Daniel Perez and Thomas Herrmann contributed equally to this work.
Cancer Cell Biology
Cancer testis antigen expression in gastrointestinal stromal tumors: New markers for early recurrence
Article first published online: 21 JUL 2008
Copyright © 2008 Wiley-Liss, Inc.
International Journal of Cancer
Volume 123, Issue 7, pages 1551–1555, 1 October 2008
How to Cite
Perez, D., Herrmann, T., Jungbluth, A. A., Samartzis, P., Spagnoli, G., Demartines, N., Clavien, P.-A., Marino, S., Seifert, B. and Jaeger, D. (2008), Cancer testis antigen expression in gastrointestinal stromal tumors: New markers for early recurrence. Int. J. Cancer, 123: 1551–1555. doi: 10.1002/ijc.23698
- Issue published online: 28 JUL 2008
- Article first published online: 21 JUL 2008
- Manuscript Accepted: 4 APR 2008
- Manuscript Received: 14 SEP 2007
- cancer testis antigen;
- gastrointestinal stromal tumor;
Cancer testis antigens (CTAs) are expressed in a variety of malignant tumors but not in any normal adult tissues except germ cells and occasionally placenta. Because of this tumor-associated pattern of expression, CTAs are regarded as potential vaccine targets. The expression of CTAs in gastrointestinal stromal tumors (GIST) has not been analyzed systematically previously. The present study was performed to analyze the expression of CTA in GIST and to determine if CTA expression correlates with prognosis. Thirty-five GIST patients were retrospectively analyzed for their expression of CTAs by immunohistochemistry using the followingmonoclonal antibodies (mAb/antigen): MA454/MAGE-A1, M3H67/MAGE-A3, 57B/MAGE-A4, CT7-33/MAGE-C1 and E978/NY-ESO-1. Fourteen tumors (40%) expressed 1 or more of the 5 CTAs tested. Fourteen percent (n = 5/35) were positive for MAGE-A1, MAGE-A3 or MAGE-A4, respectively. Twenty-six percent (n = 9/35) stained positive for MAGE-C1 and 20% (n = 7/35) for NY-ESO-1. A highly significant correlation between CTA expression and tumor recurrence risk was observed (71% vs. 29%; p = 0.027). In our study population, the high-risk GIST expressed CTAs more frequently than low-risk GIST (p = 0.012). High-risk GISTs which stained positive for at least 1 CTA, recurred in 100% (n = 25) of the cases. This is the first study analyzing CTA expression in GIST and its prognostic value for recurrence. The CTA staining could add information to the individual patient prognosis and represent an interesting target for future treatment strategies. © 2008 Wiley-Liss, Inc.
Cancer testis antigens (CTAs) are expressed in a wide variety of malignant neoplasms, but in normal adult tissues CTA expression is restricted predominantly to germ cells of the adult testis and placenta.1–4 To date, more than 80 CTAs or CTA families have been identified and their number remain growing. Several monoclonal antibodies (mAbs) for the in situ detection of CTAs have been developed recently.4 Originally, CTAs were primarily identified based on their ability to elicit autologous T-cell and/or serological responses. Other CTAs were identified by methods such as database mining. Based on their tumor-restricted expression pattern, CTAs are regarded as valuable targets for cancer immunotherapy.5 Additionally, analysis of CTA expression may also be of value for the surgical pathologist helping to discriminate malignant lesions from benign neoplasms. The prognostic significance of CTA expression in many tumors remains unknown.
Diagnosis of gastrointestinal stromal tumors (GIST) is made on the immunohistochemical detection of the KIT receptor protein (CD117 antigen).6 Mutational analysis of the KIT and PDGF receptor α genes is used in KIT-negative tumors. Recurrence risk estimation after resection of a primary GIST remains challenging and is based on a consensus approach including the parameters tumor size and mitotic activity of the tumor.7 Recently, a new classification which includes tumor localization was proposed.8 The tumor localization plays an important role in relapse risk estimation.9 Additional information about CTA expression might be useful to complement prognostic factors.
