Intensive Chemotherapy (IC) remains standard of care in fit older AML patients (pts). Oncogenetics can only partly predict the outcome of these pts. Leukemic Stem Cells (LSCs) are deemed chemoresistant. Functional assessment of blasts and/or LSCs could improve risk stratification. Mitochondrial priming of blasts relative to normal hematopoietic stem cells (HSCs), measured by cytochrome C release upon exposure to a BIM peptide has been proposed to predict CR in AML pts (Vo Cell 2012). Whether mitochondrial priming of blasts and/or LSCs has prognostic relevance in the context of recent oncogenetic risk stratification remains unknown.
To interrogate mitochondrial priming of blasts and LSCs in a prospective cohort of older AML pts treated with IC.
We studied bone marrow samples of newly diagnosed de novo or secondary AML ≥60y pts enrolled in the ALFA 1200 trial (NCT01966497). Pts received one 7+3 course followed by 2 intermediate-dose cytarabine courses. Pts were risk stratified according to ELN 2017 guidelines. Thawed BMNCs were stained with Lin, CD45, CD34, CD38, CD45RA, CD123 and CD90 antibodies then incubated with a BIM peptide at 0.1, 0.3 and 1.0 μM or DMSO. Cytochrome C release was analyzed on blasts (CD45dim/SSClow), LSCs (combining GMP-like [Lin-/CD34+/CD38+/CD123+/CD45RA+] and LMPP-like [Lin-/CD34+/CD38-/CD90-/CD45RA+] subsets, Goardon Cancer Cell2011) and pre-leukemic HSCs (Lin-/CD34+/CD38-/CD45RA-). Mitochondrial priming was determined by fitting an area under the curve of Cytochrome C release across BIM concentrations (Figure).
We analyzed 65 pts (median age 71y, M/F 34/31). ELN 2017 risk was favorable, intermediate and adverse in 43%, 17% and 40% resp. High WBC (>50 x109/L) or secondary AML did not affect blasts or LSC priming (all p>0.4). Blasts and LSC priming were moderately correlated (R2 = 0.62). Patients achieving CR/CRp (n = 43, 66%) had higher priming of LSCs (p = 0.03), but similar priming of blasts (p = 0.31) compared to patients with induction failure (n = 19) or early death (n = 3). Based on ROC curve analysis, we identified a best cutoff of 57% LSC priming; 30 (46%) pts had high LSC priming (>57%). 25/30 (83%) pts with primed LSC achieved CR/CRp compared to 18/35 (51%) with low LSC priming (p = 0.009). In a multivariate regression high LSC priming was predictive of CR/CRp (OR = 4.2 95% CI 1.3–13.8, p = 0.02) independently of WBC, sAML and ELN 2017 risk. Forty-one pts (63%) had primed blasts relative to their cognate HSCs; 5y-OS was 40% (95%CI 24–55%) in pts with relatively primed blasts vs 19% (95%CI 24–55%) in pts with unprimed blasts (p = 0.046). Relative LSC priming had no impact on OS (p = 0.9). In a multivariate Cox model, primed blasts (relative to cognate HSCs) predicted longer OS (HR = 0.47, p = 0.02) independently of non-adverse ELN 2017 risk (HR = 0.38, p = 0.002). Integration of mitochondrial priming refined ELN stratification and identified a group of 23 pts (35%) with both non-adverse ELN risk and primed blasts with very favorable outcome (5y-OS 63%, 95%CI 39% - 80%) in the context of AML of the elderly (Figure).
Mitochondrial priming is feasible in a clinical trial. Assessment of different compartments may yield different information on short-term vs long-term outcome. Our results support prospective study of mitochondrial priming in older AML pts. Integration of functional biomarkers to conventional oncogenetic classifiers will help personalize treatment in older AML pts.