Primary Acute Myeloid Leukemia Research Articles

The Mevalonate Pathway Is a Therapeutic Target in TP53 Mutant Acute Myeloid Leukemia

Introduction: Acute myeloid leukemia (AML) with mutations in the TP53 tumor suppressor gene is the most fatal of AMLs with a median overall survival of only six months due to chemotherapy resistance. To understand biologic differences in TP53 mutant (MT) AML, as compared to wildtype (WT) AML, we performed transcriptomic analysis on sorted patient samples. Gene set enrichment analysis demonstrated significant upregulation of the cholesterol biosynthesis or mevalonate pathway in TP53 MT AML. In TP53 MT solid tumor models, the mevalonate pathway plays a key role in tumorigenesis and metastasis. We hypothesized that the mevalonate pathway is essential for chemotherapy resistance of TP53 MT AML and represents a novel therapeutic target. Methods: We developed chemoresistant, isogenic TP53 MT AML cell lines using CRISPR/Cas9 technology in the MOLM14 AML cell line. We used primary TP53 MT AML patient samples to perform colony forming unit (CFU) assays, in vitro assays with a serum substitute and cytokines, and a patient derived xenograft (PDX) model of TP53 MT AML. Results: We first sought to determine if TP53 MT AML exhibits upregulation of mevalonate pathway activity in response to cytarabine (AraC), the backbone of AML therapy. We treated our isogenic TP53 MT AML cell lines for 24 hours (h) with AraC and assessed mevalonate pathway gene expression and metabolites using qRT-PCR and liquid chromatography high resolution mass spectrometry (LC-HRMS), respectively. Only TP53 MT cell lines respond to AraC with a significant upregulation of key mevalonate pathway genes and metabolites. We then determined if TP53 MT AML is sensitized to AraC by inhibition of the mevalonate pathway with a statin. We pretreated our isogenic cell lines with 24h rosuvastatin before 24h AraC and assessed cell viability with flow cytometry with AnnexinV/7AAD staining and XTT assays. We also performed CFU assays in TP53 MT AML patient samples treated with the two drugs alone or in combination and assessed CFUs after 14 days. These studies revealed a synergistic reduction in cell viability and CFUs in TP53 MT AML by rosuvastatin in combination with AraC. We next addressed the mechanism by which a statin sensitizes TP53 MT AML to AraC. Our group and others have demonstrated enhanced mitochondrial activity is associated with AML chemoresistance. We hypothesized that mevalonate pathway byproducts, known to be crucial for mitochondrial functions, contribute to the mitochondrial response to AraC. We treated TP53 MT AML cell lines as above with rosuvastatin and AraC and used seahorse technology to measure oxidative phosphorylation (OXPHOS). We also used electron microscopy, flow cytometry of mitochondrial protein, TOM20, and qRT-PCR for mitochondrial DNA content to quantify and characterize mitochondria. These data demonstrate that TP53 MT AML cell lines, compared to WT, exhibit a significant increase in OXPHOS after 24h that is due to an increase in total mitochondrial. These effects are abrogated by pretreatment with a statin. We also validated these findings in primary TP53 MT AML patient samples in vitro. Importantly, co-treatment of the isogenic TP53 MT AML cell lines with a soluble form of mevalonate or a downstream byproduct, geranylgeranyl pyrophosphate (GGPP), recovered OXPHOS and subsequent chemoresistance. Overall, this data supports a model where TP53 MT AML cells dynamically upregulate the mevalonate pathway to regulate OXPHOS and avoid DNA damage-induced cell death (Figure 1). Finally, we studied a PDX model of TP53 MT AML. Engrafted mice were treated with AraC for five days alone or in combination with high dose rosuvastatin. Mice were harvested on day eight with assessment of leukemic burden by flow, OXPHOS by seahorse and mevalonate byproducts by LC-HRMS. TP53 MT AML leukemic burden was only significantly reduced by the combination therapy. Consistent with in vitro findings, mice sacrificed three days after completing AraC continue to demonstrate enhanced OXPHOS and increased mevalonate byproducts. This response is blunted by rosuvastatin. Conclusions: These results demonstrate that TP53 MT AML requires the mevalonate pathway for chemotherapy resistance and targeting of this pathway in combination with chemotherapy may improve clinical responses. Analysis is ongoing to determine the role of GGPP in regulating the mitochondrial response to DNA damage in TP53 MT cells in order to elucidate novel therapeutic approaches.

Ferritinophagy Is a Druggable Vulnerability of Quiescent Leukemic Stem Cells

Acute myeloid leukemia (AML) poses significant challenges in terms of prognosis and treatment options. Leukemic stem cells (LSCs) contribute to therapy resistance, relapse and adverse outcomes. This study aimed to explore one of the major features of LSCs (e.g. quiescence) and its underlying molecular mechanisms, as well as identify potential targets for therapeutic intervention. Using in vivo cell labeling in patient-derived xenograft (PDX) assays, we identified a quiescent cell subpopulation with high LSC potential. Transcriptomics experiments conducted separately on quiescent and cycling cell subsets revealed gene signatures characteristic of quiescent (Q_AML_UP) and cycling cells (C_AML_UP). Using single-cell RNA sequencing, we identified clusters of quiescent leukemic cell that were highly enriched in gene signatures related to LSCs (Figure 1A). Through a machine learning approach, we identified a gene set highly specific of quiescent cells (named the QS35 gene signature) that emerged as a significant prognostic factor for poor outcomes in independent cohorts of AML patients (Figure 1B). Next, we aimed to identify new vulnerabilities of LSCs by investigating the quiescent subset of leukemic cells. Transcriptomics and proteomics approaches on quiescent vs cycling cells revealed enhanced autophagy within the quiescent subset. Additionally, the transferrin receptor TFRC and ferritin, two key regulators of iron metabolism, exhibited lower expression in the quiescent subpopulations. As intracellular iron availability is tightly regulated by ferritinophagy, a specific form of autophagy leading to the degradation of ferritin, we investigated the coordinated regulation of autophagy and iron metabolism in the quiescent LSC subset. Intriguingly, the intracellular iron pool was lower in quiescent LSCs, particularly in the CD34+CD38- compartment compared to more mature leukemic cells. Moreover, autophagy inhibition selectively induced cytotoxicity in the CD34+CD38- compartment and reduced the frequency of long-term culture-initiating cells (LTC-IC), and these effects were rescued by exogenous iron supply, underscoring the critical role of autophagy in iron bioavailability within the quiescent LSC subpopulation. To target ferritinophagy in this compartment, we depleted NCOA4, the main carrier in ferritinophagy responsible for delivering ferritin-bound iron to the autophagosome, using shRNAs in primary AML samples. Interestingly, NCOA4 was found more expressed in quiescent compared to cycling AML cells. In ex vivo assays, NCOA4 downregulation significantly reduced the frequency of LTC-IC and inhibited the clonogenic capacities of leukemic cells. Furthermore, NCOA4 invalidation abrogated leukemia initiation and self-renewal in serial transplantation experiments in vivo, particularly within the quiescent LSC population (Figure 2A). Additionally, treatment with a novel small molecule inhibitor named compound 9a that disrupts the NCOA4-ferritin interaction, depleted intracellular iron in primary AML cells, reduced their clonogenic potential and LTC-IC frequency ex vivo, and induced cell death specifically within the quiescent CD34+CD38- subset. Importantly, the viability of LSCs treated with compound 9a was rescued by the addition of iron, demonstrating that the cytotoxic effects of this compound were dependent on iron depletion. Notably, compound 9a exhibited minimal toxicity to normal hematopoietic cells ex vivo. In PDX models, in vivo short-term treatment with compound 9a decreased tumor burden and induced cell death within the CD34+CD38- LSC population (Figure 2B). Mechanistically, genetic of chemical inhibition of NCOA4 led to the accumulation of dysfunctional mitochondria, and to the generation of reactive oxygen species especially within the LSCs population. Furthermore, inhibition of NCOA4 by compound 9a decreased mitophagy, strongly suggesting that NCOA4 coordinates ferritinophagy and mitophagy pathways to increase intracellular iron, thereby specifically promoting the survival of LSCs. These findings highlight the critical vulnerability of quiescent LSCs to ferritinophagy. Thus, NCOA4 inhibition represents an innovative therapeutic approach for AML patients.

