Clinical Response of Live-Attenuated, Listeria monocytogenes Expressing Mesothelin (CRS-207) with Chemotherapy in Patients with Malignant Pleural Mesothelioma

This article is featured in Highlights of This Issue, p. 5727

Malignant pleural mesothelioma (MPM) is a rare but life-threatening disease associated with prior exposure to asbestos or other small carcinogenic fibers (1). Approximately 2,500 new cases of mesothelioma are diagnosed in the United States each year, with the majority occurring in men. Globally, mesothelioma causes approximately 43,000 deaths per year. Due to the 20- to 40-year latency period between fiber exposure and diagnosis, the incidence of mesothelioma continues to increase in countries where the commercial use of asbestos persists, with peak incidences expected between 2012 and 2014 in Australia, between 2015 and 2024 in Italy, and 2027 in Japan (2–4). South Africa once held a major asbestos mining industry, which peaked in 1977 with a production of 380,000 tons (5).

Because the majority of patients with MPM present with advanced disease, they are often not candidates for surgical resection. The only FDA-approved treatment for patients with unresectable MPM is combination pemetrexed and cisplatin chemotherapy, which demonstrates median overall survival (mOS) of 12.1 months (1, 6, 7), with a 1-year OS of 50.3%. Since approval of this chemotherapy regimen in 2004, no new agents have been approved for treatment of MPM. The severity of the disease, high mortality rate, and lack of treatment options underscore the urgent and unmet need for additional therapies to be developed for this lethal malignancy.

Immunotherapy offers the potential to improve patient outcomes by inducing specific antitumor immune responses, some of which can be directed against tumor-specific antigens. Previous studies have revealed that high levels of CD8⁺ tumor-infiltrating lymphocytes were associated with improved prognosis in patients with MPM who underwent surgical resection (8). Mesothelin (MSLN), a cell-surface tumor differentiation antigen, is an attractive target for immunotherapy due to its limited distribution in homeostatic normal human tissue (mesothelial cells lining the pleura, peritoneum, and pericardium), and broad overexpression in many cancers, including virtually all epithelial mesotheliomas (9–15). In addition, MSLN is shed into the serum of patients with mesothelioma, where it has been identified as a potential biomarker of tumor response to therapy (16–18).

Listeria monocytogenes (Lm) is a Gram-positive, facultative intracellular bacterium that potently stimulates innate and adaptive immune responses through recruitment of select myeloid cells, and activation of CD4⁺ and CD8⁺ T cell-immunity specific for encoded heterologous antigens (19–22). CRS-207 is a live, attenuated, double-deleted Lm (LADD), which is nonvirulent with the addition of an expression cassette encoding human MSLN (23). CRS-207 was constructed using precise deletions of the entire coding sequences for virulence determinants, actin assembly inducing protein A (actA), and internalin B (InlB), by homologous recombination, thus attenuating the pathogenicity of Lm in mice by 1,000-fold as compared with wild-type, while retaining its immunogenicity (24). Preclinical studies revealed that CRS-207 elicits MSLN-specific CD4⁺ and CD8⁺ T-cell responses in treated mice and cynomolgus monkeys, while demonstrating therapeutic efficacy in tumor-bearing mice (25). In clinical trials, over 400 patients with advanced cancer have received CRS-207 alone or in combination with other therapies (23, 25). A phase I study in patients with advanced mesothelioma, non–small-cell lung cancer, ovarian cancer, and pancreatic adenocarcinoma who had previously failed standard therapy revealed that CRS-207 was well tolerated at a dose of 1.0 × 10⁹ colony-forming units (CFU; ref. 23). Subsequent clinical studies administering 1.0 × 10⁹ CFU CRS-207 have thus far reported an acceptable safety profile and have demonstrated the ability to induce a robust innate and adaptive, MSLN-specific T-cell response (23, 25).

