Single-cell Protein Expression Research Articles

HER4 is a high-affinity dimerization partner for all EGFR/HER/ErbB family proteins.

Human epidermal growth factor receptors (HER)-also known as EGFR or ErbB receptors-are a subfamily of receptor tyrosine kinases (RTKs) that play crucial roles in cell growth, division, and differentiation. HER4 (ErbB4) is the least studied member of this family, partly because its expression is lower in later stages of development. Recent work has suggested that HER4 can play a role in metastasis by regulating cell migration and invasiveness; however, unlike EGFR and HER2, the precise role that HER4 plays in tumorigenesis is still unresolved. Early work on HER family proteins suggested that there are direct interactions between the four members, but to date, there has been no single study of all four receptors in the same cell line with the same biophysical method. Here, we quantitatively measure the degree of association between HER4 and the other HER family proteins in live cells with a time-resolved fluorescence technique called pulsed interleaved excitation fluorescence cross-correlation spectroscopy (PIE-FCCS). PIE-FCCS is sensitive to the oligomerization state of membrane proteins in live cells, while simultaneously measuring single-cell protein expression levels and diffusion coefficients. Our PIE-FCCS results demonstrate that HER4 interacts directly with all HER family members in the cell plasma membrane. The interaction between HER4 and other HER family members intensified in the presence of a HER4-specific ligand. Our work suggests that HER4 is a preferred dimerization partner for all HER family proteins, even in the absence of ligands.

Imputing Single-Cell Protein Abundance in Multiplex Tissue Imaging.

Multiplex tissue imaging are a collection of increasingly popular single-cell spatial proteomics and transcriptomics assays for characterizing biological tissues both compositionally and spatially. However, several technical issues limit the utility of multiplex tissue imaging, including the limited number of molecules (proteins and RNAs) that can be assayed, tissue loss, and protein probe failure. In this work, we demonstrate how machine learning methods can address these limitations by imputing protein abundance at the single-cell level using multiplex tissue imaging datasets from a breast cancer cohort. We first compared machine learning methods' strengths and weaknesses for imputing single-cell protein abundance. Machine learning methods used in this work include regularized linear regression, gradient-boosted regression trees, and deep learning autoencoders. We also incorporated cellular spatial information to improve imputation performance. Using machine learning, single-cell protein expression can be imputed with mean absolute error ranging between 0.05-0.3 on a [0,1] scale. Finally, we used imputed data to predict whether single cells were more likely to come from pre-treatment or post-treatment biopsies. Our results demonstrate (1) the feasibility of imputing single-cell abundance levels for many proteins using machine learning; (2) how including cellular spatial information can substantially enhance imputation results; and (3) the use of single-cell protein abundance levels in a use case to demonstrate biological relevance.

DIPG-51. DIFFUSE MIDLINE GLIOMAS HIJACK MYELOID CELL POPULATIONS TO INITIATE AN ANTI-INFLAMMATORY TUMOR IMMUNE MICROENVIRONMENT

Abstract BACKGROUND Diffuse midline gliomas (DMG) are universally fatal pediatric brain tumors, for which no effective treatment exists. These tumors are known for their “cold” non-inflammatory tumor-immune microenvironment (TIME), which is thought to be a major hurdle for immunotherapy efficacy. METHODS/RESULTS To further investigate the mechanisms that drive this cold environment, we allografted immunocompetent mice with Histone 3.3 Lysine 27 (H3.3K27M) and 3.1 Lysine 27 (H3.1K27M) mutant murine DMG tumors, created through intra uterine electroporation, and performed single-cell analyses on the tumors that developed. Using Cellular Indexing of Transcriptomes and Epitopes by Sequencing (CITEseq) we mapped the single-cell RNA and protein expression profiles of the TIME of these tumors and healthy pontine regions, with a dominant focus on the myeloid cell populations. Through differential expression and pathway enrichment analyses, we found that DMG-microglia populations upregulate anti-inflammatory pathways and, over time, are pushed out by infiltrating anti-inflammatory macrophages. We confirmed these findings in single-cell sequencing data from DMG patients, taken at diagnosis and at late stage disease. Next, we established microglia cultures in vitro, by differentiating monocytes from healthy donors into microglia. We discovered that treatment of these microglia with tumor-conditioned medium, taken from primary DMG cultures, diminished their pro-inflammatory potential after stimulation with lipopolysaccharides (LPS), as observed by cytokine release arrays that showed a strong downregulation of IL-6 and TNF-alpha release, amongst others. Subsequently, using a Seahorse metabolic analyzer, we found that glycoATP potential was diminished in DMG-primed microglia after LPS stimulation, further suggesting that DMG-excreted factors are responsible for blocking the pro-inflammatory potential of microglia. CONCLUSIONS Taken together, these findings increase our understanding on how DMG cells reprogram tumor-associated myeloid cells to generate an immune-cold TIME, which will aid the scientific community to find therapeutic solutions to reverse this phenomenon and improve the efficacy of immunotherapy in these devastating malignancies.

