ICELLNET v2: a versatile method for cell-cell communication analysis from human transcriptomic data

SUMMARY

Several methods have been developed in the past years to infer cell-cell communication networks from transcriptomic data based on ligand and receptor expression. Among them, ICELLNET is one of the few approaches to consider the multiple subunits of ligands and receptors complexes to infer and quantify cell communication. In here, we present a major update of ICELLNET. As compared to its original implementation, we (i) drastically expanded the ICELLNET ligand-receptor database from 380 to 1669 biologically curated interactions, (ii) integrated important families of communication molecules involved in immune crosstalk, cell adhesion, and Wnt pathway, (iii) optimized ICELLNET framework for single-cell RNA sequencing data analyses, (iv) provided new visualizations of cell-cell communication results to facilitate prioritization and biological interpretation. This update will broaden the use of ICELLNET by the scientific community in different biological fields.

AVAILABILITY AND IMPLEMENTATION

ICELLNET package is implemented in R. Source code, documentation and tutorials are available on GitHub (https://github.com/soumelis-lab/ICELLNET).

Large-scale analysis of cell-cell communication reveals angiogenin-dependent tumor progression in clear cell renal cell carcinoma

Cellular crosstalk in the tumor microenvironment (TME) is still largely uncharacterized, while it plays an essential role in shaping immunosuppression or anti-tumor response. Large-scale analyses are needed to better decipher cell-cell communication in cancer. In this work, we used original and publicly available single-cell RNA sequencing (scRNAseq) data to characterize in-depth the communication networks in human clear cell renal cell carcinoma (ccRCC). We identified 50 putative communication channels specifically used by cancer cells to interact with other cells, including two novel angiogenin-mediated interactions. Expression of angiogenin and its receptors was validated at the protein level in primary ccRCC. Mechanistically, angiogenin enhanced ccRCC cell line proliferation and down-regulated secretion of IL-6, IL-8, and MCP-1 proinflammatory molecules. This study provides novel biological insights into molecular mechanisms of ccRCC, and suggests angiogenin and its receptors as potential therapeutic targets in clear cell renal cancer.

A mathematical model of Familial Mediterranean Fever predicts mechanisms controlling inflammation

BACKGROUND

Familial Mediterranean Fever (FMF) is a monogenic disease caused by gain-of-function mutations in the MEditerranean FeVer (MEFV) gene. The molecular dysregulations induced by these mutations and the associated causal mechanisms are complex and intricate.

OBJECTIVE

We sought to provide a computational model capturing the mechanistic details of biological pathways involved in FMF physiopathology and enabling the study of the patient's immune cell dynamics.

METHODS

We carried out a literature survey to identify experimental studies published from January 2000 to December 2020, and integrated its results into a molecular map and a mathematical model. Then, we studied the network of molecular interactions and the dynamic of monocytes to identify key players for inflammation phenotype in FMF patients.

RESULTS

We built a molecular map of FMF integrating in a structured manner the current knowledge regarding pathophysiological processes participating in the triggering and perpetuation of the disease flares. The mathematical model derived from the map reproduced patient's monocyte behavior, in particular its proinflammatory role via the Pyrin inflammasome activation. Network analysis and in silico experiments identified NF-κB and JAK1/TYK2 as critical to modulate IL-1β- and IL-18-mediated inflammation.

CONCLUSION

The in silico model of FMF monocyte proved its ability to reproduce in vitro observations. Considering the difficulties related to experimental settings and financial investments to test combinations of stimuli/perturbation in vitro, this model could be used to test complex hypotheses in silico, thus narrowing down the number of in vitro and ex vivo experiments to perform.

Mitochondrial stress in advanced fibrosis and cirrhosis associated with chronic hepatitis B, chronic hepatitis C, or nonalcoholic steatohepatitis

BACKGROUND AND AIMS

Hepatitis B virus (HBV) infection causes oxidative stress (OS) and alters mitochondria in experimental models. Our goal was to investigate whether HBV might alter liver mitochondria also in humans, and the resulting mitochondrial stress might account for the progression of fibrosis in chronic hepatitis B (CHB).

APPROACH AND RESULTS

The study included 146 treatment-naïve CHB mono-infected patients. Patients with CHB and advanced fibrosis (AF) or cirrhosis (F3-F4) were compared to patients with no/mild-moderate fibrosis (F0-F2). Patients with CHB were further compared to patients with chronic hepatitis C (CHC; n = 33), nonalcoholic steatohepatatis (NASH; n = 12), and healthy controls ( n = 24). We detected oxidative damage to mitochondrial DNA (mtDNA), including mtDNA strand beaks, and identified multiple mtDNA deletions in patients with F3-F4 as compared to patients with F0-F2. Alterations in mitochondrial function, mitochondrial unfolded protein response, biogenesis, mitophagy, and liver inflammation were observed in patients with AF or cirrhosis associated with CHB, CHC, and NASH. In vitro , significant increases of the mitochondrial formation of superoxide and peroxynitrite as well as mtDNA damage, nitration of the mitochondrial respiratory chain complexes, and impairment of complex I occurred in HepG2 cells replicating HBV or transiently expressing hepatitits B virus X protein. mtDNA damage and complex I impairment were prevented with the superoxide-scavenging Mito-Tempo or with inducible nitric oxide synthase (iNOS)-specific inhibitor 1400 W.

CONCLUSIONS

Our results emphasized the importance of mitochondrial OS, mtDNA damage, and associated alterations in mitochondrial function and dynamics in AF or cirrhosis in CHB and NASH. Mitochondria might be a target in drug development to stop fibrosis progression.

Meta-Analysis of Human Cancer Single-Cell RNA-Seq Datasets Using the IMMUcan Database

The development of single-cell RNA sequencing (scRNA-seq) technologies has greatly contributed to deciphering the tumor microenvironment (TME). An enormous amount of independent scRNA-seq studies have been published representing a valuable resource that provides opportunities for meta-analysis studies. However, the massive amount of biological information, the marked heterogeneity and variability between studies, and the technical challenges in processing heterogeneous datasets create major bottlenecks for the full exploitation of scRNA-seq data. We have developed IMMUcan scDB (https://immucanscdb.vital-it.ch), a fully integrated scRNA-seq database exclusively dedicated to human cancer and accessible to nonspecialists. IMMUcan scDB encompasses 144 datasets on 56 different cancer types, annotated in 50 fields containing precise clinical, technological, and biological information. A data processing pipeline was developed and organized in four steps: (i) data collection; (ii) data processing (quality control and sample integration); (iii) supervised cell annotation with a cell ontology classifier of the TME; and (iv) interface to analyze TME in a cancer type-specific or global manner. This framework was used to explore datasets across tumor locations in a gene-centric (CXCL13) and cell-centric (B cells) manner as well as to conduct meta-analysis studies such as ranking immune cell types and genes correlated to malignant transformation. This integrated, freely accessible, and user-friendly resource represents an unprecedented level of detailed annotation, offering vast possibilities for downstream exploitation of human cancer scRNA-seq data for discovery and validation studies.

SIGNIFICANCE

The IMMUcan scDB database is an accessible supportive tool to analyze and decipher tumor-associated single-cell RNA sequencing data, allowing researchers to maximally use this data to provide new insights into cancer biology.

Multiobjective semisupervised learning with a right-censored endpoint adapted to the multiple imputation framework

Semisupervised learning aims to use additional knowledge in the search for data structure. In clinical applications, including predictive information in the construction of a data-driven classification is of major importance. This work was motivated by a study that aimed to identify different patterns of immune parameters that would be associated with relapse-free survival in a cohort of breast cancer patients. Supervised and unsupervised objectives can be concomitantly optimized using multiobjective optimization. We propose such a procedure that addresses two challenges in the semisupervised approach, that is, missing data and additional knowledge based on survival time. The former was handled by using multiple imputation and consensus clustering. Survival information was incorporated in the supervised objective through the estimation of a cross-validation error of a Cox regression. A simulation study was performed to assess the performance of the proposed procedure. On complete datasets, the performances were compared to those of an existing modified multiobjective semisupervised learning method. The added value of including the survival data in the learning process was assessed by comparing the procedure to unsupervised learning. The proposed procedure showed better performance than the existing method, notably in the selection of the number of clusters. On incomplete datasets, the procedure showed little sensitivity to most of its parameters, even though a high number of imputations and partition initialization seeds improved the performance. The performance was degraded with a high proportion of missing data (40%) and with more ambiguous data structures. Simulation results and application on real data support the conclusion that our procedure enables the construction of a classification associated with a right-censored endpoint on a possibly incomplete dataset.

