The Akt/mammalian target of rapamycin (mTOR) pathway, downstream from phosphatidylinositol 3-kinase (PI3K), mediates cell survival and proliferation. Although this pathway reportedly contributes to the progression of synovial sarcoma, its prognostic impact has not been clarified.
The authors analyzed clinicopathologic data and phosphorylation status of Akt (a serine/threonine kinase also known as protein kinase B), mTOR, the eukaryotic translation initiation factor 4E binding protein (4E-BP1), and the S6 ribosomal protein by immunohistochemical analysis of 120 formalin-fixed, paraffin-embedded samples and by Western blot analysis of 24 frozen samples from 112 patients with synovial sarcoma.
Akt, mTOR, 4E-BP1, and S6 were activated in 76.5%, 67.6%, 59.6%, and 42.6% of samples, respectively. Immunohistochemically positive phosphorylated (p) mTOR (pmTOR) and p4E-BP1 results were correlated with higher mitotic activity, and positive p4E-BP1 results were correlated with greater necrosis. No mutations around the hot spots in the PI3K catalytic subunit α (PI3KCA) and Akt1 genes were observed. In multivariate analysis of clinicopathologic parameters, frequent mitosis was a risk factor for shorter overall survival; and male sex, visceral location, larger tumor size, and frequent mitosis were identified as risk factors for shorter event-free survival. Positive pmTOR and p4E-BP1 results were correlated significantly with shorter overall survival, and positive p4E-BP1 results were correlated with shorter event-free survival in univariate analysis. Positive pAkt results were associated significantly with shorter event-free survival in multivariate analysis.
Synovial sarcoma (SS) is regarded as a relatively chemosensitive sarcoma; when the disease is advanced, the prognosis remains devastating. Phosphoinositide 3-kinase (PI3K) and the downstream Akt/mammalian target of rapamycin (mTOR) pathway have essential roles in modulating cellular functions, activating molecules like ribosomal protein S6 kinase and the eukaryotic translation initiation factor 4E (eIF4E)-binding protein (4E-BP1), which contribute to the regulation of cell size, proliferation, and survival.[1, 2] Therefore, the Akt/mTOR pathway is regarded as oncogenic and has the potential to be a therapeutic target. Several previous studies have demonstrated activation of the Akt/mTOR pathway and its contribution to cell survival and proliferation in SS cell lines[3-6] as well as the effects of the clinical application of therapeutic agents targeting mTOR in soft tissue sarcoma (STS). The prognostic impact of the activation of this pathway, however, remains to be clarified.
Although this pathway is activated in response to extracellular signals, such as growth factors and cytokines, aberration of phosphatase and tensin homolog (PTEN) is a potential activator of this pathway. In addition, activating mutations in PI3K catalytic subunit α (PIK3CA) and Akt1 have recently been recognized. PIK3CA mutations in human neoplasms have gained increasing attention since the landmark study of Samuels et al, and “hotspots” of mutations have been reported in exon 9 (E542, E545) and exon 20 (H1047). Akt1 mutation (E17K) has been reported as another activator. In this study, we investigated the phosphorylation status of Akt (a serine/threonine kinase also known as protein kinase B), mTOR, 4E-BP1, and S6 in a large series of SS and evaluated the relation between Akt/mTOR pathway activation and clinical and histopathologic features.
MATERIALS AND METHODS
Patients and Materials
We used the SS samples obtained from 112 patients who were registered at the Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan. One-hundred twenty formalin-fixed, paraffin-embedded samples were prepared (103 primary tumors, 11 recurrent tumors, and 6 metastatic tumors). The primary tumors could be compared with the recurrent lesions in 6 cases and with the metastatic lesions in 2 cases. Information about adjuvant therapy was obtained for 56 patients. Twenty-seven patients had undergone chemotherapy. Among them, samples obtained before chemotherapy, including biopsy specimens, were available in 14 patients. In the remaining 13 patients, tumor cells with degenerative changes and necrosis were not taken into consideration in histologic or immunohistochemical analysis. The best clinical response to chemotherapy was judged according to the Response Evaluation Criteria in Solid Tumors. Western blot analysis was performed on the frozen samples from 24 patients; of those, 9 samples could be compared with normal tissue samples (skeletal muscle), and 3 were paired with their corresponding recurrent or metastatic tumors. DNA extracted from frozen tumor samples from 35 patients was analyzed for gene mutations.
