RATIONALE:Only subset of patients with Non-Small Cell Lung Cancer (NSCLC) benefit from immunotherapy and this is partly due to limited understanding of how oncogenic mutations shape the tumor microenvironment (TME). OBJECTIVES:To define how EGFR and KRAS mutations influence the spatial organization and immune composition of the NSCLC TME. METHODS:We conducted 38-marker high-plex immunofluorescence to 197 NSCLC tumors (>2 million cells) stratified by EGFR and KRAS genotypes. Quantitative phenotyping and spatial analyses, including cellular neighborhoods (CN), nearest neighbors (NNs = 5-20), and spatial proximity profiling (25-100 μm), were performed to assess immune architecture and clinical correlations. MEASUREMENTS AND MAIN RESULTS:EGFR- and KRAS-mutant tumors showed higher tumor cell density and reduced immune infiltration compared with wild-type tumors. Both mutation types were associated with depletion of cytotoxic T cells, dendritic cells, and granulocytes, while EGFR-mutant tumors showed enrichment of M2-like tumor-associated macrophages (TAMs). CN-analysis identified 14 spatial clusters, with reduced cytotoxic and helper T cell-rich neighborhoods in mutant tumors. NN-analysis revealed shorter distances between M2-like TAMs in EGFR-mutant tumors and greater immune exclusion in non-mutants. Spatial proximity revealed higher densities of T-regs, TAMs near tumor cells in KRAS-mutant tumors. All spatial metrics correlated significantly with prognosis in Cox proportional hazards models, highlighting immune cell positioning as an important predictor of outcome. CONCLUSIONS:EGFR- and KRAS-mutant tumors remodel the NSCLC immune landscape, creating distinct immunosuppressive and immune-excluded niches. Spatial proteomics reveals prognostic immune architectures that may guide mutation-directed immunotherapy strategies.

Spatially organized immune hubs of T cells and antigen-presenting cells (APCs) have been linked to immune checkpoint therapy (ICT) efficacy, yet the mechanisms underlying their function remain unclear. Using CODEX multiplex imaging, we longitudinally characterized the dynamic evolution of intratumoral cellular neighborhoods (CN) defined by triad interactions of CD4 and CD8 T cells with two distinct myeloid APC populations: cDC1s and IFN-gamma-activated macrophages. We termed this CN the immunity-promoting CN (IP-CN) and tracked its progressive development during tumor rejection induced by anti-CTLA-4/anti-PD-1 therapy. A coordinated IFN-gamma; and TNF-alpha signaling signature accompanied the IP-CN assembly. Over time, the IP-CN underwent functional maturation, forming specialized sub-neighborhoods that compartmentalized proliferating T cells at the tumor periphery versus cytotoxic T effector cells interacting with tumor cell targets. Our findings reveal a spatiotemporal mechanism by which the IP-CN sustains and amplifies cytotoxic T cell responses, demonstrating how T cell-APC neighborhoods orchestrate tumor immunity.

High-grade serous ovarian cancer (HGSOC) is a highly aggressive and lethal form of ovarian cancer. Challenges to diagnosis and treatment include a lack of effective screening methods for early detection, the absence of cancer-specific symptoms, and the development of chemoresistance. The genomic instability of HGSOC, further complicated by homologous recombination deficiency (HRD), leads to heterogeneity in HGSOC tumors and patient response to treatment. This makes it challenging to develop a single, effective diagnostic and treatment approach for this disease. Proteogenomic studies have provided some insight into HGSOC biology, but a deeper understanding of the tumor proteome through chemotherapy and disease recurrence is needed. Here we have profiled the proteome of tumors from 29 HGSOC patients before and after multiple rounds of chemotherapy. The proteome of HGSOC tumors remained unchanged for individual patients even after numerous rounds of chemotherapy. Differential expression analysis revealed known and novel proteins associated with chemoresistance and HRD status, further supported by similar changes at the genetic and epigenetic levels. We found that HRD affected proteins related to immune pathways. HRD was also associated with more shared T cell receptor (TCR) CDR3 repertoires of tumor-infiltrating T cells and increased neoantigen counts. ERAP1, a protein involved in peptide trimming before antigen presentation to immune cells via MHC-class I, was overexpressed in homologous recombination proficient (HRP) tumors and negatively correlated to neoantigen count. Its potential role in immune suppression in HRP tumors makes it an attractive therapeutic target that may be effective in combination with immunotherapy, particularly for HRP tumors. 26-plex immunostaining of HGSOC whole tissues further revealed significant differences in the spatial proximities of immune cells to each other and to tumor cells based on HRD status. Through proteomic and imaging analysis, this work has shown that the immune landscape of HGSOC tumors is influenced by homologous recombination status and identified candidate drivers of HGSOC biology. ### Competing Interest Statement The authors have declared no competing interest.

