Fig. 4. α-PD-1 checkpoint blockade alters the regulatory T cell contexture and leads to immunosuppressive CD103⁺ Tregs accumulation in the glioblastoma TME.

a, UMAP⁴⁷ projection and unsupervised FlowSOM⁴⁸ clustering of the Treg population in PDG-Ink4a tumors identified five distinct subpopulations of Tregs (Pop 1–5). b, Heat map depicting the mean fluorescence intensity (MFI) of activation markers for the identified Treg subpopulations in a. c, UMAP density projections plot of Treg subpopulations from a in RT and RT + Conc.IT treatment groups. For a–c: RT n = 5, RT + Conc.IT n = 6 mice. d, Quantification of CD103⁺ Tregs (gated from CD45⁺CD11b⁺CD3⁺CD4⁺FOXP3⁺KLRG1⁻) and KLRG1⁺ Tregs (gated from CD45⁺CD11b⁺CD3⁺CD4⁺FOXP3⁺) in RT- or RT + Conc.IT-treated PDG-Ink4a tumors (Tu, tumor). e, Quantification of CD25⁺ Tregs in the CD103⁺ and KLRG1⁺ Treg populations from d. For d and e: RT CD103⁺ n = 5, RT KLRG1⁺ n = 5, RT + Conc.IT CD103⁺ n = 6, RT + Conc.IT KLRG1⁺ n = 6 mice. f–j, CD4⁺ T cells, CD25⁺ Tregs and CD103⁺ Tregs FACS-purified from RT- and RT + Conc.IT-treated PDG-Ink4a glioblastoma submitted to RNA-seq analyses. Enrichment of the Magnuson Treg gene signature⁵⁴ (f) (RT CD4⁺ n = 3, RT + Conc.IT CD4⁺ n = 3, RT CD25⁺ n = 3, RT + Conc.IT CD25⁺ n = 3, RT + Conc.IT CD103⁺ n = 3 mice). Venn diagram (g) of differentially upregulated genes in RT + Conc.IT CD103⁺ Tregs and RT + Conc.IT CD25⁺ Tregs versus RT CD25⁺ Tregs (Supplementary Table 2). Bar graph (h) of upregulated pathways identified from the 702 shared genes common to RT + Conc.IT CD25⁺ Tregs and RT + Conc.IT CD103⁺ Tregs versus RT CD25⁺ Tregs (Supplementary Table 9). Volcano plot (i) depicting log₂ fold change (x axis) versus significance (−log₁₀(P value)) of differentially expressed genes in RT + Conc.IT CD25⁺ versus RT + Conc.IT CD103⁺ Tregs (Supplementary Table 12). Bar graph (j) depicting the upregulated pathways identified from the 122 genes upregulated only in RT + Conc.IT CD103⁺ Tregs (not in RT + Conc.IT CD25⁺ Tregs) versus RT CD25⁺ Tregs (Supplementary Table 11). For g–j: RT CD25⁺ n = 3, RT + Conc.IT CD25⁺ n = 3, RT + Conc.IT CD103⁺ n = 3 mice. k, Flow cytometry quantification of CD39⁺, Ki67⁺, IFNy⁺, GrzB⁺ and GrzA⁺ FACS-purified CD8⁺ T cells (from control spleens) after 24 h of monoculture (mono) or co-culture (cocx) with CD25⁻ T cells, CD25⁺ or CD103⁺ Tregs isolated from RT + Conc.IT-treated PDG-Ink4a. Cells were stimulated with anti-CD3/anti-CD28 antibodies, and cultured at a 1:1 ratio (Tregs:CD8⁺ T cells; mono: n = 11, CD25⁻ n = 6, CD25⁺ n = 3, CD103⁺ n = 3 biologically independent samples). For all graphs, analyses were done at d12 post treatment initiation on the tumor-containing brain quadrant. Statistics: one-way ANOVA with Benjamini, Krieger and Yekutieli correction for multiple testing (d–f and k), Fisher’s exact test (two-sided; h and j) and Wald test (i) in combination with the Benjamini–Hochberg method for correction of multiple hypotheses testing (two-sided; h and j). Data are represented as mean ± s.e.m. (d–f) or ± s.d. (k). Gating strategies (g and k) depicted in Extended Data Fig. 6a.

Source data