ALBERT

Description: Bortezomib (BTZ) was recently evaluated in a randomized Phase 3 clinical trial which compared standard chemotherapy (cytarabine, daunorubicin, etoposide; ADE) to standard therapy with BTZ (ADEB) for de novo pediatric acute myeloid leukemia. While the study concluded that BTZ did not improve outcome overall, we examined patient subgroups benefitting from BTZ-containing chemotherapy using proteomic analyses. The proteasome inhibitor BTZ disrupts protein homeostasis and activates cytoprotective heat shock responses. We measured total heat shock factor 1 (HSF1) and phosphorylated HSF1 (HSF1-pSer326) in leukemic cells from 483 pediatric patients using Reverse Phase Protein Arrays. HSF1-pSer326 phosphorylation was significantly lower in pediatric AML compared to CD34+ non-malignant cells. We identified a strong correlation between HSF1-pSer326 expression and BTZ sensitivity. BTZ significantly improved outcome of patients with low-HSF1-pSer326 with a 5-year event-free survival of 44% (ADE) vs. 67% for low-HSF1-pSer326 treated with ADEB (P=0.019). To determine the effect of HSF1 expression on BTZ potency in vitro, cell viability with HSF1 gene variants that mimicked phosphorylated (S326A) and non-phosphorylated (S326E) HSF1-pSer326 were examined. Those with increased HSF1 phosphorylation showed clear resistance to BTZ vs. those with wild type or reduced HSF1-phosphorylation. We hypothesize that HSF1-pSer326 expression could identify patients that benefit from BTZ-containing chemotherapy.

Description: Our group and others have shown that acute myeloid leukemia (AML) cells have unique mitochondrial characteristics with an increased reliance on oxidative phosphorylation. Through an shRNA screen for new biological vulnerabilities in the mitochondria of AML cells, we identified the mitochondrial protease, neurolysin (NLN). NLN is a zinc metalloprotease whose mitochondrial function is not well understood and whose role in AML growth and viability has not been previously reported. We analyzed the expression of NLN in AML cells and normal hematopoietic cells. By immunoblotting, NLN was overexpressed in 80% of primary AML patient samples compared to normal hematopoietic cells. Likewise, in an analysis of gene expression databases, NLN mRNA was increased in a subset of AML patient samples, compared to normal hematopoietic cells. Next, we assessed the effects of knocking down NLN in AML cell lines (OCI-AML2, NB4, and MV4-11) using three independent shRNAs in lentiviral vectors. Target knockdown was confirmed by immunoblotting. NLN knockdown reduced growth in all three tested cell lines by 50-70%. NLN knockdown also targeted the leukemia initiating cells in vitro and in vivo as NLN knockdown reduced the clonogenic growth of AML cells (40-75%) and the engraftment of TEX cells into immune deficient mice by 85%. Taken together, these data suggest that NLN is necessary for the growth of AML cells. The role of NLN in the mitochondria is not well understood. To gain insight into NLN's mitochondrial function, we investigated NLN's protein interactors using proximity-dependent biotin labeling (BioID). The top hits in the protein-protein interaction screen were mitochondrial matrix proteins and respiratory chain subunits were particularly enriched. Therefore, we measured the effects of NLN knockdown on mitochondrial structure and function. Knockdown of NLN in AML cells reduced basal oxygen consumption without altering reactive oxygen species generation, mitochondrial membrane potential, or mitochondrial mass. No changes were seen in the total levels of respiratory chain complex subunits as measured by immunoblotting on denaturing gels. Respiratory chain complexes assemble into higher order supercomplex structures that maintain the integrity of the mitochondria and promote efficient oxidative metabolism. Therefore, we tested whether NLN is required for the formation of respiratory chain supercomplexes. As measured by blue native polyacrylamide gel electrophoresis, knockdown of NLN impaired the formation of respiratory chain supercomplexes. Through our BioID analysis, we also identified the mitochondrial Ca2+/H+ antiporter, LETM1 (leucine zipper-EF-hand containing transmembrane protein 1) as a top interactor with NLN. LETM1 is a known regulator of respiratory chain supercomplex formation. We showed that knockdown of NLN impaired LETM1 assembly, potentially explaining how NLN regulates supercomplex formation. Finally, we tested if hypoxia influences respiratory chain supercomplex formation and sensitivity to NLN inhibition. OCI-AML2 cells cultured for 72 hours under hypoxic conditions (0.2% O2) showed impaired assembly of respiratory chain supercomplexes, decreased levels of LETM1 protein, and resistance to NLN knockdown. Thus, we discovered that the mitochondrial protease NLN regulates oxidative metabolism by controlling the assembly of respiratory chain supercomplexes. Moreover, we highlight NLN as a potential new therapeutic target for AML. Disclosures Schimmer: Otsuka Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees; Medivir AB: Research Funding; Jazz Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees.

