Prognostic impact of cancer/testis antigen expression in advanced stage multiple myeloma patients

Multiple myeloma (MM) is a malignancy characterized by the proliferation of a clonal plasma cell, with a resultant increase of a single immunoglobulin and its fragments in the serum and urine, osteolytic lesions, anemia and hypercalcemia (1). High-dose chemotherapy followed by autologous stem cell transplant increases complete response rates and overall survival (OS). On the other hand, an allogeneic stem cell transplant represents a chance of cure for selected cases but also results in high mortality and morbidity (2). Currently, immunomodulatory drugs, such as thalidomide and its analogs (lenalidomide), and proteasome inhibitors (bortezomib) have been used as single or combined agents for relapsed/refractory cases and have even been used as first line therapy (3). However, despite of all these new therapeutic alternatives, MM remains an incurable disease, with a median survival of 3 years (1). Therefore, insights into the biology of aberrant plasma cell differentiation in monoclonal gammopathy of undetermined significance (MGUS) and MM offer the potential for new therapeutic approaches (4). In this context, tumor vaccines are very attractive therapeutic strategies because they may induce death of tumor cells resistant to current chemotherapy protocols. In addition, immune memory mechanisms induced by some vaccines can elicit long-term tumor immunosurveillance and could be helpful to prevent disease relapse (5).

Cancer/testis (CT) antigens are genes expressed almost exclusively in the normal human germ line and in malignancies. They have become the most extensively studied antigen group in the field of tumor immunology (6, 7). CT antigens were originally described in patients with malignant melanoma due to their ability to elicit cytotoxic T cells and humoral responses (8). Because CT antigens show restricted normal tissue expression and are highly immunogenic, they are especially attractive targets for immunotherapeutic approaches in cancer patients (5, 9).

In a study comparing genes overexpressed in malignant plasmablasts versus polyclonal plasmablasts generated from peripheral blood B cells, five of the eight most expressed genes were the CT antigens GAGE-6, MAGEA1, MAGEA2, MAGEA3, and SSX2 (10).

Recently, MAGEC1/CT7 and MAGEA3/6 were found to be frequently expressed in advanced stage MM patients (11). Higher levels of MAGEC1/CT7 and MAGEA3/6 proteins were also found to correlate with an elevated plasma cell proliferation index. These findings suggest a possible pathogenic role for such proteins in MM and also show that they could be attractive targets for immunotherapy in this disease.

This study aims to analyze the global expression of an extensive panel of CT antigens in normal tissues and bone marrow aspirates from MGUS, plasmacytomas and MM patients to detect possible prognostic markers and immunotherapeutic targets for this incurable disease.

We used RT-PCR to analyze the expression of 14 CT antigens (MAGEA1, MAGEA2, MAGEA3/6, MAGEA4, MAGEA10, MAGEA12, BAGE-1, MAGEC1/CT7, GAGE family, LAGE-1, NY-ESO-1, PRAME, SPA17, SSX1; Table 1) in 15 normal tissues, a pool of 10 normal bone marrow samples, 3 normal tonsils, bone marrow aspirates from 6 normal donors, 3 MGUS, 5 solitary plasmacytomas, 39 multiple myeloma samples and the MM cell line U266. Baseline characteristics of patients with MGUS, plasmacytoma and multiple myeloma are described in Table 2.

Different normal tissue RNA samples were analyzed for the expression of the 14 CT antigens. The SPA17 gene was found to be expressed in all seven normal bone marrow samples and all normal tissues tested. Sequencing of the SPA17 RT-PCR products from normal testis, normal mammary gland, normal bone marrow and normal plasma cells confirmed amplification of this gene. Thus, SPA17 was excluded from further analyses. The results of the analysis of the expression of the other 13 CT antigens in normal tissues are reported in more detail in Table 3.

