Mixed 20-peptide cancer vaccine in combination with docetaxel and dexamethasone for castration-resistant prostate cancer: a randomized phase II trial

Prostate cancer-related death is common among patients with metastatic castration-resistant prostate cancer (CRPC) in which the disease progresses despite androgen deprivation therapy. Although several new agents for CRPC, such as sipuleucel-T, cabazitaxel, abiraterone acetate, enzalutamide, and radium-223, have been approved recently based on their demonstrated overall survival (OS) benefits in phase 3 studies, each of these prolonged survival by only a few months [2‐6]. Thus, there remains a need for treatments that can provide stable disease control and long-term survival benefits. After the approval of sipuleucel-T for patients with asymptomatic or minimally symptomatic CRPC based on survival benefits [2], and stable responses observed with the checkpoint inhibitors ipilimumab and nivolumub in other malignancies [7‐9], immunotherapy has emerged as a viable and attractive strategy for the treatment of CRPC. However, immunotherapy alone may be unable to induce an immune response of sufficient potency to result in tumor regression when immune tolerance and large tumor burden are present.

The standard first-line chemotherapy for patients with progressive CRPC has consisted of docetaxel and oral prednisone [10, 11]. Several animal studies have examined the use of cancer vaccines in combination with docetaxel, and reported enhancement of T-cell responses with antitumor activity [12‐14]. These studies suggested that docetaxel is able to reduce the tumor burden and immune-suppressing elements, such as the increase in myeloid-derived suppressor cells (MDSC) in the tumor microenvironment, and it is well known that tumor-associated immunosuppression plays a significant role in tumor progression and resistance to immunotherapy [13].

We developed a novel cancer vaccine consisting of 20 mixed peptides (KRM-20) for patients with CRPC designed to induce cytotoxic T lymphocytes (CTL) against twelve different tumor-associated antigens (TAAs) highly expressed in prostate cancer tissues. The CTL epitopes represented by these 20 peptides are restricted to human leukocyte antigen (HLA)-A2, -A24, -A3 super type (-A3, -A11, -A31 or -A33) or -A26 of major histocompatibility complex class 1 molecules, providing coverage of the majority of patients who have different HLA alleles. A previous phase I study on KRM-20 for patients with CRPC demonstrated the feasibility, safety, and rapid and high immune responses without changes in immunosuppressive cell subsets [15].

This phase II study was designed to evaluate if docetaxel has the ability to alter components of the immune system independent of antitumor activity, and to examine the potential synergistic activity of the 20-mixed peptides vaccine in combination with docetaxel and dexamethasone.

Between July 31, 2013 and July 11, 2014, 55 chemotherapy-naïve patients with progressive CRPC were screened for enrollment at ten medical centers in Japan (Fig. 1). Fifty-one patients were enrolled and randomly assigned to receive either KRM-20 with docetaxel and dexamethasone (n = 25) or placebo (n = 26) with docetaxel and dexamethasone. Two patients in the KRM-20 arm were not treated due to the physician’s decision or death, and not included in the safety analysis set. Baseline demographic and clinical characteristics of participants, including median age, ECOG performance status, metastatic site, HLA types, Gleason scores, median PSA levels, median times from diagnosis to study entry, median IgG levels, and median CTL levels were balanced between the two arms (Table 1). Most patients (72% in the KRM-20 arm and 69.2% in the placebo arm) had bone disease, and all patients were refractory to previous hormone therapies. At the end of study treatment, 17 (73.9%) of 23 patients in the KRM-20 arm and 20 (76.9%) of 26 patients in the placebo arm had completed the study treatment (Fig. 1).

Regarding immune responses, the mean total HLA-matched peptide-specific IgG (P = 0.014, t test) and CTL (P = 0.007, t test) responses in the KRM-20 arm significantly increased after treatment, whereas the IgG and CTL responses in the placebo arm did not increase after treatment (Fig. 2a, b). The median number of HLA-matched peptides in the KRM-20 arm was 16 (8–17), and peptide-specific IgG and CTL responses matching HLA were observed in 8 (35%) of 23 patients and 5 (22%) of 23 patients, respectively (Supplementary Table 2). In the exploratory analysis for immune suppression, the numbers of both Treg and MDSC among PBMC in the two arms did not increase during treatment (Fig. 2c, d), and the number of MDSC in the KRM-20 arm significantly decreased after the treatment (P = 0.03, t test) (Fig. 2d).

