Blocking the GITR-GITRL pathway to overcome resistance to therapy in sarcomatoid malignant pleural mesothelioma

Abstract

Malignant pleural mesothelioma (MPM) is an aggressive neoplasm originating from the pleura. Non-epithelioid (biphasic and sarcomatoid) MPM are particularly resistant to therapy. We investigated the role of the GITR-GITRL pathway in mediating the resistance to therapy. We found that GITR and GITRL expressions were higher in the sarcomatoid cell line (CRL5946) than in non-sarcomatoid cell lines (CRL5915 and CRL5820), and that cisplatin and Cs-137 irradiation increased GITR and GITRL expressions on tumor cells. Transcriptome analysis demonstrated that the GITR-GITRL pathway was promoting tumor growth and inhibiting cell apoptosis. Furthermore, GITR+ and GITRL+ cells demonstrated increased spheroid formation in vitro and in vivo. Using patient derived xenografts (PDXs), we demonstrated that anti-GITR neutralizing antibodies attenuated tumor growth in sarcomatoid PDX mice. Tumor immunostaining demonstrated higher levels of GITR and GITRL expressions in non-epithelioid compared to epithelioid tumors. Among 73 patients uniformly treated with accelerated radiation therapy followed by surgery, the intensity of GITR expression after radiation negatively correlated with survival in non-epithelioid MPM patients. In conclusion, the GITR-GITRL pathway is an important mechanism of autocrine proliferation in sarcomatoid mesothelioma, associated with tumor stemness and resistance to therapy. Blocking the GITR-GITRL pathway could be a new therapeutic target for non-epithelioid mesothelioma.

Trial registration: ClinicalTrials.gov NCT00797719.

Conflict of interest statement

M.D.P. received personal fees from Astra-Zeneca, Bayer, and Actelion outside of the submitted work. Other authors declare that they have no competing interests.

© 2021. The Author(s).

Figures

Fig. 1. Characteristics of three human mesothelioma cell lines.

a Top panel: Representative image of phase contrast microscopy shows the difference of cell morphology among the three cell lines. Bottom panel: A total of 6 × 106 of CRL5820, CRL5915, or CRL5946 cells were subcutaneously or intraperitoneally injected into NOD/SCID mice. Representative images of H&E staining of xenografted tumor from mice in each treatment group. b, c Comparison of GITRL and GITR expression among mesothelioma cell lines by using qPCR (b) and immunoblot analysis (c) (n = 8). d Mesothelioma cell lines were treated with 2 μg/mL cisplatin or exposed to 7.5 Gy Cs-137 irradiation and the cell proliferation was analyzed by counting the number of viable cells 4 days after treatment (n = 6). e Representative image of colony formation assay and quantification of visualized colonies in cells 14 days after exposure to 2 μg/mL cisplatin or Cs-137 7.5 Gy irradiation (n = 6). Data were analyzed with non-parametric Kruskal–Wallis test followed by the Mann–Whitney U test with Dunns correction. (*P < 0.05, **P < 0.01, ***P < 0.001).

Fig. 2. Expression of GITRL and GITR in sarcomatoid mesothelioma cell line (CRL5946) increased when treated with cisplatin or Cs-137 irradiation.

CRL5946 cells were exposed to cisplatin 2 μg/ml for 24 h or Cs-137 irradiation 7.5 Gy then collected 4 days later. a qPCR shows that GITRL and GITR mRNA expressions are significantly increased in cisplatin-treated and Cs-137-irradiated cells. b Representative histograms show GITRL and GLTR protein expressions are increased in cisplatin-treated and Cs-137-irradiated cells. c Immunofluorescence staining for GITRL (green), GITR (red), DAPI (blue) in CRL5946 treated with cisplatin or Cs-137 irradiation. DAPI was used as a nuclear marker. *P < 0.05; **P < 0.01; vs nontreated control, determined by using Mann–Whitney U test.

Fig. 3. Microarray transcriptome analysis of GITR and GITRL signaling in CRL5946 cells.

a Flow cytometry analysis of GITRL+ or GITR+ cells sorted by anti-GITRL or anti-GITR antibody. Three subpopulations of CRL5946 cells, GITRL+, GITR+, or GITR−GITRL− cells, were separated by using the magnetic beads isolation kits. The separated cells were stained with anti-GITRL or anti-GITR antibody and analyzed by flow cytometry. The transcriptome profiling in three subpopulations of CRL5946 cells was analyzed by microarray. b Principal component analysis compared the transcriptome profile of GITRL+, GITR+, or to double negative cells. c,d IPA comparison of GITRL+ or GITR+ with double negative cells by uploading the significantly different expression of upregulated and downregulated genes (p < 0.05, FDP < 0.1) revealed predicted up- or downregulated canonical pathways (Z-score > 1, orange or < −1, blue). c and predicted disease and biological function analysis. d (Z-score > 2, red; <−2, cyan). Z-score represents the IPA trends (Z-score > 1, upregulated; Z-score < −1, downregulated). Results are evaluated by the negative log P value greater than 1.3.

