NaroNet: Discovery of tumor microenvironment elements from highly multiplexed images

Understanding the spatial interactions between the elements of the tumor microenvironment -i.e. tumor cells. fibroblasts, immune cells- and how these interactions relate to the diagnosis or prognosis of a tumor is one of the goals of computational pathology. We present NaroNet, a deep learning framework that models the multi-scale tumor microenvironment from multiplex-stained cancer tissue images and provides patient-level interpretable predictions using a seamless end-to-end learning pipeline. Trained only with multiplex-stained tissue images and their corresponding patient-level clinical labels, NaroNet unsupervisedly learns which cell phenotypes, cell neighborhoods, and neighborhood interactions have the highest influence to predict the correct label. To this end, NaroNet incorporates several novel and state-of-the-art deep learning techniques, such as patch-level contrastive learning, multi-level graph embeddings, a novel max-sum pooling operation, or a metric that quantifies the relevance that each microenvironment element has in the individual predictions. We validate NaroNet using synthetic data simulating multiplex-immunostained images where a patient label is artificially associated to the -adjustable- probabilistic incidence of different microenvironment elements. We then apply our model to two sets of images of human cancer tissues: 336 seven-color multiplex-immunostained images from 12 high-grade endometrial cancer patients; and 382 35-plex mass cytometry images from 215 breast cancer patients. In both synthetic and real datasets, NaroNet provides outstanding predictions of relevant clinical information while associating those predictions to the presence of specific microenvironment elements.

Rhesus Macaque CODEX Multiplexed Immunohistochemistry Panel for Studying Immune Responses During Ebola Infection

Non-human primate (NHP) animal models are an integral part of the drug research and development process. For some biothreat pathogens, animal model challenge studies may offer the only possibility to evaluate medical countermeasure efficacy. A thorough understanding of host immune responses in such NHP models is therefore vital. However, applying antibody-based immune characterization techniques to NHP models requires extensive reagent development for species compatibility. In the case of studies involving high consequence pathogens, further optimization for use of inactivated samples may be required. Here, we describe the first optimized CO-Detection by indEXing (CODEX) multiplexed tissue imaging antibody panel for deep profiling of spatially resolved single-cell immune responses in rhesus macaques. This 21-marker panel is composed of a set of 18 antibodies that stratify major immune cell types along with a set three Ebola virus (EBOV)-specific antibodies. We validated these two sets of markers using immunohistochemistry and CODEX in fully inactivated Formalin-Fixed Paraffin-Embedded (FFPE) tissues from mock and EBOV challenged macaques respectively and provide an efficient framework for orthogonal validation of multiple antibody clones using CODEX multiplexed tissue imaging. We also provide the antibody clones and oligonucleotide tag sequences as a valuable resource for other researchers to recreate this reagent set for future studies of tissue immune responses to EBOV infection and other diseases.

SpatialLeiden: spatially aware Leiden clustering

Clustering can identify the natural structure that is inherent to measured data. For single-cell omics, clustering finds cells with similar molecular phenotype after which cell types are annotated. Leiden clustering is one of the algorithms of choice in the single-cell community. In the field of spatial omics, Leiden is often categorized as a "non-spatial" clustering method. However, we show that by integrating spatial information at various steps Leiden clustering is rendered into a computationally highly performant, spatially aware clustering method that compares well with state-of-the art spatial clustering algorithms.

Spatial-Omics Methods and Applications

Traditional tissue profiling approaches have evolved from bulk studies to single-cell analysis over the last decade; however, the spatial context in tissues and microenvironments has always been lost. Over the last 5 years, spatial technologies have emerged that enabled researchers to investigate tissues in situ for proteins and transcripts without losing anatomy and histology. The field of spatial-omics enables highly multiplexed analysis of biomolecules like RNAs and proteins in their native spatial context-and has matured from initial proof-of-concept studies to a thriving field with widespread applications from basic research to translational and clinical studies. While there has been wide adoption of spatial technologies, there remain challenges with the standardization of methodologies, sample compatibility, throughput, resolution, and ease of use. In this chapter, we discuss the current state of the field and highlight technological advances and limitations.

