Acid ceramidase inhibition enhances BCL-2 targeting in venetoclax-resistant acute myeloid leukemia via a cytotoxic integrated stress response

Word Count: 192/200 40 Article Word Count (excluding abstract): 4469/4500 41 Main Figure/Table Count: 6/8 42 Supplemental Figure/Table Count: 4 43

Count: 54/60 44 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint

45 Resistance to combination regimens containing the BCL-2 inhibitor venetoclax in acute myeloid leukemia (AML) 46 is a growing clinical challenge for this extensively utilized agent . We previously established the anti-leukemic 47 properties of ceramide, a tumor-suppressive sphingolipid, in AML and demonstrated that upregulated expression 48 of acid ceramidase (AC), a ceramide-neutralizing enzyme, supported leukemic survival and resistance to BH3 49 mimetics. Here, we report the anti- leukemic efficacy and mechanisms of co-targeting AC and BCL-2 in 50 venetoclax-resistant AML. Analysis of the BeatAML dataset revealed a positive relationship between increased 51 AC gene expression and venetoclax resistance. Targeting AC enhanced single-agent venetoclax cytotoxicity 52 and the venetoclax + cytarabine combination in AML cell lines with primary or acquired venetoclax resistance. 53 SACLAC + venetoclax was equipotent to the combination of venetoclax + cytarabine at reducing cell viability 54 when evaluated ex vivo across a cohort of 71 primary AML patient samples. Mechanistically, SACLAC + 55 venetoclax increased ceramide to levels that trigger a cytotoxic integrated stress response (ISR), ISR-mediated 56 NOXA protein upregulation, mitochondrial dysregulation, and caspase-dependent cell death. Collectively, these 57 data demonstrate the efficacy of co-targeting AC and BCL-2 in AML and rationalize targeting AC as a therapeutic 58 approach to overcome venetoclax resistance. 59 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint

60 Acute myeloid leukemia (AML) is a devastating hematologic malignancy with limited treatment options and poor 61 survival outcomes (1). Dysregulated mitochondrial apoptosis supports AML resistance to anti-neoplastic agents 62 (2). Intracellular cell death signals converge in part at the mitochondria where cell death is regulated by the 63 coordinated activity of anti -apoptotic (e.g., BCL-2, MCL-1, BCL-xL) and pro-apoptotic (e.g., BIM, NOXA, BAX, 64 BAK) BCL-2 family proteins (3). A targetable BCL-2 vulnerability was identified in AML and spurred clinical trials 65 evaluating the BCL- 2 inhibitor venetoclax alone or in multiple combinations (4). M odestly active as a 66 monotherapy (5), venetoclax was effective in combination with hypomethylating agents or low-dose cytarabine 67 in early studies, which supported its accelerated approval in 2018 (6, 7). Positive results from the confirmatory 68 phase 3 VIALE-A and VIALE-C clinical trials evaluating venetoclax + azacitidine versus azacitidine alone or 69 venetoclax + low-dose cytarabine versus cytarabine alone supported the regular approval of venetoclax in 2020 70 (8, 9). The adoption of venetoclax-containing regimens to intensive combination chemotherapy was especially 71 beneficial for chemotherapy-ineligible AML patients. Unfortunately, responses to venetoclax -containing 72 regimens are transient and resistance is an unmet clinical challenge (10). Therefore, identifying additional 73 vulnerabilities to enhance venetoclax-containing regimens is paramount. 74 75 Sphingolipids are a heterogene ous class of structural and signaling lipids (11). Important for oncology is the 76 sphingolipid species ceramide, the tumor-suppressive properties of which were initially described in myeloid 77 leukemia as a potential therapeutic vulnerability (12). Ceramides reside at the canonical center of sphingolipid 78 metabolism and are generated via de novo synthesis, higher -order sphingolipid catabolism, or inhibition of 79 ceramide-catabolizing pathways (13). Ceramides contribute to the cytotoxic mechanism of AML therapeutics 80 including cytarabine (14), daunorubicin (15, 16) , BH3 mimetics (17), and FLT3 inhibit ors (18). We previously 81 demonstrated that exogenous ceramide supplementation through ceramide nanoliposomes augmented the 82 efficacy of venetoclax and cytarabine in combination (19). Heightened ceramide depletion is an emerging AML 83 feature due in part to upregulation of acid ceramidase (AC), a lysosomal lipid hydrolase and ceramide-84 catabolizing enzyme (20). Previous work in AML from our group found that AC upregulation supports leukemic 85 survival and resistance to chemotherapy and BH3 mimetics (17, 21) . While g enetic and pharmacologic AC 86 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint inhibition resensitized drug-resistant AML cells to chemotherapy , the effect of AC inhibition on BH3 mimetic 87 efficacy and venetoclax-containing regimens remained unknown. 88 89 In this study, we evaluated AC and BCL-2 co-targeting using cell death assays, western blotting, sphingolipid 90 profiling, and rescue studies to define the efficacy and cytotoxic mechanisms . Co-targeting AC and BCL- 2 91 induced ceramide accumulation and a cytotoxic integrated stress response leading to mitochondrial dysfunction 92 and caspase-dependent cell death. Collectively, this work highlights the importance of ceramide clearance in 93 modulating venetoclax cytotoxicity and provides the rationale for pursuing AC -targeting modalities to augment 94 venetoclax-containing regimens. 95 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint

96 Cell culture 97 Cells were cultured in RPMI-1640 (Corning, NY, USA, #10-040; for MM-6) or IMDM ( ThermoFisher, Waltham, 98 MA, USA, #12440; for MV4-11 variants) supplemented with 10% FBS (VWR, Radnor, PA, USA, #97068–085) 99 and incubated in humified incubators set to 37°C with 5% CO2. MM-6 cells were purchased from DSMZ. HL-60 100 and MV4-11 cells were obtained from ATCC (Manassas, VA, USA). Venetoclax-resistant MV4-11 cells (MV4-11 101 VEN-R) were generated by our group via continued subculture in media containing venetoclax at 0.25 µM. HL-102 60 cells were transduced to express GFP and luciferase as described in the supplemental methods . MV4-11 103 cells expressing YFP and luciferase were kindly gifted by Kenichiro Doi (22). Primary AML patient sample 104 acquisition, processing, and handling are detailed in the supplement. 105 106 Cell viability assays 107 For cell lines: cells were seeded in 96- well flat-bottom plates at 2x10 5 cells/mL and treated at the indicated 108 concentrations in a final volume of 100 μL. Following treatment, 20 μL of [3- (4,5-dimethylthiazol-2-yl)-5-(3-109 carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] (MTS) (Promega, Madison, WI, USA, #G3582) was 110 added and incubated for an additional 3 h. Absorbance was measured at 490 nm with the BioTek Cytation 3 111 plate reader (Biotek, Winooski, VT, USA). Following background subtraction, responses were normalized to 112 vehicle control which was defined as 100% . For patient samples: Single and combination d rug screening f or 113 patient samples was performed as previously described (23). Briefly, 5000 cells were seeded in 384-well plates 114 and treated at the indicated concentrations for the indicated times. Cell viability was determined using the 115 CellTiter-Glo luminescence assay (Promega, #G755B), and luminescence was measured using GloMax 116 Discover (Promega). 117 118 Flow cytometry 119 Cells were seeded at 2.5x10 5-5x105 cells/mL and treated as indicated in 6-well plates or tissue culture flasks . 120 Following treatment, cells were assayed per the manufacturer’s instructions using the Muse Annexin V & Dead 121 Cell Kit ( Cytek, Fremont, CA, USA, #MCH100105) , Muse MitoPotential Kit (Cytek, #MCH100110), or Muse 122 Count & Viability Kit (Cytek, #MCH100102) per the manufacturer’s protocol. For all assays, cells were diluted to 123 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint <5x105 cells/mL and 1000-2000 events were captured using the Cytek Guava Muse Cytometer. Data were 124 analyzed using the Muse software version 1.8.0.3. Each assay identified live, apoptotic, and dead cells. 