Enzyme-mediated depletion of methylthioadenosine restores T cell function in MTAP-deficient tumors and reverses immunotherapy resistance

Abstract

Chromosomal region 9p21 containing tumor suppressors CDKN2A/B and methylthioadenosine phosphorylase (MTAP) is one of the most frequent genetic deletions in cancer. 9p21 loss is correlated with reduced tumor-infiltrating lymphocytes (TILs) and resistance to immune checkpoint inhibitor (ICI) therapy. Previously thought to be caused by CDKN2A/B loss, we now show that it is loss of MTAP that leads to poor outcomes on ICI therapy and reduced TIL density. MTAP loss causes accumulation of methylthioadenosine (MTA) both intracellularly and extracellularly and profoundly impairs T cell function via the inhibition of protein arginine methyltransferase 5 (PRMT5) and by adenosine receptor agonism. Administration of MTA-depleting enzymes reverses this immunosuppressive effect, increasing TILs and drastically impairing tumor growth and importantly, synergizes well with ICI therapy. As several studies have shown ICI resistance in 9p21/MTAP null/low patients, we propose that MTA degrading therapeutics may have substantial therapeutic benefit in these patients by enhancing ICI effectiveness.

Figure 1.. Impact of CDKN2A GA/MTAP GA, CDKN2A GA/MTAP WT, or CDKN2A WT/MTAP WT on overall survival (OS) and time-to-treatment failure (TTF) in ICI-treated patients with melanoma and urothelial carcinoma.

A), Overall survival in DFCI melanoma cohort. B) Time-to-failure survival in DFCI melanoma cohort. C) Overall survival in DFCI urothelial carcinoma cohort. D) Time-to-failure in DFCI urothelial carcinoma cohort. E) Impact of MTAP GA on tumor-infiltrating lymphocyte density of DFCI melanoma cohort with available H&E images and genomic data. Statistical analyses: reported p values are from multivariable Cox regression analysis (OS and TTF, adjusted for prior lines of therapy, single vs. combination ICI, tumor mutational burden and ECOG PS), and Wilcoxon rank sum (TILs). WT, wild-type; GA, genomic alteration. Also see Figure S1, Table S1.

Figure 2.. Impact of PEG-MTAP or PEG-MTAN administration on L1210 tumor MTA levels, growth, immune phenotype and sensitization to ICI therapies.

A) Relative levels of MTA in L1210 tumors tissue of mice treated with one dose of PEG-MTAP as compared to vehicle. B) DBA/2 mice inoculated with L1210 cells treated with 50 mg/kg PEG-MTAP or PEG-MTAN exhibit marked tumor growth delay (p.t. EOD for 6 total doses) (n=7 for control, n=9 for MTAP, n=9 for MTAN). C) L1210 tumors inoculated in nude mice exhibit no response to treatment with 50 mg/kg PEG-MTAP (p.t. EOD for 6 total doses) (n=6 for each group). D) L1210 tumors treated with PEG-MTAP (p.t. EOD for 4 total doses) exhibit increased numbers of CD8⁺, CD4⁺ T cells and elevated markers of proliferation (n=13 for control, n=15 for MTAP). E) L1210 tumor cells treated with PEG-MTAP enzyme in combination with anti-CTLA4 or anti-OX40 antibodies (n=7 for control, n=8 for MTAP, n=7 for anti-CTLA4, n=8 for MTAP + anti-CTLA4, n=6 for anti-OX40, and n=7 for MTAP + anti-OX40) exhibit synergistic and additive efficacy respectively in controlling tumor growth. Statistical analyses: two-sided T-test. Data represent mean ± SD (a& d) and mean ± SEM (B, C, & E). Also see Figures S2–3.

Figure 3.. PEG-MTAP impact on B16 Mtap^−/− tumor growth requires CD8⁺ T cells and increases tumor infiltrating CD8⁺ and CD4⁺ T cell numbers:

the Mtap deletion elicits a deficit in PRMT5 mediated methylation and translational changes in cytokine production and antigen display as compared to the parental B16 cell line. A) The growth of B16-F10-Mtap^−/− tumors are inhibited after treatment with PEG-MTAP (n = 14 each, TIW × 2 weeks, p <0.05 at control endpoint). B, C & D) PEG-MTAP control of B16-Mtap^−/− tumor growth is reversed by deletion of CD8⁺ T cells and augmented by deletion of CD4⁺ T cells. E) B16-F10-MTAP^−/− tumors treated with PEG-MTAP (48 hrs post single dose) exhibit increased numbers of infiltrating CD8⁺, CD8⁺/Granzyme B⁺ and CD4⁺ T cells compared to vehicle treated tumors with no changes observed in B cells (n =11 for control, n=10 for MTAP). F) Deletion of Mtap from the B16-F10 cell line causes a marked loss of protein SDMA, G) modulates the levels of PRMT5 adaptor proteins and substrates and H) alters the levels of proteins associated with cytokine production/immune response and antigen processing/presentation. I) PEG-MTAP administration controls the growth of Cdkn2a/Mtap co-deleted MB49 urothelial carcinomas and increases survival. Statistical analyses: two-sided T-test, Z-score, Kaplan Meier method (*p< 0.05,**p<.01, ***p<.001). Data represent raw (a-d), mean ± SD (e, and i, middle), fold change (g, h) mean ± SEM (I, left). Also see Figure S3.

Figure 4.. MTA inhibits the proliferation and translational profile of murine CD8⁺ and CD4⁺ T cells by additional mechanisms beyond adenosine receptor signaling.

A) FACS analysis of dye dilution proliferation assays (CellTrace^™ Violet) of murine CD8⁺ and CD4⁺ T cells incubated with MTA demonstrates marked inhibition of proliferation that is reversed by co-incubation with MTAP or MTAN enzymes (quantification in Figure S4H). B) The proliferation of murine CD8⁺ and CD4⁺ T cells incubated with MTA is only slightly restored with co-incubation of adenosine receptor antagonists that block A2aR and A2bR signaling (istradefylline, GS 6201, or the combination) but fully restored with co-incubation with MTAN (quantification in Figure S4I). C) Protein expression of murine T cells incubated with MTA or adenosine (compared to untreated cells) demonstrates significant overlap of proteins involved in TCR activation and cytokine signaling. D) MTA but not adenosine increases the levels of methyltransferases and decreases the levels of demethylases and E) MTA modestly inhibits asymmetric dimethylarginine protein methylation and strongly inhibits symmetric dimethylarginine protein methylation in murine T cells. F) MTA but not adenosine strongly increases the levels of pro-apoptotic and Tp53 related proteins in murine T cells. G) Comparison by RNA SEQ of differentially expressed genes in murine T cells incubated with MTA or an PRMT5 inhibitor (LLY-283) demonstrate a strong linear correlation expression (up or down, r2 = 0.94) and both MTA and LLY-283 elicit nearly identical changes in genes related to T cell activation and proliferation H), cell death I) and in methyltransferases/demethylases J). Statistical analyses: Z- score, Pearson’s correlation. Data represent fold change (c-j). Also see Figure S4.