Epigenetic Profiling in High-Risk MDS and AML Patients Identifies Pre-Treatment Methylation Patterns That Indicate Response to Hypomethylating Agents: A Pilot Study

Abstract

Background:Azacitidine is a nucleoside analog that functions as a hypomethylating agent (HMA), approved for use in myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML). However, the complete response (CR) rate to HMA therapy is generally < 20%, indicating the need to more accurately identify patients most likely to benefit from therapy. The primary objective of this pilot study was to assess the efficacy of epigenetic DNA methylation markers in screening pre-treatment bone marrow (BM) myeloblasts to differentiate azacitidine responders from non-responders.Methods:We selected 10 patients from Moffitt Cancer Center Total Cancer Care Database and repository who had BM samples collected within 3 weeks prior to treatment with azacitidine, including 5 responders and 5 non-responders. Responders were defined as patients who received at least 4 cycles of therapy with azacitidine and achieved CR as defined by International Working Group response criteria for AML and MDS whereas non-responders were defined as patients with inferior outcome to complete or partial remission. Institutional Review Board approval was obtained prior to initiation of the study. The cohorts consisted of 5 responders and 5 non-responders. Genomic DNA was extracted from the BM samples of both groups and purified for NGS whole genome sequencing and DNA fragment libraries were created. Computational processing of sequence reads was used to quantify CpG site-specific methylation levels. Methylation scores, single nucleotide polymorphisms and short insertions/deletions from the same NGS sequence read files were compiled. Comparisons and analyses of methylation profile differences between drug response groups were conducted at the level of individual CpG sites, genes, and pathways. Clustering analysis with bootstrap iterations was performed to identify statistically significant methylation patterns between the patient groups.Results:Responders included samples from 2 patients with MDS-refractory anemia with excess blasts-2 (RAEB-2), 2 patients with secondary AML, and 1 patient with de novo AML. The median age at diagnosis was 68 years (56-82) with 2 males/3 females. Median BM myeloblast percentage (BM % blasts) was 19 (16-28). In the non-responder group, there were 4 males and 1 female with a median age of 71 (60-95), and median BM % blasts was 17 (12-43). Three patients had MDS-RAEB-2 and 2 had secondary AML. (3 patients had disease progression and 2 patients had stable disease). Duration of response varied in responder group from 5 to 19 months from the date of CR (median 10 months). Differential methylation activity calculated between non-responders and responders across CpG sites within genes discovered the 40 most hypermethylated genes in non-responders and the 40 most hypermethylated genes in responders as depicted in Figure 1. A total of 1.01 million "CCGG" total sites were analyzed, from which 88% were scored in each sample. Using an ordinate analysis technique of non-metric multidimensional scaling (NMDS), CpG methylation profiles were compared among individuals to isolate patterns conserved within groups while also differing between groups. We discovered that the separation between drug response groups was statistically significant with the 95% confidence intervals around the samples in each group being highly separated (p < 0.0004) (Figure 2). Clustering analysis with bootstrap iterations and dendrogram heatmap were created which identified 37 statistically significant distinct methylated CpG sites in the responder cohort (p < 0.0025) (Figure 3). Assessing methylation loads across genes, and grouping genes within defined pathways (Kyoto Encyclopedia of Genes and Genomes), revealed substantial hypermethylation in responders, primarily in signaling pathways mediated by ErbB, TGFβ, and estrogen. Comparison of the methylation signatures of individual genes revealed differences between responders and non-responders in ADGRB1, C3, CD5, CHCHD2, ETV6, HTT, MIR8073, and WNT7B.Conclusions:These findings indicate that a targeted assay of CpG methylation sites can provide a mechanistically-based screening tool to discern potential for response azacitidine (or resistance), which warrants confirmation in a larger confirmatory cohort. Such a clinical screening test to guide the selection of therapy could avoid exposure to a therapy with expected inferior outcomes. [Display omitted] [Display omitted] [Display omitted] DisclosuresMarsh:Genome Profiling, LLC: Employment. Sweet:Pfizer: Speakers Bureau; Novartis: Consultancy, Speakers Bureau; Karyopharm: Honoraria, Research Funding; Incyte Corporation: Research Funding; Ariad: Consultancy, Speakers Bureau. Komrokji:Incyte: Consultancy; Novartis: Consultancy, Speakers Bureau; Celgene: Membership on an entity's Board of Directors or advisory committees, Research Funding; Boehringer-Ingelheim: Research Funding.