GM-CSF-activated human dendritic cells promote Tfh1 cell polarization in a CD40-dependent manner

T follicular helper (Tfh) cells regulate humoral responses and present a marked phenotypic and functional diversity. Tfh1 cells were recently identified and associated with disease severity in infection and autoimmune diseases. The cellular and molecular requirements to induce human Tfh1 differentiation are not known. Here, using single-cell RNA sequencing and protein validation, we report that GM-CSF-activated human blood CD1c+ dendritic cells drive the differentiation of naïve CD4+ T cells into Tfh1. These Tfh1 cells displayed typical Tfh molecular features, including high levels of PD-1, CXCR5, and ICOS. They co-expressed Bcl6 and T-bet, and secreted large amounts of IL-21 and IFN-γ. Mechanistically, GM-CSF triggered the emergence of two DC sub-populations defined by their expression of CD40 and ICOS-ligand, presenting distinct phenotype, morphology, transcriptomic signature, and function. CD40highICOS-Llow DC efficiently induced Tfh1 differentiation in a CD40-dependent manner. In both patients with mild COVID-19 and latent mycobacterium tuberculosis infection, Tfh1 cells were positively correlated with a CD40highICOS-LLow DC signature in single-cell RNA sequencing of PBMC or blood transcriptomics, respectively. Our study uncovered a novel CD40-dependent Tfh1 axis with potential physiopathological relevance to infection.

Lipid-associated macrophages are induced by cancer-associated fibroblasts and mediate immune suppression in breast cancer

Tumor-associated macrophages (TAM) play a detrimental role in triple-negative breast cancer (TNBC). In-depth analysis of TAM characteristics and interactions with stromal cells, such as cancer-associated fibroblast (CAF), could provide important biological and therapeutic insights. Here we identify at the single-cell level a monocyte-derived-STAB1+TREM2high lipid-associated macrophage (LAM) subpopulation with immune suppressive capacities that is expanded in patients resistant to immune checkpoint blockade (ICB). Genetic depletion of this LAM subset in mice suppressed TNBC tumor growth. Flow cytometry and bulk-RNA sequencing data demonstrated that co-culture with TNBC-derived CAFs led to reprogramming of blood monocytes towards immune suppressive STAB1+TREM2high LAMs, which inhibit T cell activation and proliferation. Cell-to-cell interaction modeling and assays in vitro demonstrated the role of the inflammatory CXCL12-CXCR4 axis in CAF-myeloid cell crosstalk and recruitment of monocytes in tumor sites. Altogether, these data suggest an inflammation model whereby monocytes recruited to the tumor via the CAF-driven CXCL12-CXCR4 axis acquire pro-tumorigenic LAM capacities to support an immunosuppressive microenvironment.

Abstract 2105: PD-L1 high ICOSL low Secretory dendritic cells infiltrate human solid tumors

Background: Human conventional dendritic cells (cDC) are essential for the anti-tumor immune response. Their immunogenic and tolerogenic states in cancer remain controversial. Our objective was to define the range of DC activation states in vitro and to determine their relevance ex vivo in cancer.

Methods: First, in vitro, we activated healthy donor cDC with 16 different stimuli and measured cDC and T cells cytokine and chemokine secretion upon co-culture (n = 130 individual experiments) to define new maturation archetypes. Second, we established the relevance of these archetypes in cancer, ex vivo in human. We performed in depth analysis of cDC infiltrating head and neck cancer (HNSCC) by flow cytometry (n = 22 patients), transcriptomic analysis (n = 6 patients), and cDC-enriched single-cell transcriptomics (ScRNAseq) (n = 10503 cells from 2 patients). Then we performed a merged analysis of cDC from in house and public ScRNAseq datasets (n = 36 tumor samples from HNSCC, Lung, and Breast cancer, and, as comparator, n = 33 blood and juxtatumor samples, n = 19 healthy donor samples, and n = 23 samples from inflammatory diseases).

Results: In vitro, we identified two archetypes of cDC activation: (i) PD-L1highICOSLlow cDC that produce large amounts of inflammatory cytokines and chemokines, labeled Secretory cDC; (ii)PD-L1lowICOSLhigh cDC that induce a broad range of Th cytokines, labeled Helper cDC. This functional dichotomy of cDC was mutually exclusive and not receptor specific. In cancer, we identified Secretory cDC, and these cells aligned with mature LAMP3+cDC. Helper cDC were not found ex vivo. Secretory cDC simultaneously expressed stimulatory (CD40, TNFRSF9 (4-1BB)), inhibitory (CD274, CD200, IDO1), and mixed (PVR) checkpoints. Secretory cDC were associated with T cell inflammation in HNSCC, triple-negative breast cancer (TNBC), and melanoma (all p < 1.10-15), with improved overall survival in HNSCC and TNBC (all p < 0.01), and with response to checkpoint blockade in 2 melanoma cohorts (all p < 0.01).

Conclusion: We identify and characterize Secretory cDC in several solid cancers. This novel phenotypic and functional dichotomy of human cDC activation states has broad implications for all types of immunotherapies and provides rational for drug combinations.

Citation Format: Caroline Hoffmann, Floriane Noël, Maximilien Grandclaudon, Lucile Massenet-Regad, Paula Michea, Philemon Sirven, Lilith Faucheux, Aurore Surun, Olivier Lantz, Mylene Bohec, Jian Ye, Weihua Guo, Juliette Rochefort, Jerzy Klijanienko, Sylvain Baulande, Charlotte Lecerf, Maud Kamal, Christophe Le Tourneau, Maude Guillot-Delost, Vassili Soumelis. PD-L1 high ICOSL low Secretory dendritic cells infiltrate human solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2105.

Abstract 5011: Meta-analysis of human cancer single cell RNAseq datasets using the fully integrated IMMUcan database

The development of single cell RNA-sequencing (scRNAseq) technologies has greatly contributed to deciphering the tumor microenvironment (TME) landscape, and a wealth of biological data is now publicly accessible. This represents a very valuable resource to researchers in the field, offering a reference for comparison of novel results, as well as opportunities for original meta-analysis studies. However, the massive amount of biological information renders its exploitation difficult in the absence of a well-structured and annotated resource. Marked heterogeneity and variability between studies in terms of cancer type, clinical context, technological platform, data quality, number and type of cells, create additional bottlenecks. We have developed a fully integrated scRNAseq database exclusively dedicated to human cancer. It gathers 134 datasets on 51 different cancer types, annotated in 50 fields containing precise clinical, technological and biological information. We developed an original data processing pipeline organized in 4 steps: first, data collection; second, data processing, which includes quality control, sample integration, cell clustering; third, cell ontology tree of the TME, built and used to annotate the clusters in a supervised and manual manner; and fourth, interface to analyze TME in a cancer type-specific or global manner. This integrated, accessible and user-friendly resource should be of great value to the biomedical community. It represents an unprecedented level of detailed annotation, offering vast possibilities for downstream exploitation of human cancer scRNAseq data for discovery and validation studies. The database is freely accessible at: https://immucanscdb.vital-it.ch.

Citation Format: Jordi Camps, Floriane Noel, Robin Liechti, Lou Goetz, Caroline Hoffmann, Lucile Massenet-Regad, Elise Amblard, Melissa Saichi, Mahmoud M. Ibrahim, Jack Pollard, Jasna Medvedovic, Helge G. Roider, Vassili Soumelis. Meta-analysis of human cancer single cell RNAseq datasets using the fully integrated IMMUcan database [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5011.

