Phase II study of the histone deacetylase inhibitor vorinostat (Suberoylanilide Hydroxamic Acid; SAHA) in recurrent or metastatic transitional cell carcinoma of the urothelium – an NCI-CTEP sponsored: California Cancer Consortium trial, NCI 6879

Abstract

Background:

Until the advent of T cell check point inhibitors standard second-line therapy for patients with metastatic urothelial cancer (mUC) was undefined. Histone deacetylase inhibitors (HDACi) have anti-cancer activity in a variety of tumor models including modulation of apoptosis in bladder cancer cell lines. We evaluated the efficacy and toxicity of the HDACi vorinostat in patients with mUC failing first-line platinum-based therapy either in the adjuvant/neoadjuvant setting or for recurrent/advanced disease.

Methods:

Vorinostat was given orally 200 mg twice daily continuously until progression or unacceptable toxicity. The primary end point was RECIST response rate (RR); a RR > 20% was deemed interesting in a 2-stage design requiring one response in the first 12 patients to proceed to 2nd stage for a total of 37 subjects. CT or MRI scan imaging occurred every 6 weeks.

Results:

Fourteen patients were accrued characterized by: median age 66 years (43–84); Caucasian (79%); males (86%); and Karnofsky performance status ≥ 90 (50%). Accrual was terminated in the first stage as no responses were observed. Best response was stable disease (3 patients). Progression was observed in 8 patients. Two patients came off therapy prior to re-imaging and a 3^rd patient died while on treatment and was not assessed for response. Median number of cycles was 2 (range 1–11). Median disease-free survival and overall survival times were 1.1 (0.8, 2.1) & 3.2 (2.1, 14.5) months, respectively. Toxicities were predominantly cytopenias and thrombocytopenic bleeding. Two pts had grade 5 toxicity unlikely related to treatment. Two pts had grade 4 and 6 had grade 3 toxicities observed. Two patients with stable disease remained on therapy for 6+ cycles.

Conclusions:

Vorinostat on this dose-schedule had limited efficacy and significant toxicity resulting in a unfavorable risk:benefit ratio in patients with mUC. NCT00363883

INTRODUCTION

Bladder cancer is estimated to occur in 81,000 Americans in 2020 and to be result in the death of around 18,000 [1]. The large majority of all deaths from bladder cancer derive from muscle invasive cancer that recurs and/or metastasizes after local therapy [2]. Metastatic urothelial cancer is an initially chemotherapy-sensitive tumor. Platinum-based regimens remain a cornerstone of therapy for metastatic bladder cancer [3]. The median overall survival in patients treated with platinum-based regimens remains between 12 and 18 months [3]. Unfortunately, around 5% of patients are 5-year survivors [1]. More recently monoclonal antibodies directed at PD-1/PD-L1 have demonstrated activity in advanced urothelial cancer previously treated with platinum, with a single trial showing a significant improvement in overall survival for pembrolizumab compared to single agent chemotherapy[4]. Enfortumab vedotin, an antibody drug conjugate directed at nectin 4 on the surface of almost all urothelial cancer cells, and Erdafitinib, a tyrosine kinase inhibitor of FGFR1–4 with antitumor activity in cancer with FGF receptor alterations, have just received accelerated approval from the FDA [5] [6]. Despite this, a large majority of patients do not respond and rapidly become candidates for further therapy or palliative care. Given the high rate of failure of first line therapy and limited effectiveness of salvage regimens, there is strong rationale and need for exploration of new treatment options in patients with recurrent bladder cancer.