Fig. 4. GITR/GITRL signaling contributes to tumorigenicity and cell proliferation in CRL5946 mesothelioma cells.

a Schematic representation of the experimental design and analysis of sphere formation from purified GITRL+ or GITR+ cells from CRL5946 cells in vitro and in vivo. b Representative images of in vitro cells that formed mesospheres (>50 μm) and quantification of spheroid growth from the purified GITRL+, GITRL−, GITR+, or GITR− populations in ultralow attachment plate for 14 days (left panel, n = 30 for each group). The purified cells were intraperitoneally injected into NOD/SCID mice to grow mesospheres. Representative images of spheres were taken on day 30 after cell injection. The size of mesospheres over 70 μm has been quantified to compare the tumorigenicity between GITRL+/GITRL− or GITR+/GITR− (right panel, n = 6 for each group). c Schematic representation of the experimental design by using the GITR neutralizing monoclonal antibody to verify the requirement of GITR/GITRL axis for cell proliferation and tumorigenicity. d Upper panel: The CRL5946 cells were treated with or without 2 μg/mL cisplatin or 7.5 Gy Cs-137 irradiation. Compared with PBS or IgG, cell proliferation was attenuated by GITR neutralized mAb in concentration at 80 μg/mL in nontreated cells (n = 23) or at 20 μg/mL in cisplatin or Cs-137 treated cells (n = 32 for each group). Bottom panel: NOD/SCID mice were injected with nonpretreated, cisplatin-pretreated (2ug/ml for 24 h), or Cs-137 irradiation-pretreated (7.5 Gy, 24 h before injection) cells (n = 8 each treatment group) for 24 days. All mice received intraperitoneal injection of PBS, 100 μg IgG, or anti-GITR on day 0 and day 7. Spheres were decreased in anti-GITR-injected group, whereas PBS- or IgG-injected group did not. All data shown represent the mean ± standard deviation; *P < 0.05; **P < 0.01; ***P < 0.0001 vs PBS or IgG group, determined by one-way ANOVA analysis of variance with Bonferroni post-hoc test.

Fig. 5. Interruption of GITR/GITRL signaling in patient-derived sarcomatoid xenograft inhibits tumor growth.

a Tumours were collected from clinical mesothelioma patients with epithelioid MPM subtype and sarcomatoid MPM subtype and subcutaneously implanted into NOD/SCID mice on right flank region to generate xenografted mice. Mice with xenografted tumours that were the size of 30–70 mm3 received intraperitoneal PBS on day 0 and 7, 5 mg/kg cisplatin on day 0 and 7, or 400 µg anti-GITR mAb on day 0 and 200 µg on day 7. b Representative images of H&E stain for morphologic analysis and immunohistochemistry staining for GITRL and GITR expression. c Tumor size was decreased in sarcomatoid xenograft mice treated with cisplatin (n = 4) or anti-GITR mAb (n = 4) and decreased in epithelioid xenograft mice treated with cisplatin (n = 5), whereas anti-GITR mAb group (n = 4) had no effect on tumor size compared with PBS group (n = 4 of sarcomatoid xenograft, n = 6 of epithelioid xenograft). *P < 0.05; **P < 0.01 vs PBS injected group, determined by Mann–Whitney test.

Fig. 6. GITR/GITRL were highly expressed in nonepithelioid MPM and negatively correlated with survival after the SMART approach.

a Kaplan–Meier survival analysis of all MPM patients demonstrates that patients with epithelioid subtype (Epi.) had better survival than those with nonepithelioid subtype (non-Epi.) with a median survival of 28.6 (Epi.) versus 15.9 months (non-Epi.) (P < 0.0001). b GITR and GITRL levels by immunohistochemistry in epithelioid and nonepithelioid MPM; immunohistochemistry results were quantified by Definiens Tissuestudio 4.0 software. *P < 0.05, determined by Mann–Whitney U test. c Correlation between expression of GITR and GITRL (p = 0.0018, r = 0.3115). d Risk analysis of patients with epithelioid and nonepithelioid subtype who died from mesothelioma after the SMART approach to determine the correlation between GITRL and GITR and length of survival. e Kaplan–Meier survival analysis in nonepithelioid MPM showed that high expression of GITRL and GITR was associated with worse survival after SMART. In epithelioid MPM, high expression of GITRL was associated with better survival after SMART.