Background
Hepatocellular carcinoma (HCC) is the sixth most common cancer globally, and the third leading cause of cancer mortality both in Hong Kong and worldwide [
1,
2]. The outlook of patients with unresectable HCC is poor. To date, the only systemic agent that has been shown to provide survival benefit is sorafenib [
3,
4]. In parts of the world including Hong Kong, HCC patients often present with advanced disease stage, but the use of sorafenib has only been approved in recent years as standard therapy.
It has been well-established that numerous genetic abnormalities are involved in HCC; comprehensive genomic analyses shows that components of the phosphatidylinositol-3 kinase (PI3K)/Akt/mTOR pathway are dysregulated in 40-50% of HCC [
5-
7]. On the other hand, a meta-analysis of over 450 patients with HCC who received liver transplant demonstrated lower rates of recurrence and mortality for patients who received the mTOR inhibitor (mTORI), sirolimus, for immunosuppression [
8]. The expansion of mTORIs as a therapeutic strategy for HCC was also strengthened by their successes in other cancers [
9-
12]. In various HCC models, mTORIs significantly reduced tumour volume and angiogenesis, delayed tumour growth and increased survival [
5,
6,
13-
16].
Everolimus had initially been evaluated in HCC in phase I and II studies. A US study achieved an MTD of 10 mg/day [
17]; among the 25 patients enrolled, 10 achieved stable disease, one achieved partial response, and median survival was 8.4 months. In another study, Taiwanese patients tolerated only a daily dose of 7.5 mg, and the median survival was 7.7 months [
18]. However, the efficacy of everolimus in HCC has not been confirmed by the recently reported global phase III study (EVOLVE-1,
NCT01035229) [
19].
Temsirolimus is a prodrug of sirolimus; it is administered intravenously and has a long half-life of 73 hours. To date, there has been limited clinical data on the use of temsirolimus in HCC patients who often suffer from chronic liver disease. We conducted a phase I/II study of temsirolimus (Torisel®) in patients with unresectable HCC, majority of whom had concomitant hepatitis B virus-related chronic liver disease. The objectives in the phase I study were to determine dose limiting toxicity (DLT) and maximum tolerated dose (MTD). Once the MTD was determined, the phase II portion of the study was conducted to determine the activity of temsirolimus.
Although promising results have been shown with temsirolimus in a number of malignancies, there has been very limited data on potential biomarkers that could enable appropriate selection of tumours which are likely to undergo a favorable clinical response. Further, the failure to demonstrate efficacy of everolimus in the EVOLVE study has highlighted the potential importance of appropriate patient selection. Thus, in the current study, an exploratory analysis was also conducted to determine if the expression of stathmin, pS6, pMTOR and p-AKT might be predictive for response to temsirolimus in HCC.
Methods
Eligibility criteria included: Histologically/cytologically confirmed unresectable HCC; ECOG ≤2; measurable disease; life expectancy > 12 weeks; absolute neutrophil count ≥ 1.5 × 109/L, platelets ≥ 80 × 109/l, serum creatinine ≤ 150 μmol/L, total bilirubin ≤ 30 umol/l, albumin ≥ 28 g/l, alanine transaminases ≤ 5.0 × UNL (institutional upper normal limit), alkaline phosphatase ≤ 6 × UNL, prothrombin time ≤ 4 sec of ULN, and absence of clinical ascites.
The main exclusion criteria were Child’s B or C cirrhosis, use of other systemic treatments within 3 weeks prior to study entry; prior use of mTORI; significant cardiovascular disease; severe impairment of lung function; poorly controlled diabetes mellitus; and ≥ grade 2 pre-existing neuropathy.
Written consent was sought from individual patient to participate in the study and for the exploratory analysis that involved the use of tissue obtained for diagnostic purpose
. This study was approved by the Clinical Research Ethics Committee of the Joint NTEC-Review Board of the Chinese University of Hong Kong, and has been registered in ClinicalTrials.gov (Id:
NCT00321594).
