Dynamic network curvature analysis of gene expression reveals novel potential therapeutic targets in sarcoma

Network properties account for the complex relationship between genes, making it easier to identify complex patterns in their interactions. In this work, we leveraged these network properties for dual purposes. First, we clustered pediatric sarcoma tumors using network information flow as a similarity metric, computed by the Wasserstein distance. We demonstrate that this approach yields the best concordance with histological subtypes, validated against three state-of-the-art methods. Second, to identify molecular targets that would be missed by more conventional methods of analysis, we applied a novel unsupervised method to cluster gene interactomes represented as networks in pediatric sarcoma. RNA-Seq data were mapped to protein-level interactomes to construct weighted networks that were then subjected to a non-Euclidean, multi-scale geometric approach centered on a discrete notion of curvature. This provides a measure of the functional association among genes in the context of their connectivity. In confirmation of the validity of this method, hierarchical clustering revealed the characteristic EWSR1-FLI1 fusion in Ewing sarcoma. Furthermore, assessing the effects of in silico edge perturbations and simulated gene knockouts as quantified by changes in curvature, we found non-trivial gene associations not previously identified.

Clonal evolution during metastatic spread in high-risk neuroblastoma

Patients with high-risk neuroblastoma generally present with widely metastatic disease and often relapse despite intensive therapy. As most studies to date focused on diagnosis-relapse pairs, our understanding of the genetic and clonal dynamics of metastatic spread and disease progression remain limited. Here, using genomic profiling of 470 sequential and spatially separated samples from 283 patients, we characterize subtype-specific genetic evolutionary trajectories from diagnosis through progression and end-stage metastatic disease. Clonal tracing timed disease initiation to embryogenesis. Continuous acquisition of structural variants at disease-defining loci (MYCN, TERT, MDM2-CDK4) followed by convergent evolution of mutations targeting shared pathways emerged as the predominant feature of progression. At diagnosis metastatic clones were already established at distant sites where they could stay dormant, only to cause relapses years later and spread via metastasis-to-metastasis and polyclonal seeding after therapy.

Abstract 6541: Geometry of gene expression network reveals potential novel indicator in Ewing sarcoma

Oncogenic driver mutations in different pediatric sarcoma subtypes have been identified but may not be druggable. In general, identifying novel therapeutic targets and biomarkers for response remains a major challenge. We hypothesize that considering the structure of the interaction network in which the genes operate as a system is crucial for understanding a gene's role. We propose to use the protein interaction network geometry to characterize the shape of network architecture and identify key aspects of direct and indirect cooperation pertaining to the cancer network and prognosis using geometrical methods.

We model gene networks as weighted graphs where edges indicate protein-level interactions and edge weights estimate the strength of the interaction. The Human Protein Reference Database was used to define the gene network topology. RNA-Seq data from pediatric sarcoma tissues extracted from patients treated at MSK (n=12 Ewing sarcoma; n=29 osteosarcoma; n=20 desmoplastic small round cell tumor) was employed to prescribe correlation-based weights to create pediatric sarcoma subtype-specific weighted graphs. The geometry of the weighted gene networks was computed via a discrete notion of Ricci curvature.

Intuitively, the curvature provides a measure of feedback (triangles) in the network. Positive curvature reflects robust communication and ease of information transfer, while negative curvature reflects bridge-like architecture or bottlenecks of information flow. We utilized a dynamic (multi-scale) notion of curvature to quantify the functional associations between genes, computed as a function of scale between diffusion processes initially localized on each node (i.e., gene). The curvature becomes more positive on edges between communal genes and more negative on bridge-like edges between communities, until reaching the critical scale. Curvature therefore, as we demonstrate, partitions the cancer networks into functionally associated communities.

Community detection by removing bridge-edges, determined as edges with negative curvature at the critical scale, revealed sarcoma subtype-specific preferential gene associations. In particular, we agnostically found the EWSR1-FLI1 association in a cluster that was unique to the Ewing sarcoma network. Interestingly, we found ETV6 in the same community as the characteristic Ewing sarcoma EWSR1-FLI1 feature, suggesting a novel implication of ETV6 in Ewing sarcoma. These results suggest that persisting communities found by leveraging the cancer network geometry may identify potential mechanisms of drug resistance and actionable therapeutic targets.

Citation Format: Rena Elkin, Jung Hun Oh, Filemon Dela Cruz, Larry Norton, Joseph O. Deasy, Andrew L. Kung, Allen R. Tannenbaum. Geometry of gene expression network reveals potential novel indicator in Ewing sarcoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 6541.

Single-cell analysis and functional characterization uncover the stem cell hierarchies and developmental origins of rhabdomyosarcoma

Rhabdomyosarcoma (RMS) is a common childhood cancer that shares features with developing skeletal muscle. Yet, the conservation of cellular hierarchy with human muscle development and the identification of molecularly defined tumor-propagating cells has not been reported. Using single-cell RNA-sequencing, DNA-barcode cell fate mapping and functional stem cell assays, we uncovered shared tumor cell hierarchies in RMS and human muscle development. We also identified common developmental stages at which tumor cells become arrested. Fusion-negative RMS cells resemble early myogenic cells found in embryonic and fetal development, while fusion-positive RMS cells express a highly specific gene program found in muscle cells transiting from embryonic to fetal development at 7-7.75 weeks of age. Fusion-positive RMS cells also have neural pathway-enriched states, suggesting less-rigid adherence to muscle-lineage hierarchies. Finally, we identified a molecularly defined tumor-propagating subpopulation in fusion-negative RMS that shares remarkable similarity to bi-potent, muscle mesenchyme progenitors that can make both muscle and osteogenic cells.

