A Study on Preclinical Efficacy, Underlying Mechanisms, and Sensitivity Markers of a Novel Hypomethylating Agent Ntx-301 in Acute Myeloid Leukemia

Introduction: While intensive induction chemotherapy has been standard-or-care for patients with acute myeloid leukemia (AML), intensive regimens have often been limited in elderly patients with comorbidities. As an alternative option, less intensive hypomethylating agents (HMAs), decitabine (DAC) and azacitidine (AZA), are currently being used in these unfit patients. However, their low response rates and adverse events when administered alone suggest a need to develop novel HMAs that elicit enhanced efficacy and reduced mortality. A recent study demonstrated pharmacological improvements of a novel 4′-thio-modified analog, 5-aza-4′-thio-2′-deoxycytidine (hereafter NTX-301), including enhanced chemical stability and incorporation into DNA and improved preclinical efficacy (Thottassery, 2014). However, the mechanism of action (MoA) of NTX-301 has not yet been understood. Herein, we aim to thoroughly investigate the preclinical efficacy and MoAs of NTX-301 in AML through comparative analysis with DAC and AZA. To this end, we used in vitro and in vivo preclinical models of AML and performed multiomics-based analyses.

Results: We comprehensively examined viability of 200 cancer cell lines (CCLs) upon treatment with NTX-301. Consistent with the current use of HMAs as therapeutics for hematologic malignancies, this sensitivity profiling displayed the most remarkable potency of NTX-301 toward blood CCLs (OR=3.97, p=0.0003). In addition, phenotypic assays revealed that the anti-leukemic activity of NTX-301 was superior to that of DAC, which is attributed to more effective actions in inducing apoptosis, cell cycle arrest, and differentiation. Supporting the in vitro results, orally administrated NTX-301 led to more potent tumor regression, better tolerability, and survival benefits over DAC and AZA in both systemic (1.5-2.0 mg/kg, qdx5 then 2 days off, 5 days on, and 9 days off, for 3 cycles) and subcutaneous (0.2-1.5 mg/kg) xenograft models.

To assess the underlying MoAs, we interrogated global alterations at the transcriptome and methylome levels upon treatment with NTX-301 in three AML cell lines using RNA sequencing and methylation array. Methylome analyses revealed that NTX-301-induced demethylation patterns were distinguished from DAC; DAC triggered stronger global demethylation than NTX-301 did, whereas NTX-301 derived rather selective demethylation, preferentially in early-replicating regions, H3K27ac-marked regions, and non-CpG islands. In transcriptome analyses compared with DAC, NTX-301 more markedly elicited a transcriptional reversal toward a normal myeloid-like signature by increasing a differentiation signature and suppressing a leukemic stem cell signature. NTX-301 also mediated more pronounced activation of DNA damage response and the p53 pathway, which are characterized by marked induction of pH2AX and pChk1 and increased stability of p53, respectively. Given the synthetic lethality of p53 activation and BCL2 inhibition (Rongqing, 2017), stronger p53 stabilization by NTX-301 may confer more benefits in combination with venetoclax. Indeed, the combination of NTX-301 + venetoclax produced a more synergistic combination index compared with DAC + venetoclax. Strikingly, the combined NTX-301 (0.5-2.0 mg/kg) + venetoclax (50 mg/kg) achieved complete tumor remission, no notable toxicity, and prolonged survival benefits over AZA (2.5 mg/kg) + venetoclax (50 mg/kg) in preclinical models of AML.

By integrating sensitivity profiles and multiomics data of 200 CCLs, we interrogated molecular determinants associated with sensitivity to NTX-301. Intriguingly, when comparing methylomes between sensitive and resistant CCLs, we found a significantly biased global hypermethylation trend toward sensitive CCLs. A combinatorial set of the most significantly biased 352 differentially methylated regions (FDR<0.05) showed potential as a predictive sensitivity marker for NTX-301, exhibiting a significant correlation (r=0.69, p<0.0001) with sensitivity to NTX-301.

Conclusions: Our study demonstrated an improved therapeutic index of NTX-301 over traditional HMAs, providing a rationale for further clinical development of the agent as a single-agent or in combination with other agents. We also believe that our study for MoAs and biomarkers will improve our understanding of NTX-301.