Transcriptome-based identification of novel endotypes in adult atopic dermatitis

BACKGROUND

Atopic dermatitis (AD) is a frequent and heterogeneous inflammatory skin disease, for which personalized medicine remains a challenge. High-throughput approaches have improved understanding of the complex pathophysiology of AD. However, a purely data-driven AD classification is still lacking.

METHODS

To address this question, we applied an original unsupervised approach on the largest available transcriptome dataset of AD lesional (n = 82) and healthy (n = 213) skin biopsies.

RESULTS

Taking into account pathological and physiological state, a variance-based filtering revealed 222 AD-specific hyper-variable genes that efficiently classified the AD samples into 4 clusters that turned out to be clinically and biologically distinct. Comparison of gene expressions between clusters identified 3 sets of upregulated genes used to derive metagenes (MGs): MG-I (19 genes) was associated with IL-1 family signaling (including IL-36A and 36G) and skin remodeling, MG-II (23 genes) with negative immune regulation (including IL-34 and 37) and skin architecture, and MG-III (17 genes) with B lymphocyte immunity. Sample clusters differed in terms of disease severity (p = .02) and S. aureus (SA) colonization (p = .02). Cluster 1 contained the most severe AD, highest SA colonization, and overexpressed MG-I. Cluster 2 was characterized by less severe AD, low SA colonization, and high MG-II expression. Cluster 3 included mild AD, mild SA colonization, and mild expression of all MGs. Cluster 4 had the same clinical features as cluster 3 but had hyper-expression of MG-III. Last, we successfully validated our method and results in an independent cohort.

CONCLUSION

Our study revealed unrecognized AD endotypes with specific underlying biological pathways, highlighting novel pathophysiological mechanisms. These data could provide new insights into personalized treatment strategies.

Characterization and functional interrogation of the SARS-CoV-2 RNA interactome

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent of the COVID-19 pandemic, which has led to a devastating global health crisis. The emergence of variants that escape neutralizing responses emphasizes the urgent need to deepen our understanding of SARS-CoV-2 biology. Using a comprehensive identification of RNA-binding proteins (RBPs) by mass spectrometry (ChIRP-MS) approach, we identify 107 high-confidence cellular factors that interact with the SARS-CoV-2 genome during infection. By systematically knocking down their expression in human lung epithelial cells, we find that the majority of the identified RBPs are SARS-CoV-2 proviral factors. In particular, we show that HNRNPA2B1, ILF3, QKI, and SFPQ interact with the SARS-CoV-2 genome and promote viral RNA amplification. Our study provides valuable resources for future investigations into the mechanisms of SARS-CoV-2 replication and the identification of host-centered antiviral therapies.

PD-L1 and ICOSL discriminate human Secretory and Helper dendritic cells in cancer, allergy and autoimmunity

Dendritic cells (DC) are traditionally classified according to their ontogeny and their ability to induce T cell response to antigens, however, the phenotypic and functional state of these cells in cancer does not necessarily align to the conventional categories. Here we show, by using 16 different stimuli in vitro that activated DCs in human blood are phenotypically and functionally dichotomous, and pure cultures of type 2 conventional dendritic cells acquire these states (termed Secretory and Helper) upon appropriate stimuli. PD-L1highICOSLlow Secretory DCs produce large amounts of inflammatory cytokines and chemokines but induce very low levels of T helper (Th) cytokines following co-culturing with T cells. Conversely, PD-L1lowICOSLhigh Helper DCs produce low levels of secreted factors but induce high levels and a broad range of Th cytokines. Secretory DCs bear a single-cell transcriptomic signature indicative of mature migratory LAMP3+ DCs associated with cancer and inflammation. Secretory DCs are linked to good prognosis in head and neck squamous cell carcinoma, and to response to checkpoint blockade in Melanoma. Hence, the functional dichotomy of DCs we describe has both fundamental and translational implications in inflammation and immunotherapy.

Phenotypic and functional states of dendritic cells critically influence the outcome of cancer and inflammation. Authors here show by single cell transcriptomics and in vitro validation assays that dichotomous PD-L1 and ICOSL expression assign dendritic cells to secretory and helper functions, with respective predominance of inflammatory cytokine expression or T helper cytokine induction.

Negative Relationship between Post-Treatment Stromal Tumor-Infiltrating Lymphocyte (TIL) and Survival in Triple-Negative Breast Cancer Patients Treated with Dose-Dense Dose-Intense NeoAdjuvant Chemotherapy

Background: Patients with triple-negative breast cancers (TNBC) have a poor prognosis unless a pathological complete response (pCR) is achieved after neoadjuvant chemotherapy (NAC). Few studies have analyzed changes in TIL levels following dose-dense dose-intense (dd-di) NAC. Patients and methods: From 2009 to 2018, 117 patients with TNBC received dd-di NAC at our institution. We aimed to identify factors associated with pre- and post-NAC TIL levels, and oncological outcomes relapse-free survival (RFS), and overall survival (OS). Results: Median pre-NAC and post-NAC TIL levels were 15% and 3%, respectively. Change in TIL levels with treatment was significantly correlated with metabolic response (SUV) and pCR. High post-NAC TIL levels were associated with a weak metabolic response after two cycles of NAC, with the presence of residual disease and nodal involvement at NAC completion. In multivariate analyses, high post-NAC TIL levels independently predicted poor RFS and poor OS (HR = 1.4 per 10% increment, 95%CI (1.1; 1.9) p = 0.014 and HR = 1.8 per 10% increment 95%CI (1.3−2.3), p < 0.0001, respectively). Conclusion: Our results suggest that TNBC patients with TIL enrichment after NAC are at higher risk of relapse. These patients are potential candidates for adjuvant treatment, such as immunotherapy, in clinical trials.

Hubness reduction improves clustering and trajectory inference in single-cell transcriptomic data

MOTIVATION

Single-cell RNA-seq (scRNAseq) datasets are characterized by large ambient dimensionality, and their analyses can be affected by various manifestations of the dimensionality curse. One of these manifestations is the hubness phenomenon, i.e. existence of data points with surprisingly large incoming connectivity degree in the datapoint neighbourhood graph. Conventional approach to dampen the unwanted effects of high dimension consists in applying drastic dimensionality reduction. It remains unexplored if this step can be avoided thus retaining more information than contained in the low-dimensional projections, by correcting directly hubness.

RESULTS

We investigated hubness in scRNAseq data. We show that hub cells do not represent any visible technical or biological bias. The effect of various hubness reduction methods is investigated with respect to the clustering, trajectory inference and visualization tasks in scRNAseq datasets. We show that hubness reduction generates neighbourhood graphs with properties more suitable for applying machine learning methods; and that it outperforms other state-of-the-art methods for improving neighbourhood graphs. As a consequence, clustering, trajectory inference and visualization perform better, especially for datasets characterized by large intrinsic dimensionality. Hubness is an important phenomenon characterizing data point neighbourhood graphs computed for various types of sequencing datasets. Reducing hubness can be beneficial for the analysis of scRNAseq data with large intrinsic dimensionality in which case it can be an alternative to drastic dimensionality reduction.