Emerging data indicate that immunotherapies targeting checkpoint inhibitory molecules, administered in combination with cytotoxic therapies, show promise for synergistic antitumor effectiveness, through programming tumor microenvironments, enhanced antigen presentation potential by tumor-associated antigens, reduction of T cell–suppressive immune cells (in some cases), and tumor lysis followed by cross-priming (26–28). Clinical studies of combination immunotherapy and chemotherapy have revealed improved chemotherapy response rates, longer time periods to progression and tumor recurrence, using a variety of regimens for various cancers (29–34). Herein, we present the first study to investigate safety and immunogenicity of CRS-207 when combined with standard-of-care chemotherapy in patients with MPM. Results indicate that the majority of reported patients (57%) with advanced unresectable mesothelioma exhibited objective tumor responses, a higher proportion of patients compared with historical controls (25%–35%). Notably, tumor size reduction was observed in 31% of patients in response to CRS-207 alone, prior to administration of chemotherapy, and was associated with characteristics indicative of immune reprogramming of the tumor microenvironment. In addition, CRS-207 in combination with chemotherapy induced significant changes in the frequency of circulating immune cell subsets. Remodeling in the tumor microenvironment as evidenced by in situ changes in immune cell complexity and effector phenotype was observed following CRS-207 prime infusion.

Sixty patients were enrolled into two cohorts in this phase Ib clinical trial. In cohort 1, the safety and tolerability of CRS-207 in combination with pemetrexed and cisplatin were studied in patients with newly diagnosed unresectable MPM who were not eligible for surgical resection. Up to 16 subjects were originally planned for this study cohort based on calculated power to evaluate induction of immune response, plus an expansion phase with a minimum of an additional 16 subjects to obtain further safety, immune and tumor response data. Patients in cohort 2 received cyclophosphamide (200 mg/m²) 1 day prior to each CRS-207 dose. This study was not powered to compare cohorts 1 and 2. Herein, results from cohort 1 only are described.

The treatment regimen consisted of two prime infusions (prime dosing), 2 weeks apart, of CRS-207 (1 × 10⁹ CFU in 0.9% sodium chloride given i.v. over 2 hours), prior to receiving up to six 3-week cycles of pemetrexed and cisplatin chemotherapy 2 weeks later. Three weeks following completion of chemotherapy, two infusions of CRS-207 were given 3 weeks apart (boost dosing). Patients were eligible to receive CRS-207 maintenance therapy at the first follow-up visit and every 8 weeks thereafter provided there was no disease progression. Patients received acetaminophen, ibuprofen, or meperidine as needed for fever, rigors, and chills after the infusion. To ensure clearance of CRS-207, a 7-day course of oral amoxicillin or trimethoprim/sulfamethoxazole in penicillin-allergic patients was given 7 days after the second CRS-207 priming infusion and after the final dose of CRS-207 when the treatment was discontinued.

Chemotherapy with pemetrexed and cisplatin was administered 2 weeks after the second priming dose. Pemetrexed (500 mg/m² BSA in 100 mL 0.9% sodium chloride) was administered as an i.v. infusion over 10 minutes on day 1 of a 21-day cycle and cisplatin (75 mg/m² BSA in 1.0 L 0.9% sodium chloride with mannitol 30 g/L) was given over 2 hours, beginning approximately 30 minutes after the end of pemetrexed administration on day 1 of each 21 day cycle. Patients received up to six cycles of chemotherapy. If patients were unable to tolerate cisplatin, carboplatin could be substituted using prescribed dosing guidelines. The study timeline of treatment and follow-up schedule is shown in Supplementary Fig. S1.

Patients eligible for this study had histologically confirmed epitheliod or biphasic pleural mesothelioma not amenable to potentially curative surgical resection at the time of study entry. Patients with biphasic tumors with ≥50% sarcomatoid component were excluded. Patients had to be 18 years of age or older with an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1 (on a scale from 0 to 5, with higher numbers indicating greater tumor-related disability; a score of 0 means patients are asymptomatic, while 1 indicates patients have mild symptoms) with adequate renal, hepatic, pulmonary, and hematologic function (35). In addition, patients had to have measurable disease as defined by modified Response Evaluation Criteria in Solid Tumors (mRECIST) for MPM (36). Exclusion criteria included history of autoimmune disease, immunodeficiency disease, systemic immunosuppression; allergy to both penicillin and sulfa drugs; pregnant or breastfeeding women; prosthetic heart valves or major implants or devices placed within 12 months of screening.

