Single-cell Multiomics Research Articles

PP 1.16 – 00100 Single-cell multiomics analysis reveals distinct mechanisms of HIV persistence in memory CD4+ T cell subsets from tissues

PP 1.16 – 00100 Single-cell multiomics analysis reveals distinct mechanisms of HIV persistence in memory CD4+ T cell subsets from tissues

A single-cell atlas of glioblastoma evolution under therapy reveals cell-intrinsic and cell-extrinsic therapeutic targets

Recent longitudinal studies of glioblastoma (GBM) have demonstrated a lack of apparent selection pressure for specific DNA mutations in recurrent disease. Single-cell lineage tracing has shown that GBM cells possess a high degree of plasticity. Together this suggests that phenotype switching, as opposed to genetic evolution, may be the escape mechanism that explains the failure of precision therapies to date. We profiled 86 primary-recurrent patient-matched paired GBM specimens with single-nucleus RNA, single-cell open-chromatin, DNA and spatial transcriptomic/proteomic assays. We found that recurrent GBMs are characterized by a shift to a mesenchymal phenotype. We show that the mesenchymal state is mediated by activator protein 1. Increased T-cell abundance at recurrence was prognostic and correlated with hypermutation status. We identified tumor-supportive networks of paracrine and autocrine signals between GBM cells, nonmalignant neuroglia and immune cells. We present cell-intrinsic and cell-extrinsic targets and a single-cell multiomics atlas of GBM under therapy.

Fibroblast inflammatory priming determines regenerative versus fibrotic skin repair in reindeer

Adult mammalian skin wounds heal by forming fibrotic scars. We report that full-thickness injuries of reindeer antler skin (velvet) regenerate, whereas back skin forms fibrotic scar. Single-cell multi-omics reveal that uninjured velvet fibroblasts resemble human fetal fibroblasts, whereas back skin fibroblasts express inflammatory mediators mimicking pro-fibrotic adult human and rodent fibroblasts. Consequently, injury elicits site-specific immune responses: back skin fibroblasts amplify myeloid infiltration and maturation during repair, whereas velvet fibroblasts adopt an immunosuppressive phenotype that restricts leukocyte recruitment and hastens immune resolution. Ectopic transplantation of velvet to scar-forming back skin is initially regenerative, but progressively transitions to a fibrotic phenotype akin to the scarless fetal-to-scar-forming transition reported in humans. Skin regeneration is diminished by intensifying, or enhanced by neutralizing, these pathologic fibroblast-immune interactions. Reindeer represent a powerful comparative model for interrogating divergent wound healing outcomes, and our results nominate decoupling of fibroblast-immune interactions as a promising approach to mitigate scar.

Human fetal cerebellar cell atlas informs medulloblastoma origin and oncogenesis.

Medulloblastoma (MB) is the most common malignant childhood brain tumour1,2, yet the origin of the most aggressive subgroup-3 form remains elusive, impeding development of effective targeted treatments. Previous analyses of mouse cerebella3-5 have not fully defined the compositional heterogeneity of MBs. Here we undertook single-cell profiling of freshly isolated human fetal cerebella to establish a reference map delineating hierarchical cellular states in MBs. We identified a unique transitional cerebellar progenitor connecting neural stem cells to neuronal lineages in developing fetal cerebella. Intersectional analysis revealed that the transitional progenitors were enriched in aggressive MB subgroups, including group 3 and metastatic tumours. Single-cell multi-omics revealed underlying regulatory networks in the transitional progenitor populations, including transcriptional determinants HNRNPH1 and SOX11, which are correlated with clinical prognosis in group 3 MBs. Genomic and Hi-C profiling identified de novo long-range chromatin loops juxtaposing HNRNPH1/SOX11-targeted super-enhancers to cis-regulatory elements of MYC, an oncogenic driver for group 3 MBs. Targeting the transitional progenitor regulators inhibited MYC expression and MYC-driven group 3 MB growth. Our integrated single-cell atlases of human fetal cerebella and MBs show potential cell populations predisposed to transformation and regulatory circuitries underlying tumour cell states and oncogenesis, highlighting hitherto unrecognized transitional progenitor intermediates predictive of disease prognosis and potential therapeutic vulnerabilities.

Low-dose IL-2 reduces IL-21+ T cell frequency and induces anti-inflammatory gene expression in type 1 diabetes

Despite early clinical successes, the mechanisms of action of low-dose interleukin-2 (LD-IL-2) immunotherapy remain only partly understood. Here we examine the effects of interval administration of low-dose recombinant IL-2 (iLD-IL-2) in type 1 diabetes using high-resolution single-cell multiomics and flow cytometry on longitudinally-collected peripheral blood samples. Our results confirm that iLD-IL-2 selectively expands thymic-derived FOXP3+HELIOS+ regulatory T cells and CD56bright NK cells, and show that the treatment reduces the frequency of IL-21-producing CD4+ T cells and of two innate-like mucosal-associated invariant T and Vγ9Vδ2 CD8+ T cell subsets. The cellular changes induced by iLD-IL-2 associate with an anti-inflammatory gene expression signature, which remains detectable in all T and NK cell subsets analysed one month after treatment. These findings warrant investigations into the potential longer-term clinical benefits of iLD-IL-2 in immunotherapy.

Single-cell multiomics reveals the complexity of TGFβ signalling to chromatin in iPSC-derived kidney organoids

TGFβ1 plays a regulatory role in the determination of renal cell fate and the progression of renal fibrosis. Here we show an association between SMAD3 and the histone methyltransferase, EZH2, during cell differentiation; ChIP-seq revealed that SMAD3 and EZH2 co-occupy the genome in iPSCs and in iPSC-derived nephron progenitors. Through integration of single cell gene expression and epigenome profiling, we identified de novo ACTA2+ve/POSTN+ve myofibroblasts in kidney organoids treated with TGFβ1, characterised by increased SMAD3-dependent cis chromatin accessibility and gene expression associated with fibroblast activation. We have identified fibrosis-associated regulons characterised by enrichment of SMAD3, AP1, the ETS family of transcription factors, and NUAK1, CREB3L1, and RARG, corresponding to enriched motifs at accessible loci identified by scATACseq. Treatment with the EZH2 specific inhibitor GSK343, blocked SMAD3-dependent cis co-accessibility and inhibited myofibroblast activation. This mechanism, through which TGFβ signals directly to chromatin, represents a critical determinant of fibrotic, differentiated states.

