Immune-monitoring of myelodysplastic neoplasms: Recommendations from the i4MDS consortium.

Advancements in comprehending myelodysplastic neoplasms (MDS) have unfolded significantly in recent years, elucidating a myriad of cellular and molecular underpinnings integral to disease progression. While molecular inclusions into prognostic models have substantively advanced risk stratification, recent revelations have emphasized the pivotal role of immune dysregulation within the bone marrow milieu during MDS evolution. Nonetheless, immunotherapy for MDS has not experienced breakthroughs seen in other malignancies, partly attributable to the absence of an immune classification that could stratify patients toward optimally targeted immunotherapeutic approaches. A pivotal obstacle to establishing "immune classes" among MDS patients is the absence of validated accepted immune panels suitable for routine application in clinical laboratories. In response, we formed International Integrative Innovative Immunology for MDS (i4MDS), a consortium of multidisciplinary experts, and created the following recommendations for standardized methodologies to monitor immune responses in MDS. A central goal of i4MDS is the development of an immune score that could be incorporated into current clinical risk stratification models. This position paper first consolidates current knowledge on MDS immunology. Subsequently, in collaboration with clinical and laboratory specialists, we introduce flow cytometry panels and cytokine assays, meticulously devised for clinical laboratories, aiming to monitor the immune status of MDS patients, evaluating both immune fitness and identifying potential immune "risk factors." By amalgamating this immunological characterization data and molecular data, we aim to enhance patient stratification, identify predictive markers for treatment responsiveness, and accelerate the development of systems immunology tools and innovative immunotherapies.

In-depth analysis of human virus-specific CD8+ T cells delineates unique phenotypic signatures for T cell specificity prediction.

Following viral infection, the human immune system generates CD8+ T cell responses to virus antigens that differ in specificity, abundance, and phenotype. A characterization of virus-specific T cell responses allows one to assess infection history and to understand its contribution to protective immunity. Here, we perform in-depth profiling of CD8+ T cells binding to CMV-, EBV-, influenza-, and SARS-CoV-2-derived antigens in peripheral blood samples from 114 healthy donors and 55 cancer patients using high-dimensional mass cytometry and single-cell RNA sequencing. We analyze over 500 antigen-specific T cell responses across six different HLA alleles and observed unique phenotypes of T cells specific for antigens from different virus categories. Using machine learning, we extract phenotypic signatures of antigen-specific T cells, predict virus specificity for bulk CD8+ T cells, and validate these predictions, suggesting that machine learning can be used to accurately predict antigen specificity from T cell phenotypes.

Bystander CD4+ T cells infiltrate human tumors and are phenotypically distinct.

Tumor-specific T cells likely underpin effective immune checkpoint-blockade therapies. Yet, most studies focus on Treg cells and CD8+ tumor-infiltrating lymphocytes (TILs). Here, we study CD4+ TILs in human lung and colorectal cancers and observe that non-Treg CD4+ TILs average more than 70% of total CD4+ TILs in both cancer types. Leveraging high dimensional analyses including mass cytometry, we reveal that CD4+ TILs are phenotypically heterogeneous, within each tumor and across patients. Consistently, we find different subsets of CD4+ TILs showing characteristics of effectors, tissue resident memory (Trm) or exhausted cells (expressing PD-1, CTLA-4 and CD39). In both cancer types, the frequencies of CD39- non-Treg CD4+ TILs strongly correlate with frequencies of CD39- CD8+ TILs, which we and others have previously shown to be enriched for cells specific for cancer-unrelated antigens (bystanders). Ex-vivo, we demonstrate that CD39- CD4+ TILs can be specific for cancer-unrelated antigens, such as HCMV epitopes. Overall, our findings highlight that CD4+ TILs can also recognize cancer-unrelated antigens and suggest measuring CD39 expression as a straightforward way to quantify or isolate bystander CD4+ T cells.

Diagnosis of Myelodysplastic Syndromes: From Immunological Observations to Clinical Applications.

