MAGE-A4, NY-ESO-1 and SAGE mRNA expression rates and co-expression relationships in solid tumours

Background Cancer testis (CT) antigens are promising targets for cancer immunotherapies such as cancer vaccines and genetically modified adoptive T cell therapy. In this study, we evaluated the expression of three CT antigens, melanoma-associated antigen A4 (MAGE-A4), New York oesophageal squamous cell carcinoma 1 (NY-ESO-1) and sarcoma antigen gene (SAGE). Methods MAGE-A4, NY-ESO-1 and/or SAGE antigen expression in tumour samples was evaluated by quantitative real-time polymerase chain reaction (qRT-PCR). Informed consent was obtained from individuals prior to study enrolment. Results In total, 585 samples in 21 tumour types were evaluated between June 2009 and March 2018. The positive expression rates of these CT antigens were as follows: MAGE-A4, 34.6% (range, 30.7–38.7); NY-ESO-1, 21.0% (range, 17.2–25.1); and SAGE, 21.8% (range, 18.5–25.4). The MAGE-A4 antigen was expressed in 54.9% of oesophageal cancers, 37.5% of head and neck cancers, 35.0% of gastric cancers and 34.2% of ovarian cancers; the NY-ESO-1 antigen was expressed in 28.6% of lung cancers, 25.3% of oesophageal cancers and 22.6% of ovarian cancers; and the SAGE antigen was expressed in 35.3% of prostate cancers, 32.9% of oesophageal cancers and 26.3% of ovarian cancers. The most common tumour type in this study was oesophageal cancer. MAGE-A4, NY-ESO-1 and SAGE antigen expression were assessed in 214 oesophageal cancer samples, among which 24 (11.2%) were triple-positive, 58 (27.1%) were positive for any two, 59 (27.6%) were positive for any one, and 73 (34.1%) were triple negative. Conclusions Oesophageal cancer exhibited a relatively high rate of CT antigen mRNA expression positivity.


Background
Cancer testis (CT) antigens are anticipated to be optimal targets for cancer immunotherapy because their expression is limited to the testis and placenta in normal tissue [1]. Since T. Boon et al. reported that melanomaassociated antigen (MAGE), a CT antigen, was recognized by T cells [2], many researchers have studied the potential of CT antigens as cancer immunotherapy targets [3]. Not only immune checkpoint inhibitors but also genetically modified T cell therapies, such as chimeric antigen receptor (CAR) and T cell receptor (TCR)-engineered T cell therapies, have been developed in this era of cancer immunotherapy [4][5][6]. CT antigens are anticipated to be target proteins for genetically modified T cell therapy.

Statistical analysis
Pearson's chi-squared test of independence was used to evaluate associations between 2 variables. P-values of less than 0.05 were considered statistically significant. Calculations were performed with SPSS Statistics version 25 (IBM Japan, Ltd., Tokyo, Japan).  (Table 1).

CT antigen mRNA expression in tumours
Twenty-one tumour types were included in this study. The tumour types for which 5 or more samples were evaluated are listed in Tables 2-4. MAGE-A4 expression was evaluable in 575 samples. The MAGE-A4 expression rate was high in oesophageal cancer (54.9%), head and neck cancer (37.5%), gastric cancer (35.0%) and ovarian cancer (34.2%) ( Table 2).
The tumour types for which 4 or fewer samples were evaluated, which are not listed in the tables, were as follows: thyroid, small intestine, biliary tract, pancreatic, mesothelial, breast, urothelial, sarcoma, skin, multiple myeloma and unknown primary.

Co-expression of CT antigens
MAGE-A4, NY-ESO-1 and SAGE mRNA expression levels demonstrated positive relationships (Fig. 1). To exclude the influence of oesophageal cancer, which accounted for approximately half of the assessed samples, we divided the samples into oesophageal cancer and other cancers for analysis. In the non-oesophageal tumour types, significant correlations in CT antigen co-expression, except for NY-ESO-1 and SAGE co-expression, were identified (Additional Fig. 1). All 3 CT antigens were assessed in 436 samples, and three tumour types with high CT antigen expression rates are shown in Fig. 2. In oesophageal cancer, 65.9% of tumours were positive for at least one CT antigen, and 38.3% expressed 2 or 3 CT antigens. Among these CT antigen-positive tumour types, the median copy numbers of MAGE-A4 and SAGE in oesophageal cancer were higher than those in the other 2 tumour types (Additional Table 1).

