Abstract Sonic Hedgehog (SHH) signaling functions in temporal- and context-dependent manners to pattern diverse tissues during embryogenesis. The signal transducer Smoothened (SMO) is induced by sterols, oxysterols, and arachidonic acid (AA) through binding pockets in its extracellular cysteine-rich domain (CRD) and 7-transmembrane (7TM) bundle. In vitro analyses suggest SMO signaling is allosterically enhanced by combinatorial ligand binding to these pockets, but in vivo evidence of SMO allostery is lacking. Herein, we map an AA binding pocket at the top of the 7TM bundle and show that its disruption attenuates SHH- and sterol-stimulated SMO induction. A knock-in mouse model of compromised AA binding reveals that homozygous mutant mice are cyanotic, exhibit perinatal lethality, and display congenital heart disease. Surviving mutants demonstrate pulmonary maldevelopment and fail to thrive. Neurodevelopment is unaltered in these mice, suggesting that context-specific allosteric regulation of SMO signaling allows for precise tuning of pathway activity during cardiopulmonary development.

ABSTRACT:Despite great progress in understanding the genomic basis of immature T-cell acute lymphoblastic leukemia/lymphoblastic lymphoma (T-ALL) and acute leukemias of ambiguous lineage, there are still cases that lack defining genetic markers, complicating risk stratification, and limiting targeted therapeutic options. Recent studies have shown that enhancer hijacking drives oncogene activation in approximately half of T-ALL cases, with the BCL11B (BCL11 transcription factor B) enhancer frequently involved. Here, we describe a subtype of leukemia with a distinct gene-expression signature, and immunophenotype characterized by positivity for immature (CD38), myeloid (CD13), T-lymphoid (cytoplasmic [c]CD3, CD7), and B-lymphoid markers (CD19, CD79a, CD10). This subtype is defined by the t(14;16)(q32;q24) translocation, which places the FOXF1 gene and its antisense long noncoding RNA gene FENDRR under the regulatory control of the BCL11B enhancer, leading to their ectopic transcriptional activation. Common concomitant genetic lesions are loss-of-function alterations of GATA3, CDKN2A/CDKN2B deletion and activating JAK/STAT and NOTCH1 pathway mutations. Patients were predominantly children and adolescents/young adults and experienced poor treatment outcome. High-throughput drug screening of 176 compounds demonstrated efficacy of combined BCL2 apoptosis regulator family proteins and JAK/STAT signaling inhibitors. Additionally, the clinical use of tyrosine kinase inhibitors in some of these patients showed therapeutic efficacy. Collectively, these findings identify BCL11B-enhancer-mediated deregulation of FOXF1/FENDRR as a hallmark of a subtype of high-risk lineage ambiguous leukemia that is potentially amenable to targeted therapeutic intervention.

Sonic Hedgehog (SHH) signaling functions in temporal- and context-dependent manners to pattern diverse tissues during embryogenesis. The signal transducer Smoothened (SMO) is activated by sterols, oxysterols, and arachidonic acid (AA) through binding pockets in its extracellular cysteine-rich domain (CRD) and 7-transmembrane (7TM) bundle. In vitro analyses suggest SMO signaling is allosterically enhanced by combinatorial ligand binding to these pockets but in vivo evidence of SMO allostery is lacking. Herein, we map an AA binding pocket at the top of the 7TM bundle and show that its disruption attenuates SHH and sterol-stimulated SMO induction. A knockin mouse model of compromised AA binding reveals that homozygous mutant mice are cyanotic, exhibit high perinatal lethality, and show congenital heart disease. Surviving mutants demonstrate pulmonary maldevelopment and fail to thrive. Neurodevelopment is unaltered in these mice, suggesting that context-dependent allosteric regulation of SMO signaling allows for precise tuning of pathway activity during cardiopulmonary development.