Tumor location was categorized as either distal extremity, proximal extremity and trunk, or visceral organs. Each tumor was classified histologically into monophasic, biphasic, or poorly differentiated according to the World Health Organization classification. The extent of necrosis and mitosis was evaluated according to the French Federation of Cancer Centers (FNCLCC) grading system. The seventh edition of the American Joint Committee on Cancer (AJCC) staging system was applied to each case. The Institutional Review Board at Kyushu University approved this study (permission code 22-75).
The immunohistochemical study was performed as previously described. The rabbit polyclonal antibodies for phosphorylated (p) Akt (pAkt) (serine 473 [Ser473]; 1:50 dilution), pmTOR (Ser2448; 1:100 dilution), pS6 (Ser235/236; 1:100 dilution), and p4E-BP1 (threonine 37/46 [Thr37/46]; 1:400 dilution) (Cell Signaling Technology, Danvers, Mass) and the mouse antibody for Ki-67 (MIB-1) (1:100 dilution; DAKO, Carpinteria, Calif) were used as primary antibodies. Each section was evaluated independently by 2 investigators (N.S. and K.K.) and was reconfirmed by another investigator (Y.O.). The percentage of immunoreactive cells and staining intensity were evaluated in the most representative area, and the samples were judged “positive” (and, thus, the molecule was defined as “activated”) when >10% of tumor cells were stained more strongly than adjacent endothelial cells (ECs).[12, 13] The percentage of MIB-1–immunopositive cells in the most “active” area was referred to as the MIB-1 labeling index (LI).
Western Blot Analysis
The procedure for Western blot analysis was described previously. In addition to the antibodies used for immunohistochemistry, total (t) Akt (tAkt), tmTOR, t4E-BP1, and tS6 (Cell Signaling Technology) were used as the primary antibodies (1:1000 dilution). For internal controls, anti-actin (1:20,000 dilution; MP Biomedicals, Irvine, Calif) and anti-glyceraldehyde 3-phosphate dehydrogenase (anti-GAPDH) (1:10,000 dilution; Santa Cruz Biotechnology, Santa Cruz, Calif) mouse monoclonal antibodies were used. Each protein was quantified, and the extent of protein phosphorylation was calculated against its total form (eg, pAkt [tumor]/tAkt [tumor]). PTEN was quantified as follows: PTEN (tumor)/GAPDH (tumor). Each membrane included a loading control for quantification.
The mutational analysis was performed using the same protocol and primers that were used in our previous study.
Continuous variables are presented as mean ± standard deviation values. The chi-square test, the t test, the Fisher exact test, and the Mann-Whitney U test were used as appropriate to evaluate associations between 2 variables. A Tukey multiple comparison test was applied when needed. The survival correlations were illustrated with Kaplan-Meier curves, and survival analyses were performed using the log-rank test. In the multivariate analysis, a Cox proportional hazards model was used to examine factors. A 2-sided P value < .05 was considered to indicate statistical significance.
Patients and Clinicopathologic Parameters
The clinicopathologic data and the results from survival analysis of all 112 patients are summarized in Table 1 and in Figure 1. Survival data were available for overall survival (OS) in 108 patients (96.4%), who had a follow-up ranging from 1 month to 278 months (median, 51 months) and a 5-year OS rate of 62%; and data were available for event-free survival (EFS) in 65 patients, who had a follow-up ranging from 2 months to 252 months (median, 24 months). The younger (<30) patients tended to receive chemotherapy (P = .0277) and radiation (P = .0147) and to have the SS18-SSX type 2 fusion gene transcript rather than the type 1 transcript (P = .0002). Proximal tumors (P = .0013) and deep tumors (P = .009) were large in size. Frequent mitosis was observed in proximal tumors compared with distal tumors (P = .027) and in visceral tumors compared with somatic tumors (P = .0364). Large tumors exhibited frequent mitosis (P = .0435) and larger areas of necrosis (P = .0004). Poorly differentiated SS had frequent mitosis (P = .0375) and larger areas of necrosis (P = .0115) than the other (monophasic or biphasic) histologic subtypes.
Table 1. Clinicopathologic Parameters and Survival Analysisa
5-Year Survival Rate, %
No. of Patients
Abbreviations: AJCC, American Joint Committee on Cancer; Bi, biphasic synovial sarcoma; D, distal extremities; EFS, event-free survival; FNCLCC, French Federation of Cancer Centers; HPF, high-power fields; LI, labeling index; Mono, monophasic synovial sarcoma; NA, not available; OS, overall survival; P, proximal extremities and trunk; Poor, poorly differentiated synovial sarcoma; V, viscera.
