Cell Type Clustering Research Articles

FR-Match: robust matching of cell type clusters from single cell RNA sequencing data using the Friedman-Rafsky non-parametric test.

Single cell/nucleus RNA sequencing (scRNAseq) is emerging as an essential tool to unravel the phenotypic heterogeneity of cells in complex biological systems. While computational methods for scRNAseq cell type clustering have advanced, the ability to integrate datasets to identify common and novel cell types across experiments remains a challenge. Here, we introduce a cluster-to-cluster cell type matching method—FR-Match—that utilizes supervised feature selection for dimensionality reduction and incorporates shared information among cells to determine whether two cell type clusters share the same underlying multivariate gene expression distribution. FR-Match is benchmarked with existing cell-to-cell and cell-to-cluster cell type matching methods using both simulated and real scRNAseq data. FR-Match proved to be a stringent method that produced fewer erroneous matches of distinct cell subtypes and had the unique ability to identify novel cell phenotypes in new datasets. In silico validation demonstrated that the proposed workflow is the only self-contained algorithm that was robust to increasing numbers of true negatives (i.e. non-represented cell types). FR-Match was applied to two human brain scRNAseq datasets sampled from cortical layer 1 and full thickness middle temporal gyrus. When mapping cell types identified in specimens isolated from these overlapping human brain regions, FR-Match precisely recapitulated the laminar characteristics of matched cell type clusters, reflecting their distinct neuroanatomical distributions. An R package and Shiny application are provided at https://github.com/JCVenterInstitute/FRmatch for users to interactively explore and match scRNAseq cell type clusters with complementary visualization tools.

DIANA-miRGen v4: indexing promoters and regulators for more than 1500 microRNAs.

Deregulation of microRNA (miRNA) expression plays a critical role in the transition from a physiological to a pathological state. The accurate miRNA promoter identification in multiple cell types is a fundamental endeavor towards understanding and characterizing the underlying mechanisms of both physiological as well as pathological conditions. DIANA-miRGen v4 (www.microrna.gr/mirgenv4) provides cell type specific miRNA transcription start sites (TSSs) for over 1500 miRNAs retrieved from the analysis of >1000 cap analysis of gene expression (CAGE) samples corresponding to 133 tissues, cell lines and primary cells available in FANTOM repository. MiRNA TSS locations were associated with transcription factor binding site (TFBSs) annotation, for >280 TFs, derived from analyzing the majority of ENCODE ChIP-Seq datasets. For the first time, clusters of cell types having common miRNA TSSs are characterized and provided through a user friendly interface with multiple layers of customization. DIANA-miRGen v4 significantly improves our understanding of miRNA biogenesis regulation at the transcriptional level by providing a unique integration of high-quality annotations for hundreds of cell specific miRNA promoters with experimentally derived TFBSs.

A map of cis-regulatory elements and 3D genome structures in zebrafish.

The zebrafish (Danio rerio) has been widely used in the study of human disease and development, and about 70% of the protein-coding genes are conserved between the two species1. However, studies in zebrafish remain constrained by the sparse annotation of functional control elements in the zebrafish genome. Here we performed RNA sequencing, assay for transposase-accessible chromatin using sequencing (ATAC-seq), chromatin immunoprecipitation with sequencing, whole-genome bisulfite sequencing, and chromosome conformation capture (Hi-C) experiments in up to eleven adult and two embryonic tissues to generate a comprehensive map of transcriptomes, cis-regulatory elements, heterochromatin, methylomes and 3D genome organization in the zebrafish Tübingen reference strain. A comparison of zebrafish, human and mouse regulatory elements enabled the identification of both evolutionarily conserved and species-specific regulatory sequences and networks. We observed enrichment of evolutionary breakpoints at topologically associating domain boundaries, which were correlated with strong histone H3 lysine 4 trimethylation (H3K4me3) and CCCTC-binding factor (CTCF) signals. We performed single-cell ATAC-seq in zebrafish brain, which delineated 25 different clusters of cell types. By combining long-read DNA sequencing and Hi-C, we assembled the sex-determining chromosome 4 de novo. Overall, our work provides an additional epigenomic anchor for the functional annotation of vertebrate genomes and the study of evolutionarily conserved elements of 3D genome organization.

Chromatin accessibility in canine stromal cells and its implications for canine somatic cell reprogramming.

Naturally occurring disease in pet dogs is an untapped and unique resource for stem cell‐based regenerative medicine translational research, given the many similarities and complexity such disease shares with their human counterparts. Canine‐specific regulators of somatic cell reprogramming and pluripotency maintenance are poorly understood. While retroviral delivery of the four Yamanaka factors successfully reprogrammed canine embryonic fibroblasts, adult stromal cells remained resistant to reprogramming in spite of effective viral transduction and transgene expression. We hypothesized that adult stromal cells fail to reprogram due to an epigenetic barrier. Here, we performed assay for transposase‐accessible chromatin using sequencing (ATAC‐seq) on canine stromal and pluripotent stem cells, analyzing 51 samples in total, and establishing the global landscape of chromatin accessibility before and after reprogramming to induced pluripotent stem cells (iPSC). We also studied adult stromal cells that do not yield iPSC colonies to identify potential reprogramming barriers. ATAC‐seq analysis identified distinct cell type clustering patterns and chromatin remodeling during embryonic fibroblast reprogramming. Compared with embryonic fibroblasts, adult stromal cells had a chromatin accessibility landscape that reflects phenotypic differentiation and somatic cell‐fate stability. We ultimately identified 76 candidate genes and several transcription factor binding motifs that may be impeding somatic cell reprogramming to iPSC, and could be targeted for inhibition or activation, in order to improve the process in canines. These results provide a vast resource for better understanding of pluripotency regulators in dogs and provide an unbiased rationale for novel canine‐specific reprogramming approaches.

