Single-cell RNA-seq Data Research Articles

Single-cell and bulk RNA-sequencing reveal SPP1 and CXCL12 as cell-to-cell communication markers to predict prognosis in lung adenocarcinoma.

Lung adenocarcinoma (LUAD) generally presents as an immunosuppressive microenvironment. The characteristics of cell-to-cell communication in the LUAD microenvironment has been unclear. In this study, the LUAD bulk RNA-seq data and single-cell RNA-seq data were retrieved from public dataset. Differential expression genes (DEGs) between LUAD tumor and adjacent non-tumor tissues were calculated by limma algorithm, and then detected by PPI, KEGG, and GO analysis. Cell-cell interactions were explored using the single-cell RNA-seq data. Finally, the first 15 CytoHubba genes were used to establish related pathways and these pathways were used to characterize the immune-related ligands and their receptors in LUAD. Our analyses showed that monocytes or macrophages interact with tissue stem cells and NK cells via SPP1 signaling pathway and tissue stem cells interact with T and B cells via CXCL signaling pathway in different states. Hub genes of SPP1 participated in SPP1 signaling pathway, which was negatively correlated with CD4+ T cell and CD8+ T cell. The expression of SPP1 in LUAD tumor tissues was negatively correlated with the prognosis. While CXCL12 participated in CXCL signaling pathway, which was positively correlated with CD4+ T cell and CD8+ T cell. The role of CXCL12 in LUAD tumor tissues exhibits an opposite effect to that of SPP1. This study reveals that tumor-associated monocytes or macrophages may affect tumor progression. Moreover, the SPP1 and CXCL12 may be the critic genes of cell-to-cell communication in LUAD, and targeting these pathways may provide a new molecular mechanism for the treatment of LUAD.

Single-cell RNA sequencing reveals aberrant sphingolipid metabolism in non-small cell lung cancer impacts tumor-associated macrophages and stimulates angiogenesis via macrophage inhibitory factor signaling.

Sphingolipids not only serve as structural components for maintaining cell membrane fluidity but also function as bioactive molecules involved in cell signaling and the regulation of various biological processes. Their pivotal role in cancer cell development, encompassing cancer cell proliferation, migration, angiogenesis, and metastasis, has been a focal point for decades. However, the contribution of sphingolipids to the complexity of tumor microenvironment promoting cancer progression has been rarely investigated. Through the integration of publicly available bulk RNA-seq and single-cell RNA-seq data, we conducted a comprehensive analysis to compare the transcriptomic features between tumors and adjacent normal tissues, thus elucidating the intricacies of the tumor microenvironment (TME). Disparities in sphingolipid metabolism (SLM)-associated genes were observed between normal and cancerous tissues, with the TME characterized by the enrichment of sphingolipid signaling in macrophages. Cellular interaction analysis revealed robust communication between macrophages and cancer cells exhibiting low SLM, identifying the crucial ligand-receptor pair, macrophage inhibitory factor (MIF)-CD74. Pseudo-time analysis unveiled the involvement of SLM in modulating macrophage polarization towards either M1 or M2 phenotypes. Categorizing macrophages into six subclusters based on gene expression patterns and function, the SPP1+ cluster, RGS1+ cluster, and CXCL10+ cluster were likely implicated in sphingolipid-induced M2 macrophage polarization. Additionally, the CXCL10+, AGER+, and FABP4+ clusters were likely to be involved in angiogenesis through their interaction with endothelial cells. Based on multiple scRNA-seq datasets, we propose that a MIF-targeted strategy could potentially impede the polarization from M1 to M2 and impair tumor angiogenesis in low-SLM non-small cell lung cancer (NSCLC), demonstrating its potent antitumor efficacy.

Abstract LB253: Crafted experiments to evaluate feature selection methods for single cell RNA-seq data

Abstract Single-cell RNA sequencing (scRNA-seq) technologies have revolutionized our understanding of cellular compositions and tumor behavior at the single-cell level, and have been widely used in many biomedical applications, such as identifying and characterizing novel cell types/states. While numerous computational methods have been developed for analyzing scRNA-seq data, benchmarking various analytical methods remains a key challenge for two reasons. First, there is a lack of ground truth datasets where unambiguous cell type identities are known using external information. Second, simulation of scRNA-seq data tends to be overly artificial and simplistic, which may not well mimic the underlying biological complexity as well as high level of technical variability. Here, we propose the use of crafted experiments, a new approach based upon perturbing signals in a real dataset for comparing different scRNA-seq analytical methodologies. We demonstrate the effectiveness of crafted experiments in the context of a novel univariate distribution-oriented suite of feature selection methods, called GOF (Goodness of Fit). We show that GOF more frequently selects features that robustly identify crafted artificial clusters in crafted experiments and achieves similar performance compared to the field standard method using real datasets. Importantly, we show the use of crafted experiments for identifying the contexts in which each method performs the best. Crafted experiments offer valuable comparisons of scRNA-seq analysis methods, and the crafted datasets generated from this study provide a useful resource for the single-cell community. Citation Format: Siyao Liu, David Corcoran, Susana Garcia-Recio, Charles Perou, J.S. Marron. Crafted experiments to evaluate feature selection methods for single cell RNA-seq data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(7_Suppl):Abstract nr LB253.

