The Myogenesis Program Drives Clonal Selection and Drug Resistance in Rhabdomyosarcoma

Rhabdomyosarcoma (RMS) is a pediatric cancer with features of skeletal muscle; patients with unresectable or metastatic RMS fare poorly due to high rates of disease recurrence. Here, we use single cell and single nucleus RNA-sequencing to show that RMS tumors recapitulate the spectrum of embryonal myogenesis. Using matched patient samples from a clinical trial and orthotopic patient-derived xenografts (O-PDXs), we show chemotherapy eliminates the most proliferative component with features of myoblasts; after treatment, the quiescent immature population with features of paraxial mesoderm expands to reconstitute the developmental hierarchy of the original tumor. We discovered that this paraxial mesoderm population is dependent on EGFR signaling and is sensitive to EGFR inhibitors. Taken together, this data serves as a proof-of-concept that targeting each developmental state in RMS is an effective strategy for improving outcomes by preventing disease recurrence.

Quantitative single-cell proteomics as a tool to characterize cellular hierarchies

Large-scale single-cell analyses are of fundamental importance in order to capture biological heterogeneity within complex cell systems, but have largely been limited to RNA-based technologies. Here we present a comprehensive benchmarked experimental and computational workflow, which establishes global single-cell mass spectrometry-based proteomics as a tool for large-scale single-cell analyses. By exploiting a primary leukemia model system, we demonstrate both through pre-enrichment of cell populations and through a non-enriched unbiased approach that our workflow enables the exploration of cellular heterogeneity within this aberrant developmental hierarchy. Our approach is capable of consistently quantifying ~1000 proteins per cell across thousands of individual cells using limited instrument time. Furthermore, we develop a computational workflow (SCeptre) that effectively normalizes the data, integrates available FACS data and facilitates downstream analysis. The approach presented here lays a foundation for implementing global single-cell proteomics studies across the world.

ETMR-05. SINGLE-CELL RNA-SEQ OF ETMR REVEALS CELL PROGRAMS OF DEVELOPMENTAL HIERARCHY AND CELLULAR DIVERSITY IN THE TUMOR MICROENVIRONMENT

Embryonal tumors with multilayered rosettes (ETMR) are deadly brain malignancies affecting young children. No standard treatment is available and the median survival is less than 12 months. Molecularly, the disease is characterized by the miRNA C19MC cluster amplification, with the expression of multiples miRNAs related to a stem cell program. The discoveries on the purely molecular mechanisms of the disease did not help to create a bridge for new treatment strategies so far and the cellular diversity of ETMR remains poorly understood. In this study, we used single-cell RNA sequencing of murine and human tumors to describe ETMR cellular heterogeneity. Our findings support that intra-tumoral heterogeneity is mainly characterized by 4 cellular programs defining a developmental hierarchy related to different metabolic states: 1) Early quiescent NSC-like cells supported by fatty-acid oxidation 2) Late NSC and NP-like proliferative cells fueled by glycolytic metabolism; 3) Post-mitotic neuroblast-like cells, relying on oxidative-phosphorylation; 4) NSC-like proliferative cells, with metabolic plasticity and capable of performing the three types of metabolism. Tumor-specific ligand-receptor interaction analysis revealed that ETMR exchange with microglia and vascular mural cells (MC) signals related to extracellular matrix (ECM) organization (Cxcl12-CxCr4), stem cell signaling (BMPs-BMP receptors), anti-apoptosis and survival (Ntf3-Ntrk), not seen in the control brain. In addition, the vascular MC showed a cancer-associated fibroblast (CAF) phenotype, with potential prognostic implications, as previously demonstrated for other tumors. This study provides new findings to build up a more robust understanding of ETMR biology and opens space for further studies in the field.

