The Role of Adenosine Deaminase Acting on RNA 1 (ADAR1) Enzyme in the Resistance to Hypomethylating Agents

Hypomethylating agents (HMA) remain the mainstay of treatment for high-risk Myelodysplastic Syndrome (HR-MDS), but their exact mechanism of action remains largely unknown. Though HMA induce a viral mimicry response on cancer lines, this phenomenon was not associated with clinical responses when examined in primary leukemia cells. In patient-derived cancer cells differential expression and activity of ADAR1, an A-to-I RNA editing enzyme which depletes the immunogenic dsRNAs was shown to affect HMA efficacy. Therefore, we investigated if the expression of ADAR1 in primary samples from HMA-treated patients with HR-MDS is related to the dissociation between the induction of viral mimicry and clinical response to HMA.

We assessed ADAR1 levels, its two isoforms, the constitutively expressed (p110) and the interferon-inducible (p150), STAT3a and STAT3b isoforms by quantitative real-time PCR in FACS-sorted primary CD34⁺ cell subsets. Response to HMA was evaluated using the revised International Working Group Response Criteria for MDS (2023).

Non-responders to HMAs (n=24) demonstrated markedly higher pretreatment ADAR1 levels in total CD34⁺ blasts compared to responders (n=20, p=0.001). Likewise, the expression levels of both ADAR1p110 and ADAR1p150 isoforms were higher in non-responders compared to responders to HMA (p=0.001 for both). The level of cellular differentiation where AZA acts by inducing the viral mimicry state remains unidentified; therefore, we measured ADAR1 levels in FACS-sorted hematopoietic stem and progenitor cell subpopulations. We observed significantly higher pretreatment ADAR1p150 levels in the primitive Lin^-CD34⁺CD38^- compartment of non-responders (n=20) compared to responders (n=19) to HMA (p=0.003) but not in Lin^-CD34⁺CD38⁺ committed progenitors (p=0.3), erythroblasts (p=0.4), or the mature Lin⁺CD34⁺ subset (p=0.5).

Of note, at the time of relapse after HMA we observed in both total CD34⁺ and Lin^-CD34⁺CD38^- cells a significant upregulation of the expression of both ADAR1 (p=0.008 and p=0.031, respectively) and ADAR1p150 (p=0.039, p=0.031, respectively) compared to pretreatment levels (n=8). Also, early responders to HMA (at first 4 cycles, n=13) demonstrated significantly higher levels of ADAR1 at the time of both their first (p=0.026) and best response (p=0.035) compared to late responders (after 4 cycles, n=5), further supporting the involvement of ADAR1 in a slow-acting immune-mediated mechanism of action of AZA.

Recent data support the existence of a positive feedback regulatory loop between STAT3b and ADAR1, in which STAT3b promotes ADAR1 expression and in turn, ADAR1 contributes to the stabilization of STAT3b.

Νo significant associations of pretreatment levels of STAT3a with response to HMAs (p=0.23) was found in total CD34⁺ cells, but non-responders displayed marginally higher levels of STAT3b (p=0.057). We further found a strong positive correlation between the expression of STAT3b and ADAR1p150 in the CD34⁺ cells of all patients (p<0.0001), supporting the potential existence of the above feedforward loop. By contrast, no correlation of the ADAR1 levels with the mutational profile of the patients was found.

Despite two decades of clinical experience, the mode of action of HMAs remains elusive and no current therapeutic strategy can overcome resistance and/or restore the response to HMAs, whereas there is also lack of predictive biomarkers. In line with preclinical data, our findings implicate the overexpression of the ADAR1 enzyme as a potential mechanism of resistance to HMA. If confirmed in larger cohorts, our data argue for the utilization of ADAR1 as both a therapeutic target and a predictor of response to HMA.

Papaemmanuil:Isabl Inc.: Current holder of stock options in a privately-held company, Other: CEO, Patents & Royalties; TenSixteen Bio: Current holder of stock options in a privately-held company.