Our study was performed to evaluate the expression of CTAs in a series of GISTs and its impact on prognosis.
Material and methods
Our study was done in compliance with local legal and ethical regulations. All tissue samples were retrieved from the archives of the Department of Pathology at the University Hospital of Zurich. All specimens were from standard formalin-fixed paraffin-embedded tissue blocks. Hematoxylin and eosin-stained sections were reviewed independently by S.M. and A.J. to verify the diagnosis and assess the presence of representative tissue. Morphological and histological parameters such as tumor size, number of mitoses and the presence of necrotic tissue were recorded. Only tissue specimens with unequivocal diagnosis of GIST and immunohistochemical positivity for CD117 were chosen for our study in order to reliably test the significance of CTA expression.
The cases analyzed in the present study were treated in the Department of Visceral and Transplant Surgery of the University Hospital Zurich between the years 1986 and 2006. Patient clinical follow-up data were recorded retrospectively from the medical records of the Department of Visceral and Transplantation Surgery of the University Hospital of Zurich and from the records of thepatients' family physicians. The duration between the initial operation and the first recurrence, the initial tumor site and the localization of the recurrence were analyzed. Patients with initial metastatic disease were excluded from the study. All patients were initially operated with a curative intention. The risk for recurrence was classified according to the Fletcher criteria.7 Thirty-five patients [22 males (63%), 13 females (37%)] with a mean age of 53.3 ± 19.2 years (range 12–85 years) and a mean follow-up of 59.6 ± 13.4 months (range 1–258 months) were included. Tumor primary site was stomach in 22 cases (60%), small bowel in 11 patients (31%), and in one case each colon, mesenterium and adrenal gland (9%). The patients after disease relapse received treatment with Imatinib mesylate (Gleevec®, Novartis, Switzerland).10 The impact of CTA expression on the mortality was not investigated due to only few and non-GIST related deaths during the study period (n = 5/35, 14%).
All primary tumors were analyzed for their expression of CTAs by immunohistochemistry with mAbs specific for the following antigens (mAb/antigen): MA454/MAGE-A1; M3H67/MAGE-A3; 57B/MAGE-A4, CT7-33/MAGE-C1; E978/NY-ESO-1, as described previously.11–15
After antigen retrieval, sections were incubated at 4°C overnight with MA454, 57B, CT7-33 and M3H67 mAbs. Specific binding was detected by using a biotinylated horse-antimouse secondary reagent (Vector Laboratories, Burlingame, CA) followed by an avidin–biotin complex system (ABC Elite; Vector Laboratories, Burlingame, CA). The E978 binding to NY-ESO-1 was detected with the Powervision kit (Immunovision Lab, Brisbane, CA). Diaminobenzidine tetrachloride or aminoethylcarbazole served as a chromogen. The investigators (D.P. and A.J.) were blinded to the clinical data. Testicular tissue with intact spermatogenesis was used as positive control. Immunoreactivity of tumor cells was graded as follows: focal, ∼<5% of tumor cells stained; +, 5–25%; ++, >25–50%; +++, >50–75%; ++++, >75% tumor cells positive. Immunopositivity was assigned to any grade of immunoreactivity observed in a given lesion.
The continuous and nominal data are presented as mean ± standard deviation and number with percentage, respectively. Nonparametric variables were analyzed with the χ2-test. Recurrence analysis was performed by the Kaplan–Meier survival analysis, and groups were compared by the log-rank test. A stepwise Cox-regression (forward and backward likelihood ratio) was performed to analyze independent predictors of recurrence. p-Values <0.05 were considered statistically significant.