Abstract LB248: Interrogating the cancer-immunity cycle through the lens of a large-scale single-cell atlas of immune checkpoint inhibition response and resistance

Abstract Introduction. The cancer-immunity cycle (CIC; Chen et al. 2013; Mellman et al. 2023) systematizes the current understanding of the factors (both suppressive and promoting) that together determine the anticancer immune responses. While the CIC is not limited to any single therapeutic modality, the framework is useful for understanding the mechanisms of immune checkpoint inhibition (ICI) response and resistance. Given its importance in the field, it is critical to be able to systematically examine and expand the CIC (or related frameworks) using new high-throughput data. Here we intersect the CIC framework with a large-scale atlas of single-cell RNA-seq data to provide a window into the CIC factors supported across multiple tumor types. Approach. We have previously developed the Single-Cell Immune Checkpoint Inhibition Atlas (ICI Atlas) which in its original version spanned 27 datasets and 1.9 million cells across a range of tumor types and ICI therapies. Motivated by rapid growth of data in the field, here we develop version 2 of the Atlas, over twice as large and consisting of 55 datasets and 4.4 million cells with cross-dataset integration via a transformer-based deep learning model. Of the 55 datasets, 47 contain both pre- and post-treatment samples, 2 contain only pre-treatment samples, and 6 contain only post-treatment samples. The Atlas spans 124 cell types and covers 19 different cancer types as well as 31 treatment strategies including 19 ICI therapies. Results. To determine the ICI Atlas support for each factor in the CIC, we analyzed the percent of studies where the factor’s cell type-specific pre-treatment expression was significantly higher in ICI responders (for promoting factors) or in non-responders (for inhibiting factors). We found that 16 cell type-specific CIC factors (12 promoting and 4 inhibiting) had strong support in the ICI Altas (supported in ≥ 50% of relevant studies). Specific steps of the CIC were associated with well-supported promoting factors, e.g. Step 2: interferon alpha/beta signaling pathway (dendritic cells/APCs, 50-100% of studies); Step 3: TCR signaling pathway (T cells; 58%); Step 5: CXCL9/10/11 (macrophages, 50-75%) and CCL5 (T cells, 58%); Step 5a: SELL (T cells, 68%); or inhibiting factors, e.g. Step 5b: IL10 (T cells, 53%). For some factors the support is tissue-specific, e.g. in tumor samples type-I IFN signaling was higher in dendritic cells and monocytes of responders while in blood this was reversed. Interestingly, certain co-stimulatory factors showed higher expression in non-responders (e.g. OX40 in T-cells, 53% of analyses, and 4-1BBL in macrophages, 50%). Our analysis also identifies a number of new factors whose expression is associated with response to ICI across the majority of ICI Atlas studies, providing opportunities for expanding current knowledge of cancer-immune system interplay and ICI response. Citation Format: Piotr Topolewski, Zofia Kochańska, Marlena Osipowicz, Roy Ronen, Janusz Dutkowski. Interrogating the cancer-immunity cycle through the lens of a large-scale single-cell atlas of immune checkpoint inhibition response and resistance [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 2 (Late-Breaking, Clinical Trial, and Invited Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(7_Suppl):Abstract nr LB248.

Elucidating the role of FOS in modulating the immune microenvironment through fibroblast and myeloid cell regulation in locoregional recurrent HNSCC.

Head and neck squamous cell carcinoma (HNSCC) presents a significant clinical challenge, particularly due to its high propensity for locoregional recurrence. Current research underscores the need to unravel the complex interactions within the tumor microenvironment. This study addresses the critical gap in understanding how FOS modulates the immune landscape in HNSCC, with a focus on its influence on fibroblast and myeloid cell dynamics. Employing a comprehensive approach, we analyzed tissue samples from HNSCC patients and adjacent non-cancerous tissues using bulk RNA sequencing complemented by in-depth bioinformatics analyses, including gene ontology (GO), Kyoto encyclopedia of genes and genomes (KEGG) pathway analysis, and immune infiltration assessment. A pivotal aspect of our research involved dissecting single-cell RNA-seq data from GSE234933 to elucidate the cell-type-specific expression of FOS. We found that FOS expression varies significantly in different cell populations in the HNSCC tumor microenvironment, especially in fibroblasts and myeloid cells. This expression difference may reflect the different roles of these cells in tumor progression and their impact on the tumor microenvironment. Our results uncover a significant correlation between FOS expression and key immune and hypoxia-related pathways, suggesting its integral role in the tumor microenvironment. These findings not only enhance our understanding of HNSCC pathogenesis but also highlight FOS as a potential therapeutic target. This study marks a significant step towards addressing the urgent need for targeted interventions in HNSCC, particularly in the context of locoregional recurrence.

Insights for disease modeling from single-cell transcriptomics of iPSC-derived Ngn2-induced neurons and astrocytes across differentiation time and co-culture

