DNA Mismatch Repair Protein (MSH6) Correlated With the Responses of Atypical Pituitary Adenomas and Pituitary Carcinomas to Temozolomide: The National Cooperative Study by the Japan Society for Hypothalamic and Pituitary Tumors

According to the 2004 World Health Organization classification (1), pituitary tumors are classified as typical adenomas, atypical pituitary adenomas (APAs), and in very rare cases, pituitary carcinomas (PCs). Atypical adenomas are defined by their invasiveness, having a Ki-67 (MiB-1) proliferation index of 3% or more and extensive nuclear staining for the p53 protein. Pituitary carcinomas are characterized by the presence of metastases. APAs and PCs are resistant to conventional treatments and are difficult to control (2–4).

Temozolomide (TMZ) is an alkylating agent and a prodrug that is rapidly hydrolyzed into the active compound [5-(3-methyltriazeno) imidazole-4-carboxamide] under physiological conditions (5). The 5-(3-methyltriazeno)-imidazole-4-carboxamide causes DNA damage through methylation of the O⁶ position of guanine, forming the most potent cytotoxic DNA adducts. The sequence of mismatch-repair events leads tumor cells to apoptosis. In cases of newly diagnosed glioblastoma multiforme (GBM), TMZ (75 mg/m²/d for 42 days) is administered orally concomitant with fractionated radiotherapy (60 Gy total dose: 2 Gy × 5 d/wk for 6 weeks) followed by adjuvant TMZ (150 mg/m²/d × 5 days every 4 weeks) (6). Recently, it has been documented to be effective against malignant pituitary tumors resistant to other therapies (7–9).

O⁶-methyl-guanine-DNA methyltransferase (MGMT) is a DNA repair protein that reverses alkylation at the O⁶ position of guanine by transferring the alkyl group to a sulfur group of cysteine within its sequence (10). MGMT removes alkylating adducts induced by TMZ treatment and counteracts its antineoplastic action. Malignant gliomas have been reported to acquire resistance to TMZ through mutations of p53, a tumor-suppressor gene (5), and mutations of MSH6, a DNA repair protein (11, 12).

However, TMZ is currently indicated to treat only patients with malignant glioma, anaplastic astrocytoma, or GBM. At present, TMZ therapy for malignant pituitary neoplasms is not covered by public health insurance. Patients with malignant pituitary neoplasms who received TMZ therapy have borne the heavy burden of paying for their treatment.

The purpose of this study is to identify the characteristics of cases of APA and PC treated with TMZ and to provide a guidance for TMZ treatment against APA and PC.

Materials and Methods

This study was performed with approval of the ethics committee of Teikyo University Chiba Medical Center (approval number 10-28).

Surveys

A survey was distributed to members of the Japan Society for Hypothalamic and Pituitary Tumors. The questions were in regard to their therapeutic experiences and the clinical characteristics of APA and PC with or without TMZ treatment.

Tumor specimens

Stored tumor samples that were obtained from APA and PC patients treated by the members responding to the questionnaire were delivered to the Department of Pathology of the International University of the Health and Welfare Mita Hospital and were examined pathologically. We reviewed the clinical records of them and excluded cases that were introduced to TMZ monotherapy for 2 or fewer cycles. Regarding case 10, the combination therapy with TMZ and radiation was based on the standard protocol of the initial therapy for GBM (6). Response to TMZ treatment was defined as complete response (CR), partial response (PR), progressive disease (PD), and stable disease (SD) according to RECIST (Response Evaluation Criteria in Solid Tumors) version 2.0. A CR, PR, or SD was judged to be effective, a PD as ineffective. In cases of functional adenomas, the values of hormones were excluded from the TMZ response assessment.

Immunohistochemical analyses of Ki-67, MGMT, p53, MSH6, and pituitary hormones

Paraffin-embedded tissue sections were examined immunohistochemically using the mouse monoclonal antibodies, p53 (DO-7), MSH6 (Abcam, Cambridge, Massachusetts; 1:100 dilution), MGMT (MT 3.1; Millipore, Billerica, Massachusetts; 1:500 dilution), and Ki-67 (Dako, Glostrup, Denmark; 1:25 dilution). Tissue sections were stained with Ventana BenchMark ULTRA automated immunostainer using heat-induced epitope retrieval and a standard diaminobenzidine (DAB) detection kit (Ventana iVIEW DAB Universal Kit) except for MGMT immunostaining.