Pretreatment evaluation
All patients underwent complete medical history and physical examination, blood profiles including complete blood counts, renal and liver functions, fasting glucose and lipids, clotting profiles, alpha-fetoprotein (AFP), and hepatitis B surface antigen (HBsAg), hepatitis C antibody (anti-HCV), chest x-ray and CT scan of abdomen and/or other disease sites were performed.
Treatment plan
Temsirolimus was added to 250 mL of 0.9% sodium chloride and administered intravenously over 30 minutes weekly, every 3 weeks. All patients received premedication with diphenhydramine 25 mg or 50 mg IV bolus dose 30 minutes prior to temsirolimus. Standard anti-emetics included at least a 5-HT3 antagonist. Patients who were HBsAg seropositive were also given lamivudine prior to study treatment.
Phase I study
For the phase 1 study, there were 5 dose levels of temsirolimus: 10 (level −2), 15 (level −1), 20 (level 1), 25 (level 2) and 30 mg/week (level 3). Level 1 was the starting dose level.
DLT was defined during cycle 1 as: any grade 4 hematological toxicity; grade ≥3 non-hematological toxicity (excluding alopecia); grade 3 nausea, vomiting, or diarrhoea that did not respond to therapy; and treatment delay > 2 weeks.
The conventional 3 + 3 design was employed. Dose escalation was based on the modified Fibonacci method [
20]. The MTD was defined as the dose below which ≥ 2 of 3 or ≥ 2 of 6 patients experiencing DLT. A total of 10 patients were entered into the MTD to further define toxicity.
Treatment delay and modification
For each cycle, treatment was delayed if the ANC was <1.5 × 109/L or platelet count was < 75 × 109/ml on the scheduled day of drug administration. Patients who experienced grade 3 non-haematological toxicity, thrombocytopenia or febrile neutropenia, as well as grade 4 neutropenia continued to receive temsirolimus at the next lower dose level upon resolution of all toxicities to grade 1. For an individual, there could be a limit of two dose de-escalations for serious toxicity. The drug was discontinued for toxicities of the following nature: grade 4 non-hematological toxicities, thrombocytopenia/febrile neutropenia/recurrent grade 4 neutropenia despite dose reduction, as well as any haematological or non-haematological toxicity requiring interruption for ≥ 3 weeks.
Treatment was continued provided that toxicities were tolerable or until one of the following criteria applied: disease progression; intercurrent illness that prevented further treatment administration; unacceptable adverse events; patient’s decision; or investigator’s judgment.
Phase II study
Upon determination of MTD, patients were enrolled into the phase II part of the trial at MTD; the 10 patients at the MTD in phase I were included in the phase II analysis.
Definitions of response and toxicity
Tumour response assessment with CT every two cycles was assessed according to the Response Evaluation Criteria in Solid Tumors (RECIST) Committee [
21]. Toxicity was graded according to Common Toxicity Criteria of the National Cancer Institute (NCI-CTC v3).
Methodology for stathmin, pS6, pMTOR and p-AKT immunohistochemistry
Thirty-four patients had pre-treatment tissues available for this analysis. For immunohistochemistry, 5-μm tissue sections were prepared from each block. Tissue sections were deparaffinized, rehydrated and rinsed in distilled water. Antigen retrieval was done by using pressure cooker with 10 nM citrate buffer (pH 6.0) for 25 minutes. The endogenous peroxidase activity was then blocked by incubating the slides in 3% hydrogen peroxide in methanol for 10 min. The primary antibodies used in this study were STMN1 (1:50), pS6 (Ser235/236, 1:100), pMTOR (Ser2448, 1:50) and p-AKT (Ser473, clone D9E, 1:25) from Cell Signaling Technology (Danvers MA). The primary antibodies were incubated at 4°C overnight and chromogen development was performed using the DAKO EnVision System (Glostrup, Denmark) except for p-AKT, which was detected using the OptiView DAB IHC Detection Kit (Ventana Medical Systems).
