Mesothelin, a protein found on cell surfaces and in serum, has emerged as a promising target for immunotherapy-based treatment approaches for several malignancies with poor prognoses and limited treatment options. Investigating novel approaches to optimize delivery and identifying combinations of agents to synergistically improve therapeutic response will be the next steps in bringing mesothelin-targeted therapies into the clinical setting.
Mesothelin, a protein found on cell surfaces and in serum, has emerged as a promising target for immunotherapy-based treatment approaches for several malignancies with poor prognoses and limited treatment options, including mesothelioma, pancreatic cancer, and ovarian cancer. Investigating novel approaches to optimize delivery and identifying combinations of agents to synergistically improve therapeutic response will be the next steps in bringing mesothelin-targeted therapies into the clinical setting.
Mechanisms of Mesothelin Overexpression in Cancer
Mesothelin was first identified as the cell-surface antigen CAK1 in 19921 and characterized as a 40-kDa glycoprotein found on the surface of mesotheliomas and ovarian cancers.2 Although it is commonly a membrane-bound, glycosylphosphatidylinositol-linked protein, the extracellular domain of mesothelin is shed from tumor cells,3 and some data suggest that detection of serum mesothelin may be a useful marker in cancer diagnostic procedures.4
Although the biologic function of mesothelin has not been completely elucidated, it is thought to contribute to cell adhesion, differentiation, and signal transduction in cancer cells.5 Overexpression in cancer cells may promote proliferation, cell migration and spread, chemotherapy resistance, and inhibition of apoptosis through activation of multiple intracellular pathways, such as the NF-κB, MAPK, and PI3K pathways.6,7 Binding of mesothelin to cancer antigen (CA) 125, also known as mucin 16 (MUC16), in the neoplastic setting suggests its role in cell adhesion and potential as a target for anticancer therapy,8 particularly in ovarian cancer, in which mesothelin is implicated in promoting peritoneal metastases (Figure).9 Mesothelin also been shown to activate MAPK pathway signaling independent of MUC16 expression.10
Notably, mesothelin expression on normal cells is limited to mesothelial cells lining the pleura, the peritoneum, and the pericardium. This increases the potential for therapies with a lower risk of off-target toxicities.11
Strategies for Mesothelin-Targeted Therapy
Overexpression of mesothelin has been demonstrated in multiple types of cancer, with mesothelioma, epithelial cancer, and pancreatic adenocarcinoma demonstrating particularly robust expression.12 Investigators at Memorial Sloan Kettering (MSK) Cancer Center in New York, New York, and colleagues have estimated that the prevalence of mesothelin expression in solid tumors ranges from 20% to 25% in endometrial cancers to 85% to 90% in mesothelioma samples.13 High expression is associated with poor prognosis in ovarian cancer,14 cholangiocarcinoma,15 lung adenocarcinoma,16 triple-negative breast cancer,17 and resectable pancreatic ductal adenocarcinoma.18 The association between mesothelin overexpression and poor-prognosis cancer, coupled with the fact that normal expression of mesothelin is limited to cells that are dispensable, makes mesothelin an attractive target for cancer immunotherapy (Table). Approaches for targeting mesothelin that have been investigated in early-stage clinical trials include immunotoxins, vaccines, chimeric monoclonal antibodies, antibody— drug conjugates, and chimeric antigen receptor (CAR) T-cell therapy. However, much of the early clinical trial data suggest that adding immunosuppressive agents or combining mesothelin-directed therapy with chemotherapy or immunotherapy may improve therapeutic response and should be investigated in larger groups of patients.