Navtemadlin, a Novel MDM2 Inhibitor, Potentiated Venetoclax-Induced Antitumor Efficacy in TP53 Wild-Type Acute Myeloid Leukemia (AML)

Introduction: The tumor suppressor p53 ( TP53) maintains genomic integrity, and nearly 50% of cancers are characterized by a loss or mutation in the TP53 gene. Murine double minute 2 (MDM2) negatively regulates p53 expression and promotes oncogenesis in TP53 wild-type ( TP53 WT) tumors. Navtemadlin (nvtm) is a potent, selective, orally available MDM2 inhibitor that overcomes MDM2 dysregulation by restoring p53 activity and inducing apoptosis in TP53 WT tumors. AML is a heterogeneous hematological malignancy resulting from the transformation of stem cells and is characterized by poor therapy response. Thus, combination therapies are required to avoid tumor escape mechanisms, such as alterations in apoptotic machinery and metabolism. Venetoclax (ven), a BCL-2 inhibitor, is a promising therapy in AML. Overexpression of anti-apoptotic members of the BCL-2 family, including myeloid cell leukemia-1 (MCL-1) limits ven response, thus requiring MCL-1 inhibitors to further improve sensitivity. A proof-of-concept study with idasanutlin (idasa), a first-generation MDM2 inhibitor and ven demonstrated manageable safety and encouraging efficacy in relapsed/refractory AML patients (Daver et al. Blood 2023). Here, we present a preclinical evaluation of nvtm alone and in combination with ven in the setting of TP53 WT AML, with a focus on metabolic pathways, pro-survival signaling, and bone marrow (BM) niche. Methods: AML-derived cell line, MOLM-13 ( TP53 WT, FLT3-ITD) was cultured alone, with HS5 AML stromal cells or with cytokines. Cultures were treated with DMSO, nvtm (0.1-1.5 μM), or ven (0.5 μM) for short-term (4-8 hrs) or long-term (48-72 hrs) and analyzed for changes in cell cycle progression, viability, clonogenicity potential, and metabolism (glycolysis and mitochondrial oxidative phosphorylation by analyzing extracellular acidification rate (ECAR) and oxygen consumption rate (OCR) respectively). Primary AML BM mononuclear cells (MNCs) were cultured in cytokine support media. Results: In MOLM-13 cells, short-termnvtm (0.5 µM) treatment rapidly restored p53 expression and phosphorylation, with greater efficacy than idasa. Long-term treatment significantly induced apoptosis in a dose-dependent manner from 0.25 μM (p<0.01) with 80% cytotoxicity at 0.75 μM (p<0.001). Cell cycle analysis at 48 hr showed that nvtm significantly arrested cells in the sub-G0-G1 phase. Clonogenicity potential was significantly lowered in nvtm-treated cells. Metabolic alterations in the Mito Stress and Glycostress tests showed nvtm significantly lowered EACR, OCR, and mitochondrial membrane potential, as compared to idasa. We next evaluated the effects of nvtm in ven-treated MOLM-13 cells. Combination treatment significantly lowered the metabolism, cell viability, and clonogenicity potential of MOLM-13 cells, compared to nvtm or ven treatment. Immunoblotting analysis revealed a reduction in key anti-apoptosis proteins (BCL-2, MCL-1, PARP). Strikingly, nvtm downregulated ven-induced MCL-1 expression in a dose-dependent manner. Culturing MOLM-13 cells in HS5-conditioned media, transwells overlying HS5 cells, or with HS5 cells increased MOLM-13 cell resistance to nvtm. However, in HS5 co-cultures, a combination of nvtm and ven strongly re-sensitized MOLM-13 cells for apoptosis. Nvtm also significantly blocked the cytoprotective effects of AML-relevant cytokines (IL-6, SCF, FLT3 ligand, TPO, and GM-SCF). In cultures of primary AML BM-MNCs (n=11), nvtm alone at clinically relevant concentrations reduced viability (14-40%), and a combination nvtm and ven significantly enhanced AML cytotoxicity (30-52%). Conclusions: Nvtm, a potent second-generation MDM2 inhibitor, rapidly restored p53 function, induced apoptosis in p53 WT AML MOLM-13 by impeding cell cycle, viability, clonogenicity potential, and reducing both glycolysis and oxidative phosphorylation. Nvtm combined with ven significantly enhanced these effects. Importantly, this combination blocked stromal cell-mediated (contact and soluble factors) cytoprotection. These encouraging results provide a mechanistic basis for the evaluation of this combination in vivo and as a therapeutic strategy to overcome documented ven resistance mechanisms in TP53 WT AML patients.

Autophagy Induction Constitutes a Targetable Vulnerability in FLT3-Mutant Acute Myeloid Leukemia Cells Treated with FLT3-Inhibitors

Recently, we and others identified autophagy induction as a protective mechanism in FLT3-ITD mutant acute myeloid leukemia (AML) cells upon treatment with FLT3-inhibitors (FLT3i), constituting an increasing issue in clinical practice. Here, we investigated the molecular consequences of autophagy induction upon FLT3i treatment, exploring resistance mechanisms and interrogated whether autophagy inhibition can improve FLT3i antileukemic therapy. Drug response of primary FLT3-ITD AML samples ex vivo treated with FLT3i, quizartinib (AC220, n= 59) and midostaurin (PKC412, n=64), was associated with their transcriptional profile to unravel which molecular programs were enriched in poor responders. Both gene set enrichment analysis (GSEA) and single sample GSEA (ssGSEA) revealed enriched molecular signatures compatible with autophagy activation in poor responders to FLT3i, while sensitive samples enriched molecular signatures consonant with mitochondrial function. Enrichment scores for programs associated with “regulation of autophagy” and “autophagosome organization” were positively correlated to ex vivo response area under the curve values for AC220 and PKC412. To functionally address the molecular pathways associated with autophagy activation upon FLT3i treatment, MOLM-13 and MV4-11 AML cells ( FLT3-ITD) and primary AML samples, were treated with AC220 and PKC412 alone or in combination with the autophagy inhibitor chloroquine (CQ, 5-10μM). In FLT3-ITD AML models, treatment with FLT3i significantly induced the formation of acidic vesicular organelles, suggesting autophagy induction. We also observed a reduction of autophagy regulators in a dose- and time-dependent manner, evidenced by the conversion of LC3B-I to LC3B-II, with degradation of LC3B-II and p62. Furthermore, FLT3i treatment resulted in phosphorylation inhibition of AKT Ser473, mTOR Ser2448 and its downstream protein p70S6K Thr421/Ser424, and phosphorylation reduction of ULK1 Ser757. Combination of FLT3i with CQ (10 μM) significantly reduced cell viability and survival for both PKC412 and AC220 inhibitors. To rule out the off target effects associated with CQ treatment, we performed a knockdown (KD) on the key autophagy mediator ATG5 using shRNA-mediated inhibition in FLT3-ITD cells. In MOLM13 cells, FLT3i treatment in ATG5-KD cells significantly reduced the IC50 (control: 31nM vs. ATG5-KD: 17nM) and increased apoptosis induction (control: 41% vs. ATG5-KD: 58%). We validated these observations in ex vivo treated primary AML samples, where we observed that in FLT3-ITD but not in FLT3 wild-type samples, the combination of FLT3i (PKC412 and AC220) plus autophagy inhibitors (CQ) reduced cell proliferation and survival. In summary, we demonstrated that autophagy induction is a biological process intrinsically associated with drug resistance in FLT3-ITD mutant AMLs when treated with FLT3i. Hence, autophagy inhibition can be leveraged to increase the antileukemic effect of FLT3i and highlights the autophagy as a major acquired biological process associated with FLT3i therapy.

Wilms' Tumor 1 Functions As a Tumor Suppressor to Suppress FLT3-STAT Signaling and Epigenetic Remodeling in Acute Myeloid Leukemia (CALGB 8461, 9665 and 20202; Alliance)

The Wilms' Tumor 1 ( WT1) gene is a transcription factor that is recurrently mutated or commonly overexpressed in several cancer types. In acute myeloid leukemia (AML), frequent overexpression of WT1 and poor patient outcomes associated with WT1 mutations highlight its importance in the disease; however, there are no tailored treatments for these patients. Furthermore, WT1's fundamental role as either an oncogene or tumor-suppressor remains unresolved. Here we examine the roles of both wild-type and mutant WT1 in AML through epigenetic and mechanistic studies. Using our findings we propose a personalized treatment strategy. WT1 mutations are enriched in the NPM1 mut subset of AML; thus we first focused our study on a cohort of 581 patients with de novo AML and NPM1 mutations enrolled on Alliance for Clinical Trials in Oncology studies. Transcriptomic analyses revealed that WT1 mut patientsphenocopied a distinct, aberrant gene expression signature associated with FLT3 internal tandem duplications (ITD), a mutation known to activate STAT signaling and associated with poor outcome in AML. We observed that WT1 expression levels were remarkably elevated in FLT3-ITD patients and were driven by STAT5A binding to the WT1 promoter. STAT5A binding and subsequent WT1 upregulation were blocked by small molecule FLT3 inhibitors. These findings linking FLT3 activity with WT1 expression raise two possibilities: either WT1 is an oncogene cooperating with the FLT3-STAT pathway, or WT1 naturally functions to suppress FLT3 signaling in a negative feedback loop subverted by WT1 mutations. As WT1 interacts with TET2 to facilitate epigenetic remodeling and DNA binding, we performed genome-wide DNA methylation analysis on WT1 mut AML patients and observed selective hypermethylation of WT1 binding motifs consistent with loss of function. We confirmed that WT1 mutations in AML commonly cause truncation of the C-terminal DNA binding domain in our patient cohort. Using co-immunoprecipitation, ChIP-sequencing and luciferase reporter assays, we found that mutant WT1 functions as a dominant-negative, inhibiting WT1 derived from the wild-type allele. Investigation of WT1 target genes by intersecting transcriptomic and DNA methylation profiles identified hypermethylation and downregulation of miR-193a, leading to a significant downregulation of miR-193a-3p in WT1 mut AML patients. Overexpression of wild-type, but not mutated, WT1 rescued miR-193a expression in AML cell lines. Enforced expression of miR-193a significantly delayed AML onset in vivo. In addition, overexpression of miR-193a in multiple human AML cell lines and primary patient samples impaired AML cell growth and colony-forming capacity while promoting monocytic differentiation, underscoring its role as potent tumor suppressor. Therapeutic modulation of miRNA levels in cancer patients has been limited by inefficient delivery and tissue enrichment. To overcome this, we tested a novel lipid-nanoparticle (LNP) formulation of miR-193a-3p (INT-1B3), currently being investigated in a phase I clinical trial (NCT04675996). Biweekly i.v. treatments of INT-1B3 in the immunocompetent Hoxa9/Meis1 (H9M)-transduced model system prevented AML formation, highlighting the potent anti-leukemic activity of this miRNA based therapeutic. Overexpression of miR-193a-loaded LNPs downregulated FLT3 expression and suppressed STAT signaling in primary AML samples. Finally, treatment of primary AML cells with FLT3 inhibitors revealed enhanced sensitivity of WT1 mut cells in the absence of FLT3-ITD, highlighting the role of wild-type FLT3 in WT1 mut cells. In summary, we uncovered a critical negative-feedback loop maintained by WT1 to suppress FLT3 activity. Loss-of-function WT1 mutations subvert the tumor suppressor function of WT1 via failure to maintain miR-193a expression, leading to increased FLT3 expression and STAT5 signaling, subsequently impairing differentiation, increasing proliferation and disease aggressiveness in AML (Figure 1). Our findings advocate for use of FLT3 inhibition and miR-193a supplementation for treatment of WT1 mut patients, a subgroup with poor outcomes and no targeted treatment options.

