Fig 2: Systemic CD8 responses in NSCLC patients Percentage of circulating highly differentiated CD8 cells in age-matched healthy donors (n = 40) and NSCLC patients (n = 51) before undergoing immunotherapies. Relevant statistical comparisons are shown by the test of Mann–Whitney. In green, objective responders (OR). In red, no OR.ROC analysis of baseline CD8 THD quantification as a function of objective clinical responses.Dot-plot of lung cancer antigen-specific CD8 T cells obtained before the start of immunotherapies and stimulated with A549-loaded autologous DCs autologous DC in healthy donors (n = 5), G1 (n = 11) and G2 (n = 13) patients, as indicated. Relevant statistical comparisons are indicated by the test of Kruskal–Wallis.Left dot-plot, percentage of CD28-negative CD8 T cells within TAA-specific CD8 subsets in G1 (n = 11) and G2 (n = 13) patients, as indicated. Right dot-plot, same as left but with CD28-positive subsets. Green, objective responders (OR) and red, no ORs. N, number of biological replicates (independent patients).

Fig 3: Differential systemic CD4 immunity and responses to PD-1/PD-L1 blockade in NSCLC patients The scatter plot shows PD-1 expression after co-culture of CD4 T cells from healthy donors (n = 9) or NSCLC patients (n = 14), as indicated, with A459-SC3 lung cancer cells. Relevant statistical comparisons with the test of Mann–Whitney are indicated.Upper graphs, flow cytometry density plots of PD-1 and LAG-3 co-expression in CD4 T cells from healthy donors, a G1 responder (G1 R), a G1 non-responder (G1 NR), and a G2 non-responder as indicated, following stimulation with A549-SC3 cells. Percentage of expressing cells are indicated within each quadrant. Below, same as in the upper graphs but as a scatter plot of the percentage of CD4 T cells that simultaneously co-express PD-1 and LAG-3 that simultaneously co-express PD-1 and LAG-3 in G1 healthy donors (n = 10), G1 (n = 10) and G2 (n = 10) patients. Relevant statistical comparisons are shown with the test of Mann–Whitney.Upper flow cytometry histograms of Ki67 expression in CD4 T cells from the representative subjects as indicated on the right, after stimulation with A549-SC3 cells. Vertical dotted line indicates the cut-off value of positive versus negative Ki67 expression. The percentage of Ki67-expressing CD4 T cells is shown within the histograms. Below, same data represented as a scatter plot from a sample of G1 and G2 donors as indicated, with relevant statistical comparisons with the test of Mann–Whitney (n = 7–10).Proliferation of CD4 T cells stimulated by A549-SC3 cells from the indicated patient groups. CD28 expression is shown together with the proliferation marker Ki67. Percentages of cells within each quadrant are shown.Same as in (D) but in the presence of an isotype control antibody or an anti-PD-1 antibody with the equivalent sequence to pembrolizumab. The effects on CD4 T cells from a G1 and a G2 patient are shown, divided into CD28 high or low/negative subsets as indicated. Relevant statistical comparisons are shown with paired Student's t-test.Top, flow cytometry density plots of Ki67 expression in CD4 T cells from representative G1 or G2 patients after three cycles of therapy, activated by incubation with A549-SC3 cells. Below, same as above but as a dot-plot graph (n = 7–10). A comparison between proliferating CD4 T cells before and after therapy is shown in unpaired patient samples. G1 R, G1 objective responder patient. G2 NR, G2 patient with no objective responses; green, objective responders (OR) and red, no OR; Iso, treatment with an isotype antibody control; and a-PD-1, treatment with anti-PD-1 antibody. Statistical comparisons were performed with the test of Mann–Whitney.

Fig 4: Systemic circulating CD4 T cells in NSCLC patients are proficient in cytokine production with an overall Th17 profile AColumn graphs representing the percentage of CD4 T cells from NSCLC patients or age-matched healthy donors as represented in the graph, expressing the indicated cytokines after T-cell stimulation with anti-CD3/anti-CD28 antibodies. Data on total CD4 (left graph), CD28+ subsets (center graph) and CD28negative subsets (right graph) are shown. Error bars correspond to standard deviations, and bars represent means from nine independent biological replicates (healthy donors) and six independent replicates (patients).B–DSame as in (A) but representing CD4 T cells expressing only one cytokine (B), two (C) or three cytokines simultaneously (D). Error bars correspond to standard deviations, and bars represent means from five independent biological replicates (patients).Data information: Relevant statistical comparisons are shown within the graphs by the test of Kruskal–Wallis.

Fig 5: CD4 THD cells in NSCLC patients are mainly non-senescent memory subsets AScatter plot graphs of the percentage of memory phenotypes in baseline CD4 THD cells according to CD62L-CD45RA expression (% CD45RAnegative CD62Lpositive central-memory + % CD45RAnegative CD62Lnegative effector-memory cells) in a sample of healthy donors (n = 18), G1 (n = 12) and G2 (n = 19) patients. Relevant statistical comparisons are shown by one-way ANOVA followed by Tukey's test.B, CExpression of the genotoxic damage makers H2AX (B) and CD57 (C) by flow cytometry in CD4 T-cell subsets from an aged-matched healthy donor, and NSCLC G1 and G2 patients as indicated. Percentage of positivity and mean fluorescent intensities are indicated for each population. Top, histogram analysis within CD27+ CD28+ CD4 T cells, and bottom, CD27negative CD28low/negative counterparts as indicated. US, unstained control.