Chimeric antigen receptor (CAR) T cells targeting mesothelin plus pembrolizumab, and atezolizumab plus bevacizumab, have recently shown clinical efficacy in phase I trials in malignant pleural and peritoneal mesothelioma. Despite being tested in a highly selected patient population and requiring a complex engineering that can hardly be upscaled, CAR T cells combined with pembrolizumab bring the first proof of efficacy in cold solid tumors with low genomic heterogeneity, while atezolizumab–bevacizumab offers an easy-to-use combination of antiangiogenics and immunotherapy in an orphan disease.
See related article by Raghav et al., p. 2738.
See related article by Adusumilli et al., p. 2748.
Malignant pleural (MPM) and peritoneal mesothelioma (MPeM) are rare and aggressive cancers that have had limited therapeutic options with moderate survival benefit and impaired quality of life for a very long period (1). Cisplatin–pemetrexed has been the mainstay first-line treatment for nearly 17 years in MPM, with 13 months as a reference of overall survival (OS), and it has also been adopted in MPeM (2, 3). The addition of bevacizumab in selected patients with MPM has improved OS from 16.1 to 18.8 months (1). Comprehensive molecular analyses of the two main pathologic subtypes of MPM (epithelioid and the poor-prognosis sarcomatoid subtype) have not revealed any actionable targets, which contrasts with other thoracic cancers (4).
In the post-platinum setting, single-agent immunotherapy PD-1/PD-L1 blockers have a modest benefit with median progression-free survival (PFS) ranging from 2.5 to 6.2 months (reviewed in ref. 5; Supplementary Table S1). The combination of nivolumab with ipilimumab, two antibodies targeting PD-1 and CTLA4, respectively, in untreated patients improved OS as compared with standard chemotherapy in patients presenting MPM (median OS, 18.1 months vs. 14.1 months; HR, 0.74; ref. 2). The primary resistance to immune checkpoint blockers might be explained by the low PD-L1 expression and low tumor mutational burden (TMB; ref. 6). Compared with MPM, MPeM has slightly higher PD-L1 expression and TMB (5) and less 9p21 deletion that contains a cluster of IFN genes (7). However, it is unknown whether this different biology could unlock immune resistance to checkpoint blockers. To date, there are no available data on the activity of a single agent targeting PD-1/PD-L1 in MPeM.
In light of these data, novel therapies are urgently needed for the treatment of these orphan malignancies. In two recently published articles, very promising results of clinical efficacy placed chimeric antigen receptor (CAR) T cells plus pembrolizumab and bevacizumab plus atezolizumab in the spotlight, as potential breakthrough discovery treatments in MPM and MPeM, respectively (refs. 5, 8; Fig. 1).
CAR T-cell therapy has been successfully developed in hematologic malignancies, but has shown limited benefit in solid tumors so far (9). Possible explanations include the inadequacy of the CAR T-cell target with the tumor target expression, CAR T targeting a single antigen in heterogeneous tumors, or reactive inhibitory checkpoint inactivation following CAR T-cell monotherapy (9). Mesothelin is an ideal cell-surface antigen to target. It is involved in tumor invasion, and is expressed in 85% to 90% of MPM compared with lower levels in mesothelia (10). In the article by Adusumilli and colleagues in this issue, the regional administration of mesothelin-targeted CAR T cells followed by the administration of the PD-1 blocker pembrolizumab was associated with unusual survival outcomes, reaching a median OS of 23.9 months in 18 pretreated patients with MPM (8). In this trial, patients with mesothelin-positive MPM (i.e., expression of at least 10% of tumor cells) were eligible. The intrapleural administration of CAR T cells was feasible and reached long-lasting systemic circulation by preventing the sequestration of CAR T cells in the lungs. In 39% of patients, CAR T cells were detected in peripheral blood for more than 100 days. The treatment had no grade 5 adverse events, and all grade 4 events were reversible laboratory abnormalities associated with lymphodepleting chemotherapy. There were no grade 3 events such as cytokine release syndrome, neurotoxicity, or on-target, off-tumor toxicity. The best overall response in patients with measurable disease was partial response in two of 16 patients and stable disease in nine of 16 patients (8). Although the objective response rate seems lower than expected with second-line PD-1/PD-L1 blockers that range from 20% to 30%, responses/stable disease was maintained for more than six months in half of the patients receiving the immunotherapy combination. Moreover, as seen with CTLA4 blockers in prostate cancer or in dendritic vaccines, bringing T cells in a cold tumor should probably be accompanied by a blockade of other inhibitory pathways to prevent their reactive overexpression (9).