Primary objectives of the study were safety of CRS-207 and the induction of immune responses prior to treatment and at time points during and after treatment. Secondary objectives of the study were objective tumor response, time to progression, OS, and the predictive value of serum MSLN for therapeutic response. Exploratory objectives were immune subset analysis by IHC and gene-expression profiling of tumor tissue pre- and post–CRS-207 administration, the induction of antimesothelin humoral immune response and tumor marker kinetics as measured by enzyme linked immunosorbent assay (ELISA). All listed endpoints were prospectively defined.

Radiologic tumor response, progression-free survival, and time to progression were assessed by investigators as secondary endpoints using mRECIST for MPM. OS was also measured. Tumor response was assessed by CT scans of the thorax and abdomen and were obtained before start of therapy, at week 4 (i.e., after two doses of CRS-207 and before starting chemotherapy), every 6 weeks after starting chemotherapy and then every 8 weeks until disease progression. Tumor response was graded as complete response (CR: complete disappearance of measurable disease), partial response (PR: decrease in sum of measurable lesions greater than 30%), stable disease (SD: decrease in tumor size less than 30% or increase less than 20%), and progressive disease (PD: increase in tumor size of greater than 20% or appearance of new lesions). Best overall response (BOR) for SD was identified by the best tumor assessment sustained for a minimum duration of 12 weeks (84 days). A patient was considered not evaluable if lost to follow-up prior to the minimum duration from baseline. All responses were confirmed by a repeat imaging at least 4 weeks later. In addition to mRECIST for tumor response, immune-related RECIST criteria were used to assess immune responses (37, 38). Patients with PD by CT scan could continue on treatment if the patient was clinically stable and met dosing eligibility.

Safety was assessed by evaluation of adverse events and deaths, vital signs, physical examination, clinical chemistry, and hematology laboratory findings in treated patients. All adverse events were reported from the time study treatment was first administered through 28 days after the final dose of investigational product. Clinical adverse events and laboratory abnormalities were evaluated and graded using the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03 (39). Causality and the relationship between the investigational product and the occurrence of each adverse event were assessed based on the best clinical judgment of the investigator.

Optional tumor biopsies were obtained for IHC analysis from primary tumor sites and metastatic sites (as applicable) by standard technique, such as CT guided biopsy. Two to four cores of tumor tissue were obtained from formalin-fixed paraffin-embedded (FFPE) blocks. In addition, unstained archival slides were obtained from the diagnostic blocks for baseline IHC.

Peripheral blood samples were obtained from patients at baseline and over the course of treatment. Peripheral blood mononuclear cells (PBMCs), serum, and plasma were isolated and cryopreserved at each time point within 6 hours of blood collection. Immune monitoring and biomarker analyses were performed on available cryopreserved specimens.

The concentration of serum CA-125 was determined at clinical sites. The concentrations of serum mesothelin and plasma osteopontin were measured using Quantikine ELISA kits (R&D Systems) per manufacturer's protocol. Total IgG specific to mesothelin was measured as described (40). Data analysis was performed using Softmax Pro (version 6.3; Molecular Devices).

Sequential multi-plex IHC was performed as previously described (41), using two 12-antibody panels reflective lineage-selective antibodies, in addition to a panel of antibodies evaluating T-cell functionality (Supplementary Figs. S2 and S3), to identify lymphoid and myeloid immune cells in biopsy tissues following sequential antibody staining and stripping treatments (41). Coregistration and conversion to pseudofluorescence images were processed by CellProfiler (cellprofiler.org) and ImageJ (imagej.nih.gov), and visualized by Aperio ImageScope (Leica Biosystems). Single-cell–based chromogenic intensity was quantified by CellProfiler, and the number and phenotypes of cells were analyzed by FCS Express 5 Image Cytometry (Denovo Software). Antibodies used were αPD-1 (NAT105, Abcam, 1:50), αCD3 (SP7, Thermo Scientific, 1:150), αRORgt (6F3.1 EMD Millipore, 1:200), αCD56 (123C3, Santa Cruz Biotech, 1:25), αCD8 (C8/144B, Thermo Scientific, 1:100), αT-bet (H-210, Santa Cruz Biotech, 1:100), αGATA3 (L50-823, BD Biosciences, 1:100), αFoxp3 (236A/E7, eBioscience, 1:40), αCD20 (0.N.85, Santa Cruz Biotech, 1:1,000), αCD45 (H130, Thermo Scientific, 1:100), αMesothelin (EPR2685(2), Abcam, 1:500), αTryptase (AA1, Abcam, 1:20,000), αCD68 (PG-M1, Abcam, 1:50), αCSF1R (SP211, Abcam, 1:150), αDC-SIGN (DC-28, Santa Cruz Biotech, 1:100), αCD66b (G10F5, eBioscience, 1:600), αCD83 (1H4b, Abcam, 1:40), αCD163 (10D6, Thermo Scientific, 1:100), αMHC class II (SPM288, Novus Biological, 1:100), αTbr2 (Eomes) (AB2283, EMD Millipore, 1:1,000), and αKi67(SP6, Abcam, 1:500). Data sets from the lymphoid and myeloid panels were normalized based on the CD45⁺ cell number.