RRM2 enhances MYCN-driven neuroblastoma formation and acts as a synergistic target with CHK1 inhibition

High-risk neuroblastoma, a pediatric tumor originating from the sympathetic nervous system, has a low mutation load but highly recurrent somatic DNA copy number variants. Previously, segmental gains and/or amplifications allowed identification of drivers for neuroblastoma development. Using this approach, combined with gene dosage impact on expression and survival, we identified ribonucleotide reductase subunit M2 (RRM2) as a candidate dependency factor further supported by growth inhibition upon in vitro knockdown and accelerated tumor formation in a neuroblastoma zebrafish model coexpressing human RRM2 with MYCN. Forced RRM2 induction alleviates excessive replicative stress induced by CHK1 inhibition, while high RRM2 expression in human neuroblastomas correlates with high CHK1 activity. MYCN-driven zebrafish tumors with RRM2 co-overexpression exhibit differentially expressed DNA repair genes in keeping with enhanced ATR-CHK1 signaling activity. In vitro, RRM2 inhibition enhances intrinsic replication stress checkpoint addiction. Last, combinatorial RRM2-CHK1 inhibition acts synergistic in high-risk neuroblastoma cell lines and patient-derived xenograft models, illustrating the therapeutic potential.

Abstract 488: A systems biology approach to defining tumor heterogeneity and prognostic and targetable master regulator protein signatures from bulk and single-cell RNAseq in osteosarcoma

Despite many clinical trials, outcome for osteosarcoma (OS) patients remains poor, especially for those with metastatic or relapsed/refractory disease. We leveraged network-based systems biology approaches to discover Master Regulator (MR) proteins representing pharmacologically accessible, mechanistic determinants of OS cell state, and to dissect tumor transcriptional heterogeneity. MRs were identified by interrogating an OS regulatory network—generated from 87 RNAseq profiles in the TARGET OS cohort—with gene expression signatures from 149 diagnostic OS samples, using the VIPER algorithm. RNAseq profiles were generated in four sarcoma cell lines following perturbation with ~400 oncology drugs, and used to identify drugs capable of inverting MR activity profiles with the OncoTreat algorithm. Unsupervised, protein activity-based clustering of the samples identified two clusters, characterized by a significant overall survival difference (76% vs 38%, log-rank p-val 0.0041). Patients in the high-risk cluster presented aberrant activity of proteins involved in cell cycle (FOXM1, CENPF, TOP2A) and epigenetic remodeling (EZH2, KDM1A), while those in the low-risk cluster had aberrant activity of proteins involved in immune response (VAV1, CD86, CEBPE), interferon gama signaling (IRF5, MNDA), and senescence control (CREG1, TFEC). We further identified patients with metastatic disease and particularly poor outcome (<15% survival) showing high activity in stem cell-related proteins NANOG, ZSCAN4, and LIN28A. To address intra-tumor heterogeneity and immune infiltration (40-60% in the samples) in bulk RNAseq data, we inferred protein activity in previously published single-cell RNAseq profiles from 11 OS patients, using regulatory networks also produced by single-cell analysis. We identified three tumor subpopulations, co-existing in all the samples, with lineage similarity to osteoblasts, chondroblasts, and fibroblasts. The most active MR proteins (p-val < 10E-30) were highly distinct across these subpopulations: NANOG, REL, IGF2, TOP2A (osteoblasts), SSTR3, WNT1, KDM6B, EZH2 (chondroblasts), and NOD1, ERBB3, PRKDC, INSR (fibroblasts). Pharmacologically accessible MRs significantly active across the three subpopulations and in subsets of bulk RNAseq samples included EZH2, CDK2, TOP2A, HDAC5, and PRKDC. OncoTreat analysis predicted sensitivity to camptothecin, seliciclib, galunisertib, and rigosertib in both fibroblast- and chondroblast-like subpopulations. In conclusion, using network-based systems biology approaches in OS, we identified pharmacologically accessible MR subtypes differentiating survival at the bulk tissue level, and defining three distinct tumor subpopulations at the single-cell level with unique predicted drug sensitivities.

Citation Format: Somnath Tagore, Jovana Pavisic, Aaron T. Griffin, Katherine Janeway, Andrew L. Kung, Filemon Dela Cruz, Alejandro Sweet-Cordero, Inge Behroozfard, Stanley Leung, Alex Lee, Darrell Yamashiro, Julia Glade Bender, Andrea Califano. A systems biology approach to defining tumor heterogeneity and prognostic and targetable master regulator protein signatures from bulk and single-cell RNAseq in osteosarcoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 488.

Abstract 4083: Case study of single-cell protein activity-based drug prediction and validation for precision treatment of cholangiocarcinoma

