Immunology

The immune system is an intricate multi-layered defense system of cells, tissues and organs that protect the body against harm. 

Historically, the immune system was primarily recognized for this basic function. However, with the advent of techniques such as single-cell RNA sequencing, CRISPR, AI and spatial omics, scientists have begun to make sense of the complex interplay of immune cells and their role in the development and progression of diseases across almost all disease contexts, ranging from cancer to neuropsychiatry. 

In the oncology world, early adoption of such technologies has enabled researchers to identify novel immune cell functions and states within the tumor microenvironment, as well as how immune, vascular, stromal and cancer cells interact to contribute to tumor immune and therapy responses. Machine learning is enabling an increased understanding of such interactions and functions, and will enable the design of more effective immunotherapies to treat a range of cancers.

We have also entered the era of neuroimmunology, wherein a growing body of evidence has overturned the commonly held notion that the central nervous system is immune privileged. Indeed, the confluence of improved molecular biology and computational techniques has shed light on how immune crosstalk may lead to the progression of several neurologic diseases, such as multiple sclerosis, ALS, Alzheimers, and even  neuropsychiatric disorders. Although the brain remains the most challenging biological system to study, renewed interest across academia and the biotechnology industry has led to promising interventions for previously intractable conditions.

We have only begun to scratch the surface of the immune system; however, it is clear that a robust comprehension of immunology will have profound implications for our understanding of health and disease. 

Given the importance of cytokines as mediators of cell-cell communication in the immune system, we lack a global view of the cellular responses of each immune cell type to each cytokine. To address this, the researchers used novel scRNA and computational techniques to create an ‘Immune Dictionary’, a compendium of >1,400 cytokine-cell type combinations in mouse lymph nodes. This Dictionary revealed previously uncharacterized immune polarization states as well as pleitropic effects, both crucial to our understanding of immune involvement in disease and thus to enable more effective therapeutic exploitation.

Methods: The study utilized single-cell RNA sequencing (scRNA-seq) to profile the responses of over 17 immune cell types to 86 cytokines in vivo. Cytokines were administered to C57BL/6 mice, and skin-draining lymph nodes were collected to analyze cell-specific transcriptomic changes. The study generated a large-scale perturbational dataset, carefully controlling for data quality and batch consistency.

Results: The analysis revealed that most cytokines induce highly cell-type-specific responses. For example, interleukin-1β (IL-1β) triggers distinct gene programs across almost every cell type. The study also identified more than 66 cytokine-driven cellular polarization states, some of which were previously uncharacterized. Furthermore, it developed IREA software to assess cytokine activities and immune cell polarization from gene expression data, showcasing its utility in revealing cytokine networks in tumors following immune checkpoint blockade therapy and in severe COVID-19 cases.

Discussion: The paper discusses the extensive variability and specificity of cytokine responses across different immune cell types, highlighting the importance of considering cell-type-specific effects in immunological studies and therapies. It points out the novel insights into cytokine functions, pleiotropic effects, and immune cell activation states provided by the Immune Dictionary. The development of IREA is emphasized as a significant advancement for interpreting transcriptomic data in the context of immune responses.

This review  explores how merging single-cell and spatial multi-omics data with existing biological datasets can deepen our understanding of the immune system. It underscores the ways in which is integrated approach can elucidate gene regulatory networks, cell differentiation, immune cell interactions and the effects of cytokines, providing a fresh perspective on the immune system’s role in health and disease.

Methods: The paper reviews computational strategies that leverage existing biological knowledge to analyze single-cell and spatial multi-omics data. These methods include:

• Annotation of cells using reference datasets and ontologies.
• Inferring the activity of functional modules, transcription factors (TFs), and gene regulatory networks (GRNs) from single-cell RNA sequencing (scRNA-seq) data.
• Integrating data from assays like ATAC-seq for chromatin accessibility and scRNA-seq to predict context-specific GRNs and TF activity.
• Analyzing cell-cell communication (CCC) through ligand-receptor inference from single-cell transcriptomics, and incorporating spatial multi-omics data to understand spatial organization and interactions within tissues.

