Machine Learning Links T-cell Function and Spatial Localization to Neoadjuvant Immunotherapy and Clinical Outcome in Pancreatic Cancer.

Tumor molecular data sets are becoming increasingly complex, making it nearly impossible for humans alone to effectively analyze them. Here, we demonstrate the power of using machine learning (ML) to analyze a single-cell, spatial, and highly multiplexed proteomic data set from human pancreatic cancer and reveal underlying biological mechanisms that may contribute to clinical outcomes. We designed a multiplex immunohistochemistry antibody panel to compare T-cell functionality and spatial localization in resected tumors from treatment-naïve patients with localized pancreatic ductal adenocarcinoma (PDAC) with resected tumors from a second cohort of patients treated with neoadjuvant agonistic CD40 (anti-CD40) monoclonal antibody therapy. In total, nearly 2.5 million cells from 306 tissue regions collected from 29 patients across both cohorts were assayed, and over 1,000 tumor microenvironment (TME) features were quantified. We then trained ML models to accurately predict anti-CD40 treatment status and disease-free survival (DFS) following anti-CD40 therapy based on TME features. Through downstream interpretation of the ML models' predictions, we found anti-CD40 therapy reduced canonical aspects of T-cell exhaustion within the TME, as compared with treatment-naïve TMEs. Using automated clustering approaches, we found improved DFS following anti-CD40 therapy correlated with an increased presence of CD44+CD4+ Th1 cells located specifically within cellular neighborhoods characterized by increased T-cell proliferation, antigen experience, and cytotoxicity in immune aggregates. Overall, our results demonstrate the utility of ML in molecular cancer immunology applications, highlight the impact of anti-CD40 therapy on T cells within the TME, and identify potential candidate biomarkers of DFS for anti-CD40-treated patients with PDAC.

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 RNAs and proteins 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. Our results demonstrate (1) the feasibility of imputing single-cell abundance levels for many proteins using machine learning to overcome the technical constraints of multiplex tissue imaging and (2) how including cellular spatial information can substantially enhance imputation results.

Machine learning links T cell function and spatial localization to neoadjuvant immunotherapy and clinical outcome in pancreatic cancer.

Tumor molecular datasets are becoming increasingly complex, making it nearly impossible for humans alone to effectively analyze them. Here, we demonstrate the power of using machine learning to analyze a single-cell, spatial, and highly multiplexed proteomic dataset from human pancreatic cancer and reveal underlying biological mechanisms that may contribute to clinical outcome. A novel multiplex immunohistochemistry antibody panel was used to audit T cell functionality and spatial localization in resected tumors from treatment-naive patients with localized pancreatic ductal adenocarcinoma (PDAC) compared to a second cohort of patients treated with neoadjuvant agonistic CD40 (αCD40) monoclonal antibody therapy. In total, nearly 2.5 million cells from 306 tissue regions collected from 29 patients across both treatment cohorts were assayed, and more than 1,000 tumor microenvironment (TME) features were quantified. We then trained machine learning models to accurately predict αCD40 treatment status and disease-free survival (DFS) following αCD40 therapy based upon TME features. Through downstream interpretation of the machine learning models' predictions, we found αCD40 therapy to reduce canonical aspects of T cell exhaustion within the TME, as compared to treatment-naive TMEs. Using automated clustering approaches, we found improved DFS following αCD40 therapy to correlate with the increased presence of CD44+ CD4+ Th1 cells located specifically within cellular spatial neighborhoods characterized by increased T cell proliferation, antigen-experience, and cytotoxicity in immune aggregates. Overall, our results demonstrate the utility of machine learning in molecular cancer immunology applications, highlight the impact of αCD40 therapy on T cells within the TME, and identify potential candidate biomarkers of DFS for αCD40-treated patients with PDAC.

An omic and multidimensional spatial atlas from serial biopsies of an evolving metastatic breast cancer.

