Cholesterol biosynthesis-related genes reveal new opportunities for risk stratification and treatment in Acute Myeloid Leukemia

Le résumé fourni par la source

Abstract Although genetic and molecular profiling is a well-established method for risk stratification in acute myeloid leukemia (AML), implementing such strategies is often unfeasible in low- and middle-income countries (LMICs), especially those with a public healthcare system. To address this, we have previously developed a four-gene prognostic index (4-PI) for intensively treated AML patients. High 4-PI score was associated with inferior clinical outcomes and upregulation of cholesterol homeostasis genes (Ortiz et al., BJH 2023). Among the 4-PI genes, SQLE and DHCR7 encode key enzymes in the cholesterol biosynthesis pathway, suggesting their potential as therapeutic targets. These enzymes can be inhibited by existing pharmacological agents, such as FDA-approved terbinafine for SQLE, offering alternative strategies beyond statin-based approaches. Here, we first validated the 4-PI in a multicenter real-world Brazilian cohort and assessed the cytotoxic effects of SQLE and DHCR7 pharmacological inhibition in AML cells. Overall, we included 111 intensively treated Brazilian patients with de novo AML. Gene expression was assessed by RT-qPCR. In our cohort, 4-PI analysis classified patients into low (20, 18%), intermediate (43, 38.7%), and high (48, 43.2%) 4-PI groups. Complete remission rate was higher in the low-risk patients (70.0% vs. 52.1%). Cox regression analysis demonstrated that high-risk patients had lower overall survival (OS). After adjusting for age, sex, leukocyte count, and FLT3 mutational status, high 4-PI remained an independent predictor of poor OS. Afterwards, we focused on the pharmacological inhibition of SQLE and DHCR7 in AML cells. We performed MTT assays after 72h of incubation with terbinafine (0-400 µM) or NB-598 (0-128 µM), to target SQLE, or with AY-9944 (0-100 µM) to target DHCR7. IC50 for cytarabine (AraC, 0-24 µM) was also performed. Among the eight screened AML models, MV411 and MOLM13, both FLT3-ITD+ cells, showed high sensitivity to terbinafine (IC50= 59 and 78.1 µM, respectively), NB-598 (IC50= 50.4 and 51.2 µM, respectively) and AY-9944 (IC50=5.2 and 6.2 µM, respectively); whereas KG1a and Kasumi1 were less responsive (terbinafine: IC50>149 µM; NB-598: IC50>86 µM; AY-9944: IC50>26 µM). To evaluate drug synergism, we opted to investigate the combined effect of terbinafine with AraC by calculating the combination index (CI). Thus, IC25 of AraC was combined with different concentrations of terbinafine. CI values <1 indicated synergism, =1 an additive effect, and >1 antagonism. Synergistic effects were observed in MOLM13, Kasumi1 and MV411 (CI=0.79, 0.82 and 0.84, respectively), while additive effects were seen in KG1a and SET2 (CI=0.90 and 0.96), suggesting potential therapeutic benefits of this combination in AML. To investigate whether the observed synergistic effect was linked to modulation of cholesterol biosynthesis, intracellular cholesterol levels were measured using the Amplex™ Red assay after 24h of treatment with AraC alone or combined with terbinafine. Terbinafine (≥50 µM) significantly suppressed AraC-induced cholesterol accumulation in MOLM13 and MV411 cells, but not in Kasumi1, KG1a, or OCIAML3. To explore the molecular basis of terbinafine sensitivity in AML, cell lines were ranked by IC50 values and RNA-seq data was analyzed. Gene set enrichment analysis revealed that terbinafine-resistant cells upregulated genes associated with the cellular response to sterol depletion (GO:0006991), suggesting activation of compensatory pathways in less sensitive AML cells. Collectively, our data confirms the prognostic value of 4-PI in a real-world AML setting. Pharmacological inhibition of SQLE and DHCR7 decreased AML cell viability, with SQLE inhibition exhibiting synergistic effects and modulation of intracellular cholesterol levels in FLT3-ITD+ cell lines. These findings underscore the critical role of cholesterol biosynthesis and metabolism pathways as key factors influencing therapeutic response and risk prediction in AML.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.