Table S15: Relative quantification by lipid class

Table S12: gene signatures used to analyze RNA-seq data from patients with AML before/after GILT

Table S2: EnrichR analysis of FLT3 gene signatures

Supplementary materials and methods

Table S4: GSEA results in FLT3-ITD PDX AML cells ex vivo-treated with QUIZ or Veh.

Venetoclax plus azacitidine represents a key advance for older, unfit patients with acute myeloid leukemia (AML). The chemotherapy and venetoclax in elderly AML trial (CAVEAT) was first to combine venetoclax with intensive chemotherapy in newly diagnosed patients ≥65 years. In this final analysis, 85 patients (median age 71 years) were followed for a median of 41.8 months. The CAVEAT induction combined cytarabine and idarubicin with 5 dose levels of venetoclax (50-600 mg) for up to 14 days. Two additional cohorts explored adjusted-dose venetoclax (50 mg, 100 mg) with posaconazole. CAVEAT induction was well tolerated, with low mortality (4%) and limited high-grade gastrointestinal toxicity (4%). Delayed hematological recovery after consolidation was ameliorated by omitting idarubicin from post-remission therapy. The overall response rate (ORR: CR + CRh + CRi) was 75% with a median overall survival (OS) of 19.3 months (95% CI 11.1-31.3). Among de novo AML, ORR was 88% and median OS 33.1 months (95% CI 19.3-54.3). Almost one-third have not relapsed, many benefiting from prolonged treatment-free remission (median 17.9 months). CAVEAT induction was well tolerated and associated with high ORR that was durable, particularly for de novo AML. CAVEAT represents an effective time-limited treatment option for fit older patients with AML. (https://www.anzctr.org.au; ACTRN12616000445471).

Table S8: Multimodal analysis of shCEBPA versus shCTL MOLM14 cell line

Table S14: PUFA/MUFA ratios in phospholipids

Venetoclax received full approval in October 2020 for use in older, unfit patients with acute myeloid leukemia (AML) combined with either hypomethylating agents or low-dose cytarabine. This ended a semi-centennial of stalled clinical progress and initiated a new treatment option with proven capacity to enhance response and prolong survival in older patients with AML. Despite widespread use of azacitidine-venetoclax (AZA-VEN), there is increasing appreciation that this regimen is myelosuppressive and associated with a higher risk of infectious complications than AZA alone. Key principles of initial management include prevention of tumor lysis syndrome in patients at high risk and minimizing infectious complications during induction. In the post-remission phase, limiting cumulative marrow suppression by allowing sufficient time between cycles for optimal marrow recovery and truncating the duration of venetoclax exposure for those with delayed blood count recovery have emerged as important axioms of effective care. This new addition to the How I Treat series casts a clinical spotlight on important challenges and dilemmas encountered in practice. We also outline a structured framework to assist in the safe management of AZA-VEN in the clinic.

Table S7: Transcriptomics analysis of shCEBPA#2 versus shCTL

Table S13: lipidomic results by lipid species in dox-inducible and constitutives shRNAs and siRNA CTL/CEBPA, empty vector/CEBPA-OE and vehicle/QUIZ conditions in MOLM-14 cells

Supplemental figures S1-S22: Supplementary Figure S1 shows lipid metabolism dependency in FLT3-mutant AML, Supplementary Figure S2 shows anti-leukemic activity of GILT across a panel of AML PDX, Supplementary Figure S3 shows mechanisms of post-translational regulation of C/EBPα expression, Supplementary Figure S4 shows inhibition of C/EBPα phosphorylation and protein expression by FLT3i, Supplementary Figure S5 shows the implication of Ser-21 phosphorylation in post-translational regulation of C/EBPα expression, Supplementary Figure S6 shows that FLT3-ITD regulates the expression of C/EBPα and of proteins related to lipid biosynthesis in AML cell lines, Supplementary Figure S7 shows that FLT3-ITD regulates the expression of genes related to lipid biosynthesis in AML cell lines, Supplementary Figure S8 shows that C/EBPα directly regulates the transcription of lipid biosynthesis genes in AML, Supplementary Figure S9 shows the correlation between CEBPA mRNA expression and FLT3-ITD mutations, Supplementary Figure S10 shows scRNA-seq analysis from PDXTUH84, Supplementary Figure S11 shows scRNA-seq analysis from PDXTUH110, Supplementary Figure S12 shows the evolution of CEBPA_UP and FLT3-ITD_UP signatures in patients with AML treated with GILT, Supplementary Figure S13 shows that C/EBPα regulates rate-limiting lipid biosynthetic enzymes downstream of FLT3-ITD, Supplementary Figure S14 shows that C/EBPα controls lipid amount in FLT3-ITD cell lines, Supplementary Figure S15 shows that FLT3 inhibitors induce a lipid switch increasing neutral lipids dependent on C/EBPα, Supplementary Figure S16 shows that FLT3 inhibitor inhibits fatty acid synthesis fueled by glucose and glutamine, Supplementary Figure S17 shows that FLT3 inhibition decreases monounsaturated fatty acid dependent on C/EBPα, Supplementary Figure S18 shows that FLT3 inhibition increases PUFA/MUFA ratio dependent on C/EBPα, Supplementary Figure S19 shows that ferroptotic cell death induced by FLT3i is mediated by inhibition of SCD-dependent mono-unsaturated FA synthesis, Supplementary Figure S20 shows that FLT3 inhibitors unmask a vulnerability of FLT3-mutant leukemic cells to ferroptosis, Supplementary Figure S21 shows that lipid redox stress induction by GPX4 inhibition primed FLT3i activity in FLT3-ITD AML cells and Supplementary Figure S22 shows combined treatment with GILT and APR-246 in preclinical AML models in vivo.