ABSTRACT:Aberrant activation of BCL11B (BCL11B-a) defines a subtype of lineage-ambiguous leukemias with T-lymphoid and myeloid features, co-occurring activating FLT3 mutations, and a stem/progenitor immunophenotype and gene expression profile. Similar to other lineage-ambiguous leukemias, optimal treatment is unclear, and there are limited targeted therapeutic options. Here, we investigated the efficacy of B-cell lymphoma 2 (BCL-2) and FMS-like tyrosine kinase 3 (FLT3) inhibition with venetoclax and gilteritinib, respectively, in preclinical models of BCL11B-a leukemia. Despite variation in response to single-agent therapies, the combination of venetoclax plus gilteritinib (VenGilt) was highly effective in all models evaluated. BH3 profiling suggested that resistance to venetoclax monotherapy was due to the tumor-intrinsic dependence on additional BCL-2 family proteins before drug treatment. Longitudinal single-cell RNA sequencing analysis identified mitochondrial pathways and a pro-lymphoid gene expression signature as potential drivers of rare cell survival on VenGilt therapy. These data support clinical evaluation of venetoclax in combination with gilteritinib in BCL11B-a lineage-ambiguous leukemias.

Non-tuberculous mycobacteria are emerging pathogens with high intrinsic drug resistance. Among these, Mycobacterium abscessus is particularly refractory owing to its extensive array of resistance mechanisms. Here we introduce florfenicol amine (FF-NH2), a major metabolite of the antibiotic florfenicol, which acts as a prodrug with narrow-spectrum activity against M. abscessus-chelonae complex species. FF-NH2 leverages intrinsic M. abscessus resistance conferred by the transcription factor WhiB7. It avoids WhiB7-dependent resistance mediated by the O-acetyltransferase Cat and is activated by the WhiB7-dependent N-acetyltransferase Eis2 in a prodrug fashion to generate the active translational inhibitor FF acetyl (FF-ac). FF-NH2 induces Eis2 expression through WhiB7, creating a feed-forward bioactivation loop, which increases FF-ac accumulation and antimicrobial action. FF-NH2 displays antiresistance properties, can synergize with other antibiotics and mitigates toxicity linked to mammalian mitochondrial ribosome inhibition. Importantly, FF-NH2 demonstrated efficacy in a murine model of M. abscessus infection. These findings suggest intrinsic resistance can be exploited to develop safer and more effective treatments for this pathogen.

Retinoic acid (RA) is a standard-of-care neuroblastoma drug thought to be effective by inducing differentiation. Curiously, RA has little effect on primary human tumors during upfront treatment but can eliminate neuroblastoma cells from the bone marrow during post-chemo maintenance therapy—a discrepancy that has never been explained. To investigate this, we treat a large cohort of neuroblastoma cell lines with RA and observe that the most RA-sensitive cells predominantly undergo apoptosis or senescence, rather than differentiation. We conduct genome-wide CRISPR knockout screens under RA treatment, which identify bone morphogenic protein (BMP) signaling as controlling the apoptosis/senescence vs differentiation cell fate decision and determining RA’s overall potency. We then discover that BMP signaling activity is markedly higher in neuroblastoma patient samples at bone marrow metastatic sites, providing a plausible explanation for RA’s ability to clear neuroblastoma cells specifically from the bone marrow, by seemingly mimicking interactions between BMP and RA during normal development.

Drug combinations are essential to modern medicine, but their discovery remains slow and inefficient as experimental complexity expands rapidly with each additional drug tested. Although modern liquid handling systems enable complex and highly customizable experimental designs, a lack of strategies integrating these technologies with combination-specific analytical methods has limited throughput. Here we introduce Combocat, an open-source and streamlined framework that combines acoustic liquid handling protocols with machine learning-based inference to achieve ultrahigh-throughput drug combination screening. Using Combocat, we generate a reference dataset of over 800 unique combinations in a dense 10 × 10 matrix format across multiple cell types, and use this to train a predictive model that accurately infers drug combination effects from sparse data, drastically reducing the number of experimental measurements required. As proof of concept, we screened 9,045 combinations in a neuroblastoma cell line-the largest number of combinations tested in a single cell line to date-achieved using minimal resources. By integrating advanced drug dispensing technologies with predictive computational modeling, Combocat provides a scalable solution to accelerate the discovery of novel drug combinations.