125 126 Western blotting 127 Cells were seeded at 2.5x10 5-5x105 cells/mL and treated as indicated in 6-well plates or tissue culture flasks. 128 Following treatment, cells were centrifuged at 400xg for 7 min, washed in 1X PBS, and subjected to protein 129 isolation and western blotting as previously described (23). 130 131 Cytotoxicity rescue studies 132 Cells were seeded at 2.5x105-5x105 cells/mL and pretreated as indicated in 6-well plates or tissue culture flasks. 133 Following pretreatment, cells were treated with vehicle, SACLAC, venetoclax, or the combination at the indicated 134 concentrations and times and assayed. 135 136 Sphingolipid profiling 137 Cells were seeded at 2.5x10 5-5x105 cells/mL and treated as indicated in 6-well plates or tissue culture flasks. 138 Following treatment, sphingolipids were extracted, processed, and quantified by LC-MS as previously described 139 (23). 140 141 Quantification and statistical analysis 142 Statistical tests were performed as indicated in figure legends with GraphPad Prism (version 10 .4.2). Dose 143 response curves were generated and analyzed using the Prism function “log(inhibitor) vs. response—variable 144 slope (four parameters) ”. Bliss synergy analyses were performed using SynergyFinder 2.0 as previously 145 described (23) . Experiments were performed in three independent replicates with three or more technical 146 replicates. Unless otherwise stated, a representative experiment is shown with error bars representing mean ± 147 standard deviation (SD). 148 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint

149 We previously established a relationship between AC overexpression and resistance to the BH3 mimetic ABT-150 737 (17). To explore links between AC and the clinically approved BH3 mimetic venetoclax, we analyzed the 151 BeatAML database (24) and identified a positive association between increased AC mRNA expression (gene 152 name ASAH1) and increased venetoclax resistance ( Fig. 1a ). These findings prompted us to study the 153 therapeutic potential of co-targeting AC and BCL-2 in AML cell lines and primary AML patient samples using two 154 previously characterized small molecule AC inhibitors: SACLAC (22) and LCL-805 (23)). 155 156 AC inhibition enhanced BCL-2 sensitivity in venetoclax-resistant AML cells 157 Primary and acquired mechanisms contribute to venetoclax resistance (25). To model both, we evaluated the 158 efficacy of co- targeting AC and BCL- 2 in human AML cell lines with primary (MM-6) or acquired venetoclax 159 resistance (MV4-11 VEN-R; generated through continual venetoclax exposure). Compared to parental MV4-11 160 cells, MV4-11 VEN-R cells were venetoclax resistant and exhibited increased protein levels of the anti-apoptotic 161 proteins MCL-1, BCL-2, and BCL-xL as well as decreased levels of the pro- apoptotic protein BAX (Fig. 1b-c). 162 To test whether the AC targeting enhances venetoclax cytotoxicity, we exposed AML cells to SACLAC and 163 venetoclax at increasing concentrations and assessed cell viability and annexin V/7-AAD staining. SACLAC and 164 venetoclax significantly reduced cell viability (Fig. 1d) and increased the fraction of annexin V-positive cells (Fig. 165 1e) in a concentration -dependent manner in MM-6 and MV4- 11 VEN- R cells . Combination effects were 166 quantified with SynergyFinder 2.0 (26) and yielded Bliss scores indicative of strong synergy ( Fig. 1d-e ). 167 Additional flow cytometry profiling of SACLAC and venetoclax-treated cells demonstrated significantly reduced 168 live cells and increased annexin V positive, 7-AAD positive, and double-positive cells (Fig. 1f). Co-targeting AC 169 and BCL-2 also enhanced cell killing in venetoclax -sensitive human AML cell lines ( Fig. S1). Together, these 170 findings show that AC inhibition significantly enhances venetoclax cytotoxicity in venetoclax -sensitive and -171 resistant AML cells. 172 173 Co-targeting AC and BCL- 2 resulted in ceramide accumulation, caspase activation, and NOXA protein 174 upregulation 175 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint We extended our studies to investigate the cytotoxic mechanism underlying AC and BCL- 2 co- targeting. 176 Ceramide accumulation following inhibition of anti -apoptotic BCL-2 family proteins potentiates apoptotic cell 177 death (27) and ceramide nanoliposomes exacerbated venetoclax cytotoxicity in AML (19) . Thus, we next 178 assessed the impact of AC and BCL- 2 inhibition on intracellular ceramide levels. In both MM-6 and MV4 -11 179 VEN-R, SACLAC treatment increased total intracellular ceramide levels by 4- fold while venetoclax alone 180 increased levels by 1.5 -fold. Ceramides of C16 and C24:1 fatty acid chain length were the primary species 181 increased. Combining SACLAC and venetoclax resulted in significant ly greater ceramide accumulation over 182 vehicle ( MM-6=7.3-fold; MV4 -11 VEN -R=3.7-fold), SACLAC ( MM-6=1.3-fold; MV4- 11 VEN -R=1.5-fold), or 183 venetoclax (MM-6=3.8-fold; MV4-11 VEN-R=2.8-fold) (Fig. 2a). SACLAC treatment elevated hexosylceramide 184 levels in MV4-11 VEN-R cells while the drug combination reduced sphingomyelin levels in MM-6 cells (Fig. S2). 185 These findings demonstrate that AC inhibition and venetoclax treatment cooperate to elevate intracellular 186 ceramide accumulation to drive cytotoxicity. 187 188 We next determined the effects of SACLAC and venetoclax treatment on apoptotic and BCL- 2 family protein 189 levels over a time course. SACLAC enhanced venetoclax cytotoxicity in a time-dependent manner across both 190 cell lines beginning 24 h post treatment as measured by reduced live cell numbers and enhanced annexin V , 191 positive, 7-AAD, and double positive cell percentages (Fig. 2b). Cell death following AC and BCL-2 inhibition 192 was accompanied by increased PARP and caspase-3 cleavage, suggestive of caspase activation (Fig. 2c). To 193 test whether caspase activation was necessary to induce cell death with SACLAC + venetoclax, we pretreated 194 cells with the pan-caspase inhibitor Z-VAD-FMK followed SACLAC, venetoclax, or the combination. Z-VAD-FMK 195 pretreatment protected against venetoclax and SACLAC + venetoclax cell killing in MM-6 cells concomitant with 196 reduced PARP and caspase-3 cleavage (Fig. 2d-e). However, Z-VAD-FMK conferred less protection to SACLAC 197 + venetoclax in MV4- 11 VEN-R cells. Despite increased cell death, anti-apoptotic BCL-2 family protein levels 198 were not diminished. Instead, we observed increased protein levels of the pro-apoptotic BH3-only protein, NOXA, 199 a well-described endogenous MCL -1 inhibitor (Fig. 2c). These results demonstrate that co-targeting AC and 200 BCL-2 enhanced caspase-mediated cell death and upregulated the BH3-only protein NOXA. 201 202 AC and BCL-2 co-targeting induced a cytotoxic integrated stress response 203 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint To better understand how SACLAC modulates venetoclax cytotoxicity, we examined the proteome following 204 SACLAC treatment in MV4-11 cells and identified 249 differentially upregulated and 190 differentially 205 downregulated proteins ( Fig. 3a ). STRINGDb pathway analysis identified p roteins involved in “translation”, 206 “ribonucleoprotein complex biogenesis” , and “ribonucleoprotein complex assembly” Gene Ontology term 207 processes as significantly depleted following SACLAC treatment (Fig. 3b, Table S1). These data are consistent 208 with previous reports that ceramides induce the integrated stress response (ISR), which regulates translation in 209 response to cellular stress (28). ISR activation also sensitizes AML cell lines to venetoclax (28, 29). To evaluate 