Inhibitors of the programmed cell death protein (PD)‑1, and a number of other immune checkpoint inhibitors have shown efficacy as anti-cancer treatments. PD‑1 inhibitors have proved particularly effective for the treatment of non-small cell lung cancer (NSCLC), and several inhibitors, such as nivolumab and pembrolizumab, have been approved for treatment. Despite this, only a third of patients display clinical benefits and strategies to potentiate the response rates are under investigation.
Biomarkers that predict the therapeutic response will aid in the discovery of novel therapeutics to increase the response to PD‑1 inhibitors. Currently, expression of PD‑L1 is the only predictive biomarker of PD‑1 inhibitor efficacy approved by the FDA. Transcriptomic and mutation analysis of immune checkpoint inhibitor responses have been widely characterised, however epigenetic changes in methylation have not been thoroughly investigated. Epigenetic modulation by tumour cells has been demonstrated as one of a number of mechanisms of immune evasion and understanding the changes in the methylome in response to immune checkpoint inhibitors may be beneficial for improving immunotherapies.
In this study the authors sought to characterise demethylated regions (DMRs) in promoters (pDMRs) and enhancers (eDMRs) in anti‑PD‑1 responders and non-responders from NSCLC patients. They use the Infinium methylation arrays correlated with RNA expression and patient survival.
Methylation profiles from NSCLC patients before treatment with nivolumab or pembrolizumab were determined using the Infinium Methylation EPIC arrays. From the 18 patients, six were responders and 12 were non-responders. After data processing and filtering, 1437 DMRs were detected, 1007 being pDMRs and 607 eDMRs (some promoters could display enhancer activity).
1109 genes were identified immediately downstream of the 1007 pDMRs and 1173 genes that interacted with the 607 eDMRs with only 217 genes overlapping between the 2 DMR types.
KEGG pathway analysis of differentially methylated gene targets demonstrated that immune-related, oncogenic and metabolic pathways were enriched. Many of the pathways identified such as mTOR/AMPK, Hif‑1, IGF and longevity pathways are all consistent with pathways involved in cancer immunomodulation.
KEGG pathway analysis on the different types demonstrated that 45 KEGG pathways were enriched in the eDMR targets. Out of the 30 detected for all DMRs 27 of these pathways were enriched in the eDMR target genes. Only 5 pathways detected in KEGG analysis for pDMRs overlapped with all DMRs suggesting the eDMRs have a greater contribution to the epigenetics of responders and non reponders.
These results were confirmed in an independent cohort of NSCLC patients.
Antigen processing and presentation had the greatest enrichment of target genes in eDMRs and class II MHC molecules (e.g. HLA‑DM, HLA‑DO, HLA‑DP, HLA‑DQ, and HLA‑DR), which facilitate T-Cell recognition of cancer cells and had higher expression in responders compared to non-responders. eDMRs targeting HLA genes overlapped with super-enhancers identified from lung tissue, implicating methylation at super-enhancers contribute to class II MHC expression and response to PD‑1 inhibitors.
Transcriptomic analysis, by xCell, identified greater infiltration of CD4+ effector memory T-cells and CD8+ T-cells in responders compared to non-responders.
Biomarkers that may predict response were identified by correlating the methylation status with the expression data. pDMRs for CYTIP and TNFSF8 were selected for pyrosequencing analysis of a validation cohort of 56 FFPE NSCLC patients.
Methylation status of pDMRs for CYTIP and TNFSF8 in the validation cohort differed between responders and non-responders and demonstrated a significant association to inhibitor response.
PD‑L1 expression, the current biomarker, was also associated with response in the validation cohort, however, the predictive power of CYTIP or TNFSF8 pDMRs was greater.
Significantly longer progression free survival (PFS) and overall survival (OS) was demonstrated for patients with hypomethylation of CYTIP or TNFSF8 pDMRs, whilst PD‑L1 expression did not correlate with PFS or OS. Higher significance for both PFS and OS was observed when pDMRs for both were analysed.
Immune checkpoint inhibitors have shown great promise for the treatment of some cancers, however only a third of patients achieve any clinical benefit. Combination treatments and novel therapeutics could improve response, but investigations are hampered by the limitation of biomarkers for response with PD‑L1 – currently the only biomarker available. In this study the authors demonstrated differential methylation of promoters and enhancers between responders and non-responders prior to treatment with PD‑1 inhibitors. They also identified two target genes CYTIP and TNFSF8 that could predict response with greater sensitivity than the current biomarker, PD‑L1 expression.
Cho, J., Hong, M.H., Ha, S. et al. Genome-wide identification of differentially methylated promoters and enhancers associated with response to anti‑PD‑1 therapy in non-small cell lung cancer. Exp Mol Med (2020).