Adult Human Glioblastomas Harbor Radial Glia-like Cells

Radial glia (RG) cells are the first neural stem cells to appear during embryonic development. Adult human glioblastomas harbor a subpopulation of RG-like cells with typical RG morphology and markers. The cells exhibit the classic and unique mitotic behavior of normal RG in a cell-autonomous manner. Single-cell RNA sequencing analyses of glioblastoma cells reveal transcriptionally dynamic clusters of RG-like cells that share the profiles of normal human fetal radial glia and that reside in quiescent and cycling states. Functional assays show a role for interleukin in triggering exit from dormancy into active cycling, suggesting a role for inflammation in tumor progression. These data are consistent with the possibility of persistence of RG into adulthood and their involvement in tumor initiation or maintenance. They also provide a putative cellular basis for the persistence of normal developmental programs in adult tumors.

Regenerative lineages and immune-mediated pruning in lung cancer metastasis

Developmental processes underlying normal tissue regeneration have been implicated in cancer, but the degree of their enactment during tumor progression and under the selective pressures of immune surveillance, remain unknown. Here, we show that human primary lung adenocarcinomas are characterized by the emergence of regenerative cell types typically seen in response to lung injury, and by striking infidelity amongst transcription factors specifying most alveolar and bronchial epithelial lineages. In contrast, metastases are enriched for key endoderm and lung-specifying transcription factors, SOX2 and SOX9, and recapitulate more primitive transcriptional programs spanning stem-like to regenerative pulmonary epithelial progenitor states. This developmental continuum mirrors the progressive stages of spontaneous outbreak from metastatic dormancy in a mouse model and exhibits SOX9-dependent resistance to Natural Killer (NK) cells. Loss of developmental stage-specific constraint in macrometastases triggered by NK cell depletion suggests a dynamic interplay between developmental plasticity and immune-mediated pruning during metastasis.

L1CAM defines the regenerative origin of metastasis-initiating cells in colorectal cancer

Metastasis-initiating cells with stem-like properties drive cancer lethality, yet their origins and relationship to primary-tumor-initiating stem cells are not known. We show that L1CAM⁺ cells in human colorectal cancer (CRC) have metastasis-initiating capacity, and we define their relationship to tissue regeneration. L1CAM is not expressed in the homeostatic intestinal epithelium, but is induced and required for epithelial regeneration following colitis and in CRC organoid growth. By using human tissues and mouse models, we show that L1CAM is dispensable for adenoma initiation but required for orthotopic carcinoma propagation, liver metastatic colonization and chemoresistance. L1CAM^high cells partially overlap with LGR5^high stem-like cells in human CRC organoids. Disruption of intercellular epithelial contacts causes E-cadherin-REST transcriptional derepression of L1CAM, switching chemoresistant CRC progenitors from an L1CAM^low to an L1CAM^high state. Thus, L1CAM dependency emerges in regenerative intestinal cells when epithelial integrity is lost, a phenotype of wound healing deployed in metastasis-initiating cells.

Immune profiling of human tumors identifies CD73 as a combinatorial target in glioblastoma

Immune checkpoint therapy with anti-CTLA-4 and anti-PD-1/PD-L1 has revolutionized the treatment of many solid tumors. However, the clinical efficacy of immune checkpoint therapy is limited to a subset of patients with specific tumor types^1,2. Multiple clinical trials with combinatorial immune checkpoint strategies are ongoing; however, the mechanistic rationale for tumor-specific targeting of immune checkpoints is elusive. To garner an insight into tumor-specific immunomodulatory targets, we analyzed 94 patients representing five different cancer types, including those that respond relatively well to immune checkpoint therapy and those that do not, such as glioblastoma multiforme, prostate cancer and colorectal cancer. Through mass cytometry and single-cell RNA sequencing, we identified a unique population of CD73^hi macrophages in glioblastoma multiforme that persists after anti-PD-1 treatment. To test if targeting CD73 would be important for a successful combination strategy in glioblastoma multiforme, we performed reverse translational studies using CD73^-/- mice. We found that the absence of CD73 improved survival in a murine model of glioblastoma multiforme treated with anti-CTLA-4 and anti-PD-1. Our data identified CD73 as a specific immunotherapeutic target to improve antitumor immune responses to immune checkpoint therapy in glioblastoma multiforme and demonstrate that comprehensive human and reverse translational studies can be used for rational design of combinatorial immune checkpoint strategies.