AVAILABILITY AND IMPLEMENTATION

The code used to analyze the datasets and produce the figures of this article is available from https://github.com/sysbio-curie/schubness.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A multivariate modeling framework to quantify immune checkpoint context-dependent stimulation on T cells

Cells receive, and adjust to, various stimuli, which function as part of complex microenvironments forming their "context". The possibility that a given context impacts the response to a given stimulus defines "context-dependency" and it explains large parts of the functional variability of physiopathological and pharmacological stimuli. Currently, there is no framework to analyze and quantify context-dependency over multiple contexts and cellular response outputs. We established an experimental system including a stimulus of interest, applied to an immune cell type in several contexts. We studied the function of OX40 ligand (OX40L) on T helper (Th) cell differentiation, in 4 molecular (Th0, Th1, Th2, and Th17) and 11 dendritic cell (DC) contexts (monocyte-derived DC and cDC2 conditions). We measured 17 Th output cytokines in 302 observations, and developed a statistical modeling strategy to quantify OX40L context-dependency. This revealed highly variable context-dependency, depending on the output cytokine and context type itself. Among molecular contexts, Th2 was the most influential on OX40L function. Among DC contexts, the DC type rather than the activating stimuli was dominant in controlling OX40L context-dependency. This work mathematically formalizes the complex determinants of OX40L functionality, and provides a unique framework to decipher and quantify the context-dependent variability of any biomolecule or drug function.

Definition of a novel breast tumor-specific classifier based on secretome analysis

BACKGROUND

During cancer development, the normal tissue microenvironment is shaped by tumorigenic events. Inflammatory mediators and immune cells play a key role during this process. However, which molecular features most specifically characterize the malignant tissue remains poorly explored.

METHODS

Within our institutional tumor microenvironment global analysis (T-MEGA) program, we set a prospective cohort of 422 untreated breast cancer patients. We established a dedicated pipeline to generate supernatants from tumor and juxta-tumor tissue explants and quantify 55 soluble molecules using Luminex or MSD. Those analytes belonged to five molecular families: chemokines, cytokines, growth factors, metalloproteinases, and adipokines.

RESULTS

When looking at tissue specificity, our dataset revealed some breast tumor-specific characteristics, as IL-16, as well as some juxta-tumor-specific secreted molecules, as IL-33. Unsupervised clustering analysis identified groups of molecules that were specific to the breast tumor tissue and displayed a similar secretion behavior. We identified a tumor-specific cluster composed of nine molecules that were secreted fourteen times more in the tumor supernatants than the corresponding juxta-tumor supernatants. This cluster contained, among others, CCL17, CCL22, and CXCL9 and TGF-β1, 2, and 3. The systematic comparison of tumor and juxta-tumor secretome data allowed us to mathematically formalize a novel breast cancer signature composed of 14 molecules that segregated tumors from juxta-tumors, with a sensitivity of 96.8% and a specificity of 96%.

CONCLUSIONS

Our study provides the first breast tumor-specific classifier computed on breast tissue-derived secretome data. Moreover, our T-MEGA cohort dataset is a freely accessible resource to the biomedical community to help advancing scientific knowledge on breast cancer.

Compartmentalized multicellular crosstalk in lymph nodes coordinates the generation of potent cellular and humoral immune responses

Distributed throughout the body, lymph nodes (LNs) constitute an important crossroad where resident and migratory immune cells interact to initiate antigen‐specific immune responses supported by a dynamic 3‐dimensional network of stromal cells, that is, endothelial cells and fibroblastic reticular cells (FRCs). LNs are organized into four major subanatomically separated compartments: the subcapsular sinus (SSC), the paracortex, the cortex, and the medulla. Each compartment is underpinned by particular FRC subsets that physically support LN architecture and delineate functional immune niches by appropriately providing environmental cues, nutrients, and survival factors to the immune cell subsets they interact with. In this review, we discuss how FRCs drive the structural and functional organization of each compartment to give rise to prosperous interactions and coordinate immune cell activities. We also discuss how reciprocal communication makes FRCs and immune cells perfect compatible partners for the generation of potent cellular and humoral immune responses.

Machine Learning Approaches in Study of Multiple Sclerosis Disease Through Magnetic Resonance Images

Multiple sclerosis (MS) is one of the most common autoimmune diseases which is commonly diagnosed and monitored using magnetic resonance imaging (MRI) with a combination of clinical manifestations. The purpose of this review is to highlight the main applications of Machine Learning (ML) models and their performance in the MS field using MRI. We reviewed the articles of the last decade and grouped them based on the applications of ML in MS using MRI data into four categories: 1) Automated diagnosis of MS, 2) Prediction of MS disease progression, 3) Differentiation of MS stages, 4) Differentiation of MS from similar disorders.

X-linked recessive TLR7 deficiency in ~1% of men under 60 years old with life-threatening COVID-19

Autosomal inborn errors of type I IFN immunity and autoantibodies against these cytokines underlie at least 10% of critical COVID-19 pneumonia cases. We report very rare, biochemically deleterious X-linked TLR7 variants in 16 unrelated male individuals aged 7 to 71 years (mean: 36.7 years) from a cohort of 1,202 male patients aged 0.5 to 99 years (mean: 52.9 years) with unexplained critical COVID-19 pneumonia. None of the 331 asymptomatically or mildly infected male individuals aged 1.3 to 102 years (mean: 38.7 years) tested carry such TLR7 variants (p = 3.5 × 10-5). The phenotypes of five hemizygous relatives of index cases infected with SARS-CoV-2 include asymptomatic or mild infection (n=2, 5 and 38 years), or moderate (n=1, 5 years), severe (n=1, 27 years), or critical (n=1, 29 years) pneumonia. Two boys (aged 7 and 12 years) from a cohort of 262 male patients with severe COVID-19 pneumonia (mean: 51.0 years) are hemizygous for a deleterious TLR7 variant. The cumulative allele frequency for deleterious TLR7 variants in the male general population is < 6.5x10-4 We also show that blood B cell lines and myeloid cell subsets from the patients do not respond to TLR7 stimulation, a phenotype rescued by wild-type TLR7 The patients' blood plasmacytoid dendritic cells (pDCs) produce low levels of type I IFNs in response to SARS-CoV-2. Overall, X-linked recessive TLR7 deficiency is a highly penetrant genetic etiology of critical COVID-19 pneumonia, in about 1.8% of male patients below the age of 60 years. Human TLR7 and pDCs are essential for protective type I IFN immunity against SARS-CoV-2 in the respiratory tract.

A computational workflow for the detection of candidate diagnostic biomarkers of Kawasaki disease using time-series gene expression data

Unlike autoimmune diseases, there is no known constitutive and disease-defining biomarker for systemic autoinflammatory diseases (SAIDs). Kawasaki disease (KD) is one of the "undiagnosed" types of SAIDs whose pathogenic mechanism and gene mutation still remain unknown. To address this issue, we have developed a sequential computational workflow which clusters KD patients with similar gene expression profiles across the three different KD phases (Acute, Subacute and Convalescent) and utilizes the resulting clustermap to detect prominent genes that can be used as diagnostic biomarkers for KD. Self-Organizing Maps (SOMs) were employed to cluster patients with similar gene expressions across the three phases through inter-phase and intra-phase clustering. Then, false discovery rate (FDR)-based feature selection was applied to detect genes that significantly deviate across the per-phase clusters. Our results revealed five genes as candidate biomarkers for KD diagnosis, namely, the HLA-DQB1, HLA-DRA, ZBTB48, TNFRSF13C, and CASD1. To our knowledge, these five genes are reported for the first time in the literature. The impact of the discovered genes for KD diagnosis against the known ones was demonstrated by training boosting ensembles (AdaBoost and XGBoost) for KD classification on common platform and cross-platform datasets. The classifiers which were trained on the proposed genes from the common platform data yielded an average increase by 4.40% in accuracy, 5.52% in sensitivity, and 3.57% in specificity than the known genes in the Acute and Subacute phases, followed by a notable increase by 2.30% in accuracy, 2.20% in sensitivity, and 4.70% in specificity in the cross-platform analysis.

