Abstract
Purpose:
Two studies in previously-treated metastatic pancreatic cancer have been completed combining GVAX pancreas vaccine (granulocyte-macrophage colony-stimulating factor-secreting allogeneic pancreatic tumor cells) with cyclophosphamide (Cy) and CRS-207 (live, attenuated Listeria monocytogenes expressing mesothelin). In the current study, we compared Cy/GVAX followed by CRS-207 with (Arm A) or without nivolumab (Arm B).
Experimental Design:
Patients with pancreatic adenocarcinoma who received one prior therapy for metastatic disease and RECIST measurable disease were randomized 1:1 to receive treatment on Arm A or Arm B. The primary objective was to compare overall survival (OS) between the arms. Additional objectives included assessment of progression-free survival, safety, tumor responses, CA19–9 responses and immunologic correlates.
Results:
Ninety-three patients were treated (Arm A, 51; Arm B, 42). The median OS in Arms A and B were 5.9 (95% CI, 4.7, 8.6) and 6.1 (95% CI, 3.5, 7.0) months, respectively, with a hazard ratio 0.86 (95% CI, 0.55, 1.34). Objective responses were seen in three patients using immune-related response criteria (4%, 2/51, Arm A; 2%, 1/42, Arm B). The ≥grade 3 related adverse event rate while higher in Arm A (35.3% vs 11.9%) was manageable. Changes in the microenvironment, including increase in CD8+ T cells and a decrease in CD68+ myeloid cells, were observed in long-term survivors in Arm A only.
Conclusions:
While the study did not meet its primary endpoint of improvement in OS of Arm A over Arm B, the OS was comparable to standard therapy. Objective responses and immunologic changes in the tumor microenvironment were evident.
Introduction
Immunotherapy for pancreas cancer remains a challenge and single agents are unlikely to be effective in this disease with few endogenously infiltrating cytotoxic T cells, dense immunosuppressive stroma, and increased inhibitory tumor associated macrophages. Prior work with the GVAX pancreas vaccine, allogeneic pancreatic cancer cells modified to express granulocyte-macrophage colony-stimulating factory (GM-CSF), combined with low dose cyclophosphamide (Cy) to inhibit regulatory T cells administered in the neoadjuvant treatment of resectable pancreatic cancer has shown that a “vaccination” strategy is capable of inducing immune infiltrating cells into the tumor microenvironment (1). However, this is unlikely to result in clinical benefit due to the myriad of counterregulatory mechanisms that are simultaneously upregulated in response to therapy. Among the immune checkpoints that are increased in this setting is the programmed death (PD-1/PD-L1) axis. Upregulation of PD-L1 in the tumor microenvironment suppresses the activity of the tumor infiltrating lymphocytes. PD-1 inhibitors are now widely used in the clinic for more immunogenic tumors and preclinical data suggests synergy with a variety of different agents including with vaccines (2, 3). Therefore, clinical studies evaluating the potential of therapeutic combinations are warranted.
Listeria monocytogenes-expressing mesothelin (CRS-207) is a gram-positive attenuated Lm that has been modified to express the tumor associated antigen mesothelin (4–6). Mesothelin is present on the cell surface of a majority of pancreatic cancers and the induction of mesothelin-specific T cells in patients treated with GVAX pancreas has been associated with improved survival (7, 8). Listeria are intracellular organisms that have access to both MHC class I and II antigen processing pathways and have the capacity to stimulate both adaptive and innate immunity. A unique advantage of using microorganism-based constructs is that they naturally stimulate immunity via “danger signals” and their effects on toll-like receptors. Listeria induces IFNβ expression through a stimulator of interferon genes (STING)-dependent pathway. Preclinical vaccine combination strategies suggest that Lm is best used in a prime-boost approach and can serve to strengthen immune responses primed by both dendritic cell and whole cell vaccines. Studies of CRS-207 in patients with pancreatic cancer have shown the agent to be safe and capable of eliciting T cell and cytokine responses (4–6, 9). A study of GVAX pancreas prime and CRS-207 boost in patients with previously treated metastatic pancreatic cancer demonstrated a promising survival benefit over GVAX pancreas alone but a subsequent study testing the combination in treatment-refractory metastatic pancreatic cancer against chemotherapy failed to meet its primary endpoint (6). Despite these results, a prime-boost strategy remains scientifically valid. However, vaccination strategies without targeting immune checkpoints or other immunosuppressive targets in the pancreatic cancer immune microenvironment are largely being abandoned.
The combination of anti-cytotoxic T-lymphocyte associated protein 4 (CTLA-4) or anti-PD-1 with GM-CSF whole cell vaccines or Lm-based vaccines have shown synergy in murine tumor models (2, 3, 10). While neither vaccine strategy has been combined with PD-1 inhibition, GVAX pancreas has been combined with the CTLA-4 inhibitor, ipilimumab, in patients with treatment refractory advanced pancreatic cancer (11). Of the patients who received ipilimumab in combination with GVAX pancreas (N=15), 7 patients had CA19–9 responses and 3 patients had prolonged stable disease (31, 71 and 81 weeks). Objective delayed stabilization or regressions were seen despite not meeting Response Evaluation Criteria in Solid Tumors (RECIST) criteria and a one year overall survival (OS) of 27% was intriguing in a heavily pretreated patient population. Induction and maintenance of T cell responses as well as enhancement of the repertoire of mesothelin-specific T cell responses correlated with longer OS.
The current study builds on prior experience with the prime boost vaccination strategy as well as combination strategies using vaccines and immune checkpoint inhibitors. This was an open-label study, in patients with previously treated metastatic pancreatic cancer, randomized to Cy/nivolumab/GVAX pancreas vaccine followed by nivolumab/CRS-207 (Arm A) or Cy/GVAX pancreas vaccine followed by CRS-207 (Arm B) (ClinicalTrials.gov ID: NCT02243371).