Transcriptional Basis of Mouse and Human Dendritic Cell Heterogeneity

Dendritic cells (DCs) play a critical role in orchestrating adaptive immune responses due to their unique ability to initiate T cell responses and direct their differentiation into effector lineages. Classical DCs have been divided into two subsets, cDC1 and cDC2, based on phenotypic markers and their distinct abilities to prime CD8 and CD4 T cells. While the transcriptional regulation of the cDC1 subset has been well characterized, cDC2 development and function remain poorly understood. By combining transcriptional and chromatin analyses with genetic reporter expression, we identified two principal cDC2 lineages defined by distinct developmental pathways and transcriptional regulators, including T-bet and RORγt, two key transcription factors known to define innate and adaptive lymphocyte subsets. These novel cDC2 lineages were characterized by distinct metabolic and functional programs. Extending our findings to humans revealed conserved DC heterogeneity and the presence of the newly defined cDC2 subsets in human cancer.

•

Single-cell analyses reveal novel dendritic cell subsets

•

Major cDC2 subsets differentially express T-bet and RORγt

•

Distinct pro- and anti-inflammatory potential of T-bet⁺ and Tbet^– cDC2s

•

Transcriptional basis for cDC2 heterogeneity conserved across mouse and human

Engineering γδT cells limits tonic signaling associated with chimeric antigen receptors

Despite the benefits of chimeric antigen receptor (CAR)-T cell therapies against lymphoid malignancies, responses in solid tumors have been more limited and off-target toxicities have been more marked. Among the possible design limitations of CAR-T cells for cancer are unwanted tonic (antigen-independent) signaling and off-target activation. Efforts to overcome these hurdles have been blunted by a lack of mechanistic understanding. Here, we showed that single-cell analysis with time course mass cytometry provided a rapid means of assessing CAR-T cell activation. We compared signal transduction in expanded T cells to that in T cells transduced to express second-generation CARs and found that cell expansion enhanced the response to stimulation. However, expansion also induced tonic signaling and reduced network plasticity, which were associated with expression of the T cell exhaustion markers PD-1 and TIM-3. Because this was most evident in pathways downstream of CD3ζ, we performed similar analyses on γδT cells that expressed chimeric costimulatory receptors (CCRs) lacking CD3ζ but containing DAP10 stimulatory domains. These CCR-γδT cells did not exhibit tonic signaling but were efficiently activated and mounted cytotoxic responses in the presence of CCR-specific stimuli or cognate leukemic cells. Single-cell signaling analysis enabled detailed characterization of CAR-T and CCR-T cell activation to better understand their functional activities. Furthermore, we demonstrated that CCR-γδT cells may offer the potential to avoid on-target, off-tumor toxicity and allo-reactivity in the context of myeloid malignancies.

Cohesin Members Stag1 and Stag2 Display Distinct Roles in Chromatin Accessibility and Topological Control of HSC Self-Renewal and Differentiation

Transcriptional regulators, including the cohesin complex member STAG2, are recurrently mutated in cancer. The role of STAG2 in gene regulation, hematopoiesis, and tumor suppression remains unresolved. We show that Stag2 deletion in hematopoietic stem and progenitor cells (HSPCs) results in altered hematopoietic function, increased self-renewal, and impaired differentiation. Chromatin immunoprecipitation (ChIP) sequencing revealed that, although Stag2 and Stag1 bind a shared set of genomic loci, a component of Stag2 binding sites is unoccupied by Stag1, even in Stag2-deficient HSPCs. Although concurrent loss of Stag2 and Stag1 abrogated hematopoiesis, Stag2 loss alone decreased chromatin accessibility and transcription of lineage-specification genes, including Ebf1 and Pax5, leading to increased self-renewal and reduced HSPC commitment to the B cell lineage. Our data illustrate a role for Stag2 in transformation and transcriptional dysregulation distinct from its shared role with Stag1 in chromosomal segregation.

Abstract 3371: Mapping the evolution of T cell states during DLI response and resistance using single-cell data and computational tools

Donor lymphocyte infusion (DLI) is a standard of care and potentially curative immunotherapy for relapsed leukemia after allogeneic hematopoietic stem cell transplant (allo-SCT). Despite low response rates for many leukemias, chronic myelogenous leukemia (CML) has historically exhibited an exquisite DLI sensitivity, and we previously reported that durable response to DLI was associated with reversal of exhaustion of bone marrow (BM) -infiltrating T cells (Bachireddy et al., Blood 2014). Critical questions remain, however, regarding the exact transcriptional states of those T cell subtypes mediating exhaustion, anti-leukemia responses, and resistance to DLI.