Urothelial cancer is characterized by a series of key molecular aberrations including cell cycle dysregulation, apoptotic ineptitude and enhanced signal transduction, cytokine elaboration, cellular adhesion and neoangiogenesis pathway signaling [7–12]. In addition, urothelial cancer is notable for a high prevalence of gene silencing by gene or gene promoter methylation [13–15]. By maintaining the dynamic equilibrium of the acetylation status of highly conserved lysine residues on histones, histone deacetylases regulate chromatin remodeling and gene expression. DNA methylation and the hypoacetylation of core nucleosomal histone proteins lead to the tight coiling of chromatin, thereby silencing the expression of a variety of genes, including those implicated in the regulation of cell survival, proliferation, differentiation, and apoptosis [16]. Vorinostat is a linear hydroxamic acid that inhibits histone deacetylase activity (HDAC) [17,18]. It causes hyperacetylation of all core histone proteins, H2A, H2B, H3, and H4, and inhibits purified HDAC activity in vitro [17,19]. Histone deacetylase inhibitors generated significant interest as anti-cancer agents due to their ability to cause growth arrest, terminal differentiation and/or apoptosis in preclinical carcinoma models [20]. Small molecule inhibitors of HDAC, including vorinostat, induce differentiation and suppress cell proliferation in cancer cell lines including T24 bladder carcinoma cells [21–23,19].The activity of vorinostat was investigated using the 60 cell line screen at the NCI and IC50s ranging from approximately 500 nM to 5 μM were obtained [24].

HDAC inhibitors alone or in combination with other agents restore expression of silenced genes by remodeling the tightly coiled chromatin, leading to the subsequent induction of differentiation, arrest in the progression of the cell cycle, or apoptosis [25,26]. Vorinostat induces genes with a negative regulatory effect on cell cycle progression such as p21^CIP1/WAF1, p27^KIP1 and p16^INK4A [19,27–29,12]. and others that gear the cell toward programmed death. In in vitro experiments, Vorinostat induces apoptosis alone and in combination with other agents in addition to suppression of signal transduction pathways and cytokine release [30,12]. The potential molecular features of HDAC inhibition with vorinostat that mesh with urothelial or transitional cell cancer based on preclinical evidence are summarized on Table 1.

Table 1.

Vorinostat and urothelial cancer: selected molecular actions and characteristic match-ups that provide rationale for the clinical trial

Molecular pathway or process	Urothelial or “transitional” cell cancer	Vorinostat effect
Gene or gene promoter silencing due to hypermethylation	Frequent with diminished expression of gene product [13,14]	May reverse effect on gene expression and chromatin coiling [25,36]
Apoptosis		Increased apoptosis in resistant cells [19,29,37,38]
p53	Increased expression or mutation adversely prognostic [8]	Unknown
Cell cycle
p16INK4A	Reduced expression/mutation adversely prognostic [39,40]	Increased expression [29]
p21CIP1/WAF1	Reduced expression adversely prognostic [41]	Increased expression [19,28]
p27KIP1	Frequently reduced in expression and reduced expression may be adversely prognostic [42]	Increased expression [28]
pRb	Loss adversely prognostic [7]	Unknown
Angiogenesis
VEGF	VEGF and other angiogenesis factors overexpressed [11] [43]	VEGFR expression decreased [44]
Inflammation	Produce IL6 and IL8 which are associated with progression [9,10]	Reduced expression of selected cytokines [45]

Phase I clinical trials evaluated vorinostat in patients with advanced solid tumors and hematological malignancies prior to this trial [31]. The maximum tolerated dose of vorinostat was 400 mg daily or 200 mg twice daily when administered orally. The dose limiting toxicities were anorexia, diarrhea and fatigue. Encouraging anti-tumor activity has been noted in the phase I clinical trials of vorinostat. While the dose-limiting toxicities were non-hematological (anorexia, dehydration, diarrhea, and fatigue), important hematologic adverse events included anemia and thrombocytopenia which were rapidly reversible after study drug interruption. Significant antitumor activity in hematologic malignancies and solid tumors was observed in early studies. Subsequent phase II trials demonstrated the effectiveness of vorinostat in a variety of cancers, most particularly cutaneous T cell lymphoma, for which vorinostat was approved by the FDA in May 2009 [26].