Multicolor flow cytometry using four panels of lineage-specific antibodies was used to identify major lymphoid and myeloid immune cells in circulation. Cryopreserved PBMCs were thawed, and 0.5 million PBMCs were stained per reaction. Cells were incubated with Human Fc Block (BD Biosciences) for 10 minutes at room temperature, washed with phosphate-buffered saline (PBS) and incubated with LIVE/DEAD fixable aqua dead cell stain (Thermo) for 20 minutes at room temperature. Cells were washed with staining buffer (PBS containing 0.5% bovine serum albumin and 1 mmol/L EDTA) and then stained for 30 minutes at 4°C with the antibodies. Cells were washed three times with staining buffer and acquired on BD X20 LSRFortessa instrument (Becton Dickinson). Events (200,000) were acquired per sample. Data were analyzed using Cytobank. Immune cell subsets were analyzed following exclusion of dead cells and doublets.

This study was designed by the academic authors in collaboration with the sponsor (Aduro Biotech); the sponsor worked jointly with the investigators to collect and analyze the data. This study was performed in accordance with Good Clinical Practice and the Declaration of Helsinki and the research protocol was approved by the institutional review boards at each participating institution, as well as by the NIH Recombinant DNA Advisory Committee. All human participants gave written informed consent.

Demographics and baseline clinical variables for patients were summarized using descriptive statistics. Tumor response analyses were conducted on the full analysis (all patients enrolled who received at least one dose of study treatment) from cohort 1 and per protocol (patients who received at least two doses of CRS-207 followed by at least two cycles of pemetrexed and cisplatin) populations.

A total of 38 patients were enrolled in cohort 1 of this study, and the median time on treatment was 6.7 months (median time on study, 8.5 months). The median age of patients was 69 years (range, 51–82 years) encompassing 34 (89%) male and four (11%) female patients. Fifteen patients (40%) had ECOG PS 0, whereas 23 (61%) had ECOG PS 1. One patient who was initially presumed to have mesothelioma was removed from the study after receiving one prime infusion of CRS-207 when subsequent pathologic examination revealed primary lung malignancy and not pleural mesothelioma. Of the 37 patients with mesothelioma, 33 (89%) had epithelioid histology and four (11%) had biphasic disease (Table 1). All data presented herein are as of the data cutoff date of April 2018.

At data cutoff, two patients with MPM remained on treatment (N = 37). Thirty-six (97%) patients received both CRS-207 prime doses and 31 (84%) received at least one CRS-207 booster dose (Table 1). Thirty patients completed a full treatment course consisting of two prime doses, up to six cycles of chemotherapy and two boost doses (lasting 25 weeks, including end of course visit at 4 weeks after the second boosting dose of CRS-207) out of 38 who either reached 25 weeks or discontinued the study early. Of the 30 patients who completed their full treatment course, 18 (60%) received at least one maintenance dose with CRS-207. We evaluated pre- and posttreatment serum antimesothelin antibody level in 15 patients. One patient had detectable antibody level at base line. Out of 14 patients, one developed antimesothelin antibody posttreatment (data not shown). The median number of cycles of pemetrexed and cisplatin administered was six (range, 1–6), and in nine patients, carboplatin was substituted for cisplatin. Two patients with MPM did not have postbaseline tumor measurements, in one case due to clinical progression leading to death, and one due to patient withdrawal from the study after one CRS-207 dose.