To map evolving phenotypic T cell states in situ at single cell resolution, we profiled viable cells isolated from cryopreserved BM mononuclear cells from a median of 3 timepoints before and after DLI from 12 patients with relapsed CML after allo-SCT, including 6 long-term responders to DLI (R’s) and 6 nonresponders (NR’s), using single cell RNA sequencing (scRNA-seq). Using our computational tools for processing and analyzing scRNA-seq data (Azizi et al., Cell 2018), we detected 381,462 cells in total derived from 43 unique patient-timepoints that met our quality metrics.

Because DLI’s anti-leukemic efficacy derives in large part from T cell activity, we sought a more refined characterization of T cells using our tool Biscuit (Azizi et al., Cell 2018) to merge, normalize and cluster T cells. We observed a marked increase in the number of T cell clusters in post-DLI samples compared to matched pre-DLI samples (p<0.001). Both R and NR cases exhibited increases in phenotypic volume induced by DLI (p<1x10-6), suggesting DLI induces multiple, independent gene expression components in both clinical outcomes. However, at both pre- and post-DLI timepoints, phenotypic volumes in R cases were higher than that of NR cases.

Comparing T cell exhaustion between R vs NR cases, we confirmed increased T cell exhaustion signatures in R-pre T cells. Using factor analysis techniques we found that anergy, dysfunction and tolerance are shared factors driving a subset of T cells that are enriched in NRs, suggesting multiple forms of T cell dysfunction in DLI resistance. We found that the clusters dominated by post-DLI R T cells were characterized by greater diversity of T helper subsets (Th1, Tfh, Th2, Th9, and Th22) and enrichment for exhaustion, type I and II IFN pathways, proinflammatory gene sets and CD8 T cell activation. Clusters dominated by NR T cells displayed increases in Th17 and Treg signatures, anergy and tolerance.

These data suggest that (1) pretreatment T cell phenotypic diversity may be important for DLI response; (2) that DLI increases such diversity differently in R’s than in NR’s; (3) while T cell subsets exhibit some overlap pre-therapy, responders and non-responders become increasingly dissimilar post therapy.

Citation Format: Elham Azizi, Pavan Bachireddy, Vinhkhang N. Nguyen, Shuqiang Li, Donna S. Neuberg, Robert J. Soiffer, Jerome Ritz, Edwin P. Alyea III, Dana Pe'er, Catherine J. Wu. Mapping the evolution of T cell states during DLI response and resistance using single-cell data and computational tools [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3371.

Tissue-based molecular and histological landscape of acquired resistance to osimertinib given initially or at relapse in patients with EGFR-mutant lung cancers

9028

Background: Even though osimertinib (osi) is now the initial treatment for patients with EGFR-mutant lung cancers, our knowledge about mechanisms of resistance (MOR) is largely derived from patients who received osi after developing acquired resistance to initial EGFR inhibitor. Further, studies of osi resistance to date have mainly reported genotyping of plasma which suboptimally detects lineage plasticity, copy number changes, and chromosomal rearrangements. Methods: To identify MOR to osi and characterize clinical, molecular and histologic factors associated with duration of response, we identified patients with EGFR-mutant lung cancers who had next-generation sequencing performed on tumor tissue after developing acquired resistance to osi. Results: From January 2016 to December 2018, post-osi tumor tissue was collected from 71 patients (42 with paired pre-treatment specimens). See mechanisms of resistance below. Histologic transformation was identified in 19% of initial cases and 14% of all cases. When osi is given as initial treatment, with median follow up of 17 months, early emerging MOR rarely included on-target resistance mechanisms (1/16 cases of acquired EGFR G724S). Acquired alterations representing potential resistance mechanisms included CCNE1 and MYC amplifications, and mutations in MTOR and MET H1094Y. We confirmed in preclinical studies that an amino acid substitution at MET H1094 can reduce sensitivity to osi. Conclusions: In this analysis of MOR identified on NGS from tumor tissue, we found a different spectrum of resistance mechanisms to initial and later-line osi, with histologic transformation (including squamous cell transformation) a dominant MOR, particularly in the first-line setting, that cannot be identified on plasma testing. Subsequent studies are needed to assess patients with a longer time on initial osi as there may be a temporal bias to MOR, with off-target MOR emerging earlier and on-target resistance mutations later. [Table: see text]

Natural Genetic Variation Reveals Key Features of Epigenetic and Transcriptional Memory in Virus-Specific CD8 T Cells

Stable changes in chromatin states and gene expression in cells of the immune system form the basis for memory of infections and other challenges. Here, we used naturally occurring cis-regulatory variation in wild-derived inbred mouse strains to explore the mechanisms underlying long-lasting versus transient gene regulation in CD8 T cells responding to acute viral infection. Stably responsive DNA elements were characterized by dramatic and congruent chromatin remodeling events affecting multiple neighboring sites and required distinct transcription factor (TF) binding motifs for their accessibility. Specifically, we found that cooperative recruitment of T-box and Runx family transcription factors to shared targets mediated stable chromatin remodeling upon T cell activation. Our observations provide insights into the molecular mechanisms driving virus-specific CD8 T cell responses and suggest a general mechanism for the formation of transcriptional and epigenetic memory applicable to other immune and non-immune cells.