Given the preclinical and phase I trial activity of vorinostat in a variety of cancers and its mechanistic targeting of multiple regulatory pathways in urothelial cancer, we undertook a phase II study with vorinostat in patients with urothelial tumors who have progressed on or subsequent to platinum-based chemotherapy. The primary objective of the trial was to determine the response rate of advanced urothelial cancer to vorinostat in the second line setting. Secondary objectives focused on time to disease progression, overall survival, safety and toxicity and feasibility.

PATIENTS AND METHODS:

Patient Eligibility:

Eligible patients were >18 years, able to swallow pills whole without crushing, and had a pathologic diagnosis of urothelial carcinoma of the bladder or urothelium with <25% component of other cell types such as small cell, neuroendocrine or squamous cell carcinoma. Patients had metastatic or locoregionally advanced disease with lesions which were measurable, by RECIST 1.0 criteria, and were not amenable to curative surgery and/or radiation. Patients with bony metastases were allowed to participate in the study provided they also had non-osseous disease that was measurable. Patients must have had at least one prior chemotherapy regimen containing a platinum compound while those treated with a second line of chemotherapy were included provided >6 months had elapsed from the completion of the first line of chemotherapy to the start of the second. Any number of prior intravesical therapies for non-muscle invasive bladder cancer were permitted as was one experimental biologic therapy for their metastatic urothelial cancer provided this did not act through histone deacetylation or demethylation. Systemic therapy and radiation must have been completed >4 weeks prior to entering the study with recovery from related toxicities. Eligible patients had to have a Karnofsky performance score of 60% or better, a creatinine of ≤ 1.5 times the institutional upper limit of normal or creatinine clearance of ≥40 mL/min., absolute neutrophil count (ANC) ≥1500/mcL and platelet count ≥100,000/mcL. Patients with central nervous system metastases were excluded. All patients were required to use appropriate birth control. HIV-positive patients on antiviral therapy were not eligible, as well as patients on medicines known to inhibit or induce the P450 cytochrome metabolism pathway of vorinostat. The protocol was approved by the institutional review boards of the participating institutions, and all patients provided written informed consent.

Treatment with Vorinostat:

Vorinostat was initially given at 200 mg twice daily orally each day on a 21-day cycle. Premedication and as needed therapy were permitted at the clinician’s discretion. Loperamide was mandated for drug-induced diarrhea and patients were instructed regarding maintenance of oral hydration. Colony stimulating factors were not given prophylactically. Vorinostat was held for grade ≥3 diarrhea or stomatitis. Delays of up to 3 weeks for toxicity were allowed. Dose reduction for toxicity was mandated for grade ≥3 myelosuppression, febrile neutropenia or other grade ≥3 non-hematological toxicity. The first lower dose level was vorinostat 200 mg twice daily for 14 of 21 days followed by a 7-day break. The next lower dose was vorinostat 300 mg per day for 14 of 21 days in each cycle. Re-escalation was not permitted. Patients needing more than two dose reductions were removed from the study. Treatment was continued until disease progression, unacceptable toxicity or patient withdrawal.

Efficacy and Safety Evaluation:

Tumor response was evaluated by computed tomography (CT) or magnetic resonance imaging (MRI) every 2 cycles (6 weeks) using RECIST version 1.0. Progression-free survival was defined as the interval between the date of start of treatment and the date of either documentation of disease progression (either radiologic or symptomatic progression) or death owing to any cause. Patients who were alive and not known to have progressive disease were censored at the date the patient was last known to be progression-free. Overall survival was defined as the interval between the date of start of treatment and the date of death.

A physical examination was conducted, and vital signs, hematology, clinical chemistry, urine protein to creatinine ratio were assessed every cycle before treatment. Toxicity was graded every 21 days and on an ongoing basis throughout the study using NCI CTCAE, version 3.0.