Modeling the Th17 and Tregs Paradigm: Implications for Cancer Immunotherapy

CD4 + T cell differentiation is governed by gene regulatory and metabolic networks, with both networks being highly interconnected and able to adapt to external stimuli. Th17 and Tregs differentiation networks play a critical role in cancer, and their balance is affected by the tumor microenvironment (TME). Factors from the TME mediate recruitment and expansion of Th17 cells, but these cells can act with pro or anti-tumor immunity. Tregs cells are also involved in tumor development and progression by inhibiting antitumor immunity and promoting immunoevasion. Due to the complexity of the underlying molecular pathways, the modeling of biological systems has emerged as a promising solution for better understanding both CD4 + T cell differentiation and cancer cell behavior. In this review, we present a context-dependent vision of CD4 + T cell transcriptomic and metabolic network adaptability. We then discuss CD4 + T cell knowledge-based models to extract the regulatory elements of Th17 and Tregs differentiation in multiple CD4 + T cell levels. We highlight the importance of complementing these models with data from omics technologies such as transcriptomics and metabolomics, in order to better delineate existing Th17 and Tregs bifurcation mechanisms. We were able to recompilate promising regulatory components and mechanisms of Th17 and Tregs differentiation under normal conditions, which we then connected with biological evidence in the context of the TME to better understand CD4 + T cell behavior in cancer. From the integration of mechanistic models with omics data, the transcriptomic and metabolomic reprograming of Th17 and Tregs cells can be predicted in new models with potential clinical applications, with special relevance to cancer immunotherapy.

Single-cell RNA sequencing of blood antigen-presenting cells in severe COVID-19 reveals multi-process defects in antiviral immunity

COVID-19 can lead to life-threatening respiratory failure, with increased inflammatory mediators and viral load. Here, we perform single-cell RNA-sequencing to establish a high-resolution map of blood antigen-presenting cells (APCs) in 15 patients with moderate or severe COVID-19 pneumonia, at day 1 and day 4 post admission to intensive care unit or pulmonology department, as well as in 4 healthy donors. We generated a unique dataset of 81,643 APCs, including monocytes and rare dendritic cell (DC) subsets. We uncovered multi-process defects in antiviral immune defence in specific APCs from patients with severe disease: (1) increased pro-apoptotic pathways in plasmacytoid DCs (pDCs, key effectors of antiviral immunity), (2) a decrease of the innate sensors TLR9 and DHX36 in pDCs and CLEC9a⁺ DCs, respectively, (3) downregulation of antiviral interferon-stimulated genes in monocyte subsets and (4) a decrease of major histocompatibility complex (MHC) class II-related genes and MHC class II transactivator activity in cDC1c⁺ DCs, suggesting viral inhibition of antigen presentation. These novel mechanisms may explain patient aggravation and suggest strategies to restore the defective immune defence.

Systems analysis of human T helper17 cell differentiation uncovers distinct time-regulated transcriptional modules

T helper (Th) 17 cells protect from infections and are pathogenic in autoimmunity. While human Th17 cell differentiation has been defined, the global and stepwise transcriptional changes accompanying this process remain uncharacterized. Herein, by performing transcriptome analysis of human Th17 cells, we uncovered three time-regulated modules: early, involving exclusively “signaling pathways” genes; late, characterized by response to infections; and persistent, involving effector immune functions. To assign them an inflammatory or regulatory potential, we compared Th17 cells differentiated in presence or absence of interleukin (IL)-1β, respectively. Most inflammatory genes belong to the persistent module, whereas regulatory genes are lately or persistently induced. Among inflammatory genes, we identified the effector molecules IL17A, IL17F, IL26, IL6, interferon (IFN)G, IFNK, LTA, IL1A, platelet-derived growth factor (PDGF) A and the transcriptional regulators homeodomain-only protein homeobox (HOPX) and sex-determining-region-Y-box (SOX)2, whose expression was independently validated. This study provides an integrative representation of the stepwise human Th17 differentiation program and offers new perspectives toward therapeutic targeting of Th17-related autoimmune diseases.

•

Human Th17 cells are driven by early, late, and persistent transcriptional modules.

•

Human Th17 cells express IL17A, IL17F, IL26, IL6, IFNG, IFNK, LTA, IL1A, and PDGFA.

•

RORC regulates the IL17A, IL17F, IFNG, PDGFA, and IL1A expression in human Th17 cells.

•

HOPX and SOX2 contribute to the expression of IFNG by human Th17 cells.

Microbial and transcriptional differences elucidate atopic dermatitis heterogeneity across skin sites

It is well established that different sites in healthy human skin are colonized by distinct microbial communities due to different physiological conditions. However, few studies have explored microbial heterogeneity between skin sites in diseased skin, such as atopic dermatitis (AD) lesions. To address this issue, we carried out deep analysis of the microbiome and transcriptome in the skin of a large cohort of AD patients and healthy volunteers, comparing two physiologically different sites: upper back and posterior thigh. Microbiome samples and biopsies were obtained from both lesional and nonlesional skin to identify changes related to the disease process. Transcriptome analysis revealed distinct disease-related gene expression profiles depending on anatomical location, with keratinization dominating the transcriptomic signatures in posterior thigh, and lipid metabolism in the upper back. Moreover, we show that relative abundance of Staphylococcus aureus is associated with disease severity in the posterior thigh, but not in the upper back. Our results suggest that AD may select for similar microbes in different anatomical locations-an "AD-like microbiome," but distinct microbial dynamics can still be observed when comparing posterior thigh to upper back. This study highlights the importance of considering the variability across skin sites when studying the development of skin inflammation.

Interplay between SMAD2 and STAT5A is a critical determinant of IL-17A/IL-17F differential expression

Interleukins (IL)-17A and F are critical cytokines in anti-microbial immunity but also contribute to auto-immune pathologies. Recent evidence suggests that they may be differentially produced by T-helper (Th) cells, but the underlying mechanisms remain unknown. To address this question, we built a regulatory graph integrating all reported upstream regulators of IL-17A and F, completed by ChIP-seq data analyses. The resulting regulatory graph encompasses 82 components and 136 regulatory links. The graph was then supplemented by logical rules calibrated with original flow cytometry data using naive CD4⁺ T cells, in conditions inducing IL-17A or IL-17F. The model displays specific stable states corresponding to virtual phenotypes explaining IL-17A and IL-17F differential regulation across eight cytokine stimulatory conditions. Our model analysis points to the transcription factors NFAT2A, STAT5A and SMAD2 as key regulators of the differential expression of IL-17A and IL-17F, with STAT5A controlling IL-17F expression, and an interplay of NFAT2A, STAT5A and SMAD2 controlling IL-17A expression. We experimentally observed that the production of IL-17A was correlated with an increase of SMAD2 transcription, and the expression of IL-17F correlated with an increase of BLIMP-1 transcription, together with an increase of STAT5A expression (mRNA), as predicted by our model. Interestingly, RORγt presumably plays a more determinant role in IL-17A expression as compared to IL-17F expression. In conclusion, we propose the first mechanistic model accounting for the differential expression of IL-17A and F in Th cells, providing a basis to design novel therapeutic interventions in auto-immune and inflammatory diseases.

SARS-CoV-2 induces human plasmacytoid predendritic cell diversification via UNC93B and IRAK4

Several studies have analyzed antiviral immune pathways in late-stage severe COVID-19. However, the initial steps of SARS-CoV-2 antiviral immunity are poorly understood. Here we have isolated primary SARS-CoV-2 viral strains and studied their interaction with human plasmacytoid predendritic cells (pDCs), a key player in antiviral immunity. We show that pDCs are not productively infected by SARS-CoV-2. However, they efficiently diversified into activated P1-, P2-, and P3-pDC effector subsets in response to viral stimulation. They expressed CD80, CD86, CCR7, and OX40 ligand at levels similar to influenza virus-induced activation. They rapidly produced high levels of interferon-α, interferon-λ1, IL-6, IP-10, and IL-8. All major aspects of SARS-CoV-2-induced pDC activation were inhibited by hydroxychloroquine. Mechanistically, SARS-CoV-2-induced pDC activation critically depended on IRAK4 and UNC93B1, as established using pDC from genetically deficient patients. Overall, our data indicate that human pDC are efficiently activated by SARS-CoV-2 particles and may thus contribute to type I IFN-dependent immunity against SARS-CoV-2 infection.