Cholangiocarcinoma (CCA) is an aggressive biliary adenocarcinoma, with a median survival ranging from 12 to 37 months and no effective treatment options [Dhanasekaran et al. 2013]. Studies of the CCA tumor micro-environment (TME) and intratumoral heterogeneity have been limited, despite clinically significant interaction between tumor and stromal or immune components across other tumor types. Single-cell RNA Sequencing (scRNASeq) has emerged as a valuable technique to characterize the TME. Here we present a case study profiling CCA TME at the resolution of scRNASeq, and the first application of a unique OncoTarget (OTar) and OncoTreat (OTr) approach [Alvarez et al. 2018, Zeleke et al. 2020, Mundi et al. 2021] to predict and identify actionable drug targets at the single-cell level. These are both CLIA-certified algorithms for personalized drug prediction, now adapted for the first time to drug sensitivity prediction at the level of individual tumor cells. scRNASeq from a human CCA sample revealed significant tumor immune infiltration, with T-cells comprising the largest population, and fewer than 10% of cells identified as tumor cells, such that transcriptional profiles from RNA-Seq derived from whole tumor samples would be dominated by non-tumor cells. VIPER-based protein activity inference of tumor cells identified three sub-populations, not distinguished by gene expression alone [Alvarez et al. 2016]. These were characterized by upregulation of KRAS pathway, TNFa signaling with epithelial-mesenchymal transition, and upregulation of MYC targets, respectively. Consensus OTar/OTr drug prediction analysis on both scRNASeq tumor cells and bulk RNA-Seq of an engrafted PDX model from resected tumor at the time of biopsy ranked Glasdegib, Plicamycin, Flavopiridol, AT9283, and Dacinostat as the top 5 drugs with best overall tumor cell coverage. Therefore, we administered these drugs to a cohort of 8 PDX-bearing mice per treatment arm. Dacinostat and Plicamycin significantly reduced tumor growth rate (p=0.007 and p=0.03, respectively), with Dacinostat stabilizing tumor size over 28 days of treatment. Both of these drugs significantly extended survival time by Kaplan-Meier regression (p=0.001 and p=0.03, respectively). Furthermore, scRNASeq data of drug-treated PDXs showed that Dacinostat uniformly depleted all three tumor sub-populations compared to Vehicle control, whereas one of the tumor sub-clusters was resistant to Plicamycin, consistent with single-cell drug sensitivity predicted by OTr.

Given the in vivo activity of these two drugs in inhibiting tumor growth, and the effectiveness of Dacinostat across observed tumor cell phenotypes, as well as the high immune infiltration of this CCA sample, these drugs may be translated into pre-clinical and clinical trials alone and in combination with checkpoint immunotherapy.

Citation Format: Lorenzo Tomassoni, Aleksandar Obradovic, Filemon Dela Cruz, Daoqi Yu, Elise Fraser, Susan E. Bates, Yvonne Saenger, Andrea Califano. Case study of single-cell protein activity-based drug prediction and validation for precision treatment of cholangiocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 4083.

Bromodomain 4 inhibition leads to MYCN downregulation in Wilms tumor

BACKGROUND

Wilms tumor is the most common childhood kidney cancer. Two distinct histological subtypes of Wilms tumor have been described: tumors lacking anaplasia (the favorable subtype) and tumors displaying anaplastic features (the unfavorable subtype). Children with favorable disease generally have a very good prognosis, whereas those with anaplasia are oftentimes refractory to standard treatments and suffer poor outcomes, leading to an unmet clinical need. MYCN dysregulation has been associated with a number of pediatric cancers including Wilms tumor.

PROCEDURES

In this context, we undertook a functional genomics approach to uncover novel therapeutic strategies for those patients with anaplastic Wilms tumor. Genomic analysis and in vitro experimentation demonstrate that cell growth can be reduced by modulating MYCN overexpression via bromodomain 4 (BRD4) inhibition in both anaplastic and nonanaplastic Wilms tumor models.

RESULTS

We observed a time-dependent reduction of MYCN and MYCC protein levels upon BRD4 inhibition in Wilms tumor cell lines, which led to cell death and proliferation suppression. BRD4 inhibition significantly reduced tumor volumes in Wilms tumor patient-derived xenograft (PDX) mouse models.

CONCLUSIONS

We suggest that AZD5153, a novel dual-BRD4 inhibitor, can reduce MYCN levels in both anaplastic and nonanaplastic Wilms tumor cell lines, reduces tumor volume in Wilms tumor PDXs, and should be further explored for its therapeutic potential.

Comprehensive Molecular Profiling of Desmoplastic Small Round Cell Tumor

Desmoplastic small round cell tumor (DSRCT) is characterized by the EWSR1-WT1 t(11;22) (p13:q12) translocation. Few additional putative drivers have been identified, and research has suffered from a lack of model systems. Next-generation sequencing (NGS) data from 68 matched tumor-normal samples, whole-genome sequencing data from 10 samples, transcriptomic and affymetrix array data, and a bank of DSRCT patient-derived xenograft (PDX) are presented. EWSR1-WT1 fusions were noted to be simple, balanced events. Recurrent mutations were uncommon, but were noted in TERT (3%), ARID1A (6%), HRAS (5%), and TP53 (3%), and recurrent loss of heterozygosity (LOH) at 11p, 11q, and 16q was identified in 18%, 22%, and 34% of samples, respectively. Comparison of tumor-normal matched versus unmatched analysis suggests overcalling of somatic mutations in prior publications of DSRCT NGS data. Alterations in fibroblast growth factor receptor 4 (FGFR4) were identified in 5 of 68 (7%) of tumor samples, whereas differential overexpression of FGFR4 was confirmed orthogonally using 2 platforms. PDX models harbored the pathognomic EWSR1-WT1 fusion and were highly representative of corresponding tumors. Our analyses confirm DSRCT as a genomically quiet cancer defined by the balanced translocation, t(11;22)(p13:q12), characterized by a paucity of secondary mutations but a significant number of copy number alterations. Against this genomically quiet background, recurrent activating alterations of FGFR4 stood out, and suggest that this receptor tyrosine kinase, also noted to be highly expressed in DSRCT, should be further investigated. Future studies of DSRCT biology and preclinical therapeutic strategies should benefit from the PDX models characterized in this study. IMPLICATIONS: These data describe the general quiescence of the desmoplastic small round cell tumor (DSRCT) genome, present the first available bank of DSRCT model systems, and nominate FGFR4 as a key receptor tyrosine kinase in DSRCT, based on high expression, recurrent amplification, and recurrent activating mutations.