Mechanisms of therapeutic resistance and vulnerability evolve in metastatic cancers as tumor cells and extrinsic microenvironmental influences change during treatment. To support the development of methods for identifying these mechanisms in individual people, here we present an omic and multidimensional spatial (OMS) atlas generated from four serial biopsies of an individual with metastatic breast cancer during 3.5 years of therapy. This resource links detailed, longitudinal clinical metadata that includes treatment times and doses, anatomic imaging, and blood-based response measurements to clinical and exploratory analyses, which includes comprehensive DNA, RNA, and protein profiles; images of multiplexed immunostaining; and 2- and 3-dimensional scanning electron micrographs. These data report aspects of heterogeneity and evolution of the cancer genome, signaling pathways, immune microenvironment, cellular composition and organization, and ultrastructure. We present illustrative examples of how integrative analyses of these data reveal potential mechanisms of response and resistance and suggest novel therapeutic vulnerabilities.

Single-cell spatial architectures associated with clinical outcome in head and neck squamous cell carcinoma.

There is increasing evidence that the spatial organization of cells within the tumor-immune microenvironment (TiME) of solid tumors influences survival and response to therapy in numerous cancer types. Here, we report results and demonstrate the applicability of quantitative single-cell spatial proteomics analyses in the TiME of primary and recurrent human papillomavirus (HPV)-negative head and neck squamous cell carcinoma (HNSCC) tumors. Single-cell compositions of a nine patient, primary and recurrent (n = 18), HNSCC cohort is presented, followed by deeper investigation into the spatial architecture of the TiME and its relationship with clinical variables and progression free survival (PFS). Multiple spatial algorithms were used to quantify the spatial landscapes of immune cells within TiMEs and demonstrate that neoplastic tumor-immune cell spatial compartmentalization, rather than mixing, is associated with longer PFS. Mesenchymal (αSMA+) cellular neighborhoods describe distinct immune landscapes associated with neoplastic tumor-immune compartmentalization and improved patient outcomes. Results from this investigation are concordant with studies in other tumor types, suggesting that trends in TiME cellular heterogeneity and spatial organization may be shared across cancers and may provide prognostic value in multiple cancer types.

MCMICRO: a scalable, modular image-processing pipeline for multiplexed tissue imaging.

Highly multiplexed tissue imaging makes detailed molecular analysis of single cells possible in a preserved spatial context. However, reproducible analysis of large multichannel images poses a substantial computational challenge. Here, we describe a modular and open-source computational pipeline, MCMICRO, for performing the sequential steps needed to transform whole-slide images into single-cell data. We demonstrate the use of MCMICRO on tissue and tumor images acquired using multiple imaging platforms, thereby providing a solid foundation for the continued development of tissue imaging software.

Multiomics analysis of serial PARP inhibitor treated metastatic TNBC inform on rational combination therapies.

In a pilot study, we evaluated the feasibility of real-time deep analysis of serial tumor samples from triple negative breast cancer patients to identify mechanisms of resistance and treatment opportunities as they emerge under therapeutic stress engendered by poly-ADP-ribose polymerase (PARP) inhibitors (PARPi). In a BRCA-mutant basal breast cancer exceptional long-term survivor, a striking tumor destruction was accompanied by a marked infiltration of immune cells containing CD8 effector cells, consistent with pre-clinical evidence for association between STING mediated immune activation and benefit from PARPi and immunotherapy. Tumor cells in the exceptional responder underwent extensive protein network rewiring in response to PARP inhibition. In contrast, there were minimal changes in the ecosystem of a luminal androgen receptor rapid progressor, likely due to indifference to the effects of PARP inhibition. Together, identification of PARPi-induced emergent changes could be used to select patient specific combination therapies, based on tumor and immune state changes.

Leukocyte Heterogeneity in Pancreatic Ductal Adenocarcinoma: Phenotypic and Spatial Features Associated with Clinical Outcome.

