High-dimensional Profiling Research Articles

Abstract A105: PARP inhibition modulates the infiltration, phenotype, and function of tumor-associated macrophages (TAMs) in BRCA-associated breast cancer and can be augmented by harnessing the antitumor potential of TAMs

Abstract Patients with BRCA-associated triple-negative breast cancer (TNBC) have few effective treatment options. PARP inhibitors are promising, and we recently showed they induce an influx of white blood cells, including CD8+ T cells and macrophages into the tumor. The influx of CD8+ cells, mediated by activation of the STING pathway in tumor cells, contributes substantially to efficacy of PARP inhibition in mice. Strikingly, in these studies the greatest infiltration of immune cells into the tumor was macrophages. Given that objective responses to PARP inhibition have been observed in clinical trials but the benefits are transitory, we hypothesized that this was due to a suppressive tumor microenvironment, driven by tumor macrophages. To better understand the molecular basis of resistance to PARP inhibitors, we used high-dimensional single-cell immune profiling on human TNBC. We observed a ≥10-fold increase in TAMs in BRCA-associated TNBC compared to BRCA-wild-type TNBC. Using a preclinical model of BRCA1-deficient triple-negative breast cancer, we found that PARP inhibitors not only further increased TAM abundance but also induced functional and phenotypic changes associated with STING pathway activation, antigen presentation, and chemokine and cytokine signaling. PARP inhibitors increased the frequency of TAMs expressing costimulatory molecules CD80 and CD86 as well as the activation and maturation marker CD40, which are indicative of an antitumor phenotype. We also identified a novel negative feedback mechanism that limits the functionality of the anti-tumor TAMs and is consistent with induction of an immune-suppressive macrophage population. Utilizing transcriptomic, proteomic, and metabolic profiling of ex vivo cultured human myeloid cells, we identified multiple biologic processes associated with PARP inhibition, showing that these drugs directly affect macrophage states and phenotypes. Remarkably, in the preclinical BRCA1-deficient TNBC model, the novel combination of PARP inhibition with macrophage modulation significantly extended remissions obtained with PARP inhibitor therapy only, and this advantage persisted when treatment was discontinued, suggestive of a durable reprogramming of the tumor microenvironment. Moreover, CD8+ cells were required for the extension of PARP inhibitor-induced remissions, suggesting that targeting macrophages lifted the constraints imposed by protumor macrophages on CD8+ T cell-mediated tumor cell killing. We identify mechanisms related to macrophage and T-cell activation that increase PFS and provide evidence that TAMs may serve as targets for new therapeutic interventions designed to overcome PARP inhibitor resistance in BRCA-associated TNBC. Citation Format: Anita K. Mehta, Emily M. Cheney, Jessica A. Castrillon, Jia-Ren Lin, Mateus de Oliveira Taveira, Christina A. Hartl, Nathan T. Johnson, William M. Oldham, Marian Kalocsay, Sarah A. Boswell, Olmo Sonzogni, Constantia Pantelidou, Brett P. Gross, Shawn Johnson, Deborah A. Dillon, Sandro Santagata, Judy E. Garber, Nadine Tung, Elizabeth A. Mittendorf, Gerburg M. Wulf, Geoffrey I. Shapiro, Peter K. Sorger, Jennifer L. Guerriero. PARP inhibition modulates the infiltration, phenotype, and function of tumor-associated macrophages (TAMs) in BRCA-associated breast cancer and can be augmented by harnessing the antitumor potential of TAMs [abstract]. In: Proceedings of the AACR Special Conference on Tumor Immunology and Immunotherapy; 2019 Nov 17-20; Boston, MA. Philadelphia (PA): AACR; Cancer Immunol Res 2020;8(3 Suppl):Abstract nr A105.

Abstract A59: Inhibition of Glycogen Synthase Kinase 3 (GSK-3) synergizes with anti-PD-1 by potentiating tumor-infiltrated CD8+ T-cells

Abstract Anti-programmed cell death-1 (anti-PD-1) blockade in tumor therapy results in an expansion of CD8+ T cells and can restore the function of exhausted CD8+ T cells. We identified GSK-3 as a key upstream kinase of regulating PD-1 expression. Inhibition of GSK-3 in CD8+ T cells increased Tbx21 (Tbet), which suppressed Pdcd1 (PD-1) transcription (Taylor et al., Immunity 2016). In controlling spontaneous pulmonary metastasis of B16 melanoma and EL-4 lymphoma cells, GSK-3 blockade is as effective as anti-PD-1 blocking antibody (Taylor et al., Cancer Research 2018). Given the clinical importance in devising strategies to overcome tumor resistance to immune checkpoint blockade, we assessed whether the combined use of GSK-3 small-molecule inhibitor (SMI) and anti-PD-1 can overcome resistance to anti-PD-1 monotherapy. Indeed, GSK-3 SMI and anti-PD-1 synergized in therapy against solid B16 tumors that are resistant to monotherapy, leading to delayed tumor growth. Further, the synergy of GSK-3 SMI and anti-PD-1 therapy has been successfully confirmed by using adoptive transfer of OT1 cells into B16-OVA tumor-bearing mice. High-dimensional profiling of tumor-infiltrated lymphocytes (TILs) using mass cytometry (CyTOF) showed that the combination therapy gave an increase in cellularity of CD8+ effector memory T cells, which correlated to better tumor control. Transcriptome analysis of these CD8+ TILs highlighted the increased cytotoxicity as shown by significant changes of granzymes. Our findings outline a next-generation approach of potentiating PD-1 antibody therapy with SMIs of GSK-3 that synergizes as therapy by boosting the cellularity and cytotoxicity of tumor infiltrated CD8+ T cells. Citation Format: Janna Krueger, Alison Taylor, Alexandra Kazanova, Saloua Jeidane, Vinicius Motta, Ian Watson, Christopher E. Rudd. Inhibition of Glycogen Synthase Kinase 3 (GSK-3) synergizes with anti-PD-1 by potentiating tumor-infiltrated CD8+ T-cells [abstract]. In: Proceedings of the AACR Special Conference on Tumor Immunology and Immunotherapy; 2019 Nov 17-20; Boston, MA. Philadelphia (PA): AACR; Cancer Immunol Res 2020;8(3 Suppl):Abstract nr A59.