BACKGROUND:While coronavirus disease 2019 (COVID-19) is primarily a respiratory infection, few studies have characterised the immune response to COVID-19 in lung tissue. We sought to understand the pathogenic role of microenvironmental interactions and the extracellular matrix in post-mortem COVID-19 lung using an integrative multi-omic approach. METHODS:Post-mortem formalin-fixed paraffin-embedded lung tissue from fatal COVID-19 and nonrespiratory death control lung underwent multi-omic evaluation by Quantseq Bulk RNA sequencing, Nanostring GeoMx spatial transcriptomics, RNAscope, multiplex immunofluorescence and immunohistochemistry, to evaluate virus distribution, immune composition and the extracellular matrix. Markers of extracellular synthesis and breakdown were measured in the serum of 215 patients with COVID-19 and 54 healthy volunteer controls using ELISA. RESULTS:We found that severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection was restricted to the pneumocytes and macrophages of early-stage disease. Spatial analyses revealed an immunosuppressive virus microenvironment, enriched for PDL1+IDO1+ macrophages and depleted of T-cells. Oligoclonal T-cells in COVID-19 lung showed no enrichment of SARS-CoV-2 specific T-cell receptors. Collagen VI was upregulated and contributed to alveolar wall thickening and impaired gas exchange in COVID-19 lung. Serum from COVID-19 patients showed increased levels of PRO-C6, a marker of collagen VI synthesis, predicted mortality in hospitalised patients. CONCLUSIONS:Our data refine the current model of respiratory COVID-19 with regard to virus distribution, immune niches and the role of the noncellular microenvironment in pathogenesis and risk stratification in COVID-19. We show that collagen deposition is an early event in the course of the disease.

Single-cell technologies have revealed significant microglial cell heterogeneity across the human brain in both health and disease. However, the integration of high-plex protein and spatial information in single-cell approaches constitutes a challenge essential for advancing our cell biology comprehension in the neuroscience field. In the present study, we employed co-detection by indexing (CODEX), a protein multiplexed imaging technology, for the first time to unravel the association between different microglial populations and pathological features of Alzheimer’s disease (AD) in the human brain. We used a 32-plex panel of DNA-barcoded antibodies able to detect neurons, oligodendrocytes, astrocytes, myeloid cells, vascular components, and pathological markers, in the same brain tissue section. In addition, we implemented algorithms to segment morphologically complex cells and advanced data analysis pipelines. Our results provide a comprehensive mapping of the human brain cytoarchitecture, showing different cell phenotypes based on their protein expression and morphology, cell interaction dynamics, and cell spatial organizations in healthy and AD individuals. Through the identification of different microglial phenotypes, we found a specific subpopulation associated to amyloid-ß plaques in AD brains. Interestingly, this subpopulation exhibited shared properties with border-associated macrophages. This study presents a novel approach to explore spatial brain cell heterogeneity in the context of neurological diseases and brings new insights into the microglial diversity in AD.