Description: Background: Pediatric acute myeloid leukemia (Pedi-AML) accounts for approximately 15% of childhood acute leukemias and the overall survival is 60-70%. Although cytogenetics provides risk stratification, most recurrent genetic events lack agents that specifically target them. Since genetic events are revealed by the expression and activation status of proteins, we hypothesized that we could define recurrent protein expression signatures that could suggest targets for individualized therapies. Methods:Custom Reverse Phase Protein Arrays (RPPA) with 95 Pedi-AML samples and 10 normal CD34+ bone marrow samples were performed and probed with 194 antibodies to determine protein expression levels and activation status. Proteins were divided in protein functional groups (ProFnGrp) to analyze them in the context of other proteins. Progeny clustering was performed to determine the optimal number of protein clusters and principal component analysis was used to visualize the distribution between protein clusters and normal CD34+ cells. Protein networks were generated using known literature associations combined with computational derived edges from the dataset. Associations between clinical features, outcomes and protein clusters were determined. Hierarchical clustering (Figure 1) was performed on a compilation of all protein clusters into one binary matrix to identify recurrent protein expression signatures (PrSIG) that comprised similar combinations of protein constellations(PrCON). A list of proteins that were over or under expressed was made for each signature. Results: For each ProFnGrp 3 to 6 distinct clusters were recognized. All but 2 ProFnGrp (Cell Cycle, Protein Kinase C) had at least one protein cluster that was similar to normal CD34+ samples and all ProFnGrp had clusters specific to the leukemic state. Protein expression levels were overlaid onto the networks and revealed distinct expression and activation states. Hierarchical clustering showed strong co-correlation between protein clusters from various ProFnGrp and suggested 12 PrCON. Patients that expressed similar recurrent combinations of PrCON compiled 8 PrSIG. Heterogeneity was observed for overall survival (OS) and complete remission duration (RD). High OS rates were seen in PrSIG 5 compared to PrSIG 1 (8/12 vs. 3/11 alive, Median NR vs. 54 wks, P=0.09) and high relapse free survival was observed for PrSIG 3 compared to PrSIG 1 (6/6 vs. 3/8 CCR, (Median RD NR vs 39 wks, P=0.04). Rearrangements of 11q23 were overrepresented in PrSIG 2 (7/16, 43.8%) and 7 (5/19, 26.3%) and underrepresented in PrSIG 4 (1/20, 5%) and 6 (0/5, 0%) (P=0.027). Complex karyotype was highly present in PrSIG 8 (4/6, 67%) compared to the overall population (16/95, 16.7%). Hispanic ethnicity was high in PrSig 7 & 8 (53 & 67%) and low in PrSIG 1, 2 4 & 5 (20%, 25%, 5%, 8%). PrSIG membership was also correlated with FAB classification (P=0.025, M0 with PrSIG 4 and M5 with PrSIG 5) as well as with WBC, blast count, HGB and PLT. As we had both Pedi-AML and acute lymphoblastic leukemia (ALL) samples on this RPPA array (see accompanying ALL abstract), this enabled comparison between protein expression levels in both diseases. Combined re-analysis suggested 12 PrSIG and 12 PrCON; most PrSIG and PrCON were dominant for ALL or AML and only 1 PrSIG and 3 PrCON showed overlap between the two diseases. Identification of significantly altered proteins for each PrSIG can be used to select targeted therapies for combination. For example, for PrSIG1 combined inhibition of RPTOR and RICTOR may be useful as both are overexpressed. Conclusion: As hypothesized we demonstrated the existence of protein expression signatures in Pedi-AML based on strong correlations between different protein clusters. Signatures were correlated with therapeutic outcome, as well as with cytogenetics, Hispanic ethnicity, and laboratory features. Recognition of altered proteins within the signatures suggested combinations of targets that could facilitate personalized therapy. Shared protein constellations between AML and ALL indicate joint protein deregulations that could be targetable in both diseases. Figure 1. Hierarchical clustering shows recurrent patterns of 12 PrCON (y-axis) that form 8 recurrent PrSIG (x-axis). The association with Hispanic ethnicity is shown. Figure 1. Hierarchical clustering shows recurrent patterns of 12 PrCON (y-axis) that form 8 recurrent PrSIG (x-axis). The association with Hispanic ethnicity is shown. Disclosures No relevant conflicts of interest to declare.