Patients and Methods
Study Design
The study was reviewed by the local Institutional Review Boards, biosafety committees, the US Food and Drug Administration, and the National Institutes of Health Recombinant DNA Advisory Committee. The trial was conducted according to the Declaration of Helsinki and the Good Clinical Practice guidelines of the International Conference of Harmonisation. All patients provided written informed consent prior to enrollment. Interim data were reviewed by an independent data monitoring committee (DMC).This multicenter, randomized, open-label phase 2 trial recruited patients at 5 centers (Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Providence Portland Medical Center, University of California San Francisco, Abramson Cancer Center at the University of Pennsylvania, and Stanford University). Patients were randomized 1:1 to receive courses of 2 doses Cy/nivolumab/GVAX + 4 doses of nivolumab/CRS-207 (Arm A) or 2 doses Cy/GVAX + 4 doses of CRS-207 (Arm B) (Figure 1). Randomization was stratified by disease status (progressive disease or stable disease at study start) and study site.
Figure 1. Treatment schema.
Patients with previously treated, metastatic pancreatic cancer were randomized 1:1 to 2 treatment arms (Arm A: Cy/GVAX & CRS-207 + Nivolumab vs. Arm B: Cy/GVAX & CRS-207). Cycles were administered at 3 week intervals. Cyclophosphamide and nivolumab were administered on days 1, GVAX and CRS-207 were administered on days 2. A course was 6 cycles and courses could be repeated.
The primary endpoint was to compare OS between arms. Secondary endpoints included objective response rate (ORR), progression-free survival (PFS), safety, tumor marker kinetics and immunologic responses.
Patients
Eligible patients had cytologically or histologically-proven, metastatic adenocarcinoma of the pancreas, were ≥18 years old, progressed on one prior regimen administered for metastatic disease, had an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, RECIST v.1.1 measurable disease, and adequate organ function. Patients were excluded if they had brain metastases, major artificial implants or devices that could not be easily removed (portacaths and biliary stents were allowed), hepatic cirrhosis or clinical or radiographic ascites or malignant pleural effusion, thromboembolic disease within two months, HIV, hepatitis B or C, autoimmune disease, or select prior immunotherapies.
Treatment
In both arms, one treatment course consisted of six cycles at 3-week intervals (Figure 1). In Arm B, Cy (200mg/m2) was delivered intravenously (IV) one day before each GVAX treatment (first day of weeks 1 and 4). GVAX consisted of two irradiated, allogeneic, GM-CSF-secreting pancreatic adenocarcinoma cell lines (2.5 × 108 cells each; Johns Hopkins University, Baltimore, MD), combined and administered as six intradermal injections on the second day of weeks 1 and 4. CRS-207 (1 × 109 colony-forming units, Aduro Biotech, Inc., Berkeley, CA), was delivered by 1-hour intravenous infusion followed by a 4-hour observation period on the first day of weeks 7, 10, 13, and 16. In Arm A, the same Cy/GVAX/CRS-207 schedule was used plus nivolumab (3mg/kg) was administer by IV on one day prior to GVAX or CRS-207 every 3 weeks for 6 cycles. Oral antibiotics were initiated 7 days after the final CRS-207 dose of each course. Patients could receive additional courses if clinically stable.
Assessments
Physical examinations, complete blood count and chemistries were performed prior to each treatment and laboratory assessment was repeated the day after CRS-207 infusions. Safety was assessed in all patients who received any study treatment. Imaging was performed at baseline, week 10 and week 20. Tumor response was determined by investigator assessment using RECIST v1.1. Immune-related response was assessed using a modified iRECIST (12).
Immunologic Assessments
Tissue biopsies were obtained at baseline and at the end of the third cycle of therapy (after 2 doses of CY/nivolumab/GVAX and 1 dose of nivolumab/CRS-207 in Arm A and 2 doses of CY/GVAX and 1 dose of CRS-207 in Arm B). A blinded study pathologist selected biopsies containing >30% tumor cellularity for multiplex immunohistochemistry (IHC). Chromogenic sequential IHC was conducted on 5-micron formalin-fixed paraffin-embedded tissue sections as previously described (13). Three 12-plex multiplex IHC panels containing markers of lymphoid, myeloid, and T cell function were stained in the indicated order and condition shown in Supplementary Table 1. Whole slide scanning was conducted by Aperio ImageScope (Leica), image processing by CellProfiler Version 2.1.1, and visualization by ImageJ Version 1.48. Image cytometry data analysis was performed by FCS Express 6 Image Cytometry Version 6.03.0011 (De Novo Software, CA, USA) based on gating strategies previously reported (13). 16 immune cell lineages including lymphoid and myeloid cell populations were quantitatively evaluated by image cytometry according to lineage selective markers shown in Supplementary Table 2.
Cytokine Analyses
Plasma samples were obtained at baseline, at the end of the second cycle of therapy (after 2 doses of CY/nivolumab/GVAX in Arm A and 2 doses of CY/GVAX in Arm B), and at the end of the third cycle of therapy (after 2 doses of CY/nivolumab/GVAX and 1 dose of nivolumab/CRS-207 in Arm A and 2 doses of CY/GVAX and 1 dose of CRS-207 in Arm B). Human plasma samples were stored at − 80°C. A total of 50 microliters from each of the samples was used in the cytokine assay. The samples were run in single wells undiluted. The Bioplex 200 platform (Biorad, Hercules CA) was used to determine the concentration of multiple target proteins in the plasma samples. Luminex bead-based immunoassays (Millipore, Bilerica NY) were performed following the manufacturers protocols and concentrations were determined using 5 parameter log curve fits (using Bioplex Manager 6.0) with vendor provided standards and quality controls. The Millipore Human Cytokine/Chemokine Magnetic Bead Panel (Catalog# HCYTOMAG-60K-PX38 Lot# 3024390) was used to detect EGF, FGF-2, Eotaxin, TGF-α, GCSF, Flt-3L, GMCSF, Fractalkine, IFNα2, IFNγ, GRO, IL-10, MCP-3, IL-12(p40), MDC, IL-12(p70), IL-13, IL-15, sCD40L, IL-17, IL-1RA, IL-1α, IL-9, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IP-10, MCP-1, MIP-1α, MIP-1β, TNFα, TNFβ, and VEGF.