Samples of blood, tumor tissue and buccal mucosa were obtained for correlative markers but destroyed or returned due to withdrawal of NCI resources for budgetary reasons.

Statistical Considerations:

Response rate was assessed in a 2-stage Simon optimum accrual design (12+25) with an estimated accrual rate of 3 patients per month. A maximum of 37 patients were to be accrued to rule out a null hypothesized response rate of 5% versus an alternative response rate of 20%. With the proposed design, the probability of falsely declaring a regimen with a 5% response rate as warranting further study was ≤ 0.10 (alpha) and the probability of correctly declaring an agent with a 20% response rate as warranting further study was ≥ 0.90 (power). In the first stage, accrual continued until 12 patients were evaluable for response. If no responses were observed, accrual was to stop with the conclusion that this regimen of oral vorinostat is not promising for further study. If 1 or more responses were observed in the first 12 patients, an additional 25 were to be accrued. Four (4) or more responses out of 37 patients would have been considered evidence warranting further study of the regimen providing other factors, such as toxicity and survival, also appeared favorable. If 3 or fewer responses out of 37 patients are observed, further study of this regimen of vorinostat would not be warranted.

Toxicity was monitored on an ongoing basis according to guidelines based on a truncated sequential probability ratio test with theoretical parameters set to α=0.10, β=0.10, p_o=0.20, p_a=0.40, and a maximum of 37 patients. Unacceptable toxicity for the purposes of this analysis was defined as treatment related death, grade 4 non-hematological toxicity or grade 4 hematological toxicity requiring platelet transfusion or admission for treatment of neutropenic fever over 38.5 degrees Celsius. Clear evidence that the chance of unacceptable toxicity was greater than 20%, i.e. if 4 of 5, 5 of 9 or 6 of 12 patients accrued had unacceptable toxicity, then review for closure due to toxicity would be triggered.

Standard statistics were used to summarize the results. The Kaplan-Meier (KM) product-limit estimator was used to summarize overall survival (OS) and progression-free survival (PFS). Estimates of median OS and PFS were based on KM curves, as were the associated confidence intervals. All reported confidence intervals are two-sided and calculated to have a 95% chance of coverage.

RESULTS

Patient Characteristics:

Fourteen patients were accrued between June 2006 and September 2007; all were eligible and all started treatment. Patients 13 and 14 had signed consent before patient 12 reached assessment for primary endpoints. Patient demographics and clinical characteristics at diagnosis are summarized in Table 2. Median age was 66. The majority of patients were male Caucasians; 50% patients had Karnofsky performance status of 100% or 90% while no patients fell in the 60–70% range (ECOG 2); the primary site was in the bladder in all but 1 patient.

Table 2.

Patient demographics and clinical characteristics

Demographic/Characteristic	No. of Patients Total 14	%
Age, years
Median (Range)	66 (43–84)
Sex
Male	12	86
Female	2	14
Race
Caucasian	11	79
Hispanic	2	14
Black	1	7
Site of primary tumor
Bladder	13	93
Renal pelvis/ureter	1	7
Prior cystectomy	8	57
Prior chemotherapy	14	100
Adjuvant/neoadjuvant only	2	14
Metastatic only	9	64
Metastatic and Adjuvant or Neoadjuvant	3	21
Karnofsky status
>=90%	7	50
<90%	7	50

Toxicity:

Table 3 summarizes toxicities: those possibly, probably or definitively attributable to therapy. Toxicities were predominantly due to cytopenia and thrombocytopenic bleeding. For more granularity in this number of patient the outcomes including toxicity are included in Table 4. Two early deaths occurred, which were reported as unlikely related to treatment toxicity: one patient died of aspiration shortly after diagnosed with progression, and one patient suffered sudden death while on treatment. Two patients developed grade 4 attributable to treatment toxicity during the trial, and 6 had grade 3. One patient experienced unacceptable toxicity by the trial definition with grade 3 thrombocytopenia during cycle 4 of therapy with possible attribution to vorinostat therapy. In total, 4 patients experienced Grade 3 or greater thrombocytopenia attributed to vorinostat, and another patient experienced Grade 3 thrombocytopenia that was felt to be unlikely to vorinostat. Interestingly, the attribution of toxicity for patients apart from these five patients was deemed unrelated or unlikely due to therapy.