Immunotherapies targeting aspects of T cell functionality are efficacious in many solid tumors, but pancreatic ductal adenocarcinoma (PDAC) remains refractory to these treatments. Deeper understanding of the PDAC immune ecosystem is needed to identify additional therapeutic targets and predictive biomarkers for therapeutic response and resistance monitoring. To address these needs, we quantitatively evaluated leukocyte contexture in 135 human PDACs at single-cell resolution by profiling density and spatial distribution of myeloid and lymphoid cells within histopathologically defined regions of surgical resections from treatment-naive and presurgically (neoadjuvant)-treated patients and biopsy specimens from metastatic PDAC. Resultant data establish an immune atlas of PDAC heterogeneity, identify leukocyte features correlating with clinical outcomes, and, through an in silico study, provide guidance for use of PDAC tissue microarrays to optimally measure intratumoral immune heterogeneity. Atlas data have direct applicability as a reference for evaluating immune responses to investigational neoadjuvant PDAC therapeutics where pretherapy baseline specimens are not available. SIGNIFICANCE: We provide a phenotypic and spatial immune atlas of human PDAC identifying leukocyte composition at steady state and following standard neoadjuvant therapies. These data have broad utility as a resource that can inform on leukocyte responses to emerging therapies where baseline tissues were not acquired.This article is highlighted in the In This Issue feature, p. 1861.

The Human Tumor Atlas Network: Charting Tumor Transitions across Space and Time at Single-Cell Resolution.

Crucial transitions in cancer-including tumor initiation, local expansion, metastasis, and therapeutic resistance-involve complex interactions between cells within the dynamic tumor ecosystem. Transformative single-cell genomics technologies and spatial multiplex in situ methods now provide an opportunity to interrogate this complexity at unprecedented resolution. The Human Tumor Atlas Network (HTAN), part of the National Cancer Institute (NCI) Cancer Moonshot Initiative, will establish a clinical, experimental, computational, and organizational framework to generate informative and accessible three-dimensional atlases of cancer transitions for a diverse set of tumor types. This effort complements both ongoing efforts to map healthy organs and previous large-scale cancer genomics approaches focused on bulk sequencing at a single point in time. Generating single-cell, multiparametric, longitudinal atlases and integrating them with clinical outcomes should help identify novel predictive biomarkers and features as well as therapeutically relevant cell types, cell states, and cellular interactions across transitions. The resulting tumor atlases should have a profound impact on our understanding of cancer biology and have the potential to improve cancer detection, prevention, and therapeutic discovery for better precision-medicine treatments of cancer patients and those at risk for cancer.

A harmonized meta-knowledgebase of clinical interpretations of somatic genomic variants in cancer.

Precision oncology relies on accurate discovery and interpretation of genomic variants, enabling individualized diagnosis, prognosis and therapy selection. We found that six prominent somatic cancer variant knowledgebases were highly disparate in content, structure and supporting primary literature, impeding consensus when evaluating variants and their relevance in a clinical setting. We developed a framework for harmonizing variant interpretations to produce a meta-knowledgebase of 12,856 aggregate interpretations. We demonstrated large gains in overlap between resources across variants, diseases and drugs as a result of this harmonization. We subsequently demonstrated improved matching between a patient cohort and harmonized interpretations of potential clinical significance, observing an increase from an average of 33% per individual knowledgebase to 57% in aggregate. Our analyses illuminate the need for open, interoperable sharing of variant interpretation data. We also provide a freely available web interface (search.cancervariants.org) for exploring the harmonized interpretations from these six knowledgebases.

High-dimensional multiplexed immunohistochemical characterization of immune contexture in human cancers.

