Shedding Light on Immunological Research in Lyon, France. IRCI2022 Brings the Scientific World to Lyon and Bridges the Latest Immunological Findings in Cancer and Infection

At the heart of Lyon, France are the Centre International de Recherche en Infectiologie (CIRI) and the Cancer Research Center of Lyon (CRCL). Both institutes were created in the early 2010s and have departments with solidified poles in immunological research. The first Immune Responses in Cancer and Infection (IRCI) meeting was co-organized by CIRI and CRCL and held in Lyon, France in February 2017. During two and a half days, world-renowned experts in the field of immunology in cancer and/or infection came to Lyon and discussed their findings with principal investigators, students, and postdocs from the CIRI and CRCL as well as other research centers. The huge local success and international recognition of IRCI2017 mobilized the CIRI and CRCL to host another meeting 3 y later. With the advent of COVID-19, the second IRCI in 2020 was cancelled, yet leaders of the scientific organizing committee, Julien Marie (CRCL) and Uzma Ayesha Hasan (CIRI), persisted to host an in-person meeting on June 15–17, 2022 at the Ecole Normale Supérieur, Lyon, France (https://irci2022.insight-outside.fr, twitter#IRCI2022).

The meeting centered on six sessions in innate immunity, immune metabolism, immunotherapy and clinical immunology, microbiota, genetics and epigenetics, and infection. IRCI2022 brought together prominent international scientists and clinicians to discuss their recent immunological findings in fundamental and clinical research. The number of participants was limited to increase discussion between leaders in the field and attendees, thus strengthening as well as defining immunological research and therapies in the future. Funding from grant calls, as well as from institutional and corporate sponsors, reduced the registration cost. The scientific committee promoted maximal participation of students and postdocs based on abstract selection for short talks, poster sessions, and teasers. Prizes were awarded for the best poster and judged by committees of invited speakers and scientific committee members. The meeting was sold out prior to the early registration deadline with more than 400 participants, of which 30% came from outside of France. The organization incorporated complete parity for male and female international invited scientists embedded with locally selected Lyonnais speakers from the CIRI or CRCL to exemplify Lyon research. Highly appreciated by all participants was the six sessions of interactive invited and selected scientific talks, three poster sessions with comical poster teasers, and a networking mixer and gala dinner. Keynote talks were given on each day by Yasmine Belkhaid, Antonio Bertoletti, and Ruslan Medzhitov. Furthermore, Editors from Cell and Nature Immunology Reviews participated in a meet-the-editors session.

IRCI2022 increased the visibility of Lyon’s excellence in immunology and stimulated discussion between local junior cancer or infection immunologists and international experts. Most importantly, the low cost of registration and local organization permitted Lyonnais students and postdocs to meet prominent scientists. IRCI2022 was more successful than IRCI2017 and was well received by participants and invited speakers. The call for another meeting has solidified the position of IRCI as a key international immunological symposium. Shown in Table I is a summary of all of the speakers, and below we highlight the events and significant scientific takeaways.

The innate immune and DNA repair systems can induce cell death and tissue injury in some disorders, such as infections and cancers. DNA damage from cellular stress is detected by DNA sensors (such as cyclic GMP-AMP synthase [cGAS] and AIM2-like receptors [ALRs]) and constitutes an intrinsic danger signal (1, 2). Under homeostatic conditions, low levels of cytoplasmic DNA prime the innate immune system via DNA sensors. However, excessive DNA damage increases cytosolic DNA, constituting a false alarm to the immune system and leading to an enhanced constitutive expression of type I IFN (IFN-I).

Nelson O. Gekara (Umeå University, Umeå, Sweden) reported the role of DNA-binding proteins in immune priming, genome stability, and microbiota–host interactions. The ataxia telangiectasia (AT) disease model is caused by the loss of function of the DNA repair kinase AT mutated (ATM). This was used to identify the impact of genomic stress on innate immunity. Already published data showed that AT patients exhibit a basal inflammation characterized by an elevated IFN-I response and antiviral activity that significantly decreases viral replication. Consistent with these patient observations, ATM-deficient mice display a primed innate immune system that can protect against vesicular stomatitis virus infection (3). However, ATM-knockout (KO) mice are also more susceptible to LPS-induced sepsis, with an increase of IFN-β and a decrease in survival as compared with wild-type (WT) mice (4). Thus, a hyperactive innate immune system primed by mutations in ATM provides a protective boost against viruses but remains a pathological condition resulting from neurologic defects. Gekara also described the microbiota as a major extrinsic signal of the innate immune system. The human body contains 100 trillion microbes, mostly composed of commensal bacteria (5). In addition to maintaining homeostasis and providing a competitive barrier to bacterial and fungal pathogens, the commensal microbiota also controls viral infections. Oral administration of antibiotics negatively impacts the microbiota and promotes systemic infection by HSV-1 in WT C57BL/6 mice. (4, 5). Thus, commensal bacteria prime innate immunity via the cGAS–stimulator of IFN genes (STING) pathway, protecting against viruses. Activation of innate DNA sensors has been described to be restricted to the cytoplasm under stress conditions. However, DNA sensors can also be chromatin-bound nuclear proteins (6, 7). Indeed, innate DNA sensors have distinct functions according to their localization. Gekara showed that the nuclear DNA sensors cGAS and ALRs potentiate antitumor effects of DNA damage-based chemotherapy and radiotherapy. He demonstrated that cGAS or ALR-deficient mice are more resistant to bone marrow depletion upon irradiation. In humans, lower expression of cGAS and ALRs correlates with poor survival following chemotherapy (7). Finally, Gekara suggested that DNA sensors could be potential biomarkers or therapeutic targets for radiotherapy/chemotherapy responses.