For MGMT immunostaining, sections were heated up to 121°C in citrate buffer (pH 6.0) in an autoclave for 10 minutes. Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide in water (10 minutes). After each following step, sections were washed with 0.01M PBS (pH 7.4) 3 times for 15 minutes each. The tissue sections were incubated overnight at 4°C with the primary antibody. Specific binding was detected by DAB, and the sections were then counterstained with Mayer's hematoxylin.

Positive control tissues consisted of paraffin-embedded sections of colon cancer for immunostaining of MGMT, p53, and MSH6.

Assessment of immunohistochemical findings

Ki-67 and p53 were evaluated as the percentage of immunopositive cells. In the precedent literature, MGMT immunopositivity was classified into positive and negative by the cutoff value at 10% of cells (13, 14). For MGMT, The specimens were divided into two groups clearly, the immunonegative ones and the immunopositive ones showing strong staining over 50% of cells. For MSH6, the specimens were classified into four scores as follows: score 0, immunonegative; score 1, <10% of cells; score 2, 10%–50%; score 3, >50% (representative cases are shown in Figure 1). MSH6 labeling scores ≥2 were considered immunopositive and the others as immunonegative (12). Scoring was performed blindly by two independent reviewers (A.M. and R.Y.O.).

Figure 1.

Immunohistochemistry of case 3a and case 3b for MSH6. A, Immunohistochemistry of case 3a (TMZ-effective tumor) for MSH6 showing diffuse and remarkable stain (score 3). B, Immunohistochemistry of case 3b (TMZ-resistant tumor after TMZ treatment) for MSH6 showing loss of MSH6 immunoexpression (score 0). Original magnification, ×20.

Open in new tabDownload slide

Statistical analyses

We investigated correlations between immunohistochemical findings and the efficacy of TMZ using a Student's t test and Fisher's exact test in the SPSS statistical program.

Results

Surveys

A total of 205 members responded to the surveys. There were 35 cases of PC, and 12 of them had been treated with TMZ between 2006 and 2011. The characteristics of the cases are presented in Table 1. Of the 12 patients with PC, 7 were women and 5 were men; their mean age was 51.4 years (range, 23–73 years).

We found 178 cases of APA through checking duplicate subjects, and 4 of them had been treated with TMZ. There were 3 women and 1 man; their mean age was 56.8 years (range 45–72 years).

Immunohistochemistry of hormones, Ki-67, MGMT, p53, and MSH6

Three samples from 3 cases of APA and 11 samples from 10 cases of PC were analyzed finally. The clinical and immunohistochemical characteristics of the tumors are shown in Table 2. TMZ effectiveness was determined according to the response at the time of the surgery where each sample was obtained. For example, two samples from case 3 were obtained by surgical resection, one before TMZ treatment (sample 3a) and the other after acquiring TMZ resistance (sample 3b). Case 3 was a 60-year-old woman with atypical prolactinoma susceptible to TMZ. The tumor became resistant to TMZ after 10 cycles of treatment and was diagnosed as a pituitary carcinoma. She underwent the second operation, and sample 3b was obtained.

Ki-67 labeling index

The samples had large deviations of Ki-67 labeling index. There was no significant correlation between Ki-67 labeling index and the effectiveness of TMZ (P = .46, Student's t test) (Table 3).

MGMT immunoexpression

Seven specimens were positive for MGMT immunoexpression. Of those, TMZ was effective in 4 samples and ineffective in the remaining 3 samples. In the MGMT-negative samples, TMZ was effective in 4 and ineffective in 3. These data showed no apparent correlation between MGMT immunoexpression and TMZ efficacy (Table 3).

p53 labeling indices

The p53 labeling indices ≥10% were considered immunopositive and <10% as immunonegative. Five samples were p53-positive; TMZ was effective in 4 of those and ineffective in 1. Nine samples were p53-negative. In those, TMZ was effective in 4 and ineffective in 5. There was no statistical significance between p53 and TMZ efficacy (P = .24, Fisher's exact test) (Table 3).