Abstract 1515: Tracking the origins and drivers of metastatic expansion in osteosarcoma

Objective: Outcomes for patients with osteosarcoma with metastatic or relapsed disease remain poor. Multiple large-scale tumor genomic profiling efforts have been undertaken, however little is known about its spatial intermetastic heterogeneity and temporal clonal evolutionary processes.

Methods: To address this issue, we performed 30-80x whole genome sequencing of nine metastases, three pre-treatment samples, and three on-therapy resection samples from three patients with osteosarcoma. Each patient (OSCE1, OSCE2, OSCE3) had a pre-treatment and on therapy resection of primary site sample as well as at least two sites of metastatic disease sequenced. OSCE1 has multiply relapsed disease and had tissue sequenced from relapse one, two, and four, while OSCE2 and OSCE3 had lethal refractory osteosarcoma. A set of high confidence single nucleotide variants (SNV), copy number alterations (CNA), structural variations (SV), and mutational signatures were called for each sample and an evolutionary analysis was performed using the Treeomics pipeline.

Results: There was diversity in the evolutionary patterns observed across the three patients. OSCE1 and OSCE2 demonstrated linear evolutionary patterns during progression to metastatic disease while OSCE3 demonstrated parallel evolution. Whole-genome duplication events were truncal in all three samples with an average ploidy of 3.3, 3.0, and 3.3 per sample in OSCE1, OSCE2, and OSCE3 respectively. Regarding SNVs, OSCE1 had a clonal TP53, OSCE2 had a clonal SUZ12 mutation, and OSCE3 had a clonal ATRX mutation which was present in all samples. Subclonal mutations in consensus cancer genes were typically private to the individual sample. Copy number changes were a source of genomic diversity in consensus driver genes and between primary and metastatic sites. OSCE1 had persistent KEAP1 and CCNE1 amplification in relapse specimens that were not present in the pretreatment biopsy. OSCE2 had gains or amplifications in MECOM and KRAS which were truncal to all samples however gains in RUNX2 and KDR were private to the metastatic samples. In OSCE3 had focal amplifications in chromosome 7p were truncal to all samples, however focal amplifications in chromosome 12q involving CDK4 and GLI1 were present in only the pretreatment biopsy and one of four metastatic sites. Structural variants involving consensus driver genes were typically truncal to all samples for each patient.

Conclusion: Our study confirms that copy number alterations in osteosarcoma are key oncogenic drivers of treatment resistance and are the largest contributor to intertumoral genetic heterogeneity in consensus driver genes, especially when comparing primary to metastatic samples. Our findings demonstrate the dynamic nature of genomic instability processes, highlighting the importance of longitudinal sampling at the time of recurrence to define treatment effect and the changing mutational landscape in these tumors.

Citation Format: Michael David Kinnaman, Alvin Makohon-Moore, Dominik Glodzik, Max Levine, Nancy Bouvier, Filemon Dela Cruz, Meera Hameed, Paul Meyers, Elli Papaemmanuil, Andrew Kung, Christine Iacobuzio-Donahue. Tracking the origins and drivers of metastatic expansion in osteosarcoma [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1515.

Abstract B70: Clonal evolution of chemotherapy-resistant rhabdomyosarcoma via multifocal genomic analysis of pretreatment and treatment-resistant autopsy specimens

Objective: Outcomes for patients with rhabdomyosarcoma (RMS) who have relapsed or refractory disease remain poor. Multiple large-scale tumor genomic profiling efforts have been undertaken; however, little is known about its spatial intratumoral heterogeneity (ITH) and temporal clonal evolutionary processes.

Methods: To address this issue, we performed 80x whole-genome sequencing of 23 metastases and 2 pretreatment samples from two patients with lethal rhabdomyosarcoma. Multiregion and spatially separated metastases were available through our research autopsy donation program. RA 17-1 had fusion-positive alveolar RMS, while RA 17-10 had Li-Fraumeni syndrome and anaplastic embryonal RMS.

Results: We observed on average 65 and 98 coding single-nucleotide variants (SNV) per sample in RA 17-1 and RA 17-10, respectively. RA 17-1 was found to have the canonical PAX03-FOX01 fusion along with MYC-N amplification, which was truncal to all samples. A somatic TP53 mutation in combination with TP53 loss of heterozygosity (LOH) was present in all metastases sequenced at autopsy. A TP53 mutation was not identified in the pretreatment biopsy sample selected for sequencing; however, TP53 immunohistochemistry demonstrated the presence of a TP53 mutation in at least 2 out of 6 pretreatment biopsy cores. In RA 17-10 each sample exhibited numerous copy number alterations and complex structural variants, including LOH of TP53, a TSC2 mutation with LOH, and MYC amplification, which were truncal to all samples. PTCH1 amplification was private to all relapse specimens in RA 17-10. RA 17-10 had more structural and copy number variations per sample than RA 17-1. Phylogenies were derived based on SNVs and demonstrated a branched evolutionary pattern for each patient. When comparing related samples’ mutational signatures, several dynamics were apparent, including the predominance of signatures 3 and 8 (DNA double-strand break repair pathways) in metastatic samples, consistent with the fact that both patients harbored TP53 mutations.

Conclusion: Defining important elements of ITH within metastases remains a goal for genomic study in pediatrics, especially in patients with germline predisposition syndromes. Our findings demonstrate the dynamic nature of genomic instability processes, highlighting the importance of longitudinal sampling at the time of recurrence to define treatment effect and the changing mutational landscape in these tumors. The case of RA 17-1 demonstrates the limitations of targeted sequencing on a single sample at diagnosis, as the subclonal TP53 mutation was not originally identified on the clinical assay. TP53 mutations in fusion-positive RMS have rarely been described and likely contributed to this patient’s refractory disease course. Both RA 17-1 and RA 17-10 had more structural variants per sample then what has previously been reported for fusion-positive and fusion-negative RMS, respectively, likely highlighting the contribution of the TP53 mutation to genomic instability in each patient.

