FAS Gene Expression Is Epigenetically Regulated and Predicts the Responsiveness to Azacitidine In High-Risk Myelodysplastic Syndromes

Low risk myelodysplastic syndromes (MDS) CD34-positive cells exhibit high level of the death receptor Fas at their surface and abnormal Fas-dependent apoptosis. Fas expression decreases when the disease progresses to acute myeloid leukemia (AML). Based on recent evidence of higher DNA methylation level in high-risk MDS (Figueroa M et al, Blood 2010), we explored the epigenetic regulation of FAS gene during MDS evolution to AML.

We quantified FAS gene expression by RT-qPCR in bone marrow mononuclear cells (BMMNC) from 82 MDS, 37 AML and 17 controls, including 54 patients treated with azacitidine according to the FDA/EMEA schedule. Response was scored according to IWG2006 criteria for MDS and to Cheson et al (JCO, 2003) for AML. We used HL-60 and SW480 cell lines to set up DNA methylation and chromatin immunoprecipitation (ChIP) assays. DNA methylation of 36 CpG dinucleotides in the FAS promoter was explored by bisulfite-treatment of genomic DNA, PCR, cloning and sequencing. Histone modifications of FAS promoter and control genes (B2M, RAG1 and PAX6) were assessed in 4 AML-post MDS patients before and during azacitidine treatment, using acetylated H3 (AcH3) and trimethyl-lysine 9-H3 (H3K9me3) as transcriptionally active and dimethyl-lysine 4-H3 (H3K4me2) and trimethyl-lysine 27-H3 (H3K27me3) as transcriptionally repressive markers.

qRT-PCR identified a significant decrease in FAS mRNA in AML-post MDS compared to RAEB1 (P<0.05). We also noticed a trend to decreased expression of FAS mRNA in AML-post MDS compared to all-stage MDS (P=0.12). Fas protein expression at the surface of BM CD34-positive cells had no impact on overall survival in a cohort of 45 untreated patients. By contrast, a low Fas protein expression (cut-off, RFI<1.8) correlated with a higher response rate in a series of 54 MDS and AML-post MDS patients treated with azacitidine (P=0.043). In these patients, Kaplan-Meier analysis demonstrated that a low Fas protein expression also correlated with longer survival in high-risk MDS and AML-post MDS (Log-Rank test, HR 0.47 [CI95%: 0.23 – 0.96]; P=0.032), and a serial evaluation of Fas protein demonstrated a significant increase in responding patients after 6 cycles (P=0.015).

HL60 leukemic cells express Fas at their surface whereas SW480 colon cancer cells do not. FAS mRNA was higher in HL-60 (nR=2.3±0.4) than in SW-480 (nR=0.5±0.4) cells. In HL-60 cells, only 2 of the 36 studied CpG in FAS promoter were methylated, compared to 20/36 in SW-480, and AcH3 and H3K9me3 active markers were highly enriched, suggesting opened chromatin in HL-60 FAS promoter. In BM CD34-positive cells, we observed a significant increase in the methylation of 2 CpG dinucleotides in 14 AML-post MDS compared to 18 MDS (CpG # 2, P=0.05; CpG # 9, P=0.02). In 5/6 AML-post MDS samples, in vitro treatment with azacitidine induced a 1.7-fold increase of FAS mRNA, and rescued Fas-mediated apoptosis. In 4 AML-post MDS patients, a series of 6 cycles of azacitidine induced a decrease in DNA methylation of the FAS promoter in CD34-positive cells (from 28% [range: 20 – 37] to 14% [range: 11 – 15] methylated CpG) and an enrichment for H3K9me3, not for H3K27me3 marker.

An epigenetic mechanism is responsible for the down-regulation of Fas in AML-post MDS, which is corrected by azacitidine in responding patients. These results suggest that FAS gene reactivation could predict the responsiveness to azacitidine and could be a biomarker of azacitidine efficacy.

No relevant conflicts of interest to declare.