Fig. 5: Anti-tumor myeloid communications maintain CD8 + T cell cytotoxicity and migration in CDK4/6i-sensitive tumors. | Nature Communications

Fig. 5: Anti-tumor myeloid communications maintain CD8 + T cell cytotoxicity and migration in CDK4/6i-sensitive tumors.

From: Cellular interactions within the immune microenvironment underpins resistance to cell cycle inhibition in breast cancers

Fig. 5

A Schematic: communication pathway analysis revealing consequences of M2-like polarization of myeloid populations on immune-activating signals to CD8 + T cells. Cytokine communication strengths from phenotypically diverse myeloid populations to CD8 + T cells were contrasted between resistant/sensitive tumors throughout treatment (Hierarchical random effects models (HRE)+Satterthwaite t-test) and verified in discovery/validation cohorts. B Box plot: reduced CD8 + T cell activating communications from myeloid populations in ribociclib-resistant versus sensitive tumors (top panels) post-treatment (Day 180) in the discovery/validation cohorts but not under letrozole alone (NS = not significant)(HRE statistics in Supplementary Data 29)(color = Interleukin communication). Sample size: 42049 cells (Myeloid:Discovery = 10940+Validation = 16097; T-cell:Discovery = 3496+Validation = 11516), 139 cytokine receptors (from gene-ontology), 347 LR communications, 134 tumor samples (Discovery:biopsies = 59, patients = 28,Validation:biopsies = 75, patients = 27), 3 timepoints. C Box plot: reduced CD8 + T cell effector differentiation in ribociclib-resistant tumors during treatment (top) causing lower effector function post-treatment versus sensitive tumors (HRE::differentiation trend:estimate  =  −1.33,se = 3.13e-5, df = 112.7, t = −4.26, p = 2.2e-5;post-treatment:estimate=1.63,se=7.7e-5, df=72.3, t = 2.11, p = 0.035). No trend in differentiation under letrozole alone (bottom)(estimate = −1.29,se = 1.35e-4, df = 141.7, t = −0.94, p = 0.35). Effector differentiation of discovery/validation cohort (shape) measured using CD8 + T cell specific ssGSEA pathway contrasting naive/cancer-killing effector expression (GSE_22886_Naive_CD8_T_cell_vs_NK_cell_up). Sample size: 2579 CD8 + T-cells (Discovery = 1977+Validation = 602 cells), 116 tumor samples (Discovery:biopsies = 56, patients = 28;Validation:biopsies = 60, patients = 26), 3 timepoints. D Box plot: reduced CD8 + T cell recruitment communications from myeloid populations in ribociclib-resistant versus sensitive tumors (top panels) post-treatment in the discovery/validation cohorts but not under letrozole alone (HRE statistics in Supplementary Data 29)(color = integrin recruitment communication)(Sample size: as in B). E Box plot: reduced T cell abundance in ribociclib-resistant tumors (top) during treatment (x-axis) causing lower abundance post-treatment versus sensitive tumors in discovery/validation cohorts (shape)(logistic regression:logit(resistant-trend):estimate = −0.73,se = 0.046, df = 100,z = −16.06, p = 2e-16;post-treatment:estimate = 0.60,se = 0.058, df = 100,z = 10.47, p = 2e-16). Throughout letrozole alone treatment, T cell abundance was lower in resistant versus sensitive tumors (logit(difference):estimate = 0.61,se = 0.050, df = 59,z = 12.3, p = 2e-16). Sample size: 424,581 cells (T-cells = 3166),173 tumor samples,62 patients,3 timepoints. F Box plot: reduced immune response phenotype across cell types (x-axis) post-treatment in resistant versus sensitive tumors under combination ribociclib (top) but not letrozole alone (bottom)(HRE::ribociclib: [Cancer:estimate = −0.027,se = 0.001, df = 1913, t = −20.2, p = 2e-16;Diploid epithelial:estimate = −0.018,se=0.001, df = 2154, t = −12.6, p = 2e-16;Myeloid:estimate = −0.013,se = 0.002, df = 3551, t = −8.49, p = 2e-16;CD8 + T-cell:estimate = −0.026,se = 0.005, df = 6918, t = −5.24, p = 1.7e-7;CD4 + T-cell:estimate = −0.012,se = 0.002, df = 4916, t = −6.03, p = 1.7e-9;Stromal:estimate = −0.014,se = 0.0014, df = 1961, t = −9.9, p = 2e-16];letrozole[Cancer:estimate=0.014,se=0.009, df=1.8, t = 1.47, p = 0.16;Diploid epithelial:estimate = 0.002,se = 0.009, df = 1.9, t = 0.25, p = 0.81;Myeloid:estimate = −0.005,se = 0.009, df = 1.9, t = −0.61, p = 0.55;CD8 + T-cell:estimate = −0.004,se = 0.011, df = 4.1, t = −0.36, p = 0.72;CD4 + T-cell:estimate = −0.014,se = 0.009, df = 1.9, t = −1.58, p = 0.13;Stromal:estimate = −0.005,se = 0.0092, df = 1.8, t = −0.55, p = 0.59]). Points = mean(ssGSEA hallmark interferon gamma response score) per tumor in discovery/validation cohorts (shape). Sample size:108844 cells (Discovery = 43814;Validation = 65030), 53 post-treatment tumors biopsies. All box elements represent median (center line),upper/lower quantiles (hinges),1.5*inter-quartile range (whiskers). All statistical tests two-sided. Source data provided as a Source Data file.

Back to article page