Results and Discussion:

• The review presents examples from research on cancer, aging, infectious diseases, and chronic inflammation to illustrate how these computational methods have been applied to gain insights into immune system function.
• It discusses the strengths of using gene sets and databases to estimate the activity of cellular processes and the limitations due to biases in our prior biological knowledge, measurement capabilities at the single-cell level, and assumptions made in computational modeling.
• The paper emphasizes the importance of considering the dynamics, spatial organization, and multicellularity of biological processes in understanding the immune system.
• It highlights ongoing technological and algorithmic developments, such as CRISPR screens, functional proteomics, and emerging spatial multi-omics technologies, that promise to enhance our understanding of the immune system.
• Finally, the paper calls attention to future perspectives, including the development of comprehensive computational models that integrate the dynamics of biological processes and the need for experimental validation of computational hypotheses.

This comprehensive review significantly impacts the field by integrating recent insights into the cancer-immunity cycle, emphasizing the dynamic interplay between cancer cells, the immune system, and the tumor microenvironment (TME). By elucidating how various components of the immune system respond to and are influenced by cancer cells and the TME, the paper sheds light on mechanisms of current immunotherapies and identifies potential new targets for therapeutic intervention. The recognition of tumor immunological phenotypes (immunotypes) and their implications for therapy responsiveness represents a pivotal advancement towards personalized cancer immunotherapy.

Methods: The paper synthesizes a decade of research findings on the cancer-immunity cycle, including single-cell omics technologies, which have advanced our understanding of immune cell types and states within the TME. It incorporates insights into dendritic cells' roles, the significance of the tumor microenvironment, and the iterative nature of anti-cancer immune responses.

Results and Discussion:

• Cancer-Immunity Cycle Revisited: The paper updates the cancer-immunity cycle framework, highlighting the importance of dendritic cells in sustaining anti-tumor immunity and recognizing the TME's role in both supporting and suppressing the immune response.
• Tumor Immunotypes: It introduces tumor immunotypes (immune-inflamed, immune-excluded, immune desert) as a critical factor in understanding and predicting responses to immunotherapy, indicating that the immunological context of tumors can guide therapeutic strategies.
• Significance of Dendritic Cells and the TME: The review elaborates on the evolved understanding of dendritic cells beyond antigen presentation in lymph nodes, emphasizing their role in the tumor itself and the formation of tertiary lymphoid structures (TLS) that can stimulate anti-tumor responses.
• Influence of Fibroblasts and Myeloid Cells: It details how cancer-associated fibroblasts (CAFs) and myeloid cells within the TME contribute to immune suppression, highlighting potential targets for reinvigorating immune responses against tumors.‍
• Emerging Therapeutic Strategies: Discussion extends to the potential of immune checkpoint inhibitors, cancer vaccines, CAR-T cell therapies, and strategies targeting the TME to enhance immunotherapy efficacy.

This review highlights the crucial role of spatial technologies in deciphering the cellular heterogeneity and spatial architecture of the TME. In the era of cancer immunotherapy, understanding how the positioning of immune cells within the tumor landscape influences treatment efficacy is paramount. The paper assesses the strengths, limitations, and future prospects of emerging spatial technologies, emphasizing tumor immunology. It outlines the integration of these technologies with AI and multiomics datasets to offer a comprehensive view of the TME, marking a significant step forward in cancer research and potential therapeutic approaches.

Methods:

• Spatial Technologies: The review explores various spatial technologies, including spatial transcriptomics, proteomics, metabolomics, and their integration to study the TME. It discusses the transition from single-cell analyses to spatially resolved technologies that maintain the spatial context of cellular interactions and functions.
• AI and Multiomics Integration: The paper delves into the application of AI to analyze spatial datasets, emphasizing the value of integrating spatial omics with existing biological knowledge to unravel the TME's complexities.