Myelodysplastic syndromes (MDS) constitute a very heterogeneous group of diseases with a high prevalence in elderly patients and a propensity for progression to acute myeloid leukemia. The complexity of these hematopoietic malignancies is revealed by the multiple recurrent somatic mutations involved in MDS pathogenesis and the paradoxical common phenotype observed in these patients characterized by ineffective hematopoiesis and cytopenia. In the context of population aging, the incidence of MDS will strongly increase in the future. Thus, precise diagnosis and evaluation of the progression risk of these diseases are imperative to adapt the treatment. Dysregulations of both innate and adaptive immune systems are frequently detected in MDS patients, and their critical role in MDS pathogenesis is now commonly accepted. However, different immune dysregulations and/or dysfunctions can be dynamically observed during the course of the disease. Monitoring the immune system therefore represents a new attractive tool for a more precise characterization of MDS at diagnosis and for identifying patients who may benefit from immunotherapy. We review here the current knowledge of the critical role of immune dysfunctions in both MDS and MDS precursor conditions and discuss the opportunities offered by the detection of these dysregulations for patient stratification.

NY-ESO-1-specific redirected T cells with endogenous TCR knockdown mediate tumor response and cytokine release syndrome.

BACKGROUND: Because of the shortage of ideal cell surface antigens, the development of T-cell receptor (TCR)-engineered T cells (TCR-T) that target intracellular antigens such as NY-ESO-1 is a promising approach for treating patients with solid tumors. However, endogenous TCRs in vector-transduced T cells have been suggested to impair cell-surface expression of transduced TCR while generating mispaired TCRs that can become self-reactive. METHODS: We conducted a first-in-human phase I clinical trial with the TCR-transduced T-cell product (TBI-1301) in patients with NY-ESO-1-expressing solid tumors. In manufacturing TCR-T cells, we used a novel affinity-enhanced NY-ESO-1-specific TCR that was transduced by a retroviral vector that enables siRNA (small interfering RNA)-mediated silencing of endogenous TCR. The patients were divided into two cohorts. Cohort 1 was given a dose of 5×108 cells (whole cells including TCR-T cells) preconditioned with 1500 mg/m2 cyclophosphamide. Cohort 2 was given 5× 109 cells preconditioned with 1500 mg/m2 cyclophosphamide. RESULTS: In vitro study showed that both the CD8+ and CD4+ T fractions of TCR-T cells exhibited cytotoxic effects against NY-ESO-1-expressing tumor cells. Three patients and six patients were allocated to cohort 1 and cohort 2, respectively. Three of the six patients who received 5×109 cells showed tumor response, while three patients developed early-onset cytokine release syndrome (CRS). One of the patients developed a grade 3 lung injury associated with the infiltration of the TCR-T cells. No siRNA-related adverse events other than CRS were observed. Cytokines including interleukin 6 I and monocyte chemotactic protein-1/chemokine (C-C motif) ligand (CCL2)increased in the sera of patients with CRS. In vitro analysis showed these cytokines were not secreted from the T cells infused. A significant fraction of the manufactured T cells in patients with CRS was found to express either CD244, CD39, or both at high levels. CONCLUSIONS: The trial showed that endogenous TCR-silenced and affinity-enhanced NY-ESO-1 TCR-T cells were safely administered except for grade 3 lung injury. The TCR-T cell infusion exhibited significant tumor response and early-onset CRS in patients with tumors that express NY-ESO-1 at high levels. The differentiation properties of the manufactured T cells may be prognostic for TCR-T-related CRS. TRIAL REGISTRATION NUMBER: NCT02366546.

Characterization of neoantigen-specific T cells in cancer resistant to immune checkpoint therapies.

Neoantigen-specific T cells are strongly implicated as being critical for effective immune checkpoint blockade treatment (ICB) (e.g., anti-PD-1 and anti-CTLA-4) and are being targeted for vaccination-based therapies. However, ICB treatments show uneven responses between patients, and neoantigen vaccination efficiency has yet to be established. Here, we characterize neoantigen-specific CD8+ T cells in a tumor that is resistant to ICB and neoantigen vaccination. Leveraging the use of mass cytometry combined with multiplex major histocompatibility complex (MHC) class I tetramer staining, we screened and identified tumor neoantigen-specific CD8+ T cells in the Lewis Lung carcinoma (LLC) tumor model (mRiok1). We observed an expansion of mRiok1-specific CD8+ tumor-infiltrating lymphocytes (TILs) after ICB targeting PD-1 or CTLA-4 with no sign of tumor regression. The expanded neoantigen-specific CD8+ TILs remained phenotypically and functionally exhausted but displayed cytotoxic characteristics. When combining both ICB treatments, mRiok1-specific CD8+ TILs showed a stem-like phenotype and a higher capacity to produce cytokines, but tumors did not show signs of regression. Furthermore, combining both ICB treatments with neoantigen vaccination did not induce tumor regression either despite neoantigen-specific CD8+ TIL expansion. Overall, this work provides a model for studying neoantigens in an immunotherapy nonresponder model. We showed that a robust neoantigen-specific T-cell response in the LLC tumor model could fail in tumor response to ICB, which will have important implications in designing future immunotherapeutic strategies.