Description: Background: Pediatric AML (pAML) treatment outcomes can vary due to genomic heterogeneity. Thus, selecting the right drugs for a given patient is challenging. There is a need for a priori means of predicting treatment responses based on tumor "omics". Computational biology modeling (CBM) is a precision medicine approach by which biological pathways of tumorigenesis are mapped using mathematical principles to yield a virtual, interactive tumor model. This model can be customized based on a patient's omics and analyzed virtually for response to therapies. Aim: To define prediction values of a CBM precision medicine approach in matching clinical response to ADE therapy in a cohort of pAML patients. Methods: Thirty pAML patients that were treated ADE chemotherapy were utilized with information on the clinical, genomic (cytogenetics, mutations) and protein expression data from this cohort of pAML patients used for the CBM. From cytogenetics results, gene copy number variations were coded as either knocked-down (KD) or over-expressed (OE). From NGS results (2 gene panel - CEBPA, NPM1), gene mutations were coded as either loss or gain of function (LOF or GOF). For protein expression data, proteins that were 〉2sigma from the mean were coded as KD if their value was 0. Proteins with values

Description: The proteasome degrades unneeded and damaged proteins. Tumor cells highly depend on increased protein production and their degradation suggesting that malignant cells with a high proliferation index will be more sensitive to proteasome inhibition. The addition of the proteasome inhibitor Bortezomib (Velcade, 'BTZ') to standard pediatric AML chemotherapy (cytarabine, daunorubicin and etoposide, 'ADE') depleted leukemia-initiating cells in a phase 2 clinical trial in pediatric AML (pedi-AML) patients. A randomized phase 3 clinical trial was then conducted by the Children's Oncology Group (COG) comparing ADE and ADE+BTZ treatment regimes in pedi-AML (AAML1031). To determine if there were specific protein expression profiles that correlated with response to BTZ-containing chemotherapy, we analyzed key components of the proteome of pedi-AML that participated in the trial using reverse phase protein arrays (RPPA). RPPA-based expression of 293 validated antibodies was tested in 500 leukemia samples and compared to expression in CD34+ samples from healthy individuals (n=30). Among all proteins, FOXO3 expression was identified as the protein with the highest influence on outcome in the ADE group. The expression of FOXO3 was prognostic for event-free survival (EFS) in both univariate (HR = 0.56, 95% CI = 0.34-0.90, P = 0.017) and multivariate (HR = 0.55, 95% CI = 0.34-0.88, P = 0.013) analysis. All patients were divided into two clusters (low and high) based on their FOXO3 expression level using median FOXO3 expression in normal CD34+. Kaplan-Meier survival analysis showed poor OS (3 year OS 65.3% vs. 73.9%, P = 0.03) and EFS (3 year EFS, 42.8% vs. 55%, P = 0.01) in low FOXO3 expressors (n=119) compared to patients with high FOXO3 expression (n=291) (fig. 1A). Notably, the poor prognostic effect of low FOXO3 for OS was seen in ADE (3 year OS 60% vs. 72.3%, P = 0.03), but not in ADE-BTZ (3 year OS 70.3% vs. 75.3%, P = 0.23) (fig. 1B). This suggests that in particular patients with low FOXO3 may be eligible candidates for BTZ-addition. To validate our findings, we performed knockdown (KD) of FOXO3 using a short hairpin approach in OCI-AML3 (p53WT) and THP-1 (p53null) cell lines. KD FOXO3 in OCI-AML3 had a short-term growth advantage compared to controls (Day 4, P = 0.004), but not KD FOXO3 THP-1 cells suggesting a role for p53 in the FOXO3 functional pathway. KD FOXO3 cells were more resistant to doxorubicin and etoposide combination therapy than controls (P = 0.04), confirming our clinical observations. Since therapeutic regimes in AML are currently shifting towards Bcl-2 inhibition by Venetoclax (ABT-199, 'ABT'), we were eager to test whether BTZ and ABT could act in synergy, and if this is related to FOXO3 expression. Single low dose BTZ and ABT did not reduce cell numbers after 3 days, but were effective when used in combination (