If patients had recurrent or metastatic tumors, then the first obtained lesions were evaluated.
The results of immunostaining for pAkt, pmTOR, p4E-BP1, and pS6 are illustrated in Figure 2. The positive ratios for pAkt, pmTOR, p4E-BP1, and pS6 were 76.5%, 67.6%, 59.6%, and 42.6%, respectively. The cases in which ECs failed to reveal any staining were excluded from evaluation (pAkt, 1 tumor; pmTOR, 15 tumors; p4E-BP1, 6 tumors; pS6, 12 tumors). The immunostaining was evaluated in nonepithelioid cells compared with ECs, because much stronger staining was observed in the cytoplasm of epithelioid tumor cells in some biphasic SS (Fig. 2B). Of 32 biphasic SS tumors, this epithelioid-predominant staining was recognized in 19 tumors for pAkt, in 22 tumors for pmTOR, in 15 tumors for p4E-BP1, and in 19 tumors for pS6. Whereas positive staining for all of the antibodies used in this study was recognized mainly in the cytoplasm of SS cells, unequivocal nuclear staining was observed in 72% for pAkt, in 68% for pmTOR, and in 27% for p4E-BP1; and stronger staining in nuclei than in cytoplasm was recognized in 14%, 13%, and 9%, respectively. Nuclear staining for pS6 was not evident. No significant correlation was observed between these staining patterns and other parameters or prognosis (data not shown).
The results of immunohistochemistry and clinicopathologic data are summarized in Table 2. The positive results for pAkt, for pmTOR and for downstream p4E-BP1 or pS6 were correlated with each other. The positive results for pmTOR and p4E-BP1 were correlated with higher mitotic activity and FNCLCC grade. The p4E-BP1–positive tumors also tended to have larger areas of necrosis and higher AJCC stage. Immunopositivity for pmTOR and for p4E-BP1 were significant risk factors for a poorer prognosis. The Kaplan-Meier survival curves for OS according to the immunohistochemical results are displayed in Figure 3. Correspondent metastatic lesions also were analyzed in 2 patients, and local recurrences were analyzed in 6 patients. The immunostaining results from recurrent or metastatic lesions appeared to correspond with the results from primary lesions in the majority of patients, but the correspondence was not statistically significant: pAkt (7 of 8 samples; P = .1071), pmTOR (5 of 6 samples; P = .3333), p4E-BP1 (6 of 8 samples; P = .4286), and pS6 (4 of 6 samples; P > .9999).
Table 2. Immunohistochemical Results and Statistical Analysis
Twenty-seven patients received adjuvant chemotherapy (17 preoperative and 15 postoperative) and/or chemotherapy for progressive disease (7 patients). Most adjuvant chemotherapy regimens (25 of 32 patients) were a combination of doxorubicin and ifosfamide, and 12 were combinations with other drugs (mostly cisplatin and/or vincristine). Second-line chemotherapy was usually an ifosfamide-based regimen. The patients who had received postoperative chemotherapy had a better prognosis (OS, P = .2291; EFS, P = .014). There was no significant difference in immunohistochemical expression of pAkt, pmTOR, p4E-BP1, or pS6 between samples that were obtained before (n = 13) and after (n = 43) chemotherapy (P = .7166, P = .4684, P = .3087, and P = .2850, respectively; Fisher exact test). The results from the prognostic implications of these molecules were similar in stratified log-rank tests that were adjusted for the presence of chemotherapy (pAkt: OS, P = .0989; EFS, P = .1423; pmTOR: OS, P = .0019; EFS, P = .0862; p4E-BP1: OS, P = .0093; EFS, P = .0022; pS6: OS, P = .2928; EFS, P = .8436).
Information about the clinical response to chemotherapy was obtained from 15 patients, including 3 patients who had a partial response, 9 patients with stable disease, and 3 patients with progressive disease. The responses indicated no significant impact on survival (OS, P = .3995; EFS, P = .1287). Patients who had pAkt-positive tumors tended to show better responses (P = .0486; Mann-Whitney U test), but no other molecules were correlated with responses to chemotherapy (pmTOR, P > .9999; p4E-BP1, P = .7142; pS6, P > .9999).