High-dimensional single cell-based immune profiling of the tumor immune microenvironment in prostate cancer.

376 Background: Genetic lesions that drive prostate cancer (PCa) development are able to modify the immune response and tumor infiltrating immune subsets, resulting in tumor progression. We investigated the profile of the immune microenvironment in PCa by high dimensional single cell analysis. Methods: We conducted an immune profiling study based on integrated RNA single cell sequencing and multiparametric flow cytometry in order to dissect the immune landscape of PCa. CD45+ immune cells infiltrating tumoral and adjacent non tumoral tissues were isolated from patients with PCa who underwent software assisted fusion biopsy, based on MRI, and/or radical prostatectomy, and analyzed by single cell sequencing. The primary endpoint was to evaluate the effectiveness of single cell RNA sequencing on CD45+ cell sorted from tumoral and adjacent non-tumoral tissues. Secondary endpoint was the identification of tumor-driven immune changes in prostatic lesions. Results: The cohort consisted of 3 patients who underwent radical prostatectomy (RP) and 45 patients with positive prostate biopsy; the negative control was checked by pathological assessment. In patients who underwent RP the gene expression analysis identified a modulation in the abundance of several immune subsets infiltrating the tumoral tissue, when compared with the non tumoral, evident for Tumor associated macrophages (TAMs), Natural Killer cells (NK) and T regulatory cells. We then implemented a 22 parameters flow cytometry panel that we tested on fresh prostatic tissue and peripheral blood from positive PCa biopsies. We identified a subset of tumor infiltrating macrophages showing an altered gene expression profile when compared with macrophages infiltrating the non-tumoral tissue. Importantly we derived a genetic signature from this subset of tumoral TAMs that resulted to be associated with cancer progression. Conclusions: Our findings support the effectiveness of single cell RNA sequencing in the dissection of the immune landscape in PCa and identified immune changes in patients when comparing neoplastic tissue with non tumoral areas. Such data may be useful for understanding the role of immune system in PCa carcinogenesis.

Abstract P5-04-01: PARP inhibition modulates the infiltration, phenotype and function of tumor-associated macrophages (TAMs) in BRCA-associated breast cancer and can be augmented by harnessing the anti-tumor potential of TAMs

Abstract Patients with BRCA-associated triple negative breast cancer (TNBC) have few effective treatment options. PARP inhibitors are promising, and we recently showed they induce an influx of white blood cells, including CD8+ T-cells and macrophages into the tumor. The influx of CD8+ cells, mediated by activation of the STING pathway in tumor cells, contributes substantially to efficacy of PARP inhibition in mice. Strikingly, in these studies, the greatest infiltration of immune cells into the tumor was macrophages. Given objective responses to PARP inhibition have been observed in clinical trials but the benefits are transitory, we hypothesized that this was presumably due to a suppressive tumor microenvironment, driven by tumor macrophages. To better understand the molecular basis of resistance to PARP inhibitors, we used high dimensional single-cell immune profiling on human TNBC. We observed a ≥10-fold increase in TAMs in BRCA-associated TNBC compared to BRCA-wildtype TNBC. Using a pre-clinical model of BRCA1-deficient triple-negative breast cancer, we found that PARP inhibitors not only further increased TAM abundance but also induced functional and phenotypic changes associated with STING pathway activation, antigen presentation, and chemokine and cytokine signaling. PARP inhibitors increased the frequency of TAMs expressing co-stimulatory molecules CD80 and CD86 as well as the activation and maturation marker CD40, which are indicative of an anti-tumor phenotype. We also identified a novel negative feedback mechanism which limits the functionality of the anti-tumor TAMs, and is consistent with induction of an immune suppressive macrophage population. Utilizing transcriptomic, proteomic and metabolic profiling of ex vivo cultured human myeloid cells, we identified multiple biological processes associate with PARP inhibition, showing that these drugs directly affect macrophage states and phenotypes. Remarkably, in the pre-clinical BRCA1-deficient TNBC model, the novel combination of PARP inhibition with macrophage modulation significantly extended remissions obtained with PARP inhibitor therapy only, and this advantage persisted when treatment was discontinued, suggestive of a durable reprogramming of the tumor microenvironment. Moreover, CD8+ cells were required for the extension of PARP inhibitor-induced remissions, suggesting that targeting macrophages lifted the constraints imposed by pro-tumor macrophages on CD8+ T cell-mediated tumor cell killing. We identify mechanisms related to macrophage and T-cell activation that increase PFS and provide evidence that TAMs may serve as targets for new therapeutic interventions designed to overcome PARP inhibitor resistance in BRCA-associated TNBC. Citation Format: Jennifer L Guerriero, Anita K Mehta, Emily M Cheney, Jessica A. Castrillon, Jia-Ren Lin, Mateus de Oliveira Taveira, Olmo Sonzogni, Constantia Pantelidou, Christina A Hartl, William M Oldham, Nathan T Johnson, Sarah A Boswell, Marian Kalocsay, Matthew J Berberich, Sholin Mei, Dan Wang, Shawn Johnson, Brett Gross, Deborah A Dillon, Mikel Lipschitz, Evisa Gjini, Scott Rodig, Sandro Santagata, Judy E Garber, Nadine Tung, Peter Sorger, Geoffrey I Shapiro, Gerburg M Wulf, Elizabeth A Mittendorf. PARP inhibition modulates the infiltration, phenotype and function of tumor-associated macrophages (TAMs) in BRCA-associated breast cancer and can be augmented by harnessing the anti-tumor potential of TAMs [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-04-01.