Description: Background: Heat shock factor 1 (HSF1) is a key-component of the heat shock response and plays a major role in cancer biology. The heat shock response is a highly conserved mechanism that is important for cell survival under stressful conditions. It facilitates normal protein folding and guards the proteome from misfolding and aggregation. Bortezomib (BTZ) is a reversible proteasome inhibitor and was recently tested in a phase 3 clinical trial for patients with de novo pediatric acute myeloid leukemia (pedi-AML). This trial compared standard therapy (cytarabine/daunorubicin/etoposide (ADE)) to ADE plus BTZ (ADE+B). As HSF1 and BTZ act in opposing circumstances, our goal was to globally assess the phosphorylation of HSF1 at serine 326 (HSF1.pS326) and to determine if baseline HSF1.pS326 was prognostic of clinical response to ADE+B and if we could identify a subset of patients that benefitted from ADE+B. Methods: Reverse phase protein array (RPPA) probed with 222 total antibodies and 69 specific post-translationally modified proteins was used to determine the protein expression levels in 505 bulk leukemic pedi-AML samples compared to 20 CD34+ samples from healthy pediatric donors. Patients participated in the Children's Oncology Group (COG) AAML1031 Phase 3 clinical trial that compared ADE to ADE+B. Survival curves were generated using the Kaplan-Meier method. Results: HSF1.pS326 protein expression levels in the pedi-AML patient samples. Based on the HSF1.pS326 expression levels, patients were classified into two clusters; high HSF1 326 (n = 216) and low HSF1.326 (n = 289). The median of the normal CD34+ samples was used as cut-off point. Expression of HSF1.pS326 was not prognostic for outcome in patients treated with ADE (event-free survival (EFS), p = 0.51) (Fig. 1A). In patients that were treated with ADE+B however, HSF1.pS326 levels were highly prognostic, with low HSF1.pS326 associated with a better EFS (p = 0.009) (Fig. 1B). Patients with high HSF1.pS326 did not benefit from BTZ addition (EFS, p = 0.77) (Fig. 1C), whereas patients with low HSF1.pS326 levels significantly improved (EFS, p = 0.004) (Fig. 1D). Between the two patient clusters, the NPM1 mutation state was higher in patients with high HSF1.pS326 levels (p = 0.042), and the CEBPA mutation state was higher in patients with low HSF1.pS326 (p = 0.011). No other patient or disease characteristics were different between the two clusters. To functionally validate our findings, we overexpressed wild type or mutant HSF1 in 293T cells and assessed its effect on BTZ sensitivity. The mutant HSF1 constructs contained amino acid substitutions at Ser326, resulting in either a phospho-mimic HSF1 (S326E), or a phospho-dead HSF1 protein (S326A). Cell viability analysis showed clear resistance to BTZ in cells with increased phosphorylation (HSF1 S326E) (Fig. 1E - blue curve). Cells that expressed reduced phosphorylation of HSF1 (HSF1 S326A), however were highly sensitive to BTZ (Fig. 1E, red curve). This confirms our observations from the pedi-AML patients. Conclusions: Using a proteomics approach, we identified a subgroup of pedi-AML patients that benefitted from BTZ addition to ADE therapy. We hypothesize that patients with low HSF1 activation are more susceptible to protein cell stress, and protein cell stress susceptibility is amplified by the addition of BTZ with ADE. The use of HSF1.pS326 as potential biomarker could identify patients that would benefit from ADE+B therapy. Patients with high HSF1 could potentially benefit from ADE+B in combination with a HSF1 (phospho) inhibitor, suggesting a novel combinational therapy. Figure 1. Kaplan Meier survival curves for EFS based on HSF1.pS326 expression. A. HSF1.pS326 did not affect outcome in ADE treated patients (high HSF1.PS326 shown in red, low HSF1.pS326 in blue). B. HSF1.pS326 was highly prognostic for EFS ADE+B. C. Patients with high HSF1.pS326 did not benefit from BTZ (ADE shown in blue, ADE+B in green). D. Patients that expressed low HSF1.pS326 did significantly better after ADE+B. E. To assess the effect of high and low HSF1.pS326 on BTZ sensitivity, 293T cells were transfected with two HSF1.pS326 mutants that mimicked high (blue) and low (red) HSF1.pS326 patients, in addition to wild type HSF1 (orange) and an empty vector as control (green). Cell viability assays showed resistance to BTZ in the phospho-mimic (blue) mutant, whereas the phospho-dead (red) mutant was sensitive to BTZ. Figure 1. Figure 1. Disclosures Kolb: Roche- Genentech: Membership on an entity's Board of Directors or advisory committees; Servier: Membership on an entity's Board of Directors or advisory committees.

