Proteomic and Genomic Profile of High-Risk MDS After Treatment with 5-Azacytidine,

Conventional cytogenetics is the major prognostic factor in predicting outcome of myelodysplasia (MDS), but at least 50% of patients have a normal karyotype. 5-azacytidine (5-AZA) is a DNA methyltransferase inhibitor used in treatment of high-risk MDS, which is able to delay the progression to AML and improve clinical outcomes, but targets of the methylation status are poor known.

To evaluate molecular changes, at genomic or proteomic level, which can be identified as additional targets of 5-azacytidine in MDS patients with normal conventional cytogenetics.

By reverse-phase protein microarray (RPMA), we analyzed bone marrow mononuclear cells from 19 patients affected of high risk MDS, treated with 5-azacytidine (median age 71 years, M/F=12/5). Treatment consisted of 4 cycles of 100mg flat dose for 7 days+21 days of wash-out. In 7 cases the sample after 4 cycles of therapy was matched to the sample collected before starting treatment. RPMA was used to quantitatively map 45 cell signaling pathway endpoints, including survival, proliferation, drug resistance, apoptosis, and autophagy. For the first 4 patients we used also a genome-wide approach based on SNP (single nucleotide polymorphism) array (6.0 Affymetrix platform) to detect copy number and allelotype data in order to identify new genetic markers in terms of SNP, CNV and LOH.

All patients were evaluable for response one month after the 4^th cycle. Three patients were refractory and progressed to AML and 1 was a late responder (after 7 cycles). All other patients experienced hematologic improvement. After 4 cycles of 5-AZA, at proteomic level we found three main signaling pathways were increased and did not correlate with the response: 1) pro-survival signaling: PLC-y-1-Tyr783 (p=0.0017), and its upstream regulators Src-Tyr416, c-Abl-Tyr735 (p=0.002) and downstream target STAT5Tyr694 (p=0.0017) were increased, without affecting proliferative pathways, such as AKT activation status on Ser473 and Thr308 or mTORSer2448. 2) Autophagy: ATG5, Beclin 1 and LC3B were significantly elevated after treatment (p values respectively <0.0001, 0.0056 and 0.0124). Activation of the autophagy pathway occurred downstream of mTOR, since mTORSer2448, AktSer473, AktThr308, ERKThr202Tyr204 (and in general proliferation markers) were not affected. 3) Restore of cell cycle G1-S checkpoint: c-Abl-Tyr735, Rb-Ser608, Rb-Ser780Chk1-Ser345, p53, p53Ser15, and in particular HDAC1, HDAC3 were significantly elevated (p<0.001).

At genomic level, we identified 46 losses and 25 gains. In three cases we found a deletion on chromosome 20q13.12 (from 46125 kb to 46178 kb) at level of NCOA3 gene, a transcriptional coactivator protein that contains several nuclear receptor interacting domains and an intrinsic histone acetyltransferase activity. In 2 cases we found a loss in the region 22q13 (from 40871 kb to 40910 kb), at level of MKL1 gene, encoding a nuclear protein involved in transduction signals from the cytoskeleton to the nucleus. MKL1 is involved in a specific translocation event associated with acute megakaryocytic leukemia. Moreover in one case we found a deletion on chromosome 4q24 (from 105800 kb to 107200 kb) in a region involving TET2 gene. Loss of heterozygosity 4q24 and TET2 mutations have been implicated in the pathogenesis of myelodysplastic syndromes. Another patient showed autozigosity regions with an extensive CN-LOH on chromosomes 7 and 11, involving the centromers and recently their neoplastic potential in colon cancer have been reported and never, before, observed in MDS.

More informative findings can be detected using a proteomic approach in combination with the genomic one. Autophagy activation can be considered an escape pathway promoting survival in neoplastic cells and the observation that 5-AZA does not affect proliferative pathways suggests the possibility to combine it with anti-proliferative agents. The increase of HDAC1 and HDAC3 can provide the molecular rationale for the development of a combination of 5-azacytidine with HDAC inhibitors.

No relevant conflicts of interest to declare.