Abstract P5-04-07: Defining T cell dysfunctionality in breast cancer by single cell analysis: Implications for immunotherapy

Abstract Background Despite the well-known association between extensive lymphocytic infiltration in breast cancer (BC), good prognosis, and high response rates to neoadjuvant chemotherapy (CT), pathologic evaluation of tumor infiltrating lymphocytes (TILs) is currently not routine and a deep understanding of the properties of TILs, remain largely unfilled. So far, only in triple-negative (TN), PD-L1 positive BCs it has been possible to demonstrate a survival benefit with immunotherapy plus CT, thereby making the definition of possible immunotherapeutic targets across all biological subtypes a fundamental requirement. Material and methods Single-cell RNA-sequencing (scRNAseq) data from 8 BCs were downloaded from Gene Expression Omnibus dataset (GSE114725), obtaining CD45+ single cells only from the tumoral compartment. scRNAseq-guided high-dimensional profiling by 27-parameter FACS was then applied to a large cohort (n=54, including luminal-like, TN and HER2+ BCs) of early BC patients surgically treated at our Institution. We simultaneously profiled the peripheral blood, the normal and tumoral tissue from each patient and acquired using FACS Symphony A5 flow cytometer (BD Biosciences). Flow Cytometry Standard (FCS) 3.0 files were imported into FlowJo software v9, and analyzed by standard gating to remove aggregates and dead cells, and subsequently imported in FlowJo v10, biexponentially transformed, and exported for further analysis in R by a custom-made script, using an ad-hoc pipeline. Data were analyzed using the Phenograph unbiased algorithm coded in the cytofkit package. Data were further analyzed in FlowJo to determine the frequency of positive cells for each marker and the corresponding median fluorescence intensity (MFI). These values were multiplied to derive the integrated MFI; hierarchical metaclustering of all samples, based on the frequency of Phenograph clusters, was performed in R based on the Euclidean distance and Ward-linkage. Pearson correlation analysis was used to investigate the relationship between CD8+ and CD4+ clusters. Results Our scRNA-guided informative 27-colors flow cytometry panel included antibodies to define not only different lymphocyte subpopulations, but also different stages of T lymphocytes: differentiation and memory maturation (CCR7, CD45RA), activation status (HLA-DR), cytotoxicity (GZMK, GZMB), exhaustion (PD-1, TIGIT) and tissue residency (CD69, CD103). Focusing on CD8+ T cells, we observed, as expected, bona fide naïve T cells to be virtually absent at the tumor site and enriched in peripheral blood, while cytotoxic and effector memory cells were enriched in the tumor compartment. Of note, we identified a population of tissue resident memory T cells (Trm) CD69+ CD103+, CD39+, specifically enriched in the tumor, that could be further subdivided into a HLA-DR+ CD127- and a HLA-DR- CD127+ subpopulation. Interestingly, the HLA-DR+ subset exhibited more exhaustion markers (PD-1hi, TIGIT+, NKG2A+) and was significantly positively correlated with CD4+ regulatory T cells. We speculated, and deeply investigated by bulk RNA-seq, that these two subsets of Trm might represent different maturation states and that they could be reinvigorated targeting highly expressed inhibitory molecules using approved (e.g., PD-1) together with newly identified (e.g., NKG2A) immunotherapeutics. Moreover, as no difference was observed in the relative distribution of Phenograph clusters, our data may be applied to revert acquired immune escape mechanisms in all different BC biological subtypes. Conclusions We identified, among Trm CD69+ CD103+, tumor-specific CD39+, PD1+ exhausted population, novel dysfunctional NKG2A+ T cells previously not characterized in BC. This population is of extreme interest to characterize a novel potential immunotherapeutic target. Citation Format: Agnese Losurdo, Caterina Scirgolea, Emilia Mazza, Valentina Errico, Bethania Fernandes, Luca Di Tommaso, Andrea Sagona, Karolina Pilipow, Rosalba Torrisi, Giovanna Masci, Rita De Sanctis, Elisa Agostinetto, Alberto Testori, Corrado Tinterri, Massimo Roncalli, Armando Santoro, Enrico Lugli. Defining T cell dysfunctionality in breast cancer by single cell analysis: Implications for immunotherapy [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-04-07.

OSbrca: A Web Server for Breast Cancer Prognostic Biomarker Investigation With Massive Data From Tens of Cohorts.

Potential prognostic mRNA biomarkers are exploited to assist in the clinical management and treatment of breast cancer, which is the first life-threatening tumor in women worldwide. However, it is technically challenging for untrained researchers to process high dimensional profiling data to screen and validate the potential prognostic values of genes of interests in multiple cohorts. Our aim is to develop an easy-to-use web server to facilitate the screening, developing, and evaluating of prognostic biomarkers in breast cancers. Herein, we collected more than 7,400 cases of breast cancer with gene expression profiles and clinical follow-up information from The Cancer Genome Atlas and Gene Expression Omnibus data, and built an Online consensus Survival analysis web server for Breast Cancers, abbreviated OSbrca, to generate the Kaplan–Meier survival plot with a hazard ratio and log rank P-value for given genes in an interactive way. To examine the performance of OSbrca, the prognostic potency of 128 previously published biomarkers of breast cancer was reassessed in OSbrca. In conclusion, it is highly valuable for biologists and clinicians to perform the preliminary assessment and validation of novel or putative prognostic biomarkers for breast cancers. OSbrca could be accessed at http://bioinfo.henu.edu.cn/BRCA/BRCAList.jsp.