Exploratory analysis of single-cell RNA sequencing (scRNA-seq) typically relies on hard clustering over two-dimensional projections like uniform manifold approximation and projection (UMAP). However, such methods can severely distort the data and have many arbitrary parameter choices. Methods that can model scRNA-seq data as non-discrete “gene expression programs” (GEPs) can better preserve the data’s structure, but currently, they are often not scalable, not consistent across repeated runs, and lack an established method for choosing key parameters. Here, we developed a GPU-based unsupervised learning approach, “consensus and scalable inference of gene expression programs” (CSI-GEP). We show that CSI-GEP can recover ground truth GEPs in real and simulated atlas-scale scRNA-seq datasets, significantly outperforming cutting-edge methods, including GPT-based neural networks. We applied CSI-GEP to a whole mouse brain atlas of 2.2 million cells, disentangling endothelial cell types missed by other methods, and to an integrated scRNA-seq atlas of human tumors and cell lines, discovering mesenchymal-like GEPs unique to cancer cells growing in culture.

Abstract Retinoic acid (RA) is a standard-of-care neuroblastoma drug used during post-chemo consolidation therapy. Based on clinical trials from the 90s, RA benefits 10-15% of patients. It is widely believed the anti-cancer activity of RA is due to retinoid-induced differentiation of cancer cells, conclusions largely attributable to observations in cell culture. However, given RA is typically used in the minimal residual disease setting, this mechanism has never been definitively proved in patients. To better understand RA’s activity, we deployed several new technologies. First, we conducted genome-wide CRISPR modifier screens in RA-treated hyper-sensitive neuroblastoma cell lines. Surprisingly, we found that, in these cells, RA primarily decreased cell viability via apoptosis or senescence, rather than differentiation—activities in which the CRISPR screens strongly implicated bone morphogenetic protein (BMP) signaling. Using ChIP-seq and RNA-seq we showed these behaviors were mediated by the coordinated gene regulatory actions of RARA and BMP-family SMAD transcription factors. Notably, interactions between BMP signaling and RA are well established in developmental biology, where BMP signaling can tip cell fate decisions between differentiation, apoptosis, and senescence upon exposure to naturally occurring RA, behaviors that can seemingly be maintained in neuroblastoma cells. Next, we assessed the correlations between RA IC50 and the expression of all (∼20,000) genes in a panel of 19 cell lines. Remarkably, SMAD9, a critical downstream transcription factor of the BMP pathway, was the #1 most correlated gene with RA IC50 (R = -0.92, P = 4.2 × 10-6), suggesting a highly generalizable relationship between BMP signaling and RA response. By modulating SMAD9, or other components of BMP signaling, we could promote apoptosis/senescence and sensitize cells to RA. We then performed large-scale drug combination screens of RA and the drug FK506, which can amplify BMP signaling activity. RA exhibited a synergistic effect with FK506 in all 10 neuroblastoma cell lines we tested, very strikingly in some cases, suggesting it could be possible to pharmacologically amplify the activity of RA in patients. Finally, using published single-cell RNA-seq data from neuroblastoma patient samples, we found BMP signaling activity is relatively low in primary tumors, but much higher in disseminated metastatic neuroblastoma cells in the bone marrow. We confirmed this trend using immunofluorescence staining in 6 paired primary patient samples. This site-specific variability in BMP signaling activity provides the first reasonable explanation for RA’s curious clinical activity, whereby it has little effect on bulky established tumors, but has been shown to clear disseminated metastatic cells from the bone marrow during consolidation therapy. Overall, our study revealed that BMP signaling controls neuroblastoma cell fate and sensitivity to RA and that this observation is consistent with the unique clinical behaviors of this drug. Citation Format: Min Pan, Yinwen Zhang, William C. Wright, Hyeong-Min Lee, Richard H. Chapple, Xueying Liu, Jonathan Low, Duane Currier, Allister J. Loughran, Michael A. Dyer, Shondra M. Pruett, Burgess Freeman III, Taosheng Chen, Brian J. Abraham, Elizabeth Stewart, John Easton, Paul Geeleher. BMP signaling determines neuroblastoma sensitivity to retinoic acid by directing cell fate [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Pediatric Cancer Research; 2024 Sep 5-8; Toronto, Ontario, Canada. Philadelphia (PA): AACR; Cancer Res 2024;84(17 Suppl):Abstract nr B005.