Abstract Background: Cutaneous squamous cell carcinoma (cSCC), arising from proliferation of malignant epidermal keratinocytes, is the second most common non-melanoma skin cancer. cSCC accounts for 75% of all skin cancer related deaths excluding melanoma, and metastatic and immunotherapy-resistant cases pose an emerging global threat. Tumor development is a gradual process characterized by high mutational burden and an immunosuppressive tumor immune microenvironment (TiME); therefore, identifying the underlying changes across the spatial and temporal landscape of these tumors will be crucial to understanding the key cellular and molecular drivers of disease progression and response. Methods: In this longitudinal study, we used a single-cell, whole-slide approach to mapping key proteins of immune lineage, tissue architecture, cellular metabolism, and stress to study the spatial landscape of cSCC at baseline (pre-treatment) and multiple time points over the course of immunotherapy with the aim to develop a temporal atlas of the tumor. An ultra-high plex antibody panel encompassing over 50 major determinants in the tumor microenvironment was employed on the PhenoCycler®-Fusion spatial biology platform. Deep bioinformatic analyses was performed to identify cellular phenotypes, spatial neighborhoods, heterogenous functional states and cellular interactions by spatial proximity determinations. Results: Previous studies in our laboratory have identified unique phenotypes correlating with response and resistance to immunotherapies in pre-treatment cSCC biopsies. Our results expand on these findings with temporal data tracking the dynamic changes in the TiME over the course of immune checkpoint inhibitor therapy. Apart from differences in immune cell composition and spatial localization patterns, we also saw differences in macrophage polarization and expression of metabolic markers in the responder and non-responder cohorts. Conclusions: Overall, ultrahigh-plex spatio-temporal monitoring of cSCC revealed distinct signatures of response and resistance and identified key features in the TiME that reveal deeper insights into the pathobiology of the tumor. Our discovery-based approach identifies novel biomarkers for patient stratification and therapeutic modulation. Citation Format: Niyati Jhaveri, Bassem Ben Cheikh, Dmytro Klymyshyn, Ning Ma, Aditya Pratapa, James Monkman, Nadine Nelson, Arutha Kulasinghe. A spatio-temporal approach to mapping the dynamics of cutaneous squamous cell carcinoma progression and immunotherapy response: A journey through TiME [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 6872.

Abstract Background: Therapies targeting immune checkpoint have significantly transformed the treatment of several cancers, including lung cancer. However, only 20% of lung cancer patients show a response to immunotherapies. KRAS and EGFR mutations are significant factors in the development and progression of lung cancer and its response to immunotherapy. Hence, it's vital to understand intra-tumoral heterogeneity, the makeup of the Tumor Microenvironment (TME), and the cellular interactions in tumors with KRAS and EGFR mutations to uncover new insights. Methods: Lung cancer tissues from 200 patients were collected, focusing on KRAS and EGFR mutation statuses, specifically KRAS+/EGFR-, KRAS-/EGFR+, and KRAS-/EGFR-. For each patient, two samples were taken, resulting in a total of 400 tissue microarray (TMA) cores. These samples were imaged at single-cell resolution using the PhenoCycler®-Fusion system, which employed a 42-plex antibody panel targeting immune, tumor, proliferation, and apoptosis markers. The first step in the analysis was cell segmentation, performed using a fine-tuned deep learning model. This process successfully segmented a total of 2.16 million cells across the entire collection of samples. Following this, the protein expressions for each cell were calculated, and unsupervised clustering was performed, resulting in 36 distinct clusters, that were manually annotated into 14 cell phenotypes. The proportions of these cell phenotypes were then quantified for each of the 400 TMA cores and compared across the three mutation groups. The analysis also included the spatial proximity and cellular neighborhoods in relation to these mutation groups. Results: Quantitative analysis focusing on the relative abundance of cell types revealed that the ratio of tumor cells to immune cells was significantly lower in KRAS-/EGFR- tumors compared to those with KRAS+/EGFR- and KRAS-/EGFR+ mutations. Furthermore, the percentages of Regulatory T cells (Tregs) and M1 Macrophages were significantly higher in the KRAS+/EGFR- group compared to the KRAS-/EGFR+ group. Spatial proximity analysis indicated that in the KRAS-/EGFR+ group, M2 Macrophages were significantly closer to Cytotoxic T cells and Helper T cells compared to the distances observed in the other two groups. Moreover, neighborhood analysis identified 20 distinct cellular neighborhoods, each defined by specific cell-cell interactions. COX hazard analysis based on these cellular neighborhoods demonstrated notable variations among the three mutation groups, which are linked to differences in patient outcomes. Conclusions: Through single-cell spatial phenotyping, this work has shown that the spatial immune landscape of lung tumors is influenced by KRAS and EGFR mutation statuses. This presents a novel opportunity to enhance our understanding of the spatial structure of lung cancer and to identify more effective therapeutic targets. Citation Format: Rajender Nandigama, Bassem Ben Cheikh, Ning Ma, Jochen Wilhelm, Laura Klotz, Florian Eichhorn, Mark Kriegsmann, Marek Bartkuhn, Jamal Nabhanizadeh, Sascha Seidel, Thorsten Stiewe, Albrecht Stenzinger, Andreas Weigert, Mario Looso, Friedrich Grimminger, Werner Seeger, Soni Savai Pullamsetti, Niyati Jhaveri, Hauke Winter, Rajkumar Savai. Single-cell spatial landscape of the mutation-specific human lung tumor immune microenvironment [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 5508.