Single Cell Transcriptome Analysis Reveals Disease-Defining T Cell Subsets in the Tumor Microenvironment of Classic Hodgkin Lymphoma

INTRODUCTION: Classic Hodgkin lymphoma (cHL) is uniquely characterized by an extensively dominant microenvironment composed primarily of different types of non-cancerous immune cells with a rare population (~1%) of tumor cells. Detailed characterization of these cellular components and their spatial relationship is crucial to understand crosstalk and therapeutic targeting in the cellular ecosystem of the tumor microenvironment (TME). METHODS: In this study, we performed high dimensional and spatial profiling of immune cells in the TME of cHL. Single cell RNA sequencing (scRNA-seq) was performed with the 10x Genomics platform on cell suspensions collected from lymph nodes of 22 cHL patients, including 12 of nodular sclerosis subtype, 9 of mixed cellularity subtype and 1 of lymphocyte-rich subtype, with 5 reactive lymph nodes (RLNs) serving as normal controls. Illumina sequencing (HiSeq 2500) was performed to yield single-cell expression profiles for 127,786 cells. We also performed multicolor IHC and imaging mass cytometry (IMC) on TMA slides from the same patients. RESULTS: Unsupervised clustering using PhenoGraph identified 22 cell clusters including 12 T cell clusters, 7 B cell clusters and 1 macrophage cluster. While most immune cell populations were common between cHL and RLN, we observed an enrichment of cells from cHL in all 3 regulatory T cell (Treg) clusters. The most cHL-enriched cluster was characterized by high expression of LAG3, in addition to common Treg markers such as IL2RA (CD25) and TNFRSF18 (GITR), but lacked expression of FOXP3, consistent with a type 1 regulatory (Tr1) T cell population. LAG3+ T cells in cHL had high expression of immune-suppressive cytokines IL-10 and TGF-b . In vitro exposure of T cells to cHL cell line supernatant induced significantly higher levels of LAG3 in naïve T cells compared to co-culture with other lymphoma cell line supernatant or medium only. CD4+ LAG3+ T cells isolated by FACS also suppressed the proliferation of responder CD4+ T cells when co-cultured in vitro. Additionally, Luminex analysis revealed that cHL cell lines secrete substantial amounts of cytokines and chemokines that can promote Tr1 cell differentiation (e.g. IL-6). Our scRNA-seq analysis revealed that LAG3 expression was significantly higher in cHL cases with loss of major histocompatibility class II (MHC-II) expression on HRS cells as compared to MHC-II positive cases (P = 0.019), but was not correlated with EBV status or histological subtype. Strikingly, LAG3 was identified as the most up-regulated gene in cells from MHC-II negative cases compared to MHC-II positive cases. Topological analysis using multicolor IHC and IMC revealed that in MHC-II negative cases, HRS cells were surrounded by LAG3+ T cells. In these cases, the density of LAG3+ T cells in HRS cell-rich regions was significantly increased, and the average distance between an HRS cell and its closest LAG3+ T cell neighbor was significantly shorter. These associations were confirmed in an independent cohort of 166 cHL patients. Finally, we observed a trend towards an inferior disease-specific survival (DSS; P = 0.072) and overall survival (OS; P = 0.12) in cases with an increased number of LAG3+ T cells. A high proportion of LAG3+ T cells (> 20%) was identified as an independent prognostic factor for DSS by multivariate Cox regression. CONCLUSIONS: Our results reveal a diverse TME composition with inflammatory and immunosuppressive cellular components that are linked to MHC class II expression status on HRS cells (Figure). Unprecedented transcriptional and spatial profiling at the single cell level has established the pathogenic importance of HRS cell-induced CD4+ LAG3+ T cells as a mediator of immunosuppression in cHL, with potential implications for novel therapeutic approaches. Figure Disclosures Savage: Seattle Genetics, Inc.: Consultancy, Honoraria, Research Funding; BMS, Merck, Novartis, Verastem, Abbvie, Servier, and Seattle Genetics: Consultancy, Honoraria. Scott:Roche/Genentech: Research Funding; Celgene: Consultancy; Janssen: Consultancy, Research Funding; NanoString: Patents & Royalties: Named inventor on a patent licensed to NanoSting [Institution], Research Funding. Steidl:Bristol-Myers Squibb: Research Funding; Nanostring: Patents & Royalties: Filed patent on behalf of BC Cancer; Roche: Consultancy; Seattle Genetics: Consultancy; Bayer: Consultancy; Juno Therapeutics: Consultancy; Tioma: Research Funding.

High Dimensional Immune Profiling in Smoldering Multiple Myeloma Identifies Novel Organizing Features of the Tumor Microenvironment

Multiple myeloma (MM) is a malignancy of plasma cells that arises from premalignant Monoclonal Gammopathy of Undetermined Significance (MGUS) and often progresses through an asymptomatic Smoldering (SMM) phase lasting months or years before manifesting clinical symptoms warranting therapy. Current research indicates that the tumor microenvironment (TME) in the bone marrow may play a significant role in governing progression to symptomatic disease. Therefore, understanding of the interactions between malignant plasma cells and the TME in early disease states is critical in the pursuit of therapies that will prevent progression to symptomatic disease. We performed high dimensional genomic and immunologic analysis of bone marrow specimens from 73 subjects with SMM. We performed RNA-seq on the malignant plasma cells isolated by anti-CD138 magnetic bead positive selection, mass cytometry (CyTOF) and T cell receptor sequencing (TCR Seq) of CD138-depleted bone marrow mononuclear cells, and proteomics, seromic, and grand serology analysis of bone marrow plasma. These samples and assays provided a broad view of the tumor cells and the cellular and soluble components of the TME. Each of these assays identified self-organizing clusters of subjects, indicating that subgroups of SMM patients shared common characteristics in the tumor or TME populations. We then applied novel bioinformatic methods to compare data from pairs, trios, quartets, and quintets of assays to identify communities of subjects with similar immunologic and genomic characteristics. Integrated analysis of CyTOF, proteomic, and TCR Seq resolved three distinct communities with a high degree of significance. These communities shared distinct cellular and proteomic features that suggested early adaptive, activated adaptive, or innate immune characteristics. These results suggest that the continuum from MGUS to MM does not consist of a single pathway in either the tumor cells or the TME, and that complex interactions ultimately determine progression. This suite of assays (CyTOF, proteomics, and TCR Seq) may be applicable in translational and clinical studies to understand key tumor and immune determinants of SMM and lead to rationally designed therapy to replicate these conditions to prevent progression to symptomatic disease. Disclosures Cho: Genentech: Honoraria, Research Funding; Celgene: Honoraria, Research Funding; The Multiple Myeloma Research Foundation: Employment; Takeda: Research Funding; BMS: Consultancy; Agenus: Research Funding; GSK: Consultancy. Adams:Janssen Pharmaceuticals R&D: Employment, Other: Own Stock. Parekh:Foundation Medicine Inc.: Consultancy; Celgene Corporation: Research Funding; Karyopharm Inc.: Research Funding.