Single-cell multi-omics allows functional characterization of structural variants.

Single-cell multi-omics allows functional characterization of structural variants.

Abstract B016: Quantifying and dissecting pancreatic cancer cell phenotypic plasticity using lineage tracing, single-cell multiomics and CRISPR perturbations reveals novel regulators of plastic states

Abstract Pancreatic cancer is a lethal disease in part because tumor cells exist in distinct transcriptional phenotypes (e.g. basal and classical states), each with a selective ability to evade current chemotherapy regimens. Two major mechanisms have been suggested for treatment evasion: 1) intrinsic resistance of certain phenotypes to particular chemotherapy regimens and 2) plasticity of treatment sensitive phenotypes to adopt more resistant phenotypes. However, the relative contribution of these mechanisms to treatment resistance is still poorly understood. Whereas previous work has described the redistribution of tumor cell states under selective treatment pressure, there is no direct evidence that tumor cells exhibit phenotypic plasticity at steady state or with treatment. By leveraging technological advancements in single-cell methods, lineage tracing and functional genomics, we have now shown direct evidence of phenotypic state switching in human pancreatic cancer cell lines. By performing single-cell RNA-seq on 5 barcoded PDAC cell lines over a steady state timecourse and under chemotherapy selective pressure (>600k cells total), we identify unique plasticity phenotypes within these cell lines and infer regulators of these plastic states. We validate the role of several of these regulators using bulk phenotypic CRISPRi screens in these cell lines. We next perform CRISPRi perturbations along with lineage tracing and single-cell multiomics (>300k cells) to dissect the regulatory relationships that underlie these cell states. We identify several novel epithelial and mesenchymal biasing factors, including those with unique roles in the most plastic clones. Collectively, we nominate several regulators that bias PDAC cell states thus posing a paradigm whereby perturbations may be used to homogenize tumor populations towards treatment-sensitive phenotypes. We believe this approach combined with current chemotherapy regimens could benefit pancreatic cancer patients by targeting residual, resistant tumor cells in the localized and metastatic disease settings to improve patient survival. Citation Format: Arnav Mehta, Lynn Bi, Aziz Al'Khafaji, Martin Jankowiak, Milan Parikh, Mehrtash Babadi, Alex Bloemendal, Marc Schwartz, Glen Munson, Joeseph Chan, Cassandra Burdziak, Elisa Donnard, Ryan Park, Chen Lu, Philippe Rigollet, Andrew Aguirre, Vidya Subramanian, Ray Jones, Eric S. Lander, David T. Ting, Dana Pe'er, Nir Hacohen. Quantifying and dissecting pancreatic cancer cell phenotypic plasticity using lineage tracing, single-cell multiomics and CRISPR perturbations reveals novel regulators of plastic states [abstract]. In: Proceedings of the AACR Special Conference on Pancreatic Cancer; 2022 Sep 13-16; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2022;82(22 Suppl):Abstract nr B016.

Single Cell Multi-Omic Profiling of Multiple Myeloma with t(4;14) Identifies a T-Cell Population That Correlates with Clinical Outcomes

Background Multiple Myeloma (MM) is the second most common haematologic malignancy and patients with high risk MM have adverse outcomes. Translocation (4;14) occurs in 15% of MM and is associated with inferior survival. However, outcomes are heterogeneous even within this subset of patients. We set out to identify factors that correlated with progression free survival (PFS) in a cohort of t(4;14) patients by using high dimensional profiling of the malignant plasma cells (PCs) and tumor micro-environment through single cell multi-omics. Methods We analysed diagnostic bone marrow (BM) samples from 13 patients with t(4;14) MM using the ESCAPE™ platform from Singleron Biotechnologies which simultaneously measures gene and cell surface protein expression of 62 proteins in single cells. Cryopreserved BM samples were stained with antibodies and subsequently sorted on CD138 expression. The CD138 positive and negative fractions were recombined at a 1:1 ratio for analysis using the 10x Genomics 3' RNAseq kit. Resulting data were analyzed with Singleron Biotechnology's MapSuite™ single cell analytics platform. Subsequent validation of the t(4;14) findings were performed using the 5' chemistry version of ESCAPE™ with the 10x Genomics RNAseq platform. Results All patients received either bortezomib or thalidomide-based induction and had a median age of 63 years. Median PFS and overall survival (OS) were 22 and 34 months respectively. Consistent with published bulk gene expression data, MMSET was overexpressed in all PCs, while FGFR3 expression was heterogeneous. No gene or protein expression patterns within the PCs were identified that correlated with PFS or OS. In the tumor microenvironmental compartment, we identified a CD3+ population (cluster 26) that had not been previously annotated in published single cell RNA-seq data. These cells were CD45+ and CD138- at both the protein and RNA level, suggesting they are not PCs. Independent bioinformatics tools "MapCell" and "DISCO" both identified cluster 26 cells as T cells. The proportion of cluster 26 cells among non-PCs correlated negatively with PFS (P=0.04, logrank test), with samples containing a higher proportion of these cells having an inferior PFS compared to those with a lower proportion (Figure 1 panel A). We subsequently demonstrated that the T cells identified in cluster 26 could also be detected in BM samples from a separate cohort of MM patients analysed using the ESCAPE platform. These patients were treated with daratumumab in the first line or relapsed setting and included patients without t(4;14). Analysis of 15 patients from this cohort revealed that the percentage of cluster 26 cells retained its negative correlation with PFS (P=0.03, logrank test) (Figure 1 panel B).Conclusions We identified a T-cell population which may have prognostic significance in patients with t(4;14 )MM. Interestingly, this population also retains its negative correlation with PFS in a cohort of non t(4;14) patients, suggesting it may have a broader significance across cytogenetic subtypes of MM. The second cohort of patients were treated with daratumumab, hence the novel population may also be relevant in the response to anti -CD38 monoclonal antibody therapy. These data highlight the potential of single cell multi-omic analysis to identify immune micro-environmental signatures that correlate with clinical outcomes in MM. Further work is ongoing to definitively characterize these cells and explore the biological basis for their clinical impact. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Needle in a Haystack: A Pilot Study Combining Single-Cell Multiomics with Clinical NGS-MRD Sequencing to Search for Circulating Clonotypic Dedifferentiated Myeloma Cells