Integrated workflow for analysis of immune enriched spatial proteomic data with IMmuneCite

Spatial proteomics enable detailed analysis of tissue at single cell resolution. However, creating reliable segmentation masks and assigning accurate cell phenotypes to discrete cellular phenotypes can be challenging. We introduce IMmuneCite, a computational framework for comprehensive image pre-processing and single-cell dataset creation, focused on defining complex immune landscapes when using spatial proteomics platforms. We demonstrate that IMmuneCite facilitates the identification of 32 discrete immune cell phenotypes using data from human liver samples while substantially reducing nonbiological cell clusters arising from co-localization of markers for different cell lineages. We established its versatility and ability to accommodate any antibody panel and different species by applying IMmuneCite to data from murine liver tissue. This approach enabled deep characterization of different functional states in each immune compartment, uncovering key features of the immune microenvironment in clinical liver transplantation and murine hepatocellular carcinoma. In conclusion, we demonstrated that IMmuneCite is a user-friendly, integrated computational platform that facilitates investigation of the immune microenvironment across species, while ensuring the creation of an immune focused, spatially resolved single-cell proteomic dataset to provide high fidelity, biologically relevant analyses.

Multiplexing Autoradiography

Autoradiography, the direct imaging of radioactive distribution in tissue sections, is a powerful technique that has several key advantages for the validation of PET radiotracers. Using autoradiography, we can localize radiotracer uptake to neighbours of cells, and when multiplexed with additional radiotracers, fluorescent probes, or in situ tissue analysis, autoradiography can help to characterize the mechanism of radiotracer uptake and assess functional heterogeneity in tissue. In this chapter, the author outlines the basic ex vivo autoradiography protocol and shows how it can be multiplexed using dual radionuclides 18F and 14C. They also highlight where autoradiography can be combined with other technologies to provide synergistic information for interrogating spatial biology.

Integrating Tissue Microarray to GeoMx® Digital Spatial Profiler : Spatial Transcriptomics Assay with Bioinformatics Analysis

Genome-wide or high-plex gene expression is important to understand the organism, tissue, and cellular mechanism. From microarrays to next-generation sequencing (RNA-Seq) at the bulk level and then to the single cell level, gene expression studies have undergone a long transition. The current bulk gene expression and pathway-centric approach toward disease and therapeutics is moving toward spatial transcriptomics, which makes it possible to profile gene expression from the cellular microenvironment without any loss of spatial information. Spatial transcriptomics allows us to understand cellular interactions, cell type abundance, and profile expression differences between the region of interest and its microenvironment. The technology is revolutionizing oncology, developmental biology, neuroscience, preclinical studies, and many therapeutic approaches, especially immunotherapy. Taking into consideration the diverse spatial transcriptomics technologies available, the current chapter aims to delineate the NGS assay protocol for Digital Spatial Profiler (DSP) and follow bioinformatics analysis. While the workflow itself has been detailed elsewhere, in this chapter we are focusing on the integration of tissue microarrays, bioinformatics pipelines, and statistical approaches specific to GeoMx RNA assays as well as common errors that can occur while running a DSP RNA assay. The descriptions of these methods refer to the current version of the GeoMX DSP guidebook and GeoMx Data Analysis Manual, which can be downloaded from the documents section of NanoString website. These documents and user guides are continuously improved and updated; hence, it is important to regularly check the company's website for the most recent version.

In Situ Hybridization (ISH) Combined with Immunohistochemistry (IHC) for Co-detection of EGFR RNA and Phosphorylated EGFR Protein in Lung Cancer Tissue

Detection of phosphorylated proteins in tissue sections using immunohistochemistry (IHC) is a challenging task. The absence of tissue staining may be caused by either a lack of protein expression or a lack of protein activation via its phosphorylation. To address this problem, we employed Integrated Co-detection Workflow (ICW) protocol to analyze lung cancer tissue sections by combining in situ hybridization (ISH) with IHC. The target protein of interest was epidermal growth factor receptor (EGFR, also known as ErbB1 and HER1) which is the founding member of the ErbB family of receptor tyrosine kinases. Using phospho-specific antibodies specific for a phosphorylated site Y1173 of EGFR molecule allowed us to analyze IHC and ISH staining at a single cell level in lung cancer tissue. We have observed both a co-localization of IHC with ISH signals and ISH-positive cells lacking IHC labeling for phosphorylated EGFR. ICW appears to be a very powerful spatial biology technique for accurate localization of cancer cells with phosphorylated/activated and non-phosphorylated/nonactivated proteins.