CD38-Directed, Single-Chain T Cell-Engager Targets Leukemia Stem Cells through IFNγ-Induced CD38 Expression

Leukemia stem cells (LSCs), enriched in the CD34+CD38- AML population, are considered the main drivers of relapse in acute myeloid leukemia (AML). While more differentiated progenitors and bulk AML blasts reside in the CD34+CD38+ fraction, targeting CD38 in AML is less pursued, as LSCs are mainly CD38-. We show IFNγ (10 ng/ml) induced an increase in CD38 mRNA (mean fold change IFNγ over vehicle = 7.7, p=0.03, n=3) and surface CD38 (mean CD38 High% population IFNγ 56%±17 vs vehicle 16%±10, p<0.0001) in both CD34+ and CD34- AML primary blasts, and decreased the clonogenic activities of primary AML cells (colony forming unit IFNγ 50±36 vs vehicle 216±117, p=0.002, n=4) regardless of cytogenetics. scRNA and total RNA sequencing data in CD34+ AML primary blasts showed that upon IFNγ treatment, CD38 and Interferon Regulatory Factor 1 (IRF-1) were consistently the most upregulated genes compared to vehicle (>2.5 Log 2 Fold Change; p=0; FDR=0). Furthermore, gene set enrichment analysis (FDR=0) indicated IFNγ induced an upregulation of genes involved in apoptosis and DNA repair, and a suppression of genes in the pro-oncogenic Wnt beta catenin pathways. CD38 luciferase-based reporter assays, chromatin immunoprecipitation, and IRF-1 knockdown experiments in AML cells showed that, upon IFNγ treatment, IRF1 is pivotal in CD38 transcriptional regulation. Because CD4+T helper type 1 and CD8+ cytotoxic T cells release IFNγ once engaged on their target, we hypothesized a CD38-directed T cell engager will convert CD34+CD38- LSC into CD34+CD38+ blasts via the secreted IFNγ to kill LSCs and blasts. We created a single chain construct (BN-CD38), inserting a CD38 nanobody between the light and heavy chains of an anti-CD3 Fab using two short peptide linkers. Based on modeling, BN-CD38 reproduces the distance between T cell receptor and the major histocompatibility complex, mimicking the natural immunological synapse. Surface plasmon resonance and cytometry studies showed BN-CD38 had strong binding affinity to both CD38 (K D 4.77xE -10 M) and CD3 (K D 4.27xE -8 M). The IC 50 for BN-CD38 at 1:1 E:T at 24hrs in different CD38+ AML cell lines (n=6) was within the 0.01 to 0.1 ng/ml range, whereas the IC 50 of a non-binding CD38 construct (BN-CD38Mut) was 100-fold higher. By 24hrs of treatment, we noted >60% CD4+ and CD8+ T cell activation (CD69+, CD25+, IFNγ+) with BN-CD38 (e.g., CD69: 60%±27) in contrast to controls (e.g., CD69: IgG 7%±0.8 and BN-CD38Mut 11%±10, p<0.0001). In primary AML samples (n=10) BN-CD38 activated >70% T cells against autologous leukemia cells (CD69: BN-CD38 73%±16, BN-CD38Mut 10%±9, IgG 11%±10, p<0.0001). CyTOF analysis showed BN-CD38 but not controls significantly reduced CD34+CD38+ AML blasts (%change to IgG range: BN-CD38 -27 to -82% vs BN-CD38Mut 1% to 7%, p = 0.01) and CD34+CD38- LSCs (%change to IgG range: BN-CD38 -78 to -99% vs BN-CD38Mut -7 to 11%, p= 0.0003) and expanded CD8+ effector memory (p<0.01), CD8+ terminally differentiated effector memory CD45RA+ (p<0.05), NKT cells (p<0.01), and central memory Tregs (p<0.0001). Anti-human-IFNγ neutralizing antibody reverted BN-CD38 induced expression of CD38 and BN-CD38 killing activity on CD34+CD38- AML cells (p<0.05). BN-CD38 (2.5 mg/kg, i.v, n=6/group) co-injected weekly with healthy 3x10 6 T cells induced near-complete disease eradication in mice xenografted with luciferase-CD38+ THP1 (p=0.003) and U937 (p=0.004) cells. Long-term survival analysis showed 38% of BN-CD38-treated THP-1 engrafted mice achieved complete cure compared to IgG (OS median: BN-CD38 61.5d, control IgG 29d, p=0.0005). Consistently, BN-CD38 suppressed AML cell engraftment, splenomegaly, and prolonged survival in three patient-derived xenografts (PDX) with complex karyotypes including TP53 mutation (PDX1: OS median BN-CD38 undefined, BN-CD38Mut 28d (p=0.002), and IgG 27d (p=0.002); PDX2: OS median BN-CD38 39d, BN-CD38Mut and control IgG 32d (p<0.01); PDX3: OS median BN-CD38 undefined, BN-CD38Mut 92d (p=0.001), and IgG 81d (p=0.0007)). BN-CD38 did not affect normal hematopoietic stem cell clonogenicity and the development of multi-lineage human immune cells in CD34+ humanized mice. In sum, BN-CD38 induces an effective T cell synapse and IFNγ release, while concomitantly blocking the clonogenicity and inducing CD38 expression on LSCs. These studies suggest a new mechanism to unmask and target LSCs, providing the rationale for using BN-CD38 to effectively treat AML.

Addiction to Small Gtpase ARF6-Mediated Sphingolipid Homeostasis By AML but Not the Host

