Safety and efficacy of temsirolimus as second line treatment for patients with recurrent bladder cancer

Bladder cancer is the seventh most common cancer worldwide in men and occurs at a median age of 73 years old [1]. Age-standardized incidence rates are higher in men (9 per 1000.000) than in women (2.2 per 100.000), which parallel the mortality rates of 3.2 and 0.9, respectively [2]. More than half of cases are occurring in the most developed areas including Europe and North America, but significant variations can be observed depending on the countries [2]. Though several risk factors have been invoked, it is admitted that tobacco use is the most prevalent one and could be associated with a future rise in incidence [3]. This represents a serious healthcare burden as bladder cancer is associated with one of the highest treatment costs [4]. Most bladder cancers are urothelial carcinomas and include the two categories of non-muscle-invasive and muscle-invasive tumors, the latter representing 20–30% of newly diagnosed cases [5]. While non-muscle invasive tumors are usually of good prognosis, up to 25% of them progress to the invasive form of the disease [6]. Transurethral resection of the bladder is the treatment of choice for non-muscle-invasive bladder cancers and cystectomy is used for non-metastatic forms of muscle invasive tumors [7]. In the case of locally advanced tumors or in metastatic diseases, two first-line chemotherapies where cisplatin is associated with either gemcitabine (GC) or methotrexate, vinblastine, and doxorubicin (MVAC) have been approved and show overall response rates above 50% with a median progression-free survival (PFS) of 7–9 months and a median overall survival (OS) of 12–15 months [8]. Vinflunine was the only drug approved in 2009 as second line therapy based on a 2.4 months benefit as compared to best supportive care [9], emphasizing the need for new treatment options. For these patients, blockade of the PD1/PD-L1 immune checkpoint is an attractive strategy as recent phase II/III clinical trials showed significant improvement in tumor response, with a higher response rate for patients with PD-L1 positive tumor-infiltrating immune cells and a good tolerability [10]. This led to the approval of pembrolizumab, atezolizumab, durvalumab, nivolumab and avelumab as second line treatment for platinum pretreated patients [11‐15].

With the implementation of tumor collections and the development of new generation sequencing, a growing number of potential actionable mutations have been identified in solid tumors. In bladder cancers, numerous gene alterations have been reported in a fare percentage of tumor samples including PTEN deletions, mutations of FGFR3, TP53, RAS or RAF or mutations of several key factors of the PI3K/Akt/mTOR signalling pathway (reviewed in [16]). Deregulation of the PI3K/Akt/mTOR pathway was associated with remarkable efficacy of mTOR inhibitors such as rapamycin or everolimus in bladder cancer cells in vitro and in xenograft models [17‐20]. Temsirolimus or everolimus are approved for the treatment of metastatic renal cell carcinoma, breast cancer, mantle cell lymphomas and neuroendocrine pancreatic tumors [21‐25].

Two phase II trials have been conducted with everolimus in patients with urothelial cancer who were refractory to first line platinum-based chemotherapy and showed mild antitumor activity [26, 27]. The study of Seront et al. reported a prevalence of PTEN loss in non-responder patients,²¹ and only few prolonged responses were associated with the presence of specific mutations of the TSC1 gene [28]. A phase II trial was also conducted with temsirolimus in the same settings, but was stopped after the inclusion of 14 patients as no sufficient benefit was observed on OS, leading to the conclusion that temsirolimus had poor activity [29].

Here, we report the results of a multicenter, open-label, single-arm phase II trial evaluating the antitumor activity of temsirolimus in patients with relapsed bladder cancer after first-line chemotherapy. The primary objective was to evaluate the efficacy of temsirolimus in terms of two months non-progression. Secondary objectives were to evaluate PFS and OS as well as toxicity. Our study is the first to provide clinical evidence of a potential benefit of temsirolimus for the treatment of relapsed bladder cancers. We also discuss the potential use of PET scan (early-D15 and late-D56 diagnostic performances) as a predictive value of response to temsirolimus.

Management of patients with metastatic urothelial carcinoma of the bladder following treatment with platinum-based chemotherapy remains a major challenge given the difficulty to control these symptomatic patients and avoid severe adverse events. Prior the immunotherapy era, potential options in the therapeutic armamentarium included kinase inhibitors targeting the PI3K/AKT/mTOR pathway that was known to be activated in bladder cancer [16, 34]. Indeed, preclinical studies using bladder cancer cell models or xenografted mice demonstrated that inhibition of mTOR by everolimus inhibited bladder cancer cell growth in vitro and in xenografted mice [19]. Subsequently, clinical trials evaluating the mTOR inhibitors everolimus or temsirolimus in this population of patients have been performed but failed to meet their primary objectives despite some prolonged responses [26, 27, 29].