210 the link between the ISR and co- targeting AC and BCL-2, we measured ISR activation following SACLAC and 211 venetoclax treatment. Central to ISR activation is the phosphorylation of eIF2α at Ser51 (p-eIF2α) by PKR, 212 PERK, GCN2, or HRI, which results in reduced global protein synthesis and upregulation of the transcription 213 factor ATF4 (30). Both SACLAC and venetoclax increased p-eIF2α (Ser51) and ATF4 protein expression in MM-214 6 and MV4- 11 VEN-R cells ( Fig. 3c). Interestingly, SACLAC + venetoclax treatment cooperated to induce a 215 heightened ISR compared to either single agents. Paradoxically, ISR activation has been reported to support 216 (31, 32) and overcome (28, 33, 34) therapy resistance in AML . To determine whether ISR activation was 217 cytoprotective or cytotoxic, we treated with the small molecule ISR inhibitor ISRIB and assessed ISR signaling 218 and cytotoxicity with subsequent SACLAC and venetoclax treatment. ISRIB pretreatment significantly abrogated 219 SACLAC + venetoclax-mediated cytotoxicity and reduced ATF4 protein upregulation ( Fig. 3d-e). We validated 220 our findings with a separate AC inhibitor, LCL- 805, which also enhanced venetoclax -mediated cell killing in a 221 caspase- and ISR-dependent manner (Fig. S3a-c). These findings demonstrate that co-targeting AC and BCL-222 2 induces a cytotoxic ISR. 223 224 Integrated stress response activation promoted NOXA protein upregulation 225 We next studied the relationship between AC inhibition, the ISR, and BCL- 2 family proteins to understand the 226 mechanisms underlying AC and BCL- 2 co -targeting. Mitochondrial outer membrane permeabilization and 227 apoptosis are regulated by pro-apoptotic and anti-apoptotic BCL-2 proteins (3). Anti-apoptotic proteins were not 228 decreased following treatment with SACLAC or the SACLAC + venetoclax combination. Instead, we observed 229 upregulation of the pro-apoptotic BH3-only protein NOXA (Fig. 2c). ISR activation has previously been reported 230 to induce NOXA induction (28, 34, 35), which prompted us to evaluate this link in our model. To test this, we 231 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint pretreated MM-6 and MV4- 11 VEN-R cells with ISRIB and measured NOXA protein following SACLAC and 232 venetoclax treatment. ISRIB pretreatment partially prevented NOXA accumulation following SACLAC + 233 venetoclax treatment compared to cells pretreated with vehicle ( Fig. 3 e). Mechanistically, ISRIB functions 234 downstream of eIF2α phosphorylation by titrating the formation of eIF2B heterodimers to aid 235 eIF2•GTP•methionyl-initiator tRNA ternary complex formation to support translation (30). As an orthogonal 236 approach, we pharmacologically inhibited ISR sensor kinases to determine which pathway was involved in ISR 237 activation following AC inhibition. While the PERK inhibitor, AMG PERK 44, did not blunt SACLAC-induced ATF4 238 or NOXA protein upregulation, GCN2 inhibition with GCN2 -IN-1 significantly reduced ATF4 and NOXA protein 239 upregulation in both cell lines (Fig. 3f). PKR inhibition with PKR-IN-C16 also prevented ATF4 accumulation, 240 though the effect on NOXA was only observed in MV4 -11 VEN-R cells. Consistent with GCN2 involvement in 241 SACLAC-induced ISR , our proteomics data demonstrated significant enrichment of GCN1, an endogenous 242 GCN2 activator involved in the detection of colliding ribosomes (36) and GIGYF2, which is involved in the 243 ribosome-associated quality control pathway to regulate proteostasis (37) (Fig. 3a). Together, these results are 244 consistent with AC inhibition inducing a GCN2- and PKR- mediated ISR leading to NOXA protein upregulation. 245 246 Co-targeting AC and BCL-2 antagonized mitochondrial function 247 Mitochondrial reprogramming is a hallmark of venetoclax resistance (38) and targeting mitochondrial translation 248 (33) or structure (39) enhances venetoclax cytotoxicity. Because AC inhibition also antagonizes mitochondrial 249 function (40), we posited that AC and BCL- 2 inhibition converge at the mitochondria as a combined cytotoxic 250 mechanism. To test this, w e assessed the effect of co- targeting AC and BCL-2 on mitochondrial membrane 251 potential, a biomarker of mitochondrial health. While transient depolarization can be tolerated, long -term 252 depolarization commits cells to apoptotic cell death (41, 42). We treated cells with SACLAC and venetoclax and 253 assessed mitochondrial polarization with tetramethylrhodamine, methyl ester (TMRM), a fluorescent dye that 254 localizes to polarized mitochondria. We observed significant mitochondrial depolarization following SACLAC + 255 venetoclax treatment compared to single agent controls in both MM -6 and MV4-11 VEN-R cells ( Fig. 4a). To 256 determine whether the observed mitochondrial depolarization preceded cell death, we pretreated cells with 257 vehicle or Z -VAD-FMK, treated cells with SACLAC and venetoclax, and assessed mitochondrial polarization. 258 Although Z- VAD-FMK protect ed against SACLAC + venetoclax -induced cytotoxicity ( Fig. 2d), Z -VAD-FMK 259 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint pretreatment did not affect mitochondrial depolarization suggesting that depolarization was upstream of cell 260 death (Fig. 4a). As ISR inhibition protected against SACLAC + venetoclax cytotoxicity (Fig. 3b), we also tested 261 whether ISRIB affected mitochondrial depolarization induced by the combination. ISRIB pretreatment 262 significantly protected against mitochondrial depolarization induced by SACLAC + venetoclax in MM-6 and MV4-263 11 VEN- R cells ( Fig. 4 b). These findings suggest that the cytotoxic mechanism of SACLAC + venetoclax 264 converges at and leads to mitochondrial depolarization, enhanced cell death via the ISR, and downstream 265 caspase activation. 266 267 We next investigated the effect of AC and BCL -2 co- targeting on mitochondrial respiration using a 268 comprehensive mitochondrial diagnostic workflow to assess contributions to the combined cytotoxic mechanism 269 (40, 43). This approach leveraged a modified version of the creatine kinase clamp to titrate extramitochondrial 270 ATP/ADP ratios and mimic physiologically relevant ATP free energies (ΔG ATP). Following AC inhibition or 271 venetoclax treatment, cells were permeabilized with digitonin and oxygen consumption was measured after 272 respiratory stimulation by clamping ΔG ATP at -54 kJ/mol, supplementing cells with mitochondrial electron 273 transport complex -specific substrates, or treating cells with the respiratory uncoupler FCCP. In MM -6 cells, 274 treatment with two AC inhibitors, SACLAC or LCL -805, partially antagonized respiration irrespective of the 275 stimulant (Fig. 4c). Venetoclax alone was sufficient to significantly reduce respiration, which was enhanced by 276 combining venetoclax with either SACLAC or LCL -805 ( Fig. 4 c). Though SACLAC, LCL -805, venetoclax , 277 SACLAC + venetoclax, and LCL-805 + venetoclax inhibited respiration in MV4-11 VEN-R cells, the effects were 278 modest due to the poor respiratory capacity of this cell line . In contrast to MM -6, the stimulants failed to 279 substantially increase mitochondrial respiration in MV4-11 VEN-R cells (Fig. 4d). These findings are consistent 280 with our previous report that blocking mitochondrial respiration is not sufficient to induce AML cell death (44). 281 These findings highlight i) stark differences in mitochondrial respiratory capacity between MM -6 and MV4- 11 282 VEN-R cells, ii) that venetoclax alone was sufficient to reduce mitochondrial respiration to near baseline in MM-283 6 cells and respiration was very low in MV4-11 VEN-R though both cell lines remained viable, and ii i) that the 284 addition of two independent AC inhibitors resulted in additional loss of respiratory capacity beyond that which 285 was observed with venetoclax alone. 