An intensity score of 0 to 3 was assigned for the intensity of tumour cells (0, none; 1, weak; 2, intermediate; 3, strong). A proportional score was given by the estimated proportion of positive tumour cells in percentage. To assess the average degree of staining within a tumour, multiple regions were analyzed, and at least 100 tumour cells were assessed. The cytoplasmic expression was assessed by H-score system [
22]. The formula for the H-score is: Histoscore = ∑(I × Pi), where I = intensity of staining and Pi = percentage of stained tumour cells, producing a cytoplasmic score ranging from 0 to 300. The scoring was independently assessed by two assessors (AWHC and JHMT) who were not aware of the clinical outcomes.
Statistical methods
For the Phase I portion, the estimated patient number would be 14–19. For the phase II portion, the primary endpoint was progression free survival (PFS). The secondary endpoints were response according to RECIST, overall survival (OS) and toxicity. The PFS was assessed from day 1 of treatment cycle 1 to the date when objective disease progression was observed. OS was calculated from day 1 of treatment cycle 1 to the date of death. Death was regarded as a progression event in those subjects who died before disease progression. Subjects without documented objective progression at the time of the final analysis were censored at the date of their last tumour assessment. Survival curves were constructed using the Kaplan–Meier method.
The planned accrual for phase II was 30 assessable patients. Patients are considered assessable if they have completed ≥ 1 cycle of treatment or are removed from study due to disease progression. If the PFS at 3 months is ≤ 0.5, the regimen would be considered inactive. If the PFS at 3 months is ≥ 0.66, this regimen would be considered worthy of further investigation. If ≥ 18 of 30 assessable patients are observed to be progression-free by 3 months, the study would have 80% power and 0.18 significance level. An additional 6 patients (i.e. 20%) would be accrued to account for ineligibility, cancellation, major treatment violation, or other reasons. Therefore, the maximum accrual would be 36 patients (including the 10 patients from phase I at MTD). In order to observe enough events for the study, all patients would be followed up for at least 3 months.
Exploratory analysis on cytoplasmic expression of the biomarkers was viewed as hypothesis generating. The optimal cutoff for stathmin, pS6, pMTOR and p-AKT was determined by the receiver operating characteristic (ROC) curve distribution analysis [
23,
24]. Out of a total H-score of 300, the threshold for differentiating between positive and negative immunostaining were set at H-scores of 15, 120, 20 and 5 respectively; tumours were categorized as ‘low H-score’ and ‘high H-score’ depending on whether the individual score were ‘lower than or equal to’ or ‘higher than’ the respective thresholds. Response rates in terms of disease stabilization (defined as complete response [CR] + partial response [PR] + stable disease [SD] ≥ 12 weeks) and AFP drop in association with H-scores of stathmin, pS6, pMTOR and p-AKT cytoplasmic were compared using Fisher’s exact and proportional hazard model where applicable. Response assessment based on AFP was conducted for patients whose baseline AFP > 20 ng/ml and who had 2 cycles of study treatment. The drop in AFP based on baseline AFP was compared with the lowest level of AFP detected after 2 cycles of study treatment, and AFP response was defined as a > 20% decrease in AFP value [
25].
Discussion
The present study confirmed the MTD for temsirolimus in patients with chronic liver disease and advanced HCC to be 25 mg weekly, which is the approved dose for metastatic renal cell carcinoma [
9,
10]. Common adverse reactions of temsirolimus noted in this study were consistent with the reported toxicity profile of this agent, which included skin and mucosal toxicities, constitutional symptoms (fatigue, anorexia, insomnia), myelosuppression, metabolic disturbances (disturbances in glucose and lipids controls) and the uncommon but well-known occurrence of interstitial pneumonitis.
In an unselected population of advanced HCC patients, the current study reveals that the use of temsirolimus yielded a 3-month PFS of 0.47, which is lower than the pre-specified limit considered to be efficacious. The present finding is in line with that of the EVOLVE study, in which everolimus has failed to achieve the primary endpoint in improving OS in an unselected HCC patient population who had progressed on sorafenib [
19]. The discouraging result sheds light to the potential importance of suitable patient selection.