These optimistic results must be tempered by the absence of the control group that could better clarify the extent of benefit related to patient selection. Beyond the predominant proportion of epithelioid subtype, the trial design intrinsically selected patients with the more indolent, slowly progressive, or nonprogressive diseases. Out of 71 patients screened with malignant pleural disease, only 35 underwent leukapheresis and 27 were dosed (N = 23 with MPM). The median time elapsed between screening and treatment administration was three months (8). Moreover, more than 50% were not progressing after prior chemotherapy, thus making the extent of the benefit exclusively explained by CAR T cells plus immunotherapy unclear. Another confounding factor is that the majority of patients received cyclophosphamide plus pembrolizumab prior to CAR T-cell administration. Thus, it will be important to interrogate the activity of CAR T cells plus PD-1/PD-L1 blockers versus standard-of-care treatment, in particular nivolumab–ipilimumab, in an all-comer population (2). We can only hypothesize that the different mechanism of action of CAR T cells may overcome primary resistance to immunotherapy. Finally, this single-center study evaluates a treatment that requires complex engineering, and enhances capacity in interventional radiology to deliver the CAR T cells on site. That might limit the access of such therapy in a broader population and in community centers.
In contrast, Raghav and colleagues explored two easy-to-use drugs targeting PD-L1 (atezolizumab) and VEGF (bevacizumab) in 20 patients with advanced MPeM, after progression or intolerance to prior platinum–pemetrexed (5). The majority of patients experienced progression on prior platinum–pemetrexed (N = 17/20). The combination therapy was associated with an objective response rate of 40% according to RECIST v1.1 criteria, a median duration of response of 12.8 months, a median PFS of 17.6 months, and meaningful one-year OS of 85%. Moreover, this combination showed an acceptable safety profile, consistent with prior reports for each drug, both approved in other tumor types. The tested combination is hypothesized to function by bevacizumab's reversal of VEGF-mediated immunosuppression with subsequent enhancement of T cell–mediated cell killing. However, the biomarker analysis did not find any predictive value for PD-L1, TMB, microsatellite instability, angiogenesis, immune sensitivity signatures, or components of immune cells. High scores of the epithelial-to-mesenchymal transition signatures were associated with therapeutic resistance to both atezolizumab–bevacizumab and prior platinum–pemetrexed chemotherapy, even among patients with epithelioid subtypes (5).
Along with the potential selection bias inherent to a small patient population, the heterogeneous natural history of MPeM makes the true impact of a therapeutic intervention challenging to determine. Selection bias is a real issue in trials evaluating immunotherapy, notably because these patients have tumors that are more indolent. Here again, patient selection is to be highlighted, with epithelioid mesothelioma being the main pathologic subtype and only two biphasic mesotheliomas included. Moreover, 60% had prior surgery, 60% had only one line of treatment, and 60% had performance status (PS) 0. The authors addressed this issue by studying the clinical outcomes on prior platinum–pemetrexed chemotherapy in the patient population (5). However, platinum sensitivity might not recapitulate indolent biology and sensitivity to immunotherapy.
These two articles highlight important perspectives. Cell therapy targeting mesothelin warrants further investigation, but the field of such “hard immunotherapy” is highly competitive with the development of bispecific T-cell engager antibodies and tumor-infiltrating lymphocyte therapy. Their development could serve other tumors with high expression of mesothelin (ref. 10; e.g., KRAS-mutant non–small cell lung cancer). The extensive rationale of combining antiangiogenic drugs with checkpoint blockers is strengthened by the positive results of bevacizumab and atezolizumab in other hard-to-treat diseases such as advanced hepatocellular carcinoma or oncogene-addicted non–small cell lung cancer. Confirmatory results are needed to quickly implement new treatment options in our algorithms, while waiting for more innovative therapeutic approaches.
M. Aldea reports other support from Sandoz outside the submitted work. N. Chaput reports personal fees from AstraZeneca, grants from Sanofi and GSK, and other support from Cytune outside the submitted work. B. Besse reports grants from Roche-Genentech during the conduct of the study; grants from 4D Pharma, AbbVie, Amgen, Aptitude Health, AstraZeneca, BeiGene, Blueprint Medicines, BMS, Boehringer Ingelheim, Celgene, Cergentis, Cristal Therapeutics, Daiichi-Sankyo, Eli Lilly, GSK, Inivata, Janssen, Onxeo, OSE Immunotherapeutics, Pfizer, Sanofi, Takeda, and Tolero Pharmaceuticals outside the submitted work. No disclosures were reported by the other author.
Figure 1 is an unpublished original work, created by Mihaela Aldea with BioRender.com, for the express purpose of publication in this Cancer Discovery article.