Unmasking the suppressed immunopeptidome of EZH2-mutated diffuse large B-cell lymphomas through combination drug treatment

Abstract

Exploring the repertoire of peptides presented on major histocompatibility complexes (MHCs) helps identify targets for immunotherapy in many hematologic malignancies. However, there is a paucity of such data for diffuse large B-cell lymphomas (DLBCLs), which might be explained by the profound downregulation of MHC expression in many DLBCLs, and in particular in the enhancer of zeste homolog 2 (EZH2)-mutated subgroup. Epigenetic drug treatment, especially in the context of interferon-γ (IFN-γ), restored MHC expression in DLBCL. In DLBCL, peptides presented on MHCs were identified via mass spectrometry after treatment with tazemetostat or decitabine alone or in combination with IFN-γ. Such treatment synergistically increased the expression of MHC class I surface proteins up to 50-fold and the expression of class II surface proteins up to threefold. Peptides presented on MHCs increased to a similar extent for both class I and class II MHCs. Overall, these treatments restored the diversity of the immunopeptidome to levels described in healthy B cells for 2 of 3 cell lines and allowed the systematic search for new targets for immunotherapy. Consequently, we identified multiple MHC ligands from the regulator of G protein signaling 13 (RGS13) and E2F transcription factor 8 (E2F8) on different MHC alleles, none of which have been described in healthy tissues and therefore represent tumor-specific MHC ligands that are unmasked only after drug treatment. Overall, our results show that EZH2 inhibition in combination with decitabine and IFN-γ can expand the repertoire of MHC ligands presented on DLBCLs by revealing suppressed epitopes, thus allowing the systematic analysis and identification of new potential immunotherapy targets.

Graphical abstract

Figure 1.

Decitabine (DAC) and tazemetostat (TAZ) upregulated HLA protein in DLBCL cell lines. (A-D) Cells were treated with indicated 125 nM decitabine or 1 μM tazemetostat. DB and SUDHL-6 cells were assayed for HLA-A, -B, and -C or HLA-A-02 expression by flow cytometry. (E-H) Cells were treated as in panels A-D along with 100 ng/mL IFN-γ. Analysis of variance (ANOVA) was performed using either untreated (panels A-D) or IFN-γ alone (panels E-H) as control, followed by a post hoc Tukey’s test for individual experimental groups. Mean ± standard deviation (SD) is shown for 3 technical replicates per 2 biological replicates. ns, not significant. *P < .05; **P < .01; ***P < .001; ****P < .0001.

Figure 2.

Decitabine and tazemetostat activated transcription of HLA alleles on DLBCL cell lines. (A-G) SUDHL-6, DB, and SUDHL-4 cells were treated with indicated drugs (decitabine, 100 nM; tazemetostat, 1 μM; IFN-γ, 10 ng/mL). Graph of fold-change in transcript to untreated for each indicated gene (panels A, D, F) HLA-A, (panels B, E, G) HLA-B, and (panel C) HLA-C. ANOVA was performed using IFN-γ alone as control, followed by a post hoc Tukey’s test for individual experimental groups. Mean ± SD is shown for 3 technical replicates per 2 biological replicates. *P < .05; **P < .01; ****P < .0001.

Figure 3.

Decitabine and tazemetostat activated transcription of antigen-presentation genes in DLBCL cell lines. (A-F) SUDHL-6 and DB cells were treated with indicated drugs (decitabine, 100 nM; tazemetostat, 1 μM; IFN-γ, 10 ng/mL). Graph of fold-change in transcript to untreated for each indicated gene (panels A, D) B2M, (panels B, E) Tap1, (panels C, F) Tap2. Mean ± SD is shown for 3 technical replicates per 2 biological replicates. *P < .05; **P < .01; ****P < .0001.

Figure 4.

Decitabine upregulated HLA class II molecules in DLBCL cell lines. (A-C) Cells were treated with 125 nM decitabine or 1 μM tazemetostat along with 100 ng/mL IFN-γ for SUDHL-6, SUDHL-10, and WSU-DLCL-2 cell lines and were assayed for expression of HLA-DR/DQ. Mean ± SD is shown for 3 technical replicates per 2 biological replicates. (D-E) Serial dilutions of IFN-γ were performed in the presence or absence of decitabine and tazemetostat for SUDHL-6 and SUDHL-10 cell lines. Mean ± SD is shown for 3 individual replicates. *P < .05; **P < .01; ***P < .001; ****P < .0001.

Figure 5.

Epigenetic drug treatment in the presence of IFN-γ unmasked the immunopeptidome of DLBCL cell lines. (A) Cells were treated with 125 nM decitabine, 1 μM tazemetostat or the combination of both in the presence of 100 ng/mL IFN-γ. Fold change of unique identifications of HLA ligands relative to untreated cells is depicted. Error bars indicate mean plus range. Experiments were performed in duplicates. (B) Relative distribution of HLA alleles after assignment to their respective alleles through NetMHCpan 4.0. (C) Overlap analysis of all peptides by cell line and respective treatment condition. (D) Overlap of source proteins for HLA ligands shared between the SUDHL-4, DB, and SUDHL-6 cell lines (top). Overlap of 151 source proteins from overlap at the top were matched with 13 428 source proteins of the HLA class I ligandome from healthy donors as published by Marcu et al.