Statistical Analyses
Clinical Trial Analyses
The primary efficacy analysis compared OS based on the log-rank test for Arms A versus B and included all randomized subjects who received at least one dose of study treatment. The original sample size of 88 patients (44 per arm) would provide 80% power to detect a hazard ratio (HR) of 1.66, i.e. an increase in median OS from 6 to 9.96 months, assuming a 2-sided type I error rate of 15%, a uniform accrual period of 18 months, follow-up of 24 months with no loss to follow-up. In October 2015, the sample size was revised to 108 patients, which provides 80% power to detect a HR of 1.66 while allowing for a two-sided type I error of 0.14278, 18 months of accrual, 12 months of follow-up, and a 5% loss to follow-up (i.e. an effective sample size of 102 treated patients). The type I error was reduced for the final analysis to allow for type I error spending based upon O’Brien-Fleming stopping boundaries of 0.02361 at an interim efficacy analysis when approximately 50% of the expected deaths in treatment subjects had occurred (approximately 42 deaths). Accrual to Arm B of this study was stopped at 42 treated patients following an ad hoc data monitoring committee meeting on June 10, 2016 to evaluate survival data. This meeting was prompted by the negative results of the ECLIPSE study (NCT02004262), which evaluated a treatment strategy similar to Arm B of this study compared to single agent chemotherapy (6). Enrollment to Arm A was not altered (51 treated), and patients previously randomized continued on their originally assigned treatment in accordance with the protocol.
Kaplan Meier techniques were used to estimate median survival and 95% confidence intervals (CIs) for OS and PFS. The log-rank test was used to compare OS and PFS between treatment arms. The Cox proportional hazards model and Wald statistics were used to estimate HRs and CIs. Overall survival and progression-free survival were calculated from the date of randomization until the date of death and date of disease progression or death, respectively. Individuals without an event were censored at the last date of contact for OS and the date of last scan for PFS.
Correlative Studies Analyses
In chromogenic multiplex IHC analyses, given that the median survival of patients treated on a second line regimen is approximately 4–6 months (14), the patients were stratified into two groups of short-term (OS ≤ 6 months), and long-term (OS > 6 months) survivors to explore potential changes of immune complexity profiles associated with therapeutic response. Wilcoxon signed rank tests were used to determine statistically significant differences between the changes from baseline to after the third cycle within each treatment (A and B) and survival (short- and long-term) combination. Mann-Whitney U-tests were used to compare short and long term survivors within each treatment group.
Cytokine values were log2 transformed and out of range values were replaced with the detection limits. Differential expression analysis was performed on baseline values and differences relative to baseline with the R/Bioconductor package LIMMA (15). Benjamini-Hotchberg adjusted p-values for empirical Bayes moderated statistics in a multivariate model accounting for patient outcomes, trial arm, and cytokine batch were reported. Patient outcomes were binarized based on an overall survival cutpoint of 6 months in this model. A knitr report containing the R code and results for this analysis is provided as Supplemental Methods File.
Results
Study population, demographics and baseline characteristics
A total of 96 patients across 5 US sites were enrolled and randomized. Three (all assigned to Arm B) never received treatment and were excluded from the analysis. Of the remaining 93 patients who received treatment, 51 and 42 patients were assigned to Arms A and B, respectively. The median duration on treatment was 64 and 61 days, respectively. Demographics and baseline characteristics are summarized in Table 1. Ninety-seven percent of patients had progressive disease upon enrollment and 74% of patients had liver metastases. Sixty-six percent, 53%, and 18% of patients received prior 5-fluorouracil (FU)-based regimens, gemcitabine-based regimens, and both 5FU and gemcitabine-based regimens, respectively. All patients had received at least 1 prior chemotherapy, however, patients may have received 2 or more chemotherapy regimens when including treatments for localized or resected disease. Thirty-seven percent underwent a prior resection with curative intent with a smaller percentage in Arm A than in Arm B (24% vs 52%).
Table 1.
Demographics, disease, and treatment history among participants who received at least one treatment.