Single-Cell Patch-Clamp/Proteomics of Human Alzheimer's Disease iPSC-Derived Excitatory Neurons Versus Isogenic Wild-Type Controls Suggests Novel Causation and Therapeutic Targets.

Standard single-cell (sc) proteomics of disease states inferred from multicellular organs or organoids cannot currently be related to single-cell physiology. Here, a scPatch-Clamp/Proteomics platform is developed on single neurons generated from hiPSCs bearing an Alzheimer's disease (AD) genetic mutation and compares them to isogenic wild-type controls. This approach provides both current and voltage electrophysiological data plus detailed proteomics information on single-cells. With this new method, the authors are able to observe hyperelectrical activity in the AD hiPSC-neurons, similar to that observed in the human AD brain, and correlate it to ≈1400 proteins detected at the single neuron level. Using linear regression and mediation analyses to explore the relationship between the abundance of individual proteins and the neuron's mutational and electrophysiological status, this approach yields new information on therapeutic targets in excitatory neurons not attainable by traditional methods. This combined patch-proteomics technique creates a new proteogenetic-therapeutic strategy to correlate genotypic alterations to physiology with protein expression in single-cells.

Abstract 3704: Whole genome and targeted proteome analysis of single cancer cells

Abstract The simultaneous genome and proteome analysis of very rare single cells would enable addressing important biological and diagnostic questions particularly in cancer. Here, we introduce a novel genome and proteome (G&P) analytic workflow, a method allowing comprehensive analysis of single-cell DNA and targeted profiling of protein expression in single cells. As a proof of concept, we demonstrate the capability of the method by quantifying copy number of the ERBB2 gene alongside with protein expression levels of HER2 and phospho-HER2 in single cells. The G&P workflow utilizes antibodies conjugated with DNA oligonucleotides designed to enable their amplification alongside the single-cell genome facilitating parallel genome and proteome profiling by either quantitative PCR (qPCR) or sequencing. We utilized antibodies targeting cytokeratin 8/18/19, HER2 and phospho-HER (Tyr1248). Performance of each DNA-conjugated antibody was assessed by immunofluorescence staining in a variety of breast cancer cell lines (MDA-MB231, MDA-MB453, BT474 and MCF7) and sample types. Stained cells were isolated by manual micromanipulation and subjected to whole genome amplification (WGA). Protein expression levels, derived from quantification of the amplified DNA tags, were measured by qPCR. Shallow whole genome sequencing was employed to facilitate genome-wide profiling of copy number variations (CNVs), including the ERBB2 locus. We successfully developed and tested a G&P workflow allowing parallel genome-wide CNV profiling and quantification of three target proteins. G&P measurement yielded concordant results to immunofluorescence quantification for all tested markers. Furthermore, the workflow allowed direct correlation of copy numbers of the ERBB2 gene with HER2 expression in single cancer cells. Notably, the method enabled measurements of cellular and molecular responses at the single-cell level to lapatinib treatment, i.e. monitoring of the HER2 pathway activation level. Our novel multiplex G&P technique is capable of generating multiparametric measurements linking DNA copy numbers and expression levels of selected genes of interest. The method is optimized for analysis of single cells from rare cell populations and is readily translatable to study circulating tumor cells (CTCs) and disseminated cancer cells (DCCs), enabling for the first time pathway analysis via phosphoprotein analyses of these cells. Citation Format: Atieh Alyasin, Kathrin Weidele, Catherine Botteron, Christoph A. Klein, Zbigniew T. Czyż. Whole genome and targeted proteome analysis of single cancer 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 3704.

Multiplex fluorescence detection of single-cell droplet microfluidics and its application in quantifying protein expression levels.

This study presented a platform of multiplex fluorescence detection of single-cell droplet microfluidics with demonstrative applications in quantifying protein expression levels. The platform of multiplex fluorescence detection mainly included optical paths adopted from conventional microscopy enabling the generation of three optical spots from three laser sources for multiple fluorescence excitation and capture of multiple fluorescence signals by four photomultiplier tubes. As to platform characterization, microscopic images of three optical spots were obtained where clear Gaussian distributions of intensities without skewness confirmed the functionality of the scanning lens, while the controllable distances among three optical spots validated the functionality of fiber collimators and the reflector lens. As to demonstration, this platform was used to quantify single-cell protein expression within droplets where four-type protein expression of α-tubulin, Ras, c-Myc, and β-tubulin of CAL 27 (Ncell = 1921) vs WSU-HN6 (Ncell = 1881) were quantitatively estimated, which were (2.85 ± 0.72) × 105 vs (4.83 ± 1.58) × 105, (3.69 ± 1.41) × 104 vs (5.07 ± 2.13) × 104, (5.90 ± 1.45) × 104 vs (9.57 ± 2.85) × 104, and (3.84 ± 1.28) × 105 vs (3.30 ± 1.10) × 105, respectively. Neural pattern recognition was utilized for the classification of cell types, achieving successful rates of 69.0% (α-tubulin), 75.4% (Ras), 89.1% (c-Myc), 65.8% (β-tubulin), and 99.1% in combination, validating the capability of this platform of multiplex fluorescence detection to quantify various types of single-cell proteins, which could provide comprehensive evaluations on cell status.