Description: Background: Myelodysplastic syndromes (MDS) are heterogeneous clonally derived bone marrow disorders characterized by ineffective hematopoiesis and propensity to transform to acute myeloid leukemia. With greater than 15,000 new cases identified yearly, patients (pts) with MDS have a wide range of clinical manifestations and outcomes. Challenges in treating MDS include disease heterogeneity and a small number of effective treatments, particularly beyond first-line approaches and supportive care. Although mutational analysis of MDS provides prognostic information, the plethora of genetic events in a single case complicates using this information for guiding clinical therapy. We hypothesized that these genetic events coalesce into a finite number of protein expression signatures and that these would guide individualized therapy. Methods: A custom Reverse Phase Protein Array (RPPA) with 378 samples (including replicates) from 123 newly diagnosed and 76 relapsed/refractory MDS pts as well as 20 normal CD34+ bone marrow samples was created and probed with 136 antibodies to determine relative protein expression. To assess impact of source cell type on protein expression, 112 of the 378 samples (representing some replicates from 95 pts) had paired CD34+ and CD34+CD38- samples. Since proteins interact with each other and function within networks, proteins were first divided into 25 Protein Functional Groups (ProFnGp) based on their known functionality in the literature. Progeny clustering was then performed for each ProFnGp to determine the optimal number of protein clusters. Principal component analysis (PCA) was used to map global differences and similarities between protein clusters and normal CD34+ samples. Hierarchical clustering (HC) was performed on a compilation of all protein clusters in one binary matrix to identify recurrent protein expression signatures (PrSIG)that comprised similar combinations of protein constellations (PrCON). Associations between signature membership, clinical and molecular features, and outcome were assessed. Proteins that were universally over or under expressed and specific for a given signature were identified. Results: Clustering of pts for each ProFnGp revealed distinct relative expression and activation states compared to normal CD34+ samples. For each ProFnGp, 2 to 6 distinct expression clusters were identified, providing 110 protein clusters for HC. Of the 25 ProFnGp, all had MDS specific patterns and 19 had at least one cluster similar to normal CD34+ cells. HC revealed strong co-correlation between multiple groups of protein clusters from various ProFnGp and suggested 11 PrCON. Pts that expressed similar recurrent combinations of PrCON formed 11 PrSIG (Figure 1). Within PrSIG, no bias was observed in sample status (fresh or cryopreserved), source (peripheral blood or bone marrow), gender, or relapse status. Analysis of paired samples revealed 84% of CD34+ samples were present in separate PrSIG from corresponding CD34+CD38- samples, suggesting cases where CD34+ samples were distinct from CD34+CD38- samples. This suggests cell type should be considered in future analyses. Structured cluster memberships were identified, suggesting ProFnGp targets. The distinct cluster identified in PrCON 4 x PrSIG 1 revealed associations with ProFnGp including apoptosis, SMAD, PKC, mTOR, MEK, and Hippo pathways. Within this cluster, upregulation was identified in proteins including PKCα, PI3Kp110α, and SMAD6 and downregulation in SMAC, PKCd, SMAD1, SMAD4, TSC2, and NF2, suggesting targets for directed combination therapy with agents such as selective PKCα and PI3K inhibitors or SMAC mimetics. Overall summation of expression for each protein across each signature revealed many proteins with either significantly higher or lower expression relative to CD34+ controls. Conclusions: Analysis of protein expression levels in a network-based approach revealed classification of MDS pts into finite protein expression signatures based on the existence of recurrent protein constellations. Recognition of universal differentially expressed proteins, together with signature specific proteins, suggests targets for personalized and directed combinatorial therapeutics. Figure 1 HC based on binary ProFnGp cluster membership. Each vertical pt column consists of 25 of 110 protein clusters. Blue squares indicate positive cluster membership. Figure 1. HC based on binary ProFnGp cluster membership. Each vertical pt column consists of 25 of 110 protein clusters. Blue squares indicate positive cluster membership. Disclosures No relevant conflicts of interest to declare.