Characterization of cell fate probabilities in single-cell data with Palantir

Single-cell RNA sequencing studies of differentiating systems have raised fundamental questions regarding the discrete versus continuous nature of both differentiation and cell fate. Here we present Palantir, an algorithm that models trajectories of differentiating cells by treating cell fate as a probabilistic process and leverages entropy to measure cell plasticity along the trajectory. Palantir generates a high-resolution pseudo-time ordering of cells and, for each cell state, assigns a probability of differentiating into each terminal state. We apply our algorithm to human bone marrow single-cell RNA sequencing data and detect important landmarks of hematopoietic differentiation. Palantir's resolution enables the identification of key transcription factors that drive lineage fate choice and closely track when cells lose plasticity. We show that Palantir outperforms existing algorithms in identifying cell lineages and recapitulating gene expression trends during differentiation, is generalizable to diverse tissue types, and is well-suited to resolving less-studied differentiating systems.

Negative Co-stimulation Constrains T Cell Differentiation by Imposing Boundaries on Possible Cell States

Co-stimulation regulates T cell activation, but it remains unclear whether co-stimulatory pathways also control T cell differentiation. We used mass cytometry to profile T cells generated in the genetic absence of the negative co-stimulatory molecules CTLA-4 and PD-1. Our data indicate that negative co-stimulation constrains the possible cell states that peripheral T cells can acquire. CTLA-4 imposes major boundaries on CD4⁺ T cell phenotypes, whereas PD-1 subtly limits CD8⁺ T cell phenotypes. By computationally reconstructing T cell differentiation paths, we identified protein expression changes that underlied the abnormal phenotypic expansion and pinpointed when lineage choice events occurred during differentiation. Similar alterations in T cell phenotypes were observed after anti-CTLA-4 and anti-PD-1 antibody blockade. These findings implicate negative co-stimulation as a key regulator and determinant of T cell differentiation and suggest that checkpoint blockade might work in part by altering the limits of T cell phenotypes.

Epigenomic-Guided Mass Cytometry Profiling Reveals Disease-Specific Features of Exhausted CD8 T Cells

Exhausted CD8 T (Tex) cells are immunotherapy targets in chronic infection and cancer, but a comprehensive assessment of Tex cell diversity in human disease is lacking. Here, we developed a transcriptomic- and epigenetic-guided mass cytometry approach to define core exhaustion-specific genes and disease-induced changes in Tex cells in HIV and human cancer. Single-cell proteomic profiling identified 9 distinct Tex cell clusters using phenotypic, functional, transcription factor, and inhibitory receptor co-expression patterns. An exhaustion severity metric was developed and integrated with high-dimensional phenotypes to define Tex cell clusters that were present in healthy subjects, common across chronic infection and cancer or enriched in either disease, linked to disease severity, and changed with HIV therapy. Combinatorial patterns of immunotherapy targets on different Tex cell clusters were also defined. This approach and associated datasets present a resource for investigating human Tex cell biology, with implications for immune monitoring and immunomodulation in chronic infections, autoimmunity, and cancer.

CD49b defines functionally mature Treg cells that survey skin and vascular tissues

Fan et al. have identified a population of recirculating Treg cells with greater suppressive ability and a unique tissue distribution. Using single-cell RNA-seq, they place these Treg cells at the apex of the Treg developmental trajectory and show that similar cells may exist in humans.

Regulatory T (Treg) cells prevent autoimmunity by limiting immune responses and inflammation in the secondary lymphoid organs and nonlymphoid tissues. While unique subsets of Treg cells have been described in some nonlymphoid tissues, their relationship to Treg cells in secondary lymphoid organs and circulation remains unclear. Furthermore, it is possible that Treg cells from similar tissue types share largely similar properties. We have identified a short-lived effector Treg cell subset that expresses the α₂ integrin, CD49b, and exhibits a unique tissue distribution, being abundant in peripheral blood, vasculature, skin, and skin-draining lymph nodes, but uncommon in the intestines and in viscera-draining lymph nodes. CD49b⁺ Treg cells, which display superior functionality revealed by in vitro and in vivo assays, appear to develop after multiple rounds of cell division and TCR-dependent activation. Accordingly, single-cell RNA-seq analysis placed these cells at the apex of the Treg developmental trajectory. These results shed light on the identity and development of a functionally potent subset of mature effector Treg cells that recirculate through and survey peripheral tissues.