Clara Taffoni (Institut de Génétique Humaine, Montpellier, France) discussed the role of the cGAS–STING pathways in tumor responses. Tumor-derived DNA accesses the cytoplasm of innate immune cells, such as dendritic cells (DCs), key cells for antitumor responses. As DNA sensors, cGAS and STING are expressed at low levels in cancers (8). Taffoni suggested the discovery of alternatives to target inflammation in a cancer context. Interestingly, she clarified that dsDNA-dependent IFN-I responses were not restricted to DNA sensors and showed the cooperation between cGAS–STING and DNA–protein kinase (PK) pathways in cancer. Activation of DNA-PK complexes (DNA-PKcs) allows DNA repair by nonhomologous end-joining, inhibits cGAS, and activates IFN regulatory factor (IRF)3 (9, 10), leading to the expression of IFN-I–stimulated genes. Taffoni confirmed that the addition of dsDNA in the glioblastoma cell line (T98G), which does not express cGAS, increased phosphorylation of DNA-PK and IRF7 and expression of proinflammatory cytokines (IFN-β and CXCL10). The addition of a DNA-PK inhibitor (NU7441) in the presence of dsDNA in cGAS-deficient cells led to the downregulation of phosphorylated IRF3 and decreased expression of cyclic GMP-AMP, IFN-β, and CXCL10, implicating DNA-PKcs in innate immune responses. However, no evidence is known for a role of DNA-PKcs in inducing inflammatory responses upon genotoxic stress. Taffoni demonstrated that the addition of camptothecin, a topoisomerase inhibitor leading to genotoxic stress, on the T98G glioblastoma cell line induces the same effects (i.e., the upregulation of p-IRF3, IFN-β, STING, and CXCL10), but the DNA-PK inhibitor reversed the effects. Thus, DNA-PKcs induces IFN-I in cGAS-deficient cells and controls IRF3, STING, and cyclic GMP-AMP activation. Taffoni then focused on the role of cGAS in tumorigenesis. The cGAS-expressing glioblastoma volume is significantly reduced across 90 d as compared with nonexpressing cGAS glioblastoma tumors. Also, macrophage type 1 markers (TNF-α and CXCL11) increased, showing the cooperation between cGAS cancer cell expression and macrophage recruitment. Surprisingly, the high expression of PRKDC (DNA-PK) and MB21D1 (cGAS) gene expression in a long time course induces the upregulation of macrophages markers (CD68 and ITGAM) and chemokine production (CXCL10, CCL2, CCL5), leading to the reduction of survival rate (C. Taffoni, J. Marines, H. Chamma, M. Saccas, A. Bouzid, S. Guha, A.-L. Chaves Valadao, K. Polak, M. Del Rio, C. Gongora, et al., manuscript posted on bioRxiv, DOI: 10.1101/2022.06.08.495278). Thus, cGAS expression exhibits antitumor functions early in tumorigenesis through antitumor macrophage recruitment; however, DNA-PKcs and cGAS cooperation sustains inflammation and thus promotes tumorigenesis in the long term.

Other cooperative immune cells present in the tumor microenvironment (TME) have a central role in cancer-related inflammation and responses. Jonathan C. Kagan (Harvard University, Cambridge, MA) highlighted the importance of DC hyperactivators in the activation of antitumor immunity. In particular, the lipid 1-palmitoyl-2-glutaryl-sn-glycero-3-phosphocholine (PGPC), an agonist of DC hyperactivation, drives higher DC migration to the lymph node than other Food and Drug Administration–like adjuvants, such as LPS and Alum (11). Surprisingly, the combination of LPS and PGPC increased the survival rate of B16OVA-injected mice with durable protective immunity against cancer as compared with pyroptosis adjuvants (aluminum or LPS). Kagan also mentioned the first evidence of an antitumor effect of DC inflammasomes upon hyperactivation by an experiment with NLRP3-, Casp1/11-, or CCR7-deficient mice, showing no benefit of the hyperactivation adjuvant in the survival rate. Finally, he concluded that immunizations with DC hyperactivators are the most potent stimulators of Ag-specific CD8⁺ T cells. However, many defects in immunity occur in the elderly population (12), and Kagan underlined the absence of their treatment on cancer therapies. He showed that old mice lose their naive CD8⁺ T cell repertoire, and the absence of CD8⁺ T cells in young mice reduces their survival rate. Protective immunity against cancer rises from CD4⁺ T cells in old mice, whereas CD8⁺ T cells are responsible for antitumor immunity in young mice. Remarkably, DC hyperactivation rearranges the CD4⁺ T cell repertoire in the TME of old mice and generates memory T cells that last into old age, demonstrating that DC hyperactivators such as LPS in combination with PGPC induce protective immunity through memory T cells. In conclusion, DC inflammasomes induce durable and protective CD4⁺ and CD8⁺ T cell responses, and DC hyperactivators can correct age-associated defects in antitumor immunity.

