BCL2L10 (Bcl-B) Is Associated with Resistance to Azacitidine (AZA) in MDS and AML, and Is a Possible Therapeutic Target in AZA Resistant Patients

AZA is a current reference treatment for higher-risk MDS patients, also active in AML patients unfit for intensive chemotherapy. However, no biological marker predictive of AZA resistance has been clearly established. In addition, patients not responding to AZA or relapsing after a response have very poor outcome (Prebet, JCO, 2011), and require new treatments. We have generated SKM1-R, an AZA resistant cell line (Cluzeau et al., Cell cycle 2011), that exhibits increased expression of BCL2L10, an anti-apoptotic Bcl-2 family member, as a likely cause of resistance (Yasui et al., Cancer Research 2004). We tried to correlate BCL2L10 expression with response to AZA in MDS and AML patients and analysed the effect of ABT 737 a peptidomimetic inhibitor binding the BH3 domain of the BCL-2 family of proteins (especially BCL2L10), on patient leukemic cells expressing BCL2L10.

In two cohorts of 77 MDS or AML patients treated by AZA (75mg/m²/d, 7days every 4 weeks), we quantified the percentage of BCL2L10 positive bone marrow (BM) cells using flow cytometry. Cytometry assay was performed after several steps of fixation, permeabilization, primary incubation with an anti-BCL2L10 antibody (cell signalling) and secondary incubation with a donkey anti-Rabbit FITC-antibody. The first cohort included fresh BM samples from 32 higher-risk MDS (www.clinicaltrials.gov, NCT01210274) or AML patients treated with AZA. The second one is composed of 45 frozen samples from low risk MDS, higher-risk MDS or AML treated by AZA. Response and relapse after AZA initial response were assessed by IWG 2006 criteria. Cell metabolism (XTT assay) and Propidium iodide (PI) staining were performed in 10 patients treated with AZA (5 sensitive and 5 resistant) with or without ABT-737 (Abbott). Resistance to AZA was defined by primary failure or relapse after initial sensitivity to AZA

In the first patient cohort, the mean value for BM cells from AZA-resistant patients was significant higher than AZA-sensitive patients (85% vs 0% respectively, p<0.0001). In the second cohort, retrospective comparison of BM samples from low risk MDS patients, AZA-sensitive or AZA-resistant patients showed that the counts of BCL2L10 positive cells were also significant different (0%, 10% and 33% respectively, p<0.0001). In sub-group analysis on patients from cohort 2, we observed that quantification of BCL2L10 positive cells could predict failure to AZA at diagnosis (10% in sensitive patients vs 33% in resistant patients, p=0.023) and relapse after response to AZA (14% in AZA responding patients without relapse vs 43% in AZA responding patients who relapsed, p=0.0002). Using a cut-off of 50% BCL2L10 positive cells in patients from cohort 1, patients with > 50% of BCL2L10 positive cells had a 3 months OS at 51% compared to 95% in patients with ≤ 50% of BCL2L10 positive cells (p=0.0016).We next used high doses of ABT-737, alone or in combination with AZA. Cell metabolism (XTT assay) and Propidium iodide (PI) staining in the presence or absence of ABT 737 showed no effect in AZA-sensitive patients. By contrast, we observed a significant increase of PI positive cells and a significant decrease of cell metabolism in AZA-resistant patient samples stimulated with the combination of AZA + ABT-737 vs AZA alone.

In this work, quantification of BCL2L10 positive cells by flow cytometry was correlated with resistance and survival with AZA treatment in MDS and AML. This possible biomarker of resistance will have to be tested prospectively. Inhibition of BCL2L10 positive cells by ABT-737 suggests that this drug, or other drugs targeting more selectively BCL2L10 could be proposed in combination with AZA in such patients.

No relevant conflicts of interest to declare.