Abstract 4990: Regenerative origin of colorectal metastasis stem cells

Metastatic cancers invariably relapse due to the emergence of resistant tumor clones capable of self-renewal, entry into and exit from quiescence, tumor re-initiation and therapy resistance. The origins of such metastasis propagating cells (MPCs), which ultimately cause cancer death, are not well-understood.

To directly scrutinize MPCs in patient metastases, we established ex vivo organoid cultures from surgically resected, chemoresistant residual colorectal cancer (CRC) liver metastases. We show that the neuronal cell-adhesion molecule L1CAM, which is ectopically expressed in many cancer types and strongly associated with poor prognosis, is a marker of MPCs. L1CAM+ cells are largely quiescent in structured neoplastic glands in tumors, but when dissociated from their epithelial niche, proliferate to regenerate heterogeneous organoids or xenografts containing both L1CAM+ and L1CAM- progeny.

To define the relationship between L1CAM+ MPCs and Lgr5+ intestinal stem cells, we performed single cell mRNA sequencing on ~15,000 CRC organoid-derived cells from four patients. We identified only partial overlap between Lgr5+ and L1CAM+ cells. Lgr5high cells consistent with homeostatic stem cells, have low L1CAM levels, while Lgr5low transit amplifying progenitor-like cells have high L1CAM levels. In addition, we identify a separate population of L1CAMhighLgr5- cells. The data suggest that human CRC metastases are derived from an L1CAM+ population of transit-amplifying, partially differentiated cells.

L1CAM is not expressed in intact human or mouse intestinal crypts during homeostasis. However, when the intestinal epithelium is disrupted by dextran sodium sulfate-mediated colitis, L1CAM is strongly induced in cells in the middle of regenerating crypts. Intestinal epithelium specific deletion of L1CAM causes profound weight loss, poor tissue healing and reduces survival in DSS-treated mice. In turn, L1CAM knockdown/knockout in mouse or human CRC cells inhibits regeneration of organoids in vitro, subcutaneous tumors and orthotopic liver metastases in vivo. Mechanistically, L1CAM RNA expression is normally silenced in non-neuronal cells by the transcriptional repressor REST. We show that disruption of epithelial integrity by organoid dissociation or E-cadherin knockdown reduces REST binding to an L1CAM intronic enhancer, thus inducing L1CAM expression.

Our results suggest that L1CAM is dispensable for epithelial homeostasis, but is required for normal and neoplastic epithelial regeneration when tissue integrity is disrupted. During cancer progression, disseminated tumor cells at the invasion front of primary tumors, in the circulation, or in isolated residual disease following therapy, induce and depend on L1CAM for survival and eventual regrowth. Thus, L1CAM represents a crucial vulnerability of disseminated and residual MPCs that could be exploited therapeutically to treat patients with metastatic cancer.

Citation Format: Karuna Ganesh, Harihar Basnet, Kevin P. O'Rourke, Ashley M. Laughney, Lan He, Eduard Batlle, Scott W. Lowe, Dana Pe'er, Jinru Shia, Joan Massague. Regenerative origin of colorectal metastasis stem cells [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 4990.

Abstract NG03: Chromosomal instability promotes metastasis through a cytosolic DNA response