The role of NK cells in tumors was discussed by Nathalie Bendriss-Vermare (CRCL). She reminded the audience about the role of NK functions in immune surveillance at early cancer stages and how cancer cells escape NK cell surveillance during tumorigenesis. Bendriss-Vermare illustrated the importance of the DC/NK cross-talk. Precisely, TLR-activated DCs secrete IL-12 that engages STAT4 signaling in human CD56^dim NK cells, driving the ST2 expression. The addition of anti–IL-12 Ab prevents the emergence of the ST2⁺ NK cell subpopulation. Furthermore, ST2⁺CD56^dim NK cells were shown to display a unique transcriptional signature compared with conventional CD56^bright and CD56^dim NK cells. Interestingly, IL-33 synergizes with IL-12 to trigger polyfunctionality (cytokine production, proliferation, and cytotoxicity) of human ST2⁺CD56^dim NK cells. Interestingly, ST2⁺ NK cells are also found in human tumors and respond to IL-33 and IL-12 stimulation by producing IFN-γ. Using the mouse mammary E0771 tumor model, Bendriss-Vermare demonstrated the NK-dependent antitumor effect and systemic IFN-γ production induced by IL-33 and IL-12 peritumor administration. Thus, IL-33 drives the polyfunctionality of a unique ST2⁺ NK cell state with antitumor activities, infiltrating tumors, and inducing tumor rejection. Finally, a strong NK/IL-33/IFN-γ signature is associated with improved patient survival in breast cancer (unpublished data). Bendriss-Vermare and her team aim to 1) identify the DC subset involved in ST2⁺ NK cell emergence, 2) evaluate the potential of IL-33 to induce memory-like NK cells for adoptive cell transfer approaches, and 3) demonstrate the role of the NK/IL-33/DC axis in tumor control in therapeutic settings as well as in immune surveillance.

Posttranslational modifications, such as linear ubiquitination by OTULIN (OTU deubiquitinase with linear linkage specificity), and their implication on chronic inflammation and diseases were described by Geert Van Loo (Ghent University, Ghent, Belgium). He reported the regulation of LUBAC (linear Ub chain assembly complex)–mediated inflammatory signaling by OTULIN using a mouse model (OTULIN-AlfpCre) of ORAS (OTULIN-related autoinflammatory syndrome). The LUBAC complex is active in inflammatory signaling and cell death and regulated by the cytotoxic TNF, which promotes apoptosis or necroptosis (13). Hepatocyte-specific OTULIN-deficient mice develop liver cancer and hepatitis, dependent on Fas-associated death domain protein (FADD) and receptor-interacting protein kinase (RIPK). Indeed, the absence of OTULIN increases AST, ALT, and ALP levels, inducing a liver pathology, whereas genetic ablation of FADD and RIPK reverses this phenotype. Thus, OTULIN prevents liver inflammation and hepatocellular carcinoma by inhibiting FADD and RIPK1 kinase-mediated hepatocyte apoptosis (14). These observations are also demonstrated in skin inflammation and lesions when OTULIN expression is removed specifically from skin. Interestingly, OTULIN-deficient mouse skin shows enhanced keratinocyte cell death that drives inflammation and skin lesions. The deletion of FADD and MLKL genes in keratinocytes rescues the observed defects (15). Thus, OTULIN maintains skin homeostasis by inhibiting RIPK1 kinase-mediated keratinocyte death. In an unpublished study, Van Loo showed that OTULIN controls intestinal homeostasis by preventing intestinal epithelial cell apoptosis. In conclusion, OTULIN controls LUBAC-mediated inflammatory signaling pathways, such as TNF, and plays a critical role in inhibiting cells from death and preventing chronic tissue inflammation and cancer.

Several speakers discussed the importance of metabolism in controlling immune responses and tumor proliferation. Immune and tumor cells are both highly activated and characterized by high energy and nutrient demands. To sustain their proliferation and functions, these cells engage in anaerobic glycolysis, which is characterized by high glycolysis whereby the pyruvate produced from glucose is converted to lactate instead of being funneled into the mitochondria (Warburg effect) (16). Even though anaerobic glycolysis has been described as a hallmark of tumor cells, Ping-Chih Ho (University of Lausanne, Lausanne, Switzerland) opened this session by reminding the audience that cancer cells use a broad range of metabolic programs, at least in vitro (17), suggesting that uncharacterized mechanisms force tumor cells into anaerobic glycolysis. He postulated that this driving force is the selection pressure exerted by the immune cells on the developing tumor, a process called immunoediting (18). To test the importance of immunoediting on tumor cell metabolism, Ping-Chih Ho and his team used melanoma cells coming from either WT mice (edited tumor cells) or Rag KO mice (unedited tumor cells). Using both seahorse and metabolomic analyses, they showed that unedited tumor cells relied more on oxidative metabolism than did their edited counterpart while producing the same amount of ATP. Furthermore, using T cell depleting Abs at different time points, they showed that this phenotype is observed only when the T cells are present at early time points (between 2 and 5 wk after tumor injection). This study highlights the importance of immunoediting in governing the tumor metabolic switch to aerobic glycolysis, which allows tumor cells to support cell division while promoting an immunosuppressive environment through glucose depletion within the TME. The suspected mechanism relies on IFN-γ production by T cells, which drives STAT3-cMyc activation and promotes metabolic reprogramming in tumor cells.

Upon activation, immune cells also experience metabolic reprogramming to meet their energy needs. Signaling hubs allow the cell to integrate multiple signals such as nutrient availability or immune signals to dictate a cell fate that matches the environmental stimulus. The mTOR pathway, which was first described to regulate cell growth and metabolism, is among the essential signaling pathways for metabolic reprogramming (19). However, a large body of recent evidence shows that mTOR signaling is also fundamental in cell fate decisions. Indeed, several reports show the importance of mTOR in lymphocyte development and function (reviewed in Ref. 20), yet the mechanisms are still poorly characterized. Local speaker Antoine Marçais (CIRI) presented data on mTOR regulation in NK cells by cytokines. He showed that mTOR is activated synergistically by both IL-15 and IL-18, two cytokines used to generate activated NK cells for immunotherapy (21, 22). The combination of IL-15 and IL-18 treatment increased NK cell proliferation and metabolism in an mTOR-dependent manner. Deciphering the pathway involved, he demonstrated that these cytokines activate mTOR through noncanonical pathways, relying on PI3K-AKT and ERK for IL-15 and p38 MAPK for IL-18. Leonid Pobezinsky (University of Massachusetts Amherst, Amherst, MA) presented another unconventional mechanism of mTOR regulation in T cells. He showed that inhibition of microRNA let-7 in CD8⁺ T cells promotes a T cell–exhausted signature characterized, among others, by an increase in PD-1 expression and poor tumor growth control. On the contrary, an increase in let-7 expression leads to a memory signature with decreased PD-1 expression and better control of tumor growth. Comparative transcriptomic analysis of these mutants revealed dysregulation of the mTOR pathway, echoing several reports suggesting a role of microRNA in mTOR regulation (23–25). He provided evidence that let-7 inhibits the mTOR pathway, which in turn controls reactive oxygen species (ROS) production, an essential factor in terminal effector T cell differentiation (18, 26). Treatment of Let-7–deficient cells with rapamycin, an inhibitor of mTOR, or a ROS inhibitor rescued the T cell memory phenotype. These two studies highlight the importance of multiple layers of regulation to control mTOR pathway activity, a key determinant of both NK and T cell fate decisions.