Out of the 35 patients evaluable for tumor response, one patient (3%) had a CR, 19 (54%) had PR, 10 (29%) had SD, resulting in 86% overall disease control as their BOR on study. Five patients (14%) had PD as their BOR (Fig. 1A), with two of these patients categorized as PD due to initial postbaseline measurement, but subsequent scans revealed a decrease in tumor volume (−21% and −23%) that was maintained, indicating SD that did not meet the minimum duration of response of the program. The CR occurred while the patient was on maintenance CRS-207 following chemotherapy. The response pattern was varied as some patients initially experienced an increase in tumor volume before responding to treatment where some patients showed a decrease in tumor lesion size after CRS-207 treatment alone prior to chemotherapy. Overall, 89% (31/35) exhibited tumor size reduction on treatment (Fig. 1A). The median time to response was 3.5 months (95% CI, 2.3–4.9), and the median duration of response was 5.0 months (95% CI, 3.9–11.5). The durability of tumor response and time to first response is shown in Fig. 1B. In the majority of patients enrolled, PR was observed after two cycles of chemotherapy.

Tumor volume size as determined by CT scan revealed 11 of 35 (31%) patients had reduction in tumor size ranging from −1.6% to −39.2% after CRS-207 alone, prior to administration of pemetrexed/cisplatin chemotherapy (Fig. 1C). Out of the 11 patients who had tumor shrinkage following CRS-207 infusion, seven patients subsequently had partial tumor responses, and one patient had a complete response prior to chemotherapy.

The estimated median progression-free survival was 7.5 months (95% CI, 7.0–9.9). A total of three patients continued treatment after initial progression as determined by modified RECIST. The duration of continued treatment following modified RECIST progression was 4.4, 4.8, and 6.8 months. At the time of last data analysis (April 2018) with a minimum follow-up for OS of 32 months, the estimated median OS based on survival sweep was 14.7 months (95% CI, 11.2–21.9). The OS rate at 1 year was 64.9% (95% CI, 47.3%–77.9%).

The most frequently reported adverse events during the study are summarized in Table 2. There was no case of listeriosis in the cohort. Treatment-emergent adverse events following infusion of CRS-207 and prior to administration of chemotherapy (N = 38) predominantly consisted of characteristic transient side effects of chills, fever, and nausea (Table 3). In the majority of patients, these toxicities were grade 1 to 2, but one patient (3%) had grade 3 fever, one patient had grade 3 hypotension, and two patients (5%) had grade 3 rigors and chills. Grade 1 to 2 nausea was reported in 26 (68%) patients. There were no grade 4 infusion-related adverse events. Laboratory abnormalities observed after CRS-207 treatment consisted mainly of lymphopenia, which resolved quickly. Adverse events observed during treatment were as expected, and prior treatment with CRS-207 did not lead to any additional events. Importantly, there were no additive or cumulative toxicities observed from combination treatment of CRS-207 with chemotherapy. No adverse event was reported for the patient whose diagnosis was changed to NSCLC, nor for the patient who withdrew from the study following one dose of CRS-207.

The kinetics of circulating serum mesothelin (N = 31), plasma osteopontin (N = 31), and serum CA-125 (N = 35) were analyzed in a subset of patients from this cohort. Fourteen of the 17 patients with PR and the single patient analyzed with CR demonstrated a decrease in serum mesothelin at the final posttreatment time point analyzed (Fig. 2A and B). For the patients with SD, three of the eight demonstrated a decrease in levels of serum mesothelin following treatment. Two of the four patients with PD demonstrated an increase in serum mesothelin posttreatment; the changes in plasma osteopontin or serum CA-125 following treatment did not correlate with radiologic tumor responses (Supplementary Fig. S4). Levels of antibodies specific to mesothelin were not detected prior to treatment or at the BV1D1 posttreatment time point in the majority of patients (data not shown).

Phenotyping of immune cell subsets within the T-cell, B-cell, DC, monocyte, and natural killer (NK) cell lineages was performed in peripheral blood from 27 patients. The frequencies of the vast majority of subsets analyzed at 24 weeks posttreatment were comparable with pretreatment levels, indicating that changes to immune cell frequencies in the periphery were transient (data not shown). Significant changes were observed in CD4 Tem (effector memory) and monocytes during CRS-207 priming only and in naïve CD8 T cells during CRS-207 boost treatment only. CD8 Tcm (central memory), pDCs (plasmacytoid dendritic cells), and mDCs (myeloid dendritic cells) changed significantly following combination treatment with CRS-207 and chemotherapy. Monocytes changed during CRS-207 priming alone as well as following chemotherapy and CRS-207 boost (Fig. 3). Overall, a decrease was observed in CD8 Tcm and monocytes with a concomitant increase in pDCs and mDCs at 24 weeks posttreatment compared with baseline. Changes in peripheral immune subsets were not associated with response.