Table 3.

Toxicities possibly, probably or definitely attributed to vorinostat occurring in more than one patient.

Number of patients: 14	Course 1					All Courses
Adverse Event	Grade 1	Grade 2	Grade 3	Grade 4	Grade 5	Grade 1	Grade 2	Grade 3	Grade 4	Grade 5	% of any grade	% of grade 3/4
Anemia	4	2	1			6	2	2			71.4	14,3
Platelets	3		2	1		4	2	2	2		71.4	28.6
Fatigue (asthenia, lethargy, malaise)	3	4	1			4	5	2			78.6	14.3
Weight loss	1					2	1				21.4	NR
Anorexia	2	3	2			2	6	2			71.4	14.3
Dehydration	1	1				1		1			14.3	7.1
Diarrhea	3	1				3	1				28.6	NR
Nausea	5	1	1			5	1	1			50	7.1
Taste alteration (dysgeusia)		1					1				7.1	NR
Vomiting	2	2				2	1				21.4	NR
Edema: limb	1					1					7.1	NR
ALT increase						1	1				14.3	NR
AST increase						2					14.3	NR
Hypoalbuminemia	1	1				2	1				21.4	NR
Alkaline phosphatase		1				1	1				14.3	NR
Creatinine	1	1	1			2	2	1			35.7	7.1
Hyperglycemia	3	1				3	3				42.3	7.1
Muscle weakness		1					1	1			14.3	7.1
Confusion	2					2					14.3	NR
Mood alteration	1					1					7.1	NR
Neuropathy: sensory						1					7.1	NR
Pneumonitis/pulmonary infiltrates	1					1					7.1	NR
Thrombosis/embolism								1			7.1	NR

Table 4 –

Listing of 14 patients treated on the trial.

ID	# Courses Started (# Completed)	Off Treatment Reason	Best Response (RECIST v1.0)	Highest Grade Toxicity (Attributable|Any)	PFS (months)	Overall Survival (months)
001	11 (11)	Progression	Stable Disease	1 | 3	9	9.9
002	1 (0)	Toxicity	Not Evaluated	3 | 3	2	2.1
003	2 (1)	Progression	Progression	4 | 5	<1	1
004	4 (3)	Progression	Progression	3 | 4	1	4.2
005	2 (0)	Progression	Progression	2 | 2	<1	14.5
006	6 (6)	Progression	Stable Disease	3 | 3	4	81.7
007	2 (1)	Progression	Progression	1 | 2	1	3
008	4 (3)	Urinary sepsis, hemorrhage (not related)	Stable Disease	3 | 4	3	18.7
009	1 (1)	Patient Refusal	Progression	1 | 1	<1	3.2
010	2 (1)	Progression	Progression	2 | 2	1	3.2
011	2 (2)	Progression	Progression	1 | 1	1	34.9
012	2 (1)	Sudden Death	Not Evaluated	3 | 5	1	1.2
013	2 (2)	Progression	Progression	4 | 4	1	7.4
014	1 (1)	General decline	Not Evaluated	3 | 3	<1	2

Efficacy:

Eleven of 14 patients had response radiographically evaluated – 2 patients went off therapy before undergoing week 6 imaging studies, one for toxicity and the other as part of a general cancer related decline; a 3^rd patient died early while on treatment. The median number of completed cycles of therapy was 2 with the longest duration of 11 cycles. Two patients with stable disease remained on therapy for 6 and 11 cycles of treatment. No responses were seen. Of the 11 evaluated patients, 3 had stable disease as the best response and 8 had progressive disease. For all 14 patients, median PFS was 1.1 months (95% CI: 0.8–2.1) and median overall survival was 3.2 months (95% CI: 2.1–14.5).