Multiple myeloma (MM) is a heterogenous disease in which clonal evolution or preexisting reservoirs of phenotypically divergent cells may be harbingers of therapeutic resistance. Prior research has suggested the presence of myeloma cells with immunophenotypic and expression profiles dedifferentiated from that of terminally differentiated plasma cells that may be found in circulation, marrow, or potentially other lymphoid compartments; however, this has not been conclusively demonstrated in patients. The identification of such clonotypic myeloma cells could lead to a better understanding of myeloma biology and resistance, and inform therapies targeted at these populations, and single-cell technologies provide the granularity to characterize them. Next-generation sequencing (NGS)-based measurable residual disease (MRD) clinical-grade assays (eg. clonoSEQ®, Adaptive Biosciences) identify individuals’ unique heavy and light chain sequences. Theoretically this sequencing data could be overlaid with research-grade single-cell multiomics to identify clusters of clonotypic cells with variable differentiation. In this pilot study using the Enhanced Single Cell Analysis with Protein Expression (ESCAPE) platform (Singleron Biotechnologies) we aimed to identify peripherally circulating clonotypic MM cells by their NGS-MRD sequence and simultaneously determine whether their multiomic profile reflected possible dedifferentiated clusters found in paired marrow. Using scRNA-Seq (10x Genomics) data previously generated at our site from bone marrow mononuclear cells we identified MM cell clusters enriched for differentially expressed genes known to be expressed in B-cells and plasmablasts. These genes included CD19, PAX5, BCL6, EZH2, STMN1, HMGB2, MKI67 and CD27. We next sought to assay the peripheral blood mononuclear cells (PBMC) of patients (n=2) with an overrepresentation of cells in these clusters. To evaluate both the transcriptomic and immunophenotypic profiles of the PBMCs at single-cell resolution we utilized Singleron's Immune Profiling panel on the ESCAPE platform. Using Singleron's MapSuite™ analysis pipeline, we annotated cell-type classifications, and highlighted cells within the mature B-cell/plasma cell differentiation spectrum (Figure A). Within this subset, we attempted to identify gene expression reflective of the previously identified "dedifferentiated” marrow clusters with specific genes of interest, but found that there was not significant overlap. Heatmaps comparing the 5 peripheral subclusters with marrow plasma cells demonstrate some concordance with the marrow subclusters, with examination of these overlaps ongoing, but these were not the "dedifferentiated” marrow clusters. The included patients also had clonoSEQ® identification sequences of heavy chain and/or light chain available. By cross-referencing the NGS-MRD sequence data (clonoSEQ®, Adaptive Biosciences), we were able to use ESCAPE-Seq to identify definitive clonotypic circulating MM cells within the PBMC samples at single cell granularity. Among PBMCs from sample 2 we were able to identify circulating clonotypic myeloma cells with 100% match to the IgH and IgK MRD sequences, which also represented the dominant V(D)J clonotype in peripheral circulating. Concomitant single-cell surface protein analysis of this clonotypic cell population demonstrated significant expression of CD38, without significant differential expression of CD138 nor surface markers of B-cells (Figure B). While these data are preliminary and studies are ongoing, we successfully leveraged clinical-level MRD sequencing data to inform single-cell multiomics, a potentially powerful translational research pathway for any hematologic malignancy in which NGS-MRD is clinically available. We definitively identified a small subset of circulating MM cells and used the ESCAPE platform to efficiently generate a multiomics profile. This profile suggested elevated CD38 expression without elevated CD138 expression, which previously has been associated with dedifferentiated plasma cells and plasmablasts, although further analyses are ongoing. In future studies, we will enrich for B-lymphoid spectrum cells prior to single-cell analytics. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Longitudinal Single Cell Multiomic Analysis of a Patient with GATA2 Haploinsufficiency with ASXL1 Mutant Clonal Hematopoiesis