Homebuilt Imaging-Based Spatial Transcriptomics: Tertiary Lymphoid Structures as a Case Example

Spatial transcriptomics methods provide insight into the cellular heterogeneity and spatial architecture of complex, multicellular systems. Combining molecular and spatial information provides important clues to study tissue architecture in development and disease. Here, we present a comprehensive do-it-yourself (DIY) guide to perform such experiments at reduced costs leveraging open-source approaches. This guide spans the entire life cycle of a project, from its initial definition to experimental choices, wet lab approaches, instrumentation, and analysis. As a concrete example, we focus on tertiary lymphoid structures (TLS), which we use to develop typical questions that can be addressed by these approaches.

JAK/STAT3 represents a therapeutic target for colorectal cancer patients with stromal-rich tumors

Colorectal cancer (CRC) is a heterogenous malignancy underpinned by dysregulation of cellular signaling pathways. Previous literature has implicated aberrant JAK/STAT3 signal transduction in the development and progression of solid tumors. In this study we investigate the effectiveness of inhibiting JAK/STAT3 in diverse CRC models, establish in which contexts high pathway expression is prognostic and perform in depth analysis underlying phenotypes. In this study we investigated the use of JAK inhibitors for anti-cancer activity in CRC cell lines, mouse model organoids and patient-derived organoids. Immunohistochemical staining of the TransSCOT clinical trial cohort, and 2 independent large retrospective CRC patient cohorts was performed to assess the prognostic value of JAK/STAT3 expression. We performed mutational profiling, bulk RNASeq and NanoString GeoMx® spatial transcriptomics to unravel the underlying biology of aberrant signaling. Inhibition of signal transduction with JAK1/2 but not JAK2/3 inhibitors reduced cell viability in CRC cell lines, mouse, and patient derived organoids (PDOs). In PDOs, reduced Ki67 expression was observed post-treatment. A highly significant association between high JAK/STAT3 expression within tumor cells and reduced cancer-specific survival in patients with high stromal invasion (TSPhigh) was identified across 3 independent CRC patient cohorts, including the TrasnSCOT clinical trial cohort. Patients with high phosphorylated STAT3 (pSTAT3) within the TSPhigh group had higher influx of CD66b + cells and higher tumoral expression of PDL1. Bulk RNAseq of full section tumors showed enrichment of NFκB signaling and hypoxia in these cases. Spatial deconvolution through GeoMx® demonstrated higher expression of checkpoint and hypoxia-associated genes in the tumor (pan-cytokeratin positive) regions, and reduced lymphocyte receptor signaling in the TME (pan-cytokeratin- and αSMA-) and αSMA (pan-cytokeratin- and αSMA +) areas. Non-classical fibroblast signatures were detected across αSMA + regions in cases with high pSTAT3. Therefore, in this study we have shown that inhibition of JAK/STAT3 represents a promising therapeutic strategy for patients with stromal-rich CRC tumors. High expression of JAK/STAT3 proteins within both tumor and stromal cells predicts poor outcomes in CRC, and aberrant signaling is associated with distinct spatially-dependant differential gene expression.

spatialGE: A user-friendly web application to democratize spatial transcriptomics analysis

Spatial transcriptomics (ST) is a powerful tool for understanding tissue biology and disease mechanisms. However, its potential is often underutilized due to the advanced data analysis and programming skills required. To address this, we present spatialGE, a web application that simplifies the analysis of ST data. The application spatialGE provides a user-friendly interface that guides users without programming expertise through various analysis pipelines, including quality control, normalization, domain detection, phenotyping, and multiple spatial analyses. It also enables comparative analysis among samples and supports various ST technologies. We demonstrate the utility of spatialGE through its application in studying the tumor microenvironment of melanoma brain metastasis and Merkel cell carcinoma. Our results highlight the ability of spatialGE to identify spatial gene expression patterns and enrichments, providing valuable insights into the tumor microenvironment and its utility in democratizing ST data analysis for the wider scientific community.