Heterogeneity of midgut cells and their differential responses to blood meal ingestion by the mosquito, Aedes aegypti

Mosquitoes are the most notorious hematophagous insects and due to their blood feeding behavior and genetic compatibility, numerous mosquito species are highly efficient vectors for certain human pathogenic parasites and viruses. The mosquito midgut is the principal organ of blood meal digestion and nutrient absorption. It is also the initial site of infection with blood meal acquired parasites and viruses. We conducted an analysis based on single-nucleus RNA sequencing (snRNA-Seq) to assess the cellular diversity of the midgut and how individual cells respond to blood meal ingestion to facilitate its digestion. Our study revealed the presence of 20 distinguishable cell-type clusters in the female midgut of Aedes aegypti. The identified cell types included intestinal stem cells (ISC), enteroblasts (EB), differentiating EB (dEB), enteroendocrine cells (EE), enterocytes (EC), EC-like cells, cardia cells, and visceral muscle (VM) cells. Blood meal ingestion dramatically changed the overall midgut cell type composition, profoundly increasing the proportions of ISC and three EC/EC-like clusters. In addition, transcriptional profiles of all cell types were strongly affected while genes involved in various metabolic processes were significantly upregulated. Our study provides a basis for further physiological and molecular studies on blood digestion, nutrient absorption, and cellular homeostasis in the mosquito midgut.

Single-Cell Analysis of the Classical Hodgkin Lymphoma Immune Environment Reveals a Clonally-Expanded CD8+ T Cell Population with a Cytotoxic Phenotype

Introduction: Classical Hodgkin lymphoma (cHL) is characterized by a robust and complex immune cell infiltrate and the rare presence of malignant Hodgkin-Reed-Sternberg (HRS) cells. At the genetic level, HRS cells recurrently acquire alterations that lead to defective antigen presentation (β2 microglobulin mutations) and mediate T cell dysfunction (PD-L1 copy gains/amplifications) in order to subvert host immune surveillance. The clinical relevance of PD-L1 protein over-expression in cHL is clear, as response rates to PD-1 blockade therapy are extremely high among patients with relapsed/refractory (r/r) disease. Despite its remarkable efficacy, the cells that mediate response to anti-PD-1 therapy in cHL remain undefined. Recent analyses have highlighted a possible role for CD4+ T cells in mediating the clinical activity of anti-PD-1 therapy in cHL. CD4+ T cells significantly outnumber CD8+ T cells in cHL lesions and are more frequently juxtaposed to HRS cells in situ. Furthermore, HLA class II expression on HRS cells predicted higher complete response rates to PD-1 blockade therapy in r/r cHL patients. However, a candidate T cell population capable of specific reactivity to antigens expressed by HRS cells has yet to be identified. This information is critical as such T cells might be functionally reinvigorated to mediate HRS cell elimination following PD-1 blockade therapy. In order to address this key knowledge gap, we analyzed data at single cell (sc) resolution using paired RNA and T cell receptor (TCR) sequencing in 9 diagnostic cHL and 5 reactive lymph node (RLN) specimens. Methods: Sequencing was performed using the 10x Genomics Chromium Single Cell 5' Gene Expression and V(D)J workflows. B-cell depletion of each sample was achieved using CD19 microbeads and negative selection to enrich T cell populations. Reads were analyzed and aligned with CellRanger (v3.1.0) and Seurat (v3.2.0) was used to conduct clustering by a shared nearest neighbor (SNN) graph on scRNA data. TCR sequencing data was integrated using scRepertoire (v1.0.0). Results: A detailed map of the immune cell states in cHL was created using scRNA-seq (10X) data on 79,085 cells from 9 cHL (52,602 cells) and 5 RLN samples (26,484 cells) expressing a total of 21,421 genes (mean 5649 cells/sample; mean 2849 mRNA reads/cell). Dimensionality reduction and unsupervised graph-based clustering revealed 21 distinct cell type and activation state clusters, including T cells, NK cells, macrophages, and dendritic cells (Fig 1A-B). A cluster identifying HRS cells was not observed, consistent with a recently published report. Ten T cell clusters were identified (47,573 cells), including naive- and memory-like T cells, effector/cytotoxic CD8+ T cells, regulatory T cells, and T follicular helper cells. Unexpectedly, a putative exhausted T cell cluster was not clearly observed. The relative contributions of cHL and RLNs cases to these clusters are shown in Fig 1C. Paired TCR sequencing was available for 23,943 cells. Overall TCR diversity was lower among cHL samples compared to RLN specimens (Fig 1D). In cHL samples, modest clonal expansion within regulatory T cell and memory CD4+ T cell clusters was observed, but the most striking clonal expansion occurred among cells assigned to effector/cytotoxic CD8+ T cell clusters - a finding not observed in most RLN specimens (Fig 1E). Clonally-expanded effector/cytotoxic CD8+ T cells displayed high expression of granzymes (GZMA, GZMH, GZMK), cytokines (TNF, IFNG) and chemokines (CCL4/CCL5), and modest expression of exhaustion markers (PDCD1, ENTPD1, HAVCR2, ITGAE), contrasting with data from single-cell analyses of solid tumors. Clonal expansion of effector/cytotoxic CD8+ T cells was particularly robust in EBV-positive cHLs, likely due to recognition of viral-derived epitopes displayed on HRS cells (Fig 1F). Phenotypic and functional validation of key immune cell clusters in cHL specimens using spectral cytometry is underway and will be reported at the meeting. Conclusions: For the first time, our data have unveiled the nature of the T cell repertoire in cHL at single cell resolution. Our results reveal a recurrent pattern of clonal expansion within effector CD8+ cells, which may be the HRS antigen-specific T cells that mediate response to PD-1 blockade. This hypothesis requires confirmation through similar analyses of pre- and on-treatment biopsies of cHL patients receiving anti-PD-1 therapy. Disclosures Godfrey: Gilead: Research Funding; Merck: Research Funding; Verastem: Research Funding. Venkataraman:EUSA Pharma: Speakers Bureau. Smith:Janssen: Consultancy; BMS: Consultancy; TG Therapeutics: Consultancy, Research Funding; Genentech/Roche: Consultancy, Other: Support of parent study and funding of editorial support, Research Funding; Karyopharm: Consultancy, Research Funding; FortySeven: Research Funding; Pharmacyclics: Research Funding; Acerta: Research Funding; Celgene: Consultancy, Research Funding. Kline:Kite/Gilead: Speakers Bureau; Seattle Genetics: Membership on an entity's Board of Directors or advisory committees; Merck: Research Funding; Karyopharm: Membership on an entity's Board of Directors or advisory committees; Verastem: Membership on an entity's Board of Directors or advisory committees, Research Funding.