To evaluate intratumoral changes in immune cell phenotype and functionality prior to and following CRS-207 administration, multi-plex IHC (mIHC) analysis was performed, comparing the available three pairs of FFPE samples from pre- and post–CRS-207 treatment biopsies obtained within this cohort (Supplementary Table S1; Supplementary Fig. S2A–S2C). Lymphoid and myeloid cell lineages were quantified according to lineage-selective markers shown in Supplementary Table S1. Quantification of 16 immune cell lineages by multiparameter cytometric analysis (e.g., image cytometry), revealed intratumoral longitudinal changes in immune cell composition of CD45⁺ leukocytes (Supplementary Figs. S2A, S3A, and S3B). Increased percentages of CD8⁺ T cells were observed, as well as an increased CD8⁺ T-cell/regulatory T-cell ratio (Fig. 4A and B). Notably, increased NK cells were detected in posttherapy biopsies, indicating that recruitment of NK cells was facilitated by CRS-207–based immunotherapy in agreement with a previous report (23). While immune-stimulatory (M1-like)/immune-suppressive (M2-like) ratios of macrophages varied regarding degree of immune responses in the three cases analyzed (Supplementary Fig. S2B), decreased percentages of M2-associated macrophages, defined as CD68⁺CD163⁺CSF1R⁺, were observed across all cases post-immunotherapy (Fig. 4A), indicating a M1-type response correlated with therapy. Furthermore, in the postimmunotherapy biopsies, there exhibited an increased presence of CD83⁺ DCs rather than of DC-SIGN⁺ DCs (Fig. 4A and C), possibly reflecting CRS-207–mediated maturation of intratumoral DCs. To further evaluate functional status of tumor-infiltrating CD8⁺ T cells, expression of T-cell terminal differentiation/exhaustion markers, namely, eomesodermin (Eomes) and programed cell death (PD)-1, was analyzed (Supplementary Fig. S3C). Whereas patient No. 004-004 exhibited a remarkable increase of Eomes⁻ PD-1⁻ early progenitor and Eomes⁺ PD-1⁻ late effector CD8⁺ T cells, the remaining two cases exhibited moderate expansion of Eomes⁺ PD-1⁺ population in the postimmunotherapy status (Supplementary Fig. S2C). Across all three cases, postimmunotherapy CD8⁺ T cells showed high positive percentages of the proliferation marker Ki67, particularly in Eomes⁺PD-1⁺ cells, possibly reflecting reinvigoration of “exhausted” T cells, and expansion of late effector CD8⁺ T cells despite exhaustion. These observations together indicate the presence of intratumoral changes of immune complexity, potentially induced by CRS-207 immunotherapy.

In this phase Ib clinical trial, administration of CRS-207 was well tolerated, with no additive or cumulative toxicities when combined with standard-of-care pemetrexed and cisplatin chemotherapy in patients with MPM. Adverse events attributable to CRS-207 were transient, mild, and temporally related to the infusion. The side effects of CRS-207 were similar to what was observed in studies of single-agent CRS-207 and consisted mainly of grade 1 to 2 fever, chills, and nausea following infusion that quickly resolved with supportive care. Tumor and overall response endpoints demonstrated that combination treatment with CRS-207 immunotherapy and chemotherapy led to reduction of tumor size, disease control, and improved survival, including one complete response obtained during CRS-207 maintenance therapy. Compared with historical rates of 21% to 41% ORR with chemotherapy alone, the patients in this trial obtained a 57% ORR, and 86% disease control rate. The ORR appeared to translate to a durable response and survival benefit. Our results also indicate that the administration of CRS-207 either before or following chemotherapy did not result in unexpected toxicities.