MSH6 immunoexpression

MSH6 immunoexpression was positive in 10 samples; TMZ was effective in 8 of them and ineffective in 2. MSH6 immunoexpression was negative in 4 samples; TMZ was not effective in any of them. This result presents the possibility that MSH6 expression might correlate with TMZ efficacy (P = .015, Fisher's exact test) (Table 3). Illustrative immunohistochemical staining of MSH6 is demonstrated in Figure 1: MSH6 immunopositivity in TMZ-effective tumor (Figure 1A, score 3) and negative immunostaining of MSH6 in tumor acquiring TMZ resistance (Figure 1B, score 0).

Discussion

TMZ is globally approved as a first-line chemotherapeutic agent for the treatment of malignant glioma and is also approved for malignant melanomas in several countries. In addition, phase II trials of TMZ treatment in patients with metastatic renal cell cancer and metastatic breast cancer have been reported (15, 16); however, TMZ was ineffective.

In regard to malignant pituitary adenomas treated with TMZ, we read in the literature 21 case reports of PC after 2004 and 27 case reports of APA after 2006. This study evaluates the clinical and immunohistochemical characteristics of 3 APAs and 10 PCs treated with TMZ on a relatively large scale.

In the past, several different therapeutic regimens, such as N-(2-chloroethyl)-N′-cyclohexyl-N-nitrosourea and 5-fluorouracil and cisplatin and etoposide, have been used for highly aggressive pituitary adenomas or poorly differentiated neuroendocrine tumors (17–19). Kaltsas et al (17) had shown that cis-platinum and etoposide-based chemotherapies might be correlated with considerable, but variable, responses. However, the overall response to conventional chemotherapeutic agents was relatively poor, and there are no randomized studies. Ideally, TMZ treatment should be assessed in a prospective study.

The standard therapeutic dose of TMZ is 150 to 200 mg/m² daily for 5 of every 28 days as 1 cycle (7). Most cases in the present study followed this administration method. The common adverse effects of TMZ include nausea, vomiting, and constipation. Hematologic toxicities occur in <10% of patients treated with TMZ. Case 6 had developed pancytopenia after 2 cycles of TMZ treatment (150 and 200 mg/m²). The dose was decreased to 100 mg/m² and administered for another 10 months. No case in this study dropped out of TMZ therapy due to adverse events.

The MGMT protein repairs alkylated guanine of DNA and provides the resistance to TMZ (10). In malignant gliomas such as glioblastomas, subjects whose MGMT promoter is methylated obtain resistance to TMZ. TMZ has been suggested to be effective also in pituitary adenomas harboring a methylated MGMT gene promoter, as is the case in gliomas (20, 21). Salehi et al (22) reported that there was no relationship between MGMT promoter methylation status and MGMT immunoexpression. Although the changes in MGMT expression in recurring pituitary tumors could affect adenoma progression and TMZ efficacy, MGMT is not the sole molecular factor determining sensitivity to TMZ.

The p53 tumor suppressor gene is frequently mutated in human cancer. Mutant p53 is unable to transcriptionally activate MDM2 (murine double minute 2) or E3 ubiquitin ligase and results in a loss of negative feedback and accumulation of p53 protein. Thus, the immunohistochemical overexpression and accumulation of p53 protein has been used as a surrogate biomarker for p53 mutation (23). In addition, the anti-p53 antibody used in this study reacts with wild-type as well as mutant types of p53 protein. In malignant gliomas, it was reported that MGMT expression is down-regulated by wild-type p53 and p53 status is associated with the effectiveness of TMZ in addition to MGMT status (24). In pituitary adenomas, however, p53 mutations have rarely been detected. Tanizaki et al (25) compared immunostaining and mutation status of p53 in pituitary tumors and demonstrated that 2 corticotroph carcinomas with >50% p53-immunopositive cells harbored a p53 missense mutation. Our study suggested that p53 mutation and increased immunopositivity of tumor cells were not responsible for acquiring resistance to TMZ.