High Dimensional Clustering and Identification of Immune Cell Hybrid Populations in B-ALL across the Disease Continuum

Introduction: Nearly half of adults with acute lymphoblastic leukemia (ALL) will experience relapse of their disease, and those with relapsed ALL carry a worse prognosis than those who remain in remission. The influence of the tumor immune microenvironment (TME) on relapse and survival outcomes in ALL is not well understood. Mass cytometry (CyTOF) enables immune profiling based on single cell protein expression and can identify phenotypic diversity of cells in the TME that cannot reliably be detected by standard flow cytometry. Recently, T-cell/monocyte hybrid cells have been discovered in patients with uveal melanoma and latent tuberculosis infections. However, cells expressing CD19, a surface antigen expressed on B-ALL cells, and markers of other cells in the TME including T cells and tumor-associated macrophages (TAMs), have not been described in ALL or other hematologic malignancies. The presence of CD3/CD163 hybrid cells in the TME may have predictive and prognostic implications for adults with ALL, as TAMs play a role in suppressing antitumor immunity which can lead to T-cell exhaustion. In this study, we identify the presence of multilineage hybrid cells in peripheral blood (PB) and bone marrow (BM) samples of patients with B-ALL alongside accompanying evolution of the TME at each disease stage. Methods: PB (N=16) and BM aspirate (N=16) isolates were collected from patients with B-ALL at diagnosis, relapse, or remission, as well as from healthy donors (HDs, N=7) for comparison of cellular profiles. A 28-marker CYTOF panel was designed to evaluate cell surface and intracellular markers of interest in the myeloid compartment. Antibodies were purchased pre-conjugated or conjugated in house based on manufacturer guidelines (Standard BioTools). Samples were barcoded with palladium isotopes prior to staining to reduce the impact of batch effect, and were then analyzed on a first generation CyTOF system (Standard BioTools). Following background attenuation and singlet gating in FlowJo (BD), the CATALYST R package enabled deconvolution and visualization of single cell data, including dimensional reduction techniques (tSNE) and clustering of all cell populations. An iterative approach was taken to determine a stable configuration of clusters across sample populations to protect from over-fragmentation of data, further informed by K-means clustering using the elbow method. Heatmaps were then generated based on median scaled expression of markers contained in each cluster. Results: The CyTOF panel identified several different myeloid cell populations in newly diagnosed, relapsed, and remission samples in both PB and BM compared to HDs. Most interestingly, quantifiable levels of a cluster of cells co-expressing CD19, CD3 and CD163 as well as a second cluster of cells expressing CD19 and CD163 were detected in newly diagnosed and relapsed patients, but not in HDs. A single patient with the largest proportion of these cell populations in their PB at the time of relapse presented with an aggressive leukemia variant and relapsed multiple times prior to succumbing to illness. On investigation of BM aspirates, the triple positive hybrid population was found in a similar distribution amongst the newly diagnosed and relapsed patients, as well as affecting one patient who achieved remission. This patient also had a highly aggressive leukemia that was refractory to chemotherapy and immunotherapy, and only achieved this remission following CAR-T cell therapy. She subsequently underwent stem cell transplant and is in remission. Conclusions: Our findings add to the literature highlighting the utility of high parameter single cell capture techniques in detecting rare cell populations in the leukemic TME. Further, we identified two unique hybrid cell populations that are reliably detected across different disease stages in patients with B-ALL. Importantly, these populations persisted following stringent gating protocols to rule out doublets and debris. Although our sample size was too small to demonstrate a prognostic impact of these hybrid cell populations in patients with B-ALL, the identification of these populations alone is a significant finding. Further investigation and patient sample accrual are warranted to better understand the role of these cell types in tumor biology, which may pave the way towards new prognostic models that can leverage well-established immune monitoring techniques.