SARS-CoV-2 induces human plasmacytoid pre-dendritic cell diversification via UNC93B and IRAK4

Several studies have analyzed antiviral immune pathways in late-stage severe COVID-19. However, the initial steps of SARS-CoV-2 antiviral immunity are poorly understood. Here, we have isolated primary SARS-CoV-2 viral strains, and studied their interaction with human plasmacytoid pre-dendritic cells (pDC), a key player in antiviral immunity. We show that pDC are not productively infected by SARS-CoV-2. However, they efficiently diversified into activated P1-, P2-, and P3-pDC effector subsets in response to viral stimulation. They expressed CD80, CD86, CCR7, and OX40 ligand at levels similar to influenza virus-induced activation. They rapidly produced high levels of interferon-α, interferon-λ1, IL-6, IP-10, and IL-8. All major aspects of SARS-CoV-2-induced pDC activation were inhibited by hydroxychloroquine. Mechanistically, SARS-CoV-2-induced pDC activation critically depended on IRAK4 and UNC93B1, as established using pDC from genetically deficient patients. Overall, our data indicate that human pDC are efficiently activated by SARS-CoV-2 particles and may thus contribute to type I IFN-dependent immunity against SARS-CoV-2 infection.

Deciphering the combinatorial landscape of immunity

From cellular activation to drug combinations, immunological responses are shaped by the action of multiple stimuli. Synergistic and antagonistic interactions between stimuli play major roles in shaping immune processes. To understand combinatorial regulation, we present the immune Synergistic/Antagonistic Interaction Learner (iSAIL). iSAIL includes a machine learning classifier to map and interpret interactions, a curated compendium of immunological combination treatment datasets, and their global integration into a landscape of ~30,000 interactions. The landscape is mined to reveal combinatorial control of interleukins, checkpoints, and other immune modulators. The resource helps elucidate the modulation of a stimulus by interactions with other cofactors, showing that TNF has strikingly different effects depending on co-stimulators. We discover new functional synergies between TNF and IFNβ controlling dendritic cell-T cell crosstalk. Analysis of laboratory or public combination treatment studies with this user-friendly web-based resource will help resolve the complex role of interaction effects on immune processes.

Inborn errors of type I IFN immunity in patients with life-threatening COVID-19

Clinical outcome upon infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) ranges from silent infection to lethal coronavirus disease 2019 (COVID-19). We have found an enrichment in rare variants predicted to be loss-of-function (LOF) at the 13 human loci known to govern Toll-like receptor 3 (TLR3)- and interferon regulatory factor 7 (IRF7)-dependent type I interferon (IFN) immunity to influenza virus in 659 patients with life-threatening COVID-19 pneumonia relative to 534 subjects with asymptomatic or benign infection. By testing these and other rare variants at these 13 loci, we experimentally defined LOF variants underlying autosomal-recessive or autosomal-dominant deficiencies in 23 patients (3.5%) 17 to 77 years of age. We show that human fibroblasts with mutations affecting this circuit are vulnerable to SARS-CoV-2. Inborn errors of TLR3- and IRF7-dependent type I IFN immunity can underlie life-threatening COVID-19 pneumonia in patients with no prior severe infection.

Abstract 4281: Overcoming preanalytical variabilities for peripheral blood mononuclear cells and exosome profiling in the IMMUcan multicentre study

IMMUcan is a joint public-private partnership to advance our understanding of how tumors and the immune system interact and the effect of therapeutic interventions. The project will collect tumor biopsies and peripheral blood from up to 3,000 patients across 13 European countries for multi-omic analyses.

Here, we assessed the impact of sample collection and processing conditions on the quality of peripheral blood mononuclear cells (PBMCs). We also compared circulating exosomes obtained from plasma and serum to determine the ideal source for the study.

We collected peripheral blood from 9 healthy donors across 3 centers in different countries. PBMC isolation and freezing were either performed locally at 3, 8, 24, 48 and 72 hours delays or alternatively shipped to a designated biobank for central processing at 8, 24, 48 and 72 hours delays. PBMCs were assessed for yield, viability and frequencies of immune cell populations by flow cytometry. Serum vs plasma derived exosomes were compared using nanoparticle tracking analysis for size distribution and immuno-dotblot for protein levels of HER1, HER2, HER3, cMET, P4HA1, S1000A9, PD-L1, and ALIX.

Centralized processing yielded more consistent results with an overall improvement in PBMC numbers and viability. The delay between blood withdrawal and PBMC isolation had a noticeable impact on PBMC quality when it exceeded 24 hours. Between 24 and 48 hours, we observed a significant and donor dependent decrease in viability and a major increase in granulocytic contamination. Relative frequencies of NK cells were the most sensitive to processing delays during the first 24 hours. Despite a slight global decrease in viability at 24 hours, the relative proportions of other major immune subsets remained stable. Regarding exosomes, plasma-exosomes had a significantly higher modal size compared to serum-derived. Particle abundance was minimally higher in the serum-exosomes. There were no significant differences in the levels of most proteins except for HER3 and c-MET, which were significantly upregulated in plasma exosomes.

We opted to centralize PBMC processing to one biobank to minimize variability. As it was logistically impossible to ensure shipments and processing within 24 hours for all recruitment sites, we limited sample inclusion up to 48 hours. This will be accounted for by tracking the number of granulocytes in future CyTOF measurements and including the delay durations in subsequent multi-variate analyses. Exosomes from serum and plasma will be included in the study as the data showed that both sources produced exosomes of adequate quality yet were different enough in protein levels to merit further investigation.

Funding: IMI2 JU grant agreement 821558, supported by EU's Horizon 2020 and EFPIA.

Citation Format: Rami Mustapha, Valentina Pomella, Jose Vicencio, Silvia Lopez-Lastra, Christelle Bahlawane, Wim Ammerlaan, Henoch S. Hong, Marie Morfouace, Caoimhe Vallely-Gilroy, Vassili Soumelis, Emanuela Romano, Sabine Tejpar, Tony Ng, Fay Betsou, IMMUcan consortium. Overcoming preanalytical variabilities for peripheral blood mononuclear cells and exosome profiling in the IMMUcan multicentre study [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4281.

A Quantitative Multivariate Model of Human Dendritic Cell-T Helper Cell Communication

Cell-cell communication involves a large number of molecular signals that function as words of a complex language whose grammar remains mostly unknown. Here, we describe an integrative approach involving (1) protein-level measurement of multiple communication signals coupled to output responses in receiving cells and (2) mathematical modeling to uncover input-output relationships and interactions between signals. Using human dendritic cell (DC)-T helper (Th) cell communication as a model, we measured 36 DC-derived signals and 17 Th cytokines broadly covering Th diversity in 428 observations. We developed a data-driven, computationally validated model capturing 56 already described and 290 potentially novel mechanisms of Th cell specification. By predicting context-dependent behaviors, we demonstrate a new function for IL-12p70 as an inducer of Th17 in an IL-1 signaling context. This work provides a unique resource to decipher the complex combinatorial rules governing DC-Th cell communication and guide their manipulation for vaccine design and immunotherapies.

MMP2 as an independent prognostic stratifier in oral cavity cancers

Background

Around 25% of oral cavity squamous cell carcinoma (OCSCC) are not controlled by the standard of care, but there is currently no validated biomarker to identify those patients. Our objective was to determine a robust biomarker for severe OCSCC, using a biology-driven strategy.

Patients and methods

Tumor and juxtatumor secretome were analyzed in a prospective discovery cohort of 37 OCSCC treated by primary surgery. Independent biomarker validation was performed by RTqPCR in a retrospective cohort of 145 patients with similar clinical features. An 18-gene signature (18 G) predictive of the response to PD-1 blockade was evaluated in the same cohort.