Key pathways such as the mTOR pathway are also largely controlled by nutrients and metabolites in the environment. Helene Poinot (Institute of Pharmaceutical Sciences of Western Switzerland, Geneva, Switzerland) presented a study on the role of endogenous glucocorticoids in renal cancer. Glucocorticoids are molecules known for their anti-inflammatory properties and their role in glucose metabolism. Data from The Cancer Genome Atlas reveals that glucocorticoid pathway enzyme expression correlates with the clinical outcome in renal cancer patients. More precisely, HSD11B1 (11β-hydroxysteroid dehydrogenase type 1) expression, an enzyme that reduces cortisone into its active hormone cortisol, was associated with poor clinical outcome and an immune repression signature. In patients, HSD11B1 is expressed by macrophages in the TME. Using both murine models and human samples, Poinot showed that HSD11B1 decreases Ag-specific T cell activation, inhibiting the immune control of the tumor. Inhibition of HSD11B1 in myeloid immune cells restores the immune activation.

The speakers agreed on the importance of immunometabolic studies for immunotherapy design. Indeed, a better understanding of immune cell metabolism will improve cell culture and activation techniques for adoptive cell therapy. This line of research can also be used to enhance immunotherapy efficiency. In this regard, Helene Poinot concluded with data revealing that HSD11B1 inhibitors increase the efficacy of anti–PD-1 and TLR agonist treatments in a renal tumor model, supporting the idea of combined treatments in immunotherapy.

Immunotherapies have revolutionized cancer care and treatment of infections by reactivating neutralized immune responses (27, 28). In cancer, anti–PD-1/PD-L1 and anti–CTLA-4 Abs, which are known as immune checkpoint blockade (ICB) therapies, have completely changed the prognosis of several malignancies, such as melanoma, and are now used as first-line treatments (27–29). However, we still do not completely understand the function of ICB therapies on the immune response.

Ido Amit (Weizmann Institute of Science, Rehovot, Israel) discussed how the complexity of the immune system has prevented a full understanding of ICB mechanisms of action. For example, the cell types that are targeted by ICB therapies remain unknown. John Wherry (Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA) gave a talk on CD8 T cell exhaustion and addressed some of these questions. He demonstrated that exhausted CD8⁺ T cells (T_EXs), which are supposedly anergic cells, can respond quickly to anti–PD-1 Ab. T_EXs are composed of four intermediate states: T_EX^prog1, T_EX^prog2, T_EX^int, and T_EX^term (30). In mice infected with lymphocytic choriomeningitis virus (LCMV), T_EX^prog1 and T_EX^prog2 responded to the anti–PD-1 treatment. However, prolonged treatment led to terminally exhausted cells (T_EX^terms) and defective clearance of virus. Interestingly, intermittent treatment avoided this pitfall, resulting in better immune responses to LCMV. Therefore, Wherry concluded by stating that one should be cautious about the pharmacodynamic of anti–PD-1 treatment on T_EXs because the current prolonged anti–PD-1 treatment might be suboptimal and induce terminal exhaustion of T cells.

In the context of cancer, Ido Amit argued for another anti–PD-1 target cell population. Using PIC-seq technology (RNA sequencing [RNA-seq] of physically interacting cells) on human non–small cell lung carcinoma samples, a tumor-specific CD4⁺ Th cell population was identified in the TME. These Th tumor cells are primed in tumor-draining lymph nodes and are required for harnessing the antitumor response following anti–PD-1 treatment (31). Lucas Blanchard (Institut de Pharmacologie et Biologie Structurale, Toulouse, France) discussed the role of tumor-associated high endothelial venules (TA-HEVs) in response to anti–PD-1/anti-CTLA-4 combination therapy. TA-HEVs were identified as major sites of lymphocyte entry into tumors both at baseline and during treatment with combined anti–PD-1/anti–CTLA-4 combination therapy (32). Following the combination therapy, HEVs are increased in the tumor bed, suggesting a cross-talk between T cells and TA-HECs. Therapeutically modulating TA-HEVs with an agonist of lymphotoxin β receptor ameliorates the efficacy of combined ICB. TA-HEVs were identified as predictors of response and survival of melanoma patients treated with combined ICB. Mariana Diniz (Division of Infection and Immunity, University College London, London, U.K.) showed that NK cells limit T cell responses to therapeutic vaccination in mouse models of persistent hepatitis B virus (HBV) infection. She identified a population of liver-resident NK cells with high PD-L1 surface expression that inhibit therapeutic vaccine-induced virus-specific CD8 T cells. Blocking PD-L1 on cytokine-activated NK cells allows them to help virus-specific T cells to control HBV infection (33).