The frequency of CT antigen expression in MM patients is shown in Figure 1. MAGEC1/CT7, MAGEA3/6 and LAGE-1 together were expressed in 85% of our MM patients. At least one CT antigen was expressed in 92.3% of MM cases. Of the 16 cases with a lower frequency of CT antigen expression (0, 1 or 2 CT antigens expressed), 3 (7.6%) were negative for all CT antigens tested (cases 42, 59, and 130). Nine of the remaining 13 cases (approx. 70%) were positive for MAGEC1/CT7 expression. We did not find an association between age, MM isotype or International Staging System (ISS) (12) stage and the percentage of CT antigens expressed in the MM cases studied.

To evaluate possible prognostic factors among the variables studied, we performed univariate analyses including all patients. The following parameters gave a negative prognostic impact on overall survival: (i) a β2-microglobulin level of 5.5 mg/L or more (β2-microglobulin <5.5 mg/L: n = 15, median OS = 40 months; β2-microglobulin 5.5 mg/L or more: n = 22, median OS = 11 months; log-rank, P = 0.0343, data not shown), (ii) the number of CT antigens expressed (6 CT antigens or less: n = 29, median OS = 25 months; >6 CT antigens: n = 10, median OS = 4 months; log-rank, P = 0.0062), and (iii) GAGE family expression (GAGE family negative: n = 26, median OS = 40 months; GAGE family positive: n = 13, median OS = 13 months; log-rank, P = 0.0250) (Figure 2).

Subsequently, Durie and Salmon stage, ISS stage, β2-microglobulin level, albumin level, the number of CT antigens expressed, expression of MAGEC1/CT7, GAGE family, LAGE-1, SSX1, and del 13 were included in the Cox regression model. When all patients were included in the model, GAGE family expression and having >6 CT antigens expressed were identified as independent prognostic factors (Table 4). However, when only non-transplanted patients were analyzed, GAGE family expression and the expression of >6 CT antigens lost prognostic impact but MAGEC1/CT7 expression showed independent prognostic significance (P = 0.0222) confirmed by the Cox regression model (Figure 2).

In the present study, we demonstrate high frequencies of MAGEC1/CT7, LAGE-1 and MAGEA3/6 expression in advanced stage MM patients. We also show that some CT antigens have an impact in MM prognosis.

Although the number of patients analyzed is small, we were able to show that MAGEC1/CT7 is the most frequently expressed CT antigen (77%) in MM. Cho and collaborators (13) also demonstrated that MAGEC1/CT7 protein is highly expressed (82%) in MM by immunohistochemistry (IHC). In addition, these authors suggested that MAGEC1/CT7 expression in MM cells could elicit a T-cell immune response due to the increase in INF-γ secretion by lymphocytes when co-incubated with MAGEC1/CT7-transduced autologous myeloid dendritic cells and that MAGEC1/CT7-specific vaccines may boost this immunity. On the other hand, Fontecedro and colleagues (14) showed the in vivo immunogenicity of MAGEC1/CT7 and the spontaneous occurrence of anti-MAGEC1/CT7 humoral responses in 22% of MM patients by Western blot and ELISA. All of the patients were patients with advanced disease. Furthermore, Goodyear and co-workers demonstrated CD8+ T cell activity recognizing several CT antigens in many myeloma patients (15).

We share with them the opinion that MAGEC1/CT7 is a good candidate for targeted therapy in MM, with the best candidates for this approach being patients expressing MAGEC1/CT7 and with a low tumor burden, i.e. after a successful autologous stem cell transplant.

Furthermore, our study showed that MAGEC1/CT7 expression seems to occur early in disease development, because 1 of 3 MGUS and 1 of 5 plasmacytomas were positive. These findings suggest that MAGEC1/CT7 may have a biological role in the early stages of MM and may contribute to plasma cell proliferation. Dodapkar et al. (16), as well as Jungbluth and co-workers (11), demonstrated that this CT antigen is expressed in plasmacytoma (by IHC) and MGUS (by RT-PCR), respectively.