Results and Discussion:

• Technological Advances: It outlines significant advancements in spatial transcriptomics and proteomics, detailing how these methods have evolved to provide insights into the cellular and molecular architecture of tumors. The review showcases applications of these technologies in understanding immune cell positioning, tumor heterogeneity, and the influence of the TME on disease progression and therapeutic response.
• Challenges and Opportunities: The paper discusses the analytical challenges posed by the wealth of data generated by spatial technologies and the opportunities for integrating this information with AI to achieve a holistic understanding of the TME.
• Clinical Implications: It emphasizes the potential of spatial omics and AI in identifying new therapeutic targets and improving the efficacy of existing treatments. The review suggests that the integration of spatial omics data could lead to better personalized medicine approaches for cancer treatment.
• Future Perspectives: The article concludes with a discussion on the future directions of spatial omics, highlighting the need for further technological advancements, better data integration methods, and the development of AI models that can fully capitalize on the rich datasets generated by spatial technologies.

Historically, the central nervous system was believed to be isolated from the immune system – however, recent work has demonstrated that immune cells infiltrate the CNS, contribute to disease, and are enriched at anatomic borders. This paper interrogates human-CNS interfaces during development, homeostasis, and in glioblastoma, and provides a comprehensive landscape of the myeloid-CNS interface.

Here, the authors combined single-cell RNA sequencing (scRNA-seq), cellular indexing of transcriptomes and epitopes by sequencing (CITE-seq), mass cytometry, and spatial transcriptomics from 102 individuals (356K transcriptomes) to generate a molecular map of the immune system at human CNS interfaces. They find that CNS-associated macrophages (CAMs) have core transcriptional signatures across regions, with variable expression of CD169. Additionally, certain compartments also have unique cell niches (such as Kolmer cells in the choroid plexus). Furthermore, in patients receiving stem cell transplants, peripheral myeloid cells differentially engraft into the various CNS compartments. Interestingly, glioblastoma contained myeloid subpopulations, several of which demonstrated upregulation of genes responsible for iron homeostasis and angiogenesis, similar to those found in fetal myeloid cells. This work is exciting as it establishes an atlas of CNS-immune interfaces, and provides insight into therapeutic avenues for diseases caused by CAM dysfunction.

Neurons were often believed to be the sole cells in the brain responsible for driving behavior. In this paper, the authors demonstrate that a subpopulation of astrocytes – a type of glial, or supportive immune cell cell, in the brain – contribute to repetitive behaviors (perserveration), which are often seen in neurocognitive disorders. Additionally, they identify a gene, Crym, that encodes for the µ-crystallin protein and is intimately linked to such repetitive behavioral patterns. This work adds to the emerging body of literature implicating glial cells in homeostasis and neuropsychiatric disorders.

Crym encodes for the protein µ-crystallin, which has been previously associated with various neurocognitive disorders. The authors probed the expression of µ-crystallin in mice and found that it was densely expressed in the astrocytes of the striatum, and not neurons. After using CRISPR to knock out Crym in mice, the authors found that doing so had no effect on motor control or anxiety, but induced a profound increase in perseveration - repetitive behaviors that have no purpose. Perseveration is noted to be a marked component of a variety of neurocognitive disorders. Subsequently, the authors explored the underlying mechanism driving manifestations of Crym loss in the striatum, and found that this influenced the ratio of excitatory and inhibitory neuronal signals of the lateral orbitofrontal cortex, leading to perseveration. This work elucidates a hitherto unknown mechanism of astrocytes mediating neurological function, and represents new potential means for therapeutic management.

Vitamin B12 (cobalamin) is a critical vitamin critical for DNA synthesis, playing key roles in hematopoiesis and myelination. Vitamin B12 deficiency can lead to a number of conditions, including anemia, intestinal malabsorption, and a suite of neurological issues including reversible dementia and subacute combined degeneration of the spinal cord. In this paper, the authors discover an autoimmune cause of CNS-restricted vitamin B12 deficiency, due to an autoantibody targeting the transcobalamin receptor (CD32). This work identifies a new form of metabolic neurologic disease that may be responsive to immunomodulation and vitamin supplementation.