High dimensional profiling of the immune microenvironment in smoldering multiple myeloma

Multiple myeloma (MM) is a malignancy of plasma cells that arises from premalignant Monoclonal Gammopathy of Undetermined Significance (MGUS) and often progresses through an asymptomatic Smoldering (SMM) phase lasting months or years before manifesting clinical symptoms warranting therapy. Current research indicates that the immune microenvironment (IME) in the bone marrow may play a significant role in governing progression to symptomatic disease. Therefore, understanding of the interactions between malignant plasma cells and the IME in early disease states is critical in the pursuit of therapies that will prevent progression to symptomatic disease.

Abstract 4547: High dimensional proteomic profiling of immune cell subsets with data-independent acquisition mass spectrometry

Abstract Background Recent success with therapies to activate the immune system has demonstrated the utility of targeting the immune system for control of multiple cancers. These successes have also spurred interest in characterizing immune cell sub-populations to understand mechanisms of activation and suppression, and their relationship to therapeutic response. Currently, antibody-based approaches are commonly used to characterize immune cells, however these methods are limited to 30-40 markers and are driven by previous hypotheses, limiting new discovery. In this context, determination of the surface and cellular proteome of responsive populations will provide a powerful tool for insight into response mechanisms, so far hampered by low sample material availability and insufficient sensitivity of proteomic methodology. Here, we demonstrate how data-independent acquisition mass spectrometry can be used for high-dimensional characterization of immune cells, even with very limited cell numbers. Methods Primary human Cytotoxic CD8+ T cells, CD4+ T cells, CD14+ monocytes and natural killer (NK) cells, isolated from peripheral blood mononuclear cells, were prepared for mass spectrometry using standard sample preparation workflows. All samples were analyzed using 2 and 4 hour LC gradients on a C18 column coupled to a Thermo Scientific Q Exactive HF mass spectrometer in data-independent acquisition (DIA-MS) mode. DIA data was extracted using Spectronaut Pulsar X (Biognosys) both with a library generated using directDIA data searching in Spectromine as well as a Hybrid Library combining the directDIA and a publicly available resource library. Results Cytotoxic CD8+ T cells were evaluated using 100,000 cells of input material, which resulted in more than 3,500 proteins quantified in the primary cells. When combined with a CD8+ T cell resource library, the number of proteins quantified was more than 5,000. In the current experimental setup, 30,000 cells defines the lower limit of detection of CD8A and CD8B. Among other previously characterized proteins associated with CD8+ T cells, CCL5, TBX21, GZMH, PRF1, GNLY, CST7 were all detected at 30,000 cell input. Additionally, Granzyme A and B were also quantified which have classically been used, along with PRF1, as markers of lymphocyte infiltration. Data will also be presented for CD4+ T cells, CD14+ monocytes, and NK cells to further map and compare the immune cell phenotypic landscapes. Conclusions The DIA-MS platform enables deep proteomic phenotyping of sorted immune cell samples, even with limited numbers of cells. These new data sets make available broad and un-constrained biomarker investigation for deconvolution of the processes driving immune cell activation and suppression. Citation Format: Jakob Vowinckel, Tobias Treiber, Kristina Beeler, Nicholas Dupuis. High dimensional proteomic profiling of immune cell subsets with data-independent acquisition mass spectrometry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4547.

Abstract 4055: Late-differentiated effector neoantigen-specific CD8+ T cells are enriched in non-small cell lung cancer patients responding to atezolizumab treatment

Abstract There is strong evidence that immunotherapy-mediated tumor rejection can be driven by the reinvigoration of tumor-specific CD8+ T cells recognizing neoantigens derived from tumor somatic mutations (citations). Thus, it is possible that the relative abundance and/or characteristics of these tumor-reactive, mutation-specific CD8+ T cells can be used as predictive biomarkers of response to immunotherapy. However, only a fraction of these potential neoantigens are usually immunogenic and in addition, these tumor-reactive, mutation-specific CD8+ cells are present only at low frequencies in blood, making it difficult to reliably identify these effector cells. Here, mass cytometry and highly-multiplexed combinatorial tetramer staining together with cellular barcoding was used to profile immune cells in longitudinally collected PBMCs from 14 non-small cell lung cancer (NSCLC) patients treated with anti-PD-L1 (atezolizumab) antibody to compare patients with objective response (n=8) and progressive disease (n=6). Although no significant phenotypic differences were detected in bulk CD8+ T cells, greater insight was gained from a parallel analysis using highly multiplexed peptide-MHC multimer staining to screen and profile antigen-specific T cells. A longitudinal analysis was performed using peripheral blood CD8+ T cells for 800 candidate tumor neoantigens and 73 known viral-derived control peptides across all patient samples. In addition to virus antigen-specific T cells, a total of 20 different neoantigen-specific T cell populations were detected and their high dimensional profiles were compared. We found that neoantigen-specific T cells were more frequently detected in responding patients and their phenotypes were almost entirely distinct from non-responding patients. Neoantigen-specific T cells from responding patients showed a differentiated effector phenotype with high expression of KLRG1, 2B4 (CD244) and CD57, similar to CD8+ T cells associated with CMV and some types of EBV infection. In contrast, more memory-like phenotypic profiles, with high CD27 and CD127 expression, were observed for neoantigen-specific CD8+ T cells from patients with progressive disease. In addition to the utility of this approach for the ex vivo identification, characterization, and longitudinal tracking of rare tumor-specific T cells, this study supports further research into assessing whether the presence of late-differentiated neoantigen-specific T cells could be used as a predictor of response to checkpoint blockade. Citation Format: Mahesh Yadav, Michael Fehlings, Suchit Jhunjhunwala, Bill O'Gorman, Priti Hegde, Leesun Kim, Alessandra Nardin, Susan Flynn, Hermi Sumatoh, Marcus Ballinger, David Shames, Boon Heng Lee, Evan Newell. Late-differentiated effector neoantigen-specific CD8+ T cells are enriched in non-small cell lung cancer patients responding to atezolizumab treatment [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4055.