Besides ICB, other immunotherapies are emerging and hold great promise. One exciting area of cancer research seeks to identify immunogenic tumor neoantigens that can be used to prime lymphocytes and induce an adaptive immune response. Local selected speaker Stephane Depil (CRCL), founder of the ErVaccine Technologies, characterized human endogenous retrovirus (HERV)–derived epitopes that induce adaptive immune responses and used them as therapeutic strategies (https://www.ervaccinetechnologies.com/). Using in silico studies to explore The Cancer Genome Atlas database, ErVaccine Technologies found HERV epitopes that are shared between different solid tumors and acute myeloid leukemia. These HERV epitopes are immunogenic and induce polyfunctional CD8⁺ T cells that can kill tumor cells (34). HERV-specific CD8⁺ T cells can be found in samples from triple-negative breast cancer, ovarian cancer, and acute myeloid leukemia. Thus, HERV epitopes are alternative tumor-specific Ags, which are shared between patients with the same disease, and these epitopes are therapeutic targets for off-the-shelf cancer vaccines and T cell–based therapies, especially for tumors with a low/moderate tumor mutational burden, which ErVaccine Technologies is trying to develop.

The potential of adoptive transfer of γδ T cells to treat both cancer and CMV infection was discussed by Julie Dechanet-Merville (ImmunoConcEpT, University of Bordeaux, Bordeaux, France). She showed that γδ T cells can recognize common stress-induced self-antigens on tumor and CMV-infected cells. She presented unpublished results demonstrating that HLA-I H chains free of β₂-microglobulin or peptides known as “open conformers” can be targeted by γδ T cells. Open conformers are upregulated during CMV infection, and she hypothesizes that γδ T cells recognize them as stress signals. Encouraging preliminary data suggest that in vitro–amplified Vδ2⁻ γδ T cells may be used as adoptive cell therapy to fight CMV infection and glioblastoma. Mansun Law (The Scripps Research Institute, San Diego, CA) is studying the neutralizing face of a conserved antigenic surface on the E2 envelope glycoprotein of the hepatitis C virus (HCV). The E2 neutralizing face is frequently targeted by broadly neutralizing Abs utilizing the human Ab H chain variable gene IGHV1-69 in HCV patients (35). Structural data of E2:VH1-69 broadly neutralizing Ab complexes presented by Mansun Law demonstrate that it is an excellent opportunity for HCV rational vaccine design and testing.

Effective immunotherapies are saving the lives of many patients. As we have seen during this session of IRCI2022, immunologists are currently developing a better understanding of their mechanisms of action, especially at the cellular level. However, we still cannot predict the patients who will respond to the treatment. Single-cell genomics and transcriptomics will likely characterize how molecules and clinical targets interact in a complex immune response. During Ido Amits’ talk, he presented novel technologies such as spatial transcriptomics, PIC-seq, MARS-seq (massively parallel RNA single cell sequencing), and INs-seq (intracellular staining and sequencing), which allowed him to localize transcriptomic data, identify interactions between cells, and study the intracellular protein activity of cells. Their use in studies on infection, autoimmunity, and cancer have already accelerated our understanding of the immune system and even resulted in clinical trials (36) Ido Amit concluded that in the future, we will be able to create a detailed blueprint of immunity in disease by combining population-scale single-cell genomics with patient records and artificial intelligence.

The alliance between the immune system and the microbiota plays a fundamental role in the maintenance of a symbiotic functional relationship of the host with commensal microbes. More importantly, the microbiota composition and density can modulate immune and inflammatory responses locally and at distal sites in physiological and pathological conditions. This session of the IRCI2022 highlighted the latest findings on the role of the microbiota in antitumor immunity, immunotherapy responses, cancer development, and metabolic disorders.

Thomas Gajewski (The University of Chicago Medicine, Chicago, IL) opened the session by investigating the factors that influence the immune infiltration of the TME as sources of interpatient heterogeneity. His previous work showed that the activity of anti–PD-1 in multiple cancers is associated with the T cell–inflamed TME signature at baseline (37). These factors include tumor cell–intrinsic oncogenic events, such as the Wnt/β-catenin pathway that failed to recruit Batf3-lineage DCs into the tumor site when activated, the composition of the gut microbiota, and polymorphisms in immune regulatory genes. Regarding the commensal microbiota, previous analyses in metastatic melanoma patients revealed that the gut microbiome composition is associated with anti–PD-1 efficacy (38). The establishment of stable colonies of germ-free mice (GFM) reconstituted with responder or nonresponder microbiota has confirmed a causal role for the gut microbiota in regulating immunotherapy efficacy. Moreover, single-cell RNA-seq (scRNA-seq) analysis of the TME of GFM reconstituted with nonresponder microbiota showed a shift to M2 macrophages and granulocyte-like myeloid-derived suppressor cells, and a low M1/M2 gene signature ratio is associated with lack of anti–PD-1 efficiency in melanoma patients (unpublished data). Regarding germline variants, loss-of-function variants of the PKCδ gene were associated with greater immune gene signature score, and an improved tumor control was observed in mice engineered with PKCδ-KO bone marrow. Finally, scRNA-seq of tumor-infiltrating immune cells revealed a shift from M2 to M1 macrophage phenotype in PKCδ-KO mice (unpublished data).