LAGE-1 was the second most frequently expressed CT antigen (49%) in our study. Tumor cells expressing LAGE-1 can be recognized by autologous cytotoxic T lymphocytes specific for LAGE-encoded antigenic peptides presented by various HLA class I molecules (17). NY-ESO-1 and LAGE-1 exhibit a high level of sequence similarity and, using single molecule fluorescence microscopy, Purbhoo and co-workers (18) demonstrated that NY-ESO1/LAGE-1-positive tumor cells present NY-ESO-1(10-50)/LAGE-1(157-165) epitopes per cell. Recently, Sun et al. (19) identified a new LAGE-1 peptide (LAGE-A68) which is recognized by T lymphocytes on tumor cells. In our study, the frequencies at which LAGE-1 and NY-ESO-1 were found to be expressed in MM patients were 49% (19/39) and 33% (13/39), respectively, in agreement with the findings of van Baren et al. (52% and 31% respectively; 20).

NY-ESO-1 has been shown to physically interact with MAGEC1/CT7 (21). Immunoprecipitation and immunofluorescent staining confirmed that MAGEC1/CT7 and NY-ESO-1 interact and their cytoplasmic co-localization in melanoma cells. The co-expression of these two genes was also confirmed in primary tumors, such as MM and non-small cell lung cancer (21). Among the 39 MM samples analyzed, there were 13 (33%) which were positive for NY-ESO-1 expression and 30 (77%) which were positive for MAGEC1/CT7 expression. Simultaneous expression of NY-ESO-1 and MAGEC1/CT7 was observed in 26% of MM cases. This study demonstrated that the coordinated expression of CT antigens observed in many tumor types may be important for their function and for a potential role in tumorigenesis.

Previous studies in cancer patients have shown that patients with NY-ESO-1 antigen expression and humoral immune response to NY-ESO-1 tend to have advanced stage disease. For example, a higher frequency of NY-ESO-1 expression in bladder cancer was correlated with muscular wall invasion (22) and an antibody response to NY-ESO-1 was correlated with advanced stage disease (23).

LAGE-1 is among the genes identified by array analysis that correlated with resistance to proteasome inhibitor bortezomib, when pre- and post-treatment myeloma samples were studied (24). Thus, if NY-ESO-1 antigen expression was associated with advanced stages of cancer and LAGE-1 should confer resistance to the proteasome inhibitor bortezomib, it was reasoned that the expression of these two CT antigens might contribute to a poor outcome in MM and thus represent prognostic markers for this disease. In this context, Condomines and co-workers (25) demonstrated by microarray analysis an influence of CT antigens encoded by chromosome X on MM event-free survival, particularly CTAG1B (NY-ESO-1), CTAG2 (LAGE-1), MAGEA1, MAGEA2, MAGEA3, and MAGEA6.

One of the most interesting findings of this study is the frequent expression of LAGE-1, despite the significantly lower expression of its close homologue NY-ESO-1 (approximately 94% homology). Due to this homology, it is possible that LAGE-1 could also elicit an immune response against NY-ESO-1. Furthermore, LAGE-1 and NY-ESO-1 were co-expressed in 23% of the advanced stage MM cases studied.

The MAGEA3 and MAGEA6 proteins share more than 98% sequence homology, and are frequently analyzed together. MAGEA3/6 was the third most frequently expressed CT antigen in this study (41%). Jungbluth and group (11) demonstrated by RT-PCR that the MAGEA-A family was expressed in 100% of MM patients. MAGEA3/6 was expressed in 70% of stage III MM specimens by IHC. Moreover, van Baren et al. (20) found expression of MAGEA3 in 29% and MAGEA6 in 33% of MM samples (also advanced stage MM) by RT-PCR. Due to its high frequency in advanced stage MM patients, it is possible that members of the MAGE family participate in the biology of this fatal B cell malignancy.

SPA17 was the only CT antigen strongly expressed in all non-gametogenic tissues tested and this result was confirmed by automated sequencing. This finding has been described by Scanlan et al. (9) and, more recently, by Condomines et al. (25). Therefore we believe that SPA17 does not fulfill the requirements to be classified as a CT antigen and is not a likely target for immunotherapy of any cancer type.