Abstract Introduction: The diverse tumor environment of high-grade glioma, which remains refractory to treatment, demands new innovative, multi-omic approaches to characterize tumor heterogeneity and expression profiles associated with progression and therapeutic response. Approach: We have applied an integrated, multi-omic approach using paired spatial in situ sequencing (Xenium) and protein profiling (PhenoCycler® Fusion) of glioma tissue to identify distinct glioma cellular phenotypes, activation states and metabolic pathways at true single cell resolution. Our multi-omic approach used hundreds of phenotypic RNA target probes and >50 antibodies in an unbiased single-cell analysis of primary, recurrent, and IDH1 mutant/wild type tumors with a range of heterogeneity. Summary: We were able to distinguish differentially regulated genes and pathways between aggregated primary and recurrent GBM, including cell cycle pathways being upregulated in primary vs recurrent GBM glial cell clusters and a down regulation of ERBB4 signaling in vascular cell clusters in primary GBMs. Our data also suggests that SOX11, known to be involved in tumorigenesis, is >2-fold upregulated in annotated tumor cells in primary compared to recurrent tumors. We were also able to quantitatively localize the expression of Epidermal Growth Factor Receptor, variant III (EGFRvIII) to specific tumor-cell sub-types, which is important given its association with therapeutic resistance. Further, we have been able to establish the cellular neighborhoods and cell-cell and receptor-ligand relationships within the tumors. Our spatial protein analysis identified distinct tumor and immune phenotypes with varying abundance of myeloid and lymphoid populations in IDH1 wt and mt tumors. Moreover, spatial proximity analyses and cellular neighborhood analyses revealed differences in higher order organizational landscapes that may contribute to the differential outcomes across wt and mt tumor subtypes. Conclusion: Multiomic spatial analysis enables deeper characterization of the glioma cellular and functional landscape to broaden our understanding of the key TME features that contribute to disease pathogenesis and prognoses. Our study provides an analytical framework to combine RNA and protein-based spatial data for a holistic investigation into a variety of glioma subtypes and aid in the identification of novel biomarkers, spatial neighborhoods, and functional states that drive glioma progression. Citation Format: Dmytro Klymyshyn, Vaibhav Jain, Lauren Whaley, Emily Hocke, Bassem Ben Cheikh, Alan Smith, Karen Abramson, Nadine Nelson, Diane Satterfield, Elizabeth Thomas, Giselle Lopez, Seetha Hariharan, Michael Brown, Niyati Jhaveri, Roger McLendon, David Ashley, Matthew Waitkus, Simon Gregory. Multi-omic spatial analysis of the tumor microenvironment in gliomas [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 5496.

Heterogeneous resistance to immunotherapy remains a major challenge in cancer treatment, often leading to disease progression and death. Using CITE-seq and matched 40-plex PhenoCycler tissue imaging, we performed longitudinal multimodal single-cell analysis of tumors from metastatic melanoma patients with innate resistance, acquired resistance, or response to immunotherapy. We established the multimodal integration toolkit to align transcriptomic features, cellular epitopes, and spatial information to provide deeper insights into the tumors. With longitudinal analysis, we identified an "immune-striving" tumor microenvironment marked by peri-tumor lymphoid aggregates and low infiltration of T cells in the tumor and the emergence of MITF+SPARCL1+ and CENPF+ melanoma subclones after therapy. The enrichment of B cell-associated signatures in the molecular composition of lymphoid aggregates was associated with better survival. These findings provide further insights into the establishment of microenvironmental cell interactions and molecular composition of spatial structures that could inform therapeutic intervention.