Citation Format: Michael D. Kinnaman, Alvin Makohon-Moore, Nancy Bouvier, Dominik Glodzik, Max Levine, Ellie Papaemmanuil, Filemon Dela Cruz, Leonard Wexler, Andrew Kung, Christine Iacobuzio-Donahue. Clonal evolution of chemotherapy-resistant rhabdomyosarcoma via multifocal genomic analysis of pretreatment and treatment-resistant autopsy specimens [abstract]. In: Proceedings of the AACR Special Conference on the Advances in Pediatric Cancer Research; 2019 Sep 17-20; Montreal, QC, Canada. Philadelphia (PA): AACR; Cancer Res 2020;80(14 Suppl):Abstract nr B70.

Abstract B31: Multiparameter profiling of abscopal effect (AbE) in a pediatric fibrosarcoma with EML4-NTRK3 fusion

Background: Abscopal effect (AbE) is a rare phenomenon in pediatric oncology, defined as regression of tumor at distal sites following radiation therapy. AbE is widely believed to result from the local activation and expansion of tumor-specific T cells in the area radiated, followed by the migration of these T cells to distant sites. We present a case of AbE in a 1-year-old with congenital fibrosarcoma (forearm) bearing a novel EML4-NTRK3 fusion who developed bilateral lung metastasis 6 months post resection of the initial forearm tumor. The patient’s lung lesions were unresponsive to chemotherapy prompting palliative, hypofractionated radiation therapy (RT) to the right lung lesion only. One month following RT, patient showed significant decrease in size of bilateral lung lesions, with no evidence of disease at 3 months post RT. Using whole-exome sequencing (WES), high-throughput sequencing of T cell receptors (TCRseq), and quantitative multiplex immunofluorescence (qmIF), we investigated the relationship between tumor progression and metastasis, the pre- and post-RT microenvironment, and the co-evolution of the tumor-infiltrating lymphocyte (TIL) populations in this patient. We believe this unusual case offers key insights into mechanisms underlying AbE.

Methods: Formalin-fixed, paraffin-embedded (FFPE) tissue specimens from the initial diagnostic lesion (forearm), pretreatment lung tumor, and bilateral post-RT lung lesions (right (RT treated) and left (outside RT field) were obtained. An FFPE slide was stained for each specimen using traditional immunohistochemistry (IHC) for CD3 and CD8 and qmIF for 7 biomarkers including DAPI, CD3, CD4, CD8, CD68, HLA-DR, and Ki67. QmIF images were taken using MANTRA and cell density analysis was performed using inForm software and R. DNA was isolated, from the tissue specimens and peripheral blood, for profiling using WES and TCRseq, generating a multiplex snapshot of each site. The TIL populations were described by the complementarity determining region 3 (CDR3) of each TCR (as identified by Adaptive), followed by analysis of the bias, diversity, and divergence of these TIL populations using entropy-based statistics.

Results: IHC and qmIF of the primary fibrosarcoma and pre-RT lung tumor showed very low immune cell infiltration. In contrast, immune infiltration was markedly increased (both CD4+ and CD8+) post-RT, particularly in the nonirradiated lung lesion (left). TIL populations were analyzed using TCRseq, revealing little similarity between those of the primary fibrosarcoma and the recurrent lung lesions. However, the TCR repertoires comparing the post-RT lung lesions, irradiated (right) and non-irradiated (left), showed substantial similarity, sharing a larger percentage of CDR3 sequences with each other (6.7% of combined total) than with the primary tumor (≤0.6% combined total) or the pre-RT lung tumor (≤1.6% combined total). Additionally, the extent of the TIL repertoire and its binding-associated divergence from the peripheral blood was greatest in the irradiated (right) lung tumor post RT, suggesting a direct link between the evolution of trans-acting immunity and antigen-driven immunogenicity at that site. Of note, comparative WES of the initial fibrosarcoma and subsequent lung tumors identified the same EML4-NTRK3 fusion, validating their metastatic relationship, although the number of coding mutations had expanded upon recurrence.

Conclusion: Using qmIF and TCRseq, we have extensively characterized immune infiltration in the tumor microenvironment prior to and following metastasis and RT, in the first reported case of AbE in a pediatric tumor. Our data show evidence of immune stimulation elicited by RT, and antitumor T-cell specificities are shared inside and outside of the radiation field. These data offer insight into AbE and how combination of radiation and immunotherapy may be useful to induce response in pediatric tumors.

Citation Format: Robyn D. Gartrell, Jennifer S. Sims, Sunjay M. Barton, Angela Kadenhe-Chiweshe, Filemon Dela Cruz, Andrew T. Turk, Andrew L. Kung, Jeffrey N. Bruce, Yvonne M. Saenger, Darrell J. Yamashiro, Eileen P. Connolly. Multiparameter profiling of abscopal effect (AbE) in a pediatric fibrosarcoma with EML4-NTRK3 fusion [abstract]. In: Proceedings of the AACR Special Conference: Pediatric Cancer Research: From Basic Science to the Clinic; 2017 Dec 3-6; Atlanta, Georgia. Philadelphia (PA): AACR; Cancer Res 2018;78(19 Suppl):Abstract nr B31.