Abstract 4054: Mass cytometry approaches to biomarker discovery via high-dimensional antigen-specific T cell identification and profiling

Abstract immunoSCAPE leverages the high-dimensional immune profiling capabilities of mass cytometry combined with a unique technology for the identification of antigen-specific T-cells to support the development of immunotherapy strategies in cancer and immune-related diseases. In cancer, there is now strong evidence that immunotherapy-mediated tumor rejection can rely on tumor-specific (neoantigen-specific) CD8+ T-cells. Consequently, the discovery of neoantigens becomes valuable for personalized cancer immunotherapies. Although in silico pipelines exist, that are capable of predicting non-synonymous mutations potentially giving rise to tumor-specific neoantigens, it is not clear how accurate these methods are in identifying immunogenic and therapeutically relevant epitopes, since T-cell epitope usage can be influenced by many factors. Moreover, analysis of T-cells in cancer patients is challenging as it requires detecting rare antigen-specific T-cell populations in samples that are usually limited in volume and availability. By applying cytometry by time of flight in conjunction with combinatorial peptide-MHC tetramer staining and high-performance dimensional analysis tools, we are able to map broadly MHC-class I epitope with a high sensitivity for rare antigen-specific T-cells and perform concurrently in-depth characterization of these cells. We will show here the application of this technology in the context of immunotherapy, through the example of a murine in vivo tumor model responsive to checkpoint blockade inhibitors, as well as through the analysis of different human cancer samples. Together, by providing insights into the nature of neoantigen-specific T-cells, immunoSCAPE’s unique target discovery and high-dimensional immune profiling platform is a valuable tool for the development of novel diagnostic biomarkers and therapeutic strategies at different stages of drug development. Citation Format: Mathew Cobbett, Michael Fehlings, Evan Newell. Mass cytometry approaches to biomarker discovery via high-dimensional antigen-specific T cell identification and profiling [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 4054.

Abstract 527: High-dimensional profiling of the systemic immune response informs on optimal sequencing of radiotherapy (RT) and immune checkpoint blockade (ICB)

Abstract Background: Combinatorial RT-ICB potentiates anti-tumour reactivity by modulating the immune response. We therefore performed in-depth phenotypic profiling of the systemic T cell compartment following treatment with RT-ICB. Methods: We recruited 20 patients with biopsy-proven metastatic renal cell and non-small cell lung carcinoma, who were treated with a sandwich regime of ICB-RT-ICB under a prospective observational study protocol, and compared against a RT alone-treated cohort (N=10). All patients received ablative RT (8-50Gy/1-5fr) for oligoprogression and/ or local palliation. Blood samples were longitudinally collected at pre-RT, 14 d post-RT and cycle 2 ICB post-RT. Deep T cell profiling was performed by mass cytometry using a customised 41 parameter panel, together with high dimensional analysis tools. Results: Median follow-up of the overall cohort was 18mo; median duration of ICB received in the ICB-RT-ICB arm was 15mo. We observed significant diversity of the systemic T cell repertoire between patients at baseline, and this corresponded to significant interpatient heterogeneity in T cell responses specific to the central/ effector memory, EMRA and Treg subsets post-RT. Dramatic local response (complete response at 1 mo post-RT) was significantly higher in the ICB-RT-ICB cohort compared to the RT alone cohort (12/20 vs 1/10, P<0.01). This clinical phenomenon corresponded to an increased %Ki67high CD8 and CD4 T cells post-RT exclusively in the combinatorial treated cohort, which was further enhanced upon resumption of ICB (mean = 10% vs 3% [CD8]; 13% vs 2% [CD4]; P<0.01). Deeper immunophenotyping of the Ki67high subsets revealed associated high expression of GranzymeB and Eomes. Conclusions: Here, we observed changes in the T cell phenotypes that varied remarkably across all patients following RT. We further highlight a RT-dependent T cell proliferation amongst all RT-ICB-treated patients that was further enhanced by ICB in prior responders. This immune phenomenon may account for the dramatic responses to combinatorial treatment, and informs on optimal sequencing strategies for combining RT and ICB. Citation Format: Kevin L. Chua, Michael Fehlings, Pek Lim Chu, Xiao-Tian Lin, Eugenia Yeo, Kar Perng Low, Dennis Poon, Enya Ong, Wai Yee Woo, Joseph Wee, Alessandra Nardin, Gopalakrishna Iyer, Daniel S. Tan, Kee Chee Soo, Evan Newell, Melvin Chua. High-dimensional profiling of the systemic immune response informs on optimal sequencing of radiotherapy (RT) and immune checkpoint blockade (ICB) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 527.

Immune profiling of tumor-infiltrating T cells using mass cytometry.