| Cy/GVAX/CRS-207 + NIVO | Cy/GVAX/CRS-207 | ||
|---|---|---|---|
| All participants | (Arm A) | (Arm B) | |
| Characteristics | N = 93 | N = 51 | N = 42 |
| Demographics | |||
| Age at randomization (years), median (1st–3rd Q) | 64 (58–69) | 64 (58–69) | 64 (57–70) |
| Age at randomization >65, N (%) | 39 (42%) | 21 (41%) | 18 (43%) |
| Male, N (%) | 61 (66%) | 37 (73%) | 24 (57%) |
| Disease History | |||
| Years since diagnosis, median (1st–3rd Q) | 0.9 (0.6–1.3) | 0.8 (0.6–1.1) | 1.0 (0.7–1.7) |
| Years since stage IV diagnosis, median (1st–3rd Q) | 0.6 (0.3–0.9) | 0.6 (0.3–0.8) | 0.6 (0.4–1.0) |
| Initial stage, N (%)a | |||
| IB | 4 (5%) | 1 (2%) | 3 (8%) |
| IIA | 10 (11%) | 6 (12%) | 4 (11%) |
| IIB | 17 (20%) | 6 (12%) | 11 (29%) |
| III | 11 (13%) | 3 (6%) | 8 (21%) |
| IV | 45 (52%) | 33 (67%) | 12 (32%) |
| Missing | 6 (6%) | 2 (4%) | 4 (10%) |
| Disease status at randomization, N (%) | |||
| ECOG 0 | 53 (57%) | 29 (57%) | 24 (57%) |
| Progressive disease | 90 (97%) | 49 (96%) | 41 (98%) |
| Histologic gradea | |||
| Well differentiated | 4 (6%) | 3 (9%) | 1 (3%) |
| Moderately differentiated | 36 (51%) | 18 (51%) | 18 (50%) |
| Poorly differentiated | 31 (44%) | 14 (40%) | 17 (47%) |
| Missing | 22 (24%) | 16 (31%) | 6 (14%) |
| Gross tumor locationa | |||
| Head | 45 (49%) | 24 (48%) | 21 (51%) |
| Body | 11 (12%) | 4 (8%) | 7 (17%) |
| Tail | 20 (22%) | 14 (28%) | 6 (15%) |
| Body and tail | 11 (12%) | 6 (12%) | 5 (12%) |
| Other | 4 (4%) | 2 (4%) | 2 (5%) |
| Missing | 2 (2%) | 1 (2%) | 1 (2%) |
| Disease extent, N (%) | |||
| Adrenal | 1 (1%) | 1 (2%) | 0 (0%) |
| Bone | 7 (8%) | 3 (6%) | 4 (10%) |
| Liver | 69 (74%) | 39 (76%) | 30 (71%) |
| Lung | 44 (47%) | 20 (39%) | 24 (57%) |
| Lymph nodes | 28 (30%) | 14 (27%) | 14 (33%) |
| Pancreas | 31 (33%) | 17 (33%) | 14 (33%) |
| Other | 27 (29%) | 15 (29%) | 12 (29%) |
| Lab values | |||
| Albumin at randomization, median (1st–3rd Q) | 40 (37–43) | 40 (37–43) | 40 (36–43) |
| Prior treatment, N (%) | |||
| Prior curative intent surgery | 34 (37%) | 12 (24%) | 22 (52%) |
| Prior radiation therapy | 26 (28%) | 12 (24%) | 14 (33%) |
| Number of different prior chemotherapy regimens | |||
| 1 | 72 (77%) | 42 (82%) | 30 (71%) |
| 2 | 17 (18%) | 7 (14%) | 10 (24%) |
| 3 | 3 (3%) | 2 (4%) | 1 (2%) |
| 4 | 1 (1%) | 0 (0%) | 1 (2%) |
Abbreviations: Cy, cyclophosphamide; NIVO, nivolumab; Q, quartile.
Percent missing is computed out of the total number of participants. Percent within each category is calculated out of the number with available data.
Efficacy Outcomes
The primary efficacy endpoint was comparison of OS in the treated patient population (Figure 2 & Table 2). The median OS in Arms A and B were 5.9 (95% CI, 4.7, 8.6) and 6.1 (95% CI, 3.5, 7.0) months, respectively, with a hazard ratio 0.86 (95% CI, 0.55, 1.34). At the time of the analyses, 87% of the expected events had occurred. The 12 and 18 month OS rates for Arm A were 25% and 18%, respectively, and 19% and 12% for Arm B. Treatment effect by baseline characteristics is provided in Figure 3. The absence of prior radiation favored Arm A. The median PFS was approximately 2.2 months for both arms (data not shown). Twenty-eight and 19 patients received subsequent therapies in Arms A and B, respectively.
Figure 2. Overall survival.
A comparison of overall survival in months for treatment groups A versus B.
Table 2.
Efficacy outcomes: Summary of overall survival and progression outcomes.
| Cy/GVAX/CRS-207 + NIVO | Cy/GVAX/CRS-207 | Cy/GVAX/CRS-207 + NIVO/Cy/GVAX/CRS-207 | |
|---|---|---|---|
| Outcome | (Arm A) | (Arm B) | (Arm A/Arm B) |
| Overall survival (months) | |||
| N | 51 | 42 | |
| No. events | 45 | 36 | HR (95% CI) |
| Median (95% CI) | 5.88 (4.73–8.64) | 6.11 (3.52–7.00) | 0.86 (0.55–1.34) |
| Proportion alive at 6 months | 0.49 (0.35–0.62) | 0.52 (0.36–0.67) | P = 0.49 |
| Progression-free survival (months) | |||
| N | 51 | 42 | |
| No. events | 45 | 38 | HR (95% CI) |
| Median (95% CI) | 2.23 (2.14–2.33) | 2.17 (2.00–2.23) | 0.81 (0.52–1.27) |
| Proportion progression free at 3 months | 0.17 (0.08–0.31) | 0.09 (0.03–0.22) | P = 0.37 |
| Time to progression (months) | |||
| N | 51 | 42 | |
| No. events | 39 | 30 | HR (95% CI) |
| Median (95% CI) | 2.20 (2.10–2.33) | 2.20 (2.00–2.30) | 0.96 (0.59–1.56) |
| Proportion progression free at 3 months | 0.16 (0.06–0.29) | 0.12 (0.03–0.26) | P = 0.88 |
| Immune-related progression-free survival (months) | |||
| N | 51 | 42 | |
| No. events | 39 | 35 | HR (95% CI) |
| Median (95% CI) | 2.27 (2.17–2.33) | 2.23 (2.10–2.40) | 0.90 (0.56–1.43) |
| Proportion progression free at 3 months | 0.21 (0.09–0.35) | 0.10 (0.02–0.23) | P = 0.64 |
Figure 3. Treatment Effect by Baseline Characteristics.
Forest plot of the treatment effect within subgroups defined by baseline demographic, disease, and treatment characteristics. NA and NB represent the number within in each subgroup in Arms A and B, respectively. Both the hazard ratio within each subgroup and the hazard ratio of the interaction comparing the subgroups are provided. RZ = randomization.