Characterization of T-ALL-Specific Heterogenous Blast Populations Using High Resolution Single Cell Profiling

Introduction: T-cell acute lymphoblastic leukemia (T-ALL) is characterized by proliferation of immature T-cells and accounts for ~15% of pediatric ALL. T-ALL blasts are phenotypically diverse and are sub-classified into pro-, pre-, cortical and mature T-ALL based on the stage of differentiation of the leukemic clone. Early T precursor ALL (ETP-ALL) is a T-ALL subtype associated with higher risk of relapse (Raetz and Teachey, 2016). Bulk sequencing approaches have revealed valuable information about modulated genes in T-ALL; however, little is understood about the interplay between tumor cells and the immune microenvironment. We present a comprehensive single cell RNA sequencing (scRNASeq) analysis of T-ALL samples with the purpose of characterizing the heterogenous tumor and microenvironment cells in order to identify dysregulated genes in leukemia cells and investigate oncogenic signaling pathways. Methods: We profiled 16,280 cells from 5 diagnostic pediatric T-ALL bone marrow samples using the Chromium single cell transcriptomics platform (10x Genomics, Pleasanton, CA). To compare T-ALL versus healthy bone marrow profiles and identify T-ALL-specific malignant blast cell populations, we included data from 3 healthy pediatric bone marrow samples from a recent study (Caron et al, 2020). Dimension reduction using the Uniform Manifold Approximation and Projection (UMAP) approach was used to identify unique cell type clusters (Becht et al, 2018). Using the TARGET dataset (https://ocg.cancer.gov/programs/target), we further evaluated the prognostic significance of identified malignant blast-specific genes by performing Kaplan-Meier survival analysis and compared gene expression patterns and tumor microenvironment makeup between ETP-ALL and non-ETP T-ALL patient samples. Results: We successfully characterized leukemic blasts (CD7+, CD99+ and CD3D+) and other major immune cell types (T cells, B cells, monocytes, erythroid precursors) using the expression of established marker genes (Fig 1A, C). Clustering analysis revealed patient-specific leukemia blast cell clusters (Fig 1B). Differential expression analysis between CD3D+ patient-specific leukemia clusters and CD3D+ clusters comprised of normal T-cells identified a set of 385 promiscuous genes that are significantly differentially expressed between malignant and normal T-cells (p-value <0.05) despite tumor cell heterogeneity. Among the top genes upregulated in the leukemia clusters are HES4 (a downstream target of NOTCH1), CD99, RACK1, and TUBB (Fig 1D), which have previously been implicated in T-ALL leukemogenesis and chemoresistance (DeDecker et al, 2020; Cox et al, 2016; Lei et all 2016). Further, pathways analysis demonstrated significant activation (Z score >1.3, p-value <0.05) of multiple pathways associated with cancer stemness, cellular growth and proliferation including PI3K/AKT, unfolded protein response, and glycolysis, implicating these genes as oncogenic mediators in T-ALL blasts (Fig 2A). Overexpression of S100A4, IFITM2, and CD74 were significantly associated with poor survival in the TARGET cohort of 241 patients (hazard ratio 2.2, 5.8, and 7.1, respectively, p-value <0.05), indicating their prognostic significance (Fig 2B). We additionally evaluated the expression of our gene set across ETP-ALL and non-ETP T-ALL groups in the TARGET dataset. Multiple leukemic blast-specific genes including ARMH1, CD44, CD74, DNAJC1, and IFITM2 were significantly upregulated in ETP-ALL. Further, deconvolution analysis performed on ETP-ALL vs non-ETP T-ALL samples determined that tumor microenvironment cell types are differentially enriched in these T-ALL groups. We observed enrichment of memory B cells, dendritic cells, NK cells, and M2 macrophages in ETP-ALL samples as compared to non-ETP T-ALL samples. Conversely,lower levels of mast cells and T-regulatory cells were observed in ETP-ALL samples. Conclusions: We identified a gene signature characterizing heterogenous T-ALL blast populations. External validation using the TARGET dataset identified genes within the signature that are associated with poor outcomes in T-ALL. Differences in the composition of the immune microenvironment in ETP-ALL versus non-ETP T-ALL samples provide a promising area for future study. Further studies will be carried out with relapsed and non-relapsed T-ALL samples to validate this leukemia signature. Disclosures Bhasin: Canomiiks Inc: Current equity holder in private company, Other: Co-Founder.