The mismatch repair (MMR) pathway is involved in the removal of DNA base mismatches caused either by errors in DNA replication or by DNA damage. The existing findings of the DNA MMR pathway, such as MSH2, MLH1 (MutL homolog 1), PMS2 (postmeiotic segregation increased 2), and MSH6, suggest that MMR has important roles in both predisposition to tumor and also the response to therapy (26). It was suggested that MMR not only repairs DNA but also stimulates DNA damage-induced G2 checkpoint and apoptosis and that MMR has 2 functions, one in DNA repair and the other in DNA damage signal transduction (27).

Recently, inactivated mutations in the mismatch repair gene MSH6 and subsequent loss of MSH6 immunostaining were identified in a subset of postradiation and post-TMZ treatment GBM recurrences, and loss of MSH6 correlated with tumor growth during TMZ treatment (11, 12). MSH6 inactivation and mutation in GBMs during TMZ therapy both in vitro and in vivo were associated with TMZ resistance (28). A case was reported in which loss of MSH6 occurred during the progression from an APA to a PC, which might have caused resistance to TMZ treatment (29).

We showed that the lack of MSH6 immunopositivity had a significant correlation with resistance to TMZ treatment. The results of our study might also imply that preserving MSH6 function is important for the effectiveness of TMZ on malignant pituitary adenomas.

The present study includes 3 cases of post-TMZ treatment for which TMZ was ineffective (cases 3, 9, and 13). However, 2 cases (3 and 13) were MSH6-negative. Case 9 was MSH6-positive and negative for p53 and MGMT. There may be an unknown or multifactor mechanism for acquiring TMZ resistance.

Some cases discontinued TMZ treatment after 1 or 2 cycles as nonresponders. However, case 7 has followed a stable and controllable course of treatment for over 8 months with intracranial dissemination lesions. Raverot et al (8) proposed that response after 3 cycles of TMZ treatment was a predictor of TMZ effectiveness. We also recommend that TMZ should be administered for at least 3 cycles to evaluate the efficacy for APAs and PCs.

Contrary to this, 6 of the 13 patients (46%) with APAs and PCs became recurrent after having an effective response to TMZ therapy, as indicated by asterisks in Table 2. The period for which the tumors had acquired resistance to TMZ ranged from 5 to 19 months (mean 10.5 months). Long-term follow-up is necessary for those patients with initial responses to TMZ.

Because malignant pituitary neoplasms are rare, the small number of available samples is the limitation of this study. Meanwhile, TMZ treatment might improve patients' outcome because cases indicated for TMZ treatment were resistant to other conventional treatments such as surgery, drugs, or radiation therapies.

In addition, several different and alternative treatments for TMZ-resistant pituitary tumors have been sought recently. For example, proton therapy (30), bevacizumab (31), mammalian target of rapamycin inhibitor (Everolimus) (32), TMZ, and pasireotide (33), and TMZ and capecitabine (34) have been reported. Additional clinical trials to assess optimal regimens are needed.

Conclusions

The present study suggests that MSH6 fulfills a contributory role to the efficacy of TMZ treatment for malignant pituitary adenomas. It is necessary to survey more cases and evaluate multifactor analyses to fully elucidate these findings.

Acknowledgments

This work is dedicated to Dr. Toshiaki Sano, Department of Pathology, Edogawa Hospital, who continued to work for this study to the end of his life. We thank Ms. Midori Matsuda, Department of Pathology, International University of the Health and Welfare Mita Hospital, for her special contribution to the pathological evaluation.

This study was performed as the national cooperative study by the Japan Society for Hypothalamic and Pituitary Tumors and approved by the institutional ethics committee of Teikyo University Chiba Medical Center (approval number 10-28).

Disclosure Summary: We have no disclosures, and the study received no financial support.

Abbreviations

 
• APA

  atypical pituitary adenoma

 
• CR

  complete response

 
• DAB

  diaminobenzidine

 
• GBM

  glioblastoma multiforme

 
• MGMT

  O⁶-methyl-guanine-DNA methyltransferase

 
• MMR

  mismatch repair

 
• PC

  pituitary carcinoma

 
• PD

  progressive disease

 
• PR

  partial response

 
• SD

  stable disease

 
• TMZ

  temozolomide.

References