It was noted that a varied response pattern resulted from the treatment regimen. There were three patients with no response or PD prior to chemotherapy, 24 patients with initial increase in tumor lesion size followed by immediate tumor shrinkage early on during chemotherapy, and 11 of 35 (31%) patients with a decrease in tumor lesion size (ranging from −1% to −39%) from CRS-207 prime dosing alone before chemotherapy was initiated. To assess the effect of CRS-207 alone on tumor lesion size, patients underwent a CT scan prior to chemotherapy, 2 weeks after the second CRS-207 dose. Because assessments of tumor responses in mesothelioma can be difficult, the effect of CRS-207 in mediating tumor shrinkage was confirmed by serum biomarker analysis. For patients with an initial increase in tumor size, a correlative increase in immune cell infiltration at the same time point suggested pseudoprogression, prior to the reduction in tumor lesion size during chemotherapy. Furthermore, the biomarker and tumor microenvironment changes support possible single-agent activity of CRS-207 in the reduction of tumor lesions in 11 out of 35 patients, which were observed as early as the 4-week posttreatment CT scan.

Although we assessed various reported biomarkers of tumor response, including mesothelin, osteopontin, and CA-125, only changes in serum mesothelin trended with radiologic response in a subset of patients. A decrease in serum mesothelin post-treatment was observed in the majority of patients analyzed with radiologic PR. Although prior studies (42) have reported that treatment with chemotherapy can reduce serum mesothelin levels, data presented herein indicate that an immunotherapy agent by itself may lead to a decrease in mesothelin serum levels as was observed in patients who had tumor shrinkage and reduced mesothelin levels after two infusions of CRS-207 alone. Previous studies in mesothelioma and other cancers have reported usefulness of serum mesothelin as a biomarker for treatment efficacy and disease progression (43–45). Analysis of additional patients with distinct radiologic responses is required to fully assess serum mesothelin as a surrogate for tumor response. Interestingly, antimesothelin antibody were not detected in the majority of patients following CRS-207 treatment, indicating that the decrease in circulating mesothelin is due to the treatment leading to a T-cell–mediated antitumor response.

In addition to a systemic assessment based on circulating immune cells, this study evaluated longitudinal changes in tumor-infiltrating immune cells following CRS-207 prime infusion, revealing expansion and activation of intratumoral CD8⁺ T cells, which have been reported as favorable prognostic factors in MPM (8, 46). However, systemic changes in myeloid cells, including pDCs, mDCs, and monocytes, were more significant compared with CD4 or CD8 T cells, perhaps suggesting more effective engagement upstream of antigen-presenting cells. Post-CRS-207 therapy biopsies revealed an increase in NK cells, indicating that recruitment of NK cells was prompted by Lm-based vaccination as reported previously (23). Notably, among highly infiltrative myeloid populations observed in the MPM tissues, we observed postvaccination increases of CD83⁺ DCs, as well as decreased presence of CD68⁺CSF1R⁺CD163⁺ macrophages, typically associated with T-cell–suppressive and protumoral (M2) activities. Based on preclinical mouse modeling studies (Coussens lab, personal communication and manuscript in preparation), we anticipate that alternations in intratumoral immune presence associated with altered effector phenotype facilitate enhanced proliferation and differentiation of CD8⁺ T cells as revealed by phenotypic effector and Ki67 changes (Supplementary Fig. S2C), consistent with reinvigoration of “exhausted” T cells by relieving T-cell–suppressive pathways, and as observed in other malignancies following checkpoint inhibitory therapy (47–49). Based on the changes in intratumoral immune cell filtrates, combination of checkpoint inhibitors with CRS-207 might be synergistic, an approach to consider as a future therapy. These observations together support the notion that CRS-207 prime infusion was sufficient to elicit reprogramming of the tumor immune microenvironment represented by increased in situ presence of activated Ki67⁺CD8⁺ T cells, potentially as a consequence of in situ priming of tumor antigens. Priming of tumor antigen–specific T cells and favorable microenvironment could contribute to maintenance or reinvigoration of T-cell function during subsequent chemotherapy, possibly leading to therapeutic responses observed in this study.

Limitations of the results and its interpretation include the limited biopsy patient samples available for IHC. Thus, future studies with larger cohorts are required to explore potential correlations between therapeutic responses and alternations in intratumoral immune characteristics. Analysis of additional patients, particularly for serum mesothelin, will be required to follow-up on observations in this study.

The results of our study indicate that combining immune-based therapies with traditional chemotherapy could result in increased antitumor activity in patients with mesothelioma. Further evaluation of the immune changes in the tumor microenvironment will contribute to better understanding of the mechanism of action of LADD-based therapies.