In the 51 patients in Arm A, there was one partial response (PR) by RECIST v.1.1 criteria accounting for a PR rate of 2% (Figure 4, Supplementary Table 3). Six patients had stable disease (SD) resulting in a disease control rate (DCR) of 13.7%. In the 42 patients in Arm B, one patient had a PR (2.4%) and 3 patients had SD. The DCR was 9.5%. However, using immune response criteria, the PR rate in Arm A was 4% because there was an additional case of progressive disease followed by a PR. This patient’s imaging showed growth of lesions followed by regression; immunostains revealed increased PD-L1 on CD68+ cells, CD8, and CD3 T cell infiltration on the on-treatment biopsy represented in Figures 5A–D. Microsatellite instability status of tumors was not available on responding patients.
Figure 4. Maximum decrease in the sum of longest diameters in the target lesions.
Figure 5. Delayed response after RECIST v.1.1 progression and associated immune infiltration.
A-C) Baseline, week 10, and week 30 scans of a patient treated on Arm A demonstrating growth followed by regression (immune related response at week 30). D) Top panel: Archived pancreas tumor shows expression of PD-L1 membranous staining and few CD8 and CD3 T cells. Bottom panel: On treatment lung metastases shows expected pattern of PD-L1 expression on CD68 macrophages as well CD8 and CD3 T cell infiltration.
The tumor marker kinetics of CA19–9 was followed in patients with baseline abnormal levels (Supplementary Table 4). Values were considered stable or responding (<50% increase in the first 120 days) in 36.7 and 17.9% of patients in Arm A and B, respectively.
Safety Outcomes
Safety analyses included all patients who received any study agents. Adverse events were as expected for the individual components of the regimens in Arms A and B. Immune-related AEs not including vaccine or infusion related AEs are listed in Table 3. Six patients required systemic steroid intervention. The remaining nivolumab related AEs were asymptomatic laboratory investigations and 5 cases of rash. These rates are not greater than what would be expected with nivolumab administered as a single agent. Related AEs leading to study treatment discontinuation occurred in 2.2% (n=2) patients, both from Arm A. There were no treatment-related deaths. Serious adverse events (SAE) related to study treatment were reported in 6.5% (n=6) patients. The treatment-related adverse event profile for arm B was similar to prior experience with chills, pyrexia, and fatigue being the most commonly reported(4–6).
Table 3.
Immune-related adverse events.
| Cy/NIVO/GVAX+NIVO/CRS-207 | Cy/GVAX+CRS-207 | |||||
|---|---|---|---|---|---|---|
| Arm A | Arm B | All Subjects | ||||
| (N = 51) | (N = 42) | (N = 93) | ||||
| All grades | Grades 3/4 | All grades | Grades 3/4 | All grades | Grades 3/4 | |
| n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | |
| Rash/Pruritis | 5 (9.8) | 0 | 0 | 0 | 5 (5.4) | 0 |
| Hyperglycemia | 1 (2) | 1 (2) | 0 | 0 | 1 (1.1) | 1 (1.1) |
| Pneumonitis | 1 (2) | 1 (2) | 0 | 0 | 1 (1.1) | 1 (1.1) |
| Myocarditis | 1 (2) | 1 (2) | 0 | 0 | 1 (1.1) | 1 (1.1) |
| Hypothyroidism/hyperthyroidism | 2 (3.9) | 0 | 0 | 0 | 2 (2.2) | 0 |
| Arthralgia | 2 (3.9) | 0 | 0 | 0 | 2 (2.2) | 0 |
| Thrombocytopenia | 1 (2) | 1 (2) | 0 | 0 | 1 (1.1) | 1 (1.1) |
| Renal failure, acute | 1 (2) | 1 (2) | 0 | 0 | 1 (1.1) | 1 (1.1) |
| Diarrhea | 1 (2) | 1 (2) | 0 | 0 | 1 (1.1) | 1 (1.1) |
| Transaminitis | 1 (2) | 1 (2) | 0 | 0 | 1 (1.1) | 1 (1.1) |
| Pancreatitis/hyperamylasemia | 0 | 0 | 0 | 0 | 0 | 0 |
Immune Responses
Intratumoral immune complexity profiles
Longitudinal changes in intra-tumoral immune complexity and phenotypes were evaluated via multiplex IHC on the 22 available paired biopsy specimens, comparing the baseline and the post cycle 3 samples. Based on longitudinal changes of CD45+ leukocyte cell densities (Figure 6A–B, Supplementary Figure 1A–D), ratios of lymphoid to myeloid cell densities significantly increased from baseline to the end of cycle 3 in the long OS group (survival > 6 months) of Arm A, but no other subgroups (Figure 6B, Supplementary Figure 2A–B). After the end of cycle 3, the long OS group of Arm A exhibited significantly higher ratios of lymphoid to myeloid cell densities than the short OS (survival ≤ 6 months) group of Arm A (Figure 6B).
Figure 6. Longitudinal changes of immune cell complexity profiles in association with overall survival (OS).
FFPE tissue sections derived from biopsies obtained at baseline and at the post cycle (Cy 3) were subjected to immune-detection by the three 12-marker panels of lymphoid and myeloid lineages, and T cells (Supplementary Table 1). (A) Immune cell densities (cell number per mm2) of CD8+ T cell, other lymphoid lineages cells, CD68+CSF1R+ tumor associated macrophages (TAMs), CD66b+ granulocytes, and other myeloid cell lineages are shown, comparing baseline and post Cy3 status among short and long OS groups. (B) The ratios of lymphoid to myeloid cell densities are shown, comparing baseline and post Cy3 status among short and long OS groups. (C, E) Multiplex IHC images for T cell lineages (C) and myeloid cell lineages (E), comparing baseline and post Cy3 of ADU-CL-06–001-022 in arm A. Color annotations and scales were shown. (D, F) Frequency of CD8+ T cells (D) and CD68+ myeloid cells (F) are exhibited as cell percentages of total CD45+ immune cells, comparing baseline and post Cy3 status among short and long OS groups. * and ** indicate P < 0.05 and 0.01, respectively.