We attempted to correlate the International Staging System (ISS) MM stage, which is based on albumin and β2-microglobulin serum levels at diagnosis, with the number of CT antigens expressed. The hypothesis tested was that a lower number of CT antigens expressed would correlate with a lower ISS stage. In our group of MM patients, only three cases were classified as stage 1 and fewer than three CT antigens were found to be expressed in all of them. However, due to the small number of cases in this category, we were unable to demonstrate a correlation between ISS stage and the number of CT antigens expressed.

To evaluate possible prognostic factors among the variables studied, we performed OS univariate analyses including all patients. A β2-microglobulin level of 5.5 mg/L or more, expression of >6 CT antigens and GAGE family expression had a negative impact on OS.

Using a Cox regression model, GAGE family expression and expression of >6 CT antigens were found to be independent prognostic factors. The finding that a worse prognosis is associated with the expression of a larger number of CT antigens supports the hypothesis of the cumulative deleterious effect of CT antigen expression on MM disease progression. In addition, expression of MAGEC1/CT7 showed independent prognostic significance in non-transplanted patients, suggesting that it can be a target, and also a marker, of advanced stage disease. This is now the second study demonstrating the potential prognostic relevance of CT antigen expression in MM; the other was recently published by Condomines and co-workers (25).

In this study, expression of >6 CT antigens or of the GAGE family represent important prognostic factors when all patients were analyzed. On the other hand, expression of MAGEC1/CT7 only represents such a factor for non-transplanted patients. It is possible that transplantation proceedings have interfered in these results. Seven out of 8 MM transplanted cases were positive for MAGEC1/CT7 expression (Table 5). In these cases, the autologous stem cell transplant possibly overcame poor outcome related to the expression of this gene. Therefore, only after exclusion of transplanted patients did we find that MAGEC1/CT7 expression correlated with a poor prognosis (non-transplanted, MAGEC1/CT7-negative: n = 8; median OS = not reached; non-transplanted, MAGEC1/CT7-positive: n = 23; median OS = 7 months; log rank, P = 0.0222).

On the other hand, only 2 out of 8 transplanted MM cases were positive for GAGE family expression (Table 5). Consequently, the exclusion of transplanted patients eliminated a large number of GAGE family-negative cases. Accordingly, transplantation proceedings (and not the absence of GAGE-family mRNA) justified the better prognosis seen in the GAGE-negative group (non-transplanted, GAGE-negative: n = 20, median OS = 7 months; non-transplanted, GAGE-positive: n = 11; median OS = 12 months; log rank, P = 0.0894).

Likewise, no MM transplanted patients presented expression of >6 CT antigens (Table 5). Thus, transplantation proceedings (and not the low number of CT antigens expressed) were responsible for the increased overall survival in these patients (non-transplanted, 6 CT antigens or less: n = 21, median OS = 13 months; non-transplanted, >6 CT antigen: n = 10, median OS = 4 months; log rank, P = 0.1032).

To date, CT antigens have been studied because they are immunogenic and are considered promising target molecules for cancer vaccines. Based on our findings, MAGEC1/CT7, LAGE-1 and MAGEA3/6 are good candidates for immunotherapy in MM, since together they are expressed in 85% of patients. Several clinical trials vaccinating with MAGEA3 (which shows homology with MAGEC1/CT7 and MAGEA6), MAGEA12 and NY-ESO-1 (which shows homology with LAGE-1) are currently open for enrollment for patients with MM (26). It will also be of interest to investigate if some CT antigens, such as the GAGE family or MAGEC1/CT7, might participate in MM pathogenesis.