Distinct Minor Neoplastic Follicular T-Cell Subsets Dictate Lymphoma Fates through Functional Specialization

Background: The presence and role of follicular T-cell populations other than T follicular helper (Tfh) cells, such as T follicular regulatory (Tfr) and cytotoxic (Tfc) cells, are gaining increasing attention in certain pathological states. However, the ecosystem of follicular T cells in the tumor microenvironment (TME) has not been fully elucidated. In particular, the significance of minor follicular T-cell subsets in the neoplastic follicular environment remains elusive. Here, we aimed to reveal the landscape of follicular T-cell alterations in various cancers, with a particular emphasis on the follicular lymphoma (FL) TME. Methods: We analyzed single-cell RNA/TCR sequencing data of >500,000 human T cells from FL (obtained from four cohorts) and 25 other cancer types, as well as homeostatic and reactive lymph nodes (LNs), to construct a comprehensive single-T-cell atlas. We investigated differentially expressed genes, RNA velocity, and TCR clonality using this atlas. To determine the functions of neoplastic follicular regulatory (Tnfr) and cytotoxic (Tnfc) T cells, we performed in vitro cytokineproduction and co-culture assays, in combination with cell activation/suppression, cell division, and apoptosis assays, using human FL samples. With the PhenoCycler-Fusion system, we conducted multiplex digital spatial profiling (DSP) of 169 FL samples from two independent cohorts (now being extended to 242 FL samples from three cohorts) for >25 antibodies. We also performed single-cell spatial and protein expression profiling and prognostic analysis. Results: In FL, distinct minor neoplastic follicular T-cell subsets-Tnfr and CD4 (Tnfc4) and CD8 (Tnfc8) Tnfc cells-increased relative to homeostatic LNs. The TCR repertoire analysis revealed that Tnfr cells shared clonotypes with conventional effector regulatory T (Trg) and Tfh cells, whereas Tnfc4 and Tnfc8 cells shared clonotypes with Tfh cells and effector and exhausted (Tcex) cytotoxic CD8 T cells, respectively. In line with these findings, the RNA velocity survey suggested that Tnfr, Tnfc4, and Tnfc8 cells originated from Trg, Tfh, and naïve-like CD8 T cells, respectively. Tnfr and Tnfc cells expressed higher levels of effector genes, including those involved in cytokine release, chemokine response, migration, and PD-1 signaling, than their reactive LN counterparts. The pan-cancer survey revealed that Tfr and CD4 Tfc cells were exclusive to FL, whereas the prevalence and gene expression profiles of CD8 Tfc cells varied across cancers. Tnfr cells were marked by abundant expression of IL10 and IL21, whereas Tnfc cells displayed a unique phenotype, as they concomitantly expressed markers of effector Tfh (e.g., CXCL13, CXCR5, and PDCD1), naïve/stem (e.g., CCR7 and TCF7), central memory (e.g., CD27, CD28, and SELL), and tissue-resident memory (e.g., ITGAE) cells. Hierarchical clustering demonstrated that Tnfc8 cells had transcriptional profiles similar to those of melanoma TCF1 +PD-1 +CD8 + stem-like T cells. DSP of FL detected Tnfr and Tnfc cells frequently localized within and around neoplastic follicles, forming a cellular neighborhood that allowed them to interact closely. Tnfr cells were distributed predominantly near Tfh cells. The functional co-culture assays demonstrated that Tnfr cells suppressed Tfh-cell activation and division, thereby inhibiting Tfh-mediated malignant B-cell activation and survival. Tnfc8 cells showed a higher cell division capability than that of Tcex cells, suggesting that Tnfc8 cells function as a pool of CD8 T cells in neoplastic follicles. The prognostic analysis revealed that Tnfr and Tnfc cell proportions correlated with early disease relapse (i.e., POD24) and predicted a significantly longer time-to-relapse ( P <0.05 for Tnfr and <0.001 for Tnfc cells) in FL. In the multivariate analysis, the prognostic impact of these two cell subsets was independent of the FLIPI. The prognostic analysis findings were confirmed using a validation cohort. Conclusions: Our multi-omics approach identified the expansion of minor neoplastic follicular T-cell subsets that carry unique transcriptional and functional profiles and robust prognostic impacts. These findings deepen our understanding of the biological and immunological roles of non-Tfh follicular T cells in the lymphoma TME and highlights their clinical potential for patient risk stratification and future therapeutic interventions.

Characterization and decontamination of background noise in droplet-based single-cell protein expression data with DecontPro.

Assays such as CITE-seq can measure the abundance of cell surface proteins on individual cells using antibody derived tags (ADTs). However, many ADTs have high levels of background noise that can obfuscate down-stream analyses. In an exploratory analysis of PBMC datasets, we find that some droplets that were originally called 'empty' due to low levels of RNA contained high levels of ADTs and likely corresponded to neutrophils. We identified a novel type of artifact in the empty droplets called a 'spongelet' which has medium levels of ADT expression and is distinct from ambient noise. ADT expression levels in the spongelets correlate to ADT expression levels in the background peak of true cells in several datasets suggesting that they can contribute to background noise along with ambient ADTs. We then developed DecontPro, a novel Bayesian hierarchical model that can decontaminate ADT data by estimating and removing contamination from these sources. DecontPro outperforms other decontamination tools in removing aberrantly expressed ADTs while retaining native ADTs and in improving clustering specificity. Overall, these results suggest that identification of empty drops should be performed separately for RNA and ADT data and that DecontPro can be incorporated into CITE-seq workflows to improve the quality of downstream analyses.