Single Cell RNA-Seq Reveals Deranged Developmental Hierarchy with Coexisting Oncogenic States and Immune Evasion Programs in ETP T-ALL

Introduction: Early T-cell precursor acute lymphoblastic leukemia (ETP T-ALL) is a distinct subtype of T-ALL characterized by higher rates of relapse and induction failure. Large-scale genetic sequencing studies have identified frequently mutated oncogenes and gene fusions in ETP T-ALL, while bulk transcriptome analyses have revealed expression features resembling myeloid precursors and myeloid malignancies. However, the contributions of intra-tumoral functional heterogeneity and microenvironment to tumor biology and treatment failure remain unknown. Methods: We performed full-length single-cell RNA-sequencing of 5,077 malignant and normal immune cells from bone marrow or blood from five patients with relapsed/refractory ETP T-ALL (based on immunophenotyping, all with NOTCH1 mutations), before and after targeted therapy against NOTCH1. These patients were enrolled on a phase I trial with the γ-secretase inhibitor (GSI) BMS-906024 (NCT01363817). Expression of selected genes was validated by RT-PCR, flow cytometry and immunohistochemistry. Results: Single cell transcriptome analyses revealed a deranged developmental hierarchy characterized by co-expression of stemness programs in multiple malignant cells implying ineffectual commitment to either lymphoid or myeloid lineage. Most ETP T-ALL cells co-expressed HSC (hematopoietic stem cell), CMP (common myeloid progenitor) and CLP (common lymphoid progenitor) signatures simultaneously (Pearson correlation: CLP-CMP: R= 0.41, p < 2.2e-16; HSC-CLP: R= 0.53; p < 2.2e-16; HSC-CMP: R = 0.39, p <2.2e-16). Only a fraction of cells (less than 15%) demonstrated mutually exclusive CLP or HSC signatures. In contrast, CLP, CMP and HSC signatures were not co-expressed and always negatively correlated in normal bone marrow cells (CLP-CMP: R= -0.11, p < 2.2e-16; HSC-CLP: R= -0.38; p < 2.2e-16; HSC-CMP: R = -0.67, p <2.2e-16). Direct targeting of NOTCH1 as the driving oncogene has shown disappointing results in the clinical setting due to the rapid development of resistance. PI3K activation has been shown as a genetic mechanism of Notch resistance, however it is unclear if transcriptional rewiring can give rise to PI3K dependent cells after Notch inhibition. To address this question, we predicted the activity of signaling pathways in single cells after Notch inhibitor treatment using PROGENy. Most single cells demonstrated loss of Notch signaling. PI3K signaling activity was the most anti-correlated signaling pathway to Notch signaling (Pearson correlation: R= -0.51, p < 2.2e-16). Of note, this population preexisted at a frequency of ~30% in the untreated population, coexisting with cells with high Notch activation. Analysis of the immune microenvironment revealed an oligoclonal T-cell population in ETP T-ALL compared to normal donor T-cells. CD8+ T-cells from ETP patients expressed markers of T-cell exhaustion (PDCD1, TIGIT, LAG3, HAVCR2). Analyses of expression levels of the respective ligands on leukemic blasts and the predicted interaction with their receptors on endogenous CD8+ T-cells demonstrated the highest interaction score between HAVCR2 and its ligand LGALS9. LGALS9 was universally expressed in all leukemic cells, which was confirmed by flow cytometry staining in leukemic blasts and IHC staining in bone marrow of 8 patients with ETP T-ALL and 7 patients with T-ALL. T-ALL supernatant increased expression levels of the exhaustion markers HAVCR2,TIGIT and decreased effector marker GZMB in polyclonal activated normal donor CD8+ T-cells (RT-PCR). This effect was abrogated by neutralizing LGALS9 and could be rescued with recombinant LGALS9. Conclusion: We identified deranged developmental hierarchy characterized by co-expression of stemness programs in multiple malignant cell states and ineffectual commitment to either lymphoid or myeloid lineage in ETP T-ALL. Leukemic blasts demonstrate preexisting heterogeneity of diverse oncogenic states as evidenced by opposing PI3K and Notch activity, suggesting possible novel combination therapies. Notch inhibition abolishes the Notch high state without effecting the PI3K active state. Finally, we demonstrate a possible role for HAVCR2-LGALS9 interactions in causing CD8+ T-cell dysfunction in ETP T-ALL patients, which may provide a novel therapeutic strategy in this disease. Disclosures Silverman: Takeda: Consultancy; Servier: Consultancy, Research Funding. Lane:AbbVie: Research Funding; Stemline Therapeutics: Research Funding; N-of-One: Consultancy. DeAngelo:Glycomimetics: Research Funding; Amgen, Autolus, Celgene, Forty-seven, Incyte, Jazzs, Pfizer, Shire, Takeda: Consultancy; Blueprint: Consultancy, Research Funding; Novartis: Consultancy, Research Funding; Abbvie: Research Funding. Lohr:Celgene: Research Funding; T2 Biosystems: Honoraria.