RUNX1 Mutation Leads to Megakaryocyte-Primed Hematopoietic Stem Cell Expansion and Familial Platelet Disorder

Backgrounds: Recently it is shown that there exists a subpopulation of hematopoietic stem cells (HSCs) with relatively high expression of VWF and CD41+ at the apex of the hematopoietic stem cell hierarchy. This subset of HSCs, termed mega-HSCs, can give rise to megakaryocytes and platelets directly by bypassing the traditional trajectory of megakaryocyte development from HSC via MPP and MEP. To date, aside from phenotypic marker and transplantation studies, there has been limited understanding of the mechanisms involved in the regulation of mega-HSCs. Runx1 belongs to the RUNT domain transcription factors, and is a key regulator of hematopoiesis especially for the megakaryocyte and platelet differentiation. Loss of RUNX1 in mice causes thrombocytopenia through a blockage of megakaryocyte maturation. One allele RUNX1 loss-of-function mutation is associated with familial platelet disorder (FPD) with a predisposition to developing leukemia. Here we hypothesize that Runx1 plays a role in regulating mega-HSCs to impact on platelet generation, and correcting RUNX1 mutation that causes FPD can therapeutically rescue thrombocytopenia in a mouse model. Methods: We have examined the hematopoiesis and cellularity of various bone marrow (BM) stem/progenitor populations and peripheral blood (PB) in two mouse models. Firstly, conditional knock-out of Runx1flox/flox was mediated by Mx1-Cre upon poly I:C induction. Secondly, the tetracycline-inducible RUNX1 S291fsX300 mutation was knock-in at the collagen a1 locus and the mice was crossed to MLL-PTD knock-in. The mutant RUNX1 is only expressed when the mice are fed with doxycycline and the RUNX1 mutant is "corrected" upon doxycycline withdrawn. In the second model, PB and BM were also tracked after a removal of DOX-induced RUNX1 mutation. The LSK CD150+ HSCs from the mouse BM were isolated by FACS sorting, and 10x Genomics' single-cell RNA-seq (scRNA-seq) analyses were performed to define and track HSPCs. We applied the Louvain algorithm to the scRNA-seq data to identify the cell type clusters and annotated the cell types using the marker genes of each cell cluster. MAST was employed to identify the cell-type specific, differentially expressed genes. Results: In the Runx1 KO mice, two weeks after deletion of Runx1 platelet count showed an ~2-fold decrease to 400~600 k/ul in PB. In the LSK CD150+CD48- or the LSK CD34- FLT3- compartment of BM, the CD41+ mega-HSCs increased ~2-3 fold. In the RUNX1 mutant-on model, mutant mice developed thrombocytopenia 16 weeks after DOX induction with the average platelet count dropping to ~600 k/ul and being maintained at this level. In transplant recipients, the RUNX1 mutant-on mice contained drastically increased mega-HSCs compared with RUNX1 mutant-off mice, synchronous with a decrease of the platelet count in PB. Consistent with these observations, sc-RNAseq data show that in Runx1 KO, among the sequenced LSK CD150+ cells ~70% are HSCs and ~15% are MPP2. Among the HSCs, mega-primed HSCs consist ~33% while non-mega primed HSCs are ~67%. The mega-primed HSCs are distinct by virtual of upregulated platelet-related genes and relative dormant cell cycle status. In the RUNX1 mutant-on model, we found that among the sequenced LSK CD150+ cells ~50% are HSCs and ~12% are MPP2. Interestingly, the mega-primed HSCs are significantly increased in RUNX1 mut-on HSCs compared with mut-off HSCs. Similarly, we saw highly upregulated platelet-driven pathways in mega-HSCs in the mutant-on HSCs. Finally, one month after a withdraw of DOX from the RUNX1 mutant expressing mice when the RUNX1 mutant gene became undetectable, the platelet count returned from ~600 to ~1,000 k/ul in PB, a reversal of the thrombocytopenia phenotype caused by the RUNX1 S291fsX300 expression. Importantly, the correction of Runx1 mutation significantly decreased the proportion of mega-HSCs and restored platelet related pathways in this subset of HSCs. Conclusions: Mega-HSCs contain a high level of platelet-driven gene expression. In addition to its role in regulating the HSC-MPP-MEP mediated megakaryocyte development, RUNX1 is important in regulating mega-HSCs by maintaining proper expression of mega-platelet leaning genes. Correction of RUNX1 mutation that causes FPD can rescue mega-HSC population and revert FPD, providing a rationale for future treatment strategies by gene editing in RUNX1 mutation bearing FPD patients. Disclosures No relevant conflicts of interest to declare.

Impact of data preprocessing on cell-type clustering based on single-cell RNA-seq data

BackgroundAdvances in single-cell RNA-seq technology have led to great opportunities for the quantitative characterization of cell types, and many clustering algorithms have been developed based on single-cell gene expression. However, we found that different data preprocessing methods show quite different effects on clustering algorithms. Moreover, there is no specific preprocessing method that is applicable to all clustering algorithms, and even for the same clustering algorithm, the best preprocessing method depends on the input data.ResultsWe designed a graph-based algorithm, SC3-e, specifically for discriminating the best data preprocessing method for SC3, which is currently the most widely used clustering algorithm for single cell clustering. When tested on eight frequently used single-cell RNA-seq data sets, SC3-e always accurately selects the best data preprocessing method for SC3 and therefore greatly enhances the clustering performance of SC3.ConclusionThe SC3-e algorithm is practically powerful for discriminating the best data preprocessing method, and therefore largely enhances the performance of cell-type clustering of SC3. It is expected to play a crucial role in the related studies of single-cell clustering, such as the studies of human complex diseases and discoveries of new cell types.

FSHB Transcription is Regulated by a Novel 5' Distal Enhancer With a Fertility-Associated Single Nucleotide Polymorphism.