AEs in the two arms during treatment are summarized in Supplementary Table 3. The most common AEs (occurring in more than 40% patients in one or both arms) were injection site reactions, alopecia, neutropenia, and peripheral neuropathy. AEs of grade 3 or higher developed in similar frequencies between the two arms: 16 (70%) of 23 patients in the KRM-20 arm and 18 (79%) of 26 patients in the placebo arm. Grade 5 events during treatment were observed in 2 patients with pneumonia in the placebo arm. The addition of KRM-20 did not increase toxicity. The dose of docetaxel was reduced due to hematological toxicity in a similar proportion in each patient arm (17.5% in the KRM-20 arm and 15% in the placebo arm).

After the median follow-up of 8.3 months (IQR 5.0–13.3), 45 (88%) of 51 patients had disease progression or died: 22 (88%) in the KRM-20 and 23 (85%) in the placebo arm. Based on investigator assessment of disease response and progression using PCWG2 or RECIST criteria, partial response was observed in 2 (8%) patients in the KRM-20 arm and 3 (12%) patients in the placebo arm. No complete responses were observed in either arm. The median PFS time was 8.9 months (95% CI 4.9–12.2) in the KRM-20 arm and 7.4 months (95% CI 5.3–12.5) in the placebo arm (Fig. 3a), but this difference was not significant (HR 1.0; 95% CI 0.6–1.9; P = 0.87). At the data cut-off date of July 1, 2017 after a median follow-up of 32.4 months (IQR 16.3–36.1), 24 (47%) patients in the ITT population had died: 12 (48%) patients in the KRM-20 arm and 12 (46%) patients in the placebo arm. The median OS time in the placebo arm was 30.5 months (95% CI 23.9–38.8), but it has yet to be reached in the KRM-20 arm (95% CI 16.4-not reached); it was estimated to be 37.7 months using the adjusted HR (Fig. 3b). Although patients in the KRM-20 arm had a slightly longer OS than those in the placebo arm, there was no between-group difference in median OS time (HR 0.81; 95% CI 0.3–2.1; P = 0.83). To further investigate the effects of KRM-20 with docetaxel and dexamethasone, Cox proportional hazards regression analysis was performed to find factors that can predict disease response in the KRM-20 arm. Univariate Cox analysis demonstrated % lymphocytes (P = 0.006) and PSA level (P = 0.003) to be significantly associated with survival. The factors with an HR less than 0.5 in the univariate analysis were included in the multivariate analysis of the model. In the KRM-20, ≥ 26% lymphocytes (HR 0.3; 95% CI 0.08–0.93; P = 0.04) and PSA levels < 11.2 ng/ml (HR 0.15; 95% CI 0.03–0.58; P = 0.004) were found to be significantly favorable factors for OS (Table 2). Consequently, the median OS of patients with ≥ 26% lymphocytes or PSA levels < 11.2 ng/ml was significantly longer than that in their counterparts in the KRM-20 (median OS, not reached vs 16.4 months; P = 0.02 or median OS, not reached vs 16.3 months; P = 0.003, respectively; Fig. 3c, d).

This randomized phase II trial of KRM-20 in combination with docetaxel and dexamethasone for patients with CRPC demonstrated a similar PSA decline, increased immune responses, and reduction in MDSC compared with docetaxel and dexamethasone treatment alone. PSA decline in several clinical trials for CRPC have been associated with longer survival [19]. However, improvements in OS in patients with CRPC may be observed without differences in PSA response rates [10], and PSA decline is not considered a true response. These results suggest the difficulty in clinical assessment of the phase II study setting for CRPC.

Active specific immunotherapy using either TAAs or their peptides has failed to provide clinical benefits for cancer patients in the large number of clinical trials since the 1990s, and the mechanisms involved in the failure have not been clarified [20‐22]. Furthermore, this study did not demonstrate a longer OS or PFS than with the placebo. To understand the mechanisms involved in the failure of KRM-20, we predicted subsets of patients who will respond to KRM-20 with good clinical effects. Subgroup analysis suggested that this treatment is favorable for CRPC patients with ≥ 26% lymphocytes or PSA levels < 11.2 ng/ml. Although further validation is required, this finding is novel and helpful to prolong survival in patients with CRPC treated by peptide vaccinations.