Overexpression of mesothelin occurs in approximately 95% of malignant mesotheliomas with the epithelioid subtype, and mesothelin- targeted therapy has been of interest because mesothelioma is characterized by poor prognosis and has few FDA-approved treatment options.19
Results from early studies in mesothelioma showed that although the immunotoxin SS1P, which is composed of a mesothelin-targeting antibody fragment genetically fused to a truncated fragment of Pseudomonas exotoxin A, was generally shown to be safe; most patients developed neutralizing antibodies against the toxin portion, thereby limiting long-term efficacy.20,21 Addition of the immunosuppressive agents pentostatin and cyclophosphamide reduced the proportion of patients who developed anti-SS1P antibodies after the first treatment cycle and induced major tumor regression in 3 of 10 evaluable patients.21
In another phase I trial, the maximum tolerated dose of SS1P given with pemetrexed and cisplatin induced partial responses in 10 of 13 patients with unresectable, stage III or IV mesothelioma, and objective radiologic responses were associated with significant decreases in serum mesothelin.22 Recent preclinical research showed that SS1P increased the sensitivity of AE17M mouse mesothelioma cells to CTLA-4 inhibitors, and the combination of systemic anti—CTLA-4 therapy and intratumoral SS1P induced tumor regression and improved survival over monotherapy with either agent in a mouse model of mesothelioma, suggesting that SS1P and CTLA-4 inhibitors may have synergistic antitumor activity.23
LMB-100 (also known as RG7787), a recombinant Pseudomonas exotoxin A—based immunotoxin, was evaluated in a phase I, single-center study of patients with previously treated advanced pancreatic ductal adenocarcinoma. The agent demonstrated antitumor activity at the maximum tolerated dose (65 μg/ kg when given with nab-paclitaxel [Abraxane]), including a confirmed partial response in 1 of 8 patients and a >50% decrease in CA19-9, a serum biomarker often used to assess treatment response in pancreatic cancer, in 5 of 8 patients.24 An ongoing phase II trial is investigating LMB-100 as monotherapy and in combination with nab-paclitaxel in patients with advanced pancreatic adenocarcinoma (NCT02810418), and an ongoing phase I trial is investigating LMB-100 as monotherapy or with nab-paclitaxel in patients with malignant mesothelioma (NCT02798536).
Investigators at the Sidney Kimmel Cancer Center at Johns Hopkins University in Baltimore, Maryland, were the first to identify mesothelin as a target of vaccine-induced CD8-positive T-cell response in patients with pancreatic adenocarcinoma treated with GVAX, a granulocyte-macrophage colony-stimulating factor—secreting whole pancreatic tumor cell vaccine that expresses multiple antigens.25
Subsequently, the investigators collaborated with Aduro Biotech to develop CRS-207, a live-attenuated, double-deleted Listeria monocytogenes strain engineered to express mesothelin, and proposed administering GVAX first as a priming vaccine followed by CRS-207 as a boosting vaccine in patients with previously treated metastatic pancreatic adenocarcinoma.
Although the combination of cyclophosphamide (administered with GVAX to inhibit regulatory T cells), GVAX, and CRS-207 led to an improvement in overall survival over cyclophosphamide and GVAX alone,26 results from a phase IIb study showed that neither CRS-207 nor CRS-207 plus GVAX with cyclophosphamide improved survival over physician’s choice of chemotherapy.27 Based on these results, Aduro discontinued the development programs for the vaccine.
Chimeric Monoclonal Antibody
Amatuximab (previously MORAb-009) is a chimeric, high-affinity monoclonal immunoglobulin G1 κ anti-mesothelin antibody that prevents adhesion of mesothelin-expressing tumor cells to MUC16 and elicits cell-mediated cytotoxicity, particularly when combined with chemotherapy.28
A phase II trial of amatuximab, pemetrexed (Alimta), and cisplatin to treat 89 patients with unresectable malignant pleural mesothelioma showed partial responses in 40% and stable disease in 51%, as well as a median overall survival of 14.8 months.29 However, results of a placebo-controlled, phase II trial showed no improvement in overall survival with amatuximab plus gemcitabine in patients with untreated, unresectable stage III or IV pancreatic cancer (NCT00570713), and results from a subsequent imaging trial using radiolabeled amatuximab showed that the antibody had considerably lower uptake in pancreatic tumors than in mesothelioma.30 Results also showed that some patients had higher uptake in distant metastatic sites, suggesting that response to amatuximab is tissue dependent.