Chromosomal instability (CIN) is a hallmark of cancer and it results from ongoing errors in chromosome segregation during mitosis. While CIN is a major driver of tumor evolution, its role in metastasis has not been established. Here we show that CIN promotes metastasis by sustaining a tumor-cell autonomous response to cytosolic DNA. Errors in chromosome segregation create a preponderance of micronuclei whose rupture spills genomic DNA into the cytosol. This leads to the activation of the cGAS-STING cytosolic DNA-sensing pathway and downstream noncanonical NF-κB signaling. Genetic suppression of CIN significantly delays metastasis even in highly aneuploid tumor models, whereas inducing continuous chromosome segregation errors promotes cellular invasion and metastasis in a STING-dependent manner. Using single-cell RNA sequencing, we uncover a CIN-induced transcriptional switch from a proliferative and metabolically active state to a mesenchymal phenotype associated with inflammatory pathways, offering an opportunity to target chromosome segregation errors for therapeutic benefit. Our work reveals an unexpected link between CIN, cytosolic DNA sensing pathways, and metastasis. The use of an isogenic system has enabled us to dissect the role of CIN from that of aneuploidy. Importantly, while we do not discount the role of CIN in generating karyotypic heterogeneity that can serve as the substrate for natural selection, our work demonstrates that continuous chromosome missegregation is also required to replenish cytosolic DNA pools leading to chronic upregulation of inflammatory pathways. In non-transformed settings, cytosolic DNA sensing is incompatible with viability. Unlike normal cells, chromosomally unstable cells are awash with cytosolic DNA and have adapted to coexist with a chronically active cGAS-STING pathway by suppressing downstream type I interferon signaling and instead upregulating the alternative NF-κB pathway. Persistent STING activation mediates carcinogen-induced tumor formation and we now show that tumor cells co-opt this otherwise lethal program to spread to distant organs. The evolutionary benefit of the noncanonical pathway might justify the scarcity of inactivating mutations in cGAS and STING among human cancers. The emergence, and subsequent tolerance, of CIN represents an important bottleneck during tumor evolution. Our single-cell analysis revealed that CIN induces a transcriptional switch whereby cells shift from a proliferative and highly metabolic state, ideally suited for primary tumor growth, to a chromosomally unstable and mesenchymal state associated with upregulation of inflammatory pathways. These two largely mutually exclusive states likely account for the reversibility in chromosome missegregation rates seen in primary tumors and metastases, and provide an explanation for the negative effect of aneuploidy during early tumorigenesis. Interestingly, this mutual exclusivity was recently observed in a pan-cancer genomic analysis of metastatic tumors, and it leads us to suggest that CIN underlies the subset of metastases that are characterized by EMT and inflammation. By providing a mechanistic link between CIN and metastasis, our work opens new avenues to target chromosomally unstable tumors for therapeutic benefit.

Citation Format: Samuel F. Bakhoum, Bryon Ngo, Ashley L. Bakhoum, Julie-Ann Cavallo, Charles J. Murphy, Peter Ly, Pragya Shah, Roshan K. Sriram, Thomas B.k. Watkins, Neil K. Taunk, Mercedes Duran, Chantal Pauli, Christine Shaw, Kalyani Chadalavada, Vinagolu K. Rajasekhar, Giulio Genovese, Subramanian Venkatesan, Nicolai J. Birkbak, Nicholas McGranahan, Mark Lundquist, Quincy LaPlant, John H. Healey, Olivier Elemento, Christine H. Chung, Nancy Y. Lee, Marcin Imielinski, Gouri Nanjangud, Dana Pe'er, Don W. Cleveland, Simon N. Powell, Jan Lammerding, Charles Swanton, Lewis C. Cantley. Chromosomal instability promotes metastasis through a cytosolic DNA response [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 NG03.

Recovering Gene Interactions from Single-Cell Data Using Data Diffusion

Single-cell RNA sequencing technologies suffer from many sources of technical noise, including under-sampling of mRNA molecules, often termed "dropout," which can severely obscure important gene-gene relationships. To address this, we developed MAGIC (Markov affinity-based graph imputation of cells), a method that shares information across similar cells, via data diffusion, to denoise the cell count matrix and fill in missing transcripts. We validate MAGIC on several biological systems and find it effective at recovering gene-gene relationships and additional structures. Applied to the epithilial to mesenchymal transition, MAGIC reveals a phenotypic continuum, with the majority of cells residing in intermediate states that display stem-like signatures, and infers known and previously uncharacterized regulatory interactions, demonstrating that our approach can successfully uncover regulatory relations without perturbations.

The Human Cell Atlas

The recent advent of methods for high-throughput single-cell molecular profiling has catalyzed a growing sense in the scientific community that the time is ripe to complete the 150-year-old effort to identify all cell types in the human body. The Human Cell Atlas Project is an international collaborative effort that aims to define all human cell types in terms of distinctive molecular profiles (such as gene expression profiles) and to connect this information with classical cellular descriptions (such as location and morphology). An open comprehensive reference map of the molecular state of cells in healthy human tissues would propel the systematic study of physiological states, developmental trajectories, regulatory circuitry and interactions of cells, and also provide a framework for understanding cellular dysregulation in human disease. Here we describe the idea, its potential utility, early proofs-of-concept, and some design considerations for the Human Cell Atlas, including a commitment to open data, code, and community.