Therapeutic targeting of PGBD5-induced DNA repair dependency in pediatric solid tumors

Despite intense efforts, the cure rates of childhood and adult solid tumors are not satisfactory. Resistance to intensive chemotherapy is common, and targets for molecular therapies are largely undefined. We have found that the majority of childhood solid tumors, including rhabdoid tumors, neuroblastoma, medulloblastoma, and Ewing sarcoma, express an active DNA transposase, PGBD5, that can promote site-specific genomic rearrangements in human cells. Using functional genetic approaches, we discovered that mouse and human cells deficient in nonhomologous end joining (NHEJ) DNA repair cannot tolerate the expression of PGBD5. In a chemical screen of DNA damage signaling inhibitors, we identified AZD6738 as a specific sensitizer of PGBD5-dependent DNA damage and apoptosis. We found that expression of PGBD5, but not its nuclease activity-deficient mutant, was sufficient to induce sensitivity to AZD6738. Depletion of endogenous PGBD5 conferred resistance to AZD6738 in human tumor cells. PGBD5-expressing tumor cells accumulated unrepaired DNA damage in response to AZD6738 treatment and underwent apoptosis in both dividing and G₁-phase cells in the absence of immediate DNA replication stress. Accordingly, AZD6738 exhibited nanomolar potency against most neuroblastoma, medulloblastoma, Ewing sarcoma, and rhabdoid tumor cells tested while sparing nontransformed human and mouse embryonic fibroblasts in vitro. Finally, treatment with AZD6738 induced apoptosis and regression of human neuroblastoma and medulloblastoma tumors engrafted in immunodeficient mice in vivo. This effect was potentiated by combined treatment with cisplatin, including substantial antitumor activity against patient-derived primary neuroblastoma xenografts. These findings delineate a therapeutically actionable synthetic dependency induced in PGBD5-expressing solid tumors.

Abstract 4655: Local drug activation: Making cytotoxics safer for local tumors

Only 1-2% of systemic chemotherapies actually reach a localized tumor while the rest cause myelosuppression and sometimes even cardiotoxicity. There is a critical need to deliver cytotoxic agents to the site of the tumor and reduce overall systemic toxicity. Shasqi's Therapy (Shasqi Tx) is a two-pronged approach based on bioorthogonal chemistry consisting of i) a drug-activating gel and ii) a systemic prodrug. The biocompatible gel, modified with tetrazine (Tz), is injected at tumor site. Then, a trans-cyclooctene (TCO)-modified prodrug of a cytotoxic agent such as doxorubicin (Dox), with attenuated activity, is given systemically. The prodrug concentrates at the tumor site through a covalent cycloaddition reaction between TCO and the Tz located at the gel. The active drug is spontaneously released over multiple days, providing sustained local delivery directly to the tumor compartment. Remaining inactive prodrug is rapidly cleared by systemic clearance routes, thereby minimizing off-target effects. The attenuation of drug activity allows us to maximize the total drug dose while reducing systemic side effects, as the active drug predominantly localizes to the gel site. To this effect, NSG-H mice were injected with Tz-modified alginate gel (TAG) and systemic TCO-doxorubicin (TCO-Dox; prodrug of Dox). The maximum tolerable dose (MTD) of TCO-Dox was >12x the single MTD of Dox (control). When given as 5 daily doses, TCO-Dox dosage could be increased to >38x of Dox MTD, without any adverse effects. Preliminary serum and tissue bioanalysis provides a basis for the improved safety of Shasqi Tx over traditional chemotherapy. TAG-injected BALB/c mice received TCO-Dox intravenously. In serum 5 minutes after injection more than >90% of the circulating anthracycline remained as intact prodrug while ~7% had been converted to Dox. Finally, we tested Shasqi Tx's safety profile in a pilot canine study with a local spontaneous adenocarcinoma. Standard Dox therapy led to disease progression (tumor growth) as well as a drastic drop in body weight. Then, multiple cycles of Shasqi Tx given at higher doses reduced tumor size and led to stable disease for more than 9 months, gain in body weight and without major side effects. Serial evaluations by echocardiogram did not reveal any signs of cardiotoxicity. Collectively, our preliminary toxicity studies suggest that Shasqi Tx enhances delivery of toxic drugs to a target site while limiting exposure in off-target tissues in small and large animals. In addition, the increase in therapeutic index allows greater doses of toxic drugs to be delivered safely.

Citation Format: Sangeetha Srinivasan, Filemon Dela Cruz, Andrew Kung, Ervin Gaviria, Ethan Miller, Nathan Yee, Maksim Royzen, Jose Mejia-Oneto. Local drug activation: Making cytotoxics safer for local tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4655.

Pediatric Sarcoma Data Forms a Unique Cluster Measured via the Earth Mover's Distance

In this note, we combined pediatric sarcoma data from Columbia University with adult sarcoma data collected from TCGA, in order to see if one can automatically discern a unique pediatric cluster in the combined data set. Using a novel clustering pipeline based on optimal transport theory, this turned out to be the case. The overall methodology may find uses for the classification of data from other biological networking problems.

Abstract 4367: Accelerating prediction of tumor vulnerabilities using next-generation cancer models

The mapping of cancer genomes is rapidly approaching completion. The genomic information encoded by individual patients’ tumors should, in principle, provide a guide for predicting dependencies, but our ability to do so is suboptimal. The challenge stems from the absence of clinical data relating genotypes with dependencies since most cancer mutations are rare and our arsenal of cancer drugs is incomplete. If it was possible to build a preclinical ‘cancer dependency map’ at a scale that captured the genomic diversity of cancer (for instance, models of all genotypes tested for genetic and small-molecule dependencies), it should be feasible to improve dependency predictions. New technologies (e.g. CRISPR/Cas9 libraries) make such an effort now feasible. However, we lack a sufficient diversity of cancer models derived directly from patient samples to reflect the genetic diversity of cancer and the ability to systematically create functional data for each cancer patient to expand the map.