Abstract Background Targeted immune checkpoint inhibitors (ICI) with anti-PD-1/PD-L1 therapy offer durable treatment of mucosal head and neck squamous cell cancer (HNSCC), in both human papillomavirus-positive (HPV+) and negative (HPV-) patients. However, currently available biomarker signatures for targeted ICI therapies have limited predictive value. Our recent ultrahigh-plex profiling of HNSCC tissue with 100+ cancer hallmarks of tumor and immunobiology uncovered distinct spatial domains that serve as defining factors for clinical response and resistance. Methods Our unbiased analysis of whole-slide metastatic HNSCC tumors is from a clinical cohort of patients treated with Pembrolizumab/Nivolumab. The cohort consisted of patients with a range of outcomes from complete vs partial vs progressive disease responses to ICI therapy. We first characterized the tumor microenvironment using our ultrahigh-plex protein panel with 100+ antibodies on the PhenoCycler®-Fusion platform. To expand upon our biomarker discovery, we included multiomic cancer hallmarks with a multimodal protein/RNA detection panel. Targeted spatial RNA detection was performed to complement and augment the microenvironment characterization achieved by our protein panel. To further consolidate the multiomic data, we leveraged MaxFuse, a state-of-the art computational framework that integrates multimodal spatial and single-cell expression data. Results Our multiomic spatial phenotyping uncovered diverse tumor regions, each with distinct biomarker expression that is reflected across modalities including protein, RNA, and metabolic activity, indicating regions likely associated with resistance to immunotherapy. Our multiomic data integration also revealed spatial signatures associated with different tissue compartments, such as the tumor and non-tumor associated tertiary lymphoid structures. Conclusions We demonstrate a multi-pronged approach that incorporated both novel experimental and computational techniques for elucidating tumor microenvironment in HNSCC tissue prior to ICI-based immunotherapy. Our multiomic approach provides deeper characterization of the HNSCC at the transcriptomic and proteomic level incorporating depth across the entire transcriptome and single-cell spatial resolution of key protein determinants for predicting and furthering our understanding of immunotherapy response to ICI therapy. Citation Format: Aditya Pratapa, Lydia Hernandez, Bassem Ben Cheikh, Niyati Jhaveri, Arutha Kulasinghe. Ultrahigh-plex spatial phenotyping of head and neck cancer tissue uncovers multiomic signatures of immunotherapy response [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 5503.

Abstract Background: Pancreatic cancer has a 5-year survival of only 12%. This is due to the lack of effective treatments and virtually no early detection methods. To compound this already poor prognosis, African Americans face a 20% higher incidence and worse mortality compared to non-African American patients. With a substantial portion of the data and tools employed for studying pancreatic cancer failing to adequately represent this demographic we aim to gain more insight into the differences associated with worse outcomes for African Americans with pancreatic cancer. To achieve this, we conducted an in-depth analysis of tumor mutational burden using whole exome sequencing and the cutting-edge PhenoCycler®-Fusion 2.0 spatial biology platform, to uncover disparities in spatial immune and metabolic phenotypes. Methods: Using an ultrahigh-plex discovery panel of markers encompassing cell lineage, immune checkpoints, tissue architecture, activation, metabolism, proliferation, and stress, we sought to analyze the differences in spatial phenotypes, cellular neighborhoods and functional pathways across the two groups. Additionally, we performed laser capture microdissection on a cohort of patients containing equal numbers of African American and non-African American pancreatic cancer samples. Tumor, stroma, and adjacent normal areas were identified by a board-certified pathologist. These annotations were used to laser capture on 10-micron sections with 7 sections for each patient. We performed whole exome sequencing on these sections and different areas and compared between groups. Results: Through our investigation using these two technologies, we can find differences in the tumor microenvironment between African Americans and non-African Americans. We can see potential associations in the increased mortality and incidence of this group while also building a foundation of racially diverse data to be used for future studies. Whole exome sequencing has revealed distinct mutational patterns among African American patients, involving common genes that exhibit unique mutation profiles when compared to those previously observed in non-African American patients. Moreover, whole-slide spatial phenotyping reveals the differential tumor-immune landscapes and the key cellular and molecular niches that contribute to tumor progression and prognosis. These findings underscore the imperative need for a more systems biology exploration of the variations within these groups, with the aim of developing more effective and personalized treatment approaches. Citation Format: Daniel James Salas-Escabillas, Ning Ma, Bassem Ben Cheikh, Aditya Pratapa, Thais Pichardo, Kyra Langley, Julie Clark, Nina Steele, Niyati Jhaveri, David Kwon, Howard C. Crawford. Mutational analysis and spatial phenotyping to decipher racial disparities in pancreatic adenocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3651.

