Abstract
Abstract
Abstract 599
Aberrant epigenetic silencing of genes through aberrant promoter hypermethylation, as occurs with tumor suppressor genes (TSG), has been implicated in the pathogenesis of MDS and other myeloid malignancies and has been clinically targeted by hypomethylating agents. To date, most of the studies investigating hypermethylation of TSG in hematologic malignancies targeted empirically selected gene promoters but with the advent of methylation arrays, global analysis of methylation pattern became technically possible. We applied methylation arrays (Illumina ®) allowing for simultaneous analysis of 25K CpG sites, focusing on the WHO defined subentity of MDS/MPD, CMML, because of possible efficacy of hypomethylating agents in this disease and the need to identify diagnostic markers and predictors of response. We hypothesized that by comparing CMML patients to patients with similar monocytoid entities, we would be able to establish an epigenetic signature that was consistent across these diagnoses. We studied patients with CMML (N=26), JMML (N=22) and monocytoid forms primary AMLs (N=16; M4 N=9 and M5 N=7) to controls (N=28). In addition we studied 35 patients with advanced and 37 low risk MDS and 9 with MDS/MPN. We developed an analytic algorithm that included establishment of the methylome of normal marrow to define normal/physiologic methylation status for each of CpG islands. These parameters were used as a reference for analysis of concordantly hypermethylated genes in patients, using methylation status as either a continuous (β, where β is proportional to the percentage of cells with methylated status at the locus) or dichotomized variable (where hypermethylation was defined as a β-value greater than the 97th percentile of controls). Each disease was individually compared with controls in order to establish genes aberrantly hypermethylated within the specific entity and the established methylome of each entity was compared with that of other entities. As expected, comparison of the average methylation level across all genes showed no significant differences between groups. Among all subgroups, there were only 58 genes that were consistently hypermethylated; the majority of genes were uniquely hypermethylated in each of the disease subgroups. When CMML and JMML were examined as exemplary conditions, global methylation analysis demonstrated that there was concordant hypermethylation in 25%, 50% and 75% of CMML patients in 1086, 13 and 0 CpG sites, respectively. In contrast, there was a great deal of concordant methylation in JMML with 3796, 1006, 176 of methylated promoters concordant in 25%, 50% and 75% of patients, respectively. The genes that were the most consistently hypermethylated in each entity were selected for further analysis. In JMML, the most consistently hypermethyated genes included LHX6, CDK10, ITGA2B and RAP1GA1. These genes were differentially hypermethylated in JMML compared to CMML (p<.0001 for each gene). In CMML, examples of the most consistently hypermethylated promoters included RPL36, BCORL, GPR171 and HAPLN1; hypermethylation of GPR171 and HAPLN1 clearly distinguished CMML from JMML (p<.0001, p=.0075, respectively.) GPR171 was hypermethylated in significantly more CMML patients than patients with M4 and M5 (p<.0001). In contrast, HAPLN1 was hypermethylated in more patients with M4 and M5 than with CMML (87% of patients). This finding led us to speculate that methylation of HAPLN1 may be a marker associated with disease progression. In fact, HAPLN1 was hypermethylated in 67% of patients with CMML1, compared with 78% of patients with CMML2.
We also compared the whole epigenome profile of each subentity to each other. We selected genes whose average methylation level in a disease entity was greater than the cutoff of 2 standard deviations above the mean of controls. This resulted in selection of 550 genes in CMML patients, of which 230 were also part of the conserved epigenetic pattern of M4, while only 146 were part of the conserved epigenetic pattern of M5. M4 and M5 showed more similarity with each other, sharing 355 genes within their epigenetic profiles.
In conclusion there are few shared epigenetic changes among the monocytoid/ myelomonocytoid malignancies; however, epigenetic changes in these entities are largely unique to each entity. These data suggest that methylation analysis may be useful to supplement histomorphologic diagnostic criteria in distinguishing between these monocytoid malignancies.
Disclosures:
No relevant conflicts of interest to declare.