Fourteen tumors (40%) were positive for CTAs by immunohistochemistry. The CTA expression pattern varied from focal (Fig. 1a) to homogeneous (Fig. 1b) expression. In most cases a heterogeneous expression pattern was observed. The expression rate for multiple CTAs is shown in Table I. The CTA expression pattern for each patient positive for at least 1 CTA is depicted in Table II. No patient was positive for all 5 CTAs tested. CTA expressing GIST were predominantly high risk tumors (p = 0.012). The CTA expression indicates early tumor recurrence in patients with GIST. Recurrence-free survival was significantly reduced in patients expressing CTAs (p = 0.027) over the study period (Fig. 2). Most importantly, patients developing recurrence were in 71% of cases (n = 10/14) positive for at least 1 CTA. In sharp contrast, all 10 (29%; n = 10/35) patients which were recurrence free during the observation period were negative for mAbs MA454, M3H67, 57B, CT-33 and E978 staining. Recurrence free survival for CTA negative patients at 1, 2 and 5 years was 85, 74 and 66% compared to 70, 51 and 25% for CTA positive patients. Median freedom from recurrence was 32 months in CTA positive and not reached during the study period in CTA negative patients. Accordingly, the expression of CTAs correlated with a significantly higher tumor recurrence rate (p = 0.027) over the study period (Fig. 2).
|Number of CTAs expressed||Immunoreactive GIST|
|1 CTA||6 (17%)|
|2 CTA||3 (9%)|
|3 CTA||1 (3%)|
|4 CTA||4 (11%)|
|5 CTA||0 (0%)|
|Positive for at least 1 CTA||14 (40%)|
|Patient no.||MAGE-A1, MA454||MAGE-A3, M3H67||MAGE-A4, 57B||MAGE-C1, CT7-33||NY-ESO-1, E978|
|Total immunoreactive tumors||5/35 (14%)||5/35 (14%)||5/35 (14%)||9/35 (26%)||7/35 (20%)|
Tumor relapses occurred in 16 cases either as local recurrence (n = 12) or distant metastases (n = 4). Five patients (n = 5) died during the study period due to severe co-morbidities [cerebrovascular insult (n = 2), myocardial infarction (n = 1), other tumor (n = 2)].
The GISTs were classified according to the Fletcher criteria7 into nineteen (54%) high risk tumors, 5 (14%) intermediate risk tumors, 8 (23%) low risk tumors and 3 (9%) very low risk tumors. We confirmed the previously published data,7 indicating that tumors in the high risk group have a more aggressive course of disease and a higher recurrence rate (p = 0.001). High-risk patients showed tumor relapse in 63% (n = 12/19) of the cases. However, a 5-year recurrence free survival of 29% compared to 63% does not predict recurrence perfectly. Therefore, we analyzed if the CTA expression status of GISTs along with the established recurrence risk criteria is a useful marker to predict tumor recurrence. The statistical analysis revealed that all patients with CTA negative and non-high risk tumors (n = 10) remained free of recurrence during the whole study period (Fig. 3). This was a significantly lower recurrence rate than in those patients with a CTA positive or high-risk tumor (p = 0.002). We entered the variables localization, resection status (R0, R1, R2), positivity for at least 1 CTA, very low risk GIST, low risk GIST, intermediate risk GIST and high risk GIST into a stepwise Cox-regression. Both forward- and backward regression identified CTA expression (p = 0.03, HR = 3.3, 95% CI 1.1–9.7) and high risk (p = 0.006, HR = 8.4, 95% CI = 1.9–37.8) as independent predictors of recurrence.
The current study was performed to evaluate the expression of CTAs in GIST and its significance as prognostic marker. Despite recent advances in the diagnostic approach to GIST,6–9, 16 to estimate their risk of recurrence remains challenging. Currently, GISTs are subdivided into high risk, intermediate risk, low risk and very low risk GISTs according to the criteria published by Fletcher et al.7 Indeed, one should be aware that the biological behavior of certain cases cannot be predicted. In the past several immunohistochemical markers have been developed for GIST.17 However, none of them could be identified as an independent prognostic factor. In the present study CTA expression was found to be an independent prognostic factor in terms of tumor recurrence. Therefore, immunohistochemical analysis of CTA protein expression could add in the future important prognostic information to the established criteria for tumor malignancy. Because the medical treatment of GIST has been revolutionized by the introduction of imatinib mesylate (Gleevec, Novartis Switzerland), it is of utmost importance to accurately assess the risk of recurrence to provide the most effective treatment.18–20 At present, adjuvant treatment with imatinib mesylate for 1 year after resection of a high-risk GIST is considered by many oncologists. A recent study showed that adjuvant treatment with imatinib 400 mg daily after complete resection of a primary GIST measuring at least 3 cm and expressing KIT increased the 1-year recurrence free survival from 83% without adjuvant treatment to 97%.21
Because of their tumor-associated expression pattern, CTAs have become a prime focus for immunotherapeutic approaches against cancer in recent years. Although there is evidence of CTA expression on the mRNA level in non-neoplastic tissue,22 their protein expression is restricted to placenta and testis.22 Immunotherapy strategies focusing on CTAs target exclusively CTA positive tumor cells without affecting normal tissue.