Intratumoral CD39+CD8+ T Cells Predict Response to Programmed Cell Death Protein-1 or Programmed Death Ligand-1 Blockade in Patients With NSCLC.

INTRODUCTION: Programmed cell death protein-1 (PD-1) and programmed death-ligand 1 (PD-L1) blockade is currently widely used in the treatment of metastatic NSCLC. Despite available biomarker stratification, clinical responses vary. Thus, the search for novel biomarkers with improved response prediction is ongoing. Previously, using mass cytometry or cytometry by time-of-flight (CyTOF), our group demonstrated that CD39+CD8+ immune cells represent tumor antigen-specific, cytotoxic T cells in treatment-naive NSCLC. We hypothesized that accurate quantitation of this T cell subset would predict immunotherapy outcome. METHODS: To translate this to a clinical setting, the present study compared CyTOF data with a range of clinically relevant methods, including conventional immunohistochemistry (IHC), multiplex IHC or immunofluorescence (mIHC), and gene expression assay by NanoString. RESULTS: Quantification using mIHC but not conventional IHC or NanoString correlated with the CyTOF results. The specificity and sensitivity of mIHC were then evaluated in a separate retrospective NSCLC cohort. CD39+CD8+ T cell proportion, as determined by mIHC, successfully stratified responders and nonresponders to PD-1 or PD-L1 inhibitors (objective response rate of 63.6%, compared with 0% for the negative group). This predictive capability was independent from other confounding factors, such as total CD8+ T cell proportion, CD39+ lymphocyte proportion, PD-L1 positivity, EGFR mutation status, and other clinicopathologic parameters. CONCLUSIONS: Our results suggest that the mIHC platform is a clinically relevant method to evaluate CD39+CD8+ T cell proportion and that this marker can serve as a potential biomarker that predicts response to PD-1 or PD-L1 blockade in patients with NSCLC. Further validation in additional NSCLC cohorts is warranted.

Human Tumor-Infiltrating MAIT Cells Display Hallmarks of Bacterial Antigen Recognition in Colorectal Cancer.

Growing evidence indicates a role for the gut microbiota in modulating anti-tumor treatment efficacy in human cancer. Here we study mucosa-associated invariant T (MAIT) cells to look for evidence of bacterial antigen recognition in human colon, lung, and kidney carcinomas. Using mass cytometry and single-cell mRNA sequencing, we identify a tumor-infiltrating MAIT cell subset expressing CD4 and Foxp3 and observe high expression of CD39 on MAIT cells from colorectal cancer (CRC) only, which we show in vitro to be expressed specifically after TCR stimulation. We further reveal that these cells are phenotypically and functionally exhausted. Sequencing data show high bacterial infiltration in CRC tumors and highlight an enriched species, Fusobacteria nucleatum, with capability to activate MAIT cells in a TCR-dependent way. Our results provide evidence of a MAIT cell response to microbial antigens in CRC and could pave the way for manipulating MAIT cells or the microbiome for cancer therapy.

Partial absence of PD-1 expression by tumor-infiltrating EBV-specific CD8+ T cells in EBV-driven lymphoepithelioma-like carcinoma.