A postdoctoral fellow in the laboratory of Bertrand Routy, Meriem Messaoudene (University of Montreal, Montreal, QC, Canada) first mentioned that the gut microbiome is now part of the “hallmarks of cancer” and represents a novel biomarker of the response to immunotherapies. Hence, antibiotherapy (ATB) can blunt the efficacy of anti–PD-1, and strategies to overcome ATB-related dysbiosis are needed. Messaoudene showed the positive preclinical results in immuno-oncology associated with the clinical trial DAV132-CL-1006 (NCT03710694) that is based on the use of a charcoal adsorbent-based product (DAV132, DaVolterra) that sequesters ATB and prevents microbiota destruction. The results showed that the gut microbiome of healthy human volunteers treated with DAV132+ATB (ceftazidime-avibactam) restores the antitumor response to anti–PD-1 when transferred to tumor-bearing GFM, compared with ATB treatment alone. Then, Messaoudene presented his recent work published last April in Cancer Discovery (39) wherein the authors identified the antitumor activity of the natural polyphenol castalagin, which promoted CD8 T cell infiltration within the TME in mice, as well as its prebiotic potential to circumvent anti–PD-1 resistance in GFM reconstituted with nonresponder microbiota.

Jonathan Schertzer (McMaster University, Hamilton, ON, Canada) aimed to understand the connection between how blood glucose may interact with host–microbe symbiosis in the context of metabolic diseases. The blood glucose level is part of the emerging risk factors of death in several pathological contexts, such as cancer and vascular diseases, and is regulated by postbiotics such as LPS. Interestingly, his laboratory discovered that metabolic endotoxemia (a low-level LPS increase in the blood during metabolic disease) may be beneficial or detrimental depending on the acylation status of lipid A in LPS that dictates changes in blood glucose control in obese mice (40). Underacylated LPS can antagonize other LPS forms, setting up the potential use of different types of LPS as postbiotics to mitigate metabolic inflammation. Independent of obesity, Schertzer showed that hyperglycemia is necessary and sufficient to promote death from enteric infection by Citrobacter rodentium in mice. Mechanistically, hyperglycemia increased activation of the Wnt/β-catenin in the distal gut, leading to diarrhea death, and treatment with Wnt pathway inhibitor improved survival during enteric infection.

Giorgio Trinchieri (Center of Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD) gave an overview of the different studies that identified bacterial species correlating with a favorable clinical response to anti–PD-1. Data from five different U.S. melanoma cohorts, batch correction, and meta-analysis allowed Trinchieri to reveal shared taxonomic patterns associated with favorable and unfavorable response to anti–PD-1 (41). Machine learning approaches trained on the same melanoma patients’ microbiota composition can predict the ability of patients to respond to anti–PD-1 with some accuracy. Regarding underlying mechanisms, a high neutrophil-to-lymphocyte ratio was associated with poor survival after anti–PD-1 and with a fecal microbiota enriched for detrimental taxa, largely Gram-negative. Interestingly, the host (human) fecal transcriptome showed in progressor patients a proinflammatory gene signature from shed myeloid cells (DCs, macrophages, neutrophils) in part dependent on LPS, NF-κB activation, myeloid cell attracting chemokines (CCL2, CCL4, CXCL8), and ROS production (41). Trinchieri ended his presentation by showing results of clinical proof-of-concept studies in anti–PD-1-treated melanoma patients demonstrating that fecal microbiota transplant (42) (NCT03341143) or diet alteration (43) may be used clinically to improve the success rate of cancer immunotherapy.

Lars Vereecke (VIB Center for Inflammation Research, Ghent, Belgium) was the last short talk of the microbiota session, and he discussed the role of host–microbiota interactions during colorectal cancer (CRC) development. Using an original murine model of spontaneous microbiota-dependent invasive CRC (Zeb2^IEC-Tg/+ mice), the Vereecke laboratory identified that the adherent-invasive Escherichia coli pathobiont bacteria (pks+ E. coli CCR20 strain, a human CRC isolate) accelerates CRC disease development, and that loss of the bacterial type 1 pilus adhesins FimH and FmlH prevented such aggravation. Interestingly, in monocolonized Zeb2^IEC-Tg/+ mice, or in a minimal microbiota setting, CCR20 did not promote CRC progression, suggesting that CCR20 requires a complex bacterial ecosystem and adhesion-mediated tissue binding to promote CRC progression. The genotoxic effect of CCR20 was highlighted by an increase of DNA damage responses (γH2AX-positive cells within crypts) in Zeb2^IEC-Tg/+ mice. Finally, Vereecke identified that goblet cell–deficient mice develop spontaneous CRC, suggesting a major role of this population in CRC carcinogenesis.

The genetics/epigenetics session showcased speakers presenting data on the COVID-19 pandemic, especially exploring the genetics of individual susceptibility, which is a major risk factor.

Data suggest that the SARS-CoV2 pandemic originated from Rhinolophus bats in China (44); however, bats do not show pathogenic symptoms from infection. Locally selected speaker Lucie Etienne (CIRI) presented studies on the evolutionary history of SARS-CoV-2 protein molecular interactome in bats and primates containing a human genetic history. Identifying common molecular features and differences of the SARS-CoV-2 interactome between bats and primates might explain the pathogenicity in humans. Using the Detection of Genetic INNovation (DGINN) pipeline, which infers the speed of evolution of a given gene, the evolution of 334 SARS-CoV-2 interacting proteins was studied. Twenty proteins were identified as being positively selected during evolution specifically in bats, 64 specifically in primates, and 17 in both organisms. Three genes captured the attention of Lucie Etienne. TMPRSS2, which is required for the fusion of the viral particle to the cell, evolved quickly in primates but not in bats, suggesting marks of ancient CoV epidemics and lesser involvement of TMPRSS2 in bats. FYCO1 (FYVE and coiled-coil domain autophagy adaptor protein 1) is a protein involved in vesicle formation and autophagy. It evolved under positive selection in primates, but not in bats. FYCO1 polymorphism is associated with severe COVID-19 (45) and may be involved in viral egress and/or in inflammation/pathogenesis. Finally, RIPK1 has been under positive selection in bats but not in humans. This protein is essential for inflammation and cell death/survival. RIPK1 is activated by TNF receptor engagement and can switch between activating the NF-κB pathway or triggering apoptosis (46). Therefore, it can regulate prosurvival or death activities in bats but also induce tolerance to viruses and inflammation. The positive selection of these genes could represent important evolutionary virus/host determinants of symbiosis. Etienne concluded her talk presenting clues that explain the resistance and/or susceptibility of severe COVID-19 illness. A higher virus tolerance and less inflammation in combination with host-specific determinants (such as RIPK1) could be the key to understanding variable immune responses in COVID-19.