Introduction: GATA2 haploinsufficiency syndrome (G2HS) is an autosomal dominant germline predisposition state with broad phenotypic variability and high risk of developing myeloid neoplasms (74.3%, median age of onset 17 years; Homan CC, et al. 2021). Mechanisms resulting in myeloid transformation remain to be elucidated. To identify molecular changes and relevant drivers of gene expression prior to progression, we performed serial, single cell multiomics on a proband with G2HS and ASXL1 mutant clonal hematopoiesis (CH) over 7 years. Case report: The proband was diagnosed with G2HS at age 19 years when he presented with recurrent infections, scrotal and penile lymphedema and cytopenias. Bone marrow (BM) biopsy at diagnosis was hypocellular with minimal atypia. Next generation sequencing (NGS) identified a novel pathogenic variant in GATA2 (NM_001145661.2: c.1143+200_1198delinsC; Dorn J et al, 2017) confirmed to be germline, and a somatic pathogenic variant in ASXL1 (NM_015338.6: c.2638dupA; p.Thr880Asnfs*2) with a variant allele frequency (VAF) of 11%, indicative of CH. Methods: After IRB approval, clinical data was abstracted from the medical records. Four aliquots of BM mononuclear cells from BM biopsies obtained 6 months apart were subjected to targeted NGS, single cell RNA (scRNA-seq) and scATAC-seq. Preprocessing and poor-quality cell filtering were performed using the Cell Ranger pipeline. For scRNA-seq, filtered counts were clustered and annotated using Seurat v4.1.0 onto a CITEseq PBMC reference (Hao Y, et al. 2021). Genes were ranked within each cluster between samples based on their average log-fold change, and gene set enrichment analysis (GSEA) was performed using the fgsea package v1.16.0. In parallel, filtered scATAC-seq data was integrated with scRNA using the cluster-label transfer from Signa v1.5.0 and Seurat v4.1.0. Using MAESTRO (Wang C, et al. 2020), we predicted the transcriptional factors (TF) regulating the expression from differentially expressed genes and differential peaks between samples based on the enrichment scores from algorithms LISA (Qin Q, et al. 2020) and GIGGLE (Layer RM, et al. 2018) for RNAseq and ATACseq respectively. Results: Serial blood cell counts, blood indices and BM biopsies over 7 years remained stable (macrocytosis, monocytopenia and neutropenia), with the BM demonstrating stable hypocellularity (10-20%) with minimal atypia and with no evidence for progression to myelodysplastic syndrome (MDS). ASXL1 mutant VAF also remained stable (11 to 16% at last measurement). Quantitative flow cytometric assessments in blood revealed decreases in B and NK cells. Analysis of scRNA/scATAC-seq successfully identified monocytes, B-cells, CD4, CD8 and NK Cells yielding comparable counts to those obtained by flow cytometry. The G2HS-associated cytopenias resulted in very small numbers of monocytes and B cells being identified, and hence differential gene expression analysis was limited to CD4+, CD8+ T cells and NK cells, between time points 1 and 3 (Figure 1A). Differentially expressed genes showed aberrant enrichment of downregulated genes in pathways related to hematopoietic stem cell (HSC) maturation in all three populations, while upregulated genes were enriched in cell cycle and innate immune responses. Analysis of regulatory drivers identified BRD4 as the main TF controlling the differentially expressed genes in CD4+ T cells and NK cells, and ERG in CD8+ T cells. The expression of these TF did not change between timepoints. (Figure 1B) Discussion: We present a GATA2 haploinsufficient proband who has remained stable and MDS-free over 7 years, even with ASXL1 mutant CH. Despite the stability in his blood counts and BM, single cell analysis has identified progressive qualitative changes impacting relevant pathways including HSC maturation, cell cycle regulation and innate immune responses in T and NK cells. In addition, our scATAC-seq data predicts BRD4 and ERG to be important epigenetic readers/TF that regulate these changes. We plan to study larger cohorts of patients and additional time points to understand transcriptional and epigenetic changes induced by GATA2 haploinsufficiency and elucidate the role of ASXL1 mutant CH in disease progression. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Single-Cell Multi-Omics Analysis Reveals the Role of CX3CR1+ GNLY+ CD8+ T Cells in Improving Survival of Patients with Relapse/Refractory Hematologic Malignancies after Haplo+Cord HSCT

Background Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a curative treatment for hematological malignancies. Previous works have shown superior survival outcomes of sequential transplantation of haploidentical stem cell and unrelated cord blood (haplo+cord HSCT) among recipients with relapsed/refractory leukemia, compared with single cord HSCT. However, the underlying mechanisms of immune reconstitution remain largely unexplored. Methods we first performed scRNA-seq and scATAC-seq on bone marrow mononuclear cells (BMMCs) from nine patients with relapsed/refractory leukemia underwent haplo+cord or single cord HSCT to profile the immune landscapes at single cell resolution. Second, we compared the differences of gene expression and chromatin accessibility between the subsets of immune cells. Finally, by integrating multi-omics data, we explored the gene expression regulation mechanism behind the differences between two transplant group. Results We observed distinct composition and function of global immune landscapes (T, B, natural killer cells and monocytes) between two transplant strategies, where immune cells from haplo+cord HSCT involved in interferon signaling pathway and response to virus. Analysis on T cells revealed two cytotoxic CD8+ T cells populations, enriched in haplo+cord HSCT (expressing GNLY, GZMB and PRF1) and enriched in single cord HSCT (expressing CXCR4 and GZMK) respectively. Interestingly, the augmented expression of GNLY and CX3CR1 suggested a dual property of effector and memory of haplo+cord HSCT enriched CD8+ T cells that potentially offers long-term protection for patients post-transplantation. The transcription factor IRF7, IRF9 and TBX21 played the key role on regulating immune signal transduction of CD8+ T cell subsets after haplo+cord HSCT, which might be the mechanism of transcriptional regulation for cytotoxic CD8+T cell function. Finally, single cell multi-omics analysis delineate regulatory network controlling lineage-specific gene expression after HSCT, illustrated by the example of GNLY expression in CD8+ T cells regulated by an enhancer cluster within CTCF-mediated chromatin loop. Conclusions These data collectively constitute the transcriptomic and epigenetic landscape of bone marrow post-transplantation in hematopoietic malignancies at single-cell resolution, demonstrating that a CX3CR1+ GNLY+ CD8+ T cells characterized by effector and memory property served the improved survival of patients receiving haplo+cord HSCT. keywords Single cell multi-omics, CX3CR1+ GNLY+ CD8+T cells, haplo+cord HSCT, single cord HSCT, relapsed/refractory leukemia

Single-Cell Multi-Omics of Peripheral Blood Reveals Tiding and Evolution of Responsive and Resistant Clones upon Daratumumab-Based Treatments in Multiple Myeloma and Plasma Cell Leukemia