Description: Background: Acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) are both heterogeneous diseases. The underlying changes that results in leukemia are due to developmental, genetic, or environmental effects, and are mostly mediated by changes in protein expression or modification. We hypothesize that there is a finite number of patterns of protein expression and protein pathway utilization, whose perturbations result in the hallmarks of cancer. In this study we performed differential proteomics of ALL and AML, with the goal to understand the underlying (disease-specific) cellular changes of AML and ALL, as well as to identify protein utilizations that are shared between AML and ALL. Method: Reverse phase protein arrays (RPPA) was generated for 230 strictly validated antibodies using samples from 130 ALL and 241 AML patient samples, and 10 CD34+ samples from healthy controls. Expression levels were normalized relative to the normal CD34+ cells. Due to some inherent considerations of the traditional hierarchical clustering (HC) (e.g. HC weighs all proteins equally in all situations, HC is agnostic to all known functional relationships between proteins, HC requires that all data be considered and placed into a group), expression data was analyzed using the MetaGalaxyanalysis. This approach starts with the allocation of proteins into 31 protein functional groups (PFG). Progeny clustering was applied to identify an optimal number of protein clusters within each PFG. Block clustering identified protein clusters that recurrently co-occurred (protein constellation (CON)), and for each subgroup of patients that expressed similar combinations of protein constellations (patient signature (SIG)). Proteins that were differentially expressed were identified using the student's T-test or ANOVA, and a Bonferroni adjusted P-value (0.05/ 230 = 0.00021739). Results: The MetaGalaxy approach identified a substantial amount of structure across the data set (Figure 1), with an optimal number of 12 CON (horizontally) and 13 SIG (vertically). The majority of SIG were correlated with either ALL (SIG 1, 3, 4, 5) or AML (SIG7, 8, 9, 10, 11, 12) (annotation bar Figure 1), although SIG2, 6, and 13 contained a mixture of both (P〈 0.001). Similarly, CON1, 2, 3 were mostly associated with ALL, CON6, 8, 9 and 11 with AML, while CON4, 5, 7 and 10 were observed in both diseases. To understand more about the protein signaling utilizations deregulated proteins were identified for each CON and SIG. For example, CON1 was associated with PFG Apoptosis Occurring (e.g. CASP9-cl330, PARP1), autophagy (e.g. PRKAA1_2, PRKAA1_2-pTyr172), and apoptosis BH3 (e.g. BCL2, BAD-pSer112). In ALL, signature membership of CON5 was associated with a superior overall survival and complete remission duration (P= 0.016; P= 0.035). CON4 was associated with a high rate of early deaths (P = 0.041), but not with a higher frequency of relapses (P = 0.520). In AML, signatures were predictive for OS and CR, with SIG7, 10, 12 as being favorable vs SIG2, 6, 9, 11, 13 as being unfavorable. CON4 was predictive of late relapses (≥ 2 yr.). Interestingly, CON5 was associated with a trend toward inferior CR duration in AML (P= 0.093), whereas this CON5 was favorably prognostic in ALL. Conclusion: In this study we confirmed our original hypothesis that there is a finite number of SIG in ALL and AML. Although, ALL and AML are both hematological diseases that share many molecular events, SIG and CON membership were significantly correlated with ALL and AML, confirming that protein expression patterns for the majority of cases of ALL ≠ AML. However, given that some CON were associated with both disease, this indicates that common features between both also exist. Proteins or pathways with similar utilization in both diseases may allow for information on clinical utility from one disease to be transitive to the other, while those with differential utilization are likely to be uninformative with respect to clinical utility in the other disease. Figure 1. MetaGalaxy analysis for AML and ALL. Each row represents one protein clusters (n = 123), each column represents one patient (n = 371). Blue indicates membership for that particular protein cluster. Annotation bar shows strong correlation with disease (yellow = B-ALL, pink = T ALL, blue = AML). Figure Disclosures No relevant conflicts of interest to declare.