A panel of normal tissues (testis, placenta, skeletal muscle, bladder, lung, spleen, heart, brain, thymus, uterus, stomach, mammary gland, pancreas, prostate and colon; Clontech, Palo Alto, CA, USA), a pool of ten normal bone marrow samples from Clontech (Palo Alto, CA, USA), six normal bone marrow aspirates (from donors for allogeneic stem cell transplants) and three normal tonsils (obtained from children submitted to tonsillectomy) were used to test all the primer sets. Bone marrow aspirates from three monoclonal gammopathies of undetermined significance (MGUS), five solitary plasmacytomas, and 39 multiple myeloma samples (2 Durie and Salmon stage II and 37 stage III) were obtained between June 2002 and April 2006 from the Hospital Sao Paulo UNIFESP/EPM, Sao Paulo, Brazil. All of these patients were newly diagnosed and treatment naïve at the time of bone marrow harvest. The cases were classified according to the International Staging System (12) as 3 stage 1, 13 stage 2 and 20 stage 3 cases (Table 2). In the 39 MM patients, bone marrow biopsies showed 10% to 98% of plasma cells (median of 80%). Patients were initially treated with standard chemotherapy protocols, namely melphalan and prednisone (MP) or vincristine, doxorubicin and dexamethasone (VAD). Eight patients also underwent melphalan-based autologous stem cell transplantation. This study was approved by the institution's Ethical Review Committee and all patients provided written informed consent prior to collection of bone marrow samples. The MM cell line U266 (TIB-196, ATCC, Manassas, USA) was also analyzed.

Total RNA was extracted using the TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. The RNA was recovered from the aqueous phase by ethanol precipitation and the pellets were dissolved in 30 µL of diethyl pyrocarbonate (DEPC) treated water. First-strand cDNA synthesis, primed with oligo (dT) and 2 µg of RNA template, was reverse transcribed (Superscript II, Invitrogen, Carlsbad, CA, USA) following the manufacturer's instructions and diluted 10x in water. After cDNA synthesis, the expression of β-actin was determined in all samples. The 14 CT antigens described in Table 1 were evaluated by RT-PCR and 2% agarose gel electrophoresis. Normal testis was used as template for positive controls in all RT-PCR reactions. PCR reactions were performed in 0.2 mL tubes and carried out in a total volume of 25 µL containing 2 µL of cDNA template, 1x PCR buffer, 1 mM MgCI2, 0.2 µM dNTPs, 10 pmol each of the forward and reverse primers, and 1.25 U of Platinum Taq DNA Polymerase (Invitrogen, Carlsbad, CA, USA). The PCR steps were performed in an Eppendorf MasterCycler Gradient thermocycler (Hamburg, Germany) and primer sequences and cycle conditions are given in Table 1. Initial denaturation at 94˚C for 2 minutes and single step elongation at 72˚C for 7 minutes were used in all cycles. The GAGE family was analyzed using consensus primers because its members have high sequence homology. All reactions were undertaken in duplicate and amplifications of target fragments were considered positive, independent of the intensity of the bands.

Amplification of SPA17 for automated sequencing was carried out with the BigDye Terminator Cycle Sequencing Ready Reaction Kit (Perkin Elmer Applied Biosystems, Foster City, CA) using an ABI Prism 377 DNA Sequencer (Perkin Elmer, Norwalk, CT, USA).

Associations between the variables studied were tested by the Pearson chi-square test. Overall survival (OS) time was calculated from the date of diagnosis of MM until death or last follow-up. Actuarial probabilities of OS were estimated according to the Kaplan-Meier method and the curves were compared using the log-rank test. The Cox regression model was also employed to evaluate which variables could be considered independent prognostic factors on OS in this group of patients. The level of significance for all statistical tests was 5%. Statistical analysis was performed using SPSS 8.0 software.

This work was supported by grants from FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo, São Paulo, Brazil, 04/13213-3). V.C.C.A. was also supported by FAPESP (04/12855-1). The authors thank Dr. Andrew J. G. Simpson, from the Ludwig Institute for Cancer Research, for all the support and critical review of this manuscript. The authors also thank Fleury Laboratories, São Paulo, Brazil, for the immunological characterization of MM cases.