Background: Treatment with anti-CD38 antibodies is a standard of care in newly diagnosed and relapsed/refractory multiple myeloma (MM). Clonal tiding and evolution as well as genomic heterogeneity are major drivers towards drug-resistant disease. However, modes of resistance to anti-CD38 directed therapies are currently unknown and might include loss of antigen expression and/or exhaustion of the immune system. We performed longitudinal single-cell multi-omic analyses of circulating tumor cells (CTC) and non-malignant cells to decipher clonal heterogeneity and tiding upon response and resistance to daratumumab (Dara)-based therapies in MM and plasma cell leukemia (PCL). Methods: Peripheral blood mononuclear cells (MNCs) were isolated from 7 patients (PCL n=5, MM n=2) before and after exposure to Dara. Samples were collected at four different time points: 1. before administration of Dara (CD38-naïve). 2. Four days after first application of Dara (CD38-exposed). 3. After relapse from Dara (CD38-refractory). 4. After an additional line of non-CD38 targeted therapy following relapse from Dara. In total, we analyzed 17 PB samples. To demonstrate that results from PB are comparable to bone marrow (BM), four paired PB and BM samples were included (21 samples in total). MNCs were subjected to single cell RNA (scRNAseq), B-cell (scBCRseq), and T-cell receptor sequencing (scTCRseq). Surface protein expression was measured using a custom panel of 57 DNA-barcoded antibodies. Sequencing files were processed with CellRanger (10x Genomics) and data were analyzed with the Seurat toolkit for single cell sequencing. Cell types were annotated with SingleR based on the Human Primary Cell Atlas built-in references. ScRepertoire was used for longitudinal scTCRseq and scBCRseq data. Copy number variations (CNVs) in CTCs were detected with inferCNV, and cellular interactions were investigated with CellPhoneDB. Results: In total, we sequenced 130.038 MNCs (PB: 108.236, range 414-8700/patient; BM: 21802, range 3375-8516/patient, Figure 1A). Malignant plasma cells were identified using scBCRseq and restricted light chain expression from scRNAseq. Even in MM patients without morphological evidence for CTC, we could identify and characterize single malignant plasma cells in PB. By inferring genome-wide CNVs from scRNAseq, we were able to characterize intra-patient genomic heterogeneity. Longitudinal tracking of CTC revealed tiding of clones characterized by distinct CNV profiles and transcriptomes upon Dara-exposure and resistance. Early transcriptional changes in CTCs upon Dara-exposure included inflammatory signaling (e.g. upregulation of IFI6, IFI35, IFIT1, IFITM1). Analysis of CD38 protein surface expression using DNA-barcoded antibodies showed significantly decreased levels after Dara exposure. To rule out that this might have resulted from competitive binding of the diagnostic antibody and Dara, we repeated analyses after discontinuing Dara and after application of a further line of therapy. In contrast to CD38 RNA expression, surface protein expression decreased significantly upon exposure and relapse, but recovered after an additional line of therapy. Expression of other druggable antigens (BCMA, SLAMF7, GPRCD5, FcRL5, Figure 1B) was not affected by treatment with Dara. Next, we investigated co-evolution of non-malignant cells. Longitudinal scTCRseq showed concomitant expansion and stagnation of TCRs. Receptor-ligand interaction analyses demonstrated that exposure to Dara resulted in significant differences in CD38-PECAM1-mediated interactions between CTCs with non-malignant B-cells and CD8+ T-cells. We detected recurrently increased immunosuppressive receptor-ligand interactions between malignant plasma cells and CD4+, CD8+, and regulatory T-cells via CD86-CTLA4 and NECTIN3-TIGIT axis upon Dara exposure. Conclusion: We demonstrate that single-cell multi-omic analyses of PB provide a minimal invasive approach to study genomic heterogeneity and evolution of malignant plasma cells and non-malignant cells of the immune system. Longitudinal analyses reveal different patterns of response and resistance to targeted myeloma therapies and might affect future therapeutic decision-making. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

A Novel Role for PLK1 in Leukemia Stem Cell Function in Acute Myeloid Leukemia

Acute myeloid leukemia (AML) is an aggressive hematologic malignancy with poor prognosis. While most patients achieve initial remission following standard chemotherapy, the majority will relapse and ultimately die. Poor outcomes in AML are tied to the properties of leukemia stem cells (LSC), which drive therapy resistance and relapse. For successful therapeutic targeting, it is critical to understand the pathways that operate in LSC function and survival and how they are impacted by drug activity. We carried out an innovative stemness-focused high-throughput drug screen employing a hierarchical AML cell model and a Nanostring-based LSC signature assay to probe for genes and pathways that impact LSC function. All 4 small molecule inhibitors of Polo-like Kinase 1 (PLK1) included in the compound library were among the top hits. PLK1 inhibitor treatment selectively targeted the CD34⁺CD38⁻ fraction of OCI-AML-8227 cells, which is enriched for quiescent LSCs. PLK1 is an important cell cycle regulator; although there are reports for cell cycle-independent roles, they have not been systematically characterized. Importantly, the role of PLK1 has not been previously studied in the context of cancer stem cells. To investigate the role of PLK1 in LSCs, we performed single-cell (sc) multi-omics incorporating both scRNAseq and scATACseq on the CD34⁺CD38⁻ compartment of OCI-AML-8227 cells. Enrichment analysis of previously-identified chromatin accessibility signatures of primitive long-term hematopoietic stem cells (LT-HSC) and progenitors was used to identify cell clusters with distinct stemness and differentiation properties. Interestingly, PLK1 mRNA expression in OCI-AML-8227 was restricted to cells enriched for the LT-HSC epigenetic signature (Fig 1A). In healthy human bone marrow, PLK1 expression was restricted to committed progenitors and not detected in HSCs (Fig 1B). These data strongly suggest a previously unexplored role of PLK1 in regulating stemness that may be specific LSCs and not normal HSCs. To determine the effects of PLK1 against LSCs in vivo, we carried out PLK1 inhibitor treatment using the specific PLK1 inhibitor volasertib in a cohort of 9 primary AML samples in PDX models. Volasertib treatment significantly reduced leukemic burden in 7 of 9 samples, all of which had a high LSC17 score (associated with poor prognosis and standard chemotherapy response). To quantify effects on LSCs, we serially transplanted 2 samples at limiting doses into untreated secondary recipients and demonstrated a 5- to 25-fold reduction in LSC frequency following volasertib treatment, indicating that functional LSCs were targeted by PLK1 inhibition. Together, these data strongly suggest that PLK1 impacts LSC function and stemness, and that PLK1 inhibitors are acting through an uncharacterized cell-cycle independent mechanism to target quiescent LSCs. To discover novel PLK1 signaling pathways, we carried out proximity-dependent biotinylation (PDB) to characterize the PLK1 interactome in serum-starved cells. Preliminary analysis has revealed a number of previously uncharacterized PLK1 interactors, including components of autophagy and apoptosis machinery, that are currently being validated in the hematopoietic cell context. Our study highlights PLK1 as a specific vulnerability of LSCs, and furthers the understanding of malignant PLK1 signaling in non-cell cycle related processes. Further insights into the underlying mechanisms will enable the development of effective therapies to target these pathways and improve patient outcomes in AML. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Integrative Analysis of Blast Phase Chronic Myelomonocytic Leukemia Reveals Distinct Transformation Phenotypes Predictive of Drug Response