Abstract Signatures of BRCAness are found in many osteosarcoma (OS) tumors, driving an interest in PARP inhibition (PARPi) for OS treatment. Although PARPi have shown limited efficacy as OS monotherapy, rational combination strategies have not yet been explored for this disease. We performed a genome-scale loss-of-function CRISPR-Cas9 screen in two OS cell lines in the presence and absence of olaparib, a small molecule PARPi. ATM knockout (KO) scored highly as a potent synergizer of PARPi across both cell lines. Here we report this screen finding as well as in vitro validation of PARPi and ATM inhibition (ATMi) synergy in OS models. OS cell lines SAOS2 and CAL72 were infected with the Avana genome-scale lentiviral CRISPR library, targeting 20,000 genes with 4 uniquely barcoded guides (sgRNAs) per gene. After infection and antibiotic selection, cells were grown in an IC40 dose of olaparib or matched control for 18 days. Following DNA sequencing of barcodes and rigorous quality control, sgRNA dropout hits were calculated against the control arm. ATM KO sensitizing to PARPi was first validated utilizing a CRISPR-based genetic approach. ATM KO in CAL72 and SAOS2 was performed via lentiviral infection. After KO confirmation by western immunoblotting (WB), ATM KO and control KO cells were treated with PARPi in a dose range of 19.5 nM-10 µM. ATP-based and Incucyte assays were used to quantify differential cell viability. Secondly, the combination of a small molecule ATMi and a PARPi was studied in a panel of OS cell lines. Synergistic cytotoxicity was quantified using the Zero Interaction Potency (ZIP) model. DNA damage levels after PARPi and ATMi (by chemical inhibition and genetic KO) were assessed by WB for γH2AX, a marker of DNA double-stranded breaks (DSBs). In our genome-scale CRISPR screens, ATM arose as a top druggable hit across both lines. Reassuringly, PARP1 KO scored as a top resistance mediator to PARPi, confirming biologic relevance of the screen and supporting the known PARP-trapping mechanism of these drugs in OS. At low-throughput, ATM KO lines trended toward increased sensitivity to PARPi in short-term viability assays. Combined chemical inhibition yielded highly synergistic ZIP scores across OS cell lines; synergy was achieved with both drugs in the low nanomolar range. γH2AX levels were substantially increased in the setting of combined PARPi and ATMi (via chemical inhibition or genetic KO) as compared to inactivation of either PARP or ATM alone, indicating increased presence of DNA DSBs. In conclusion, these data demonstrate synergy of ATM and PARP inhibition, providing promise for their potential combined use in the treatment of OS. Accumulation of DNA DSBs in response to PARP and ATM inactivation suggests a shared role in DNA damage repair as the molecular basis of synergism. Ongoing studies are investigating anti-tumor potential in in vivo OS models and exploring the mechanistic impact of the combination in OS tumors. Citation Format: Sona N. Kocinsky, Janeala J. Morsby, William C. Wright, Monika Wierdl, Caroline S. Wechsler, Gabriela Alexe, Paul Geeleher, Kimberly Stegmaier, Lillian M. Guenther. ATM and PARP combined inhibition demonstrate synergistic antitumor efficacy in osteosarcoma models [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3912.