The results of our multicenter single-arm phase II trial showed for the first time a clinical benefit of temsirolimus for the treatment of patients with relapsed bladder cancer after first-line chemotherapy. In terms of efficacy, our study showed a two-month non progression rate of approximately 50% and an overall survival that reached approximately 7 months (5.2–9.5 CI95%). This result differs from the poor activity of the sole other trial evaluating temsirolimus as a single agent in that indication, with a median time to progression of 2.5 months and an overall survival of 3.5 months [29]. This could easily be explained by the small number of patients who were assessed (14 patients) due to the premature closing of the trial because the endpoint was not met, but it could also be attributed to a higher proportion of patients with advanced disease and poor prognosis [29]. The objective response rate of 6.7% that we observed for temsirolimus was however lower than response rates obtained in phase II/III trials testing tubulin poisons in second line setting. In the SECAVIN trial comparing cabazitaxel and vinflunine, objective response rates were of 13% and 30%, respectively [35]. In the recent KEYNOTE-045 trial comparing chemotherapy to the immune checkpoint inhibitor pembrolizumab, a response rate of 11.4% was observed in the chemotherapy arm (paclitaxel, docetaxel or vinflunine) [12]. In terms of toxicity, both trials showed a similar safety profile with equivalent rates (20%) of grade 3–4 adverse events (mainly hematologic toxicity and fatigue), indicating an acceptable tolerance in patients with an ECOG status ≤2 and a rate of visceral metastases of approximately 80%, which is highly representative of such kind of population.

The first phase II trial with everolimus reported a clinical activity with a 2-month non progression rate of 27% in a cohort of 28 assessable patients [26]. The second study was performed in a larger cohort of 45 patients and reported a 2-month non-progression rate of 51%, which is almost identical to the results of our study, despite the enrollment of a significant proportion of patients who received 2 or more additional lines of treatment following first line platinum-based chemotherapy [27].

More recently, a dual inhibitor targeting PI3K and mTORC1/2 complexes, BEZ235 was tested in urothelial carcinoma after failure of platinum-based chemotherapy [36]. Notwithstanding an unfavorable toxicity profile, few patients experienced a clinical benefit [36]. Other trials investigating the PI3K inhibitor BKM-120 or the mTORC1/2 inhibitor AZD8055 are currently ongoing and results are pending.

Together with previous studies, the results of our clinical trial reinforce the notion that blockade of the PI3K/AKT/mTOR pathway could be beneficial for a specific subset of platinum-refractory patients. However, identification of robust predictive marker of response to mTOR inhibitors is still needed to select potential responder patients. In the case of everolimus, PTEN loss was associated with poor response to the drug, whereas responder patients showed a reduction in the expression levels of proteins involved in angiogenesis [26]. Interestingly, everolimus induced some spectacular response with treatment duration of more than 26 months for one patient [27]. It was further shown that this patient harbored mutations of negative regulator of mTOR such as TSC1 that could result in a significant change in response duration as demonstrated further in other patients [37]. Similar to that study we identified 4 patients experiencing a durable response to temsirolimus of more than 30 weeks (31, 33, 43 and 52 weeks, respectively). It is possible that these patients could also harbor TSC1 mutations or other alterations of the PI3K/AKT/mTOR pathway. Though it is feasible, measuring the activity of the PI3K/AKT/mTOR pathway would need to be performed in metastases to ensure a better predictive value [36]. In the present study, we also evaluated the potential predictive value of early PET scan in the tumor response to temsirolimus, as it was suggested to be a more accurate diagnostic tool as compared to CT scan [38]. We found that early PET scan (performed at 15 days) could predict response to temsirolimus at 2.8 months in less than half of treated patients, indicating that it cannot be used as a reliable marker.

When our study was started, no result of trials investigating immune checkpoint inhibitors targeting PD1 or PD-L1 was available. However, these trials led to the approval of 5 immune checkpoint inhibitors (pembrolizumab, atezolizumab, durvalumab, nivolumab and avelumab) in platinum pretreated patients based on durable response with objective response rates varying from 16 to 21% and overall survival ranging from 8 to 13 months and a superiority to chemotherapies [11‐15]. These studies also revealed that response to anti-PD-L1 inhibitors was associated with increased levels of PD-L1 expression on immune cells and relied on mutation load within the tumors [11], which probably explains why these immunotherapies are actually efficient in only 25% of this specific population of patients (Reviewed in [39]). In that context mTOR inhibitors should be considered as a relevant option for patients who are not responding to immune checkpoint inhibitors and could either be used after relapse or in combination as it is currently being tested [40]. It is difficult to predict what the response rate would be for immune checkpoint-pretreated patients, a situation that could not be envisaged at the time of our study design, nor it is possible to predict response when immune checkpoints inhibitors would be combined to mTOR inhibitors. However, clinical trials testing these strategies will probably provide us with these results in the near future.

Until recently there were only few options for patients with metastatic bladder cancers after failure to cisplatin-based chemotherapy. The recent approval of several immunotherapies really made a difference with durable responses that were not observed before and they will certainly occupy a major place in the armamentarium. Despite the positivity of this trial suggesting that termsirolimus might be a potential option for a subset of patients who are refractory to first line platinum-based chemotherapy, the risk of adverse events linked to the use of this mTOR inhibitor would have to be considered when such an option is envisaged for such a fragile population of patients, given that robust biomarkers could be identified to select the potential responders [41].