Spatial proteomics of human diabetic kidney disease, from health to class III

AIMS/HYPOTHESIS

Diabetic kidney disease (DKD) is the leading cause of chronic and end-stage kidney disease in the USA and worldwide. Animal models have taught us much about DKD mechanisms, but translation of this knowledge into treatments for human disease has been slowed by the lag in our molecular understanding of human DKD.

METHODS

Using our Spatial TissuE Proteomics (STEP) pipeline (comprising curated human kidney tissues, multiplexed immunofluorescence and powerful analysis tools), we imaged and analysed the expression of 21 proteins in 23 tissue sections from individuals with diabetes and healthy kidneys (n=5), compared to those with DKDIIA, IIA-B and IIB (n=2 each) and DKDIII (n=1).

RESULTS

These analyses revealed the existence of 11 cellular clusters (kidney compartments/cell types): podocytes, glomerular endothelial cells, proximal tubules, distal nephron, peritubular capillaries, blood vessels (endothelial cells and vascular smooth muscle cells), macrophages, myeloid cells, other CD45+ inflammatory cells, basement membrane and the interstitium. DKD progression was associated with co-localised increases in inflammatory cells and collagen IV deposition, with concomitant loss of native proteins of each nephron segment. Cell-type frequency and neighbourhood analyses highlighted a significant increase in inflammatory cells and their adjacency to tubular and αSMA+ (α-smooth muscle actin-positive) cells in DKD. Finally, DKD progression showed marked regional variability within single tissue sections, as well as inter-individual variability within each DKD class.

CONCLUSIONS/INTERPRETATION

Using the STEP pipeline, we found alterations in protein expression, cellular phenotypic composition and microenvironment structure with DKD progression, demonstrating the power of this pipeline to reveal the pathophysiology of human DKD.

Artificial intelligence in surgical pathology - Where do we stand, where do we go?

Surgical and neuropathologists continuously search for new and disease-specific features, such as independent predictors of tumor prognosis or determinants of tumor entities and sub-entities. This is a task where artificial intelligence (AI)/machine learning (ML) systems could significantly contribute to help with tumor outcome prediction and the search for new diagnostic or treatment stratification biomarkers. AI systems are increasingly integrated into routine pathology workflows to improve accuracy, reproducibility, productivity and to reveal difficult-to-see features in complicated histological slides, including the quantification of important markers for tumor grading and staging. In this article, we review the infrastructure needed to facilitate digital and computational pathology. We address the barriers for its full deployment in the clinical setting and describe the use of AI in intraoperative or postoperative settings were frozen or formalin-fixed, paraffin-embedded materials are used. We also summarize quality assessment issues of slide digitization, new spatial biology approaches, and the determination of specific gene-expression from whole slide images. Finally, we highlight new innovative and future technologies, such as large language models, optical biopsies, and mass spectrometry imaging.

Digital spatial profiling identifies the tumor center as a topological niche in prostate cancer characterized by an upregulation of BAD

Prostate cancer is characterized by a high degree of intratumoral heterogeneity. However, little is known about the spatial distribution of cancer cells with respect to specific functional characteristics and the formation of spatial niches. Here, we used digital spatial profiling (DSP) to investigate differences in protein expression in the tumor center versus the tumor periphery. Thirty-seven regions of interest were analyzed for the expression of 47 proteins, which included components of the PI3K-AKT, MAPK, and cell death signaling pathways as well as immune cell markers. A total of 1739 data points were collected from five patients. DSP identified the BCL-2 associated agonist of cell death (BAD) protein as the most significantly upregulated protein in the tumor center. BAD upregulation was confirmed by conventional immunohistochemistry, which furthermore showed a phosphorylation of BAD at serine 112 indicating its inactivation. Knockdown of BAD in prostate cancer cells in vitro led to decreased cell viability and colony growth. Clinically, high BAD expression was associated with a shorter time to biochemical recurrence in 158 mostly high-risk prostate cancer patients. Collectively, our results suggest that the tumor center is a topological niche with high BAD expression that may drive prostate cancer progression.