BackgroundTrans-differentiation of human-induced pluripotent stem cells into neurons via Ngn2-induction (hiPSC-N) has become an efficient system to quickly generate neurons a likely significant advance for disease modeling and in vitro assay development. Recent single-cell interrogation of Ngn2-induced neurons, however, has revealed some similarities to unexpected neuronal lineages. Similarly, a straightforward method to generate hiPSC-derived astrocytes (hiPSC-A) for the study of neuropsychiatric disorders has also been described.ResultsHere, we examine the homogeneity and similarity of hiPSC-N and hiPSC-A to their in vivo counterparts, the impact of different lengths of time post Ngn2 induction on hiPSC-N (15 or 21 days), and the impact of hiPSC-N/hiPSC-A co-culture. Leveraging the wealth of existing public single-cell RNA-seq (scRNA-seq) data in Ngn2-induced neurons and in vivo data from the developing brain, we provide perspectives on the lineage origins and maturation of hiPSC-N and hiPSC-A. While induction protocols in different labs produce consistent cell type profiles, both hiPSC-N and hiPSC-A show significant heterogeneity and similarity to multiple in vivo cell fates, and both more precisely approximate their in vivo counterparts when co-cultured. Gene expression data from the hiPSC-N show enrichment of genes linked to schizophrenia (SZ) and autism spectrum disorders (ASD) as has been previously shown for neural stem cells and neurons. These overrepresentations of disease genes are strongest in our system at early times (day 15) in Ngn2-induction/maturation of neurons, when we also observe the greatest similarity to early in vivo excitatory neurons. We have assembled this new scRNA-seq data along with the public data explored here as an integrated biologist-friendly web-resource for researchers seeking to understand this system more deeply: https://nemoanalytics.org/p?l=DasEtAlNGN2&g=NES.ConclusionsWhile overall we support the use of the investigated cellular models for the study of neuropsychiatric disease, we also identify important limitations. We hope that this work will contribute to understanding and optimizing cellular modeling for complex brain disorders.

Identifying gene expression programs in single-cell RNA-seq data using linear correlation explanation

ObjectiveGene expression analysis through single-cell RNA sequencing (scRNA-seq) has revolutionized our understanding of gene regulation in diverse cell types, tissues, and organisms. While existing methods primarily focus on identifying cell type-specific gene expression programs (GEPs), the characterization of GEPs associated with biological processes and stimuli responses remains limited. In this study, we aim to infer biologically meaningful GEPs that are associated with both cellular phenotypes and activity programs directly from scRNA-seq data. MethodsWe applied linear CorEx, a machine-learning-based approach, to infer GEPs by grouping genes based on total correlation optimization function in simulated and real-world scRNA-seq datasets. Additionally, we utilized a transfer learning approach to project CorEx-inferred GEPs to other scRNA-seq datasets. ResultsBy leveraging total correlation optimization, linear CorEx groups genes and demonstrates superior performance in identifying cell types and activity programs compared to similar methods using simulated data. Furthermore, we apply this same approach to real-world scRNA-seq data from the mouse dentate gyrus and embryonic colon development, uncovering biologically relevant GEPs related to cell types, developmental ages, and cell cycle programs. We also demonstrate the potential for transfer learning by evaluating similar datasets, showcasing the cross-species sensitivity of linear CorEx. ConclusionOur findings validate linear CorEx as a valuable tool for comprehensively analyzing complex signals in scRNA-seq data, leading to deeper insights into gene expression dynamics, cellular heterogeneity, and regulatory mechanisms.

Identification of Key Genes in Angiogenesis of Breast and Prostate Cancers in the Context of Different Cell Types.

Angiogenesis involves the development of new blood vessels. Biochemical signals start this process in the body, which is followed by migration, growth, and differentiation of endothelial cells that line the inside wall of blood vessels. This process is vital for the growth of cancer cells and tumors. We started our analysis by composing a list of genes that have a validated impact in humans with respect to angiogenesis-related phenotypes. Here, we have investigated the expression patterns of angiogenesis-related genes in the context of previously published single-cell RNA-Seq data from prostate and breast cancer samples. Using a protein-protein interaction network, we showed how different modules of angiogenesis-related genes are overexpressed in different cell types. In our results, genes, such as ACKR1, AQP1, and EGR1, showed a strong cell type-dependent overexpression pattern in the two investigated cancer types, which can potentially be helpful in the diagnosis and follow-up of patients with prostate and breast cancer. Our work demonstrates how different biological processes in distinct cell types contribute to the angiogenesis process, which can provide clues regarding the potential application of targeted inhibition of the angiogenesis process.

Multi-Omics Profiling Reveals Phenotypic and Functional Heterogeneity of Neutrophils in COVID-19.

Accumulating evidence has revealed unexpected phenotypic heterogeneity and diverse functions of neutrophils in several diseases. Coronavirus disease (COVID-19) can alter the leukocyte phenotype based on disease severity, including neutrophil activation in severe cases. However, the plasticity of neutrophil phenotypes and their relative impact on COVID-19 pathogenesis has not been well addressed. This study aimed to identify and validate the heterogeneity of neutrophils in COVID-19 and evaluate the functions of each subpopulation. We analyzed public single-cell RNA-seq, bulk RNA-seq, and proteome data from healthy donors and patients with COVID-19 to investigate neutrophil subpopulations and their response to disease pathogenesis. We identified eight neutrophil subtypes: pro-neutrophil, pre-neutrophil, immature neutrophil, and five mature neutrophil subpopulations. The subtypes exhibited distinct features, including diverse activation signatures and multiple enriched pathways. The pro-neutrophil subtype was associated with severe and fatal disease, while the pre-neutrophil subtype was particularly abundant in mild/moderate disease. One of the mature neutrophil subtypes showed consistently large fractions in patients with different disease severity. Bulk RNA-seq dataset analyses using a cellular deconvolution approach validated the relative abundances of neutrophil subtypes and the expansion of pro-neutrophils in severe COVID-19 patients. Cell-cell communication analysis revealed representative ligand-receptor interactions among the identified neutrophil subtypes. Further investigation into transcription factors and differential protein abundance revealed the regulatory network differences between healthy donors and patients with severe COVID-19. Overall, we demonstrated the complex interactions among heterogeneous neutrophil subtypes and other blood cell types during COVID-19 disease. Our work has great value in terms of both clinical and public health as it furthers our understanding of the phenotypic and functional heterogeneity of neutrophils and other cell populations in multiple diseases.