Compared to the baseline, increased lymphoid cell densities were notably observed after cycle 3 in the long OS group of Arm A (Figure 6B and Supplementary Figure 2A), including helper T cell and natural killer cell densities (Supplementary Figure 2C–D). The percentages of CD8+ T cells in total CD45+ cells at cycle 3 were significantly higher in the long OS group of Arm A, compared with the short OS group of Arm A (Figure 6C–D). Comparative analysis of CD8+ T cell functional status revealed that the long OS group of Arm A showed high frequency of PD-1– Eomes+ CD8+ T cells at baseline and PD-1– Eomes– CD8+ T cells after cycle 3, compared to the short OS group of Arm A (Supplementary Figure 2E–H).
Among myeloid cell lineages, CD68+ myeloid cells showed decreased cell percentages after cycle 3 in the long OS group of Arm A, but not in the short OS group of Arm A or any group in Arm B (Figure 6E–F). PD-L1 expression was observed in both CD45– and CD45+ populations (Supplementary Figures 3A and 3B). PD-L1 expression on myeloid cell lineages increased after cycle 3 in the long OS group of Arm A.
Cytokine data
Cytokine assays were performed on human serum taken serially from patients with metastatic pancreatic cancer at baseline, after 2 doses Cy/nivolumab/GVAX (Arm A) or 2 doses Cy/GVAX (Arm B), or after 2 doses Cy/nivolumab/GVAX + 1 dose of nivolumab/CRS-207 (Arm A) or 2 doses Cy/GVAX + 1 dose of CRS-207 (Arm B). Comparing the rise and fall of 38 cytokine and chemokine levels from baseline values to the time point after two or three cycles of treatment allowed us to determine immunotherapy’s effects on host innate and adaptive immunity over time. Even though cytokines in general act locally, we postulated that the tendency to have a certain established cytokine profile might be reflected peripherally in the systemic blood circulation. Key results looked at the longitudinal change in cytokine expression profiles based on: 1) trial arm and 2) overall survival. We obtained clinical data that allowed us to correlate comprehensive biomarker data with overall survival. IP-10 expression increased at the third cycle in relative to baseline in Arm A more than its increase in Arm B (log fold change 0.9, FDR adjusted p-value 0.1, Supplemental Figure 4A). At the same timepoint, IL-6 expression had a greater expression above its baseline expression in the patients with <6 months overall survival than in patients with whose survival was ≥6 months overall survival (log fold change 1.5, FDR adjusted p-value of 0.09, Supplemental Figure 4B).
Discussion
The discovery of an immunotherapy combination effective against pancreatic cancer remains elusive. However, each iteration of a therapeutic combination, provides insights into the potential targets that might ultimately lead to an active therapy. While the present study did not meet the ambitious goal of showing a survival advantage to the addition of a PD-1 inhibitor to a prime-boost vaccination strategy, signs of clinical activity coupled with evidence of immunologic changes in the peripheral blood and tumor microenvironment support the ongoing investigation of strategies aimed to simultaneously induce immune cell migration into the tumor and promote their tumor killing capacity.
While the overall survival did not differ between the two arms of the study, the overall survival of approximately 6 months without chemotherapy is similar to the OS seen in the NAPOLI trial, which led to the FDA approval of 5-FU with nanoliposomal irinotecan(14). Using immune related response criteria, there were 2 responses in Arm A, and 1 response in Arm B. The disease control rate, 12 month OS rate, 18 month OS rate, and CA19–9 tumor marker kinetics all favored the nivolumab containing arm. Treatment in Arm A appeared to favor patients who did not have prior radiation. Theoretically, radiation could suppress the host’s ability to mount subsequent immune responses. Due to limited tissue biopsies, differences attributable to factors in the tumor microenvironment could not be delineated.
Interrogation of the tumor microenvironment and peripheral blood confirmed the hypothesis that immunotherapy combinations could exert a biologic effect in patients with pancreatic cancer. These biomarker changes mirror the inflammation induced by the treatments. In the present study, multiplex IHC interrogation of the paired biopsies suggest the following changes in tumors of the long OS group treated with vaccine and anti-PD-1: 1) increased CD8+ T cell density; 2) induction of later effector CD8+ T cells; 3) decreased myeloid cell densities; and 4) increased PD-L1 expression on myeloid cells. The myeloid compartment, in particular, is an active area of investigation for the treatment of pancreatic cancer. However, the relatively small sample size limited the ability to conduct multivariate analyses, and additional studies will be necessary to confirm the hypothesized associations between treatment and overall survival.
Decreased CSF1R+ TAMs and myeloid cells in the tumors obtained during therapy were associated with a more activated T cell phenotype and a longer survival. There are current studies in pancreatic cancer proposed or ongoing targeting CSF1R in combination with PD-1 inhibition with and without vaccination strategies. Other agents targeting the myeloid compartment such as adenosine antagonists, indoleamine-pyrrole 2,3-dioxygenase (IDO) inhibitors, chemokine receptor inhibitors, and epigenetic modulators are also actively being investigated. Longitudinal changes were also noted in the T cell compartment with a more activated T cell phenotype being associated with better outcomes suggesting the possibility that T cells may still play a role in tumor biology even in later stage pancreatic cancer patients.
In this study, we note that IL-6 expression trended towards a greater increase in expression after the third cycle of treatment relative to baseline in patients with survival <6 months compared to patients with survival ≥6 six months. Although this trend failed to reach statistical significance, it is consistent with the literature associating high IL-6 expression with poor outcomes in pancreatic cancer, including promoting human pancreatic cancer development and acting as an independent risk factor for progression to hepatic metastasis in pancreatic cancer patients (16, 17). Identifying this specific inflammatory cytokine that is associated with poor pancreatic cancer survival may lead to improved therapeutic strategies, as IL-6 may be a good potential target for pancreatic cancer treatment.