The Human Cell Atlas

The recent advent of methods for high-throughput single-cell molecular profiling has catalyzed a growing sense in the scientific community that the time is ripe to complete the 150-year-old effort to identify all cell types in the human body. The Human Cell Atlas Project is an international collaborative effort that aims to define all human cell types in terms of distinctive molecular profiles (such as gene expression profiles) and to connect this information with classical cellular descriptions (such as location and morphology). An open comprehensive reference map of the molecular state of cells in healthy human tissues would propel the systematic study of physiological states, developmental trajectories, regulatory circuitry and interactions of cells, and also provide a framework for understanding cellular dysregulation in human disease. Here we describe the idea, its potential utility, early proofs-of-concept, and some design considerations for the Human Cell Atlas, including a commitment to open data, code, and community.

Abstract 592: Distinct cellular mechanisms mediate anti-CTLA-4 and anti-PD-1 checkpoint blockade

Checkpoint blockade is able to achieve durable responses in a subset of patients, however the biological variables that distinguish responders from non-responders are not well understood. Furthermore, we lack a satisfying comprehension of the underlying mechanisms of anti-CTLA-4 and anti-PD-1 induced tumor rejection. Given that PD-1 and CTLA-4 attenuate T cell activity through distinct mechanisms that are separated spatially and temporally, we hypothesized that responses to anti-CTLA-4 and anti-PD-1 are driven by distinct mechanisms. To address this hypothesis we utilized mass cytometry to comprehensively profile the effect of checkpoint blockade on tumor immune infiltrates in murine tumor models. This approach allows for the interrogation of greater than 40 analytes at single cell resolution. We demonstrate that high dimensional mass cytometry analysis enables unsupervised identification of biologically relevant tumor infiltrating immune populations with high sensitivity and specificity.

Using this approach we analyzed immune infiltrates of MC38 and B16BL6 murine tumors in mice treated with anti-CTLA-4, anti-PD-1, or control antibodies. In both tumor models we identify 15 distinct T cell populations with 0.5% or greater frequency. The T cell populations identified in MC38 and B16BL6 tumors were highly congruent. Notably, some but not all of these T cell populations were responsive to checkpoint blockade. A subset of tumor infiltrating CD8 T cell populations expanded following both anti-CTLA-4 and anti-PD-1. Conversely, a subset of regulatory T cell populations contracted following both anti-CTLA-4 and anti-PD-1. Interestingly, we observed expansion of a Th1-like CD4 effector T cell population only in response to anti-CTLA-4 treatment. Thus, we find that anti-PD-1 predominantly engages subsets of tumor infiltrating CD8 T cells whereas anti-CTLA-4 engages both the CD4 and CD8 effector compartments.

Our findings indicate that anti-CTLA-4 and anti-PD-1 utilize distinct cellular mechanisms to induce tumor rejection. These findings highlight the importance of expanding our mechanistic understanding of immunotherapeutic approaches for the rational design of combinatorial therapeutic approaches. Furthermore, these results demonstrate that mass cytometry analysis can be utilized to identify biologically relevant tumor infiltrating T cell populations.

We acknowledge the MDACC core facility NCI Support Grant P30CA16672.

Citation Format: Spencer C. Wei, Jacob H. Levine, Dana Pe'er, James P. Allison. Distinct cellular mechanisms mediate anti-CTLA-4 and anti-PD-1 checkpoint blockade [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 592. doi:10.1158/1538-7445.AM2017-592

Abstract 977: Phenotypic analysis of single-cell breast cancer inhibition data reveals insights into EMT

Background: A leading model of cancer metastasis is epithelial-to-mesenchymal transition (EMT). We sought to determine whether single-cell inhibition data targeting potential mediators of EMT could uncover mechanistic insights into the EMT process.

Methods: EMT was artificially induced on Py2T murine breast cancer cells by TGFb treatment. Additionally, a unique drug inhibitor was added to each well of a multiplexed CyTOF experiment. 37 transcription factors and cell surface markers were measured in each cell to assess epithelial and mesenchymal states, SMAD, AKT, and MAPK signaling activity, cell cycle regulation, and apoptosis pathway activation. The final single-cell dataset consisted of 300 inhibition and control conditions (cell populations), which we aimed to characterize in relation to one another with respect to effect on EMT.