Introduction: Blastic transformation of chronic myelomonocytic leukemia (CMML) is a common and terminal event. Blast phase CMML (BP-CMML) displays clinical and genetic heterogeneity, but its clonal evolution and transcriptional dynamics are poorly understood. Improved stratification would better characterize transformation phenotypes and identify opportunities for personalized therapy. Methods: Blasts/HSPCs were FACS sorted (SSCloCD45wk) from 43 BP-CMML patients and 7 age-matched healthy controls for bulk RNA-sequencing. Blast immunophenotype by flow cytometry was integrated with clinical metadata in an unsupervised Random Forest classifier. Clonal architecture was investigated using the Tapestri DNA+Protein single cell platform and a bespoke pipeline. Drug testing comprised Cell Titer Glo assay of primary BP-CMML cells. Results: BP-CMML blasts displayed 2090 differentially expressed genes vs control HSPCs. Upregulated genes (n=1424) included the AP-1 transcription factors (JUN, JUNB, JUND), ATF3, and ID1; and included 110 genes from our previous single cell RNA seq of BP-CMML, including CXCL8, HLX, LGALS1 and MKNK1. We sought genes/pathways upregulated in both bulk and single cell datasets, prioritizing by known function and availability of inhibitors. We identified 12 targets, including RAF mitogenic effectors (VEGFA, AKT, PI3K, MTOR), the pERK downstream effector MKNK1, HIF1A, STAT3/5 and NFKB. BP-CMML cells were consistently sensitive to AKT, MKNK1 and HIF1A inhibition, irrespective of genotype or prior therapy, with an additive effect at therapeutically relevant doses for combined AKT/MKNK1 inhibition. Observed heterogeneity between BP-CMML cases suggested distinct clinicopathological phenotypes. We integrated clinical metadata with blast immunophenotype (59 independent features) for unsupervised learning by Random Forest, then overlaid RNA-seq. This clearly separated 2 groups ("C1/C2"), with distinctive clinical and molecular features (eg C2 enriched for ASXL1 (p=0.004) and RUNX1 (p=0.044) mutations; C1 for NPM1 (p=0.031)), but no difference in survival. C2 displayed an immature, stem-like phenotype (higher expression of CD34, CD117 and enriched HSC transcriptional signature). C1 reflected a more mature myelomonocytic phenotype (higher CD123, CD13, CD135, CD56) with enrichment of monocytic/granulocytic gene signatures (Fig A). Amongst genes dysregulated in BP-CMML, 1213 and 402 were unique to C1 or C2, respectively. C1 was enriched for myeloid genes (CD14, S100A8/9, TLR4) and related pathways; C2 for ribosomal, mitogenic (MAP3K8) and FOS family. Genie3 highlighted distinct regulatory networks in C1 and C2, driven by myeloid differentiation (SPI1, TFEB, KLF4) and proto-oncogene (JUND, ATF3, ETV6) transcription factors, respectively. Intriguingly, they displayed different/opposing signatures for resistance to azacitidine (C2>C1; p=0.004), venetoclax (C1>C2; p=0.007) and cytarabine (C1>C2; p=0.0003). We validated for C1 absolute and relative (vs C2) ex vivo resistance to venetoclax, daunorubicin and cytarabine, but saw no difference in azacitidine response. Multiomic single cell genotyping (n=8) revealed patterns of clonal expansion linked to secondary genetic events, with all harboring acquired mutations in RAS pathway or IDH1/2. Concurrent profiling of 45 surface proteins uncovered subtype-specific differences in origin and extent of differentiation block arising from blasts. Dominant genetic clones in C2 were restricted to phenotypically primitive (CD34+117+) cells with little maturation beyond. By contrast, fully evolved clones in C1 were widely distributed from primitive to mature cells, indicating transformation initiating upstream with a 'leaky' differentiation block and continued multilineage contribution of blasts to the pool of maturing cells (Fig B). Conclusion: Integration of multimodal data reveals distinct CMML transformation phenotypes associated with unique transcriptional states and drug sensitivities, mirroring the refractoriness of BP-CMML to AML therapy. Subgroups share mitogenic gene expression and ex vivo sensitivity to AKT/MKNK1 inhibition as a promising novel combination therapy. Finally, we describe different routes of clonal evolution, confirming that BP-CMML blasts can retain extensive differentiation capacity to supplant multilevel hematopoiesis after transformation. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Single Cell Multiomic Analysis Reveals Relapsed and Refractory Multiple Myeloma Cells Associated with 1q, TP53, and PHF19 alterations That Affect Subclonal Chromatin Accessibility