Background Neuroblastoma is a common pediatric cancer, where preclinical studies suggest that a mesenchymal-like gene expression program contributes to chemotherapy resistance. However, clinical outcomes remain poor, implying we need a better understanding of the relationship between patient tumor heterogeneity and preclinical models.Results Here, we generate single-cell RNA-seq maps of neuroblastoma cell lines, patient-derived xenograft models (PDX), and a genetically engineered mouse model (GEMM). We develop an unsupervised machine learning approach ("automatic consensus nonnegative matrix factorization" (acNMF)) to compare the gene expression programs found in preclinical models to a large cohort of patient tumors. We confirm a weakly expressed, mesenchymal-like program in otherwise adrenergic cancer cells in some pre-treated high-risk patient tumors, but this appears distinct from the presumptive drug-resistance mesenchymal programs evident in cell lines. Surprisingly, however, this weak-mesenchymal-like program is maintained in PDX and could be chemotherapy-induced in our GEMM after only 24 h, suggesting an uncharacterized therapy-escape mechanism.Conclusions Collectively, our findings improve the understanding of how neuroblastoma patient tumor heterogeneity is reflected in preclinical models, provides a comprehensive integrated resource, and a generalizable set of computational methodologies for the joint analysis of clinical and pre-clinical single-cell RNA-seq datasets.

Combination chemotherapy is crucial for successfully treating cancer. However, the enormous number of possible drug combinations means discovering safe and effective combinations remains a significant challenge. To improve this process, we conduct large-scale targeted CRISPR knockout screens in drug-treated cells, creating a genetic map of druggable genes that sensitize cells to commonly used chemotherapeutics. We prioritize neuroblastoma, the most common extracranial pediatric solid tumor, where ~50% of high-risk patients do not survive. Our screen examines all druggable gene knockouts in 18 cell lines (10 neuroblastoma, 8 others) treated with 8 widely used drugs, resulting in 94,320 unique combination-cell line perturbations, which is comparable to the largest existing drug combination screens. Using dense drug-drug rescreening, we find that the top CRISPR-nominated drug combinations are more synergistic than standard-of-care combinations, suggesting existing combinations could be improved. As proof of principle, we discover that inhibition of PRKDC, a component of the non-homologous end-joining pathway, sensitizes high-risk neuroblastoma cells to the standard-of-care drug doxorubicin in vitro and in vivo using patient-derived xenograft (PDX) models. Our findings provide a valuable resource and demonstrate the feasibility of using targeted CRISPR knockout to discover combinations with common chemotherapeutics, a methodology with application across all cancers.