Table S3: Clinical characteristics of samples from patients with AML used in PDX assays

Table S6: Proteomic analysis of FLT3-ITD AML cell lines treated with vehicle or QUIZ

Table S1: Transcriptome analysis of FLT3-ITD cell lines after FLT3 inhbition

Table S5: GSEA results in FLT3-ITD PDX AML cells in vivo-treated with GILT or Veh.

Table S10: EnrichR analysis of CEBPA_UP signature

Background: BCL2 pro-survival proteins (BCL2, MCL1, BCL-XL) function as key regulators of oncogenic survival. BH3-mimetic drugs bind to pro-survival proteins (e.g. venetoclax to BCL2), enhancing BH3-only proteins to induce BAX/BAK dependent apoptosis. MCL1 is overexpressed in acute myeloid leukemia (AML) and increased expression has been observed in leukemias resistant to venetoclax-based therapy. Co-targeting BCL2 and MCL1 with selective BH3-mimetics has synergistic activity in diverse AML cases, including TP53-mutated disease (Thijssen, Blood, 2021). Clinical development of MCL1 inhibitors has been hampered by treatment-associated rises in serum troponin (Yuda, Commun Med, 2023). To overcome this, a novel antibody drug conjugate (S227928; Servier-Novartis) has been designed, comprising a monoclonal antibody directed at human CD74, bridged by a linker to an MCL1 selective BH3-mimetic. CD74 was chosen for its high expression on AML cells and low expression on critical organs, including cardiomyocytes. In this study, we highlight the activity of S227928 in combination with venetoclax in models of AML. Aims: To assess the pre-clinical efficacy and CD74 target selectivity of S227928 alone and in combination with venetoclax in AML. Methods: CD74 expression was determined by APC labelled anti-human CD74 (Clone REA1103;Miltenyi). CRISPR-CAS9 gene editing was performed to generate CD74-/- AML cell lines. Drug sensitivity assays were evaluated by flow cytometry. Patient derived xenografts were performed in NSG-SG3 mice. S227928, isotype-MCL1i and unconjugated anti-CD74 were from Servier/Novartis. Results: Expression of CD74 was highest on EOL-1>MV4;11, MOLM13, OCI-AML3>>K562 cells. Pro-apoptotic S227928 activity by 72 h was highest in EOL-1 cells (LC50<10 nM) and least in K562 cells (LC50~10 µM). Combined venetoclax and S227928 was synergistic in AML cell lines, including those with TP53 deleted by CRISPR. Cell death was significantly reduced in BAX/BAK or CD74 deficient cells, confirming a mitochondrial apoptosis mechanism specific to CD74 expressing cells. In patient samples, CD74 expression was highest in B-lymphoid (CD20+) and monocytic blasts (CD64+ or CD14+) > leukemic stem and progenitor cells (CD34+ and/or CD117+) > myeloid blasts (CD33+) >> T lymphoid cells (CD3+). Single-agent S227928 activity in primary AML was observed in 3/14 cases (LC50<100 nM). In contrast, S227928 + venetoclax was active in 9/14 primary AML cases, comparable to non-conjugated MCL1 inhibitor + venetoclax (activity in 9/14) and greater than either unconjugated anti-CD74 + venetoclax (activity in 3/14) or an isotype-conjugated MCL1 inhibitor control (activity in 0/14). Concordant with CD74 expression being highest in monocytoid leukemias, S227928 + venetoclax appeared most potent against monocytic lineage AML. Expression of CD74 on non-leukemic CD34+ progenitors was significantly lower than expression in primary AML samples. Concordantly, viability of normal CD34+ cells was not significantly impacted by either S227928 alone or in combination with venetoclax, in contrast to the cytotoxic impact of cytarabine or daunorubicin on CD34+ cell viability. To assess activity of S227928 + venetoclax in vivo, mice were transplanted with AML cells expressing CD74, CD14 and CD64 and harboring biallelic TP53 V173M and V143M variants. Cohorts were treated with 1) vehicle, 2) S227928 (30 mg/kg IV q 14 days), 3) isotype-MCL1i q 14 days + venetoclax 50 mg/kg days 1-14 (week 1-2) and weekdays (week 3-8), 4) unconjugated anti-CD74 q 14 days + venetoclax, 5) S227928 (10 mg/kg q 14 days) + venetoclax or 6) S227928 (30 mg/kg q 14 days) + venetoclax for 8 weeks. Analysis of bone marrow 2 weeks after commencing treatment revealed blast reduction greatest for the S227928/veneteclax combination arms. Overall survival was longest for S227928/venetoclax combination therapy (median 78 and 96 days), compared to vehicle (median 41 days), S227928 (median 49 days), isotype-MCL1i + venetoclax (median 59 days) or unconjugated anti-CD74 + venetoclax (median 49 days). Conclusion: S227928 is a novel CD74 ADC that combines synergistically with venetoclax to enhance killing in pre-clinical models of AML. Importantly, this combination is active in TP53 mutated and monocytoid AML, sub-groups associated with venetoclax failure. These findings support the clinical development of S227928 alone and in combination in patients with AML.