Discussion
In this study, CT antigen expression was assessed in 585 tumour samples by quantitative real-time polymerase chain reaction (qRT-PCR). Among these tumour samples, 20-30% exhibited MAGE-A4, NY-ESO-1 and/or SAGE expression. The MAGE-A, NY-ESO-1 and SAGE expression rates in this study were comparable to those reported previously [11,12,[21][22][23][24][25][26][27]. Among the 585 tumour samples, 214 oesophageal cancer samples were evaluable for the expression of all 3 CT antigens. This evaluation revealed a high CT antigen co-expression rate in oesophageal cancer. CT antigens are promising targets for cancer immunotherapy. For example, NY-ESO-1-specific TCRengineered T cell therapy has shown promising antitumour responses in clinical trials [17,28,29]. Our data will be useful for considering the next cancer immunotherapy target. In addition, CT antigen expression and/or anti-CT antigen antibodies may have potential as biomarkers. Indeed, some reports have examined the impact of these factors on survival [30][31][32]. The impact of CT antigens on survival is controversial, possibly because of differences in tumour type, tumour stage and/ or tumour burden. Non-targeted antigen-specific T cell responses and/or antibody production, known as antigen spreading, often occurs during cancer immunotherapy. As antigen spreading may be helpful in guiding the  response to immunotherapy early in the treatment course [33,34], further assessments of CT antigens as prognostic factors are expected. The reason that CT antigen co-expression is high in oesophageal cancer has not yet been clarified. The expression of MAGE-A and NY-ESO-1 is mediated by demethylation of their promoters [35][36][37][38]. As MAGE-A, NY-ESO-1 and SAGE antigens are located in the q28 region on the X chromosome, demethylation of common promoters or those that are located nearby may occur. In this study, a high CT antigen expression rate was observed in oesophageal cancer. In general, CT antigen expression seemed to increase as the tumour progressed. However, one report suggested that CT antigen expression was high in low-grade oesophageal cancer [24]. The oesophagus may be more susceptible to demethylation than other organs.
We identified 5 reports about the co-expression of CT antigens in oesophageal cancer in international journals. Among these 5 reports, 4 assessed CT antigen expression by immunohistochemistry (IHC) [21,22,24,25] and 1 assessed it by PCR [23]. IHC can be performed on formalin-fixed, paraffin-embedded samples, enabling us to study a large number of samples retrospectively. However, CT antigen expression assessment via IHC can lack confidence. The anti-NY-ESO-1 antibody D8.38 recognizes not only NY-ESO-1 but also L antigen family member 1 (LAGE-1), which is also called NY-ESO-2 [21]. In addition, 57B, an anti-MAGE antibody often used to assess MAGE expression by IHC, cannot distinguish between members of the MAGE-A subfamily [21,39]. Forghanifard et al. [23] assessed CT antigen expression in oesophageal squamous cell carcinoma by PCR and reported a positive relationship between MAGE-A4 and NY-ESO-1 and between MAGE-A4 and LAGE-1. However, their report showed a MAGE-A4 expression rate of 90.2% in oesophageal squamous cell cancer. Although the assessment method differed (IHC vs PCR), the positive rate of MAGE-A4 expression was excessively high compared with that indicated in previous reports. The positive rates of MAGE-A4, NY-ESO-1 and SAGE expression in our study were comparable to those reported previously. In addition, the number of samples assessed for CT antigen co-expression in oesophageal cancer was larger than that in the study reported by Forghanifard et al. [23] [40]. The frequency of NY-ESO-1 mRNA expression in prostate cancer was 11.8% in our study. Latent prostate cancer might be involved.
This study has some limitations. First, all samples were assessed in a single institute. This strategy assured consistent methods and yielded reliable results, but the universality of our assessment was not confirmed. Second, details of histological types were not collected, because this study aimed to assess CT expression in tumour samples obtained from patients who hoped to enrol in clinical studies of CT antigen-targeting cancer immunotherapies. However, histological differences may affect the rate of CT antigen expression even in cancers of the same primary organ. For example, MAGE-A4 was more frequently expressed in lung squamous cell carcinoma than in lung adenocarcinoma [41], and the NY-ESO-1 expression rates in synovial sarcoma and myxoid round cell liposarcoma were higher than those in other types of soft tissue sarcoma [42]. Among patients enrolled in this study, oesophageal cancer was the most common type. In Japan, oesophageal squamous cell carcinoma accounts for approximately 90% of oesophageal cancers, and oesophageal adenocarcinoma is rare [43]. Thus, the CT antigen expression rate in oesophageal carcinoma in this study could be interpreted to reflect mainly oesophageal squamous cell carcinoma. Third, qRT-PCR analyses do   was defined as MCV-1 positivity and MCV-4 positivity [16], NY-ESO-1 positivity was defined as E978 positivity [15,19]. SAGE IHC was not assessed because of a lack of an appropriate antibody. IHC sensitivity and specificity were 64 and 75% for MAGE-A4 mRNA assessment, and 60 and 93% for NY-ESO-1 mRNA assessment, respectively (Additional Table 2); for approximately 40% of mRNA expression-positive tumour samples, protein production could not be confirmed. Despite these limitations, the large number of tumour samples, especially oesophageal cancer samples, is a strength of this study.

Conclusions
This study assessed MAGE-A4, NY-ESO-1 and/or SAGE antigen expression in 585 tumour samples. Oesophageal cancer exhibited a high rate of CT antigen mRNA expression and a high rate of CT antigen mRNA coexpression.