Background Cancer cells' unique metabolic demand has emerged as a promising therapeutic target in Acute Myeloid Leukemia (AML). The vesicular trafficking and signaling gatekeeper, ADP-Ribosylation Factor 6 (ARF6), belongs to the RAS superfamily of small GTPases and has been shown to regulate the homeostasis of lipids such as PIP2, phosphatidic acids, and cholesterol. Here we describe AML cells' unusual addiction to ARF6-mediated sphingolipid homeostasis and its clinical significance. Experiments and Results ARF6 knockdown (KD) or CRISPR-mediated ARF6 knockout (KO) inhibited the proliferation of AML cell lines, THP-1 and MV4-11, as well as the colony formation of primary AML CD34 + cells. Surprisingly, Tamoxifen-induced whole body KO of Arf6 in adult Arf6 f/f;Rosa-CreERT 2 mice led to healthy and grossly normal mice with a normal lifespan. Adult mice lacking ARF6 show no abnormality in complete blood count, urine test, or histology. Hence, ARF6 is genetically indispensable for AML, but not for the host. To confirm AML's addition to ARF6 in vivo, we introduced Arf6 KO in an MLL-AF9-induced AML model in C57BL6/J mice. We are currently following up with the disease progression between Arf6-KO AML group and control AML group, and expect to obtain results before the 2023 ASH meeting. After discovering that ARF6 is genetically necessary for AML but not needed for the host, we analyzed its clinical significance. Since ARF6 mutations are not detected in AML samples, we focused on analyzing its expression. With univariate analysis, we show that high ARF6 expression (top 50%) is significantly associated with worse overall survival (OS) and worse disease-free survival (DFS) of AML patients in the TCGA-NEJM-2013 (TCGA) cohort (N=200) and the Münchner Leukämielabor (MLL) cohort (N=579). In multi-variate analysis, high ARF6 expression is significantly associated with worse OS in the BeatAML cohort (N=562) and shows a trend towards worse OS and DFS in the TCGA cohort. Following a hypothesis-driven approach, we investigated the role of ARF6 in regulating sphingolipid homeostasis by HPLC-MS lipidomic analyses of AML cells (THP-1) with or without ARF6 depletion. Lipidomic profiles revealed that ARF6 KD led to an increase of the overall ceramide/sphingomyelin ratio, suggesting a defect in sphingolipid homeostasis. We then analyzed the expression of sphingolipid metabolic enzymes and observed that ARF6 KD downregulated sphingomyelin synthase 1 and 2 (SGMS1 and SGMS2) expression, hindering ceramide-to-sphingomyelin conversion. Supporting this, we found a positive linear correlation between ARF6 and SGMS1/2 expression among the patients in TCGA cohort, confirming their correlation. High expression of the ARF6 effectors SGMS1/2 is also associated with worse OS and DFS in the TCGA cohort and the MLL cohort in univariate analysis, resembling that of ARF6. Downstream of ceramide accumulation, ARF6 KD inhibited the pro-survival AKT pathway, confirming the pro-apoptotic roles of ceramide in cancers. The anti-proliferative effect of ARF6 depletion can be rescued by sphingomyelinase inhibitor GW4869 (blocking ceramide production) or by expressing a constitutively active form of AKT (AKT1 E17K). To pharmacologically target ARF6, we developed two generations of ARF6 small molecular inhibitor pro-drugs, named NAV-2729 and A6-4471. NAV-2729 and A6-4471 showed high specificity to ARF6 without inhibiting other small GTPases, with in vitro IC 50 in the lower μM range. Both NAV-2729 and A6-4471 inhibited AML cell line proliferation and reduced primary AML cell colony formation in lower μM range. Cord blood normal CD34 + cells are not sensitive to ARF6 inhibitors, but non-specific toxicity was observed at higher concentrations. When tested in vivo with THP-1 xenografts in NSG mice, both inhibitors reduced circulating AML cells without obvious toxicity to the host. Conclusion ARF6 is necessary for AML proliferation and its higher expression is associated with worse patient survival. However, ARF6 is dispensable for normal host physiology. Mechanistically, ARF6 maintains critical ceramide/sphingomyelin balance in AML by regulating SGMS1/2 expression. Pharmacologically inhibiting ARF6 suppressed AML proliferation in vitro and in vivo in a genotype-agnostic fashion. Our study implicates ARF6 as a potentially actionable and safe target in AML.

Novel Therapeutics Targeting Acute Myeloid Leukemia (AML) Stem Cells Identified through High-Throughput Screening

Acute myeloid leukemia (AML) is an aggressive form of blood cancer. Despite the use of cytotoxic standard-of-care drugs, patients often succumb to the disease partially due to the inability of medically unfit patients to withstand the cytotoxic treatments, regrowth from minimal residual disease and the chemo-resistant nature of leukemic stem cells (LSCs). Hence, novel therapies should focus on targeting the unique biology of LSCs to eliminate and avoid reoccurrence. To overcome challenges inherent of screening rare LSCs in vitro such as culturing and expanding, we optimized the conditions for a 4-week in vitro large-scale expansion (>600 million bulk) and enrichment of the CD34+ LSC-containing fraction (>90% purity) for a primary human AML sample (OCI-AML-8227), functionally validated to be enriched for LSCs in long-term xenotransplant assays (Eppert et al., 2011). Next, we performed a high-throughput screen of 11,140 chemical molecules in 3 stages. First, the viability of AML CD34+ cells and healthy cord blood (CB) CD34+ cells was read out at 1-2 doses using a CellTiter-Glo® Luminescent assay. 61 compounds had >70% inhibition of 8227 CD34+ cells and <30% inhibition on CB CD34+ cells. Next, we determined the dose response and refined the hits to 33 potent compounds with LC50 < 1 μM, including novel compounds and classes previously shown to target bulk and leukemic stem cells in AML. We are now presenting the follow up hits identified from the third stage of validation where we determined the LC 50 specifically in the CD34- (blast) and CD34+CD38- (LSC-enriched) OCI-AML-8227 populations using flow cytometry. We identified 25 novel anti-LSC compounds with high efficacy against CD34+CD38- AML cells (LC 50 < 500 nM). Venetoclax was among the top hits, a compound that revolutionized treatment for poor prognosis AML patients due to its anti-LSC activities, providing internal validation of the screen. We then tested these candidates in a second LSC-enriched model with poor prognosis, OCI-AML-20, to ensure the response of compounds on LSCs is not exclusive to OCI-AML-8227 and to investigate the influence of the microenvironment on drug efficiency. A total of 5 hits have significant responses (>50% reduction in LSC enriched populations) and nontoxic to stroma. Furthermore, we refined our top candidates to 7 compounds based on their efficiencies in the two LSC-enriched models (OCI-AML-20 and 8227) and low toxicity on stroma. To gain insights on their mechanisms of action in LSCs, the two models were treated with the 7 candidates and cytarabine as a control for 3 days and apoptosis was measured by flow cytometry. Out of 7 candidates, 5 induced apoptosis in the LSC-enriched fractions, suggesting elimination of LSCs through apoptosis. The remaining compounds may eliminate LSCs through different mechanisms. From these results, we focused on the three leading compounds and validated them for toxicity on CD34+ hematopoetic stem and progenitor cord blood cells (HSPCs) vs CD34+CD38- (LSC-enriched) OCI-AML-8227 by flow cytometry. Although slight toxicity was observed in HSPCs at higher doses, the LC 50 for the stem cell vs LSC-enriched populations differ by ~6-40-fold, indicating that a significantly lower concentration can be used to eradicate LSCs while sparing HSPCs (LC 50 for CD34+ population CB vs 8227; Compound A: 1558 nM vs 83 nM, Compound B: >2000 nM vs 305 nM and Compound C: 612 nM vs 16 nM, respectively). To functionally validate if HSPCs proliferation and differentiation was impaired and if leukemic progenitors were eradicated, a colony forming unit assay was performed for 12 days after 6 days of treatment. Candidates were found to eliminate leukemic progenitors by 50% or more, reducing leukemia initiating potential and having minimal or no impact on CB progenitor functionality. Overall, compound A is a potential candidate to move forward due to its ability to target and eliminate LSCs through apoptosis in two LSC-enriched models, its low toxicity on stroma and normal cord blood cells. This candidate is classified as an indole, a class shown by Pabst et al., 2014 to reduce CD34+CD15- AML cells in a drug screen for aryl hydrocarbon modulators, suggesting a potential mechanism for LSC elimination through apoptosis. We now aim to examine LSC eradication in a panel of genetically defined primary AMLs to be able to determine the broad applicability of this compound and translate the preliminary results for clinical use.

KDM6A Modulates Anti-Tumor Immune Response By Integrating Immunogenic Cell Death in Human Acute Myeloid Leukemia

Identifying tumor cell-epigenetic regulation of immune activation in acute myeloid leukemia (AML) would be instrumental in developing next-generation of targeted therapy. AML cells have low mutational burden; inefficient ‘cross-presentation’ through conventional dendritic cells (DC) and immune activation are the major challenges for AML immunomodulation. ‘Immunogenic cell death’ (ICD) represents changes in the composition of the tumor cell surface as well as the release of ‘damage associated molecular patterns’ (DAMPs) in the tumor microenvironment that stimulates the dysfunctional immune system. DAMPs in turn operate on a series of receptors expressed by DCs and natural killer (NK) cells and contribute to long-lasting, protective anti-tumor immunity. For example, plasma membrane exposure of endoplasmic-reticulum (ER)-translocated calreticulin (ecto-CRT), which functions as an ‘eat me’ signal, is an important hallmark of ICD. Previous study suggested that calreticulin (CALR) promotes immunity and type I interferon-dependent survival in mice with AML (Fucikova, Blood 2016; Chen, Oncoimmunology 2017). In addition, ecto-CRT on malignant blasts correlated with improved NK cell-mediated cytotoxicity in AML patients with superior overall survival (Truxova, Haematologica 2020). Ecto-CRT has also been identified as an endogenous sterile ligand for NK cells (Sen Santara, Nature 2023). Nevertheless, molecular regulation of ICD in the context of epigenetic derangements and inflammation and its implication in AML immunotherapeutic modulation is poorly understood. KDM6A is a histone 3 lysine 27 demethylase that plays tumor suppressor function in AML (Gozdecka, Nat Genet 2018; Sera, Blood 2021). KDM6A escapes X-chromosome inactivation, and loss-of-function deletion or point mutations associate with resistance to ‘3+7‘ chemotherapy (Stief, Leukemia 2020). We have identified that in comparison to control AML cells, deficiency of KDM6A ( KDM6A-kd) associated with daunorubicin/etoposide-induced ICD, which was manifested by significant increase (2-3 fold , P<0.05) in ecto-CRT, plasma membrane exposure of pE1F2α, HSP70 and HSP90 with extracellular release of HMGB1 and ATP. Similarly, pharmacological inhibition of JmjC catalytic function using GSK-J4 sensitized human AML lines as well as primary AML patient bone marrow-derived CD34 + hematopoietic stem/progenitor cells to ICD. Induction in ICD was accompanied by an elevated type I interferon production that was functionally active in promoting interferon regulated gene expression. Interestingly, treatment with acute recombinant IFN-β caused an increase in activated STAT1 and inflammatory gene expression in KDM6A-kd AML cells compared to control AML. Corroborating these results, gene set enrichment analysis revealed a significant enrichment (NES=1.708, P=0.007) of upregulated type I interferon signatures (n=40) in Kdm6a -/- primary leukemia cells. Re-analysis of ChIP-seq results indicated 1647 Kdm6a occupied genes, which were upregulated upon Kdm6a loss. Gene ontology (GO) analysis of the 1647 genes further suggested an enrichment ( P<0.001) of type I interferon-responsive GO terms, indicating Kdm6a role in interferon signaling. Ecto-CRT level relies on premortem ER stress. In OHSU AML, KDM6A expression correlates with ER stress/unfolded protein response. Mechanistically, quantitative-ChIP and promoter-reporter assays indicated that KDM6A-kd de-repressed ICD induced expression of ATF4 and HSPA5 through chromatin and transcriptional remodeling. In agreement with these findings, AML blasts-healthy immune cells ex vivo co-culture experiments revealed that increase in ICD in KDM6A-kd AML was associated with activation (~2-foldincrease inCD80 +CD86 +, P<0.05) and phagocytic function of DCs. Loss of KDM6A was shown to bestow ‘BRCAness’ (Boila, Leukemia 2023), and PARP inhibition itself can cause ICD and augment tumor immunity. We observed that there was a significant induction in ICD and expression of NKG2D ligands ULBP1, ULBP5 in KDM6A-kd AML cells treated with olaparib. In addition, olaparib and GSK-J4 treatment showed additive effects in ICD and immune activation using primary AML CD34 + cells. Together, we illustrate that KDM6A regulates ICD through epigenetic mechanisms, while KDM6A deficient AML subtypes could be sensitized with immunomodulatory targeted therapy.