Using phage immunoprecipitation sequencing (PhIP-seq), the authors were able to screen for autoantibodies from a patient who presented with CNS dysfunction against a library of >700K peptide sequences, and found an autoantibody targeting the B12 transporter (CD320). The patient had nearly undetectable levels of B12 in her CSF. Beyond her, the authors found anti-CD320 autoantibodies in 7 other cases (of neurologic deficits of unknown etiology), all of which targeted the same epitope. Subsequently, they screened 85 grad students and found anti-CD320 autoantibodies in 6% of that population; additionally, they analyzed a cohort of 132 samples from multiple sclerosis patients and similarly found a 6% seropositivity rate. This work is exciting as it reinforces how autoimmunity remains such an unplumbed domain, and plays a critical role across disease manifestations. Plenty of questions still remain about anti-CD320 autoantibodies and their role in CNS B12 deficiency; however, these findings elucidate a previously unknown cause of neurologic disease.

Multiple sclerosis (MS) is a chronic neuroinflammatory disease marked by foci of demyelination and neuronal damage throughout the central nervous system. Microglia are critical resident phagocytic cells in the CNS and have been known to form nodules and phagocytose myelin fragments in MS; however, their role in initiating CNS lesions remained unclear. Here, the authors demonstrate that microglia nodules are associated with more severe MS pathology, and have a differential gene expression profile as compared to those found in stroke. Additionally, they found that certain microglia nodules encapsulate partially demyelinated axons. Together, the authors propose that cytokines and antibodies may activate microglia nodules in MS and may lead to a phenotype with a propensity to MS lesion formation.

Here, the authors compared microglia nodules in multiple sclerosis to those found in stroke and surrounding non-nodular white matter using RNA sequencing of laser-microdissected tissue and immunohistochemistry. They found that microglia nodules in MS (as compared to stroke) were 1) correlated with more severe MS, 2) upregulate expression of certain genes similar to those found within MS, 3) upregulate certain genes associated with immune responses and lymphocytic activation, 4) are closely localized with nearby lymphocytes, 5) contain partially demyelinated axons, 6) are associated with key immunologic drivers, including IgG transcription and complement activation, and 7) had a hypermetabolic phenotype, with increased transcription of pro-inflammatory cytokines and reactive oxygen species genes. Thus, the authors hypothesize that microglia nodules – with a combination of cytokine and immunoglobulin activation along with phagocytosis of oxidized phospholipids – may lead to worsening demyelinating MS lesions, while those found in stroke will not.

Several genetic metabolic syndromes can be treated by hematopoietic stem cell transplantation after myeloablative conditioning. In contrast, although microglia damage has been linked to neurologic disease, bone marrow transplantation has been ineffective in such conditions due to poor cell incorporation into the CNS. Here, the authors developed a promising microglia replacement approach using circulation-derived myeloid cells (CDMCs), which broadly engrafted in the brain, and generated microglia-like cells more effectively than traditional bone marrow transplant. This represents a promising cell replacement approach for treating several neurologic diseases.

Given the variable efficiencies of microglia replacement (7 - 70%) after bone marrow transplant, the authors sought to evaluate modulation of the microglial niche to improve transplant outcomes. They found that inhibition of CSF1R (a critical receptor for microglial survival) – four weeks after bone marrow transplant – significantly improved engraftment and survival of microglia-like cells in the brain (90%+). These cells populated the entire brain and spinal cord, and histologic analyses showed these cells had a microlia-like morphology. Molecular analyses revealed that these incorporated cells had 1296 genes differentially expressed as compared to microglia; further characterization showed these had a distinct morphological and functional state as well, with increased phagocytosis and motility. Additionally, they found that purified hematopoietic stem cells (as compared to whole bone marrow - a heterogeneous population) resulted in the highest degree of chimerism and resultant microglia-like cells in the CNS. Finally, they showed that employing their CSF1R inhibition strategy + HSC transplantation reduced neuroinflammation and improved motor behavior while extending life span in a mouse model of progressive neurodegeneration.

Single-cell spatial omics tools are leading the frontier in analyzing tissue sections in 2D: single-cell spatial transcriptomics, proteomics, epigenomics and metabolomics. These tools are being expanded into 3D by rebuilding tissues via serial section alignment or though tissue-clearing techniques. Changes in spatial measurements over time (spatio-temporal measurements) are being integrated to create a dynamic picture of the immune complement using techniques such as time-controlled barcoding.