IMMU-34. SYSTEMIC IMMUNE SUPPRESSION IN GLIOBLASTOMA IS TARGETED BY CONCURRENT VEGF AND PD-1 INHIBITION IN A DOSE-DEPENDENT MANNER: TRANSCRIPTOMIC FINDINGS THROUGH CITE-SEQ FROM A RANDOMIZED CONTROLLED TRIAL

Abstract INTRODUCTION Anti-PD1 blockade has had poor efficacy in recurrent glioblastoma (rGBM) across three RCTs. VEGF, a proangiogenic factor that is upregulated in rGBM, contributes to tumor-associated immunosuppression and is widely targeted through bevacizumab. Preclinical data indicate a potential dose-dependent effect of anti-VEGF therapy on immunomodulation. Hence, a combination of anti-PD-1 and anti-VEGF agents is a promising approach for rGBM, for which an RCT was conducted. This work sought to utilize a multimodal approach capturing single-cell gene and protein expression by ‘cellular indexing of transcriptomes and epitopes by sequencing’ (CITE-Seq) to interrogate the systemic immunological changes in response to concurrent VEG and PD-1 inhibition. METHODS Patients with rGBM were administered nivolumab (240mg IV q2weeks) plus either standard-dose bevacizumab (10mg/kg) or low-dose bevacizumab (3mg/kg) IV q2weeks, with 1:1 randomization. Stratification factors were age, performance status, extent of resection, and MGMT methylation status. Primary endpoint was overall survival (OS) at 12 months. CITE-Seq was performed on 16 patients across trial arms, both pre- and on-therapy, yielding total 32 immunological profiles. For control, CITE-seq was done on three patients receiving anti-VEGF therapy alone. RESULTS 90 patients were enrolled with median age of 60.6 years and 45 patients per arm. Both arms had similar OS at 12 months (P=0.14). Significantly higher OS was found in posthoc analysis for patients aged ≥60 years with standard dose. Distinct immune cell populations were identified, and differential changes in myeloid-derived suppressor cells (MDSCs) found across trial arms on-therapy. Network medicine approach identified VEGF as target to reduce MDSCs, with VEGF primarily expressed by responders’ MDSCs. Differential gene expression analysis identified increased pro-inflammatory signatures in standard-dose arm. CONCLUSIONS Standard dose bevacizumab was associated with increased inflammatory response and reduction of immunosuppressive MDSC. These mechanistic insights potentially explain survival benefit seen with concurrent VEGF and PD-1 inhibition in elderly rGBM patients.

Yolk sac cell atlas reveals multiorgan functions during human early development.

The extraembryonic yolk sac (YS) ensures delivery of nutritional support and oxygen to the developing embryo but remains ill-defined in humans. We therefore assembled a comprehensive multiomic reference of the human YS from 3 to 8 postconception weeks by integrating single-cell protein and gene expression data. Beyond its recognized role as a site of hematopoiesis, we highlight roles in metabolism, coagulation, vascular development, and hematopoietic regulation. We reconstructed the emergence and decline of YS hematopoietic stem and progenitor cells from hemogenic endothelium and revealed a YS-specific accelerated route to macrophage production that seeds developing organs. The multiorgan functions of the YS are superseded as intraembryonic organs develop, effecting a multifaceted relay of vital functions as pregnancy proceeds.

Of tenants and nomads: The faces of memory T cells

Memory Tcells comprise circulating and tissue-resident subsets. In this issue of Immunity, Evrard etal. generate an imputed high-dimensional, single-cell protein expression atlas of memory CD8+ Tcells, providing insights into stable differentiation markers and organ-specific expression patterns.

Single-cell protein expression profiling resolves circulating and resident memory T cell diversity across tissues and infection contexts

Memory CD8+ Tcells can be broadly divided into circulating (TCIRCM) and tissue-resident memory T (TRM) populations. Despite well-defined migratory and transcriptional differences, the phenotypic and functional delineation of TCIRCM and TRM cells, particularly across tissues, remains elusive. Here, we utilized an antibody screening platform and machine learning prediction pipeline (InfinityFlow) to profile >200 proteins in TCIRCM and TRM cells in solid organs and barrier locations. High-dimensional analyses revealed unappreciated heterogeneity within TCIRCM and TRM cell lineages across nine different organs after either local or systemic murine infection models. Additionally, we demonstrated the relative effectiveness of strategies allowing for the selective ablation of TCIRCM or TRM populations across organs and identified CD55, KLRG1, CXCR6, and CD38 as stable markers for characterizing memory Tcell function during inflammation. Together, these data and analytical framework provide an in-depth resource for memory Tcell classification in both steady-state and inflammatory conditions.