Biomarker assessments of tumor specimens is widely used in cancer research to audit tumor cell intrinsic as well as tumor cell extrinsic features including the diversity of immune, stromal, and mesenchymal cells. To comprehensively and quantitatively audit the tumor-immune microenvironment (TiME), we developed a novel multiplex immunohistochemistry (mIHC) platform and computational image processing workflow using a single formalin-fixed paraffin-embedded (FFPE) tissue section. Herein, we validated this platform using nine matched primary newly diagnosed and recurrent head and neck squamous cell carcinoma (HNSCC) sections sequentially subjected to immunodetection with a panel of 29 antibodies identifying malignant tumor cells, and 17 distinct leukocyte lineages and their functional states. Image cytometric analysis was applied to interpret chromogenic signals from digitally scanned and coregistered light microscopy-based images enabling identification and quantification of individual tumor cells, structural features, immune cell phenotypes and their functional state. In agreement with our previous study via a 12-plex imaging mIHC platform, myeloid-inflamed status in newly diagnosed primary tumors associated with significantly short progression free survival, independent of lymphoid-inflamed status. Spatial distribution of tumor and immune cell lineages in TiME was also examined and revealed statistically significant CD8+ T cell exclusion from tumor nests, whereas regulatory T cells and myeloid cells, when present in close proximity to tumor cells, highly associated with rapid cancer recurrence. These findings indicate presence of differential immune-spatial profiles in newly diagnosed and recurrent HNSCC, and establish the robustness of the 29-plex mIHC platform and associated analytics for quantitative analysis of single tissue sections revealing longitudinal TiME changes.

Innate αß T Cells Mediate Antitumor Immunity by Orchestrating Immunogenic Macrophage Programming.

Unconventional T-lymphocyte populations are emerging as important regulators of tumor immunity. Despite this, the role of TCRαß+CD4-CD8-NK1.1- innate αß T cells (iαßT) in pancreatic ductal adenocarcinoma (PDA) has not been explored. We found that iαßTs represent ∼10% of T lymphocytes infiltrating PDA in mice and humans. Intratumoral iαßTs express a distinct T-cell receptor repertoire and profoundly immunogenic phenotype compared with their peripheral counterparts and conventional lymphocytes. iαßTs comprised ∼75% of the total intratumoral IL17+ cells. Moreover, iαßT-cell adoptive transfer is protective in both murine models of PDA and human organotypic systems. We show that iαßT cells induce a CCR5-dependent immunogenic macrophage reprogramming, thereby enabling marked CD4+ and CD8+ T-cell expansion/activation and tumor protection. Collectively, iαßTs govern fundamental intratumoral cross-talk between innate and adaptive immune populations and are attractive therapeutic targets. SIGNIFICANCE: We found that iαßTs are a profoundly activated T-cell subset in PDA that slow tumor growth in murine and human models of disease. iαßTs induce a CCR5-dependent immunogenic tumor-associated macrophage program, T-cell activation and expansion, and should be considered as novel targets for immunotherapy.See related commentary by Banerjee et al., p. 1164.This article is highlighted in the In This Issue feature, p. 1143.

Biology Needs Evolutionary Software Tools: Let's Build Them Right.

Research in population genetics and evolutionary biology has always provided a computational backbone for life sciences as a whole. Today evolutionary and population biology reasoning are essential for interpretation of large complex datasets that are characteristic of all domains of today's life sciences ranging from cancer biology to microbial ecology. This situation makes algorithms and software tools developed by our community more important than ever before. This means that we, developers of software tool for molecular evolutionary analyses, now have a shared responsibility to make these tools accessible using modern technological developments as well as provide adequate documentation and training.

Open pipelines for integrated tumor genome profiles reveal differences between pancreatic cancer tumors and cell lines.

We describe open, reproducible pipelines that create an integrated genomic profile of a cancer and use the profile to find mutations associated with disease and potentially useful drugs. These pipelines analyze high-throughput cancer exome and transcriptome sequence data together with public databases to find relevant mutations and drugs. The three pipelines that we have developed are: (1) an exome analysis pipeline, which uses whole or targeted tumor exome sequence data to produce a list of putative variants (no matched normal data are needed); (2) a transcriptome analysis pipeline that processes whole tumor transcriptome sequence (RNA-seq) data to compute gene expression and find potential gene fusions; and (3) an integrated variant analysis pipeline that uses the tumor variants from the exome pipeline and tumor gene expression from the transcriptome pipeline to identify deleterious and druggable mutations in all genes and in highly expressed genes. These pipelines are integrated into the popular Web platform Galaxy at http://usegalaxy.org/cancer to make them accessible and reproducible, thereby providing an approach for doing standardized, distributed analyses in clinical studies. We have used our pipeline to identify similarities and differences between pancreatic adenocarcinoma cancer cell lines and primary tumors.