The pituitary gonadotropins, follicle-stimulating hormone (FSH) and luteinizing hormone, signal the gonads to regulate male and female fertility. FSH is critical for female fertility as it regulates oocyte maturation, ovulation, and hormone synthesis. Multiple genome-wide association studies (GWAS) link a 130 Kb locus at 11p14.1, which encompasses the FSH beta-subunit (FSHB) gene, with fertility-related traits that include polycystic ovary syndrome, age of natural menopause, and dizygotic twinning. The most statistically significant single nucleotide polymorphism from several GWAS studies (rs11031006) resides within a highly conserved 450 bp region 26 Kb upstream of the human FSHB gene. Given that sequence conservation suggests an important biological function, we hypothesized that the region could regulate FSHB transcription. In luciferase assays, the conserved region enhanced FSHB transcription and gel shifts identified a binding site for Steroidogenic factor 1 (SF1) contributing to its function. Analysis of mouse pituitary single-cell ATAC-seq demonstrated open chromatin at the conserved region exclusive to a gonadotrope cell-type cluster. Additionally, enhancer-associated histone markers were identified by immunoprecipitation of chromatin from mouse whole pituitary and an immortalized mouse gonadotrope-derived LβT2 cell line at the conserved region. Furthermore, we found that the rs11031006 minor allele upregulated FSHB transcription via increased SF1 binding to the enhancer. All together, these results identify a novel upstream regulator of FSHB transcription and indicate that rs11031006 can modulate FSH levels.

72 - Single Cell RNA Sequencing Points to Altered IL-10 Signalling and Metabolism in Hyperinsulinemia-Driven Pancreatic Cancer Initiation

Epidemiological studies have shown hyperinsulinemia is independently associated with increased pancreatic ductal adenocarcinoma (PDAC) morbidity and mortality. We recently showed direct in vivo evidence that hyperinsulinemia promotes PDAC. Using a mouse model of PDAC development (Ptf1aCreER;LSL-KrasG12D) crossed with Ins1+/-;Ins2-/- alleles that result in ∼50% reduced fasting insulin, we found a ∼50% reduction in pancreatic intraepithelial neoplasia (PanIN) precancerous lesions compared with Ins1+/+;Ins2-/- control mice. Mice have 2 insulin genes, Ins1 and Ins2. Ins1 expression is restricted to β-cells where it contributes to ∼1/3 of secreted insulin, while Ins2 is evolutionarily conserved. Here, we studied PanIN development in a ‘genetically humanized’ strain of Ptf1aCreER;LSL-KrasG12D mice lacking the rodent specific Ins1 gene, and with varying Ins2 allele dosage. Experimental Ptf1aCreER;LSL-KrasG12D;Ins1-/-;Ins2+/- (PK- Ins1-/-;Ins2+/-) mice had significantly fasting lower insulin levels, without affecting glucose homeostasis, and lower body weight compared to control Ptf1aCreER;LSL-KrasG12D;Ins1-/-;Ins2+/+ (PK- Ins1-/-;Ins2+/+) mice. Blind quantification of (relatively rare) PanIN area showed that PK-Ins1-/-;Ins2+/- mice had a relative lower PanIN/total pancreas area when compared with PK- Ins1-/-;Ins2+/+ mice. scRNAseq identified 14 cell type clusters and, between the genotypes, we found genes involved in metabolic pathways like Ucp2, Cebpb were differentially expressed in acinar and β-cells, while genes involved in IL-10 signalling (Ccl5, Cxcl1, Cxcl2) were differentially expressed in T cells, macrophages and dendritic cells. There were also fewer tumour-promoting M2 macrophages in mice with reduced circulating insulin. Our data suggest reducing circulating insulin may affect PanIN development by affecting the metabolic and immune signallings within the acinar cells and immune cells.

MYC Promotes Bone Marrow Stem Cell Dysfunction in Fanconi Anemia.

Bone marrow failure (BMF) in Fanconi anemia (FA) patients results from dysfunctional hematopoietic stem and progenitor cells (HSPCs). To identify determinants of BMF, we performed single-cell transcriptome profiling of primary HSPCs from FA patients. In addition to overexpression of p53 and TGF-β pathway genes, we identified high levels of MYC expression. We correspondingly observed coexistence of distinct HSPC subpopulations expressing high levels of TP53 or MYC in FA bone marrow (BM). Inhibiting MYC expression with the BET bromodomain inhibitor (+)-JQ1 reduced the clonogenic potential of FA patient HSPCs but rescued physiological and genotoxic stress in HSPCs from FA mice, showing that MYC promotes proliferation while increasing DNA damage. MYC-high HSPCs showed significant downregulation of cell adhesion genes, consistent with enhanced egress of FA HSPCs from bone marrow to peripheral blood. We speculate that MYC overexpression impairs HSPC function in FA patients and contributes to exhaustion in FA bone marrow.

A simple, scalable approach to building a cross-platform transcriptome atlas.

Gene expression atlases have transformed our understanding of the development, composition and function of human tissues. New technologies promise improved cellular or molecular resolution, and have led to the identification of new cell types, or better defined cell states. But as new technologies emerge, information derived on old platforms becomes obsolete. We demonstrate that it is possible to combine a large number of different profiling experiments summarised from dozens of laboratories and representing hundreds of donors, to create an integrated molecular map of human tissue. As an example, we combine 850 samples from 38 platforms to build an integrated atlas of human blood cells. We achieve robust and unbiased cell type clustering using a variance partitioning method, selecting genes with low platform bias relative to biological variation. Other than an initial rescaling, no other transformation to the primary data is applied through batch correction or renormalisation. Additional data, including single-cell datasets, can be projected for comparison, classification and annotation. The resulting atlas provides a multi-scaled approach to visualise and analyse the relationships between sets of genes and blood cell lineages, including the maturation and activation of leukocytes in vivo and in vitro. In allowing for data integration across hundreds of studies, we address a key reproduciblity challenge which is faced by any new technology. This allows us to draw on the deep phenotypes and functional annotations that accompany traditional profiling methods, and provide important context to the high cellular resolution of single cell profiling. Here, we have implemented the blood atlas in the open access Stemformatics.org platform, drawing on its extensive collection of curated transcriptome data. The method is simple, scalable and amenable for rapid deployment in other biological systems or computational workflows.