CTEC: a cross-tabulation ensemble clustering approach for single-cell RNA sequencing data analysis.

Cell-type clustering is a crucial first step for single-cell RNA-seq data analysis. However, existing clustering methods often provide different results on cluster assignments with respect to their own data pre-processing, choice of distance metrics, and strategies of feature extraction, thereby limiting their practical applications. We propose Cross-Tabulation Ensemble Clustering (CTEC) method that formulates two re-clustering strategies (distribution- and outlier-based) via cross-tabulation. Benchmarking experiments on five scRNA-Seq datasets illustrate that the proposed CTEC method offers significant improvements over the individual clustering methods. Moreover, CTEC-DB outperforms the state-of-the-art ensemble methods for single-cell data clustering, with 45.4% and 17.1% improvement over the single-cell aggregated from ensemble clustering method (SAFE) and the single-cell aggregated clustering via Mixture model ensemble method (SAME), respectively, on the two-method ensemble test. The source code of the benchmark in this work is available at the GitHub repository https://github.com/LWCHN/CTEC.git.

SEGCECO: Subgraph Embedding of Gene expression matrix for prediction of CEll-cell COmmunication.

Recent advances in single-cell RNA sequencing technology have eased analyses of signaling networks of cells. Recently, cell-cell interaction has been studied based on various link prediction approaches on graph-structured data. These approaches have assumptions about the likelihood of node interaction, thus showing high performance for only some specific networks. Subgraph-based methods have solved this problem and outperformed other approaches by extracting local subgraphs from a given network. In this work, we present a novel method, called Subgraph Embedding of Gene expression matrix for prediction of CEll-cell COmmunication (SEGCECO), which uses an attributed graph convolutional neural network to predict cell-cell communication from single-cell RNA-seq data. SEGCECO captures the latent and explicit attributes of undirected, attributed graphs constructed from the gene expression profile of individual cells. High-dimensional and sparse single-cell RNA-seq data make converting the data into a graphical format a daunting task. We successfully overcome this limitation by applying SoptSC, a similarity-based optimization method in which the cell-cell communication network is built using a cell-cell similarity matrix which is learned from gene expression data. We performed experiments on six datasets extracted from the human and mouse pancreas tissue. Our comparative analysis shows that SEGCECO outperforms latent feature-based approaches, and the state-of-the-art method for link prediction, WLNM, with 0.99 ROC and 99% prediction accuracy. The datasets can be found at https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE84133 and the code is publicly available at Github https://github.com/sheenahora/SEGCECO and Code Ocean https://codeocean.com/capsule/8244724/tree.

Integration of Single-Cell and Bulk RNA-seq Data to Identify the Cancer-Associated Fibroblast Subtypes and Risk Model in Glioma.

Cancer-associated fibroblasts (CAFs) are an important component of the stroma. Studies showed that CAFs were pivotally in glioma progression which have long been considered a promising therapeutic target. Therefore, the identification of prognostic CAF markers might facilitate the development of novel diagnostic and therapeutic approaches. A total of 1333 glioma samples were obtained from the TCGA and CGGA datasets. The EPIC, MCP-counter, and xCell algorithms were used to evaluate the relative proportion of CAFs in glioma. CAF markers were identified by the single-cell RNA-seq datasets (GSE141383) from the Tumor Immune Single-Cell Hub database. Unsupervised consensus clustering was used to divide the glioma patients into different distinct subgroups. The least absolute shrinkage and selection operator regression model was utilized to establish a CAF-related signature (CRS). Finally, the prognostic CAF markers were further validated in clinical specimens by RT‒qPCR. Combined single-cell RNA-seq analysis and differential expression analysis of samples with high and low proportions of CAFs revealed 23 prognostic CAF markers. By using unsupervised consensus clustering, glioma patients were divided into two distinct subtypes. Subsequently, based on 18 differentially expressed prognostic CAF markers between the two CAF subtypes, we developed and validated a new CRS model (including PCOLCE, TIMP1, and CLIC1). The nomogram and calibration curves indicated that the CRS was an accurate prognostic marker for glioma. In addition, patients in the high-CRS score group had higher immune infiltration and tumor mutation burden levels. Moreover, the CRS score had the potential to predict the response to immune checkpoint blockade (ICB) therapy and chemotherapy. Finally, the expression profiles of three CAF markers were verified by RT‒qPCR. In general, our study classified glioma patients into distinct subgroups based on CAF markers, which will facilitate the development of individualized therapy. We also provided insights into the role of the CRS in predicting the response to ICB and chemotherapy in glioma patients.

Abstract 2706: Identification and characterization of cancer-associated polymorphonuclear cell subsets in mice