A multivariate analysis of clinicopathologic parameters demonstrated that frequent mitosis (P = .0108) was a poor prognostic risk factor for OS; and male sex (P = .0041), visceral location (P = .002), large tumor size (P = .0093), and frequent mitosis (P = .0063) were poor prognostic risk factors for EFS. MIB-1 LI, AJCC stage, and FNCLCC grade were excluded in this multivariate analysis, because they are determined or affected by other parameters. Each immunohistochemical parameter was adjusted by the above 4 clinicopathologic factors (sex, location, tumor size, and mitotic figures) (Table 3), and the multivariate analysis indicated that only positive pAkt results were associated significantly with shorter EFS.
Table 3. Multivariate Survival Analysis for Immunohistochemical Parameters Adjusted by Sex, Location, Tumor Size, and Mitosis
Akt and mTOR, as well as the phosphorylated form of each, were detected in almost all tumor samples (Fig. 4A). The phosphorylation scores for immunohistochemically positive samples were higher than those for negative samples, but the difference was not statistically significant (positive samples vs negative samples: Akt [mean ± standard deviation phosphorylation score]: 0.51 ± 0.27 vs 0.30 ± 0.25; P = .2424; mTOR: 0.38 ± 0.33 vs 0.17 ± 0.12; P = .3083; 4E-BP1: 0.50 ± 0.32 vs 0.29 ± 0.42; P = .1643; S6: 0.62 ± 0.32 vs 0.34 ± 0.22; P = .1088).
The tumor samples paired with their non-neoplastic counterparts are illustrated in Figure 4B. Densitometric analysis demonstrated that Akt was phosphorylated to a greater extent in tumor samples than in non-neoplastic tissue, with mean ± standard deviation values of 0.46 ± 0.27 and 0.14 ± 0.2, respectively (P = .0376; t test). mTOR was phosphorylated in a similar tumor-predominant fashion (tumor vs non-neoplastic tissue: 0.44 ± 0.24 vs 0.19 ± 0.15; P = .0727). 4E-BP1 and S6 phosphorylation was less intense in tumors than in non-neoplastic tissue (4E-BP1: 0.27 ± 0.32 vs 0.48 ± 0.23; P = .0108; S6: 0.27 ± 0.34 vs 0.33 ± 0.3; P = .1618), although 6 of 9 samples had neither the phosphorylated form nor the total form of S6 in normal tissue. All tumor samples had higher PTEN expression compared with normal tissue.
Three tumor samples were compared with their recurrent or metastasized lesions (Fig. 4C). The expression and phosphorylation patterns of Akt, mTOR, and PTEN were similar. The expressions of p4E-BP1 and pS6 were reduced in the recurrent lesions.
We identified no mutations (0 of 35 samples) around the hot spots in the Akt1 gene (E17) or the PIK3CA gene (E542, E545, and H1047).
It has been established that the Akt/mTOR pathway is highly activated in various malignant tumors. Several previous studies have demonstrated the activation of this pathway and its contribution to cell survival and proliferation in SS in vitro.[3, 5] However, the expression profiles of the molecules along the Akt/mTOR pathway have been analyzed in a relatively small number of clinical materials.[5, 13, 14] Some authors investigated Akt/mTOR pathway involvement in STS series and demonstrated a correlation between Akt activation and a worse prognosis or a greater probability of metastasis13; whereas, to our knowledge, no investigations have focused on SS as a single entity, and the prognostic impact of this pathway on SS has remained to be clarified.
In previous large studies, several definite adverse clinical prognostic factors were identified. We observed that frequent mitosis, sex (male), visceral location, and large tumor size were poor prognostic factors in multivariate analysis. High mitotic activity[16, 17] and large tumor size[16-19] were consistently reported as adverse prognostic factors for SS. To our knowledge, the prognostic implications of visceral locations have not been assessed; however, truncal/proximal tumors reportedly confer a worse prognoses than tumors of the distal extremities.[16, 18] We also observed that male sex was an unfavorable prognostic feature. This difference in survival according to sex also was observed in previous large studies.[17, 20] Takenaka et al documented that fewer women than men presented with metastasis. In our series, women tended to have lower AJCC stage than men (P = .0945), but no other parameters were correlated with sex.