Introduction: Myeloma is a plasma cell malignancy in which a high-risk set of patients can be defined by genomic markers including cytogenetic groups such as loss of 1p, gain or amplification (gain/amp) 1q, and TP53 abnormalities. Increasingly, other markers are also being identified that are associated with progression such as PHF19. These markers are highly prevalent at relapse, but can exist in the earlier stages of the disease. Despite these cytogenetic groups being well studied in relation to patient outcomes, there is still not yet a clear association with subclonal structure and subclonal chromatin accessibility. Methods: Bone marrow aspirates from Smoldering Multiple Myeloma (SMM; n=10), Newly Diagnosed Multiple Myeloma (NDMM; n=22), and Relapsed/Refractory Multiple Myeloma (RRMM; n=17) patients after treatment with Immunomodulatory, proteasome inhibitors, and immunotherapies. All these samples underwent CD138+ sorting and single cell multiomic sequencing (RNA-seq and ATAC-seq; 10X Genomics). The multiomic single cell data were integrated using Seurat via weighted nearest neighbors (WNN). Copy number variants (CNVs) were generated using InferCNV. Cell clusters were evaluated for cytogenetic markers, enrichment in disease progression, gene expression, and other markers of high-risk disease. In total, 335,025 high quality plasma cells were examined across the 49 patient samples. Results: Integration of all 49 patient samples resulted in 25 clusters representing diverse subsets of malignant and normal plasma cells. We identified a subtype of plasma cells that were associated with RRMM disease and were denoted relapse/refractory-associated plasma cells (RRPCs). The relative proportion of this cluster across patients significantly increased from SMM to NDMM (P = 0.025) and from NDMM to RRMM (Fig. 1A, P < 0.001). Patients with gain/amp 1q or TP53 mutations had a greater proportion of RRPCs (P = 0.029 and P = 0.023, respectively). When evaluated further for regulatory mechanisms of cell cycle, the RRPCs had increased expression of PHF19 (Log2FC = 2.79, P < 0.001) and EZH2 (Log2FC = 2.46, P < 0.001), a downstream effector of PHF19, which affects histone methylation and chromatin accessibility. RNA-seq data generated more distinct clusters per sample compared to ATAC-seq data (3 RNA clusters vs. 2 ATAC clusters, P=0.001). We identified frequent subclonal 1q amplifications and 17p deletions across SMM, NDMM, and RRMM patients. Samples with alterations on 1p, 1q, 8q, 9q, or 11q, had more than one distinct cluster in both the RNA (OR=5.96, P = 0.034) and ATAC data (OR = 12.27, P = 0.001). 1p alterations alone were associated with samples containing more than 2 distinct RNA clusters (OR=5.96, P = 0.034). Furthermore, the number of distinct ATAC clusters correlated with the total number of 1p, 1q, 8q, 9q, and 11q CNV events (PCC=0.35, PCC p-value=0.016). Genomic positions indicated key genes on these chromosomes that correlate with ATAC clustering are TENT5C (1p) and MYC (8q), and PHF19 (9q) as they relate to subclonal chromatin accessibility and disease risk. Notably, PHF19 is located on 9q and had increased copy number for RRPCs compared to other clusters (Fig. 1B, P < 0.001). Likewise, and mirroring the patient level TP53 mutational status, the inferred TP53 CNV status was lower in RRPC cells compared to other clusters (Fig. 1B, P < 0.001). In RRMM patients, inferred CKS1B CNV status was associated with GPRC5D expression (PCC = 0.27, P < 0.001), a potential target for myeloma with gain/amp 1q. Conclusion: We have generated the largest dataset of CD138+ myeloma cell multiomics to date comprising 335,025 cells across 49 patients at different stages of progression. Based on these data, expression of high-risk markers are largely found in a subset of cells from any given patient and this subset increases during myeloma progression. These cells, denoted RRPCs, have increased expression of PHF19 and EZH2 and are associated with 9q and PHF19 copy number changes which have measurable effects on chromatin accessibility. Furthermore, greater proportions of RRPCs were found in patients with 1q and TP53 alterations and RRPCs tended to have lower inferred TP53 copy number. Using RNA clusters, ATAC clusters, and inferred CNV status, we found that 1p, 1q, 8q, 9q, and 11q alterations accounted for subclonal chromatin accessibility differences and GPRC5D was identified as a potential target for gain/amp 1q clones. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Single-Cell Multiomics of Pre- and Post- Remission Bone Marrow from a Pediatric Patient with JMML Reveal Distinct Progenitors and Inflammatory Monocytes Which Are Eliminated By Chemotherapy

Juvenile myelomonocytic leukemia (JMML) is a rare myeloproliferative/myelodysplastic disorder with a poor prognosis. Unlike many other hematologic malignancies, it remains challenging to rapidly diagnose, and levels of residual disease are challenging to measure. Ongoing efforts to characterize the pathophysiology of JMML have yielded significant insight, but a complete picture of the cellular hierarchy of the disease has been elusive. To obtain an unsupervised view of the heterogeneity of hematopoietic cell types in JMML, we performed single-cell RNA multiomics on serial samples from a 4-year-old patient who presented with fevers, hepatosplenomegaly, leukocytosis, monocytosis, elevated Hemoglobin F, and bone marrow blast count of 7%. Clinical sequencing revealed mutations in PTPN11 (VAF 66%), NF1 (VAF 17%), MYB (VAF 12%), and SH2B3. We obtained a sample for multiomic profiling following his initial cycles of chemotherapy when PTPN11 variant frequency was 32%, and he remained critically ill. The subsequent sample (6 weeks later) was collected when no mutated PTPN11 was detectible, and his clinical picture had significantly improved. Both samples were processed using gradient centrifugation, and a subset of the cells captured was immunomagnetically enriched CD34+ cells. A panel of 141 oligonucleotide-tagged antibodies was used to label the cells prior to single-cell capture. The resulting libraries were sequenced using standard short-read sequencing, and reads were demultiplexed and counted using common workflows. Single-cell profiles were visualized using UMAP. Cells were annotated using a highly performant machine-learning algorithm that classifies cells using a healthy reference of human BM, which we have previously validated using publicly available data. Examining all cells together, we observed changes in two key cell types over time (Figure 1). First, we noted a large population of monocyte-like cells with increased expression of genes associated with complement, inflammation (including IL1beta), and KRAS signaling in the pre-remission sample, which nearly completely disappeared in the second sample. Moreover, surface expression of CD39 was significantly higher in these cells than in all other cells measured (Figure 2). These data suggest CD39 may represent a novel biomarker to monitor treatment response in JMML or as a potential target for therapy. Second, we observed two distinct hematopoietic stem cell (HSC) populations in the pre-remission sample and only one in the remission sample. Differential gene expression analysis between the two pre-remission HSC clusters revealed enrichment of specific genes previously associated with JMML progenitors in one, including HOPX. This population of HSCs also exhibited higher expression CD7, which correlated with the patient's flow cytometric data from the same time point. We hypothesize that CD7/HOPX expressing HSCs in the first sample were JMML progenitor cells eliminated by therapy. Our work demonstrates the power of single-cell multiomics to dissect cellular heterogeneity in diseases with complex clonal hierarchies. Moreover, these data can also be a resource for identifying potential biomarkers and drug targets. Specifically, our data suggest that inflammatory monocytes may be a distinct target for therapy, which may benefit critically ill patients like ours. Future efforts will be focused on profiling additional patients and rigorously characterizing the mutational state of individual cells. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