OBJECTIVES: Lymphoepithelioma-like carcinoma (LELC) is an uncommon lung cancer, typically observed in young, non-smoking Asian populations. LELC is associated with Epstein-Barr virus (EBV) infection of lung tumor cells of epithelial origin, suggesting a carcinogenic role of EBV as observed in nasopharyngeal carcinoma (NPC). Here, we studied the antigen specificity and phenotype of EBV-specific CD8+ T cells in blood and tumor of one LELC patient positive for EBV infection in lung tumor cells. METHODS: Using multiplex MHC class I tetramers, mass cytometry and mRNA sequencing, we studied EBV-specific CD8+ T cells at the transcriptomic and phenotypic levels in blood and tumor tissues of the LELC patient. RESULTS: Lymphoepithelioma-like carcinoma lung tumor cells were positive for EBV infection. In both blood and tumor tissues, we detected two populations of EBV-specific CD8+ T cells targeting the EBV lytic cycle proteins: BRLF1 and BMLF1. Transcriptomic analyses of these two populations in the tumor, which can be considered as tumor-specific, revealed their distinct exhausted profile and polyclonal TCR repertoire. High-dimensional phenotypical analysis revealed the distinct phenotype of these cells between blood and tumor tissues. In tumor tissue, EBV-specific CD8+ TILs were phenotypically heterogeneous, but consistently expressed CD39. Unexpectedly, although the LELC tumor cells expressed abundant PD-L1, these tumor-specific CD8+ tumor-infiltrating lymphocytes (TILs) mostly did not express PD-1. CONCLUSION: Epstein-Barr virus-specific CD8+ TILs in EBV-driven tumor are heterogeneous and partially lack PD-1 expression, suggesting that anti-PD1/PD-L1 immunotherapy may not be an appropriate strategy for disinhibiting EBV-specific cells in the treatment of LELC patients.

Ontogeny of different subsets of yellow fever virus-specific circulatory CXCR5+ CD4+ T cells after yellow fever vaccination.

Monitoring the frequency of circulatory CXCR5+ (cCXCR5+) CD4+ T cells in periphery blood provides a potential biomarker to draw inferences about T follicular helper (TFH) activity within germinal center. However, cCXCR5+ T cells are highly heterogeneous in their expression of ICOS, PD1 and CD38 and the relationship between different cCXCR5 subsets as delineated by these markers remains unclear. We applied class II tetramer reagents and mass cytometry to investigate the ontogeny of different subsets of cCXCR5+ T cell following yellow fever immunization. Through unsupervised analyses of mass cytometry data, we show yellow fever virus-specific cCXCR5 T cells elicited by vaccination were initially CD38+ICOS+PD1+, but then transitioned to become CD38+ICOS-PD1+ and CD38-ICOS-PD1+ before coming to rest as a CD38-ICOS-PD1- subset. These results imply that most antigen-specific cCXCR5+ T cells, including the CD38-ICOS-PD1- CXCR5+ T cells are derived from the CXCR5+CD38+ICOS+PD1+ subset, the subset that most resembles preTFH/TFH in the germinal center.

Mucosal-associated invariant T cells promote inflammation and intestinal dysbiosis leading to metabolic dysfunction during obesity.

Obesity is associated with low-grade chronic inflammation promoting insulin-resistance and diabetes. Gut microbiota dysbiosis is a consequence as well as a driver of obesity and diabetes. Mucosal-associated invariant T cells (MAIT) are innate-like T cells expressing a semi-invariant T cell receptor restricted to the non-classical MHC class I molecule MR1 presenting bacterial ligands. Here we show that during obesity MAIT cells promote inflammation in both adipose tissue and ileum, leading to insulin resistance and impaired glucose and lipid metabolism. MAIT cells act in adipose tissue by inducing M1 macrophage polarization in an MR1-dependent manner and in the gut by inducing microbiota dysbiosis and loss of gut integrity. Both MAIT cell-induced tissue alterations contribute to metabolic dysfunction. Treatment with MAIT cell inhibitory ligand demonstrates its potential as a strategy against inflammation, dysbiosis and metabolic disorders.

A Targeted Multi-omic Analysis Approach Measures Protein Expression and Low-Abundance Transcripts on the Single-Cell Level.

High-throughput single-cell RNA sequencing (scRNA-seq) has become a frequently used tool to assess immune cell heterogeneity. Recently, the combined measurement of RNA and protein expression was developed, commonly known as cellular indexing of transcriptomes and epitopes by sequencing (CITE-seq). Acquisition of protein expression data along with transcriptome data resolves some of the limitations inherent to only assessing transcripts but also nearly doubles the sequencing read depth required per single cell. Furthermore, there is still a paucity of analysis tools to visualize combined transcript-protein datasets. Here, we describe a targeted transcriptomics approach that combines an analysis of over 400 genes with simultaneous measurement of over 40 proteins on 2 × 104 cells in a single experiment. This targeted approach requires only about one-tenth of the read depth compared to a whole-transcriptome approach while retaining high sensitivity for low abundance transcripts. To analyze these multi-omic datasets, we adapted one-dimensional soli expression by nonlinear stochastic embedding (One-SENSE) for intuitive visualization of protein-transcript relationships on a single-cell level.