Functional Deconvolution of the Leukemia Stem Cell Pool Uncovers Co-Existing Alternative Stem Cell Driven Hierarchies

Leukemia stem cells (LSC) are not functionally homogenous, rather constitute a complex reservoir of multiple subtypes with distinct genetics and functional potential, that fuels disease progression in acute myeloid leukemia (AML). Understanding the intricacies within the LSC pool is critical for informed design of LSC targeting therapies. However, such in-depth studies are hampered by several technical challenges such as the scarcity of LSC and the lack of reliable cell surface markers to isolate viable LSC or distinct types of LSC. To overcome these challenges, we turned to the patient-derived OCI-AML22 model. This model includes functionally, transcriptionally, and epigenetically characterized LSC broadly representative of LSC extracted from primary AML samples. Focusing on the pool of LSC, we deconvoluted the multi-layered heterogeneity, using a panel of in vivo limiting dilution and single cell assays combined with single cell multiome analysis. We uncovered the co-existence of alternative hierarchies driven by distinct LSC subtypes that differ in quiescence depth, differentiation potential, repopulation capacities and that can be prospectively enriched. Notably, we captured the transcriptomic footprint of these newly discovered LSC subtypes within large primary AML cohorts, and even concomitantly within the same AML patient across different LSC fractions, highlighting the broad biological relevance of our findings.

Single-Cell Proteogenomics Analysis of AML Samples Treated with Decitabine and Venetoclax Reveals Divergent Treatment Escape Mechanisms

Background: The combination therapy of hypomethylating agents and venetoclax (HMA+VEN) has emerged as a standard of care for acute myeloid leukemia (AML) patients unfit for intensive chemotherapy. Nonetheless, significant clinical challenges persist, including relapse and treatment resistance. Previous studies have revealed genetic ( RAS/RTK pathway mutations and TP53 mutations) and phenotypic (monocytic differentiation) resistance mechanisms in the context of HMA+VEN therapy. Despite these advancements, our comprehensive understanding of the HMA+VEN resistance mechanisms remains incomplete. To gain further insights into the resistance mechanism in HMA+VEN therapy, we implemented single-cell proteogenomics analysis on longitudinally collected samples from AML patients treated with decitabine+venetoclax (DAC+VEN) trial. Methods: We performed a high-throughput, single-cell targeted DNA-antibody sequencing (DAb-seq) using Tapestri platform from Mission Bio Inc. Sixty-three AML samples sequentially collected from 33 unique patients who participated in DAC+VEN clinical trial (NCT03404193) were analyzed. Our custom DNA panel covered 37 recurrently mutated genes in AML, whereas the custom antibody panel included either 15 cell surface proteins (N = 20 patients) or 48 proteins (N = 13 patients). Cellular phenotypes were inferred based on the patterns of cell surface protein expression. Median 2,767 cells were sequenced per sample, with total 196,719 cells sequenced for this study. Results: The median age of the cohort was 70 years (range 21-84). Twenty-four (73%) patients had primary AML and 9 (27%) had secondary or therapy-related AML. Based on the FAB classification, 13 patients had M0/M1/M2, 6 had M4/M5, one had M7, and the rest had unknown subtypes. The most frequent gene mutation at baseline was DNMT3A (45%), followed by NPM1 (39%), TET2 (30%), and TP53 (24%). Among the 33 patients, 32 had evaluable treatment response; 27 (82%) were defined as responders by achieving CR or CRi, whereas 5 (15%) were non-responders. Six patients experienced relapse after responding to the therapy, and their baseline-relapse pairs were analyzed by the single-cell platform. Using the composition of cellular phenotype data, AML samples were classified into either stem-like (N = 5), progenitor-like (N = 1), predominantly monocytic (N = 4), predominantly erythroid (N = 2) and mixed population AML (mixed population of stem, progenitor, monocyte or erythroid-like population, N = 21). Genotype-phenotype analysis showed a significant correlation between the KRAS mutation and the mixed phenotype (p= 0.0403), as well as between the U2AF1 mutation and the mixed phenotype (p= 0.0471). There was no significant association between treatment response and AML phenotype. Analysis of six baseline-relapse pairs showed that relapse was associated with phenotypic shift of AML cells. Stem-like cells were largely eradicated at relapse and replaced by more differentiated cells. Among the six baseline-relapse pairs, three showed monocytic shift, one showed erythroid shift and two showed mixed monocytic/erythroid shift. Phenotypic shifts at relapse were not always associated with expansion of new genetic mutations, except for in two cases: one had monocytic shift accompanied with expansion of KRAS mutation, and another had erythroid shift accompanied with expansion of NRAS mutation. Conclusions: Phenotypic transition from stem-like cells to more differentiated cells (either monocytic or erythroid) was frequently observed in AML relapse following DAC+VEN therapy. These findings are consistent with the known dependency of these cells on non-BCL-2 anti-apoptotic proteins such as MCL-1 and BCL-xL. Close monitoring of these phenotypic alterations throughout the course of HMA+VEN treatment may potentially facilitate the precise identification of cellular populations predisposed to relapse.

Epigenetic Modulation Enhances the Therapeutic Potential of All-Trans Retinoic Acid in Acute Myeloid Leukemia

Epigenetic dysregulation has been strongly associated with the development and progression of acute myeloid leukemia (AML). While all-trans retinoic acid (ATRA), a derivative of vitamin A, has demonstrated significant anti-cancer effects in acute promyelocytic leukemia (APL-AML), its clinical efficacy in non-APL AMLs has been limited. This limited response can be attributed, in part, to the epigenetic silencing of genes targeted by or involved in the ATRA pathway, rendering them unresponsive to the drug. Consequently, our research aims to overcome the epigenetic barriers in AML by combining ATRA with small molecules that possess epigenetic modulating properties. To investigate the activation of the retinoic acid pathway in response to ATRA and a library of over 650 small molecules with epigenetic activity, we transduced HEK-293 cells with the p-GreenFire-RARE-Tk-Luc reporter construct. Through the screening process, we have identified 11 compounds that show promise as potential candidates for combination therapy with ATRA in non-APL AML. Notably, when combined with ATRA, two specific compounds, a dual PI3K HDAC inhibitor, and a pan-PKC inhibitor, demonstrated a significant 4-5 fold increase in RARE activity, indicating a synergistic effect. To determine the impact of selected compounds on the functional level, we evaluated the viability, differentiation, and phosphorylation status of ERK1/2, AKT1, MAPK9, p38, and MAPK9 in various cell lines, including HL-60, KG1a, BMNC, and primary MSC and AML cells. The dual PI3K HDAC inhibitor and a pan-PKC inhibitor were noncytotoxic to HL60, BMNC, and primary MSC. Interestingly, dual PI3K HDAC inhibitor with and without ATRA reduced the viability of KG1α by more than 95%. Moreover, dual PI3K HDAC inhibitor and a pan-PKC inhibitor, when combined with ATRA, lead to a 2-3 fold increase in the differentiation of HL60 cells and a 2-4 fold increase in the differentiation of AML cells ex vivo compared to ATRA in combination with the LSD1 inhibitor TCP (treatment control protocol) and untreated cells from AML subtypes M1, M2, and M4. PI3K HDAC inhibitor combined with ATRA significantly reduced p-MAPK and p-p38 levels while increasing the phosphorylation of p53 in HL-60 cells. Furthermore, RNA sequencing and chromatin profiling by Cleavage Under Targets and Tagmentation (CUT&Tag) were performed to gain additional insights into the molecular mechanisms involved. Based on obtained results, the combination of ATRA with specific dual PI3K HDAC and pan-PKC inhibitors holds promise as a potential therapeutic approach for AML. Nevertheless, further investigations, including studies utilizing patient-derived xenograft models, are warranted to validate the efficacy and feasibility of this proposed treatment strategy.