CTAs have been used in numerous clinical trials where patients with advanced CTA positive disease were immunized with CTA derived peptides, protein, or CTA encoding DNA.23–26 Thereafter, CTA specific immune responses have been observed frequently. Some reports with clinical disease remission after successful CTA specific immune response were documented.24, 25 So far, the best studied CTAs in terms of clinical application are MAGE-3 and NY-ESO-1.23–25
Our data document that CTA expression in GIST, even with focal staining, are prognostic markers for recurrence. Consequently, the analysis of CTA expression could be of high interest for the pathologist and clinician in estimating the individual prognosis of a GIST patient.
Although most of the mAbs of the present study have been used in extensive immunohistochemical analyses, their specificity has to be regarded with caution. This is best exemplified by mAb 57B. Although mAb 57B was generated to MAGE-A3, it was later considered to be a poly-MAGE reagent and is now regarded as reactive to MAGE-A4.13, 27, 28
The problem in the determination of the fine specificity of anti-CTA reagents is caused by the high homology of the MAGE-A family members and mAb 57B might serve as an example for the specificity of other anti-CTA reagents. However, no immunoreactivity with antigens outside the CTA spectrum has been reported for any of these mAbs.
A previous study identified the expression of several CTAs in 4 soft tissue sarcoma patients, of which, 1 was a GIST patient.29 Our study for the first time systematically analyses the expression of CTAs in a series of GIST and extends previous observations.
Previous immunohistochemical studies of tumors have shown heterogeneous CTA expression in different cancers, ranging from tumors with only single positive cells to neoplasms with a generally homogeneous expression pattern. CTA are certainly not the only markers of interest for the clinician and pathologist in the diagnosis of GIST. In addition to CTAs, there are obviously other markers, i.e., KIT (CD117),6 CD 34,30 and Ki-67.31
In conclusion, our study was performed in order to determine the expression of 5 different CTAs and draw attention to the potential use of CTA analysis for the assessment of malignancy in GIST. Future prospective analyses including other CTA will show if their expression in GIST can be used as a prognostic marker. Immunotherapy strategies targeting CTA should be explored in GIST refractory to standard treatment with imatinib mesylate and sunitinib malate.
- 21the American College of Surgeons Oncology Group (ACOSOG) Intergroup Adjuvant GIST Study Team. Adjuvant imatinib mesylate increases recurrence free survival (RFS) in patients with completely resected localized primary gastrointestinal stromal tumor (GIST): North American Intergroup Phase III trial ACOSOG Z9001. J Clin Oncol 2007; 25(18S): 10079., , , , , , , , , and
- 22The cancer/testis genes: review, standardization, and commentary. Cancer Immun 2004; 4: 1., , .
- 23LUD 00–009: phase 1 study of intensive course immunization with NY-ESO-1 peptides in HLA-A2 positive patients with NY-ESO-1-expressing cancer. Cancer Immun 2007; 7: 16., , , , , , , , , , , , , .
- 26Multi-center, double-blind, randomized, placebo-controlled phase II study to assess the efficacy of recombinant MAGE-A3 vaccine as adjuvant therapy in stage IB/II MAGE-A3-positive, completely resected, non-small cell lung cancer (NSCLC). J. Clin. Oncol 2006; 24(18S): 7019., , , , , , , , , .
- 29Antigens recognized by autologous antibodies of patients with soft tissue sarcoma. Cancer Immun 2005; 5: 4., , , , , , , .