Abstract 865: Using machine learning to detect heterogeneity in single cell multi-omics datasets

Abstract Background: While scRNA-seq helps in obtaining high-resolution views of single-cell heterogeneity through characterization of the functional state of cells, our understanding of the cellular properties and population architectures of heterogeneous tissues will be greatly advanced by the multi-omics investigation of single cells. Recently, several computational methods have been developed to integrate data from multiple different single cell experiments measuring individual analytes. But unambiguous inference that a cellular phenotype is caused by a genotype can only be achieved by their measurement from the same single cell. To address this gap, we have developed the Tapestri multi-omics workflow to analyze the RNA and DNA information from the same cell. Methods: After pre-processing of the reads, cell calling is done by identifying the good barcodes using both DNA and RNA reads. DNA reads are processed to identify genetic variants in each cell. The variant cell matrix is then filtered for data completeness to ensure only high-quality data is used in downstream processing. Ploidy is estimated using by normalizing DNA reads, genetic variants and ploidy information together is used to identify subclones. The RNA reads are log normalized and scaled within each cell. Next we set the mean expression of each transcript to 0 and scale the variance to 1. This avoids downstream analysis being skewed by high expressors. We then trained a random forest classifier to identify significantly differentially expressed transcripts across subclones which were identified using genetic variants and ploidy information. Results: Using the top differentially expressed transcripts we performed dimensionality reduction followed by clustering of cell types. The resulting visualization showed how well the genotypic and transcriptomic datasets integrated with one another. We tested this method on a model system with Raji and KG1 cell lines titrated at 50:50 ratio and were clearly able to associate the transcriptional variation with the genotypic variation of the 2 different cell lines. The method was also validated on a PBMC sample to ensure robustness of methods. We were able to identify the different cell types present the sample and were able to overlay that information with genetic variants to identify sub-clones in the identified cell types. Citation Format: Saurabh Gulati, Shu Wang, Saurabh Parikh, Ben Liu, Kaustubh Gokhale, Manimozhi Manivannan, Sombeet Sahu, Dong Kim, Anup Parikh. Using machine learning to detect heterogeneity in single cell multi-omics datasets [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 865.

Abstract 6669: Cellular neighborhoods predict pembrolizumab response in cutaneous T cell lymphoma

Abstract Cutaneous T-cell lymphoma (CTCL) is a rare, incurable CD4+ T cell malignancy of the skin with a 5-year survival rate of less than 30% in advanced stages. Immune checkpoint inhibitors, such as anti-PD-1 antibodies, have shown dramatic clinical efficacy in multiple advanced cancers, but the majority of cancer patients do not respond to these treatments. The clinical use of immunotherapies will increase considerably in the near future; therefore, predictive biomarkers of response to stratify patients for treatment are needed to limit potentially devastating adverse effects and reduce costs for healthcare systems. A clinical trial of the anti-PD-1 antibody pembrolizumab in CTCL showed that 38% of patients have a durable clinical response. However, standard tests, including comprehensive immunohistochemistry and single-cell quantification of PD-1 expression, have so far failed to identify a predictive biomarker for pembrolizumab response in this cohort. We reasoned that deep profiling of the CTCL tumor microenvironment (TME) using CODEX–a novel technology that allows for highly multiplexed tissue microscopy with >50 simultaneous parameters–could provide insight into the mechanisms of pembrolizumab response and enable prediction. We analyzed the CTCL TME using a tissue microarray of matched biopsies taken before and after pembrolizumab therapy in 7 responders and 7 non-responders. Imaging of 55 markers allowed discriminating malignant CD4+ tumor cells from reactive CD4+ T cells based on nuclear size and differential expression of CD7, CD25 and Ki-67. Unsupervised machine learning followed by supervised curation identified 21 different cell type clusters with spatial information. Integrating these data using advanced computational analysis revealed 10 distinct, conserved cellular neighborhoods (CNs) in the CTCL TME that changed in frequency and distribution during pembrolizumab therapy. In responders, effector-type CNs, including a tumor/dendritic cell CN and a tumor/CD4+ T cell CN, were significantly increased after treatment. In contrast, in non-responders, an immunosuppressive-type CN enriched in regulatory T cells was significantly increased before and after therapy. Importantly, the global frequencies in the tissues of the cell types defining these CNs were not different between patient groups. In addition, RNA sequencing of matched tissue sections revealed higher expression of effector-type cytokines and chemokines in responders. In sum, highly multiplexed analysis of the CTCL TME architecture in combination with RNA sequencing allows discovering novel, predictive spatial biomarkers of immunotherapy response and will pave the way for future studies that functionally address the identified cell types and cellular interactions. Citation Format: Christian M. Schürch, Darci J. Phillips, Belén Rivero Gutierrez, Magdalena Matusiak, Salil S. Bhate, Graham L. Barlow, Steven P. Fling, Nirasha Ramchurren, Robert H. Pierce, Martin A. Cheever, Michael S. Khodadoust, Robert West, Youn H. Kim, Garry P. Nolan. Cellular neighborhoods predict pembrolizumab response in cutaneous T cell lymphoma [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 6669.