2607 Background: Immunotherapy recent successes have opened new avenues for the treatment of cancer and the presence of tumor-specific CD8+ T cells in tumor-bearing individuals offer a promising therapeutic target. However, the detection and profiling of such T cells are challenging due to the need to detect rare antigen-specific T cell subpopulations in patient samples that are limited in size thus making it difficult to exploit these parameters for predictive signatures of clinical response. Moreover, the identification and analysis of neoantigen-specific CD8+ T-cells in tumor-bearing individuals is challenging due to the small pool of such cells. Methods: In order to identify therapy-relevant tumor antigens and to facilitate a concurrent in-depth characterization of cells directed towards these targets, immunoSCAPE leverages the high-dimensional immune profiling capabilities of cytometry by time of flight (CyTOF) combined with a unique technology allowing the identification rare antigen-specific T-cell subsets. Results: We applied this technology to patient tumor-infiltrating lymphocytes from human cancer samples and tumor-derived neoantigens recognized by T-cells were identified and characterized. Interestingly, the majority of patient-derived tumor infiltrates consisted of tumor-unrelated T-cells characterized by a diverse phenotype. Strikingly, the expression of CD39 was absent from these bystander cells, suggesting that CD39 could be a useful biomarker for the identification of putative tumor-reactive T cells. Conclusions: Simultaneous immune profiling revealed that tumor-unrelated, bystander CD8+ T-cells are phenotypically different in human tumor infiltrates and identified CD39 as a putative marker of neoantigen-specific T-cells. By providing insights into the nature, frequency and phenotype of antigen-specific T-cells, immunoSCAPE’s unique target discovery and high-dimensional immune profiling platform is a valuable tool for the development of novel diagnostic and therapeutic strategies in immunotherapy.

Osmium-Labeled Microspheres for Bead-Based Assays in Mass Cytometry.

Polystyrene beads are broadly applied in flow cytometry. Implementing bead-based assays in mass cytometry is desired but hampered by the lack of an elemental label required for their detection. In this study, we introduce stable osmium tetroxide labeling as a universal approach for generating functionalized beads readily detectable by mass cytometry. We demonstrate the utility of osmium-labeled beads for signal spillover compensation in mass cytometry, and, strikingly, their application in quantitative Ab-binding capacity assays combined with high-dimensional profiling of human PBMC enabled the systematic assessment of receptor expression profiles across large numbers of cellular phenotypes. This analysis confirmed increased monocytic Siglec-1 expression in active systemic lupus erythematosus patients and, additionally, revealed interrelated reductions of CD4 expression by regulatory and memory CD4 T cells and HLA-DR expression by myeloid dendritic cells, pointing toward defective cross-talk at the immunological synapse that may limit immune responses in systemic lupus erythematosus. By converting conventional flow cytometry beads into beads suitable for mass cytometry, our approach paves the way toward the broad implementation of bead-based assays in high-dimensional cell profiling studies by mass cytometry in biomedical research.

Glycogen synthase kinase 3 (GSK-3) synergizes with immune-checkpoint blockade (ICB) to overcome tumor resistance due to unique tumor infiltrating T-cell subsets as revealed by mass cytometry (CyTOF)

Abstract We have shown that the serine/threonine kinase glycogen synthase kinase-3 (GSK-3) is a major regulator of programmed cell death-1 (PD-1) transcription and expression in T-cells (Taylor et al., 2016, Immunity). Further, small molecule inhibitors (SMIs) increased Tbx21 transcription for enhanced Tbet expression which suppressed Pdcd1 transcription in CD8+cytolytic T-cells. SMIs against GSK-3 are as effective as anti-PD-1 blocking antibody in the control of spontaneous pulmonary metastasis of B16 melanoma and EL-4 lymphoma cells (Taylor et al., 2018, Cancer Research). This introduced a new chemical approach to down-regulate PD-1 for the treatment of cancer. In addition, GSK-3 effects in Tbet expression can induce the expression of interferon-γ1 (IFNγ1) and granzyme B (GZMB) for enhanced cytolytic T-cell (CTL) killing of tumors suggesting that GSK-3 SMIs might synergize with check-point blockade. We now show that GSK-3α/β inhibition can cooperate with anti-PD-1 blockade to overcome tumor resistance. Using high dimensional profiling of tumor infiltrated lymphocytes (TILs) with mass cytometry (CyTOF), we show that the synergy of GSK-3 SMI and anti-PD-1 is accompanied by 2 specific findings: (i) an increase in the presence of a unique subset of CD8+ TIL effector-memory T-cells beyond that observed with anti-PD-1 alone, and (ii) the significant decrease of a subset of CD4+ suppressor regulatory T-cells (Tregs) not observed with anti-PD-1 alone. Our findings offer a small molecule approach for synergy with anti-PD-1 that leads to an increase in CD8+ T-cells and the marked decrease of Tregs in tumors.

701 High dimensional peripheral blood immune cell profiling of atopic dermatitis patients unveiled IL21 cytokine activation

Atopic dermatitis (AD) is the most common chronic inflammatory skin disease affecting 7% of adults in the United States. While skin and blood profiling of AD primarily highlighted its Th2/ -skewing, novel Th2-targeting strategies are able to induce clearance in <40% of moderate-to-severe patients. To expand upon the current knowledge on the systemic immune activation of AD, we measured the frequencies of 13 cytokines, 4 transcription factors and activation markers in circulating skin-homing (CLA+) and systemic (CLA-) CD4+/CD8+ activated T-cell subsets using mass cytometry or CyTOF, that allows simultaneous profiling of more than 40 markers. High profiling mass cytometry enabled us to measure not only the defining Th/Tc cytokines Th/c1/IFN-g/Th/c2/IL-13/Th/c17/IL-17A/Th/c22/IL-22,Th/c/IL-9 in CD4+/CD8+ T-cells but also TNFa/GM-CSF/IL-2/IL-4/IL-5/IL-31/IL-10/IL-21 producing T cells in blood of moderate-to-severe AD patients (n=20) and controls (n=15). These cytokine-producing T-cell subsets were also highly correlated with disease severity (SCORAD). In addition to the Th2/Tc2 and Th22/Tc22 polarization, the blood of AD patients was characterized by significant increases in IL21 producing CD4+/CD8+ T-cells (p<0.01) in both CLA+/CLA- compartments. IL-21 production was primarily detected among Th/c2 and Th/c22 in AD but not in control subjects, as well as in a unique population that mainly produces IL-21. Positive correlations with SCORAD were observed in CLA+IL21+CD4+/CD8+ T cells (r=0.61; p<0.01). The frequency of these CLA+IL21+CD4+/CD8+ T cells was also correlated with that of the CLA+Th/c2/Th/c22 cells (p<0.001), suggesting the Th/c2/Th/c22 as potentially the source of IL21 in AD. In conclusion, high dimensional profiling of peripheral blood with mass cytometry revealed a possible role of systemic IL21-skewing in moderate-to-severe AD.

