prior to meiosis, and also during the pachytene and telophase portions of meiosis. Other cell cycle proteins have shown similar differences in expression in the two sexes. For instance, cyclin A1 deficiency in male germ cells inhibits meiosis but does not affect that in female germ cells[14]. Similarly, deficiencies of Hsp70-2 or A-myb suppress the development of male germ cells and male fertility but do not affect female development[15]. Therefore, we speculate that MAGE-b4 may similarly play an important role in the control of the cycle of male germ cells, while playing a less dramatic role in female germ cells.

Another member of the MAGE family Magphinin plays an important role in germ cell development. Mouse Trophinin is a membrane protein that plays an adhesive role in the process of zygotic implantation into the uterine endometrium; an alternative transcript of the Trophinin gene encodes Magphinin[16]. Although the Trophinin protein is not homologous with the MAGE family members, the mouse Magphinin protein shares high homology with Necdin, Dlxin and NRAGE. Of these, Necdin can bind transcription factor E₂F-1, which is responsible for inducing the expression of cyclin and promoting cell cycle progression from G1 to S. Thus, it is possible that the Necdin homolog, Magphinin, acts similarly to bind E₂F-1 and inhibit cell proliferation. Magphinin has three alternatively spliced forms: magphinin-α, β and γ. Northern blot analysis reveals that Magphinin protein is expressed in mouse brain, ovary, testis, and epididymis. Western blot analysis indicates that in mouse ovary and epididymis, Magphinin is derived from alternative translation of the Trophinin transcript beginning at the second start codon (AUG). The Magphinin protein sequence contains a nuclear localization signal that allows it to enter the nucleus and inhibit cell proliferation. Immunohistochemical studies suggest that the localization of Magphinin protein varies between male and female germ cells at various stages of the cell cycle. Before meiosis, Magphinin-β is mainly distributed in the cytoplasm of male germ cells. After the first meiosis, when primary spermatocytes differentiate into secondary spermatocytes, Magphinin is localized in both the cytoplasm and the nucleus, and after this nuclear translocation, cell division terminates. Based on this, we speculate that Magphinin regulates the cell cycle during the formation of male spermatocytes. In the female germ cell, the distribution of Magphinin is somewhat different. When the oocyte has only single- or double-layer vesicles, intracellular Magphinin (especially Magphinin-γ) is strictly cytoplasmic. When the oocyte has divided into multi-layer vesicles, the protein is only located in nucleus. In this case, meiosis terminates at stage G2, suggesting that Magphinin also controls the formation of the female ovum.

May these expression patterns and functional roles be generalized to other members of the MAGE family? This remains to be shown. Indeed, at present, little is known about the functions of MAGE family proteins inside germ cells. Further studies will show whether human MAGE proteins control the cell cycle in manners similar to the actions of MAGE-b4 and Magphinin.

Below, we will discuss several specific examples of MAGEs and what is known of their cellular functions.

MAGE-A4

In terms of cancer biology, yeast two-hybrid studies identified binding between the MAGE-A4 protein and a cancer protein: the gann ankyrin repeat protein (also called Gankyrin, PSMD10 or p28). This MAGE-A4-specific binding is mediated by its C-terminus[17]. Gankyrin, which consists of six gann ankyrin repeats and a 38 amino acid N-terminus, can bind the cancer-suppressing protein Rb, the S6 ATP enzyme subunit of the 26S protein, and cell cycle-dependent kinase 4 (Cdk4). Gankyrin expression is increased in the livers of hepatocellular carcinoma patients; this increase is seen at the earliest stages of tumor genesis[18,19]. Overexpression of Gankyrin can increase the phosphorylation and degradation of Rb, as well as immortalization of NIH/3T3 cells. In addition, the binding of Gankyrin to Cdk4 counteracts the cancer-inhibiting functions of p16^INK4A and p18^INK4C[20]. Therefore, Gankyrin plays an important role in controlling cell cycle during liver cancer tumorigenesis. Studies have shown that exogenous MAGE-A4 can partly inhibit the adhesion-independent growth of Gankyrin-overexpressing cells in vitro and suppress the formation of migrated tumors from these cells in nude mice. This inhibition is dependent upon binding between MAGE-A4 and Gankyrin, suggesting that interactions between Gankyrin and MAGE-A4 inhibit Gankyrin-mediated carcinogenesis[17].