Abstract Background: Neutrophils are the main type of polymorphonuclear cells (PMN) and comprise most immune cells in the body. PMN cells are upregulated in cancer and accumulate in the tumors. Tumor-associated PMNs comprise a heterogeneous population consisting of immunologically active mature neutrophils and immature polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs). The phenotype of these cells evolves with disease progression and is implicated in modulating immune response. However, there is yet to be a consensus to define PMN populations in tumors and blood. Methods: Here, we used single-cell RNA-seq, bulk RNA sequencing, flow cytometry, and immunohistochemistry (IHC) analysis to characterize cancer-associated PMN-MDSCs in tumor-bearing mice. Results: Single-cell RNA-seq data from lung tumors and healthy individuals revealed five metaclusters (MC): MCA, MCB, MCC, MCD, and MCE. MCA and MCB made up 34.4 and 65.1% of PMNs in healthy tissues, respectively. MCA, MCB, MCC, MCD, and MCE comprised 3.8%, 23.5% 3.3% 14.6%, and 54.7% of PMNs in tumors, respectively. Genes enriched in MCA and MCE were associated with MDSCs and myeloid progenitor cells. MCB and MCC were associated with inflammatory responses mediated by TNF-alpha signaling and neutrophil degranulation. MCD was associated with a CD33 myeloid cell phenotype and had high expression levels of CD274, Pdcd1lg2, Havcr2, and Tigit genes. Cells from MCE were associated with protein translation signatures and had elevated levels of Nt5e. Gene expression signatures from single-cell RNA-seq data were compared with bulk transcriptomic data from healthy mouse prostates and Pten-null prostate tumors characterized by high PMN cell infiltration. Gene set enrichment analysis showed enrichments after androgen deprivation therapy and Cluster MCE was associated with castration-resistant tumors. Flow cytometric and IHC analysis confirmed the enrichment of PMN cells expressing elevated levels of proteins identified from gene expression analysis including CD73, PD-L1, and PD-L2. Conclusion: This study offers further insights into the distinct subsets of PMN-MDSCs present in tumors. Citation Format: Marco A. De Velasco, Yurie Kura, Kazuko Sakai, Takashi Kikuchi, Yasunori Mori, Takafumi Minami, Kazuhiro Yoshimura, Kazutoshi Fujita, Kazuto Nishio, Hirotsugu Uemura. Identification and characterization of cancer-associated polymorphonuclear cell subsets in mice [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2706.

Abstract 151: BMP signaling determines neuroblastoma cell fate and sensitivity to retinoic acid

Abstract While retinoic acid (RA) has been successfully used for leukemia treatment for decades, the attempt to treat solid tumors with RA remains challenging, with less than 10% of neuroblastoma (NB) patients achieving complete remission when treated with RA alone. It has been believed the anti-cancer activity of retinoids is due to retinoid-induced terminal differentiation, however, what determines cell response or whether differentiation is the main effect of RA has never been fully understood. To better understand RA’s activity, we conducted genome-wide CRISPR modifier screens in RA-treated hyper-sensitive NB cell lines. Surprisingly, we found that, in these cells, RA primarily decreased cell viability via apoptosis or senescence, rather than differentiation—activities in which the CRISPR screens strongly implicated bone morphogenetic protein (BMP) signaling. BMP activation promoted apoptosis/senescence and sensitized cells to RA. Conversely, BMP inhibitors and SMAD9 (a critical transcription factor of the BMP pathway) knockout enhanced RA’s ability to induce differentiation but reduced cell sensitivity to RA. Our ChIP-seq and RNA-seq data showed these behaviors were mediated by the coordinated gene regulatory actions of RARA and BMP-family SMAD transcription factors. Furthermore, in a panel of 19 cell lines we screened with RA, we assessed the correlations between RA IC50 and the expression of all (~20,000) genes (publicly available data from the GDSC and Depmap). Remarkably, SMAD9 was the number 1 and number 10 most correlated gene, respectively, with RA IC50 (ranked by Pearson correlation coefficient; GDSC R = -0.92, P = 4.2 × 10−6; Depmap R = -0.81, P = 4.8 × 10−4). We also performed comprehensive large-scale drug screens in these cell lines with combinations of RA and a BMP activator FK506. RA exhibited a synergistic effect with FK506 in all the NB cell lines, very strikingly in some cell lines. All these data suggest BMP signaling is generally required for RA sensitivity and BMP activators are promising candidates in combination with RA to treat NB. Using published bulk and single cell RNA-seq data from NB patient samples, we found BMP signaling activity is relatively low in primary tumors, but well maintained in disseminated NB cells derived from bone marrow. This explains why RA is clinically used as a maintenance therapy and can only successfully treat the minimal residual disease and suggests that tumor microenvironment is a critical factor in determining cell response to RA. Overall, our study revealed that BMP controls NB cell fate and sensitivity in RA treatment. Notably, interactions between BMP signaling and RA are well established in developmental biology, where BMP signaling can tip cell fate decisions between differentiation, apoptosis, and senescence upon exposure to endogenous RA. Our data suggest that this developmental process can be maintained in NB cells, unveiling a novel mode of anti-neoplastic action. Citation Format: Min Pan, Yinwen Zhang, William C. Wright, Hyeong-Min Lee, Richard H. Chapple, Xueying Liu, Jonathan Low, Duane Currier, Allister J. Loughran, Dyer A. Dyer, Shondra M. Pruett, Burgess Freeman, Taosheng Chen, Brian J. Abraham, Elizabeth Stewart, John Easton, Paul Geeleher. BMP signaling determines neuroblastoma cell fate and sensitivity to retinoic acid [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 151.