Abstract While retinoic acid (RA) has been successfully used for leukemia treatment for decades, the attempt to treat solid tumors with RA remains challenging, with less than 10% of neuroblastoma (NB) patients achieving complete remission when treated with RA alone. It has been believed the anti-cancer activity of retinoids is due to retinoid-induced terminal differentiation, however, what determines cell response or whether differentiation is the main effect of RA has never been fully understood. To better understand RA’s activity, we conducted genome-wide CRISPR modifier screens in RA-treated hyper-sensitive NB cell lines. Surprisingly, we found that, in these cells, RA primarily decreased cell viability via apoptosis or senescence, rather than differentiation—activities in which the CRISPR screens strongly implicated bone morphogenetic protein (BMP) signaling. BMP activation promoted apoptosis/senescence and sensitized cells to RA. Conversely, BMP inhibitors and SMAD9 (a critical transcription factor of the BMP pathway) knockout enhanced RA’s ability to induce differentiation but reduced cell sensitivity to RA. Our ChIP-seq and RNA-seq data showed these behaviors were mediated by the coordinated gene regulatory actions of RARA and BMP-family SMAD transcription factors. Furthermore, in a panel of 19 cell lines we screened with RA, we assessed the correlations between RA IC50 and the expression of all (~20,000) genes (publicly available data from the GDSC and Depmap). Remarkably, SMAD9 was the number 1 and number 10 most correlated gene, respectively, with RA IC50 (ranked by Pearson correlation coefficient; GDSC R = -0.92, P = 4.2 × 10−6; Depmap R = -0.81, P = 4.8 × 10−4). We also performed comprehensive large-scale drug screens in these cell lines with combinations of RA and a BMP activator FK506. RA exhibited a synergistic effect with FK506 in all the NB cell lines, very strikingly in some cell lines. All these data suggest BMP signaling is generally required for RA sensitivity and BMP activators are promising candidates in combination with RA to treat NB. Using published bulk and single cell RNA-seq data from NB patient samples, we found BMP signaling activity is relatively low in primary tumors, but well maintained in disseminated NB cells derived from bone marrow. This explains why RA is clinically used as a maintenance therapy and can only successfully treat the minimal residual disease and suggests that tumor microenvironment is a critical factor in determining cell response to RA. Overall, our study revealed that BMP controls NB cell fate and sensitivity in RA treatment. Notably, interactions between BMP signaling and RA are well established in developmental biology, where BMP signaling can tip cell fate decisions between differentiation, apoptosis, and senescence upon exposure to endogenous RA. Our data suggest that this developmental process can be maintained in NB cells, unveiling a novel mode of anti-neoplastic action. Citation Format: Min Pan, Yinwen Zhang, William C. Wright, Hyeong-Min Lee, Richard H. Chapple, Xueying Liu, Jonathan Low, Duane Currier, Allister J. Loughran, Dyer A. Dyer, Shondra M. Pruett, Burgess Freeman, Taosheng Chen, Brian J. Abraham, Elizabeth Stewart, John Easton, Paul Geeleher. BMP signaling determines neuroblastoma cell fate and sensitivity to retinoic acid [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 151.

Abstract The battle against many cancers and infectious diseases has long been hindered due to the complexity of finding potent and effective drug combinations. With each new drug considered, the number of combinations to potentially test increases exponentially, posing substantial challenges in screening throughput. These challenges further intensify when accounting for the number of doses of each drug that need to be tested in a drug combination matrix (known as matrix density). There is a pressing need to screen large amounts of combinations at sufficient density to discover new therapies for diseases like cancer, but this has traditionally been out of reach. However, the recent widespread adoption of acoustic liquid handling robots has shown promise to overcome these obstacles by allowing for intricate drug screening template designs which were previously not possible to make. Despite these advances, the throughput achieved by these technologies has been limited due to lack of broadly accessible protocols and analytical tools for drug combination screening. We present Combocat, an end-to-end platform that allows for substantial increases in throughput of drug combination screens by combining experimental protocols that can be deployed for acoustic liquid handlers, machine learning algorithms for data imputation, and software that allows for in-depth analysis of results. We first generated a reference dataset of over 250,000 unique drug combination measurements in multiple cancer cell lines. The combination data were collected in a dense format (10×10 combination matrices) using a novel drug-drug template and achieved a dramatic increase in throughput compared to conventional methods. We then used this dataset to build a computational model which allowed us to accurately estimate drug combination effects using sparse measurements and imputing non-measured values with machine learning. The sparse measurements are collected in 1536-well microplates and substantially boost the throughput capabilities of drug-drug screens. As proof-of-concept, we used our method to screen a preclinical model of neuroblastoma with 9,045 drug combinations. This represents 10% the scale of the largest drug combination studies ever reported, achieved using a fraction of the resources, and in dense formats. We validated our findings by re-screening top hits using the fully-measured, non-imputed method and demonstrate the accuracy of our platform. The Combocat platform’s documentation and codebase is open-source, and we also make a GUI available for interactive exploration of screening results. By integrating advanced experimental and computational methods, we provide a generalizable pipeline that will expedite synergy screens and the drug combination discovery process for many diseases. Citation Format: William C. Wright, Paul Geeleher. An ultrahigh-throughput synergy screening platform enables discovery of novel drug combinations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 4936.