Single-Cell Multi-Omics to Infer Hierarchical Ancestry and Organization of Subclones in Acute Lymphoblastic Leukemia

Single-Cell Multi-Omics to Infer Hierarchical Ancestry and Organization of Subclones in Acute Lymphoblastic Leukemia

Unveiling Highly Fluctuating Single-Cell Transcriptome Dynamics at a Novel Bifurcation Point in a Human Lymphoid Pathway

Background: Human hematopoiesis was considered a hierarchical stepwise process, but on the basis of recent progress in single-cell RNA-sequencing (scRNA-seq) it is now considered a continuous process. However, the uncertainty or fluctuation of single-cell transcriptome dynamics during differentiation was not considered. To address this issue, we developed a new computational methodology that converts single-cell splicing kinetics into an advection-diffusion model in a latent cell state using deep learning technology, which enabled us to quantitatively capture the fluctuation in cell state dynamics. Because the human lymphoid pathway and bifurcation points for dendritic cells (DC)/monocytic lineage remain incompletely defined, we applied our method to explore the lymph-DC pathway in the human lympho-hematopoietic system. Materials and methods: CD34+ or CD34+CD38+CD45RA+CD19-CD7-CD10+ cells (CD10SPs) were sorted from a mixture of frozen human cord blood samples, and scRNA-seq and single-cell chromatin accessibility analysis (scATAC) were performed. The data were applied to the variational inference of cell state dynamics with fluctuation (VICDYF). Subsequently, the obtained findings were validated by novel mesenchymal stem cell (MSC)-like coculture system that supports comprehensive human hematopoiesis. Results: MSC-like cultures of various lymphoid populations showed that the CD10SP population, which had been thought to be B/NK progenitors, was a heterogenous population in which lineage diversification from multi-lymphoid progenitors toward lymphoids and DCs was suspected to occur. Our insilico analysis of CD34+ cells with VICDYF suggested a B/plasmacytoid DC (pDC) bifurcation with high fluctuating transcriptome dynamics in the CD10-expressing fraction among various levels of fluctuating bifurcation points. Analysis of CD10SPs identified the B/pDC bifurcation as a highly fluctuating transcriptome state in the IL-7Rα+ fraction in association with the LFA-1 expression dynamics and pDC differentiation bias. Consistent with these findings, B/pDC common progenitors were identified by single cell culture, and the LFA-1-high population gave rise to a higher percentage of pDCs than proB cells, while the LFA-1-low population produced a higher percentage of proB cells than pDCs in the c-kit+IL-7Rα+ fraction. Of note, the epigenetic regulator SFMBT2 was specifically expressed in the B/pDC bifurcation region. The scATAC of the CD34+ population revealed that the B/pDC bifurcation with large fluctuation coincided with the high accessibility of both sides of the differentially accessible chromatin regions for both B and pDC lineages. This study revealed a new landscape of the human lymphoid pathway that includes crucial bifurcation points with highly fluctuating single-cell transcriptome dynamics. Discussion: Our computational methodology discovered a novel B/pDC bifurcation as high fluctuating dynamics. The insilico predictions were confirmed by the biological findings, using MSC-like cultures. Although the underlying molecular mechanism remains elusive, single-cell multiomics analysis suggests that the estimated large fluctuation is associated with epigenetic changes in chromatin regions for both B and pDC lineages. VICDYF also identified the highly fluctuated transcriptome dynamics at the putative lineage bifurcation point during pancreatic development. Conclusion: Our results provide a new model of differentiation - fluctuation based differentiation - which reconciles continuous and stepwise models. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal

STEM-10. TEMPORAL MULTI-MODAL SINGLE-CELL ANALYSES REVEAL DYNAMIC INTERACTIONS BETWEEN GLIOBLASTOMA AND CAR-T CELLS AND IMMUNOLOGIC MODULATORS OF CANCER STEM CELL STATE

Abstract CAR-T cell therapy is a promising new immunotherapy for a number of difficult-to-treat cancers, however, it has yet to yield broad success in glioblastoma (GBM). In particular, tumor heterogeneity presents a major therapeutic challenge, and a detailed understanding of the complex interplay between different neoplastic, non-neoplastic, and CAR-T cells is critical for developing better treatments. Using a patient-derived GBM organoid model of CAR-T cell therapy, we performed single-cell multi-omics to examine the longitudinal dynamics of the adaptive tumor response, changes in cell states, and evolution of cell-to-cell interaction networks. We find that all tumor cell types - neoplastic and non-neoplastic - respond to CAR-T cell activity, and they generate to an initially anti-tumor, but subsequently pro-tumor and immune-inhibitory microenvironment, which is accompanied by eventual CAR-T cell dysfunction and exhaustion. Unexpectedly, CAR-T cell activity also leads to attenuation of glioma stem-like states in both antigen-positive and antigen-negative neoplastic cells and reduces their proliferation via diffusible factors, of which IFNɣ is required but not sufficient. These findings are supported by analyses in patient samples from CAR-T cell therapy clinical trials, and they are consistent across both de novo and recurrent tumors with different somatic mutational landscapes. Our study unravels how the complex heterogeneity of GBM interacts with CAR-T cell therapy, and we identify previously unappreciated possibilities to affect antigen-negative neoplastic cells in ways that may be further augmented for enhanced therapeutic efficacy.