Abstract 2499: Window-of-opportunity trial of metastatic pancreatic cancer reveals potential mechanisms of response to targeting RAS signaling

Abstract KRAS is mutated in ~85% of pancreatic ductal adenocarcinoma (PDAC) patient's tumors, suggesting that it is a key node of PDAC targeted therapy. RAF/MEK/ERK signaling is downstream of KRAS signaling. Preclinical studies targeting MEK showed promises but clinical trials failed to show the benefit compared to Standard-of-Care treatment chemotherapy. To elucidate the mechanism of action and resistance to several targeted therapies including MEK inhibition in human PDAC, we conducted a Window-Of-Opportunity trial for Metastatic pancreatic cancer (WOOM trial). In this trial, metastatic PDAC patients received 10 days of Cobimetinib, a FDA approved MEK inhibitor, therapy and specimens of pre- and post- treatment were obtained by biopsy to evaluate the response to targeted treatment using deep multi-omic analytics including DNA-seq, RNA-seq, multiplex IHC, cyclic IF and Digital Spatial Profiling (DSP). The clinical grade KI67 staining positivity was used as a readout of tumor response. We found three of 15 patient’s tumors had significantly decreased KI67 staining positivity after Cobimetinib treatment and defined them as Responsive tumors. Genomic sequence showed that KRAS allelic imbalance and KrasG12R mutations were seen significantly higher frequency in Responsive tumors compared to Non-responsive tumors. RNA-seq analysis showed that these Responsive tumors were more likely to be classical-like subtype compared to Non-responsive tumors. DSP analysis confirmed the downregulation of ERK phosphorylation after treatment in all examined cases consistent with effective target inhibition. Interestingly, MEK phosphorylation was upregulated after treatment paradoxically. This suggests that paradoxical upstream pathway activation including RAF activation could be potential resistance mechanism. MYC protein expression change was significantly correlated with KI67 protein expression change, which suggests MYC downregulation could contribute to action of Cobimetinib. We also found that Cobimetinib treatment increased the immunosuppressive TREM2+ Macrophage population from single cell RNA-seq data. Furthermore, PDX from biopsy recapitulated patient's response to therapy and could be a great tool for testing new therapeutic strategies. Here, we summarize the multi-omics data and discuss potential mechanisms of response and resistance to targeting RAS signaling and related microenvironmental changes. Citation Format: Motoyuki Tsuda, Colin Daniel, Xiaoyan Wang, Carl Pelz, Hayley Zimny, Alexander Smith, John Muschler, Xi Li, Tugba Yildiran Ozmen, Furkan Ozmen, Dove Keith, Christopher Corless, Koei Chin, Jonathan Brody, Charles Lopez, Gordon Mills, Rosalie Sears. Window-of-opportunity trial of metastatic pancreatic cancer reveals potential mechanisms of response to targeting RAS signaling [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2499.

Abstract 5459: Inflammatory gene-based subtyping of osteosarcoma patients reveals the association of inflammation in tumor microenvironment with better survival

Abstract Introduction:Inflammation has a significant impact on the tumor microenvironment (TME) of osteosarcoma (OS). Nevertheless, due to the high heterogenecity and variable inflammatory characteristics among OS patients, it is still unclear which inflammatory factors or immune/stroma cell types are crucially associated with OS prognosis. Therefore, it is essential to explore the immune infiltration pattern and the related core inflammatory genes in OS. Method:We combined bulk RNA-seq data of 88 osteosarcoma (OS) patients from the TARGET database and single cell RNA-seq data in GSE152048 from the NCBI Gene Expression Omnibus. Firstly, we calculated the inflammatory gene expression matrix of bulk RNA-seq data by combining 80 inflammatory gene sets (3427 genes in total) from the MSigDB using the gene set variation analysis (GSVA). Secondly, consensus clustering was executed to evaluate the stability of clustering for the above matrix. Furthermore, the overall survival and differential expression analyses were used to screen for core inflammatory genes. Immune infiltration patterns among groups were profiled with CIBERSORT. Finally, we employed the scRNA-seq series to evaluate cell-level information to identify the survival-associated cell subpopulations using the Seurat package. The t-SNE was used for dimensionality reduction and cluster identification. Results:88 osteosarcoma (OS) patients were clustered into two distinct groups according to the consensus clustering of our formed GSVA inflammatory gene expression matrix. Group 1 (43 patients) exhibited evident inflammation features, while Group 2 (45 patients) displayed limited inflammation features. Patients from Group 1 had significantly better overall survival than those from Group 2. The difference in immune infiltration pattern of two groups were examined using bulk RNA-seq data. Group 1 showed a higher population of macrophages and CD8+ T cells compared to Group 2, implying high inflammatory features might lead to a better immune activation of TME in patients. To explore the factors and cell types that contributes to restricted immune infiltration and activation in OS TME, scRNA-seq data of 11 OS patient tissues were analyzed. Consistently with bulk RNA seq analysis, the 338 up-regulated genes from Group 1 were mainly expressed in immune cells including macrophages and myeloid cells. Interestingly, the 64 up-regulated genes from Group 2 were found nearly uniformly expressed in two type of non-immune cells: myoblasts and chondroblastic cells. Conclusion:Inflammatory gene-based subtyping of osteosarcoma (OS) patients reveals that the inflammation in TME led by myeloid-derived immune cells including macrophage is associated with better prognosis. Furthermore, myoblasts and chondroblastic cells might contribute to the limited inflammation response in OS TME. Citation Format: Liu Yang, Yalan Sheng, Shuangying Qiao, Desheng Hu, Debajyoti Chowdhury, Hiu Fung Yip, Zheng Chen, Yun He, Aiping Lu, Fangfei Li. Inflammatory gene-based subtyping of osteosarcoma patients reveals the association of inflammation in tumor microenvironment with better survival [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 5459.