Machine learning modeling and prognostic value analysis of invasion-related genes in cutaneous melanoma

In this study, we aimed to develop an invasion-related risk signature and prognostic model for personalized treatment and prognosis prediction in skin cutaneous melanoma (SKCM), as invasion plays a crucial role in this disease. We identified 124 differentially expressed invasion-associated genes (DE-IAGs) and selected 20 prognostic genes (TTYH3, NME1, ORC1, PLK1, MYO10, SPINT1, NUPR1, SERPINE2, HLA-DQB2, METTL7B, TIMP1, NOX4, DBI, ARL15, APOBEC3G, ARRB2, DRAM1, RNF213, C14orf28, and CPEB3) using Cox and LASSO regression to establish a risk score. Gene expression was validated through single-cell sequencing, protein expression, and transcriptome analysis. Negative correlations were discovered between risk score, immune score, and stromal score using ESTIMATE and CIBERSORT algorithms. High- and low-risk groups exhibited significant differences in immune cell infiltration and checkpoint molecule expression. The 20 prognostic genes effectively differentiated between SKCM and normal samples (AUCs >0.7). We identified 234 drugs targeting 6 genes from the DGIdb database. Our study provides potential biomarkers and a risk signature for personalized treatment and prognosis prediction in SKCM patients. We developed a nomogram and machine-learning prognostic model to predict 1-, 3-, and 5-year overall survival (OS) using risk signature and clinical factors. The best model, Extra Trees Classifier (AUC = 0.88), was derived from pycaret's comparison of 15 classifiers. The pipeline and app are accessible at https://github.com/EnyuY/IAGs-in-SKCM.

Abstract 3116: Spatial single-cell atlas of stage III colorectal cancer

Abstract Background: Spatial heterogeneity and multicellular interactions within the tumor microenvironment drive tumor progression and response to therapy, yet, tumor molecular profiling often requires dissociation of cells, losing the spatial context. Spatial proteomic technologies enable capturing single-cell information with the spatial context allowing us to link cancer-causing mutations to outcomes in cell signaling, responses and survival that will aid in diagnosis, prognostication, and treatment of cancer. We aim to capture single cells in intact tissue sections of stage III colorectal tumors from 52 patients to spatially characterize cell types, interactions, and organization of the tumor microenvironment in relation to genetic alterations and clinical outcomes. Methods: Using Hyperion Imaging Mass Cytometry (IMC), we profiled 16 protein markers across 170 tissue regions, capturing molecular signatures of tissue architecture, tumor cells, and immune cells. We developed an analysis pipeline to quantify single-cell protein expression and identify cell phenotypes. We preserved spatial information to characterize the interactions between neighboring cells and identify multicellular communities. We also generated whole exome sequencing data, RNAseq data and histopathological images from sections of the same tissue blocks. We compared cellular and spatial features between tumors harboring prognostic genomic evens such as TP53 deletion and MSI and between patients stratified by clinical outcomes such as survival and recurrence. Results: We spatially profiled over 800,000 cells to identify 10 tumor, immune and stromal cell phenotypes, and validated identified cell types against matched histopathology and CIBERSORTx deconvoluted RNAseq. To quantify cell-cell interactions, we measured the distance between neighboring cells and performed neighborhood enrichment analysis. We characterized the multicellular organization of the tumor microenvironment to identify 7 consistent cell neighborhoods of up to 10 adjacent cells. We quantified p53 expression in individual tumor cells and found that TP53 deletion (all heterozygous) had no significant association with p53 expression or p53+ tumor cell abundance, suggesting spatial proteomics reveals additional information absent in traditional sequencing. We find that MSI is exclusive of TP53 deletion and that MSI positive tumors had few p53+ tumor cells, increased CD8+ T cell composition and B cells that were further away from p53+ tumor cells compared to MSI negative tumors. Conclusions: Using Hyperion IMC to acquire a spatial single-cell atlas of stage III colorectal cancer enabled us to characterize the tumor microenvironment by defining cell types, quantifying cell interactions, and identifying multicellular communities. These features enable the linkage of prognostic genetic alternations and clinical outcomes with changes in the tumor microenvironment. Citation Format: Andrew Su, Minh Tran, HoJoon Lee, Anuja Sathe, Xiangqi Bai, Richard Cruz, Lance Pflieger, Quan Nguyen, Hanlee P. Ji, Terence Rhodes. Spatial single-cell atlas of stage III colorectal cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3116.