Previously reported analyses of the peripheral T cell receptor (TCR) of the patients in this study treated with the prime-boost vaccination and nivolumab demonstrated a net diversification of the TCR repertoires, further supporting a biologic effect of this regimen(18). However, in comparison, patients treated with GVAX and the CTLA-4 antagonist, ipilimumab, experienced a greater change in TCR repertoire. Furthermore, treatment with ipilimumab but not nivolumab resulted in a high number of expanded clones which was associated with longer survival. The differential effects of these immune checkpoint inhibitors on the TCR repertoire and clinical benefit of immune checkpoint inhibitor combinations in other diseases support ongoing evaluation of this strategy. Using the vaccination strategies combined separately with nivolumab or ipilimumab has led to a low but objective regression rate, albeit sometimes delayed, in patients with pancreatic cancer. A current enrolling study is evaluating dual checkpoint inhibition with a vaccination approach (NCT03190265). Ongoing and future strategies with GVAX with or without CRS-207, will focus on exploratory studies testing new strategies, such as modifying macrophages, to optimize immune and clinical responses. These vaccine platforms with PD-1 inhibitors are already being combined with an IDO inhibitor (NCT03006302), CSF1R antibody (NCT03153410), and stereotactic body irradiation (NCT03161379). Future strategies should also consider the observation that loss of antigen-presentation molecules contribute to therapeutic resistance (19).
In summary, while the current study did not meet its primary endpoint, the median OS of approximately 6 months, 12 month OS rate of 25%, and occasional responses are of interest in this patient population. Treatment emergent changes in the tumor microenvironment, including expansion of functional CD8+ T cells and a decrease in in TAMs and myeloid cells, were associated with better overall survival. The data supports further evaluation of combination strategies to be tested in clinical studies that will continue to collect informative biologic correlates and improve the clinical response rates, durability of response, and survival.
Supplementary Material
Translational Relevance.
Immunotherapy has shown limited success in pancreatic cancer indicating a need for synergistic combinations. A “prime-boost” immunotherapy strategy utilizing cyclophosphamide with the pancreas vaccine GVAX, composed of pancreatic cancer cells engineered to secrete granulocyte-macrophage colony stimulating factor (GM-CSF), followed by CRS-207, an attenuated Listeria monocytogenes which expresses the tumor- associated antigen mesothelin, is safe and induces T cell responses. Here we added the PD-1 inhibitor nivolumab to this combination in a randomized two-arm phase II study to measure improvements in overall survival (OS) as well as immunologic changes in the tumor microenvironment. While the addition of nivolumab (Cy/nivolumab/GVAX followed by nivolumab/CRS-207) did not increase overall survival over Cy/GVAX followed by CRS-207), objective responses were seen and the overall survival was comparable to existing treatments. Furthermore, immunologic changes in the tumor microenvironment were evident. Further studies will examine combinations that will maximize immunologic responses and increase patient survival.
Acknowledgements
The study and analyses were funded by the Pancreatic Cancer Action Network-AACR Research Acceleration Network Grant, supported by the Fredman Family Foundation, Grant Number 14–90-25-LE, a Stand Up To Cancer-Lustgarten Foundation Pancreatic Cancer Convergence Dream Team Translational Research Grant (SU2C-AACR-DT14–14) (D. Le, E. Jaffee, T. Tsujikawa, L. Coussens), Aduro Biotech(D. Le, E. Jaffee), Bristol-Myers Squibb(D. Le, E. Jaffee), the Bloomberg Kimmel Institute(D. Le, E. Jaffee), and NCI/NIH SPORE P50CA062924 (D. Le, E. Jaffee) and CA62924(D. Le, E. Jaffee). Kelly Gemmill assisted in manuscript preparation. Stand Up to Cancer is a division of the Entertainment Industry Foundation. The abovementioned Dream Team grant is administered by the American Association for Cancer Research, the Scientific Partner of SU2C.
Footnotes
Conflict of Interest Statement: T. Tsujukawa is a consultant for Konica Minolta and Ono Pharmaceutical. T. Crocenzi has received research support from Bristol-Myers Squibb and AstraZeneca. R.A. Anders is on the scientific advisory board or receives research support from Merck, Bristol-Myers Squibb, FLX Bio, Incyte and Adaptive Biotech. E.J. Fertig is a consultant for Champions Oncology. K.A. Reiss has research support from Clovis, Bristol-Myer Squibb, Tesaro and Lilly Oncology. R.H. Vonderheide reports having received consulting fees or honoraria from Apexigen, AstraZeneca, Celgene, Genentech, Janssen, Lilly, Medimmune, Merck and Verastem. He has received research funding from Apexigen, Fibrogen, Inovio, Janssen, and Lilly. He is an inventor on a licensed patent relating to cancer cellular immunotherapy and receives royalties from Children’s Hospital Boston for a licensed research-only monoclonal antibody. A.H. Ko has received research support from Aduro Biotech and Bristol-Myers Squibb. G.A. Fisher serves on advisory boards for or received honoraria from Merck, Roche/Genentech, Aduro Biotech, and FortySeven. He serves on Data Safety Monitoring Committees for Celgene, CytomX, Silenseed, Terumo, and AstraZeneca. D.G. Brockstedt is a former employee of and is a shareholder in Aduro Biotech. L.M. Coussens is a paid consultant for Cell Signaling Technologies, received reagent and/or research support from Plexxikon Inc., Pharmacyclics, Inc., Acerta Pharma, LLC, Deciphera Pharmaceuticals, LLC, Genentech, Inc., Roche Glycart AG, Syndax Pharmaceuticals Inc., and NanoString Technologies, and is a member of the Scientific Advisory Boards of Syndax Pharmaceuticals, Carisma Therapeutics, Zymeworks, Inc, Verseau Therapeutics, and Cytomix Therapeutics, Inc. E.M. Jaffee and Johns Hopkins University have the potential receive royalties from Aduro Biotech which owns the license to GVAX and Listeria Monocytogenes-mesothelin. E. Jaffee is the Chief Medical Advisor for the Lustgarten Foundation and is active on the scientific advisory boards of Genocea, Adaptive Biotech, DragonFly, CSTONE, Achilles, Parker Institute and the National Cancer Advisory Board. She receives research funding from Aduro Biotech, Amgen, Bristol-Myers Squibb, Hertix and Corvus. D.Le serves on the advisory boards for Merck and Bristol-Myers Squibb and has received research funding from Merck, Bristol-Myers Squibb, Aduro Biotech, Curegenix, Medivir and Nouscom. She has received speaking honoraria from Merck.