CD38 KO/CD38-CAR Human Primary Natural Killer Cells Enhance Cytotoxicity Against CD38-Expressing Primary Lymphoma, Leukemia, and Myeloma

Natural killer (NK) cells have strong cancer immunotherapeutic potential that can be further augmented by engineering with a chimeric antigen receptor (CAR). CD38, a transmembrane glycoprotein involved in signal transduction and cellular adhesion, is highly expressed on the surface of the hematologic malignancies multiple myeloma (MM), acute myeloid leukemia (AML), Burkitt Lymphoma (BL) and T-cell acute lymphoblastic leukemia (T-ALL). FDA approved CD38 monoclonal antibodies (mAb) daratumumab and isatuximab are used to treat MM in patients, and have been tested to treat other CD38 expressing malignancies. NK cells play an imperative role in the efficacy of these antibodies and have natural antitumor activity against these cancers. However, because NK cells also highly express CD38, anti-CD38 mAb can cause NK cell fratricide, limiting their efficacy. Here, we generated multiple CD38 scFv based on isatuximab sequences and generated fratricide-resistant primary CD38-CAR NK cells using CRISPR/Cas9 genome editing and AAV6 gene delivery. Using this approach, we inserted the CD38-CAR gene into the CD38 locus to efficiently generate simultaneous CD38 KO and CD38-CAR KI with a mean expression level of %60. The CAR-expressing NK cells could be expanded on mbIL21-expressing feeder cells to a clinically relevant number of cells with no decline in CAR-expression level during two weeks of expansion confirming the lack of CAR mediated fratricide. We evaluated transduction efficiency and expression of CD38-CAR, and compared the cytotoxic function of two constructs with CD8 transmembrane (TM)/stalk and 41BB signaling domains and reverse heavy and light chains against AML, MM, T-ALL and BL immortalized cell lines and primary patient samples. We enhanced CD38 expression in target cancer cells by pre-treatment with all-trans retinoic acid (ATRA), a metabolite of vitamin A that pushes cells to maturation, consequently upregulating CD38. The sequential treatment of target cells with 10nM ATRA for 48 hours enhanced the cytotoxic activity of the CD38 KO/CD38-CAR-NK cells against primary MM, BL and AML samples. This would be particularly important for patients with low expression of CD38 on their cancers. The enhanced anti-tumor activity of the CAR was also studied and observed by CyTOF. Additionally, to better understand the therapeutic potential of CD38-CAR NK cells compared to CD38-CAR T cells we generated CD38 KO/CD38-CAR T cells using the same engineering approach and comparing their cytotoxic function. We showed that both CD38-CAR-T and CD38-CAR-NK cells have comparable anti-tumor activity compare to their paired non-edited wild type cells. Finally, we investigated the need for a CD38 KO to develop a fratricide-resistant therapy by comparing the cytolytic and metabolic functions of CD38 +/AAVS1 KO and CD38 KO CD38-CAR NK cells. We conclusively report a CD38-CAR NK cell therapy with enhanced metabolism and cytotoxicity toward a variety of hematologic malignancies, which can be further augmented by combination treatment with ATRA to target CD38-low malignancies.

Psychiatric and Substance Use Disorders Are Independent Predictors of Treatment Response and Outcomes in United States Veterans with Newly Diagnosed Acute Myeloid Leukemia Treated with Venetoclax Combinations

BACKGROUND: Venetoclax (VEN) combinations have become the standard frontline therapy for patients with newly diagnosed acute myeloid leukemia (AML) who are deemed unable to tolerate intensive induction chemotherapy, either due to age or comorbidities. However, outcomes in the real-world setting have been inferior to those observed in clinical trials. Preexisting psychiatric diagnoses and substance use disorders (SUDs) have previously been associated with increased mortality in patients with cancer, but their prognostic influence in veterans with AML is unclear. The aim of this study was to measure the prevalence of psychiatric and SUDs in veterans with AML treated with VEN combinations, determining the impact of these conditions on treatment response and outcomes. METHODS: This was a retrospective cohort study of veterans with newly diagnosed AML who received frontline therapy with VEN combinations within the national Veterans Affairs (VA) Healthcare System. To measure the prevalence of psychiatric and SUDs, we used the Centers for Medicare and Medicaid Services (CMS) Chronic Conditions Data Warehouse algorithm, which tracks health conditions by administrative claims data with International Classification of Diseases (ICD) diagnostic and procedural codes. Our endpoints were rates of complete remission or complete remission with incomplete marrow recovery (CR/CRi), early mortality (EM, or death within 60 days of treatment initiation), intensive care unit (ICU) admissions, and overall survival (OS). We evaluated the relationship between comorbid medical conditions and outcomes using Kaplan-Meier analysis and Cox proportional hazards regression models, first unadjusted, and then adjusted for all covariates. Patients who underwent allogeneic stem cell transplantation (HSCT) were excluded from statistical analyses. RESULTS: We identified 452 veterans treated with VEN up to April 1, 2022 (Table 1). In combination with VEN, 60% (n=270) received azacitadine, 37% (n=168) received decitabine, and 3% (n=14) received low-dose cytarabine. Median age was 74.3 years (IQR: 71.2-78.5) with 52% ≥75 years. By European LeukemiaNet (ELN) 2017 risk classification, 61% (n=274) were classified as adverse risk. Forty-six percent (n=206) of veterans had a preexisting psychiatric diagnosis, and 19% (n=84) had SUDs; 11% (n=49) had both comorbidities. Psychiatric disorders were significantly more prevalent in younger veterans (68% in <65 years vs 50% in 65-74 years vs 39% in ≥75 years, p<0.003); same was true for SUDs (38% in <65 years vs 26% in 65-74 years vs 9% in ≥75 years, p<0.001). For the entire cohort, CR/CRi was achieved by 57% (n=257), and only 3% (n=12) received HSCT. Zero veterans with SUDs and 8 with psychiatric diagnoses were transplanted. At the end of study period, 69% (n=310) had died. Median OS was 216 days (95% CI: 195-254). The 60-day mortality rate was 20% (n=87), of which 8 veterans had primary refractory AML. ICU admissions after treatment initiation was significantly higher in veterans with SUDs compared to those without history of SUDs (5.0 versus 2.2 per 5 person-years, p<0.001). On multivariable analyses (Table 2), both psychiatric and SUDs were independently associated with decreased odds of CR/CRi. Veterans with psychiatric disturbances had a 1.97 times higher hazard of EM (95% CI: 1.16-3.39, p=0.01) and a 1.28 times higher hazard of death (95% CI: 1.00-1.64, p=0.05) compared to veterans without concurrent psychiatric disorders. CONCLUSION: Psychiatric diagnoses and SUDs were not only prevalent among veterans with AML, but they also negatively impacted treatment response, EM, and OS-independently of age, sociodemographic variables, markers of disease risk, and VEN combination. Veterans with SUDs also had significantly higher ICU admissions after treatment initiation for AML. Compared to published clinical trials data, veterans experienced inferior outcomes. Veterans with psychiatric and SUDs may have difficulty adhering to regimens, reemergence of their mental instability and illness that is triggered by AML diagnosis and its treatment, or biologic mechanisms associated with these comorbidities. Better understanding of these factors contributing to health disparities and targeting these comorbidities with supportive care interventions alongside AML-specific therapy may improve outcomes and warrant further study.