Necdin

Necdin, first identified in 1991, is the best-characterized member of the MAGE protein family. The gene was isolated from P19 neural cells; the encoded protein consists of 325 amino acids and shares about 30% homology with other MAGE proteins[21]. Studies have shown that overexpression of Necdin may induce cell cycle arrest in NIH3T3 and SAOS2 cell lines, suggesting that Necdin functions in cell cycle arrest and maintenance of cell stability. In vivo, Necdin can interact with cell cycle promoting proteins such as the SV40 big-T protein, adenovirus EIA and transcription factor E₂F-1, which acts as a cell cycle regulator by trans-activating the relevant genes in an Rb-regulated pathway. The latter binds to E₂F-1 protein during G1 stage and inhibits the binding ability of E₂F-1, leading to inhibition of gene activation. Similarly, Necdin can bind to E₂F-1 and initiate conformational changes to decrease E₂F-1 binding and inhibit cell growth[22]. In addition, Necdin can bind to and inhibit p53, which normally induces cell cycle arrest and cell death[23]. Thus, Necdin plays important roles in inhibiting cell growth and apoptosis through its interactions with E₂F-1 and p53.

The human Necdin gene is localized at 15q^11-13, an area genetically associated with the neurological behavior disorder, Prader-willi Syndrome (PWS). Newborn PWS patients generally suffer respiratory failure and myasthenia, and adolescent PWS patients show psychonosema, sexual dysfunction and obesity. The PWS chromosomal region contains several genes, including Necdin and MAGE-L2. Indeed, knockouts of the mouse Necdin homolog show symptoms similar to human PWS[24], suggesting that Necdin plays a role in proper development, and that lack of Necdin may be involved in the pathogenesis of PWS.

MAGE-D1

The MAGE-D1, also named as NRAGE or Dlxin, plays important roles in mediating apoptosis and transcription. NRAGE interacts with p75^NTR, a TNF receptor responsible for binding the Trk receptor and forming a complex that facilitates binding of neurotrophin, which in turn activates the Trk receptor. p75^NTR is also capable of mediating cell death. NRAGE-p75^NTR binding, identified by yeast two-hybrid screening, occurs through an 80 amino acid intracellular segment of p75^NTR located near the plasma membrane. Overexpression of NRAGE inhibited the interaction between p75^NTR and the Trk receptor and induced cell apoptosis. In contrast, overexpression of the Trk receptor increased binding between Trk and p75^NTR, leading to inhibition of NRAGE-p75^NTR complex-induced cell death. This indicates that NRAGE-p75^NTR and Trk receptor-p75^NTR binding are mutually exclusive.

Another NRAGE-binding protein, XIAP, can inhibit apoptosis by binding to activated Caspases. The NRAGE-XIAP complex accelerates the decomposition of XIAP, suggesting that NRAGE accelerates cell death by degrading XIAP and activating Caspases[4].

Indeed, members of the MAGE-D1 subgroup can bind to several proteins that control apoptosis and cell cycle[4]. The MAGE-D1 repeat regions can directly interact with Msx2, Dlx7 and Dlx5. Msx generally acts as a transcription inhibitor, whereas Dlx acts as a transcriptional activator. In addition, Msx1 and Msx2 play roles in the control of cell cycle, and Msx2 can promote cell death. The effect of MAGE-D1 on Msx remains unknown, but MAGE-D1 is necessary for Dlx5 to promote transcription.

CONCLUSION

Although subgroup I and II MAGE genes are expressed in different tissue-specific patterns, all family members contain the MAGE homology domain (MHD), suggesting some functional conservation. Biochemical analyses of these two subgroups have provided some insights into the physiological effects of the MAGE genes. Embryonic cells have much less CpG methylation at the MAGE genes than do somatic cells[25,26]. Similarly, MAGE genes in tumor cells are hypomethylated; indeed, the entire tumor genome is generally hypomethylated[10]. Thus, it is likely that MAGE gene expression in tumor tissues is the result of tumor genesis, not a cause. It is reasonable to speculate that this family of proteins functions during embryonic development, and that the genes are subsequently deactivated, perhaps by methylation. During neoplastic transformation, these genes are re-activated, expressed, and may become antigenic targets that are recognized and attacked by the immune system[8]. Therefore, MAGE genes take part in the immune process by targeting some early tumor cells for immune destruction. Consequently, these genes should be studied further in terms of their various functions as they relate to the pathogenic mechanism of tumors, immunotherapy, and other important fields.