Azacytidine and Decitabine Exhibit Differential Effects on Cytotoxicity and Methylation and Exhibit Class Synergy with Hdaci in Models of Ptcl

Introduction: The PTCL represent a challenging group of lymphoid malignancies with a poor outcome. The PTCL appear to be the prototypical 'epigenetic disease', based on the following observations: (1) it is the one disease for which HDAC inhibitors have been shown to have universal activity; (2) multiple lines of evidence have begun to establish that the PTCL are characterized by recurrent mutations in important genes governing genome wide methylation, including TET2, IDH1/2, and DNMT3. While commonly invoked as a reason to explain sensitivity to HDAC inhibitors, the fact is only hypomethylating agents and not HDAC inhibitors, would be expected to affect this underlying biology.

Methods: We have compared Azacitidine (AZA) and Decitabine (DEC) activities in a panel of seven cell lymphoma (TCL) cell lines (H9, HH, P12, PF382, C5MJ, TLMo1 and MT-2) with regard to their dose response effects on cell viability, apoptosis, DNMT protein expression and DNA methylation. We have also evaluated the preclinical merits of combining the hypomethylating (HoMe) agents and HDAC inhibitors (Romidepsin and Belinostat) using cytotoxicity assay (CellTiter Glo). We have performed RNASeq analysis on TCL cell lines to identify unique gene sets perturbed by the AZA-Romidepsin combination.

Results: AZA and DEC demonstrated different effects on cell viability, apoptosis, DNMT levels and DNA methylation. Significant heterogeneity in DEC sensitivity was seen among the lines studies, which was not noted with AZA exposure. AZA was consistently more potent than DEC at high drug concentration (> 1μM) but DEC was active at lower drug concentration (< 0.1μM). The difference in activity was mirrored by the dose dependent effects of AZA and DEC on markers of apoptosis. A strong correlation between DNMT depletion and DNA methylation was noted for both drugs, though the concentration:effect relationships were different. Although decitabine was more effective at lower concentration a more sustained depleting effect on the DNMT1 expression was observed in azacitidine treated cells when compared to the decitabine treated cells. Among the synergistic interactions between the two classes of drugs (HoMe and HDACi), the combination of romidepsin or belinostat plus decitabine was more synergistic than the combination of romidepsin or belinostat plus azacitidine. Not all cell lines exhibit the same degree of toxicity with the combination, with synergy being most notable in HH and H9, while MT-2 appears no better than additively cytotoxic. Moreover, not all cell line lines exhibit the same degree of toxicity upon exposure to romidepsin in combination with repetitive addition of azacitidine over a 96 hour time period. However > 50% hypometylation was observed in all seven cell lines at concentration of 250 nM of azacitidine indicating the synergy was mainly dependent on HDACi and not HoMe agent concentrations. Gene expression profile analysis of untreated and combination treated TCL cell lines estimated that 38, 409 and 1215 genes were differentially expressed when untreated cells were compared to romidepsin, azacitidine and combination treated cells respectively. Statistically over-representation of down-regulated genes related to the cholesterol (8.77 fold) and steroid (6.68 fold) metabolic process and upregulated genes related to gamete generation (5.96 fold) was identified in the combination gene signature. In particular, the expression of 39 genes coding for cancer testis antigens was found upregulated.

Conclusions: These data suggest that HoMe based combinations are active and synergistic in models of TCL. Concentrations that induce DNMT loss may be a valuable biomarker in identifying optimal doses for clinical development. These data established important differences between the HDAC inhibitors and hypomethylating agents. Assumptions these drugs within a class are identical is inaccurate, though clinically they have been hard to differentiate. GEP reveals they affect vast areas of cellular biology, and is likely no singular mechanism will explain the basis for synergy. However, key information with regard to the emergence of cancer testes antigen expression may suggest future rational combinations.