Results

Among 29 deregulated molecules identified in a secretome analysis, including chemokines, cytokines, growth factors, and molecules related to tumor growth and tissue remodeling, only soluble MMP2 was a prognostic biomarker. In our validation cohort, high levels of MMP2 and CD276, and low levels of CXCL10 and STAT1 mRNA were associated with poor prognosis in univariate analysis (Kaplan-Meier). MMP2 (p = .001) and extra-nodal extension (ENE) (p = .006) were independent biomarkers of disease-specific survival (DSS) in multivariate analysis and defined prognostic groups with 5-year DSS ranging from 36% (MMP2highENE+) to 88% (MMP2lowENE-). The expression of 18 G was similar in the different prognostic groups, suggesting comparable responsiveness to anti-PD-1.

Conclusion

High levels of MMP2 were an independent and validated prognostic biomarker, surpassing other molecules of a large panel of the tumor and immune-related processes, which may be used to select poor prognosis patients for intensified neoadjuvant or adjuvant regimens.

Dissecting Effects of Anti-cancer Drugs and Cancer-Associated Fibroblasts by On-Chip Reconstitution of Immunocompetent Tumor Microenvironments

A major challenge in cancer research is the complexity of the tumor microenvironment, which includes the host immunological setting. Inspired by the emerging technology of organ-on-chip, we achieved 3D co-cultures in microfluidic devices (integrating four cell populations: cancer, immune, endothelial, and fibroblasts) to reconstitute ex vivo a human tumor ecosystem (HER2⁺ breast cancer). We visualized and quantified the complex dynamics of this tumor-on-chip, in the absence or in the presence of the drug trastuzumab (Herceptin), a targeted antibody therapy directed against the HER2 receptor. We uncovered the capacity of the drug trastuzumab to specifically promote long cancer-immune interactions (>50 min), recapitulating an anti-tumoral ADCC (antibody-dependent cell-mediated cytotoxicity) immune response. Cancer-associated fibroblasts (CAFs) antagonized the effects of trastuzumab. These observations constitute a proof of concept that tumors-on-chip are powerful platforms to study ex vivo immunocompetent tumor microenvironments, to characterize ecosystem-level drug responses, and to dissect the roles of stromal components.

A multiscale signalling network map of innate immune response in cancer reveals cell heterogeneity signatures

The lack of integrated resources depicting the complexity of the innate immune response in cancer represents a bottleneck for high-throughput data interpretation. To address this challenge, we perform a systematic manual literature mining of molecular mechanisms governing the innate immune response in cancer and represent it as a signalling network map. The cell-type specific signalling maps of macrophages, dendritic cells, myeloid-derived suppressor cells and natural killers are constructed and integrated into a comprehensive meta map of the innate immune response in cancer. The meta-map contains 1466 chemical species as nodes connected by 1084 biochemical reactions, and it is supported by information from 820 articles. The resource helps to interpret single cell RNA-Seq data from macrophages and natural killer cells in metastatic melanoma that reveal different anti- or pro-tumor sub-populations within each cell type. Here, we report a new open source analytic platform that supports data visualisation and interpretation of tumour microenvironment activity in cancer.

The complexity of the innate immune response to cancer makes interpretation of large data sets challenging. Here, the authors provide an integrated multi-scale map of signalling networks representing the different immune cells and their interactions and show its utility for data interpretation.

Microbe-host interplay in atopic dermatitis and psoriasis

Despite recent advances in understanding microbial diversity in skin homeostasis, the relevance of microbial dysbiosis in inflammatory disease is poorly understood. Here we perform a comparative analysis of skin microbial communities coupled to global patterns of cutaneous gene expression in patients with atopic dermatitis or psoriasis. The skin microbiota is analysed by 16S amplicon or whole genome sequencing and the skin transcriptome by microarrays, followed by integration of the data layers. We find that atopic dermatitis and psoriasis can be classified by distinct microbes, which differ from healthy volunteers microbiome composition. Atopic dermatitis is dominated by a single microbe (Staphylococcus aureus), and associated with a disease relevant host transcriptomic signature enriched for skin barrier function, tryptophan metabolism and immune activation. In contrast, psoriasis is characterized by co-occurring communities of microbes with weak associations with disease related gene expression. Our work provides a basis for biomarker discovery and targeted therapies in skin dysbiosis.

Atopic dermatitis (AD) and psoriasis (PSO) are associated with dysbiosis. Here, by analyses of skin microbiome and host transcriptome of AD and PSO patients, the authors find distinct microbial and disease-related gene transcriptomic signatures that differentiate both diseases.

Aberrant fucosylation enables breast cancer clusterin to interact with dendritic cell-specific ICAM-grabbing non-integrin (DC-SIGN)

Clusterin is a glycoprotein able to mediate different physiological functions such as control of complement activation, promotion of unfolded protein clearance and modulation of cell survival. Clusterin is overexpressed in many types of cancers and a large body of evidence suggests that it promotes carcinogenesis and tumor progression. We have previously described a novel clusterin glycoform present in human semen, but not in serum, highly enriched in terminal fucose motifs. Here we show that human luminal breast cancer (LBC) clusterin also bears terminal fucosylated glycans, conferring clusterin the ability to interact with DC-SIGN, a C-type lectin receptor expressed by myeloid cells. This clusterin glycosylation pattern was absent or diminished in non-involved juxtatumoral tissue, suggesting that fucosylated clusterin might represent a cancer associated glycoform. We also found that DC-SIGN is expressed by luminal breast cancer intratumoral macrophages. Moreover, experiments performed in vitro using semen fucosylated clusterin and monocyte derived macrophages showed that the interaction of semen clusterin with DC-SIGN promoted a proangiogenic profile, characterized by a high production of VEGF, IL-8 and TNF-α. Our results reveal an unexpected complexity on the structure and function of secretory clusterin produced by tumors and suggest that fucosylated clusterin produced by luminal breast cancer cells might play a role in tumor progression by promoting the release of pro-angiogenic factors by intratumoral macrophages.

A multivariate Th17 metagene for prognostic stratification in T cell non-inflamed triple negative breast cancer

A diversity of T helper (Th) subsets (Th1, Th2, Th17) has been identified in the human tumor microenvironment. In breast cancer, the role of Th subsets remains controversial, and a systematic study integrating Th subset diversity, T cell inflammation, breast cancer molecular subtypes, and patient prognosis, is lacking. In primary untreated breast cancer samples, we analyzed 19 Th cytokines at the protein level. Eight were T cell-specific, and subsequently measured in 106 prospectively-collected untreated samples. The dominant Th cytokines across all breast cancer samples were IFN-γ and IL-2. Th2 cytokines (IL-4, IL-5, IL-13) were expressed at low levels and not associated with any breast cancer subtype. Th17 cytokines (IL-17A and IL-17F) were up-regulated in triple negative breast cancer (TNBC), specifically in T cell non-inflamed tumors. In order to get insight into prognosis, we exploited the METABRIC transcriptomic dataset. We derived Th1, Th2, and Th17 metagenes based on manually curated Th signatures, and found that a high Th17 metagene was of good prognosis in T cell non-inflamed TNBC. Multivariate Cox modeling selected the Nottingham Prognostic Index (NPI), Th2 and Th17 metagenes as additive predictors of breast cancer-specific survival, which defined novel and highly distinct prognostic groups within TNBC. Our results reveal that Th17 is a novel prognostic composite biomarker in T cell non-inflamed TNBC.

Integrating immune cell and tumor molecular diversity is an efficient strategy for prognostic stratification of cancer patients.