Esteban Ballestar (Josep Carreras Institute, Barcelona, Spain) and Anne Puel (Imagine, Laboratory of Human Genetics of Infectious Diseases, New York, NY) linked inborn errors of immune-related genes to increased susceptibility to severe forms of COVID-19. Ballestar presented data on patients with common variable immunodeficiency (CVID), the most frequent symptomatic primary Ab deficiency that is characterized by severe deficiency of switched memory B cells (47). CVIDs are mostly sporadic, and patients develop recurrent and chronic respiratory infections. Because CVID patients do not mobilize neutralizing immune responses to SARS-COV2 vaccines, Ballestar examined the consequences of COVID-19 in CVID patients. Droplet-based scRNA-seq analysis was performed on CVID patients and healthy donors before, during, and after the SARS-CoV2 infection. Impaired IFN-γ and TNF-α production in both T cell and NK cell subsets and impaired activation of the BCR-associated NF-κB pathway in naive B cells was observed in CVID patients as compared with healthy donors. Additionally, an enrichment in IFN-I responses of the monocyte compartment, which is similarly found in severe COVID-19 cases, and persistent hyperactivation of several inflammasome complexes were also observed during convalescence in CVID patients as compared with healthy patients. Future studies will validate and explore the epigenetic dysregulation for the hyperactivation of the inflammasome in SARS-CoV2 patients.

As a member of the international consortium COVID Human Genetic Effort, Puel aims to discover the human genetic and immunological basis of the various clinical forms of SARS-CoV-2 infection. Indeed, severe COVID-19 is associated with epidemiologic factors (48). Analysis of large cohorts of severe and benign COVID-19 patients revealed that ∼3.5% of them carry inborn errors in the IFN-I response pathway, which are known mechanisms of susceptibility to severe viral infections (34). Autoantibodies neutralizing specific cytokines, such as IFN-α/β/ω/γ, IL-6, or IL-17A/F, induce autoimmune phenocopies of inborn errors of immunity (49). Exploring the potential role of autoimmune phenocopies of inborn errors of immunity in susceptibility to severe COVID-19, Puel found that 10% of severe COVID-19 patients display higher titers of IFN-I neutralizing autoantibodies that block the viral protective effect of IFN-α (50). Moreover, neutralizing autoantibodies against IFN-α/β/ω were found in ∼15–20% of patients with critical pneumonia following SARS-CoV2 infection (51). Interestingly, anti–IFN-I autoantibodies sharply increase in patients older than 70 y and are highly present in patients with genetic variants that impair thymic development or thymic T cell selection, such as IPEX (FOXP3 mutation) or APS-1 (AIRE mutation) (50). Therefore, severe COVID-19 appears to be tightly linked with defective thymus T cell selection. Additionally, 80% of patients with NF-κB2 variants presented neutralizing anti–IFN-I autoantibodies and ∼40% of these patients suffered from various viral diseases. Five patients with NF-κB2 mutations were reported to suffer from severe or critical COVID-19 disease.

In addition to the COVID-19 studies, the role of epigenetics in the control of immune cells and pathologies was assessed. Ballestar presented interesting data on a pair of identical twins, one healthy and the other one presenting CVID. They had no identified pathogenic gene variants. Interestingly, defective demethylation occurred during the naive-to-memory B cell transition in CVID, which disrupted B cell function and germinal center reactions, leading to clinical symptoms. Indeed, the dynamics of DNA methylation/demethylation controls the differentiation of B cells, particularly the transition from naive to memory B cells (52). Ballestar demonstrated that CVID patients show impaired DNA demethylation during the naive-to-memory B cell transition and present hypermethylation in B cells as compared with healthy subjects (53). Consequently, CVID patients display a highly heterogeneous memory B cell compartment. Using whole-exome sequencing, Laurent Genestier (CIRI) and his team identified a recurrent Jarid2 depletion in NKT cells in six patients with peripheral T cell lymphoma). Conditional deletion of Jarid2 in mouse T cells (CD4cre⁺Jarid2 ^f/fp53^−/− mice) expanded peripheral NKT cells and rearrangement of the epigenome. Interestingly, the conditional deletion of Jarid2 in mouse T cells increased incidences of T and NKT cell lymphomas, suggesting that Jarid2 is a tumor suppressor gene for T and NKT cells. Jarid2 induced hypermethylation of DNA and closed conformation of the chromatin- to control-specific related genes such as IGF2BP genes.

Julien Marie (CRCL) presented the role of TGF-β in Th17 cell program maintenance with consequences on the development of inflammation-induced cancer. This work revealed that TGF-β, known to play a key role in the differentiation of Th0 cells to Th17 cells, also functions in differentiated Th17 cells. Using mouse models with loss and gain of TGF-β mutations selectively in Th17 cells, this work demonstrated that differentiated Th17 cells need continuous TGF-β signaling to sustain their Th17 program. TGF-β in the gut conditions Th17 cells at an epigenetic level and in a reversible manner. Subsequently, in the absence of TGF-β provided by the gut, Th17 cells lose their Th17 program to become pathogenic cells and promote spontaneous intestinal adenocarcinoma development.