Multiplex MHC Class I Tetramer Combined with Intranuclear Staining by Mass Cytometry.

Antigen-specific CD8+ T cells play a crucial role in the host protective immune response against viruses, tumors, and other diseases. Major histocompatibility complex (MHC) class I tetramers allow for a direct detection of such antigen-specific CD8+ T cells. Using mass cytometry together with multiplex MHC class I tetramer staining, we are able to screen more than 1000 different antigen candidates simultaneously across tissues in health and disease, while retaining the possibility to deliver an in-depth characterization of antigen-specific CD8+ T cells and associated phenotypes. Here we describe the method for a MHC class I tetramer multiplexing approach together with intracellular antibody staining for a parallel phenotypic cell characterization using mass cytometry in human specimens.

A Subset of Type I Conventional Dendritic Cells Controls Cutaneous Bacterial Infections through VEGFα-Mediated Recruitment of Neutrophils.

Skin conventional dendritic cells (cDCs) exist as two distinct subsets, cDC1s and cDC2s, which maintain the balance of immunity to pathogens and tolerance to self and microbiota. Here, we examined the roles of dermal cDC1s and cDC2s during bacterial infection, notably Propionibacterium acnes (P. acnes). cDC1s, but not cDC2s, regulated the magnitude of the immune response to P. acnes in the murine dermis by controlling neutrophil recruitment to the inflamed site and survival and function therein. Single-cell mRNA sequencing revealed that this regulation relied on secretion of the cytokine vascular endothelial growth factor α (VEGF-α) by a minor subset of activated EpCAM+CD59+Ly-6D+ cDC1s. Neutrophil recruitment by dermal cDC1s was also observed during S. aureus, bacillus Calmette-Guérin (BCG), or E. coli infection, as well as in a model of bacterial insult in human skin. Thus, skin cDC1s are essential regulators of the innate response in cutaneous immunity and have roles beyond classical antigen presentation.

Reverse-engineering flow-cytometry gating strategies for phenotypic labelling and high-performance cell sorting.

Motivation: Recent flow and mass cytometers generate datasets of dimensions 20 to 40 and a million single cells. From these, many tools facilitate the discovery of new cell populations associated with diseases or physiology. These new cell populations require the identification of new gating strategies, but gating strategies become exponentially more difficult to optimize when dimensionality increases. To facilitate this step, we developed Hypergate, an algorithm which given a cell population of interest identifies a gating strategy optimized for high yield and purity. Results: Hypergate achieves higher yield and purity than human experts, Support Vector Machines and Random-Forests on public datasets. We use it to revisit some established gating strategies for the identification of innate lymphoid cells, which identifies concise and efficient strategies that allow gating these cells with fewer parameters but higher yield and purity than the current standards. For phenotypic description, Hypergate's outputs are consistent with fields' knowledge and sparser than those from a competing method. Availability and implementation: Hypergate is implemented in R and available on CRAN. The source code is published at http://github.com/ebecht/hypergate under an Open Source Initiative-compliant licence. Supplementary information: Supplementary data are available at Bioinformatics online.

Mapping of γ/δ T cells reveals Vδ2+ T cells resistance to senescence.

BACKGROUND: Immune adaptation with aging is a major of health outcomes. Studies in humans have mainly focus on αß T cells while γδ T cells have been neglected despite their role in immunosurveillance. We investigated the impact of aging on γδ T cell subsets phenotypes, functions, senescence and their molecular response to stress. METHODS: Peripheral blood of young and old donors in Singapore have been used to assess the phenotype, functional capacity, proliferation capacity and gene expression of the various γδ T cell subsets. Peripheral blood mononuclear cells from apheresis cones and young donors have been used to characterize the telomere length, epigenetics profile and DNA damage response of the various γδ T cell subsets phenotype. FINDINGS: Our data shows that peripheral Vδ2+ phenotype, functional capacity (cytokines, cytotoxicity, proliferation) and gene expression profile are specific when compared against all other αß and γδ T cells in aging. Hallmarks of senescence including telomere length, epigenetic profile and DNA damage response of Vδ2+ also differs against all other αß and γδ T cells. INTERPRETATION: Our results highlight the differential impact of lifelong stress on γδ T cells subsets, and highlight possible mechanisms that enable Vδ2+ to be resistant to cellular aging. The new findings reinforce the concept that Vδ2+ have an "innate-like" behavior and are more resilient to the environment as compared to "adaptive-like" Vδ1+ T cells.