Abstract High-risk neuroblastoma is one of the most difficult to treat pediatric cancers, with a survival rate of only about 50% and significant long-term consequences from current chemotherapy. Preclinical models such as cell lines and mice are the backbone of drug development and experimental-mechanistic oncology, but the development of new treatments is hampered in part by a lack of understanding of the direct clinical relevance of such data collected in preclinical models. Despite this, few formal methods have been developed to determine how these various models represent/resemble primary patient tumors. Here, we present the first comprehensive single-cell RNA-seq analysis of neuroblastoma across an extensive cohort of patient tumors and a variety of preclinical model systems (n = 126 total samples assembled - the largest cohort of its kind). By building an innovative unsupervised machine learning method, which we term “automatic consensus nonnegative matrix factorization” (acNMF), we have integrated and contrasted the transcriptional landscapes of patient tumors with those of cell lines, patient-derived xenografts (PDX), and genetic mouse models (GEMM).Using these tools, we discovered the dominant adrenergic gene expression programs found in neuroblastoma patient tumors were preserved across all preclinical models. However, the presumptive chemo-resistant mesenchymal-like programs, while identifiable in cell lines, were primarily restricted to subpopulations of cancer-associated fibroblasts and Schwann-like cells in vivo. Surprisingly, a mesenchymal-like program could be acutely chemotherapy-induced in GEMM and was evident in pre-treated patient and PDX samples, suggesting a previously uncharacterized mechanism of therapy escape. In addition to these core findings, our computational tools were able to further delineate the classical neuroblastoma adrenergic and mesenchymal gene expression programs, discovering for example, novel subpopulations of cancer associated fibroblasts. These behaviors were conserved across tumors and most preclinical models, which we validated by RNA in situ hybridization, which is a high resolution ultra sensitive spatial transcriptomics technology. Our work cautions against overreliance on traditional preclinical models without recognizing their limitations and we offer a nuanced, high-resolution view of neuroblastoma pre-clinical systems for advancing therapeutic development. We have launched an open-source web resource, featuring this integrated map to aid the scientific community in further exploration of these data and hypothesis generation (available at http://pscb.stjude.org). Citation Format: Rich Chapple, Xueying Liu, Sivaraman Natarajan, Margaret I. Alexander, Yuna Kim, Anand Patel, Christy W. LaFlamme, Min Pan, William C. Wright, Hyeong-Min Lee, Yinwen Zhang, Meifen Lu, Selene C. Koo, Courtney Long, John Harper, Chandra Savage, Melissa D. Johnson, Thomas Confer, Walter J. Akers, Michael A. Dyer, Heather Sheppard, John Easton, Paul Geeleher. A novel unsupervised machine learning model applied to neuroblastoma single-cell RNA-seq data reveals a drug-induced mesenchymal-like gene expression program [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 868.

CYP3A5 is a cytochrome P450 (CYP) enzyme that metabolizes drugs and contributes to drug resistance in cancer. However, it remains unclear whether CYP3A5 directly influences cancer progression. In this report, we demonstrate that CYP3A5 regulates glucose metabolism in pancreatic ductal adenocarcinoma. Multi-omics analysis showed that CYP3A5 knockdown results in a decrease in various glucose-related metabolites through its effect on glucose transport. A mechanistic study revealed that CYP3A5 enriches the glucose transporter GLUT1 at the plasma membrane by restricting the translation of TXNIP, a negative regulator of GLUT1. Notably, CYP3A5-generated reactive oxygen species were proved to be responsible for attenuating the AKT–4EBP1–TXNIP signaling pathway. CYP3A5 contributes to cell migration by maintaining high glucose uptake in pancreatic cancer. Taken together, our results, for the first time, reveal a role of CYP3A5 in glucose metabolism in pancreatic ductal adenocarcinoma and identify a novel mechanism that is a potential therapeutic target.