Immunopositivity for pAkt, pmTOR, and p4E-BP1 also was identified as an adverse prognostic factor in this study, whereas immunopositivity for pS6 was not. The positive results for pAkt, pmTOR, and downstream p4E-BP1 and pS6 were correlated with each other, suggesting that these molecules are activated successively in a “pathway.” The univariate prognostic factors pmTOR positivity (+) and p4E-BP1 (+) were correlated with higher mitotic activity, and p4E-BP1 (+) was correlated with a larger extent of necrosis. This pathway may be associated with cellular proliferation through the phosphorylation of 4E-BP1. In our previous study, we demonstrated that these phosphorylated proteins were expressed in the cytoplasm of leiomyosarcoma cells, whereas strong nuclear staining was observed in a certain number of SS samples. Moreover, some tumors had extensive staining in the cytoplasm of epithelioid SS cells. This epithelioid-dominant staining pattern was mentioned in a previous study. However, its significance remains unclear, because this pattern had no correlation with other clinicopathologic parameters in the current study. Patients who had pAkt-positive tumors had a better clinical response to chemotherapy than those who had tumors that were negative for pAkt, and this tendency was not observed for the other molecular markers. This is probably because of the diverse functions of Akt that are not related to mTOR, such as cellular proliferation, survival, and antiapoptosis.
The expression of pAkt and pmTOR in tumor tissue also was demonstrated in Western blot analysis, although the correlation with immunohistochemistry results was not statistically significant, probably because of the small number of samples. Several patients had lower expression of p4E-BP1 and pS6 in tumors than in non-neoplastic tissues. This may correspond to the finding that the immunohistochemically positive ratios of these proteins were relatively low. Another explanation is that they may be activated by kinases other than mTOR[21, 22] in skeletal muscle tissue, which was used as a control in the current study. These proteins had lower expression in recurrent or metastatic lesions than in primary tumors according to our Western blot analysis; however, the cause of this phenomenon remains unknown.
Abnormalities in several molecules may be responsible for activation of the Akt/mTOR pathway. We observed somewhat increased PTEN expression in tumor tissue proportional to Akt expression. PTEN expression may be induced as a result of negative feedback from activation of the Akt pathway rather than being responsible for the activation. Another factor that may activate this pathway is a mutation in PIK3CA or in Akt1. We screened 35 SS samples but did not detect any mutations. These results are consistent with the recent findings of some investigators who reported only rare PIK3CA mutation in SS.[5, 23, 24]
Previous studies have indicated another potential factor in the Akt/mTOR pathway: the activation of receptor tyrosine kinases or cytokine receptors. Epidermal growth factor receptor (EGFR),[14, 25, 26] insulin-like growth factor-1 receptor (IGF1R),[14, 27] and platelet-derived growth factor receptor α (PDGFRα)[6, 28, 29] have been proven responsible for Akt/mTOR pathway activation in SS, and some preclinical studies have lent credence to targeting this pathway.[6, 29] Clinical trials of molecular targeting drugs also have been undertaken in STS. Ridaforolimus, an mTOR inhibitor, produced significantly prolonged progression-free survival in a clinical trial targeting advanced STS, although no significant difference was indicated for OS. Pazopanib, a multitargeted kinase inhibitor, also exhibited potential effectiveness against STS, especially in a group of patients with SS; and Hosaka et al demonstrated that this drug inhibited the growth of SS cells accompanied by the suppression of the Akt pathway. Moreover, Ho et al revealed that PDGFRα, a pazopanib target, can mediate intrinsic resistance to rapamycin in SS. These findings suggest that the activation mechanism of the Akt/mTOR pathway is multimodal and that we should seek a combination of drugs targeting not only mTOR but also the upstream receptors and the feedback loops.
At this point, we cannot conclude that any of the markers we studied are useful in making the decision whether targeting this pathway is the superior strategy for each patient. In the phase 2 study of ridaforolimus, pAkt, pS6, and 4E-BP1 were not identified as good predictors of a clinically beneficial response. Immunohistochemically “negative” patients were able to benefit from targeted therapies, because some extent of immunostaining (weaker than or equal to that in ECs) was observed in most tumor cells, even among the “negative” patients in our study. At the least, pS6 would not be suitable based on the results from our immunohistochemical analysis, in which no clinicopathologic parameters were correlated with pS6 positivity. The receipt of chemotherapy seemed to have little impact on this pathway, and targeted therapy against this pathway is a potential candidate for patients who have received conventional chemotherapy.
In conclusion, we have demonstrated that the Akt/mTOR pathway, including downstream 4E-BP1, is activated and is associated with aggressive clinical behavior in SS. These findings support the validity of molecular therapy targeting this pathway in patients with SS.
Dr. Oda is supported by a Grant-in-Aid for Scientific Research (B) (21390107) from the Japan Society for the Promotion of Science.