In this study, assessment of HLA-matched peptide-specific IgG and CTL was conducted primarily to evaluate immune responses to the combination therapy with KRM-20. A significant increase in both HLA-matched peptide-specific IgG and CTL responses was observed in patients receiving the combination of KRM-20 with docetaxel and dexamethasone compared with patients in the placebo arm. The administration of cytotoxic chemotherapy agents, such as docetaxel, induces bone marrow suppression, and has been believed to negatively affect immune responses induced by the cancer vaccine. However, recent data suggest that many widely used chemotherapeutic agents have beneficial immunomodulatory effects thorough several mechanisms, including cytokine production, T cell infiltration of tumors, and maturation of dendritic cells [23‐25]. The dose and scheduling of docetaxel in combination with cancer vaccines remain unknown. In a murine model, the combination of repeated melanoma tumor cell vaccinations with a more clinically relevant dose of docetaxel did not impede T-cell activation [26]. A previous randomized phase II trial of a cancer vaccine in combination with the clinical dose of docetaxel for metastatic CRPC patients did not inhibit vaccine-specific T cell reactivity [27, 28], and the present study demonstrated that KRM-20 in combination with docetaxel and dexamethasone increased HLA-matched peptide-specific IgG and CTL responses. Several clinical trials have suggested that cancer vaccines alter the clinical effects of subsequent docetaxel. An early phase trial of an adenovirus-based vaccine targeting p53 in patients with non-small cell lung cancer reported higher objective responses to salvage chemotherapy initiated after vaccine treatment [29]. Similarly, in a survival analysis of the pivotal randomized phase III IMPACT trial, patients who received sipuleucel-T prior to docetaxel demonstrated longer survival than patients receiving placebo prior to docetaxel (P = 0.03) [2]. However, although not exactly different from the current trial, two phase III trials (VITAL-1 and VITAL-2) of GVAX which was composed of two human prostate cell lines LNCaP and PC3 as antigen source, were failed to demonstrate survival benefit. VITAL-2 compared the combination of GVAX plus docetaxel with standard docetaxel and prednisone in men with symptomatic CRPC. The VITAL-2 study was terminated early due to increased deaths in the vaccine arm. The VITAL-1 was also terminated based on a result of less than a 30% chance of meeting an improved survival end point. Regarding the GVAX failure, it has been pointed out that the lack of placebo, docetaxel dose, and timing are not taken into account [30].

Another concern is the immune suppression caused by the continuous administration of low-dose dexamethasone with the peptide-based cancer vaccine. However, recent clinical trials found that combination therapy of low-dose dexamethasone with a peptide-based vaccine induced positive CTL responses, and a longer PSA PFS and OS than the peptide-based vaccine alone [31, 32].

In the current study, the number of MDSC, but not Treg, in patients in the KRM-20 arm decreased after treatment. Immune inhibition caused by MDSC in cancer patients has been reported in a number of studies. The increase in circulating MDSC also correlates with PSA levels and tumor burden in patients with CRPC [33]. Several preclinical studies reported that docetaxel combined with a cancer vaccine depleted circulating or tumor-infiltrating MDSC with CTL responses in murine tumor models by direct alteration of MDSC signaling, phenotype, and function [12, 13, 34]. These findings suggest potential clinical benefits by addition of docetaxel to the current immunotherapy.

The limitation of the present study was that the similar PSA decline, increase in immune responses, and decrease of MDSC in the KRM-20 arm did not lead to a longer PFS or OS than those in the placebo arm. This may have been due to the small number of patients and study design not comparing the difference in survival between the two arms. Further randomized clinical trials with a study design where the primary endpoint is the comparison of survival between KRM-20 in combination with or without docetaxel are needed.

The safety and tolerability were similar to the known profiles for docetaxel and peptide-based vaccines [10, 15]. The most common AEs were grade 1 or 2 injection site reactions in both arms. The main cause of grade 3 or higher AEs was hematology toxicity due to docetaxel, and these AEs developed in almost 40% of the patients in each arm. The addition of KRM-20 in combination with docetaxel and dexamethasone was feasible without increased toxicity.

In conclusion, KRM-20 in combination with docetaxel and dexamethasone for patients with CRPC resulted in a similar PSA decline, increased immune responses, and reduced MDSC compared with docetaxel and dexamethasone treatment alone. Subgroup analysis suggested that this treatment is favorable for CRPC patients with ≥ 26% lymphocytes or PSA levels < 11.2 ng/ml. Although patients in the KRM-20 arm had a slightly longer OS than those in the placebo arm, this study did not demonstrate any survival benefits. Further large-scale clinical trials comparing OS are required to confirm the clinical benefits of this treatment.