Pediatric solid tumors arise from diverse tissues during development and exhibit a wide range of molecular, cellular and genetic features. This diversity, combined with the low incidence of pediatric cancer makes it increasingly difficult to personalize therapy for individual patients based on the unique features of their tumors. Therefore, well-credentialed preclinical models that capture the diversity and heterogeneity of pediatric solid tumors are essential for identifying molecular targeted therapeutics for precision medicine. Here, we report 281 orthotopic patient derived xenografts (O-PDXs) from 224 patients representing 24 different types of pediatric solid tumors. We have performed genomic characterization of the O-PDXs and compared them to their corresponding patient tumors. To demonstrate the feasibility and utility of using such a diverse collection of O-PDXs in preclinical studies, we performed a preclinical pediatric precision medicine trial based on the NCI-COG Pediatric MATCH trial enrollment criteria. We also tested molecular targeted therapy for a novel oncogenic fusion recently reported in pediatric melanoma and precision drug delivery using nano-liposomal irinotecan. Our studies demonstrate the value of large, well-credentialed preclinical models for future precision medicine in pediatric oncology using single agents, drug combinations and novel drug formulations. Translational Relevance This study demonstrates the value of utilizing fully characterized preclinical models of pediatric solid tumors to evaluate the response to precision medicine approaches. Our results demonstrate the importance of performing comprehensive preclinical testing using multiple orthotopic patient derived xenografts to validate and prioritize vulnerabilities identified through genomic or integrated analyses which can be translated into clinical trials. Importantly, this study identified combinations using nano-liposomal irinotecan in a precision drug delivery approach that may benefit pediatric solid tumor patients. In addition, all models and their associated data are made freely available to the scientific community through the Childhood Solid Tumor Network. ### Competing Interest Statement The authors have declared no competing interest.

Rearrangments in Histone-lysine-N-methyltransferase 2A (KMT2Ar) are associated with pediatric, adult and therapy-induced acute leukemias. Infants with KMT2Ar acute lymphoblastic leukemia (ALL) have a poor prognosis with an event-free-survival of 38%. Herein we evaluate 1116 FDA approved compounds in primary KMT2Ar infant ALL specimens and identify a sensitivity to proteasome inhibition. Upon exposure to this class of agents, cells demonstrate a depletion of histone H2B monoubiquitination (H2Bub1) and histone H3 lysine 79 dimethylation (H3K79me2) at KMT2A target genes in addition to a downregulation of the KMT2A gene expression signature, providing evidence that it targets the KMT2A transcriptional complex and alters the epigenome. A cohort of relapsed/refractory KMT2Ar patients treated with this approach on a compassionate basis had an overall response rate of 90%. In conclusion, we report on a high throughput drug screen in primary pediatric leukemia specimens whose results translate into clinically meaningful responses. This innovative treatment approach is now being evaluated in a multi-institutional upfront trial for infants with newly diagnosed ALL.

Neuroblastoma is a common pediatric cancer, where preclinical studies suggest that a mesenchymal-like gene expression program contributes to chemotherapy resistance. However, clinical outcomes remain poor, implying we need a better understanding of the relationship between patient tumor heterogeneity and preclinical models. Here, we generated single-cell RNA-seq maps of neuroblastoma cell lines, patient-derived xenograft models (PDX), and a genetically engineered mouse model (GEMM). We developed an unsupervised machine learning approach ('automatic consensus nonnegative matrix factorization' (acNMF)) to compare the gene expression programs found in preclinical models to a large cohort of patient tumors. We confirmed a weakly expressed, mesenchymal-like program in otherwise adrenergic cancer cells in some pre-treated high-risk patient tumors, but this appears distinct from the presumptive drug-resistance mesenchymal programs evident in cell lines. Surprisingly however, this weak-mesenchymal-like program was maintained in PDX and could be chemotherapy-induced in our GEMM after only 24 hours, suggesting an uncharacterized therapy-escape mechanism. Collectively, our findings improve the understanding of how neuroblastoma patient tumor heterogeneity is reflected in preclinical models, provides a comprehensive integrated resource, and a generalizable set of computational methodologies for the joint analysis of clinical and pre-clinical single-cell RNA-seq datasets.