There has been limited ability to identify biomarkers for appropriate utilization of mTORIs. In the phase I study of everolimus, 11 HCC patients had pre-treatment tumour tissues available for assessment, one patient achieved PR and the tumour showed moderate to high levels of p-AKT, p-MTOR and pS6 [
17]. The key effector in the PI3K/Akt/mTOR pathway is mTOR, which has a critical role in regulating cell proliferation, survival and angiogenesis [
27,
28]. PIK3CA has also been suggested as a predictive marker for effective mTOR inhibition in breast cancer [
29,
30], unfortunately, a recent report on endometrial cancer did not support this [
31]. Further, the reported rate of mutations in the PIK3CA gene has been inconsistent in HCC varying from 0-35% [
32,
33]. Activated PI3K propels two downstream effectors: mTOR complex 2 (mTORC2) and Akt. Akt activates mTORC1 which in turn activates downstream effector, the serine/threonine kinase, S6K1. S6K1 participates in numerous cellular processes central to promoting cell proliferation, cell growth and cell cycle progression [
34,
35]. Phosphorylated mTOR and p-S6K is elevated in approximately 40% of HCC [
6,
27,
36]. It has been observed that loss of PTEN, the negative regulator of PI3K, results in robust activation of this pathway [
37,
38], and stathmin, encoded by the signature gene STMN1, has been suggested to be a more accurate immunohistochemical marker of the PTEN signature [
39]. These data have prompted us to explore the possibility of stathmin, pAKT, pMTOR and pS6 as potential biomarkers for response.
The present exploratory analyses show pMTOR to be the only marker associated with disease stabilization effect of temsirolimus. Although some studies suggested that pMTOR overexpression may have prognostic impact independent of temsirolimus, studies in different tumour types have reported conflicting results [
40-
42]. Specifically, a study in HCC patients undergoing orthotopic liver transplantation reported mTOR pathway to be active in 40% of the patients, but none of the biomarkers [PTEN, p-AKT, p-mTOR, p-p70S6K and p-4EBP-1] were associated with survival [
43]. In this current study, assessment of pMTOR in relation to presence of vascular invasion and tumour grading was attempted; unfortunately, 22 of the 34 tumour analyzed were biopsy samples which limits detail pathological assessment.
On the other hand, the effect of rapalogs on Akt may vary with drug dose, with lower doses increasing Akt activation while higher doses diminishing Akt activity [
44,
45]. In addition, the effect on Akt also varies with cell type [
46]. Thus, determining the clinical effects of different dosages of mTORIs could be an important tactic to overcoming such limitation.
Further, combining mTORIs with other systemic agents could improve clinical efficacy. The combination of everolimus and sorafenib has been reported to synergistically inhibit proliferation and tumor growth in HCC cell lines and xenografts [
14]. A phase I study of this combination in advanced HCC patients yielded an encouraging 8% PR and 60% SD [
47]. In addition, studies have shown that the activation of Akt markedly increases the resistance against microtubule-directed cytotoxic agents while mTORIs could inhibit this resistance [
48,
49].
Acknowledgements
The study was sponsored by Pfizer Corporation Inc. (drug support, funding of imaging required in the study, and funding for personnel for data entry and data analysis), and the Chinese University of Hong Kong Direct Grant for Research (Grant Ref No. 2012.1.011). The investigators were responsible for data collection, data analysis, data interpretation, and writing of the report.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
WY, FKFM, KFT and SCHY designed research directions and protocols. WY, SLC, EPH, JK, LL, CMC, JWYH, BM and SCHY acquired clinical data. JHMT, AWHC and KF. To conducted biomarker correlative analyses. WY, FKFM, HL and KL analyzed and interpreted data. WY, JWYH, JHMT, EPH and KL wrote the manuscript. All authors read and approved the final manuscript.