Abstract 3779: Multiomic single cell analysis of normal human bone marrow identifies a unique stem and progenitor population that expands in AML

Abstract Hematopoietic stem and progenitor cells (HSPCs) can generate a diversity of blood cells throughout the human lifespan. Although these cells have been evaluated using both sorted (Corces et al. Nat. Genet. 2016) and single cell assays (Pellin et al Nat. Commun. 2019), there remains uncertainty in the degree of heterogeneity within HSPC subpopulations and their associated differentiation trajectories. The phenotypic diversity within HSPCs needs to be better characterized in order to understand the pathogenesis of blood disorders including hematologic malignancies. To address this need, we characterized healthy bone marrow mononuclear cells (BMMCs) with whole transcriptome analysis (WTA) and surface marker evaluation. We hypothesized that by utilizing concurrent RNA and surface marker analysis (n=35), we can improve HSPC clustering and characterize specific phenotypic states along unique hematopoietic differentiation trajectories. Three healthy BMMC samples were stained with antibody conjugated oligonucleotides (BD® Abseq) and analyzed using BD Rhapsody™. We filtered 8,070 high quality cells for 2,508 HSPCs, myeloid cells, and lymphocyte precursors. The antibody-derived tags (ADTs) obtained from BD® Abseq comprised 33 of the most informative features (n=2000) and resulted in more stable clustering as determined by within sum of squares (WSS = 898 verses 934 for WTA alone for 20 clusters). Additionally, we designed a targeted HSPC panel (n=500 genes) with BD® Abseq which identified similar cell clusters compared to the WTA alone and WTA plus ADT data (rand index = 0.88). HSPC clustering identified putative hematopoietic stem cell (HSC), common myeloid progenitor (CMP), and megakaryocyte-erythroid progenitor (MEP) clusters that expressed canonical surface markers. Of interest, we identified candidate novel myeloid, T-cell, and B-cell primed precursors and new surface marker expression gradients that align with specific differentiation trajectories. The results of this analysis will be presented. The HSPC clusters were converted into a signature matrix using Cibersortx (Newman et al. Nat. Biotechnol. 2019), and bulk acute myeloid leukemia (AML) and healthy samples were deconvolved into respective healthy cell clusters. We subsequently performed multivariate Cox proportional hazard analysis, and observed that high levels of healthy cluster 8 (H8; HR 3.40, 95% confidence interval 1.24-9.34), the candidate lymphoid-primed multipotent progenitor (LMPP), and low levels of healthy cluster 6 (H6; HR 0.26, 95% confidence interval 0.12-0.58), a candidate erythrocyte precursor, at diagnosis were associated with worse overall survival. Deconvolution using sorted healthy sub-populations (Corces et al. Nat. Genet. 2016) only identified erythrocyte precursors as statistically relevant (HR 0.27, 95% confidence interval 0.12-0.61). Of note, H8 had a distinct gene expression profile compared to that identified for the sorted LMPP sub-population using differential gene expression analysis. In summary, we identified novel cell type clusters and surface marker associations using combined single cell WTA and surface marker analysis (BD® Abseq). We were able to correlate cell types with both canonical and novel surface markers, and deconvolution analysis provided preliminary insights into their clinical relevance in AML. Citation Format: Asiri Ediriwickrema, Sreejith Ramakrishnan, Margaret Nakamoto, Smita Ghanekar, Bogdan Luca, Aaron Newman, Andrew Gentles, Ravindra Majeti. Multiomic single cell analysis of normal human bone marrow identifies a unique stem and progenitor population that expands in AML [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 3779.

Abstract 2750: An efficient multiplexing approach to avoid batch effects in single cell RNA- and Ab-seq studies demonstrated in a mouse model of chronic inflammation and obesity

Abstract Obesity can weaken the body's immune system and reduce its ability to fight off infections. Current analysis on immune cells from obese mouse models suggests that the immune cells suffer from chronic inflammation. To understand the root cause of this problem, we used single cell sequencing to examine thousands of immune cells isolated from primary and secondary lymphoid organs as well as adipose tissue and compared a diet-induced obesity mouse model with control mice. Cells were stained with 30 BD® AbSeq DNA-barcoded antibodies to enable multiomic analysis, i.e. examining protein alongside mRNA expression in tandem. We also utilized DNA-barcoded universal antibodies from the BD® Single-Cell Multiplexing Kit, which allowed us to combine 8 samples from different mice and tissue types into a single pooled sample, significantly reducing experimental scale and cost while eliminating potential batch effects. The pooled samples were loaded on the BD Rhapsody™ system to perform cell lysis and individual mRNA and cell barcoding, allowing measurement of ~400 immune-related mRNAs and 30 proteins at the single cell level. We were able to efficiently de-multiplex the pooled samples after sequencing. The targeted mRNA and AbSeq panel provided robust clustering of immune cell types and showed that genes related to critical immune responses, including inflammation and lymphocyte activation, are differentially regulated in specific immune-cell subsets in the obese mouse model. Using this multiomic analysis of genes differentially regulated in immune cells from different tissues, we propose a model to explain the immuno-phenotype we observed in obese mouse. For Research Use Only. Not for use in diagnostic or therapeutic procedures. BD, the BD Logo, and Rhapsody are trademarks of Becton, Dickinson and Company or its affiliates. © 2019 BD. All rights reserved. Citation Format: Hye-Won Song, Gisele Baracho, Ian Taylor, Stephanie Widmann, Stefanie Mortimer. An efficient multiplexing approach to avoid batch effects in single cell RNA- and Ab-seq studies demonstrated in a mouse model of chronic inflammation and obesity [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 2750.

Single-Nucleus RNA Sequencing of the Hypothalamic Arcuate Nucleus of C57BL/6J Mice After Prolonged Diet-Induced Obesity.