286 287 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint AC inhibition enhanced the efficacy of clinically relevant venetoclax and cytarabine combination 288 Venetoclax is FDA-approved in combination with low -dose cytarabine for AML patients ineligible for intensive 289 chemotherapy (45) and the efficacy of this combination is enhanced by exogenous ceramide supplementation 290 in various AML models (19). To explore how AC inhibition augments relevant AML therapeutics, we expanded 291 our study to include drug combinations with cytarabine. MM-6 and MV4-11 VEN-R cells were treated with the 292 three-drug combination of SACLAC + venetoclax + cytarabine, each two-drug permutation, and single agent or 293 vehicle controls. As single agents, SACLAC, venetoclax, and cytarabine modest ly reduced cell viability and 294 increased the fraction of annexin V-positive cells. The two-drug combinations of SACLAC + venetoclax and 295 SACLAC + cytarabine exhibited similar efficacy to venetoclax + cytarabine. Importantly, the triple -drug 296 combinations resulted in significantly decreased cell viability and increased cytotoxicity compared to two-drug 297 combinations or single-agent controls (Fig. 5a-b). These data show that AC inhibition enhanced the cytotoxicity 298 of venetoclax and cytarabine as single agents and significantly augmented the cytotoxicity of the venetoclax + 299 cytarabine combination. 300 301 Co-targeting AC and BCL-2 increased cytotoxicity in primary AML patient samples 302 We extended our efficacy studies beyond cell lines and performed ex vivo drug screening of SACLAC, 303 venetoclax, and cytarabine across a cohort of primary AML patient and healthy control samples. Our AML cohort 304 consisted of peripheral blood mononuclear cells (PBMCs) and bone marrow (BM)-derived cells from 71 patients. 305 We also screened 7 PBMC and 2 BM samples from healthy individuals . We previously reported patient 306 demographic and clinical information for this cohort (23). W e first determined IC50 values for SACLAC, 307 venetoclax, or cytarabine. SACLAC was significantly more toxic toward samples derived from AML patients 308 versus healthy control samples (Fig. 6a). Venetoclax and cytarabine IC50 values were highly heterogenous and 309 the mean IC50 values were lower in AML samples versus healthy controls but not statistically significant (Fig. 6b-310 c). We next tested the SACLAC + venetoclax combination across the AML specimens and healthy control 311 samples. In line with our cell line data, SACLAC + venetoclax treatment exhibited greater efficacy than SACLAC 312 or venetoclax alone. The AC inhibitors SACLAC and LCL-805 also reduced colony formation potential in primary 313 AML patient samples ( Fig. S3d-e). Importantly, SACLAC + venetoclax was significantly more toxic to AML 314 specimens than healthy control samples ( Fig. 6d). We also compared the SACLAC + venetoclax combination 315 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint versus the clinically utilized venetoclax + cytarabine combination and found that they comparably reduced cell 316 viability in patient samples (Fig. 6e). Bliss synergy analysis revealed enhanced synergy scores of the SACLAC 317 + venetoclax combination versus venetoclax + cytarabine for most patient samples ( Fig. 6f). Together, these 318 findings demonstrate promising efficacy of SACLAC + venetoclax at reducing primary AML patient sample cell 319 viability. 320 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint

321 The adoption of venetoclax-containing regimens improved the treatment landscape for AML patients unfit for 322 intensive combination chemotherapy (45). Though initially effective, 30-40% of patients are refractory to 323 treatment and most patients develop resistance (25). Thus, understanding regulators of venetoclax efficacy in 324 AML is of high importance (10, 25). In this work, we demonstrated the efficacy of co- targeting AC and BCL-2 325 across AML cell lines and primary patient samples. Pharmacologic AC inhibition enhanced both the efficacy of 326 venetoclax and venetoclax + cytarabine in venetoclax -resistant AML cell lines. Central to the cytotoxic 327 mechanism were dramatic increases of intracellular ceramides and induction of a cytotoxic ISR leading to NOXA 328 accumulation, mitochondrial depolarization, and caspase-dependent cell death. 329 330 Metabolic rewiring is an emergent AML hallmark, and dysregulated sphingolipid metabolism is becoming 331 recognized as an important feature (20, 46) . Ceramides are tumor-suppressive sphingolipids that play a key 332 cytotoxic role in AML-relevant therapeutics. For example, the anthracycline daunorubicin is used extensively for 333 AML induction therapy and the intracellular ceramide it generates contributes to its cytotoxic mechanism (15, 334 16). This “ceramide-killing effect” is not limited to chemotherapeutic drugs. Elevated ceramide levels, induced by 335 antagonizing ceramide glycosylation or sphingosine kinases, also improve BH3 mimetic efficacy across various 336 human leukemias (47). Despite this, f ew studies have targeted ceramide metabolism to augment venetoclax 337 efficacy. We previously showed that venetoclax treatment in combination with exogenous C6- ceramide 338 nanoliposomes (CNL) enhanced intracellular ceramide accumulation to support the combined cytotoxic 339 mechanism (19). We also previously identified the ceramide-catabolizing lipid hydrolase, AC, as a mediator of 340 BH3 mimetic ABT -737 efficacy in AML (17). However, the link between AC targeting and venetoclax in AML 341 remained unexplored until now. 342 343 Ceramides are hydrolyzed by ceramidases into sphingosine, which serves as the substrate for sphingosine 344 kinase-mediated phosphorylation to S1P. Conversely, S1P is dephosphorylated by S1P phosphatases to 345 generate sphingosine, which is acylated by ceramide synthases to produce ceramides. In the original cell -346 intrinsic rheostat model, increased ceramides and decreased S1P are considered pro-death while decreased 347 ceramide and increased S1P are considered pro-survival (48). Although the rheostat model has been updated 348 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint to reflect new research (e.g., different ceramide species exert distinct functions; inside-out S1P paracrine 349 signaling) (48), therapeutic strategies aimed at increasing total intracellular ceramides or decreasing S1P 350 production are promising strategies for AML as single agents or in combination with venetoclax (22, 28, 49). CNL 351 enhanced venetoclax and cytarabine efficacy in AML (19) and has now entered a phase I clinical trial for 352 relapsed/refractory AML (NCT04716452). Others demonstrated the utility of inhibiting sphingosine kinase to 353 augment venetoclax killing in AML (28). We extend this work by 1) targeting AC, which is upstream of 354 sphingosine kinase in the salvage pathway, 2) evaluating AC and BCL- 2 co-targeting in venetoclax-resistant 355 AML, 3) linking synergy directly to the ISR, 4) extending studies to a large cohort of primary patient samples, 356 and 5) assessing AC inhibition in the context of the FDA -approved venetoclax + cytarabine combination. 357 Collectively, these studies demonstrate that inhibiting flux through the sphingolipid salvage pathway or 358 supplementing with exogenous ceramide are promising approaches to improve venetoclax cytotoxicity in 359 venetoclax-sensitive and -resistant AML cells. Whereas we evaluated sphingolipid changes following SACLAC 360 and venetoclax treatment in these cells, the complete set of sphingolipid metabolic alterations and flux changes 361 in primary and acquired venetoclax resistance is unknown and an open area of study. 362 363 ISR signaling is emerging as a crucial regulator of BH3 mimetic efficacy. Indeed, ISR activation was sufficient to 364 enhance venetoclax toxicity in venetoclax -resistant AML cell lines (33). Moreover, the sphingosine kinase 365 inhibitor MP-A08 sensitized AML cells to venetoclax-induced cytotoxicity via a PKR-mediated cytotoxic ISR (28). 