Bystander CD8+ T cells are abundant and phenotypically distinct in human tumour infiltrates.

Various forms of immunotherapy, such as checkpoint blockade immunotherapy, are proving to be effective at restoring T cell-mediated immune responses that can lead to marked and sustained clinical responses, but only in some patients and cancer types1-4. Patients and tumours may respond unpredictably to immunotherapy partly owing to heterogeneity of the immune composition and phenotypic profiles of tumour-infiltrating lymphocytes (TILs) within individual tumours and between patients5,6. Although there is evidence that tumour-mutation-derived neoantigen-specific T cells play a role in tumour control2,4,7-10, in most cases the antigen specificities of phenotypically diverse tumour-infiltrating T cells are largely unknown. Here we show that human lung and colorectal cancer CD8+ TILs can not only be specific for tumour antigens (for example, neoantigens), but also recognize a wide range of epitopes unrelated to cancer (such as those from Epstein-Barr virus, human cytomegalovirus or influenza virus). We found that these bystander CD8+ TILs have diverse phenotypes that overlap with tumour-specific cells, but lack CD39 expression. In colorectal and lung tumours, the absence of CD39 in CD8+ TILs defines populations that lack hallmarks of chronic antigen stimulation at the tumour site, supporting their classification as bystanders. Expression of CD39 varied markedly between patients, with some patients having predominantly CD39- CD8+ TILs. Furthermore, frequencies of CD39 expression among CD8+ TILs correlated with several important clinical parameters, such as the mutation status of lung tumour epidermal growth factor receptors. Our results demonstrate that not all tumour-infiltrating T cells are specific for tumour antigens, and suggest that measuring CD39 expression could be a straightforward way to quantify or isolate bystander T cells.

Adaptive NKG2C+CD57+ Natural Killer Cell and Tim-3 Expression During Viral Infections.

Repetitive stimulation by persistent pathogens such as human cytomegalovirus (HCMV) or human immunodeficiency virus (HIV) induces the differentiation of natural killer (NK) cells. This maturation pathway is characterized by the acquisition of phenotypic markers, CD2, CD57, and NKG2C, and effector functions-a process regulated by Tim-3 and orchestrated by a complex network of transcriptional factors, involving T-bet, Eomes, Zeb2, promyelocytic leukemia zinc finger protein, and Foxo3. Here, we show that persistent immune activation during chronic viral co-infections (HCMV, hepatitis C virus, and HIV) interferes with the functional phenotype of NK cells by modulating the Tim-3 pathway; a decrease in Tim-3 expression combined with the acquisition of inhibitory receptors skewed NK cells toward an exhausted and cytotoxic phenotype in an inflammatory environment during chronic HIV infection. A better understanding of the mechanisms underlying NK cell differentiation could aid the identification of new immunological targets for checkpoint blockade therapies in a manner that is relevant to chronic infection and cancer.

CD161 Defines a Functionally Distinct Subset of Pro-Inflammatory Natural Killer Cells.

CD161 is a C-type lectin-like receptor expressed on the majority of natural killer (NK) cells; however, the significance of CD161 expression on NK cells has not been comprehensively investigated. Recently, we found that CD161 expression identifies a transcriptional and innate functional phenotype that is shared across various T cell populations. Using mass cytometry and microarray experiments, we demonstrate that this functional phenotype extends to NK cells. CD161 marks NK cells that have retained the ability to respond to innate cytokines during their differentiation, and is lost upon cytomegalovirus-induced maturation in both healthy and human immunodeficiency virus (HIV)-infected patients. These pro-inflammatory NK cells are present in the inflamed lamina propria where they are enriched for integrin CD103 expression. Thus, CD161 expression identifies NK cells that may contribute to inflammatory disease pathogenesis and correlates with an innate responsiveness to cytokines in both T and NK cells.