Prolonged obesity is associated with blunted feeding and thermogenic autonomic responses to leptin, but cardiovascular responses to leptin are maintained. This state of selective leptin resistance is, therefore, proposed to contribute to the pathogenesis and maintenance of obesity-associated hypertension. Cells of the arcuate nucleus of the hypothalamus detect leptin, and although the cellular and molecular mechanisms remain unclear, altered arcuate nucleus biology is hypothesized to contribute to selective leptin resistance. Male C57BL/6J mice were fed a high-fat diet (HFD) or chow from 8 to 18 weeks of age, as this paradigm models selective leptin resistance. Nuclei were then isolated from arcuate nucleus for single-nucleus RNA sequencing. HFD caused expected gains in adiposity and circulating leptin. Twenty-three unique cell-type clusters were identified, and Ingenuity Pathway Analysis was used to explore changes in gene expression patterns due to chronic HFD within each cluster. Notably, gene expression signatures related to leptin signaling exhibited suppression predominantly in neurons identified as the Agouti-related peptide (Agrp) subtype. Ingenuity Pathway Analysis results were also consistent with alterations in CREB (cAMP response element-binding protein) signaling in Agrp neurons after HFD, and reduced phosphorylated CREB was confirmed in arcuate nucleus after prolonged HFD by capillary electrophoresis-based Western blotting. These findings support the concept that prolonged HFD-induced obesity is associated with selective changes in Agrp neuron biology, possibly secondary to altered CREB signaling.

2061-P: Islet Single Cell Transcriptomes Reveal Cellular and Gene Regulatory Alterations in Type 2 Diabetes Pathogenesis

Pancreatic islet (dys)function and failure is central to both type 2 diabetes (T2D) genetic risk and pathogenesis. Islet cellular heterogeneity has impeded precise understanding of the specific gene(s) and pathway(s) in each cell type that are altered by and/or may contribute to pathophysiology in the prediabetic (PD) and type 2 diabetic (T2D) states. To identify cellular and molecular features of islet dysfunction in PD and T2D states, we completed single cell transcriptome profiling of 137k human islet cells from 11 ND, 6 PD, and 11 T2D donors (⋃5k/individual). Single cell transcriptomes aggregated into distinct clusters of endocrine (alpha, beta, delta, PP/gamma), exocrine (acinar, stellate, ductal), endothelial, and immune cell types. Targeted analysis of each endocrine cell type identified a proliferative alpha cell sub-population comprising ∼0.15-0.3% of endocrine cells and variable insulin production and endoplasmic reticulum stress beta cell states. We compared the proportion of islet endocrine cell type (sub)populations sampled between T2D, PD, and ND donors and found that T2D samples contained significantly fewer beta cells and more alpha cells than ND and, to a lesser degree, PD samples. Delta and PP/gamma cell proportions were similar between disease states. Importantly, no (sub)populations were ND, PD, or T2D-specific. We compared aggregate T2D, PD, and ND cell type-specific transcriptomes to identify differentially expressed genes in each cell type. Genes related to exocytic granule assembly/organization, including those induced by T2D risk variants (DGKB) or T2D state (GAP43), were among 99 genes induced in T2D beta cells. 103 repressed genes included those associated with glucose transport and implicated as T2D GWAS effector genes such as SLC2A2, SLC5A1, and STARD10, a gene whose deletion in mice impairs islet insulin secretion. Studies are ongoing to assess the functional role(s) of these and additional genes in islet (dys)function. Disclosure N. Lawlor: None. R. Kursawe: None. M.L. Stitzel: None. Funding American Diabetes Association/Pathway to Stop Diabetes (1-18-ACE-15 to M.L.S.)

SAT-298 Integrative Single-Cell Transcriptomic and Epigenomic Landscape of Mouse Anterior Pituitary Cell Types

The pituitary gland is a critical regulator of the neuroendocrine system. To further our understanding of the classification, cellular heterogeneity, and regulatory landscape of pituitary cell types, we performed and computationally integrated single cell (SC)/single nucleus (SN) resolution experiments capturing RNA expression, chromatin accessibility, and DNA methylation state from mouse dissociated whole pituitaries. Both SC and SN transcriptome analysis and promoter accessibility identified the five classical hormone-producing cell types (somatotropes, gonadotropes (GT), lactotropes, thyrotropes, and corticotropes). GT cells distinctively expressed transcripts for Cga, Fshb, Lhb, Nr5a1, and Gnrhr in SC RNA-seq and SN RNA-seq. This was matched in SN ATAC-seq with GTs specifically showing open chromatin at the promoter regions for the same genes. Similarly, the other classically defined anterior pituitary cells displayed transcript expression and chromatin accessibility patterns characteristic of their own cell type. This integrated analysis identified additional cell-types, such as a stem cell cluster expressing transcripts for Sox2, Sox9, Mia, and Rbpms, and a broadly accessible chromatin state. In addition, we performed bulk ATAC-seq in the LβT2b gonadotrope-like cell line. While the FSHB promoter region was closed in the cell line, we identified a region upstream of Fshb that became accessible by the synergistic actions of GnRH and activin A, and that corresponded to a conserved region identified by a polycystic ovary syndrome (PCOS) single nucleotide polymorphism (SNP). Although this locus appears closed in deep sequencing bulk ATAC-seq of dissociated mouse pituitary cells, SN ATAC-seq of the same preparation showed that this site was specifically open in mouse GT, but closed in 14 other pituitary cell type clusters. This discrepancy highlighted the detection limit of a bulk ATAC-seq experiment in a subpopulation, as GT represented ~5% of this dissociated anterior pituitary sample. These results identified this locus as a candidate for explaining the dual dependence of Fshb expression on GnRH and activin/TGFβ signaling, and potential new evidence for upstream regulation of Fshb. The pituitary epigenetic landscape provides a resource for improved cell type identification and for the investigation of the regulatory mechanisms driving cell-to-cell heterogeneity.Additional authors not listed due to abstract submission restrictions: N. Seenarine, M. Amper, N. Jain (ISMMS).