366 We also identified the ISR as a mediator of efficacy between AC and BCL -2 co- targeting through rescue 367 experiments with the ISR inhibitor ISRIB . Our data demonstrate that PKR inhibition blunted ATF4 and NOXA 368 induction following AC inhibition. Ceramides have previously been shown to induce PKR activation and blunt 369 protein synthesis (50). In addition to PKR, our findings highlight an additional link between AC inhibition, 370 ceramides, and ISR signaling through GCN2 as we found that GCN2 inhibition with GCN2 -IN-1 also blocked 371 ATF4 and NOXA induction. Indeed, a past study evaluated ceramide binding partners and predicted binding 372 between ceramide and GCN1 (51), a critical regulator of GCN2 activity (52). In support of this GCN2 model, we 373 observed that SACLAC treatment led to increased protein levels of GCN1 and GIGYF2, which serve to activate 374 GCN2 and respond to ribosomal stress (37). Collectively, these findings reveal a novel relationship between 375 ceramide-induced ribosomal stress and the ISR. 376 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 377 Venetoclax resistance is associated with upregulation of anti-apoptotic proteins such as MCL-1 and BCL-xL (10, 378 25). Unfortunately, BCL-xL inhibition in leukemia is limited by on -target dose-limiting thrombocytopenia (53). 379 Strategies targeting MCL-1 are actively being investigated preclinically and clinically though development may 380 be hindered by cardiac toxicities (54). ISR activation is linked to both MCL -1-dependent and independent 381 mechanisms of improving venetoclax killing. For example, sphingosine kinase inhibition enhanced venetoclax 382 killing through a PKR/ATF4/NOXA/MCL-1 axis (28). In a separate study, targeting mitochondrial translation with 383 antibiotics enhanced venetoclax killing by inhibiting mitochondrial respiration and causing ISR-mediated 384 glycolytic impairment independent of MCL-1 or BCL-xL protein expression changes (33). In our model, MCL-1 385 or BCL-xL protein levels did not decrease in response to AC inhibition and/or venetoclax treatment in MM-6 and 386 MV4-11 VEN-R cells. Instead, cell killing from the combination co- occurred with ISR-mediated upregulation of 387 the endogenous MCL-1 inhibitor, NOXA, and loss of mitochondrial membrane potential . The observation that 388 MM-6 and MV4- 11 VEN-R cells survive acute venetoclax exposure despite reduced mitochondrial respiration 389 upon single-agent venetoclax treatment suggests that AC -mediated enhancement of venetoclax efficacy may 390 engage additional mechanisms beyond further inhibition of mitochondrial respiration. The findings here are 391 consistent with our recent report that venetoclax-resistant AML cells hydrolyze ATP to maintain a mitochondrial 392 membrane potential and resist venetoclax-induced cell death (44). Instead of targeting respiration, targeting 393 mitochondrial membrane potential was key to inducing AML cell death (44). This may suggest that AC inhibition 394 enhances venetoclax cytotoxicity by prevent ing cells from hydrolyzing ATP to maintain a mitochondrial 395 membrane potential and resist BCL-2 targeting. Collectively, the cytotoxic mechanism of AC targeting and BCL-396 2 inhibition likely incorporates both ISR-mediated NOXA protein upregulation and more complex effects on 397 mitochondrial polarization. 398 399 Intensive chemotherapy combining cytarabine and an anthracycline remains the backbone of AML induction 400 therapy (45). We previously reported that cytarabine and daunorubicin selection pressure promotes a pro -401 survival sphingolipid composition that reduces intracellular ceramide levels in part through increased AC activity 402 and protein expression (14). T hese adaptations highlight a potentially exploitable vulnerability to improve 403 chemotherapy efficacy in AML. Indeed, our efficacy studies demonstrated that AC inhibition synergized with 404 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint cytarabine alone. The polychemotherapy consisting of SACLAC + venetoclax + cytarabine was more efficacious 405 than either doublet combination or single agent control. The pro-survival changes in sphingolipid metabolism in 406 chemotherapy-resistant AML also include other pathways that decrease intracellular ceramide levels (14), thus 407 warranting additional studies that evaluate other sphingolipid modulating drugs (e.g., glucosylceramide synthase 408 inhibitors, sphingosine kinase inhibitors, ceramide kinase inhibitors) in the context of venetoclax -containing 409 regimens. 410 411 In summary, we identified AC as a regulator of venetoclax efficacy in venetoclax-resistant AML. Pharmacologic 412 inhibition of AC was sufficient to enhance venetoclax -mediated killing via a cytotoxic ISR and mitochondrial 413 impairment. AC inhibition also increased AML sensitivity to cytarabine and the venetoclax + cytarabine 414 combination. These findings support the continued development of sphingolipid inhibitors to augment approved 415 AML drugs. 416 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint Acknowledgments: The authors would like to acknowledge and thank the patients and their families who 417 supported our studies. We thank Gemma Fabrias for providing SACLAC and Galina Diakova, the UVA Partners 418 in Discovery Team, the UVA Office of Clinical Research Non -Treatment Research Operations, and the UVA 419 Biorepository and Tissue Research Facility in the consenting of patients, specimen procurement, specimen 420 processing, data abstraction, and providing access to molecular and clinical data (IRB-HSR #18445). This work 421 was supported by the Veteran’s Administration (VA Merit Review, I BX001792 [to CEC]) and a Research Career 422 Scientist Award, IK6BX004603 [to CEC]) and the National Institutes of Health (NIH) award numbers P01 423 CA171983 (to TPL and CEC), Cancer Center Support Grant P30 CA044579 (to TPL), R01 AI139072 (to CEC), 424 F31 CA271809 (to JU), and F99 CA284252 (to JU) This work was also supported by the UVA Robert R. Wagner 425 Fellowship (to JU). The content is solely the responsibility of the authors and does not necessarily represent the 426 official views of the National Institutes of Health, Department of Veterans Affairs, or the United States 427 Government. This work is dedicated in loving memory to the late sphingolipid pioneer Dr. Mark Kester. 428 429 Contributions: JU, SFT, DJF, and TPL conceptualized the study design. JU, SFT, JJPS, MT, TMD, GDCV, 430 MMM, JTH, RTA, URG, AS, BBP, ISL, BR, KFW, and TEF were responsible for experimental work, data 431 collection, and data analysis. JU, KFW, and BBP performed proteomic, statistical, and bioinformatic analyses. 432 KAJ, FGB, MCC, and DFC provided scientific resources and subject matter expertise. DJF and TPL provided 433 project oversight. JU, CEC, and TPL were responsible for funding acquisition. JU drafted the original manuscript 434 draft. All co-authors provided edits, reviewed, and approved the final manuscript. 435 436 Competing Interests: DJF has received research funding, honoraria, and/or stock options from AstraZeneca, 437 Dren Bio, Recludix Pharma, and Kymera Therapeutics. TPL has received Scientific Advisory Board membership, 438 consultancy fees, honoraria, and/or stock options from Keystone Nano, Flagship Labs 86, Dren Bio, Recludix 439 Pharma, Kymera Therapeutics, and Prime Genomics. MCC owns shares in Keystone Nano. Ceramide 440 nanoliposome (CNL) development has been licensed to Keystone Nano. There are no conflicts of interest with 441 the work presented in this manuscript. Other authors declare no competing interests. The funders had no role in 442 the design of this study; in the collection, analyses, or interpretation of data; in the writing of this manuscript; or 443 in the decision to publish the results. 444 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 445 Data Availability Statement: Data needed to evaluate conclusions from this work are present in the paper. 446 Molecular and clinical characteristics from the primary AML patient samples were previously reported (23). All 447 proteomics data will be deposited in ProteomeXchange six months after publication. Summary data for pathway 448 analysis is available in Table S1. Other datasets generated during and/or analyzed during the current study are 449 available from the corresponding author on reasonable request. 450 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint

451 1. Kantarjian H, Borthakur G, Daver N, DiNardo CD, Issa G, Jabbour E, et al. Current status and research 452 directions in acute myeloid leukemia. Blood Cancer Journal. 2024;14(1):163. 453 2. Krawiec K, Strzałka P, Czemerska M, Wiśnik A, Zawlik I, Wierzbowska A, et al. Targeting Apoptosis in 454 AML: Where Do We Stand? Cancers [Internet]. 2022; 14(20). 455 3. Kale J, Osterlund EJ, Andrews DW. BCL-2 family proteins: changing partners in the dance towards death. 456 Cell Death & Differentiation. 2018;25(1):65-80. 457 4. Pan R, Hogdal LJ, Benito JM, Bucci D, Han L, Borthakur G, et al. Selective BCL-2 Inhibition by ABT-199 458 Causes On-Target Cell Death in Acute Myeloid Leukemia. Cancer Discovery. 2014;4(3):362-75. 459 5. Konopleva M, Pollyea DA, Potluri J, Chyla B, Hogdal L, Busman T, et al. Efficacy and Biological 460 Correlates of Response in a Phase II Study of Venetoclax Monotherapy in Patients with Acute Myelogenous 461 Leukemia. Cancer Discovery. 2016;6(10):1106-17. 462 6. DiNardo CD, Pratz K, Pullarkat V, Jonas BA, Arellano M, Becker PS, et al. Venetoclax combined with 463 decitabine or azacitidine in treatment-naive, elderly patients with acute myeloid leukemia. Blood. 2019;133(1):7-464 17. 465 7. Wei AH, Strickland SA, Hou J -Z, Fiedler W, Lin TL, Walter RB, et al. Venetoclax Combined With Low -466 Dose Cytarabine for Previously Untreated Patients With Acute Myeloid Leukemia: Results From a Phase Ib/II 467 Study. Journal of Clinical Oncology. 2019;37(15):1277-84. 468 8. Wei AH, Montesinos P, Ivanov V, DiNardo CD, Novak J, Laribi K, et al. Venetoclax plus LDAC for newly 469 diagnosed AML ineligible for intensive chemotherapy: a phase 3 randomized placebo -controlled trial. Blood. 470 2020;135(24):2137-45. 471 9. DiNardo Courtney D, Jonas Brian A, Pullarkat V, Thirman Michael J, Garcia Jacqueline S, Wei Andrew 472 H, et al. Azacitidine and Venetoclax in Previously Untreated Acute Myeloid Leukemia. New England Journal of 473 Medicine. 2020;383(7):617-29. 474 10. Nwosu GO, Ross DM, Powell JA, Pitson SM. Venetoclax therapy and emerging resistance mechanisms 475 in acute myeloid leukaemia. Cell Death & Disease. 2024;15(6):413. 476 11. Hannun YA, Obeid LM. Sphingolipids and their metabolism in physiology and disease. Nature Reviews 477 Molecular Cell Biology. 2018;19(3):175-91. 478 12. Obeid LM, Linardic CM, Karolak LA, Hannun YA. Programmed Cell Death Induced by Ceramide. 479 Science. 1993;259(5102):1769-71. 480 13. Morad SAF, Cabot MC. Ceramide- orchestrated signalling in cancer cells. Nature Reviews Cancer. 481 2013;13(1):51-65. 482 14. Kao L- P, Morad SAF, Davis TS, MacDougall MR, Kassai M, Abdelmageed N, et al. Chemotherapy 483 selection pressure alters sphingolipid composition and mitochondrial bioenergetics in resistant HL- 60 cells. 484 Journal of Lipid Research. 2019;60(9):1590-602. 485 15. Bose R, Verheij M, Haimovitz-Friedman A, Scotto K, Fuks Z, Kolesnick R. Ceramide synthase mediates 486 daunorubicin-induced apoptosis: An alternative mechanism for generating death signals. Cell. 1995;82(3):405 -487 14. 488 16. Jaffrézou JP, Levade T, Bettaïeb A, Andrieu N, Bezombes C, Maestre N, et al. Daunorubicin‐ induced 489 apoptosis: triggering of ceramide generation through sphingomyelin hydrolysis. The EMBO Journal. 490 1996;15(10):2417-24-24. 491 17. Tan S-F, Liu X, Fox TE, Barth BM, Sharma A, Turner SD, et al. Acid ceramidase is upregulated in AML 492 and represents a novel therapeutic target. Oncotarget; Vol 7, No 50. 2016. 493 18. Dany M, Gencer S, Nganga R, Thomas RJ, Oleinik N, Baron KD, et al. Targeting FLT3- ITD signaling 494 mediates ceramide-dependent mitophagy and attenuates drug resistance in AML. Blood. 2016;128(15):1944-495 58. 496 19. Khokhlatchev AV, Sharma A, Deering TG, Shaw JJP, Costa -Pinheiro P, Golla U, et al. Ceramide 497 nanoliposomes augment the efficacy of venetoclax and cytarabine in models of acute myeloid leukemia. The 498 FASEB Journal. 2022;36(10):e22514. 499 20. Ung J, Tan S -F, Fox TE, Shaw JJP, Vass LR, Costa -Pinheiro P, et al. Harnessing the power of 500 sphingolipids: Prospects for acute myeloid leukemia. Blood Reviews. 2022;55:100950. 501 21. Tan S-F, Dunton W, Liu X, Fox TE, Morad SAF, Desai D, et al. Acid ceramidase promotes drug resistance 502 in acute myeloid leukemia through NF- kappaB-dependent P -glycoprotein upregulation. Journal of Lipid 503 Research. 2019;60(6):1078-86. 504 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 22. Pearson JM, Tan S-F, Sharma A, Annageldiyev C, Fox TE, Abad JL, et al. Ceramide Analogue SACLAC 505 Modulates Sphingolipid Levels and MCL-1 Splicing to Induce Apoptosis in Acute Myeloid Leukemia. Molecular 506 Cancer Research. 2020;18(3):352-63. 507 23. Ung J, Tan S -F, Fox TE, Shaw JJP, Taori M, Horton BJ, et al. Acid Ceramidase Inhibitor LCL- 805 508 Antagonizes Akt Signaling and Promotes Iron- Dependent Cell Death in Acute Myeloid Leukemia. Cancers 509 [Internet]. 2023; 15(24). 510 24. Bottomly D, Long N, Schultz AR, Kurtz SE, Tognon CE, Johnson K, et al. Integrative analysis of drug 511 response and clinical outcome in acute myeloid leukemia. Cancer Cell. 2022;40(8):850-64.e9. 512 25. Dhakal P, Bates M, Tomasson MH, Sutamtewagul G, Dupuy A, Bhatt VR. Acute myeloid leukemia 513 resistant to venetoclax-based therapy: What does the future hold? Blood Reviews. 2023;59:101036. 514 26. Ianevski A, Giri AK, Aittokallio T. SynergyFinder 2.0: visual analytics of multi-drug combination synergies. 515 Nucleic Acids Research. 2020;48(W1):W488-W93. 516 27. Beverly Levi J, Howell Lauren A, Hernandez-Corbacho M, Casson L, Chipuk Jerry E, Siskind Leah J. 517 BAK activation is necessary and sufficient to drive ceramide synthase- dependent ceramide accumulation 518 following inhibition of BCL2-like proteins. Biochemical Journal. 2013;452(1):111-9. 519 28. Lewis AC, Pope VS, Tea MN, Li M, Nwosu GO, Nguyen TM, et al. Ceramide- induced integrated stress 520 response overcomes Bcl-2 inhibitor resistance in acute myeloid leukemia. Blood. 2022;139(26):3737-51. 521 29. Hua L, Yang N, Li Y, Huang K, Jiang X, Liu F, et al. Metformin sensitizes AML cells to venetoclax through 522 endoplasmic reticulum stress—CHOP pathway. British Journal of Haematology. 2023;202(5):971 -84. 523 30. Costa-Mattioli M, Walter P. The integrated stress response: From mechanism to disease. Science. 524 2020;368(6489):eaat5314. 525 31. Jin P, Wang X, Jin Q, Zhang Y, Shen J, Jiang G, et al. Mutant U2AF1 -Induced Mis-Splicing of mRNA 526 Translation Genes Confers Resistance to Chemotherapy in Acute Myeloid Leukemia. Cancer Research. 527 2024;84(10):1583-96. 528 32. Williams MS, Amaral FMR, Simeoni F, Somervaille TCP. A stress-responsive enhancer induces dynamic 529 drug resistance in acute myeloid leukemia. The Journal of Clinical Investigation. 2020;130(3):1217-32. 530 33. Sharon D, Cathelin S, Mirali S, Di Trani JM, Yanofsky DJ, Keon KA, et al. Inhibition of mitochondrial 531 translation overcomes venetoclax resistance in AML through activation of the integrated stress response. 532 Science Translational Medicine. 2019;11(516):eaax2863. 533 34. Jin S, Cojocari D, Purkal JJ, Popovic R, Talaty NN, Xiao Y, et al. 5- Azacitidine Induces NOXA to Prime 534 AML Cells for Venetoclax-Mediated Apoptosis. Clinical Cancer Research. 