The role of high-dimensional profiling of the systemic immune response on optimal sequencing of radiotherapy (RT) and immune checkpoint blockade (ICB).

13 Background: Combinatorial RT-ICB potentiates anti-tumour reactivity by modulating the immune response. We therefore performed in-depth phenotypic profiling of the systemic T cell compartment following treatment with RT-ICB. Methods: We recruited 20 patients with biopsy-proven metastatic renal cell and non-small cell lung carcinoma, who were treated with a sandwich regime of ICB-RT-ICB under a prospective observational study protocol, and compared against a RT alone-treated cohort (N = 10). All patients received ablative RT (8-50 Gy/1-5 fr) for oligoprogression and/or local palliation. Blood samples were longitudinally collected at pre-RT, 14 d post-RT and cycle 2 ICB post-RT. Deep T cell profiling was performed by mass cytometry using a customised 41 parameter panel, together with high dimensional analysis tools. Results: Median follow-up of the overall cohort was 18 mo; median duration of ICB received in the ICB-RT-ICB arm was 15 mo. We observed significant diversity of the systemic T cell repertoire between patients at baseline, and this corresponded to significant interpatient heterogeneity in T cell responses specific to the central/effector memory, EMRA and Treg subsets post-RT. Dramatic local response (complete response at 1 mo post-RT) was significantly higher in the ICB-RT-ICB cohort compared to the RT alone cohort (12/20 vs 1/10, P &lt;0.01). This clinical phenomenon corresponded to an increased % Ki67highCD8 and CD4 T cells post-RT exclusively in the combinatorial treated cohort, which was further enhanced upon resumption of ICB (mean = 10% vs 3% [CD8]; 13% vs 2% [CD4]; P &lt;0.01). Deeper immunophenotyping of the Ki67high subsets revealed associated high expression of GranzymeB and Eomes. Conclusions: Here, we observed changes in the T cell phenotypes that varied remarkably across all patients following RT. We further highlight a RT-dependent T cell proliferation amongst all RT-ICB-treated patients that was further enhanced by ICB in prior responders. This immune phenomenon may account for the dramatic responses to combinatorial treatment, and informs on optimal sequencing strategies for combining RT and ICB.

Single-Cell Correlations of Intron, mRNA, and Protein Content in Human Immune-Immune Cells

The response of biological systems to external stimuli are governed by complex regulatory pathways that often create a large distribution of single-cell responses. Examining this heterogeneity of gene expression, including the single-cell distributions of content and the correlations between multiple genes of interest, can yield valuable information about the sensitivity, kinetic response, and regulatory mechanisms of the genes. Here we develop and apply a fully automated microscopy platform to digitally count the gene expression response of intron bursting sites, single mRNA, and protein content of model cells to external stimuli. Single-molecule intron and mRNA content are measured though the use of single-molecule fluorescence in-situ hybridization (smFISH) while protein content is quantified though the use of antibody probes. Automation of the microscopy platform and image analysis allows thousands of individual cells to be analyzed. To mimic immune response to bacterial infection, THP-1 cells are stimulated with lipopolysaccharide (LPS) and several genes of interest (including IL1β, TNF-α) are simultaneously monitored to examine the distribution, kinetics, and correlation of the single-cell response to this pathogenic stimulus. Multiple correlations, including intron-mRNA-protein content for a single gene and mRNA-mRNA content between multiple genes, are examined and quantified with high temporal resolution over the course of the LPS stimulation. We are working to apply this platform towards viral infections, with the goal of identifying unique signatures capable of accurately distinguishing between viral and bacterial infections. Additionally, work developing a new microfluidic platform enabling high-dimensional profiling of tens-hundreds of individual genes in single-cells will be highlighted.

Dissecting alterations in human CD8+ T cells with aging by high-dimensional single cell mass cytometry

We investigated the effect of aging on the multi-dimensional characteristics and heterogeneity of human peripheral CD8+ T cells defined by the expression of a set of molecules at the single cell level using the recently developed mass cytometry or Cytometry by Time-Of-Flight (CyTOF) and computational algorithms. CD8+ T cells of young and older adults had differential expression of molecules, especially those related to cell activation and migration, permitting the clustering of young and older adults through an unbiased approach. The changes in the expression of individual molecules were collectively reflected in the altered high-dimensional profiles of CD8+ T cells in older adults as visualized by the dimensionality reduction analysis tools principal component analysis (PCA) and t-distributed stochastic neighbor embedding (t-SNE). A combination of PhenoGraph clustering and t-SNE analysis revealed heterogeneous subsets of CD8+ T cells that altered with aging. Furthermore, intermolecular quantitative relationships in CD8+ T cells appeared to change with age as determined by the computational algorithm conditional-Density Resampled Estimate of Mutual Information (DREMI). The results of our study showed that heterogeneity, multidimensional characteristics, and intermolecular quantitative relationships in human CD8+ T cells altered with age, distinctively clustering young and older adults through an unbiased approach.