Characterization and clinical relevance of CD161⁺ and TIGIT⁺ CD4 T cells. (A) Representative gating of CD161⁺, TIGIT⁺, and PD-1⁺ CD4 T cells (after exclusion of Treg and MAIT cells). (B) Percentage of cells (non-Treg/non-MAIT) expressing CD161, TIGIT, at least CD161 or TIGIT (Boolean gating), or PD-1, among total CD4 T cells from healthy donors (n = 20), or from patients (at month 3 after allo-HSCT) with persistent remission (n = 20) or with subsequent tumor relapse (n = 20). (C) Uniform Manifold Approximation and Projection (UMAP) plots showing total CD4 T-cell clustering. Outlined cell subsets (from manual gating) are colored, whereas the ungated CD4 T cells are shown in gray (left). Expression levels of CD161, TIGIT and PD-1 are depicted (right), showing a poor overlap between CD161⁺ and TIGIT⁺ cells, but a strong overlap between TIGIT⁺ and PD-1^high cells. (D) Frequencies of cells expressing various IRs among the indicated CD4 T-cell subsets. PD-1 cell surface density is also shown (expressed as median metal intensity, MMI) (bottom right). (E) Frequencies of HLA-DR⁺ and intranuclear Ki-67⁺ cells among CD161- and TIGIT-expressing CD4 T cells, compared with control (ctrl) cells that do not express CD161 and TIGIT. (F) Frequencies of CD161⁺ and TIGIT⁺ CD4 T cells were compared with regard to sex and various clinical parameters (ie, conditioning regimen, donor type, CMV reactivation, and acute GVHD). RIC, reduced intensity conditioning; MAC, myeloablative conditioning; MRD, matched related donor; MUD, matched unrelated donor. (G) FlowSOM clustering of PBMCs from the 40 patients at M3 after HSCT was performed based on lineage-defining markers (supplemental Figure 4). Frequencies of cluster 34 (left; corresponding to CD161⁺ CD4 T cells) among CD4 T cells, in nonrelapsing and relapsing patients. Relapse-free survival (RFS; right) assessed using the Kaplan-Meier approach in patients with high (>22%) vs low (<22%) frequencies of C34 in CD4 T cells. Censored patients are represented by symbols. P values by Mann-Whitney (B,F,G), Wilcoxon (E), or log-rank (G) tests: *P < .05; **P < .01; ****P < .0001.