DISCUSSION:

In this prospective phase II trial, vorinostat had an unacceptable risk:benefit ratio in patients with advanced urothelial cancer that had progressed on platinum-based chemotherapy. Efficacy was minimal and did not meet criteria for continuation to the second stage of the trial. Toxicity was strikingly more than what was predicted by phase I studies with 9 of 14 patients experiencing grade 3 or greater toxicity. Interestingly, the trial proceeded contemporaneously with several phase I and II studies in which vorinostat was combined with chemotherapy [32]. The toxicities we observed in this trial were greater than in patients with other diagnoses given combination therapy with vorinostat and platinum doublet chemotherapy, although the addition of vorinostat to carboplatin and paclitaxel increased the rate of grade 4 thrombocytopenia to 18% over 3% in the control arm. Disproportionately greater toxicity was evident when Spanish investigators undertook a phase II study in urothelial cancer patients using doses of gemcitabine and carboplatin developed and largely well tolerated in patients with non-small cell lung cancer. This dose proved very toxic in patients with urothelial cancer and the dose of both drugs had to be reduced for further trials. Exactly why patients with urothelial cancer may be more susceptible to toxicity from systemic agents is not clear, but a number of factors are postulated, including renal impairment and co-morbid conditions.

The hypothesis for this trial was based on relatively strong preclinical evidence that suggested HDAC inhibition would be beneficial in urothelial cancer based on a match of molecular processes characterized within cancer tissue and the putative effects of vorinostat on those pathways in models (Table 1.). Despite this the agent proved inefficacious and toxic in the dose and schedule given to urothelial cancer patients. The impact of vorinostat on the pathways and processes postulated to be important was to be investigated with a series of correlative studies for which biospecimens were collected but returned or destroyed due to withdrawal of NCI funds. Other investigators have sort to define populations that may benefit from HDAC inhibitors. Based on preclinical evidence from bladder cancer xenografts, Grivas et al selected for patient with CREB binding protein and or E300 mutations using tumor tissue genomic analysis and treated then with the HDAC inhibitor mocetinostat [33]. Effiacy was insufficient to expand the trial beyond 14 patients with one partial response and 2 patients with stable disease. Toxicity most commonly was reflected as nausea and fatigue and dose delays and reductions were common.

Given the epigenetic modulatory effects of HDAC inhibitors, subsequent trials have evaluated the use of these agents as modulators of response in other genitourinary cancers when given at relatively low doses. For example, the addition of oral etinostat to high-dose interleukin-2 therapy in renal cell cancer yielded an overall response rate of 37%, somewhat higher than historical experience[34]. This is now the basis for the Hoosier Oncology Group GU17–289 randomized phase II study of HDIL2 with or without etinostat in renal cell cancer (NCT03501381). Further evaluation of vorinostat has been undertaken in combination with T cell check point inhibitors in advanced urothelial cancer. The IUSCC 0551 trial evaluated the combination of vorinostat at two dose levels of vorinostat (100 and 200 mg, PO daily 2 weeks ON and one week OFF) with pembrolizumab in 2 cohorts of patients: those without prior exposure to immunotherapy agents targeting PD-1 or PD-L1 and those who have progressed on these agents. The trial tests whether addition of vorinostat increases the response to pembrolizumab in immunotherapy naïve individuals and whether the addition of this HDAC inhibitor can sensitize patients who have previously been treated and induce a response with re-challenge [35].

CONCLUSION:

Vorinostat in this dose-schedule proved surprisingly toxic and had limited efficacy in patients with advanced urothelial carcinoma who have progressed despite prior platinum-based chemotherapy. Further evaluation of lower dose regimens in combination with cytotoxic or target specific agents including immunotherapies may be warranted.

Availability of data and material:

All data was available to all authors.