In a preclinical study of the effects of amatuximab on pancreatic cancer metastasis, high-mesothelin expressing pancreatic cancer cells were injected into mice to generate a murine model of peritoneal metastases. The results showed that although amatuximab alone did not reduce overall tumor formation, it suppressed the development of metastases. When amatuximab was used as an adjuvant with gemcitabine, investigators observed a synergistic reduction in tumor mass and metastasis compared with gemcitabine alone.31 Based on these findings, the authors suggested a possible role for amatuximab in the prevention of peritoneal dissemination of pancreatic cancer following resection.
With some of the shortcomings of immunotoxins (eg, immunogenicity) and monoclonal antibodies (eg, poor tumor penetrance), investigators have aimed to research antibody—drug conjugates, which combine the specificity of an antibody with the cytotoxicity of a drug.
Anetumab ravtansine (also known as BAY 94-9343), a fully human antimesothelin monoclonal antibody conjugated via disulfide linkage to the microtubule-targeting toxophore DM4, showed highly selective binding to mesothelin and potent and selective cytotoxicity of mesothelin-expressing cells. The antibody—drug conjugate also demonstrated specific localization to and growth inhibition in mesothelin-positive tumors in subcutaneous and orthotopic xenograft models and induced a bystander effect on neighboring mesothelin- negative tumor cells.32
Results of a phase I trial of patients with solid tumors that were refractory to standard treatment showed rates of partial response, stable disease, and disease control of 18%, 47%, and 65%, respectively, at the maximum tolerated dose.33 Results also showed that 5 of 10 patients with pleural mesothelioma had partial responses to anetumab ravtansine as secondline therapy, which led to development of a phase II trial (NCT02610140) that will investigate the efficacy of anetumab ravtansine as second-line therapy for mesothelin-overexpressing malignant pleural mesothelioma.
Another ongoing phase Ib multi-indication trial (NCT03102320) is investigating objective response to anetumab ravtansine (plus gemcitabine in patients with pancreatic adenocarcinoma) and maximum tolerated dose in patients with advanced solid tumors (non—small cell lung cancer, triple-negative breast cancer, gastric or gastroesophageal junction adenocarcinoma, thymic carcinoma, or pancreatic adenocarcinoma).34 To investigate the relationship between mesothelin expression and response to anetumab ravtansine, histologically confirmed mesothelin expression is a requirement for enrollment, and patients will be stratified into a high-mesothelin-expression group (cohort A; ≥30% positive tumor cells with 2+ or 3+ staining intensity) or a low- to mid-mesothelin-expression group (cohort B; ≥5% positive cells at all intensities and <30% positive tumor cells with 2+ or 3+ membrane staining intensity).
Chimeric Antigen Receptor T-cell Therapy
CAR T-cell therapy, which involves genetic alteration of a patient’s autologous T cells to express a CAR specific to a tumor antigen, has shown promise in hematologic cancers, but success has been limited in solid tumors. However, earlystage clinical trials of mesothelin-targeted CAR T-cell therapy are investigating ways to increase efficacy through reduction of immunogenicity, combination approaches with checkpoint inhibitors, and novel designs to increase penetrance into the tumor.
Investigators at MSK engineered mesothelin- targeted CAR T cells (IcasM28z) including completely human components to reduce risk of immunogenicity and an iCaspase-9 safety “suicide” switch that can be activated to kill all the CAR T cells if the patient experiences an unexpected toxicity. “The novelty of our study is that the CAR T cells target the cancer cell-surface protein, mesothelin, which is expressed on the majority of cancer cells, and they are delivered directly to the tumor site using regional delivery techniques,” said Prasad S. Adusumilli, MD, deputy chief of thoracic service at MSK and coauthor of the study, in a press release.35 “If this approach is successful, 2 million patients with mesothelin-expressing solid tumors per year in the United States will be eligible for this treatment.”