1. Advancements in Spatial Transcriptomics:
   1. The evolution from cell suspensions to highly multiplexed in situ techniques like smFISH, seqFISH, and MERFISH, achieving unprecedented resolution in spatial gene expression mapping.
   2. The development of in situ sequencing (ISS) methods, combining the benefits of next-generation sequencing with spatial context.
   3. Commercial platforms such as Visium Spatial Gene Expression and GeoMx Digital Spatial Profiler facilitating broader application in cancer research with their user-friendly technical and analytic pipelines.
2. Rise of Spatial Proteomics:
   1. Spatial proteomics gaining traction as a vital tool for visualizing details within the TME beyond conventional histopathology, driven by methods like multiplexed ion beam imaging (MIBI) and imaging mass cytometry (IMC).
   2. Advances in using FFPE tissues for spatial proteomics, expanding the scope of studies to include a wider range of human tissues.
3. Spatial Analysis of Tumor Metabolism:
   1. Emerging focus on mapping the spatial distribution of metabolites within tissues, with advancements in techniques like TOF-SIMS and MALDI for lipidomics.
   2. Integrative approaches like scSpaMet, combining metabolite imaging with IMC for cell type-specific metabolic profiling.
4. Visualizing Tumor Heterogeneity in Three Dimensions:
   1. Development of 3D spatial technologies like STARmap and ExSeq, enabling detailed exploration of cellular heterogeneity and tissue architecture in three dimensions.
   2. The advent of tissue-clearing techniques combined with high-resolution imaging, offering new avenues for intact tissue analysis in 3D.

The integration of spatial resolution into single-cell datasets has emerged as a powerful approach to unravel the complex architecture of the tumour microenvironment. Multiple tiers of -omics datasets can be integrated to enable a comprehensive analysis across tumour landscapes. However, key factors that pose challenges:

1. Sample preparation and quality
   1. Reliance on freshly frozen samples: many spatial transcriptomic approaches require high-quality, freshly-frozen samples. For many large-scale human studies, only fixed samples are feasible.
   2. Preservation: ensuring spatial integrity and RNA quality is preserved during sample preparation is a challenge.
2. Cell complexity
   1. Rare cells and dissociation sensitivity: technologies may miss rare cell types or cells that cannot withstand dissociation, skewing the representation of cellular populations
   2. Cell segmentation - complex tissue architecture and variations in cell morphology make it difficult to define precise cell boundaries (particularly for cell types with  cell bodies distant from the nucleus (e.g. microglia and neurons)). The accuracy of accurately segmented cells completely dictates the quality of the data outputs. 
   3. Cell lineage identification - there is currently no consensus technique or combination of markers to do this, some approaches analyze complex gene expression patterns, others look at mRNA and protein levels and the characteristics of neighboring cells in the cellular niche. 
3. Batch effects
   1. Heterogeneous single-cell sequencing datasets generated across multiple tissue locations, time and conditions (in different labs) - the integration of such datasets together is essential for unravelling the full complexities of disease and biology, and has been necessary in large efforts such as the Human Cell Atlas. However, the intrinsic differences in measured gene expression among experimental settings contributed by factors such as sequencing protocols, library preparation, sample donors, tissue of origin, sampling time and condition inevitably create complex, nested batch effects that can diminish the value of data integration by confounding biological signals
4. Data integration and interpretation
   1. Multiomics Data Integration: Combining spatial transcriptomics with proteomics, metabolomics, and other data types presents challenges in harmonizing diverse datasets for integrated analysis.
   2. Interpreting Complex Interactions: Deciphering the intricate cellular interactions and microenvironmental structures from spatial datasets requires advanced bioinformatics approaches.
5. Clinical translation
   1. Scaling and Validation: Translating spatial transcriptomics findings into clinical practice requires the development of scalable, cost-effective methodologies validated for clinical relevance.
   2. Feature and Marker Identification: Identifying the most informative spatial features and markers for clinical use poses challenges in reducing complexity while retaining meaningful insights.