Use of PD-L1 and ICOSL to discriminate innate and adaptive dendritic cells in the head and neck cancer microenvironment

e14184

Background: Inflamed head and neck squamous cell carcinomas (HNSCC), have a better prognosis independently of the treatment modalities, and have an increased response rate to PD(L)1 blockade. Dendritic cells (DC) infiltration is associated to T cell infiltration and DC are critical in the immune response to cancer, but little is known on HNSCC infiltrating DC. The objectives were to determine the differences between DC infiltrating inflamed vs non inflamed tumors, to understand which mechanisms modulate DC phenotypes and which are the associated functions. Methods: 22 untreated HNSCC were characterized by flow cytometry for their immune infiltrate. RNA sequencing of sorted tumor infiltrating myeloid subsets and blood DC were analyzed for differentially expressed genes. The TMod database (Grandclaudon et al.), testing the effect of various stimuli on DC and subsequently on co-cultured T cells, was used to decipher DC and T cell modulation. Results: The level of CD3 infiltration was positively associated to the level of DC infiltration and PDL1 expression on myeloid cells, and negatively to the level neutrophils and of ICOSL expression on myeloid cells. PDL1high ICOSLneg DC from inflamed tumors were matched with the “innate DC” phenotype described in vitro, highly efficient at secreting cytokines and chemokines but poorly activating naïve CD4 T cells. DC from non-inflamed tumor resembled immature medium DC. A third type of DC, “adaptive DC”, expressing high levels of ICOSL and low of PDL1, poorly efficient at cytokine and chemokine secretion, but highly efficient at activating T cells is was described in vitro, but not found in tumors. RNAseq analysis confirmed this innate polarization and highlighted that those cells present a mature and NFkB activated phenotype. Conclusions: Inflamed HNSCC contain stimuli able to induce the recruitment and the differentiation of DC into mature, NFkB activated “innate” DC, highly efficient at secreting cytokines and chemokines but poorly activating naïve CD4 T cells. These data partly explain the superiority of CPS over TPS for the prediction of response to PD(L)1 blockade. More importantly, the stimuli identified here may be used as adjuvants to induce this DC « innate » maturation in non-inflamed tumors, but should only be used in combination with anti-PD(L)1 immunotherapy.

Lenalidomide in combination with intravenous rituximab (REVRI) in relapsed/refractory primary CNS lymphoma or primary intraocular lymphoma: a multicenter prospective 'proof of concept' phase II study of the French Oculo-Cerebral lymphoma (LOC) Network and the Lymphoma Study Association (LYSA)†

BACKGROUND

Primary central nervous system lymphomas (PCNSLs) are mainly diffuse large B-cell lymphomas (DLBCLs) of the non-germinal center B-cell (non-GCB) subtype. This study aimed to determine the efficacy of rituximab plus lenalidomide (R2) in DLBCL-PCNSL.

PATIENTS AND METHODS

Patients with refractory/relapsed (R/R) DLBCL-PCNSL or primary vitreoretinal lymphoma (PVRL) were included in this prospective phase II study. The induction treatment consisted of eight 28-day cycles of R2 (rituximab 375/m2 i.v. D1; lenalidomide 20 mg/day, D1-21 for cycle 1; and 25 mg/day, D1-21 for the subsequent cycles); in responding patients, the induction treatment was followed by a maintenance phase comprising 12 28-day cycles of lenalidomide alone (10 mg/day, D1-21). The primary end point was the overall response rate (ORR) at the end of induction (P0 = 10%; P1 = 30%).

RESULTS

Fifty patients were included. Forty-five patients (PCNSL, N = 34; PVRL, N = 11) were assessable for response. The ORR at the end of induction was 35.6% (95% CI 21.9-51.2) in assessable patients and 32.0% (95% CI 21.9-51.2) in the intent-to-treat analysis, including 13 complete responses (CR)/unconfirmed CR (uCR; 29%) and 3 partial responses (PR; 7%). The best responses were 18 CR/uCR (40%) and 12 PR (27%) during the induction phase. The maintenance phase was started and completed by 18 and 5 patients, respectively. With a median follow-up of 19.2 months (range 1.5-31), the median progression-free survival (PFS) and overall survival (OS) were 7.8 months (95% CI 3.9-11.3) and 17.7 months (95% CI 12.9 to not reached), respectively. No unexpected toxicity was observed. The peripheral baseline CD4/CD8 ratio impacted PFS [median PFS = 9.5 months (95% CI, 8.1-14.8] for CD4/CD8 ≥ 1.6; median PFS = 2.8 months, [95% CI, 1.1-7.8) for CD4/CD8 < 1.6, P = 0.03).

CONCLUSIONS

The R2 regimen showed significant activity in R/R PCNSL and PVRL patients. These results support assessments of the efficacy of R2 combined with methotrexate-based chemotherapy as a first-line treatment of PCNSL.

CLINICAL TRIALS NUMBER

NCT01956695.

TLR1/2 orchestrate human plasmacytoid predendritic cell response to gram+ bacteria

Gram+ infections are worldwide life-threatening diseases in which the pathological role of type I interferon (IFN) has been highlighted. Plasmacytoid predendritic cells (pDCs) produce high amounts of type I IFN following viral sensing. Despite studies suggesting that pDCs respond to bacteria, the mechanisms underlying bacterial sensing in pDCs are unknown. We show here that human primary pDCs express toll-like receptor 1 (TLR1) and 2 (TLR2) and respond to bacterial lipoproteins. We demonstrated that pDCs differentially respond to gram+ bacteria through the TLR1/2 pathway. Notably, up-regulation of costimulatory molecules and pro-inflammatory cytokines was TLR1 dependent, whereas type I IFN secretion was TLR2 dependent. Mechanistically, we demonstrated that these differences relied on diverse signaling pathways activated by TLR1/2. MAPK and NF-κB pathways were engaged by TLR1, whereas the Phosphoinositide 3-kinase (PI3K) pathway was activated by TLR2. This dichotomy was reflected in a different role of TLR2 and TLR1 in pDC priming of naïve cluster of differentiation 4+ (CD4+) T cells, and T helper (Th) cell differentiation. This work provides the rationale to explore and target pDCs in bacterial infection.

Wnt1 silences chemokine genes in dendritic cells and induces adaptive immune resistance in lung adenocarcinoma

Lung adenocarcinoma (LUAD)-derived Wnts increase cancer cell proliferative/stemness potential, but whether they impact the immune microenvironment is unknown. Here we show that LUAD cells use paracrine Wnt1 signaling to induce immune resistance. In TCGA, Wnt1 correlates strongly with tolerogenic genes. In another LUAD cohort, Wnt1 inversely associates with T cell abundance. Altering Wnt1 expression profoundly affects growth of murine lung adenocarcinomas and this is dependent on conventional dendritic cells (cDCs) and T cells. Mechanistically, Wnt1 leads to transcriptional silencing of CC/CXC chemokines in cDCs, T cell exclusion and cross-tolerance. Wnt-target genes are up-regulated in human intratumoral cDCs and decrease upon silencing Wnt1, accompanied by enhanced T cell cytotoxicity. siWnt1-nanoparticles given as single therapy or part of combinatorial immunotherapies act at both arms of the cancer-immune ecosystem to halt tumor growth. Collectively, our studies show that Wnt1 induces immunologically cold tumors through cDCs and highlight its immunotherapeutic targeting.

Biology and prognostic impact of clonal plasmacytoid dendritic cells in chronic myelomonocytic leukemia

Islands of CD123^high cells have been commonly described in the bone marrow of patients with chronic myelomonocytic leukemia (CMML). Using a multiparameter flow cytometry assay, we detected an excess of CD123⁺ mononucleated cells that are lineage-negative, CD45⁺, CD11c^-, CD33^-, HLA-DR⁺, BDCA-2⁺, BDCA-4⁺ in the bone marrow of 32/159 (20%) patients. Conventional and electron microscopy, flow cytometry detection of cell surface markers, gene expression analyses, and the ability to synthesize interferon alpha in response to Toll-like receptor agonists identified these cells as bona fide plasmacytoid dendritic cells (pDCs). Whole-exome sequencing of sorted monocytes and pDCs identified somatic mutations in genes of the oncogenic RAS pathway in the two cell types of every patient. CD34⁺ cells could generate high amount of pDCs in the absence of FMS-like tyrosine kinase 3-ligand (FLT3L). Finally, an excess of pDCs correlates with regulatory T cell accumulation and an increased risk of acute leukemia transformation. These results demonstrate the FLT3L-independent accumulation of clonal pDCs in the bone marrow of CMML patients with mutations affecting the RAS pathway, which is associated with a higher risk of disease progression.