In the phase I trial (NCT02414269), the investigators used cyclophosphamide for preconditioning, followed by an interventional radiology procedure to inject the IcasM28z CAR T cells into the pleural cavity of 21 patients with malignant pleural disease (19 with mesothelioma, 1 with metastatic lung cancer, and 1 with metastatic breast cancer). The CAR T cells persisted in the peripheral blood in 13 patients during the 38-week evaluation, and this presence was associated with tumor regression on imaging studies and >50% decrease in blood levels of a mesothelin-related peptide.35 In addition, 14 patients received PD-1 checkpoint inhibitors (up to 21 cycles) off protocol as the next line of therapy, and positron emission tomography (PET) scans showed that 2 patients had complete metabolic responses, 5 had partial responses, and 4 had stable disease.
Although the authors cautioned that the phase I trial did not allow for evaluation of long-term efficacy, they stated that combining strategies, such as interventional radiology, T-cell genetic engineering, and immunotherapy agents, should be investigated further in mesothelioma and other aggressive treatment-resistant cancers that have few therapeutic options.
Pancreatic tumors have several features, including desmoplastic reaction, poor vascularization, large infiltration of myeloid cells, and scarcity of T cells in the tumor microenvironment, that have made treatment—including CAR T-cell therapy—especially difficult, according to Gregory L. Beatty, MD, PhD, director of translational research at the Pancreatic Cancer Research Center at the University of Pennsylvania (Penn) in Philadelphia.36
Beatty and his colleagues at Penn developed a mesothelin-specific CAR that incorporates the mouse monoclonal antibody SS1, the costimulatory CD137 domain to enhance proliferation and survival, and a CD3ζ signaling domain to induce T-cell activation.36 They engineered the CAR T cells with an RNA platform and used electroporation to transiently express the CAR in T cells, which is thought to yield better safety than permanent expression of the CAR, such as through a lentiviral platform and transduction. Results of a phase I study showed that 2 of 6 patients with chemotherapy-refractory metastatic pancreatic ductal adenocarcinoma had stable disease following infusions of the RNA CAR T cells (3 times per week for 3 weeks), and these patients had progression-free survivals of 3.8 months and 5.4 months.37 The 18F-2-fluoro- 2-deoxy-D-glucose (FDG)—PET/computed tomography (CT) imaging showed that the metabolically active volume remained stable for 1 month in 3 patients and decreased by 69.2% at 1 month in 1 patient who had biopsy-proven mesothelin expression and significant liver tumor burden at baseline.36 In this patient, all liver lesions had a complete reduction in FDG uptake at 1 month, but the primary lesion was unchanged. According to Beatty, the collective results of this study suggested that the tumor response to CAR T-cell therapy varies among anatomic locations because of the difference in immune escape mechanisms.36
Preliminary exploration into potential mechanisms of action and resistance showed that although the CAR T cells trafficked to the tumor in some patients, the levels of CAR T cells within the tumor were generally low and that the T-cell influx within the tumor microenvironment was highly heterogeneous. Additionally, patterns of T-cell infiltration varied: In some cases, T cells were not present, whereas in other cases, the T cells were present but did not engage tumor cells or engaged but did not kill tumor cells.
Beatty also noted that although most pancreatic tumor specimens express mesothelin, many patients exhibited intracellular expression, which does not work for mesothelin- targeted CAR T-cell therapy that targets mesothelin on the cell surface. He noted that the patient who had an FDG-PET/CT response had expression of mesothelin on the surface of the tumor cell, and tumor mesothelin expression is an inclusion criterion for enrollment in the ongoing clinical trial of CAR T-cell therapy initiated by Penn (NCT03323944).
Overall, results of phase I and II trials have shown the potential of mesothelin as an important target for immunotherapy in multiple types of mesothelin-expressing cancer. To optimize patient outcomes from a mesothelin-targeted approach, further research is needed to validate the best methods for assessment of mesothelin as a biomarker for inclusion in future clinical trials and predictor of response, as well as rationally designed combination approaches with chemotherapy, targeted therapy, or other immunotherapies.