Background
The incidence of renal cell cancer (RCC) has been increasing at a slow but steady rate for the past decade, and the 5-year survival for patients with metastatic disease is only 5%-15% [
1]. The treatment of metastatic RCC (mRCC) has been considerably changed over the last 5 years due to the anti-tumor efficacy of two groups of targeted agents, namely agents that inhibit vascular endothelial growth factor (VEGF) signaling pathways and that inhibit mammalian target of rapamycin (mTOR) [
2]. Everolimus (RAD001), a new mTOR inhibitor, has been recommended as the second-line treatment option after failure of treatment by VEGF receptor (VEGFR) tyrosine kinase inhibitors [
2,
3]. An independent central review of a phase III trial of everolimus in mRCC patients suggested that treatment with everolimus was associated with a statistically significant improvement in progression-free survival (PFS) as compared to placebo treatment (4.9 months versus 1.9 months respectively) [
4]. Therefore, the mTOR pathway provides an ideal scenario for switching drug classes upon disease progression. However, how to characterize tumor or patient factors so as to choose potential patients who will benefit from mTOR inhibitors may need more investigations.
mTOR is a member of the PI3K-related protein kinase family and acts as a hub for regulating key oncogenic processes including cell proliferation, survival and angiogenesis [
2]. Autocrine binding of growth factors like VEGF to their receptor tyrosine kinases on RCC tumor cells activates PI3K, which leads to membrane localization and activation of the cytoplasmic kinase AKT. Signaling from the activated PI3K/AKT pathway increases levels of mTOR, which activates at least two separate downstream key substrates, eukaryotic initiation factor 4E (eIF4E) binding protein-1 (4EBP1) and p70 ribosomal protein S6 kinase 1 (p70S6K1) [
2]. Activation of p70S6K1 can phosphorylate the 40S ribosomal protein S6 (S6RP), and then enhances the translation of mRNAs [
2]. Phosphorylation of 4EBP1 leads to its dissociation from eIF4E, which mediates initiation of mRNAs translation [
2]. The phosphorylation status of S6RP or 4EBP1 is thus often used as a measure of mTOR activity in laboratory studies. However, whether these mTOR-associated molecules, namely AKT, mTOR, S6RP and 4EBP1, could be used as predictors of efficacy for everolimus treatment remained elusive and need further investigations.
Recently PI3K/AKT/mTOR pathway and several important markers in the pathway have been under stepwise investigation in a variety of cancers [
5‐
7], but clinical data about the prognostic value of these biomarkers in renal cancer is still limited, especially regarding the efficacy predictors. We have conducted an open-label phase 1b study of everolimus in Chinese patients with mRCC resistant to VEGFR tyrosine kinase inhibitor (TKI) therapy, suggesting that everolimus is generally well tolerated and provides clinical benefit to Chinese patients with anti-VEGF-refractory mRCC [
8]. However, predictive biomarkers for future selection of patients who may benefit from everolimus have not been identified.
We hypothesized that the activation of mTOR-associated signaling molecules may predict the efficacy of everolimus in mRCC patients. In the present study, we attempt to explore the strategy to select appropriate patients who will obtain best efficacy from everolimus therapy by retrospectively examining the expression of mTOR signaling-associated molecules, including AKT, mTOR, S6RP and 4EBP1.
Discussion
Our study has shown that everolimus is also effective in Chinese patients with mRCC, comparable to those previous reports [
4,
8,
14,
15]. More importantly, we have identified the optimal patients who may benefit from everolimus. We found that the patients with positive expression of phospho-mTOR or phospho-S6RP may show a higher clinical benefit rate or a longer progression-free survival time to everolimus treatment. Our study thus suggests that expression status of phospho-mTOR and phospho-S6RP may be applied as potential efficacy predictors for everolimus therapy in mRCC patients and indicators for selection of everolimus-responsive mRCC patients.
mTOR exerts functions mainly by activating its downsteam targets S6RP and 4EBP1 that control mRNA translation and protein synthesis [
2]. Our study found that 14/18 and 15/18 of mRCC patients are positive for expression of phospho-mTOR and phospho-S6RP respectively, and both groups included 10 patients experienced clinical benefit (71.4% and 66.7%, respectively) from everolimus. Cho et al. analyzed 20 samples with advanced RCC (12 primary and 8 metastatic specimens) who were treated with temsirolimus, another mTOR inhibitor [
12]. They reported a positive association between phospho-S6RP expression and clinical response to temsirolimus. In their study, the numbers of patients with low, intermediate and high expression of phospho-S6RP were 4, 5 and 11, respectively. All 4 patients with low phospho-S6RP expression had progressive diseases. Three of 5 patients (60%) with intermediate expression of phospho-S6RP and 7 of 11 patients (64%) with high expression of phospho-S6RP experienced clinical benefit from temsirolimus. The average CBR of patients with intermediate or high expression of phospho-S6RP in their study was 62%, which is similar to our results (66.7%). A trend toward a positive association between positive phospho-S6RP expression and clinical benefit from everolimus was also noted in the present study, although the difference was not significant. The discrepancy may be attributed to the different methods, the different numbers of patients and the different categories of patients (mRCC patients in our study) in Cho’s and our study. Additionally, in vitro data revealed the presence of another S6K, most likely p90rsk, which may be directly phosphorylated and activated by ERK1/2 [
16]. This suggests that S6K phosphorylation might be sometimes independent of mTOR activation, which may also contribute to the discrepancy described above. However, there was a significantly longer median PFS in patients with positive expression of phospho-S6RP as compared to patients with negative expression, which indicated that the expression status of phospho-S6RP should still be a predictive factor of efficacy to everolimus because of the greater importance of PFS than CBR when evaluating the efficacy of drug in clinical trials in advanced cancer patients. In both of Cho’s and our studies, the sample size was small. So the association of phospho-S6RP with clinical response to mTOR inhibitors may need to be validated in larger cohort.