References
- 1.Lutz ER, Wu AA, Bigelow E, Sharma R, Mo G, Soares K, et al. Immunotherapy converts nonimmunogenic pancreatic tumors into immunogenic foci of immune regulation. Cancer Immunol Res. 2014;2(7):616–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Quezada SA, Peggs KS, Curran MA, Allison JP. CTLA4 blockade and GM-CSF combination immunotherapy alters the intratumor balance of effector and regulatory T cells. J Clin Invest. 2006;116(7):1935–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Soares KC, Rucki AA, Wu AA, Olino K, Xiao Q, Chai Y, et al. PD-1/PD-L1 blockade together with vaccine therapy facilitates effector T-cell infiltration into pancreatic tumors. J Immunother. 2015;38(1):1–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Le DT, Brockstedt DG, Nir-Paz R, Hampl J, Mathur S, Nemunaitis J, et al. A live-attenuated Listeria vaccine (ANZ-100) and a live-attenuated Listeria vaccine expressing mesothelin (CRS-207) for advanced cancers: phase I studies of safety and immune induction. Clin Cancer Res. 2012;18(3):858–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Le DT, Wang-Gillam A, Picozzi V, Greten TF, Crocenzi T, Springett G, et al. Safety and survival with GVAX pancreas prime and Listeria Monocytogenes-expressing mesothelin (CRS-207) boost vaccines for metastatic pancreatic cancer. J Clin Oncol. 2015;33(12):1325–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Le DT, Picozzi VJ, Ko AH, Wainberg ZA, Kindler H, Wang-Gillam A, et al. Results from a Phase IIb, Randomized, Multicenter Study of GVAX Pancreas and CRS-207 Compared with Chemotherapy in Adults with Previously Treated Metastatic Pancreatic Adenocarcinoma (ECLIPSE Study). Clin Cancer Res. 2019;25(18):5493–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Laheru D, Lutz E, Burke J, Biedrzycki B, Solt S, Onners B, et al. Allogeneic granulocyte macrophage colony-stimulating factor-secreting tumor immunotherapy alone or in sequence with cyclophosphamide for metastatic pancreatic cancer: a pilot study of safety, feasibility, and immune activation. Clin Cancer Res. 2008;14(5):1455–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hassan R, Thomas A, Alewine C, Le DT, Jaffee EM, Pastan I. Mesothelin Immunotherapy for Cancer: Ready for Prime Time? J Clin Oncol. 2016;34(34):4171–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Le DT, Dubenksy TW Jr., Brockstedt DG. Clinical development of Listeria monocytogenes-based immunotherapies. Semin Oncol. 2012;39(3):311–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kim VM, Blair AB, Lauer P, Foley K, Che X, Soares K, et al. Anti-pancreatic tumor efficacy of a Listeria-based, Annexin A2-targeting immunotherapy in combination with anti-PD-1 antibodies. J Immunother Cancer. 2019;7(1):132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Le DT, Lutz E, Uram JN, Sugar EA, Onners B, Solt S, et al. Evaluation of ipilimumab in combination with allogeneic pancreatic tumor cells transfected with a GM-CSF gene in previously treated pancreatic cancer. J Immunother. 2013;36(7):382–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Seymour L, Bogaerts J, Perrone A, Ford R, Schwartz LH, Mandrekar S, et al. iRECIST: guidelines for response criteria for use in trials testing immunotherapeutics. Lancet Oncol. 2017;18(3):e143–e52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Tsujikawa T, Kumar S, Borkar RN, Azimi V, Thibault G, Chang YH, et al. Quantitative Multiplex Immunohistochemistry Reveals Myeloid-Inflamed Tumor-Immune Complexity Associated with Poor Prognosis. Cell Rep. 2017;19(1):203–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wang-Gillam A, Li CP, Bodoky G, Dean A, Shan YS, Jameson G, et al. Nanoliposomal irinotecan with fluorouracil and folinic acid in metastatic pancreatic cancer after previous gemcitabine-based therapy (NAPOLI-1): a global, randomised, open-label, phase 3 trial. Lancet. 2016;387(10018):545–57. [DOI] [PubMed] [Google Scholar]
- 15.Ritchie ME, Phipson B, Wu D, Hu YF, Law CW, Shi W, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Research. 2015;43(7). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Feurino LW, Zhang YQ, Bharadwaj U, Zhang RX, Li F, Fisher WE, et al. IL-6 stimulates Th2 type cytokine secretion and upregulates VEGF and NRP-1 expression in pancreatic cancer cells. Cancer Biology & Therapy. 2007;6(7):1096–100. [DOI] [PubMed] [Google Scholar]
- 17.Kim HW, Lee JC, Paik KH, Kang J, Kim J, Hwang JH. Serum interleukin-6 is associated with pancreatic ductal adenocarcinoma progression pattern. Medicine (Baltimore). 2017;96(5):e5926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hopkins AC, Yarchoan M, Durham JN, Yusko EC, Rytlewski JA, Robins HS, et al. T cell receptor repertoire features associated with survival in immunotherapy-treated pancreatic ductal adenocarcinoma. JCI Insight. 2018;3(13). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pandha H, Rigg A, John J, Lemoine N. Loss of expression of antigen-presenting molecules in human pancreatic cancer and pancreatic cancer cell lines. Clin Exp Immunol. 2007;148(1):127–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.