PDTM-21. MULTI-MODAL PROFILING OF PEDIATRIC HIGH-GRADE GLIOMA SINGLE CELLS USING PATCH-SEQ

Pediatric and adult high-grade gliomas are characterized by extensive intra-tumoral transcriptional heterogeneity. When measured by single cell RNA sequencing, gliomas reveal themselves as continuums of stemness and differentiation programs with important implications for therapy, but to date this transcriptional information has not been directly linked to physiological behaviors of cells. Recent work from our group establishes the electrical integration of glioma cells into neural circuitry. One subpopulation of glioma cells participates in glutamatergic synaptic communication with neurons, and a distinct subpopulation of cells sense and respond to extracellular potassium flux of neuronal networks by an entirely distinct mechanism. Our data support a model in which both modes of electrical communication are critical to glioma growth, but current associations between the electrophysiological properties of a cell, its transcriptional profile and developmental state are correlational in nature. Patch-seq is needed to clarify the relationship between transcriptional profiles of quiescent/cycling stem-like cells and the observed electrophysiological behaviors. Put more simply, patch-seq will clarify where the synaptically-connected glioma cells exist along a developmental hierarchy. METHODS Here, we adapt a recently described technique called patch-seq to record the electrophysiological profiles of individual pediatric high-grade glioma cells by whole cell patch-clamp and subsequently isolate their mRNA for single cell sequencing by smart-seq2 and analysis using Seurat. In this way, we couple electrophysiological and transcriptomic profiles to unambiguously assign functional identities to cells with transcriptional profiles along a developmental hierarchy. RESULTS We report the successful adaptation of patch-seq for use with patient-derived diffuse intrinsic pontine glioma (DIPG) xenografts in acute brain slice preparations, enabling evaluation of single glioma cells integrated in intact neural circuitry. CONCLUSIONS Data synthesizing the electrophysiological and transcriptomic profiles of single glioma cells in the context of the developmental hierarchy will be presented.

PDTM-08. SINGLE-CELL GENOMICS REVEALS TUMOR-INITIATING PROGENITORS AND ONCOGENIC CASCADES DURING TUMORIGENESIS AND RELAPSE IN MEDULLOBLASTOMA

Progenitor heterogeneity and identities underlying tumor initiation and relapse in medulloblastomas, the most common malignant pediatric brain tumor, remain elusive. Here, by utilizing single-cell analysis at different stages of tumorigenesis, we demonstrated a developmental hierarchy of diverse progenitor pools in sonic hedgehog (SHH)-medulloblastomas. Unexpectedly, we identified Olig2-expressing progenitors as transit-amplifying cells at the onset of tumorigenesis. Although Olig2+ cells become quiescent stem-like progenitors in full-blown tumors, they are highly enriched in therapy-resistant and recurrent medulloblastomas. High-level OLIG2 expression is associated with poor outcome in human SHH-medulloblastomas. Depletion of mitotic Olig2+ progenitors or Olig2-ablation impeded tumorigenesis. Transcriptome and chromatin-occupancy profiling revealed that Olig2 modulates the chromatin landscape and activates oncogenic networks including HIPPO-Yap/Taz and Aurora-A/MycN pathways. Co-targeting these oncogenic pathways induced tumor growth arrest. Together, our results raise the unanticipated possibility that glial lineage-associated Olig2-expressing progenitors are tumor-initiating cells during medulloblastoma tumorigenesis and relapse, suggesting Olig2-driven oncogenic networks as potential therapeutic targets.