Abstract While transcription factor C/AAT-enhancer binding protein a (C/EBPa) is critical for normal and leukemic differentiation, its role on cell and metabolic homeostasis is largely unknown in cancer. Here, multi-omics analyses uncovered a coordinated activation of C/EBPa and Fms-like tyrosine kinase 3 (FLT3) that increased lipid anabolism in vivo and in patients with FLT3-mutant acute myeloid leukemia (AML). Mechanistically, C/EBPa regulated FASN-SCD axis to promote fatty acid (FA) biosynthesis and desaturation. We further demonstrated that FLT3 or C/EBPa inactivation decreased mono-unsaturated FA incorporation to membrane phospholipids through SCD downregulation. Consequently, SCD inhibition enhanced susceptibility to lipid redox stress that was exploited by combining FLT3 and glutathione peroxidase 4 inhibition to trigger lipid oxidative stress, enhancing ferroptotic death of FLT3-mutant AML cells. Altogether, our study reveals a C/EBPa function in lipid homeostasis and adaptation to redox stress, and a previously unreported vulnerability of FLT3-mutant AML to ferroptosis with promising therapeutic application

PURPOSE A prospective phase II study examined the safety and efficacy of venetoclax combined with low-dose cytarabine (LDAC) in AML at first measurable residual disease (MRD) or oligoblastic relapse. METHODS Patients with either MRD (≥1 log10 rise) or oligoblastic relapse (blasts 5%-15%) received venetoclax 600 mg once daily D1-28 plus LDAC once daily D1-10 in 28-day cycles. The primary objective was MRD response in the MRD relapse cohort or complete remission (CR/CRh/CRi) in the oligoblastic relapse cohort. RESULTS Forty-eight adults with either MRD (n = 26) or oligoblastic (n = 22) relapse were enrolled. Median age was 67 years (range, 18-80) and 94% had received previous intensive chemotherapy. Patients received a median of four cycles of therapy; 17% completed ≥12 cycles. Patients with oligoblastic relapse had more grade ≥3 anemia (32% v 4%; P = .02) and infections (36% v 8%; P = .03), whereas grade 4 neutropenia (32 v 23%) or thrombocytopenia (27 v 15%) were comparable with the MRD relapse cohort. Markers of molecular MRD relapse included mutant NPM1 (77%), CBFB::MYH11 (4%), RUNX1::RUNX1T1 (4%), or KMT2A::MLLT3 (4%). Three patients with a log10 rise in IDH1/ 2 (12%) were included. By cycle 2 in the MRD relapse cohort, a log10 reduction in MRD was observed in 69%; 46% achieved MRD negative remission. In the oligoblastic relapse cohort, 73% achieved CR/CRh/CRi. Overall, 21 (44%) underwent hematopoietic cell transplantation. Median overall survival (OS) was not reached in either cohort. Estimated 2-year OS rate was 67% (95% CI, 50 to 89) in the MRD and 53% (95% CI, 34 to 84) in the oligoblastic relapse cohorts. CONCLUSION For AML in first remission and either MRD or oligoblastic relapse, venetoclax plus LDAC is well tolerated and highly effective.