Comparative Effectiveness of Reinduction Approaches for Pediatric Relapsed Acute Myeloid Leukemia

Introduction: Intensive frontline chemotherapy achieves initial remission rates above 80% in pediatric acute myeloid leukemia (AML), however, approximately half of patients relapse, and overall survival (OS) after relapse is poor. There is a paucity of comparative effectiveness data on salvage regimens to inform selection of optimal post-relapse therapy. The objective of this analysis was to compare outcomes following relapse by prevalent reinduction regimens in a multi-center real-world dataset. Methods: Patients 18 years and younger who had received frontline chemotherapy for AML between 2011-2019 at 12 pediatric institutions were eligible for inclusion. Trained personnel completed manual chart abstraction to obtain demographics, diagnosis, treatment, and outcome information. The analysis included patients who achieved morphologic remission and then experienced disease relapse; patients with primary refractory AML and those who pursued palliative therapy only after relapse were excluded. Prevalent reinduction regimens were classified as (1) intensive without anthracycline [e.g. fludarabine/cytarabine and high-dose cytarabine-containing regimens], (2) intensive with anthracycline (e.g. CPX-351, mitoxantrone/cytarabine and cytarabine/daunorubicin/etoposide), and (3) non-intensive (including azacytidine or decitabine-based regimens). Primary outcomes were OS after relapse and end reinduction disease response defined by morphologic remission (<5% blasts) and measurable residual disease (MRD)-negative remission (<0.1% blasts). Covariate distributions were compared by reinduction regimen using chi-square tests. Survival analyses used Kaplan-Meier methods and multivariable cox proportional hazards models. End induction response was evaluated using logistic regression. Adjustment was made for covariates found to both differ by reinduction regimen and be associated with outcome, regardless of statistical significance. Results: Of 530 patients, 157 (29.6%) experienced a relapse at a median of 9.7 months (IQR 6.5-16.3) after initial diagnosis. Median post-relapse follow-up was 2.2 years (IQR 1.1-3.2). One- and 2-year OS rates from relapse were 44.9 ± 4.1% and 38.6 ± 4.1%, respectively. Of 157 patients with relapse, 129 from 9 institutions had available information on post-relapse treatment and pursued reinduction with curative intent. Patients were 48.8% female, 27.1% Hispanic, 14.7% non-Hispanic Black and had a median age of 7.1 years (IQR 2.70-14.2) at relapse. Most patients (89.2%) relapsed after completion of frontline therapy, and 24% had undergone hematopoietic cell transplantation (HCT) in first clinical remission. Reinduction regimens were as follows: 79 (61.2%) received intensive reinduction without anthracycline, 34 (26.4%) received intensive reinduction with anthracycline and 16 (12.4%) received non-intensive reinduction. Patients with a non-intensive reinduction were more likely to have relapsed within one year of initial diagnosis (50% versus 34%, p = 0.23) and undergone transplant prior to relapse (75.0% versus 16.8%, p < 0.001). Covariates were balanced across intensive induction regimens. One-year OS was 50.3 ± 4.6%, with 63 (40.1%) achieving morphologic remission and 40 (31.0%) achieving an MRD-negativity after first reinduction cycle. Seventy-three patients (56.9%) underwent HCT after relapse; 11 (15.1%) received a second transplant. OS curves and adjusted outcome comparisons by reinduction regimen are shown in the Figure and Table. One-year OS was 51.1 ± 5.8%, 60.2 ± 9.0% and 26.7 ± 11.4% for intensive therapy without anthracycline, with anthracycline and non-intensive, respectively. However, after adjustment hazard ratios (HRs) were not statistically significant [HR = 0.85 (p = 0.65) and 1.77 (p = 0.23) for intensive with anthracycline and non-intensive, compared to non-intensive without anthracycline, respectively]. Conclusions: After controlling for patient, disease and treatment characteristics, we did not identify a significant difference in remission rates or OS by reinduction approaches for pediatric relapsed AML. Although anthracycline-based reinduction did not improve OS, adjusted point estimates suggest that anthracyclines may improve end reinduction disease response. Ongoing cohort expansion will enable comparative effectiveness analyses of specific reinduction regimens.

Local Oxygen Dictates Hematopoietic Cell Growth and Function

Functional hematopoietic stem (HSC) and progenitor cells (HPC) are responsible for the replenishment of blood/immune cells, are a source of treatment for diseases, and when dysregulated are a source of hematological neoplasia. To identify factors that impact HSC/HPC functional competency, we performed a transcriptomic screen which revealed that human cord blood (CB) HSCs/HPCs capable of early engraftment in an NSG mouse model of transplantation exhibited enriched expression of gene programs that vary with changes in oxygen (O 2) availability compared to non-transplanted input cells. This is physiologically and clinically relevant as HSCs/HPCs and diseased hematopoietic cells such as acute myeloid leukemia (AML) are exposed to O 2 tensions ranging from <1% to 21% as they function in various anatomical sites or are harvested for clinical applications. Thus, there are translational and practical rationales for understanding the growth and functional properties that hematopoietic cells exhibit across the range of O 2 they experience physiologically or artificially, which we examined here. First we found significantly different frequencies of HSCs/HPCs and colony forming units (CFU) in human bone marrow (BM), mobilized peripheral blood (PB), and CB, which have varied O 2 tensions, suggesting O 2 availability may impact hematopoiesis. To directly examine this, human CB CD34+ cells were split and plated in CFU assays or expanded in growth stimulating conditions at 1% O 2, 3% O 2, 5% O 2, 14% O 2 (physiologic tensions), or 21% O 2 (processing tension) . HPCs expanded best and formed more differentiated CFUs in high physiologic O 2 tensions and extra physiologic O 2 tensions, while HSCs expanded better and more potent CFUs were formed in a mid-physiologic O 2 tension (5%). This suggests that O 2 dependent pathways are critical to HSC/HPC growth but there may be a “sweet spot” for early HSCs/HPCs that allows for proliferation while maintaining self-renewal and potency to maintain a pool of primitive cells, possibly mirroring in vivo O 2 responses. In in vivo functional HSC/HPC analyses, NSG mice transplanted with CB units expanded in physiologic O 2 exhibited significantly better early neutrophil recovery and long-term myeloid reconstitution compared to input, suggesting that growth and maintenance of HSC/ myeloid HPC functional competency is better balanced in physiologic O 2 compared to extra physiologic tensions. Platelet recovery and lymphoid reconstitution were significantly better in mice receiving uncultured control CB, suggesting expansion does not affect all HSC/HPC subpopulations equally. Mechanistically, cells grown in low O 2 tensions (1-3%) had higher frequencies of early HSCs/HPCs, significantly lower intracellular ROS levels, a significantly higher percentage of cells in cell cycle G0, and lower levels of phosphorylated γH2AX than those grown in higher O 2. These data suggest that while to some extent HSCs/HPCs require O 2 to proliferate, they also exhibit properties associated with exhaustion when maintained in higher O 2. We next examined the effects varying O 2 tensions have on diseased hematopoietic cells using primary mouse AML cells or oncogene transformed human CB CD34+ cell lines. AML cells proliferated significantly more rapidly and formed more CFUs in higher O 2 tensions, suggesting that O 2 sensing pathways may be a dependency in AML that could affect disease progression as local O 2 tensions become perturbed. AML cells grown in lower O 2 exhibited lower resting levels of phosphorylated γH2AX (measured by flow cytometry) and lower O 2 consumption rates (measured by Seahorse metabolic flux analyzer), suggesting they are less susceptible to cellular stress. In line with this, AML cells maintained in low O 2 exhibited significantly better proliferation and recovery after treatment with the chemotherapeutic cytarabine compared to those in higher O 2 tensions. Thus, diseased cells growing in lower O 2 niches may be less susceptible to treatment and therefore may be a source of chemoresistance or relapse. These data reveal important roles for O 2 in balancing hematopoietic cell growth with normal and diseased cell function. Manipulating O 2 tension or O 2 sensing pathways could be used to improve HSC/HPC expansion for transplantation or to treat disease. Transcriptomic studies of HSCs/HPCs and AML cells are ongoing to determine the O 2 sensing mechanisms that dictate cellular function.

Identifying Surface Protein Markers for AML Immunotherapy

Background: Acute Myeloid Leukemia (AML) is a hematologic malignancy characterized by clonal expansion and differentiation arrest of myeloid precursor cells in the bone marrow. It results in the accumulation of abnormal and immature myeloblasts, which impairs the production of normal blood cells, including erythrocytes, leukocytes, and platelets. Chimeric Antigen Receptor T-cell therapy (CAR-T therapy) and Antibody-Drug Conjugate (ADC) immunotherapies have shown significant success in the treatment of hematological malignancies. However, their efficacy in AML is limited due to the disease's genetic diversity, lack of uniform antigen targets and immunosuppressive microenvironment. Single-cell RNA sequencing (scRNA-seq) is a powerful technique that can be utilized for surface antigen discovery in various cell types, including immune cells and cancer cells. Methods: To identify potential AML cell-surface markers, we employed a comprehensive approach that involved the utilization of 3' single-cell RNA sequencing (scRNA-seq), bulk RNA sequencing (RNA-seq), and enhanced whole exome sequencing (eWES). A total of 21 samples were analyzed, comprising 18 distinct AML patients and 3 healthy bone marrow aspirates, serving as controls. Results: Using enhanced whole exome sequencing (eWES) on tumor/normal paired samples, we defined the significant driver mutations within our sample cohort, notably DNMT3A (N=4, 21%), FLT3 (N=3, 16%), TET2 (N=3, 16%), KRAS (N=2, 11%), and GATA2 (N=2, 11%). Subsequently, we conducted scRNAseq on 21 samples, yielding a total of 185,106 high-quality cells, which were effectively clustered into 38 unbiased groups. Each sample demonstrated an average capture of 8862 cells, showcasing robust data acquisition. Our scRNAseq analysis revealed a diverse representation of major lymphoid and myeloid lineages, with prominent populations of AML blasts, T/NK cells, and progenitor cells. Notably, tumor clusters exhibited highest copy number variations and formed separate clusters from other non-tumor immune cells. To uncover key functional pathways, we conducted pathway enrichment analysis comparing tumor cells (N=131,196 cells) to non-tumor cells (N=53,910 cells). Our findings demonstrated a significant overrepresentation of cancer hallmarks, such as Allograft rejection (FDR= 0.000002), IL2/STAT5 signaling (FDR= 0.0000005), IFN-gamma response (FDR= 0.00000001) and apoptosis (FDR= 0.00000001) within the AML blasts, suggesting their potential involvement in AML pathogenesis. To identify AML specific markers, we developed analysis pipeline and performed an unbiased systematic search of the scRNAseq dataset for genes with unique expression patterns in tumor cells compared to the healthy bone marrow controls. Our analysis identified 302 differentially expressed genes, with 8 genes displaying relative cell and tissue specificity. Notably, among these genes, we unveiled key potential targets with cell surface specificity, including novel markers alongside previously known ones. Noteworthy candidates, such as CD86, CSF1R, CD33, CLEC12A, and FLT3, were prominently and significantly expressed, showcasing their potential significance as targets for engineered immunotherapy. To validate and consolidate our discoveries, we conducted a cross-referencing exercise by comparing our scRNAseq data against bulk RNA data from primary human AML samples and cell lines. Bulk global proteomics and flow cytometry-based studies are planned to confirm our genetic discoveries. Summary: Our rigorous computational and experimental approach provides valuable insights into tumor-specific cell surface proteins, which hold promise as potential targets for innovative and precision immunotherapeutic interventions in the realm of AML treatment.