2020;26(13):3371-83. 535 35. Weller S, Toennießen A, Schaefer B, Beigl T, Muenchow A, Böpple K, et al. The BCL -2 inhibitor ABT-536 199/venetoclax synergizes with proteasome inhibition via transactivation of the MCL- 1 antagonist NOXA. Cell 537 Death Discovery. 2022;8(1):215. 538 36. Ryoo HD. The integrated stress response in metabolic adaptation. Journal of Biological Chemistry. 539 2024;300(4). 540 37. Vind AC, Genzor AV, Bekker -Jensen S. Ribosomal stress -surveillance: three pathways is a magic 541 number. Nucleic Acids Research. 2020;48(19):10648-61. 542 38. Guièze R, Liu VM, Rosebrock D, Jourdain AA, Hernández -Sánchez M, Martinez Zurita A, et al. 543 Mitochondrial Reprogramming Underlies Resistance to BCL-2 Inhibition in Lymphoid Malignancies. Cancer Cell. 544 2019;36(4):369-84.e13. 545 39. Chen X, Glytsou C, Zhou H, Narang S, Reyna DE, Lopez A, et al. Targeting Mitochondrial Structure 546 Sensitizes Acute Myeloid Leukemia to Venetoclax Treatment. Cancer Discovery. 2019;9(7):890-909. 547 40. Fisher-Wellman KH, Hagen JT, Neufer PD, Kassai M, Cabot MC. On the nature of ceramide-mitochondria 548 interactions – Dissection using comprehensive mitochondrial phenotyping. Cellular Signalling. 2021;78:109838. 549 41. Gottlieb E, Armour SM, Harris MH, Thompson CB. Mitochondrial membrane potential regulates matrix 550 configuration and cytochrome c release during apoptosis. Cell Death & Differentiation. 2003;10(6):709-17. 551 42. Zorova LD, Popkov VA, Plotnikov EY, Silachev DN, Pevzner IB, Jankauskas SS, et al. Mitochondrial 552 membrane potential. Analytical Biochemistry. 2018;552:50-9. 553 43. Fisher-Wellman KH, Davidson MT, Narowski TM, Lin C- T, Koves TR, Muoio DM. Mitochondrial 554 Diagnostics: A Multiplexed Assay Platform for Comprehensive Assessment of Mitochondrial Energy Fluxes. Cell 555 Reports. 2018;24(13):3593-606.e10. 556 44. Hagen JT, Montgomery MM, Aruleba RT, Chrest BR, Krassovskaia P, Green TD, et al. Acute myeloid 557 leukemia mitochondria hydrolyze ATP to support oxidative metabolism and resist chemotherapy. Science 558 Advances. 2025;11(15):eadu5511. 559 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 45. Döhner H, Wei AH, Appelbaum FR, Craddock C, DiNardo CD, Dombret H, et al. Diagnosis and 560 management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. 561 Blood. 2022;140(12):1345-77. 562 46. Mishra SK, Millman SE, Zhang L. Metabolism in acute myeloid leukemia: mechanistic insights and 563 therapeutic targets. Blood. 2023;141(10):1119-35. 564 47. Casson L, Howell L, Mathews LA, Ferrer M, Southall N, Guha R, et al. Inhibition of Ceramide Metabolism 565 Sensitizes Human Leukemia Cells to Inhibition of BCL2-Like Proteins. PLOS ONE. 2013;8(1):e54525. 566 48. Newton J, Lima S, Maceyka M, Spiegel S. Revisiting the sphingolipid rheostat: Evolving concepts in 567 cancer therapy. Experimental Cell Research. 2015;333(2):195-200. 568 49. Powell JA, Lewis AC, Zhu W, Toubia J, Pitman MR, Wallington-Beddoe CT, et al. Targeting sphingosine 569 kinase 1 induces MCL1-dependent cell death in acute myeloid leukemia. Blood. 2017;129(6):771-82. 570 50. Ruvolo PP, Gao F, Blalock WL, Deng X, May WS. Ceramide Regulates Protein Synthesis by a Novel 571 Mechanism Involving the Cellular PKR Activator RAX. Journal of Biological Chemistry. 2001;276(15):11754-8. 572 51. Kota V, Szulc ZM, Hama H. Identification of C6- ceramide-interacting proteins in D6P2T Schwannoma 573 cells. PROTEOMICS. 2012;12(13):2179-84. 574 52. Masson GR. Towards a model of GCN2 activation. Biochemical Society Transactions. 2019;47(5):1481-575 8. 576 53. Wei Y, Cao Y, Sun R, Cheng L, Xiong X, Jin X, et al. Targeting Bcl-2 Proteins in Acute Myeloid Leukemia. 577 Frontiers in Oncology. 2020;10. 578 54. Tantawy SI, Timofeeva N, Sarkar A, Gandhi V. Targeting MCL- 1 protein to treat cancer: opportunities 579 and challenges. Frontiers in Oncology. 2023;13. 580 581 582 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint Figure Legends: 583 584 Fig. 1: Pharmacologic inhibition of acid ceramidase enhanced venetoclax cytotoxicity in venetoclax -585 resistant acute myeloid leukemia cell lines 586 A Correlation analyses comparing venetoclax area under the curve (AUC) values and ASAH1 mRNA expression 587 (n=332) from BeatAML2.0 (http://www.vizome.org/aml2/). Dotted lines represent linear regression with 95% 588 confidence interval. B MV4-11 Parental and MV4-11 VEN-R cells were subjected to increasing concentrations 589 of venetoclax for 48 h and cell viability was assessed by MTS. Dotted lines represent 95% confidence interval. 590 C Immunoblotting of MV4-11 Parental and MV4-11 VEN-R cells. β-actin expression served as the loading control 591 for immunoblots. D, E MTS cell viability, flow cytometry annexin V staining, and Bliss synergy scores 592 (SynergyFinder 2.0) were measured following DMSO, SACLAC, venetoclax, or combination treatment at the 593 indicated concentrations in MM -6 (48 h) and MV4- 11 VEN-R cells (72 h). Bliss scores greater than 10 were 594 considered synergistic, between -10 and 10 were considered additive, and less than -10 were considered 595 antagonistic. F MM-6 and MV4- 11 VEN -R cells were treated with DMSO, SACLAC, venetoclax, or the 596 combination for 48 h or 72 h, respectively, and annexin V and 7- AAD population percentages were measured 597 by flow cytometry. Significance was assessed by two- way ANOVA with Tukey’s multiple comparisons test 598 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint comparing live cells (annexin− 7- AAD−) across treatment groups. Data are presented as mean ± SD of three 599 independent experiments. ABT-199 = venetoclax (VEN). Concentrations in brackets are µM. ****p<0.0001, ns = 600 not significant. 601 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 602 Fig. 2: SACLAC and venetoclax induced ceramide accumulation and caspase- mediated cell death 603 characterized by NOXA accumulation 604 A MM-6 and MV4 -11 VEN-R cells were treated with DMSO, SACLAC, venetoclax, or the combination, and 605 ceramide levels were measured by liquid chromatography -mass spectrometry. Significance was assessed by 606 two-way ANOVA with Tukey’s multiple comparisons test. B, C Flow cytometry annexin V and 7 -AAD profiling 607 and immunoblotting of MM -6 and MV4- 11 VEN -R cells treated with DMSO, SACLAC, venetoclax, or the 608 combination at the indicated concentrations for 12, 24, 48, and 72 h. Quantitative data are the mean of three 609 independent experiments. Significance was assessed by two- way ANOVA with Tukey’s multiple comparisons 610 test comparing live cells (annexin− 7 -AAD−) across treatment groups. D Flow cytometry annexin V profiling of 611 cells pretreated with DMSO or Z-VAD-FMK for 2 h followed by DMSO, SACLAC, venetoclax, or the combination 612 treatment for 48 h. Significance was assessed by unpaired t -test with Welch’s correction. E Immunoblotting of 613 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint MM-6 cells pretreated with DMSO or Z- VAD-FMK for 2 h followed by DMSO, SACLAC, venetoclax, or the 614 combination treatment for 48 h. β -actin expression served as the loading control for immunoblots. Data are 615 presented as mean ± SD. SAC = SACLAC; VEN = venetoclax; Z-VAD = Z-VAD-FMK. Concentrations in brackets 616 are µM. *p<0.05, **p<0.01, ***p<0.005, ****p<0.0001, ns = not significant. 