Analyzing the similarity between drug inhibitions amounts to a novel type of clustering problem that involves computing the similarity between diverse cell populations generated by each inhibitor. Traditional methods for comparing cell populations are not robust to the intra-population heterogeneity we observed amongst cells undergoing EMT. Thus, we developed Phenotypic Earth Mover’s Distance (PhEMD). This method for comparing cell populations leverages the insight that only a limited number of “cell subtypes” (e.g. mesenchymal, epithelial, transitional) are observed in unperturbed and perturbed EMT. By classifying each cell as one of these distinct subtypes using community-detection based clustering, PhEMD represents an inhibition or control condition as its relative abundance of each cell subtype. It then uses Earth Mover’s Distance (EMD) to compare two relative abundance distributions (i.e. heterogeneous cell populations). PhEMD thus derives a single value representing the dissimilarity between two inhibition conditions. Using PhEMD as measure of dissimilarity between each pair of inhibition and control conditions, we constructed an inhibitor-inhibitor graph and used graph clustering to identify groups of inhibitors that had similar effects to one another.

Results: PhEMD analysis revealed that MEK, EGFR and Src inhibitors significantly halted EMT by generating far fewer mesenchymal cells and maintaining a large epithelial cell subpopulation. PhEMD also revealed that different mTOR, PI3K, and Akt inhibitors tended to have similar effects to one another and collectively generated a distinct, transitional cell subpopulation with altered pS6 expression. The pS6 dysregulation may be explained by the fact that ribosomal protein S6 is downstream of PI3K/AKT/mTOR, and the intermediate levels of both E-cadherin and vimentin suggest that this subset of cells may have been halted mid-transition. Finally, several Aurora Kinase and CDK inhibitors resulted in a relatively high percentage of apoptotic cells, suggesting these kinases may be important cell cycle regulators in the context of EMT.

Citation Format: William S. Chen, Nevena Zivanovic, Dana Pe'er, Bernd Bodenmiller, Smita Krishnaswamy. Phenotypic analysis of single-cell breast cancer inhibition data reveals insights into EMT [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 977. doi:10.1158/1538-7445.AM2017-977

Innate Immune Landscape in Early Lung Adenocarcinoma by Paired Single-Cell Analyses

To guide the design of immunotherapy strategies for patients with early stage lung tumors, we developed a multiscale immune profiling strategy to map the immune landscape of early lung adenocarcinoma lesions to search for tumor-driven immune changes. Utilizing a barcoding method that allows a simultaneous single-cell analysis of the tumor, non-involved lung, and blood cells, we provide a detailed immune cell atlas of early lung tumors. We show that stage I lung adenocarcinoma lesions already harbor significantly altered T cell and NK cell compartments. Moreover, we identified changes in tumor-infiltrating myeloid cell (TIM) subsets that likely compromise anti-tumor T cell immunity. Paired single-cell analyses thus offer valuable knowledge of tumor-driven immune changes, providing a powerful tool for the rational design of immune therapies. VIDEO ABSTRACT.

Measuring Signaling and RNA-Seq in the Same Cell Links Gene Expression to Dynamic Patterns of NF-κB Activation

Signaling proteins display remarkable cell-to-cell heterogeneity in their dynamic responses to stimuli, but the consequences of this heterogeneity remain largely unknown. For instance, the contribution of the dynamics of the innate immune transcription factor nuclear factor κB (NF-κB) to gene expression output is disputed. Here we explore these questions by integrating live-cell imaging approaches with single-cell sequencing technologies. We used this approach to measure both the dynamics of lipopolysaccharide-induced NF-κB activation and the global transcriptional response in the same individual cell. Our results identify multiple, distinct cytokine expression patterns that are correlated with NF-κB activation dynamics, establishing a functional role for NF-κB dynamics in determining cellular phenotypes. Applications of this approach to other model systems and single-cell sequencing technologies have significant potential for discovery, as it is now possible to trace cellular behavior from the initial stimulus, through the signaling pathways, down to genome-wide changes in gene expression, all inside of a single cell.

Bayesian Inference for Single-cell Clustering and Imputing

Single-cell RNA-seq gives access to gene expression measurements for thousands of cells, allowing discovery and characterization of cell types. However, the data is noise-prone due to experimental errors and cell type-specific biases. Current computational approaches for analyzing single-cell data involve a global normalization step which introduces incorrect biases and spurious noise and does not resolve missing data (dropouts). This can lead to misleading conclusions in downstream analyses. Moreover, a single normalization removes important cell type-specific information. We propose a data-driven model, BISCUIT, that iteratively normalizes and clusters cells, thereby separating noise from interesting biological signals. BISCUIT is a Bayesian probabilistic model that learns cell-specific parameters to intelligently drive normalization. This approach displays superior performance to global normalization followed by clustering in both synthetic and real single-cell data compared with previous methods, and allows easy interpretation and recovery of the underlying structure and cell types.