Description: Acute myeloid leukemia (AML) cells and stem cells have unique mitochondrial characteristics with an increased reliance on oxidative phosphorylation (OXPHOS). To identify new biological vulnerabilities in the mitochondrial proteome of AML cells, we conducted an shRNA screen and identified neurolysin (NLN), a zinc metalloprotease whose mitochondrial function is not well understood and whose role in AML has not been previously reported. To begin our investigation into the role of NLN in AML, we analyzed NLN gene expression in a database of 536 AML and 73 normal bone marrow samples. NLN was overexpressed in 41% of AML samples. Overexpression of NLN in primary AML cells compared to normal hematopoietic cells was confirmed by immunoblotting. To validate the results of the screen and to determine whether NLN is required for AML growth and viability, we knocked down NLN in the leukemia cell lines OCI-AML2, MV4-11, NB4, and TEX with shRNA. NLN knockdown reduced leukemia growth and viability by 50-70%. Moreover, knockdown of NLN in AML cells reduced the clonogenic growth of leukemic cells in vitro and the engraftment of AML cells into mouse marrow after five weeks by up to 80% and 85%, respectively. The mitochondrial function of NLN is largely unknown, so we identified NLN's mitochondrial protein interactors in T-REx HEK293 cells using proximity-dependent biotin labeling (BioID) coupled with mass spectrometry (MS). This screen identified 73 mitochondrial proteins that preferentially interacted with NLN and were enriched for functions including respiratory chain complex assembly, respiratory electron transport, and mitochondrion organization. Therefore, we assessed the effects of NLN knockdown on OXPHOS. NLN knockdown reduced basal and maximal oxygen consumption, but there were no changes in the levels of individual respiratory chain complex subunits. To understand how NLN influences OXPHOS, we examined the formation of respiratory chain supercomplexes (RCS). Respiratory chain complexes I, III, and IV assemble into higher order quaternary structures called RCS, which promote efficient oxidative metabolism. NLN knockdown significantly impaired RCS formation in T-REx HEK293, OCI-AML2, and NB4 cells, which was rescued by overexpressing wild-type shRNA-resistant NLN. RCS have not been previously studied in leukemia. Therefore, we analyzed their levels in primary AML patient samples and normal hematopoietic cells. RCS assembly was increased in a subset of AML patient samples and positively correlated with NLN protein expression (R2 = 0.83, p 〈 0.05), suggesting that NLN mediates RCS assembly in AML. To investigate how NLN may be regulating RCS assembly, we analyzed our BioID results to identify NLN interactors that are known regulators of supercomplex formation. Among the top interactors was the known RCS regulator, LETM1. Knockdown of NLN in AML cells impaired LETM1 assembly. Of note, knockdown of LETM1 also reduced growth and oxygen consumption of AML cells. As a chemical approach to evaluate the effects of NLN inhibition on AML cells, we used the allosteric NLN inhibitor R2, (3-[(2S)-1-[(3R)-3-(2-Chlorophenyl)-2-(2-fluorophenyl)pyrazolidin-1-yl]-1-oxopropan-2-yl]-1-(adamantan-2-yl)urea), whose anti-cancer effects have not been previously reported. R2 reduced viability of AML cells, as well as two primary AML culture models, 8227 and 130578. R2 impaired RCS formation in OCI-AML2, NB4, 8227, and primary AML cells. Moreover, R2 reduced the CD34+CD38- stem cell enriched population in 8227 cells, reduced LETM1 complex assembly, and impaired OXPHOS in OCI-AML2 and 8227 cells. Finally, we assessed the effects of inhibiting NLN in mice engrafted with primary AML and normal hematopoietic cells in vivo. Treatment of mice with R2 reduced the leukemic burden in these mice without toxicity. Moreover, inhibiting NLN targeted the AML stem cells as evidenced by reduced engraftment in secondary experiments. In contrast, inhibiting NLN did not reduce the engraftment of normal hematopoietic cells. Collectively, these results demonstrate that inhibition of NLN preferentially targets AML cells and stem cells as compared to normal hematopoietic cells. In summary, we defined a novel role for NLN in RCS formation. We show that RCS are necessary for oxidative metabolism in AML and highlight NLN inhibition as a potential therapeutic strategy. Disclosures Minden: Trillium Therapetuics: Other: licensing agreement. Chan:Agios: Honoraria; AbbVie Pharmaceuticals: Research Funding; Celgene: Honoraria, Research Funding. Schimmer:Medivir Pharmaceuticals: Research Funding; Novartis Pharmaceuticals: Consultancy; Jazz Pharmaceuticals: Consultancy; Otsuka Pharmaceuticals: Consultancy.