The regulatory importance of CD8⁺ T cells in fighting infection was highlighted by both Lion Uhl (University of Oxford, Oxford, U.K.), who presented how paracrine signaling of IFN-γ mediates T–T cell interaction upon infection, and Carlson Tsui (University of Melbourne, Melbourne, VIC, Australia), who presented unpublished data on the role of c-Myb in regulating CD8⁺ T cell exhaustion in chronic LCMV infection.

The original conception of this meeting in 2020 was for the speakers of this session to highlight the cross-talk between viruses and cancer immunology. However, the conversation quickly turned to a discussion of the SARS-CoV-2 pandemic. Even as the global campaign to contain the SARS-CoV-2 outbreak focused on Ab-mediated vaccine technology, a theme of this session highlighted the often-overlooked T cell response to COVID-19. Departing from her work investigating the role of HBV and hepatocarcinoma, Mala Maini (University College London, London, U.K.) described how she pivoted her research into the pandemic response as one of the lead investigators in a longitudinal study of London healthcare workers (HCWs), who underwent weekly SARS-CoV-2 PCR testing, serum collection for Ab titration, and health questionnaires for self-reported COVID symptoms during the lockdown prior to vaccine distribution. Maini and colleagues were particularly interested in seronegative individuals testing negative via PCR, postulating that these individuals had been exposed to the virus but were clearing the infection subclinically by memory T cell responses to cognate Ag. T cells from seronegative HCWs exhibited strong memory T cell response to the replication-transcription complex (RTC) compared with infected cohort peers, who mounted a reduced memory T cell response that was comparable to control PBMCs collected prior to the start of the COVID19 pandemic (54). Furthermore, these RTC-specific T cells coincided with an upregulation of the blood biomarker IFI27, an early identifier of SARS-CoV-2. These results indicate the presence of an “abortive infection” in seronegative HCWs, suggesting that this cohort of T cells was primed by prior infection by other coronaviruses. Maini emphasized the benefit of expanding cross-reactive T cells targeting the highly conserved early expressed RTC proteins and suggested that a mucosal vaccination strategy expanding RTC-specific T cells could be a useful therapeutic in anticipating future coronavirus outbreaks.

In the context of T cell biology, the answer to how the immune system fine tunes the effector response to non-self threats may lie in γδ T cells, or tissue-resident lymphocytes that recognize and present single Ag in an MHC-independent manner. These γδ T cells play a crucial role in maintaining tissue homeostasis through novel mechanisms of immunosurveillance. One of the pioneers in the field of γδ T cell biology, who founded Gamma Delta Therapeutics, and who recently introduced a novel concept of “normality sensing,” Adrian Hayday (King’s College London, Francis Crick Institute, London, U.K.) presented his study of the mouse epidermis, where Vγ5Vδ1⁺ dendritic epidermal T cell sensory activity was dependent on the expression of PD-L1–like protein Skint1 by keratinocytes in steady state (55). Additionally, Hayday underscored how the capacity of tissue-resident γδ T cells to distinguish normality from pathology extends beyond the mouse model and into human clinical relevance where tissue-resident γδ T cells occupy an antitumoral niche within the TME, with promising prognosis in both breast and lung cancer. His group found an enrichment of Vδ1 T cells in human breast epithelium, which contain an increased signature for antitumoral cytokines such as IFN-γ (56), as well as non–small cell lung cancer, which is notably resistant to immune checkpoint inhibitors (57). In each case there was a clear and significant association with positive clinical outcomes. Hayday emphasized how γδ T cells’ innate-like, non–MHC-restricted effector functions, including cytotoxicity, offer the opportunity (now being realized) of using them as allogeneic therapies derived from healthy donors, as opposed to autologous grafts of αβ T cells from patients themselves. Furthermore, Hayday placed the importance of γδ T cell biology in the context of the COVID-19 infection wherein a longitudinal COVID immunophenotyping study of seropositive patients found an increase in adaptive Vδ1 T cell responses that positively correlated with viral titer. This finding accompanies an overall loss of T cell function (particularly among IFN-I– and IFN-II–producing subsets), alongside a marked increase in IFN-γ–induced protein of 10 kDa (IP-10), a chemokine induced upon viral infection, in COVID-19 seropositive patients as compared with healthy donors (58). These phenotypes in COVID-19 patients not only offer key insights into the innate–T cell cross-talk, emphasizing the importance of γδ T cells that function as intermediates between the two systems, but also introduce a clinical diagnostic tool for estimating the severity of COVID-19 disease.

Interestingly, a key factor in the diagnosis and prevention of severe COVID-19 disease burden is the IFN-I response. Local speaker Sophie Trouillet-Assant (CIRI) highlighted that reduced IFN-I was found in 20% of patients hospitalized with severe COVID-19 clinical presentation. She postulated that the IFN-I response in the nasal mucosa may predict clinical outcomes of infection. Using NanoString RNA-seq technology from blood samples in parallel with FilmArray nested PCR on nasal swabs from patients tested for SARS-CoV2, her group found a positive correlation between IFN-I signature and nasal viral load (59). Emphasizing the data presented earlier in the meeting by Anne Puel, her group observed neutralizing autoantibodies against IFN-I in both blood and nasopharyngeal mucosa (51), suggesting that autoantibodies against both IFN-α₂ and IFN-ω may compromise intranasal antiviral immunity in the primary stages of infection. This hypothesis was supported by work in a model of SARS-CoV-2 infection in human airway epithelium cells where autoantibody blockage of IFN-I led to increases in viral replication in vitro (59). Trouillet-Assant’s findings suggested that a patient’s intranasal IFN-I/III autoantibody signature may serve as a useful diagnostic in predicting SARS-CoV-2 case severity.

The authors have no financial conflicts of interest.