In an attempt to overcome these obstacles, we have established an industry-scale pipeline to generate new cancer models directly from patient samples, a “Cancer Cell Line Factory”. We have processed over 620 samples from 400 patients across 16 cancer types through this pipeline with a 25% success rate overall. To optimize conditions for each tumor type, we have systematically compared published cell line generation methods with standard approaches and captured all information with a data management system that will enhance the ability to predict optimal ex vivo propagation conditions for future samples. In all, we report the successful derivation of over 100 new genomically confirmed cancer and normal cell lines, including a series of unique pediatric cancer models derived from rare tumors.

We hypothesized that novel patient-derived cultures could be used to enhance dependency predictions. To test this hypothesis, we tested dependencies of 65 of these novel cultures against an identical set of 440 small molecules that were previously tested against 860 existing cancer cell lines. Our results suggest that dependency data generated with novel cell cultures is potentially backwards-compatible with existing small molecule dependency datasets. Finally, we demonstrate proof-of-concept that such new models can successfully used in CRISPR-Cas9 screens and integrate results with small molecule sensitivities to uncover CDK4 and XPO1 dependencies in a rare pediatric undifferentiated sarcoma. In aggregate, these proof-of-concept studies demarcate a path by which pre-clinical dependency maps may enhance clinical dependency predictions from genomic data alone.

Citation Format: Yuen-Yi Tseng, Andrew Hong, Paula Keskula, Shubhroz Gill, Jaime Cheah, Grigoriy Kryukov, Aviad Tsherniak, Francisca Vazquez, Glenn Cowley, Coyin Oh, Anson Peng, Abeer Sayeed, Rebecca Deasy, Peter Ronning, Philip Kantoff, Levi Garraway, Mark Rubin, Calvin Kuo, Sidharth Puram, Adi Gazdar, Filemon Dela Cruz, Adam Bass, Nikhil Wagle, Keith Ligon, Katherine Janeway, David Root, Stuart Schreiber, Paul Clemons, Aly Shamji, William Hahn, Todd Golub, Jesse S. Boehm. Accelerating prediction of tumor vulnerabilities using next-generation cancer models. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4367.

Solid tumor differentiation therapy - is it possible?

Genetic and epigenetic events within a cell which promote a block in normal development or differentiation coupled with unregulated proliferation are hallmarks of neoplastic transformation. Differentiation therapy involves the use of agents with the ability to induce differentiation in cells that have lost this ability, i.e. cancer cells. The promise of differentiation-based therapy as a viable treatment modality is perhaps best characterized by the addition of retinoids in the treatment of acute promyelocytic leukemia (APML) revolutionizing the management of APML and dramatically improving survival. However, interest and application of differentiationbased therapy for the treatment of solid malignancies have lagged due to deficiencies in our understanding of differentiation pathways in solid malignancies. Over the past decade, a differentiation-based developmental model for solid tumors has emerged providing insights into the biology of various solid tumors as well as identification of targetable pathways capable of re-activating blocked terminal differentiation programs. Furthermore, a variety of agents including retinoids, histone deacetylase inhibitors (HDACI), PPARγ agonists, and others, currently in use for a variety of malignancies, have been shown to induce differentiation in solid tumors. Herein we discuss the relevancy of differentiation-based therapies in solid tumors, using soft tissue sarcomas (STS) as a biologic and clinical model, and review the preclinical data to support its role as a promising modality of therapy for the treatment of solid tumors.

Terminal differentiation and loss of tumorigenicity of human cancers via pluripotency-based reprogramming

Pluripotent cells can be derived from various types of somatic cells by nuclear reprogramming using defined transcription factors. It is, however, unclear whether human cancer cells can be similarly reprogrammed and subsequently terminally differentiated with abrogation of tumorigenicity. Here, using sarcomas we show that human-derived complex karyotype solid tumors: (1) can be reprogrammed into a pluripotent-like state as defined by all in vitro criteria used to define pluripotent stem cells generated from somatic cells; (2) can be terminally differentiated into mature connective tissue and red blood cells; and (3) terminal differentiation is accompanied with loss of both proliferation and tumorigenicity. We go on to perform the first global DNA promoter methylation and gene expression analyses comparing human cancers to their reprogrammed counterparts and report that reprogramming/differentiation results in significant epigenetic remodeling of oncogenes and tumor suppressors, while not significantly altering the differentiation status of the reprogrammed cancer cells, in essence dedifferentiating them to a state slightly before the mesenchymal stem cell differentiation stage. Our data demonstrate that direct nuclear reprogramming can restore terminal differentiation potential to human-derived cancer cells, with simultaneous loss of tumorigenicity, without the need to revert to an embryonic state. We anticipate that our models would serve as a starting point to more fully assess how nuclear reprogramming overcomes the multitude of genetic and epigenetic aberrancies inherent in human cancers to restore normal terminal differentiation pathways. Finally, these findings suggest that nuclear reprogramming may be a broadly applicable therapeutic strategy for the treatment of cancer.

A transgenic, mesodermal specific, Dkk1 mouse model recapitulates a spectrum of human congenital limb reduction defects

Congenital limb reduction defects occurring in isolation of other developmental abnormalities continue to be an important medical problem in which little progress has been made. Herein we generated transgenic mice expressing Dkk1 in an appendicular mesodermal pattern. Prx1-Dkk1 mice recapitulate a full spectrum of human congenital limb reduction defects, without other developmental issues, and have normal life-spans. Importantly, a close examination of the inheritance pattern suggests that there is a significant degree of incomplete penetrance as progeny of phenotypically positive or phenotypically negative, but genotypically positive Prx1-Dkk1 mice, consistently give rise to both phenotypically positive mice and phenotypically normal-appearing mice. Thus, this heterogeneous phenotype is reproducible with each generation regardless of the phenotype of the parents. We further go on to identify that mesenchymal stem cells from Prx1-Dkk1 mice have limited proliferative ability, but normal differentiation potential, which may explain the mechanism for the limb reduction defects observed. We believe Prx1-Dkk1 mice may prove useful in the future to study the mechanisms underlying the development of congenital limb reduction defects.

