Abstract
Diffuse large B-cell lymphoma (DLBCL) is the most common form of lymphoma type. DLBCL shows considerable clinical and biological heterogeneity. The international prognostic index (IPI) remains the most used tool to stratify patients in different risk groups but does not reflect DLBCL biological heterogeneity. Therefore, much research is currently focused on the identification of new prognostic markers for more specific patients’ risk stratification and on the development of therapeutic approaches to improve outcome. Epigenetic alterations are involved in lymphoma. Interestingly, epigenetic alterations are reversible and drugs to target some of them have been developed. With the aim to identify new and relevant prognostic factors that allow the stratification of patients with DLBCL we investigated the gene expression profile of 130 epigenetics regulators in two independent cohorts of patients with newly-diagnosed DLBCL homogeneously treated (respectively 233 and 181 cases). Using the Maxstat R function and Benjami-Hochberg multiple testing correction we found that 10 probe sets had a prognostic value for overall survival (OS) including: BRD1, CARM1, BRPF3, CDYL, DNMT3A, DOT1L, HDAC2, PRMT5, SETD8 and SP140. Using multivariate Cox analysis we found that 3 of these genes remained independent prognostic factors: DNMT3A, DOT1L and SETD8. We used these 3 genes to develop a risk score (EpiScore) based on their expression level in the cohort of 233 DLBCL. EpiScore allowed splitting the patients in 3 groups with significant different OS values: group 1 (low risk, low DNMT3A, DOT1L and SETD8 expression), group 2 (intermediate risk, high expression of one of the three genes) and group 3 (high risk, high expression of two or all three genes). EpiScore prognostic value was validated in two other independent cohorts of patients with DLBCL (181 and 69 patients respectively) We then showed that EpiScore was an independent predictor of survival when compared with previously described prognostic factors, such as the IPI, germinal center B cell and activated B cell molecular subgroups, gene expression-based risk score (GERS) and DNA repair score. As gene expression profiling (GEP) is not a technical approach widely performed in routine practise for all DLBCL newly diagnosed we analysed the pattern of expression of the ten epigenetic genes by immunohistochemistry on formalin-fixed paraffin-embedded (FFPE) tissue sections in a cohort of 65 patients with de novo previously untreated DLBCL. Our results indicate that these epigenetic related proteins are commonly overexpressed in DLBCL compared to reactive lymphoid tissues and may be important for DLBCL pathogenesis. We showed that overexpression of CARM1 and DNMT3A was significantly associated with reduced event free survival. We then designed a new risk score Epi-ImmunoScore (Epi-IS) based on the expression level of CARM1 and DNMT3A by immunohistochemistry. Epi-IS was predictive of OS in DLBCL and allowed splitting patients in two groups (high and low risk). Finally, using gene set enrichment analysis (GSEA) an HDAC gene signature was significantly enriched in the DLBCL samples included in the EpiScore high-risk group and that a significant enrichment of genes encoding for HDAC class II, multiple drug resistance and NOTCH pathways in DLBCL samples with DNMT3A overexpression. According to these data, we compared the response to HDAC inhibitor (SAHA) in DLBCL cell lines with high EpiScore versus low EpiScore and showed that DLBCL cell lines with high EpiScore were significantly more sensitive to SAHA than those low EpiScore. We concluded that EpiScore and Epi-IS, easily evaluated in the routine practice on FFPE tissue sections identified high-risk patients with DLBC. We have also recognized relevant therapeutic targets and identified a number of candidate drugs with potential therapeutic efficiency in DLBCL patients. All these findings may orient future preclinical intervention strategies in DLBCL.