Single-Cell Profiling of Tumor Immune Microenvironment Reveals Immune Irresponsiveness in Gastric Signet-Ring Cell Carcinoma

Gastric cancer (GC) is a major cancer type characterized by high heterogeneity in both tumor cells and the tumor immune microenvironment (TIME). One intractable GC subtype is gastric signet-ring cell carcinoma (GSRCC), which is associated with poor prognosis. However, it remains unclear what the GSRCC TIME characteristics are and how these characteristics may contribute to clinical outcomes. We enrolled 32 patients with advanced GC of diverse subtypes and profiled their TIME using an immune-targeted single-cell profiling strategy, including (i) immune-targeted single-cell RNA sequencing (scRNA-seq, n = 20 patients) and (ii) protein expression profiling by a targeted antibody panel for mass cytometry (CyTOF, n = 12 patients). We also generated matched V(D)J sequencing of T and B cells along CD45+ immunocytes. We find that compared to non-GSRCC, the GSRCC TIME appears to be quiescent where both CD4+ and CD8+ T cells are difficult to be mobilized, which further impairs the proper functions of B cells. CXCL13, mainly produced by Tfh, Th17, and exhausted CD8+ T cells, is a central coordinator of this transformation. We show that CXCL13 expression can predict the response to immune checkpoint blockade in GC patients, which may be related to its effects on tertiary lymphoid structures. Our study provides a comprehensive molecular portrait of immune cell compositions and cell states in advanced GC patients, highlighting adaptive immune irresponsiveness in GSRCC and a mediator role of CXCL13 in TIME. Our targeted single-cell transcriptomic and proteomic profiling represents a powerful approach for TIME-oriented translational research.

Tet2- and Tet3- Mediated Cytosine Hydroxymethylation in Six2 Progenitor Cells in Mice Is Critical for Nephron Progenitor Differentiation and Nephron Endowment.

Epigenetic changes have been proposed to mediate nephron endowment during development, a critical determinant of future renal disease development. Hydroxymethyl cytosine, an epigenetic modification important for gene regulation, is abundant in the human kidney, but its physiologic role and the role of DNA demethylase enzymes encoded by the Tet1 , Tet2 , or Tet3 , which mediate cytosine hydroxymethylation, are unclear. By genetically deleting Tet1 , Tet2 , or Tet3 in nephron progenitors in mice, the authors showed that combined Tet2 and Tet3 loss in nephron progenitors cause defective kidney development, leading to kidney failure and perinatal death. Tet2 and Tet3 deletion also caused an alteration in demethylation and expression of genes critical for nephron formation. These findings establish that Tet2- and Tet3 -mediated cytosine hydroxymethylation in nephron progenitors plays a critical role in nephron endowment. Nephron endowment is a key determinant of hypertension and renal disease in later life. Epigenetic changes have been proposed to mediate fetal programming and nephron number. DNA cytosine methylation, which plays a critical role in gene regulation, is affected by proteins encoded by the ten-eleven translocation (TET) DNA demethylase gene family ( Tet1 , Tet2 , and Tet3 ), but the roles of TET proteins in kidney development and nephron endowment have not been characterized . To study whether epigenetic changes-specifically, active DNA hydroxymethylation mediated by Tet1 , Tet2 , and Tet3- are necessary for nephron progenitor differentiation and nephron endowment, we generated mice with deletion of Tet1 , Tet2 , or Tet3 in Six2-positive nephron progenitors cells (NPCs). We then performed unbiased omics profiling, including whole-genome bisulfite sequencing on isolated Six2-positive NPCs and single-cell RNA sequencing on kidneys from newborn mice. We did not observe changes in kidney development or function in mice with NPC-specific deletion of Tet1 , Tet2 , Tet3 or Tet1 / Tet2 , or Tet1 / Tet3 . On the other hand, mice with combined Tet2 and Tet3 loss in Six2-positive NPCs failed to form nephrons, leading to kidney failure and perinatal death. Tet2 and Tet3 loss in Six2 -positive NPCs resulted in defective mesenchymal to epithelial transition and renal vesicle differentiation. Whole-genome bisulfite sequencing, single-cell RNA sequencing, and gene and protein expression analysis identified a defect in expression in multiple genes, including the WNT- β -catenin signaling pathway, due to a failure in demethylation of these loci in the absence of Tet2 and Tet3 . These findings suggest that Tet2- and Tet3 -mediated active cytosine hydroxymethylation in NPCs play a key role in kidney development and nephron endowment.

Illustrating Heterogeneity in AML By Leveraging Multimodal Analysis of DNA, Surface Proteins, and Intracellular Proteins in Single Cells