MicroRNAs in chromosomal translocation-associated solid tumors: learning from sarcomas

Sarcomas are malignant solid tumors of mesenchymal origin which consist of 10-15% of all pediatric malignancies and associated with significantly high mortality rates despite current therapies. Oncogenic fusion genes, resulting from non-random chromosomal translocations, characterize a subset of sarcomas including rhabdomyosarcoma, the Ewing's sarcoma family of tumors, and synovial sarcoma. As investigators gain further insight into the role that fusion genes play in the development and progression of sarcomas, we are slowly uncovering novel molecules and pathways that are proving to be essential for the growth and maintenance of sarcomas and other malignancies. MicroRNAs (miRs) have been implicated in a diversity of human diseases including cancer. Only recently, has miR deregulation been shown to be an important component of sarcomagenesis. This review summarizes the recent discoveries tying miR deregulation to sarcoma biology and will discuss the potential and feasibility of miRs as novel therapeutic adjuncts to current therapies. The methodological approaches utilized in the study of miR biology and development of miR-based treatment regimens can serve as a paradigm for future investigations in other translocation-associated solid tumors.

Malignant Pheochromocytomas and Paragangliomas: A Phase II Study of Therapy with High-Dose 131I-Metaiodobenzylguanidine (131I-MIBG)

Thirty patients with malignant pheochromocytoma (PHEO) or paraganglioma (PGL) were treated with high-dose 131I-MIBG. Patients were 11-62 (mean 39) years old: 19 patients males and 11 females. Nineteen patients had PGL, three of which were multifocal. Six PGLs were nonsecretory. Eleven patients had PHEO. All 30 patients had prior surgery. Fourteen patients were refractory to prior radiation or chemotherapy before 131I-MIBG. Peripheral blood stem cells (PBSCs) were collected and cryopreserved. 131I-MIBG was synthesized on-site, by exchange-labeling 131I with 127I-MIBG in a solid-phase Cu2+-catalyzed exchange reaction. 131I-MIBG was infused over 2 h via a peripheral IV. Doses ranged from 557 mCi to 1185 mCi (7.4 mCi/kg to 18.75 mCi/kg). Median dose was 833 mCi (12.55 mCi/kg). Marrow hypoplasia commenced 3 weeks after 131I-MIBG therapy. After the first 131I-MIBG therapy, 19 patients required platelet transfusions; 19 received GCSF; 12 received epoeitin or RBCs. Four patients received a PBSC infusion. High-dose 131I-MIBG resulted in the following overall tumor responses in 30 patients: 4 sustained complete remissions (CRs); 15 sustained partial remissions (PRs); 1 sustained stable disease (SD); 5 progressive disease (PD); 5 initial PRs or SD but relapsed to PD. Twenty-three of the 30 patients remain alive; deaths were from PD (5), myelodysplasia (1), and unrelated cause (1). Overall predicted survival at 5 years is 75% (Kaplan Meier estimate). For patients with metastatic PHEO or PGL, who have good *I-MIBG uptake on diagnostic scanning, high-dose 131I-MIBG therapy was effective in producing a sustained CR, PR, or SD in 67% of patients, with tolerable toxicity.

A yeast screen system for aromatase inhibitors and ligands for androgen receptor: yeast cells transformed with aromatase and androgen receptor

Endocrine disruptors are hormone mimics that modify hormonal action in humans and animals. It is thought that some endocrine disruptors modify estrogen and androgen action in humans and animals by suppressing aromatase activity. Aromatase cytochrome P450 is the key enzyme that converts C19 androgens to aromatic C18 estrogenic steroids. We have developed a novel aromatase inhibitor screening method that allows us to identify antiaromatase activity of various environmental chemicals. The screen was developed by coexpressing the human aromatase and the mouse androgen receptor in yeast cells, which carry the androgen-responsive ss-galactosidase reporter plasmid. Functional expression of aromatase in yeast has been demonstrated using the [3H]-water release assay with intact cells as well as with yeast microsomes. The aromatase activity could be blocked by known aromatase inhibitors such as aminoglutethimide (AG). Yeast-produced androgen receptors were able to transactivate a yeast basal promoter linked to an androgen-responsive element in response to androgens. The resultant triple yeast transformant responded to the treatment of testosterone, androstenedione, or 5 alpha-dihydrotestosterone (5 alpha-DHT). In the absence of the aromatase inhibitor AG, transcriptional activation was observed only for the nonaromatizable androgen 5 alpha-DHT. However, the two aromatizable androgens (testosterone and androstenedione) induced the reporter activity in the presence of AG. Using this yeast-based assay, we confirmed that two flavones, chrysin and alpha-naphtholflavone, are inhibitors of aromatase. Thus, this yeast system allows us to develop a high-throughput screening method, without using radioactive substrate, to identify aromatase inhibitors as well as new ligands (nonaromatizable androgen mimics) for the androgen receptors. In addition, this screening method also allows us to distinguish nonandrogenic aromatase inhibitors from inhibitors with androgenic activity. This yeast screening method will be useful to screen environmental chemicals for their antiaromatase activity and for their interaction with androgen receptor.