Abstract 6614: Integrating single-cell RNA-seq and large genome-wide association study data to identify colorectal cancer cell-of-origin populations

Abstract Background: Single-cell RNA sequencing (scRNA-seq) enables the precise identification of distinct cell populations and the measurement of gene expression within each population. Recently, integrating scRNA-seq and genome-wide association study (GWAS) data has shown promise in identifying cell-of-origin populations in human diseases. Colorectal cancer (CRC) progresses through a phenotypic continuum from normal colon to adenoma, and ultimately CRC. However, the specific disease cell-of-origin populations underlying each stage of CRC development at a single-cell resolution have not been well studied. Additionally, while transcriptome-wide association studies (TWAS) have identified more than 200 putative risk genes in CRC, the precise disease cell-of-origin populations associated with these genes remain elusive. Materials and methods: We analyzed scRNA-seq datasets from 31 normal colon, 8 colon serrated polyps, and 15 conventional adenoma tissues of individuals participated the Colorectal Molecular Atlas Project (COLON MAP). We next conducted an integrative analysis of the aforementioned scRNA-seq data, with extensive epigenomics data and summary statistics from a large CRC GWAS conducted among European ancestry populations (78,473 CRC cases and 107,143 controls) using single-cell disease relevance score (scDRS), to investigate cell-of-origin populations underlying the CRC development. We conducted correlation analysis between gene expression and individual cell risk scores derived from scDRS. Additionally, we performed differential expression analysis comparing disease cell-of-origin populations with other cell populations at each stage. Results: We identified multiple cell-of-origin populations associated with CRC, including absorptive cells (ABS) (P < 1 × 10-4) and goblet cells (GOB) (P < 7 × 10-3) in the normal colon, ABS (P < 2 × 10-3), enteroendocrine cells (EE) (P < 0.04), and serrated-specific cells (SSC) (P < 1 × 10-4) in serrated polyps, and ABS (P < 9 × 10-3) and SSC (P < 6 × 10-4) in adenomas. Among the previously reported 205 risk genes from TWAS, a majority show high correlations of their expressions with the risk scores of the single cells (median value R2 = 0.07, 0.14, 0.137 at normal colon, serrated polyps, adenoma, respectively). Moreover, 57 of these TWAS-reported genes (27.8% of 205 genes) exhibited evidence of differential expression in disease cell-of-origin populations compared to other cell populations at a nominal p-value < 0.05 in at least one stage (Binomial test, P < 0.01 for all). Conclusion: This study reveals the cell-of-origin populations that influence the development of CRC, and provides evidence of the role of risk genes in these populations. These findings are crucial for identifying causal genes and understanding the cellular mechanisms that drive the biology and etiology of CRC. Citation Format: Chao Li, Zhishan Chen, Qing Li, Jungyoon Choi, Quanhu Sheng, Wanqing Wen, Xiang Shu, Xiao-ou Shu, Jirong Long, Qiuyin Cai, Bhuminder Singh, Martha J. Shrubsole, Ken Lau, Wei Zheng, Xingyi Guo. Integrating single-cell RNA-seq and large genome-wide association study data to identify colorectal cancer cell-of-origin populations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6614.

Abstract 2681: Blockade of SIRPα in glioblastoma cells inhibits mitochondrial fatty acid β-oxidation and increase immune cell mediated cytotoxicity

Abstract Glioblastoma (GBM) is the most malignant brain tumor with limited immunotherapeutic options, likely due to significant GBM-associated systemic and intra-tumoral immunosuppression. The binding of SIRPα on myeloid cells to CD47 on cancer cells restricts immuno-phagocytosis against cancer cells. However, less attention has been given to SIRPα is also expressed in cancer cells, and the role of SIRPα on cancer cells is still unclear. Immunohistochemistry (IHC) staining was used to examine the expression of SIRPα in normal brain tissues and GBM. A single-cell database on the Broad Institute Portal was utilized to analyze the SIRPα expression in different cell types. SIRPα antisense morpholino was applied for gene knockdown. Established stable SIRPα overexpressed GBM cell lines were used to examine the mitochondrial function through the Seahorse Mito Stress test. In vitro, immune cell-mediated cytotoxicity was measured by Agilent xCELLigence Real-Time Cell Analysis. Patient biopsies show a 2-fold increase in SIRPα expression in GBM tumors compared to the normal adjacent (n=14-21; p < 0.01). Single-cell RNA-seq data on GBM tumors shows that SIRPα is expressed not only in the myeloid compartment but also in malignant cells. Studies have shown that mitochondrial function and fatty acid β-oxidation (FAO) play a critical role in GBM progression. Respirometry data revealed that blocking SIRPα in murine GBM cell line (CT2A) decreases mitochondrial ATP production. Seahorse Mito Stress test combined with fuel inhibitors showed that SIRPα expression mainly affects the usage of FAO (approximately 50%). In addition, stably, the GFP-tagged SIRPα overexpressed GBM cell line showed an increase of mitochondrial respiration with enhanced FAO ratio compared to the GFP control cell line. These suggested that SIRPα regulates mitochondrial FAO. Studies have shown that tumor-intrinsic FAO confers cell resistance to immune cell-mediated cytotoxicity. Interestingly, by coculturing with microglia, the innate immune population in the brain, our data showed that SIRPα blockade on GBM enhanced microglia-mediated cytotoxicity. Our data showed that targeting SIRPα on GBM may be a prospective therapeutic option by inhibiting FAO in mitochondria and increasing immune cell-mediated cytotoxicity. Citation Format: Yu-Ting Tsai, Glenn J. Lesser, David R. Soto-Pantoja. Blockade of SIRPα in glioblastoma cells inhibits mitochondrial fatty acid β-oxidation and increase immune cell mediated cytotoxicity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2681.