Introduction: Recent advancements in precision medicine, although highly promising, present a major technical challenge to researchers due to disease heterogeneity. The emergence of single-cell technologies has greatly refined the resolution in which sample diversity can be investigated, enhancing the efficiency of selecting appropriate molecular targets. Additionally, applying multiomics analysis on single cells would further improve the understanding of cell-to-cell heterogeneity by providing unique insights into cellular and genetic composition. Methods: Using a two-step droplet microfluidic technology, the Mission Bio TapestriⓇ platform enables high-throughput targeted DNA sequencing in single cells to obtain single-nucleotide variation (SNV) and copy number variation (CNV) information. By leveraging this technology, a novel workflow was developed to detect protein expression in addition to DNA genotypes in the same single cells. In this approach, cells are labeled with oligonucleotide-conjugated antibodies prior to single-cell encapsulation and DNA target amplification on the Tapestri system. Sequencing libraries are then prepared from both antibody oligonucleotides and the amplified DNA sequences, followed by identification of single-cell DNA genotypes and protein signatures from the sequencing readout. The number of protein targets can range from a few targets to over 40, which is beyond the limit for a single flow cytometry run. Results: This workflow has been successfully demonstrated on cell lines as well as clinical samples. In an acute myeloid leukemia (AML) sample, combined single-cell SNV, CNV, and protein expression data illustrated the heterogeneity within the sample. The data were able to clearly distinguish CD3+ T cells and CD19+ B cells without pathogenic SNVs and CNVs from CD34hiCD11blo and CD34loCD11bhi subpopulations that did contain the same pathogenic SNVs and CNVs. Conclusion: We further developed this workflow to include simultaneous detection of intracellular proteins and surface proteins along with targeted sequencing in single cells. The inclusion of intracellular protein detection enables measurement of pivotal proteins in cancer mechanisms. We believe that this novel multiomic technology will facilitate new discoveries inthe complex relationship between genotype and phenotype, enabling a better understanding of disease biology, and subsequently improve the design of diagnostics and therapies.

AML-160 Multimodal Analysis of DNA, Surface Proteins, and Intracellular Proteins in Single Cells

Recent advancements in precision medicine, although highly promising, present a major technical challenge to researchers due to disease heterogeneity. The emergence of single-cell technologies has greatly refined the resolution in which sample diversity can be investigated, enhancing the efficiency of selecting appropriate molecular targets. Additionally, applying multiomics analysis on single cells would further improve the understanding of cell-to-cell heterogeneity by providing unique insights into cellular and genetic composition. Using a two-step droplet microfluidic technology, the Mission Bio Tapestri® platform enables high-throughput targeted DNA sequencing in single cells to obtain single-nucleotide variation (SNV) and copy number variation (CNV) information. By leveraging this technology, a novel workflow was developed to detect protein expression in addition to DNA genotypes in the same single cells. In this approach, cells are labeled with oligonucleotide-conjugated antibodies prior to single-cell encapsulation and DNA target amplification on the Tapestri system. Sequencing libraries are then prepared from both antibody oligonucleotides and the amplified DNA sequences, followed by identification of single-cell DNA genotypes and protein signatures from the sequencing readout. The number of protein targets can range from a few targets to over 40, which is beyond the limit for a single flow cytometry run. This workflow has been successfully demonstrated on cell lines as well as clinical samples. In an acute myeloid leukemia (AML) sample, combined single-cell SNV, CNV, and protein expression data illustrated the heterogeneity within the sample. The data were able to clearly distinguish CD3+ T cells and CD19+ B cells without pathogenic SNVs and CNVs from CD34hiCD11blo and CD34loCD11bhi subpopulations that did contain the same pathogenic SNVs and CNVs. We further developed this workflow to include simultaneous detection of intracellular proteins and surface proteins along with targeted sequencing in single cells. The inclusion of intracellular protein detection enables measurement of pivotal proteins in cancer mechanisms. We believe that this novel multiomic technology will facilitate new discoveries in the complex relationship between genotype and phenotype, enabling a better understanding of disease biology, and subsequently improve the design of diagnostics and therapies.

A Single-Cell Atlas of Tumor-Infiltrating Immune Cells in Pancreatic Ductal Adenocarcinoma.

Pancreatic ductal adenocarcinoma (PDAC) is one of the most lethal malignancies with limited treatment options. To guide the design of more effective immunotherapy strategies, mass cytometry was employed to characterize the cellular composition of the PDAC-infiltrating immune cells. The expression of 33 protein markers was examined at the single-cell level in more than two million immune cells from four types of clinical samples, including PDAC tumors, normal pancreatic tissues, chronic pancreatitis tissues, and peripheral blood. Based on the analyses, we identified 23 distinct T-cell phenotypes, with some cell clusters exhibiting aberrant frequencies in the tumors. Programmed cell death protein 1 (PD-1) was extensively expressed in CD4+ and CD8+ T cells and coexpressed with both stimulatory and inhibitory immune markers. In addition, we observed elevated levels of functional markers, such as CD137L and CD69, in PDAC-infiltrating immune cells. Moreover, the combination of PD-1 and CD8 was used to stratify PDAC tumors from The Cancer Genome Atlas database into three immune subtypes, with S1 (PD-1+CD8+) exhibiting the best prognosis. Further analysis suggested distinct molecular mechanisms for immune exclusion in different subtypes. Taken together, the single-cell protein expression data depicted a detailed cell atlas of the PDAC-infiltrating immune cells and revealed clinically relevant information regarding useful cell phenotypes and targets for immunotherapy development.