Single-cell spatial analysis of proteins is rapidly becoming increasingly important in revealing biological insights. Here, we present a protocol for automated high-plex multi-slide immunofluorescence staining and imaging of human head and neck cancer formalin-fixed paraffin-embedded (FFPE) sections using PhenoCycler-Fusion 2.0 technology. We describe steps for preparing human head and neck cancer FFPE tissues, staining with a panel of immunophenotyping markers, and Flow Cell assembly. We then detail procedures for setting up for a PhenoCycler-Fusion run, post-run Flow Cell removal, and downstream analyses.For complete details on the use and execution of this protocol, please refer to Jhaveri et al.1

Cell annotation is a crucial methodological component to interpreting single cell and spatial omics data. These approaches were developed for single cell analysis but are often biased, manually curated and yet unproven in spatial omics. Here we apply a stemness model for assessing oncogenic states to single cell and spatial omic cancer datasets. This one-class logistic regression machine learning algorithm is used to extract transcriptomic features from non-transformed stem cells to identify dedifferentiated cell states in tumors. We found this method identifies single cell states in metastatic tumor cell populations without the requirement of cell annotation. This machine learning model identified stem-like cell populations not identified in single cell or spatial transcriptomic analysis using existing methods. For the first time, we demonstrate the application of a ML tool across five emerging spatial transcriptomic and proteomic technologies to identify oncogenic stem-like cell types in the tumor microenvironment.

Head and neck squamous cell carcinomas (HNSCCs) are the seventh most common cancer and represent a global health burden. Immune checkpoint inhibitors (ICIs) have shown promise in treating recurrent/metastatic disease with durable benefit in similar to 30% of patients. Current biomarkers for HNSCC are limited in their dynamic ability to capture tumor microenvironment (TME) features with an increasing need for deeper tissue characterization. Therefore, new biomarkers are needed to accurately stratify patients and predict responses to therapy. Here, we have optimized and applied an ultra-high plex, single-cell spatial protein analysis in HNSCC. Tissues were analyzed with a panel of 101 antibodies that targeted biomarkers related to tumor immune, metabolic and stress microenvironments. Our data uncovered a high degree of intra-tumoral heterogeneity intrinsic to HNSCC and provided unique insights into the biology of the disease. In particular, a cellular neighborhood analysis revealed the presence of six unique spatial neighborhoods enriched in functionally specialized immune subsets. In addition, functional phenotyping based on key metabolic and stress markers identified four distinct tumor regions with differential protein signatures. One region was marked by infiltration of CD8+ cytotoxic T cells and overexpression of BAK, a proapoptotic regulator, suggesting strong immune activation and stress. Another adjacent region within the same tumor had high expression of G6PD and MMP9, known drivers of tumor resistance and invasion, respectively. This dichotomy of immune activation-induced death and tumor progression in the same sample demonstrates the heterogenous niches and competing microenvironments that may underpin variable clinical responses. Our data integrate single-cell ultra-high plex spatial information with the functional state of the TME to provide insights into HNSCC biology and differential responses to ICI therapy. We believe that the approach outlined in this study will pave the way toward a new understanding of TME features associated with response and sensitivity to ICI therapies.