Abstract 7016: Plastic epigenome in youth: Insights from open chromatin in stem cells

Abstract The presence of plasticity in the epigenomes of young individuals is suggested by the aggressive phenotypes observed in cancers among adolescent and young adults (AYAs), including gastric cancers. To demonstrate the presence of this plasticity and identify the underlying epigenomic characteristics in gastric stem cells during youth, we examined organoids established from murine stomach tissue and conducted single-cell RNA-seq and ATAC-seq using murine antrum-pyloric tissues at ages of 3 weeks and 50 weeks, corresponding to human infancy and middle-age adult, respectively. Organoids established from infant mice showed higher dependence to niche factors, and stronger response to cytokine signals. Ligand-receptor analysis using single-cell RNA-seq data of tissues showed that infant epithelial stem cells and stromal cells had stronger interactions. Both supported the plastic nature of infant gastric epithelial cells. Single-cell ATAC-seq enabled us to identify epithelial stem cell population in infant and middle-aged mice, and 71 and 118 chromatin peaks were specific to infant and adult stem cells, respectively. Motif enrichment analysis showed an enrichment of binding sites for bZIP family transcription factors in infants, and an enrichment of binding sites of a chromatin remodeling factor and a family of transcription factors in adults. Additionally, comparison between differentiated epithelial cells, marked by the pit cell marker Muc5ac, and stem cells, marked by Lgr5, showed more pronounced differences: 46,772 peaks specific to pit cells and 39,076 peaks specific to stem cells. Interestingly, regions specific to pit cells were enriched in binding sites of the chromatin remodeling factor and the family of transcription factors, identical to the open regions specific to adult stem cells. This suggested that the chromatin accessibility profile of aged stem cells approaches to that of differentiated cells, leading to a loss of stemness. The single-cell RNA-seq data also revealed an increase in interactions between differentiated epithelial cells and fibroblasts in aged mice, as opposed to infant, suggesting the emergence of aberrant interactions in aged stomach tissue. Collectively, our data suggested that chromatin accessibility of gastric stem cells undergo an epigenomic shift during aging, approaching to the profiles of differentiated cells and losing their plasticity, and that a specific chromatin remodeling factor and a family of transcription factors could be involved in the process. Citation Format: Yu-Yu Liu, Chihiro Takeuchi, Satoshi Yamashita, Toshikazu Ushijima. Plastic epigenome in youth: Insights from open chromatin in stem cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7016.

Abstract 6212: A differential ligand-receptor network inference method to identify alterations in communication between myeloid cells and CD8 T cells in response to PancVAX neo-epitope peptide vaccine, anti-CTLA-4 and anti-PD-1 antibodies in a murine model of pancreatic ductal adenocarcinoma

Abstract Pancreatic Ductal Adenocarcinoma (PDAC) is a deadly cancer with a low tumor mutational burden that leads to the generation of few neo-epitopes derived from mutated genes that can be recognized by cytotoxic T cells. The immunosuppressive microenvironment of PDAC comprises suppressive cell populations that include myeloid cells, cancer associated fibroblast subpopulations, and regulatory T cells that reduce T cell infiltration and cytotoxic function. We have developed a murine personalized cancer vaccine to enhance cytotoxic T cell function against PDAC, PancVAX (Kinkead et al, JCI Insight 2018). Survival benefit with PancVAX was dependent on the addition of immune checkpoint inhibitor (ICI) therapy to increase the number of cytotoxic CD8 T cells infiltrating tumors and maintain their cytotoxic function (Huff et al, JCI Insight 2023). Yet, the impact of cell-cell communication on effector T cell function arising from the immunosuppressive microenvironment of PDAC remains unknown. We previously developed a ligand-receptor inference method called Domino that infers inter- and intra-cellular signaling networks in single-cell RNA-seq data (Cherry et al, Nat Biomed Eng 2021). However, this technique is limited to data from single conditions. Understanding how this combination immunotherapy alters cellular signaling networks requires new computational tools that identify changes in regulatory networks between treatment conditions. We developed capabilities for assessment of differential signaling with Domino founded on a basis of differential signal receipt. Active receptors are identified by receptor expression in recipient cells that is correlated with downstream transcription factors in the intracellular signaling cascade. Identification of causative ligands for receptor activation is assessed by a probability of interaction for ligand-receptor interactions between cell types. This probability metric considers mean expression of ligand and receptor genes as well as the number of sender cells for comparison across treatment groups. Comparing communication networks in PancVAX alone to PancVAX with ICI, PancVAX with ICI enhances cell-cell adhesion and interferon gamma signaling in cytotoxic CD8 T cells towards exhausted CD8 T cells. Macrophages in PancVAX with ICI enhance expression of immune checkpoint ligands towards CD8 T cells, despite the presence of blocking ICI antibodies. Intracellular signaling downstream of PD-1 in exhausted T cells shows diminished numbers of transcription factors linked to PD-1 signaling. This software informs prioritization of inhibitory signals for modulation in future immunotherapy regimens and can be applied generally to other tumor subtypes and treatment contexts. Citation Format: Jacob T. Mitchell, Atul Deshpande, Amanda L. Huff, Chris Cherry, Sushma Nagaraj, Kavita Krishnan, Dmitrijs Lvovs, Jennifer H. Elisseeff, Elizabeth M. Jaffee, Neeha Zaidi, Elana J. Fertig. A differential ligand-receptor network inference method to identify alterations in communication between myeloid cells and CD8 T cells in response to PancVAX neo-epitope peptide vaccine, anti-CTLA-4 and anti-PD-1 antibodies in a murine model of pancreatic ductal adenocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6212.