As one of the two downstream substrates of mTOR pathway, 4EBP1 activation, similar to S6RP, was also suggested to be a prognostic factor for survival and predictor for clinical outcomes in malignancies [
17‐
21]. However, the mechanism of eIF4E activation still remains controversial. Most recently, Sun et al. reported a diverse pattern of phospho-4EBP1 as compared to phospho-S6RP regarding their association with tumor grade and disease stage [
22]. Nawroth et al. also found that mTOR or AKT expression or activation only regulated phosphorylation of S6K1 but not 4EBP1 [
23]. These findings suggested that phospho-4EBP1 and phospho-S6RP may not be activated equally in the mTOR pathway. This was supported by the results of our study, showing that levels of phospho-4EBP1 had no impact on the CBR and PFS in mRCC patients treated with everolimus. In addition, function of eIF4E may also be enhanced as a result of signaling through the RAS/MEK/ERK pathway besides the PI3K/AKT/mTOR pathway [
24]. Therefore, the function and activity of 4EBP1 in mRCC still requires further investigation.
As an upstream regulator of mTOR, AKT plays a central role in the activation of mTOR and is expected to behave synchronously with mTOR. However, we found that expression status of phospho-AKT did not affect the CBR and median PFS of patients who were treated with everolimus. We proposed that AKT may not be the only pathway that activates mTOR. A recent study has suggested that the inhibition of MEK1/2 results in activation of AKT but not mTOR/S6K1 or 4EBP1 [
23]. Since the current study included only 18 mRCC patients, future investigations evaluating the association of phospho-mTOR and phospho-S6RP with clinical response to everolimus in mRCC patients with larger sample size should be encouraged.
In the study with 419 clear cell RCC patients covering all stages of disease, cumulative number of altered biomarkers in mTOR pathway is an independent predictor of clinical outcomes [
11]. Our analysis based on the combined expression of phosphorylated markers showed that co-expression of phospho-mTOR, phospho-S6RP and/or phospho-4EBP1 (either two or three) markers may be associated with a longer PFS. This finding indicated that combining examinations of multiple markers may improve the predictive value of these markers regarding response to targeted therapy with mTOR inhibitors, which need to be supported by further validating evidence from larger cohort.
Ideal study for identifying predictive biomarkers for clinical response to mTOR inhibitors may need to examine the expression of these markers after inhibitor treatment. Due to the initial objective of our study and the unwillingness of patients to provide samples after treatment, the final effects of everolimus on the proposed predictive markers in these patients were unavailable at present. Since a standard treatment schedule was used in the clinical trial [
8], it may be predicted that the mTOR targets were effectively inhibited by everolimus, which, however, need validations in future trials. In both of our study and Cho’s [
12], one apparent limitation is the small sample size. To avoid this limitation, multi-center investigations may be required for evaluation of the significance of these biomarkers in predicting clinical response to mTOR inhibitor therapy. Quantitative methods used in IHC analysis may also be a cause of discrepancies between studies. Our study used a similar strategy to quantify the staining signals [
11,
12] and evaluated the significance of these biomarkers by divided the staining results into negative or positive categories. Digital quantitative techniques for IHC results may finally help to reconcile the differences between studies. However, our study, at least, provided a potential selecting strategy for oncologists who need to treat mRCC patients who failed to the first-line anti-VEGF treatment.
Acknowledgements
We thank Dr. Taoyong Chen for language editing and for assistance in revision process. We also appreciate Drs. Lili Mao, Bin Lian, Bixia Tang and Xuan Wang for assistance in acquisition of data. This work was supported by grants from the National Natural Science Foundation of China (81172196, 81102068, 81272991), the Doctoral Fund of Ministry of Education of China (20110001120070, 20120001110048), the Beijing Nova Program (Z121107002512042, 2013027), the Program for Beijing Science Topics (Z111107058811101), the Program for Beijing Medical Disciplines Leaders (2011-2-25), and the Peking-Tsinghua Center for Life Sciences and Peking University Cancer Hospital Research Foundation (2013–21).
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SL, YK, LS, and JG participated in the study design. SL and YK carried out the immunohistochemical analyses. SL, ZC, XS, and CC participated in data collection and analysis. All authors participated in the interpretation and manuscript writing. SL, YK, and JG participated in editing and proof reading. All authors read and approved the final manuscript.