Rational Combinatorial Targeting By Adapter CAR T Cells (AdCAR-T) Prevents Antigen Escape in Acute Myeloid Leukemia (AML)

Targeting AML by chimeric antigen receptor T cells (CAR-T) has been challenging due to promiscuous expression of AML-associated antigens on healthy hematopoietic stem and progenitor cells as well as a high degree of inter- and intratumoral heterogeneity. Consequently, we did not yet see the dramatic clinical successes as in B-phenotypic malignancies treated with CD19 or BCMA directed CAR-T. In the present study, we analyzed inter- and intratumoral heterogeneity of AML-associated antigens in 30 primary pediatric AML samples at single cell resolution. We identified CD33, CD38, CD371, IL1RAP and CD123 to be most frequently expressed. Notably, high variability was not only observed across the different patient samples, but also among leukemic cells of the same patient, which indicates that multiplexed targeting approaches will be required to target all leukemic cells. To address this need, we utilized our modular Adapter CAR (AdCAR) platform enabling precise qualitative and quantitative control over CAR-T function. We show highly efficient and target specific activity for newly generated adapter molecules (AMs) against CD33, CD38, CD123, CD135 and CD371, both in vitro and in vivo. Importantly, we prove that inherent intratumoral heterogeneity in antigen expression translates into antigen escape and therapy failure to mono-targeted CAR-T therapy. For the first time, we demonstrate in a PDX model that rational combinatorial targeting by AdCAR-T can cure heterogenic disease. In conclusion, we elucidate the clinical relevance of heterogeneity in antigen expression in pediatric AML and present a novel concept for precision immunotherapy by combinatorial targeting, utilizing the AdCAR platform.

NAMPT Haploinsufficiency Is a Therapeutic Vulnerability in High-Risk Myeloid Malignancies

Monosomy 7 (−7) and del7q are the most common high-risk cytogenetic abnormalities and occur across the spectrum of myeloid disorders. -7/del(7q) also co-occur with other high-risk factors such as mutations or deletion of TP53. Standard, non-specific chemotherapy drugs are less effective in these high-risk patients, resulting in a median survival rate of approximately one year. To identify novel, targeted agents for these neoplasms, we sought “collateral lethal” genes, i.e. essential, druggable genes encoded on chromosome 7 that are haploinsufficient upon their loss of heterozygosity. To this end, we first identified essential genes located on chromosome 7. We mined genome-wide CRISPR/Cas9 screen data from 27 human acute myeloid leukemia (AML) cell lines within the DepMap database and identified 65 chromosome 7-encoded, candidate essential genes in AML. To determine those essential genes with known inhibitors, we intersected these 65 candidates with the Drug-Gene Interaction and canSAR.ai databases. We found that nicotinamide phosphoribosyltransferase ( NAMPT), the rate-limiting enzyme that catalyzes the first step of NAD + biosynthesis from nicotinamide, was one of four essential and druggable targets on chromosome 7. Notably, NAMPT was the only essential gene on chromosome 7 with a first-in-class orally bioavailable inhibitor, KPT 9274, currently in clinical trials (NCT04914845). Moreover, NAMPT DNA copy number, mRNA, and protein levels all negatively correlated with cell line sensitivity to NAMPT loss, implicating NAMPT as a collateral lethal candidate gene. We confirmed that isogenic human AML cell lines CRISPR-engineered to partially reduce NAMPT DNA copy led to reduced NAMPT protein levels, indicating NAMPT is a haploinsufficient gene. Cells with ~50% NAMPT levels were significantly more sensitive to the KPT 9274 treatment, and became less viable and more apoptotic, compared to control locus edited cells. To determine whether TP53 status could alter this vulnerability, we generated cell lines with both TP53 loss and 50% NAMPT loss. We found TP53 mutational status did not attenuate the observed effects of NAMPT inhibition on cell survival. Viability assays with primary patient AML samples confirmed that -7/del(7q) status imparts significantly elevated sensitivity to KPT 9274 treatment than samples with diploid chromosome 7 alleles. In summary, our findings reveal NAMPT heterozygosity as a therapeutic vulnerability in high-risk myeloid neoplasms that warrants clinical follow-up. Further, we provide a framework centered on data-mining for uncovering collateral lethal genes in other cancers with recurrent chromosomal arm level deletions.

AML Inhibits Lipid Uptake By the Liver through an HGF/c-MET Dependent Mechanism Which Results in an Increased Supply of Free Fatty Acid to the Tumor

Acute myeloid leukemia (AML) requires a high metabolic turnover to allow for constant expansion. We and others have previously shown that primary human AML blasts rely on fatty acid (FA)-oxidation for survival and proliferation. (Shafat MS, et al. 2017, Jones CL, et al. 2018). The liver plays a major role in carbohydrate, protein, amino acid, and lipid metabolism (Hodson L, and Gunn PJ. 2019). Given that AML is a FA dependent tumor, we have studied the interactions between AML and the liver, and specifically how these affect lipid metabolism, the availability of pro-tumoral FFA and ultimately the mechanism by which this supports AML growth. To understand the role of lipid metabolism we used two syngeneic models of AML (HOXA9/Meis1 or MN1). When engrafted with AML cells, mice body weight during the latter stages of disease. After sacrifice, the size of epididymal and inguinal fat pads were significantly reduced in mice with AML compared to controls. Mice with AML also had elevated free fatty acid (FFA) and reduced glucose in the plasma. We and others have shown that AML induces lipolysis of white adipose tissue for uptake by blasts and resulting in tumor expansion (Shafat MS, et al. 2017, Ye H, et al. 2016). Since the liver is the master regulator of available FA in the plasma, we isolated livers from AML engrafted mice. Transcriptomics revealed down-regulation of genes involved in FA uptake and metabolism. RT-qPCR of RNA isolated from liver samples confirmed these results showing down-regulation of FA transport proteins, (FABP1, SLC27A2, SLC27A5, and CD36) and FA metabolism genes (CPT1, ACADM, and HMGCS2). Furthermore, primary hepatocytes isolated and cultured with AML conditioned media showed a down-regulation of FA uptake and metabolism genes. These data support the hypothesis that AML secretes a factor that alters lipid metabolism in the liver. To discover the factor that AML secretes, and which regulates FA uptake and metabolism in the liver, a proteome profile of the serum from AML engrafted mice was compared to conditioned media from AML cultured cells. Pathway analysis revealed hepatocyte growth factor (HGF), IL-1B and Granulocyte colony stimulating factor (G-CSF) as possible key factors involved in liver FA metabolism (Jing Y, et al. 2019, Kaibori M, et al. 2002). In vitro gene expressions analysis of primary hepatocytes revealed that addition of HGF to the medium, but not Il-1b or G-CSF, inhibited expression of FA uptake genes. In addition, using Seahorse mitochondrial stress kit we confirmed that primary hepatocytes decreased FA oxidation when cultured with AML conditioned media or HGF. Next, we confirmed that AML derived HGF was responsible for inhibition of hepatocyte FA genes by using a neutralizing antibody to HGF which blocked the transcriptional changes previously observed in hepatocytes cultured in AML conditioned media. This showed that AML derived HGF downregulates hepatocyte FA uptake and metabolism. Finally, to determine the role of AML in the uptake of FA by hepatocytes, cultures of primary murine hepatocytes were pretreated with AML conditioned media or HGF with and without neutralizing antibody. A BODIPY conjugated long chain FA was then added to the culture. Images analysis showed hepatocytes had reduced lipid uptake in the presence of AML condition media or HGF, but this was restored with the addition of HGF neutralizing antibody. Here we report that AML inhibition of lipid uptake in the liver, through an HGF dependent mechanism, increases the availability of plasma FA and forms a fundamental part of the AML disease. Ongoing work to develop our understanding of how AML increases access to FA to aid proliferation, may aid in the development of new therapeutics to target liver metabolism in combination with other AML therapies.