Immune checkpoint inhibitor (ICI) therapy has drastically improved the treatment strategies for mucosal head and neck squamous cell cancer (HNSCC). To increase treatment response rates for highly targeted ICI therapies, predictive and prognostic biomarkers are being actively explored. Specifically, the cytokine expression patterns in the tumor microenvironment (TME) are now recognized as key to understanding immune responsive and resistant phenotypes. Cytokines play an essential role in the regulation of the TME, specifically in modulating the proliferation and differentiation of immune cells. Here, we have studied spatial signatures of various cytokines within the TME of metastatic/recurrent HNSCC tumors treated with Pembrolizumab/Nivolumab. We utilized the PhenoCycler-Fusion to perform whole-slide, single cell resolution spatial phenotyping of the TME of HNSCC tumors from a cohort of n=40 patients. The discovery cohort consisted of patients who had complete vs. partial vs. stable vs. progressive responses to ICI therapy. Transcriptomic profiling of more than 60 RNA targets for various subfamily of chemokines, interleukins, and immune cell lineages was achieved using Akoya’s novel high plex RNA detection technology. Our study identified distinct spatial signatures that implicate certain cytokines in either tumor progression or regression. Specifically, we have identified areas of high and low CXCL9 and CXCL10 expression in several tumor regions that reflect immune-cell landscapes associated with resistance and sensitivity to immunotherapy. Our study demonstrates the power of unbiased spatial phenotyping with whole-slide imaging to identify biomarkers associated with response to ICI therapy in HNSCC.

Background We previously used scRNASeq and CyTOF to define some of the cellular and molecular changes that occur during immune checkpoint therapy (ICT)-induced rejection of our well-characterized mouse T3 MCA sarcoma model in syngeneic mice. Here, we used the Phenocycler (CODEX) multiplex imaging system with a 35-plex antibody panel to characterize the spatial changes in the lymphoid and myeloid cell populations that result in either tumor outgrowth in mice treated with control antibody (cmAb) or successful tumor rejection in mice treated with the combination of αPD-1 and αCLA-4 (ICT). Methods T3 tumor-bearing mice were treated with cmAb or combo ICT, tumors were harvested at different time points, fresh frozen, sectioned, and subjected to CODEX multiplex imaging. Large-field images encompassing the entire tumor sections were processed using an Akoya-developed pipeline. A hierarchical cell clustering approach was used to profile 4,051,156 cells resulting in the identification of 12 cell types. Results When analyzed at day 10 (one day before ICT-induced rejection becomes detectable) tumors from combo ICT-treated mice displayed a marked increase in the percentage of iNOS+ macrophages, Ly6G+ neutrophils, and a corresponding decrease of CX3CR1+ macrophages, Tregs and tumor cells compared to matched tumors from cmAb treated mice. Although the overall frequencies of CD4+ and CD8+ T cells were not significantly altered at this time point by combo-ICT, they were found in areas of higher density compared to tumors from control mice. Interestingly, whereas KI67+ T cells localized to collagen-rich areas of the tumor stroma, Granzyme B+ T cells accumulated inside the tumor region. Nearest neighbor analysis revealed that combo ICT changed the landscape of heterotypic cellular interactions. Specifically, whereas in cmAb-treated mice, tumor cells and CX3CR1+ macrophages were the two main cell types interacting with one another, in combo ICT-treated tumors, interactions between CD4+ T cells, cDC2, CD206+ macrophages, iNOS+ macrophages, and CD8+ T cells predominated. Cellular neighborhood (CN) mapping indicated the formation of 8 distinct spatially organized CNs across the tumor tissue. Among these, we identified CN-1 that showed increases in frequency after combo ICT and was enriched for T cells, NK cells, and cDC1. Longitudinal images from tumors harvested on days 6 to 13 showed the expansion of the lymphoid-rich CN-1 during tumor rejection. Conclusions These results not only confirm our previous findings of lymphoid and myeloid compartment remodeling during successful ICT but also now provide a detailed view of the dynamic spatial changes leading to successful ICT.