Characterization of enhancer of zeste homolog 2 (EZH2) expression, activity, and association with the tumor immune microenvironment in olfactory neuroblastoma (ONB).

Abstract

e14655 Background: ONB is a rare sinonasal cancer with limited therapeutic options. Increased expression and up-regulation of EZH2, a core protein of the epigenetic modifier methyltransferase Polycomb repressive complex 2, have been associated with worse prognosis in several cancer types. EZH2 inhibitors are currently approved in Epithelioid Sarcoma (ES) and Follicular Lymphoma. We examined the expression and activity of EZH2 in ONB, as compared to ES (where EZH2 inhibition is standard-of-care), and its impact on the immune microenvironment. Methods: ONB (N = 35) and ES (N = 78) were tested at Caris Life Sciences (Phoenix, AZ) with NextGen Sequencing of DNA (592 genes or whole exome) and RNA (whole transcriptome). PD-L1 expression (22C3; Positive (+): TPS ≥1%) was assessed by IHC. Transcriptomic signatures predictive of immunotherapy response (T-cell inflamed [TCI] score) and MAPK pathway activation (MPAS) were applied. Immune cell fractions were estimated using quanTIseq. A transcriptional ‘ EZH2 Repression Score' (ERS, arbitrary units: AU) was examined, with high ERS suggestive of lower EZH2 activity (Morel et al, {2021}). ONB were stratified by their median ERS (high -H, N=17 v low -L, N=18). Three additional score variants were applied (equally weighted, immune regulatory genes excluded, or both) and all produced similar results. ONB were stratified by the median EZH2 transcripts per million (TPM, -H v -L). Mann-Whitney U, Fisher’s Exact/χ2 tests were applied as appropriate with p-values adjusted for multiple comparisons ( p <.05). Results: No pathogenic EZH2 mutations were observed in ONB; EZH2 expression was higher compared to ES (44.4 v 14.1 TPM, p < 0.001), yet ERS was not significantly different (5111 v 5143 AU, p = 0.76). There was no association between EZH2 expression and ERS (Pearson’s correlation = .1). In ONB, there was no difference in the prevalence of PD-L1+ between ERS-H v -L (20% v 20%). ERS-H were associated with increased expression of β2M (2110 v 1019 TPM, p = .01) and immune checkpoint genes: CTLA4 (1.4 v 0.5, p = .03), PD-1 (1.1 v 0.3, p = .04), PD-L1 (4.7 v 1.6, p < .001) and VISTA (22.1 v 6.5, p < .001). ERS-H was associated with increased M2 macrophage infiltrate (5.0 v 2.5%, p < .001). ERS-H had increased MPAS (1.99 v -.42 AU, p < .001) and increased prevalence TCI tumors (29 v 0%, p < .001), while EZH2-H tumors were not associated with differences in immune infiltrate, TCI, MPAS, nor with immune checkpoint gene expression compared to EZH2-L. Conclusions: EZH2 mutations were not detected in ONB or ES, though similar EZH2 activity (by ERS) was observed in both cohorts, thus making ONB an intriguing candidate for EZH2 modulating therapies. High ERS in ONB was associated with expression of immune checkpoint genes suggesting that inhibiting EZH2 activity could enhance the susceptibility of ONB to immune checkpoint blockade.