H. Kindler is a consultant/advisory board member for AstraZeneca, Aldeyra, Boehringer Ingelheim, Bristol-Myers Squibb, Astellas, Erytech, Five Prime, Ipsen, Kyowa, and Paredox. S. Honarmand is an employee of and holds ownership interest (including patents) in Aduro Biotech, Inc. E.M. Jaffee reports receiving commercial research grants from AduroBiotech and Bristol-Myers Squibb; holds ownership interest (including patents) in AduroBiotech; and is a consultant/advisory board member for CSTONE, DragonFly, and Genocea. L.M. Coussens reports receiving commercial research grants from Acerta Pharma, Deciphera Pharmaceuticals, Roche Glycart AG, and Syndax Pharmaceuticals; reports receiving other commercial research support from Plexxikon, Pharmacyclics, Acerta Pharma, Deciphera Pharmaceuticals, Genentech, Roche Glycart AG, Cell Signaling Technologies, and NanoString Technologies; and is a consultant/advisory board member for Pharmacyclics, Syndax Pharmaceuticals, Carisma Therapeutics, Verseau Therapeutics, Zymeworks, Melvin and Bren Simon Cancer Center at Indiana Universtiy, Koch Institute for Integrated Cancer Research at MIT, Salk Institute Cancer Center, Bloomberg-Kimmel Institute for Cancer Immunotherapy, Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, Dana-Farber Cancer Center Breast SPORE, Cancer Research Institute, The V Foundation for Cancer Research, Cancer Research United Kingdom, Starr Cancer Consortium, NIH/NCI, and Cell Signaling Technologies. D.G. Brockstedt was an employee of and holds ownership interest (including patents) in Aduro Biotech. No potential conflicts of interest were disclosed by the other authors.

Clinical Trial Registry Number NCT01675765 (50).

Conception and design: R. Hassan, T. Jahan, A. Thomas, E. Jaffee, L.M. Coussens, D.G. Brockstedt

Development of methodology: R. Hassan, T. Jahan, T. Tsujikawa, S. Kumar, E. Jaffee, L.M. Coussens, D.G. Brockstedt

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): R. Hassan, E. Alley, H. Kindler, S. Antonia, T. Jahan, N. Nair, E. Lemmens, T. Tsujikawa, G. Choe, A. Thomas, K. McDougall, L.M. Coussens

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): R. Hassan, E. Alley, H. Kindler, S. Antonia, T. Jahan, S. Honarmand, N. Nair, C.C. Whiting, A. Enstrom, T. Tsujikawa, S. Kumar, A. Thomas, E. Jaffee, L.M. Coussens, D.G. Brockstedt

Writing, review, and/or revision of the manuscript: R. Hassan, E. Alley, H. Kindler, S. Antonia, T. Jahan, S. Honarmand, N. Nair, A. Enstrom, T. Tsujikawa, S. Kumar, A. Thomas, E. Jaffee, L.M. Coussens, D.G. Brockstedt

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): R. Hassan, S. Honarmand, C.C. Whiting, E. Lemmens, G. Choe, K. McDougall

Study supervision: R. Hassan, T. Jahan, S. Honarmand, K. McDougall, A.L. Murphy, D.G. Brockstedt

Other (oversight of clinical trial operations): A.L. Murphy

A special thank you is extended to the referring physicians, participating investigators, and most of all the patients and their families. The authors also thank Jocelyn Hybiske, PhD, an independent consultant, for editorial services. This study was sponsored by Aduro Biotech, Inc. (Berkeley, CA). The Coussens laboratory acknowledges support from UL1 TR000128 (T. Tsujikawa) and NCI/NIH (R01CA15531 and U54CA163123), the Department of Defense Breast Cancer Research Program (W81XWH-11-1-0702), the Susan G Komen Foundation (KG111084), the Brenden-Colson Center for Pancreatic Health, and a Stand Up To Cancer–Lustgarten Foundation Pancreatic Cancer Convergence Dream Team Translational Research Grant (SU2C-AACR-DT14-14) to L.M. Coussens. A. Thomas is supported by the intramural program of the NIH (ZIA BC 011793). R. Hassan is supported in part by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research (ZIA BC010816).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.