617 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 618 Fig. 3: SACLAC and venetoclax upregulated NOXA via a cytotoxic integrated stress response 619 A Volcano plot of significantly enriched (n=249) or depleted (n=190) proteins from SACLAC versus DMSO -620 treated MV4- 11 cells. Significance was defined using an FDR (q)<0.1 cutoff. B Biological Process (Gene 621 Ontology) analysis of significantly downregulated proteins from panel A. C Immunoblotting of cells treated with 622 DMSO, SACLAC, venetoclax, or the combination for 12 or 24 h. D Flow cytometry annexin V profiling of cells 623 pretreated with DMSO or ISRIB for 2 h followed by DMSO, SACLAC, venetoclax, or the combination treatment 624 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint for 48 h. Significance was assessed by unpaired t -test with Welch’s correction. E Immunoblotting of cells 625 pretreated with DMSO or ISRIB for 2 h followed by DMSO, SACLAC, venetoclax, or combination treatment for 626 24 or 48 h. F Immunoblotting of cells pretreated with DMSO, PKR-IN-C16 (PKR inhibitor), AMG PERK 44 (PERK 627 inhibitor), or GCN2 -IN-1 (GCN2 inhibitor) for 2 h followed by DMSO or SACLAC treatment for 24 h. β -actin 628 expression served as the loading control for immunoblots. Data are presented as mean ± SD. SAC = SACLAC; 629 VEN = venetoclax. Concentrations in brackets are µM. *p<0.05, ***p<0.005, ns = not significant. 630 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 631 Fig. 4: Effect of AC and BCL-2 inhibition on mitochondrial function and membrane potential 632 A, B MM-6 and MV4-11 VEN-R cells were pretreated with DMSO, Z -VAD-FMK, or ISRIB for 2 h then treated 633 with DMSO, SACLAC, venetoclax, or the combination for 48 h. Mitochondrial membrane potential was assessed 634 by flow cytometry. Significance was assessed by unpaired t-test with Welch’s correction. C, D MM-6 and MV4-635 11 VEN-R cells were treated with DMSO, SACLAC, LCL-805, venetoclax, SACLAC + venetoclax, or LCL-805 + 636 venetoclax. Oxygen consumption was measured in digitonin (0.02 mg/mL) permeabilized cells in the absence of 637 substrates (basal), with ATP free energy clamped at -54 kJ/mol (ΔGATP-54), or following the addition of complex 638 I substrates (octanoyl -carnitine [Oct -Carn], glutamate [Glut]), complex II substrates (succinate [Succ]), the 639 respiratory uncoupler FCCP [FC], or the complex I inhibitor rotenone. Note different y -axis ranges in panels C 640 versus D. Data are presented as mean ± SD. SAC = SACLAC; LCL = LCL- 805; VEN = venetoclax. 641 Concentrations in brackets are µM. *p<0.05, **p<0.01, ***p<0.005, ****p<0.0001, ns = not significant. 642 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 643 Fig 5: Acid ceramidase inhibition augmented the cytotoxicity of venetoclax and cytarabine 644 A MM-6 and MV4 -11 VEN -R cells were treated with DMSO, SACLAC, venetoclax, cytarabine, doublet 645 combinations, and triplet combinations at the indicated concentrations for 48 h then MTS cell viability was 646 assessed. B Flow cytometry annexin V profiling of MM-6 and MV4-11 VEN-R cells treated with DMSO, SACLAC, 647 venetoclax, cytarabine, doublet combinations, and triplet combinations at the indicated concentrations for 48 h. 648 Significance was assessed by Welch’s ANOVA with Dunnet’s multiple comparisons test. Data are presented as 649 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint mean ± SD. SAC = SACLAC; VEN = venetoclax; Ara -C = cytarabine. Concentrations in brackets are µM. 650 *p<0.05, **p<0.01, ***p<0.005, ****p<0.0001, ns = not significant. 651 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 652 Fig. 6: SACLAC enhanced venetoclax cytotoxicity in primary acute myeloid leukemia patient samples 653 A-C Primary patient samples (n=71) and healthy PBMC (n=7) or BM (n=2) control samples were treated with 654 increasing concentrations of SACLAC, venetoclax, or cytarabine for 48 h after which cell viability was assessed 655 by CellTiterGlo. IC 50 values were determined for each drug using the GraphPad Prism “log(inhibitor) vs. 656 normalized response -- Variable slope” function. Dotted line represents the highest concentration tested and 657 samples on the dotted line did not reach IC 50 using tested concentrations. Significance was assessed with the 658 Mann-Whitney test. D Primary patient samples and healthy PBMC or BM control samples were treated with 659 DMSO, SACLAC, venetoclax, or the combination for 48 h after which cell viability was assessed by CellTiterGlo. 660 Significance was assessed by unpaired t-test with Welch’s correction. E, F Primary patient samples were treated 661 with DMSO, SACLAC, cytarabine, venetoclax, or doublet combinations for 48 h after which cell viability was 662 assessed by CellTiterGlo. Significance was assessed by unpaired t-test with Welch’s correction (E) or Wilcoxon 663 matched-pairs signed rank test (F). Bliss synergy scores were calculated using SynergyFinder 2.0. Bliss scores 664 greater than 10 were considered synergistic, between - 10 and 10 were considered additive, and less than - 10 665 were considered antagonistic. Healthy control bone marrow samples were denoted with an open triangle and all 666 other healthy control samples were PBMC-derived. Data are presented as mean ± SD. SAC = SACLAC; VEN = 667 venetoclax; Ara-C = cytarabine. Concentrations in brackets are µM. *p<0.05, **p<0.01, ***p<0.005, ****p<0.0001, 668 ns = not significant. 669 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 670 Fig. S1: Pharmacologic inhibition of acid ceramidase enhanced venetoclax cytotoxicity in venetoclax -671 sensitive acute myeloid leukemia cell lines 672 HL-60 Luc2GFP (A) and MV4-11 Luc2YFP (B) cells were treated with increasing concentrations of SACLAC, 673 venetoclax, or the combination for 48 h and cell viability was assessed by MTS. Bliss synergy scores were 674 calculated using SynergyFinder 3.0. Bliss scores greater than 10 were considered synergistic, between -10 and 675 10 were considered additive, and less than - 10 were considered antagonistic. Data are presented as mean ± 676 SD. 677 678 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 679 Fig. S2: Effect of co-targeting acid ceramidase and BCL-2 on hexosylceramide and sphingomyelin levels 680 MM-6 and MV4-11 VEN-R cells were treated with DMSO, SACLAC, venetoclax, or the combination for 24 h, and 681 hexosylceramide (A) or sphingomyelin (B) levels were measured by liquid chromatography-mass spectrometry. 682 Significance was assessed by two- way ANOVA with Tukey’s multiple comparisons test . *p<0.05, **p<0.01, 683 ***p<0.005, ****p<0.0001, ns = not significant. 684 685 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint 686 Fig. S3: LCL- 805 enhanced venetoclax- mediated cytotoxicity via caspase activation and a cytotoxic 687 integrated stress response 688 A MM-6 and MV4-11 VEN-R cells were treated with DMSO, LCL-805, venetoclax, or the combination for 24 h or 689 72 h, respectively, and cell viability was assessed by MTS. Bliss synergy scores were calculated using 690 SynergyFinder 2.0. Bliss scores greater than 10 were considered synergistic, between - 10 and 10 were 691 considered additive, and less than -10 were considered antagonistic. B Immunoblotting of MM-6 cells treated 692 with DMSO or LCL -805 at the indicated concentrations and times. β -actin expression served as the loading 693 control for immunoblots. C Flow cytometry profiling of annexin V/7- AAD negative cells pretreated MM -6 cells 694 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint were pretreated with DMSO, Z -VAD-FMK, or ISRIB for 2 h followed by DMSO, LCL- 805, venetoclax, or 695 combination treatment for 24 h. Significance was assessed with unpaired Welch’s t-test with Holm -Šídák 696 correction. D Clonogenic capacity of primary AML patient samples treated with DMSO, SACLAC, LCL- 805, 697 venetoclax, SACLAC + venetoclax, or LCL-805 + venetoclax at the indicated concentrations. Significance was 698 assessed with unpaired Welch’s t-test with Holm-Šídák correction. E Clinical data for patient samples 1265 and 699 1341. Data are presented as mean ± SD. SAC = SACLAC; ABT- 199 = venetoclax (VEN). Concentrations in 700 brackets are µM. *p<0.05, **p<0.01, ****p<0.0001, ns = not significant. 701 702 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprint (whichthis version posted June 7, 2025. ; https://doi.org/10.1101/2025.06.06.657881doi: bioRxiv preprint