Fig. 6: Interleukin 15 (IL-15) addition overcomes CDK4/6i-induced immune suppression, boosting CD8 + T cell activation and cancer control.

A Ribociclib immune-suppression shown by decreased peripheral white blood cell (WBC) counts in serial blood samples (points) of FELINE patients during combination ribociclib treatment but not letrozole alone (generalized additive model (GAM):non-linear trend::ribocicli:eff.df = 2.27, F = 12.52, p = 4.6e-7;letrozole:eff.df = 1.0(linear), F = 0.045, p = 0.83(trend; not significant:NS))(solid line = GAM treatment-specific trend,shaded = 95% confidence interval(+/-1.96*SE)). No pre-treatment difference between arms (log-linear model+ANOVA:estimate = 0.073,se = 0.081, df = 59, t = 0.91, p = 0.37). Sample size = 408 blood draws, 62 patients (Treatment:ribociclib = 39,letrozole = 23) across 7 timepoints. B Ribociclib (color) reduced patient-derived CD8 + T cell viability (top) and area (bottom) after 163-hour monoculture (log-linear model(IL-15 = 0 ng/mL):ATP:estimate = −0.48,se = 0.079, df = 2, t = −6.04, p = 0.0038;Area:estimate = −0.44,se = 0.03, df = 2, t = −14.8, p = 0.0045). IL-15 overcame inhibition, with IL-15 > 1 ng/mL restoring viability/area above DMSO control (GAM:IL-15 activation::ATP:eff.df = 3, F = 760.87, p = 2e-16;Area:eff.df = 2.94, F = 611.5, p = 2e-16). GAM characterized non-linear dose-dependent IL-15 effect with/without ribociclib (solid line = expectation;shaded regions = 95% confidence interval (expectation + /-1.96*se)). Sample size = 48 measurements, 3 experimental replicates under 16 treatment (8 IL-15 levels(0-10 ng/mL)+/-1uM ribociclib). C IL-15 (x-axis) slowed cancer growth of 4 cell lines (panels:CAMA-1/MDA-MB-134 ribociclib-resistant/sensitive pairs) in patient-derived T cell cocultures (circles;seeding-ratio::cancer:T-cell=4:1) but not monocultures (triangles)(GAM::IL-15|coculture:CAMA-1:resistant:edf = 2.00, F = 1323.00, p = 1e-16,CAMA-1:sensitive:edf = 2.00, F = 543.25, p = 1e-16,MD-AMB-134:resistant:edf = 2.00, F = 146.85, p = 1e-16,MDAMB134:sensitive:edf = 1.70, F = 693.80, p = 1e-16). Cocultured cancer growth was reduced less by IL-15 stimulation under ribociclib treatment (yellow circles) versus DMSO control cocultures (blue circles)(GAM::IL-15|ribo:CAMA-1:resistant:edf = 1.88, F = 56.15, p < 1e-16,CAMA-1:sensitive:edf = 1.00, F = 9.61, p = 2.7e-5,MD-AMB-134:resistant:edf = 1.88, F = 16.07, p = 1.5e-6,MDAMB134:sensitive:edf = 1.95, F = 51.77, p < e-16). Higher dose IL-15 controlled ribociclib-treated coculture cancer growth. Points = replicate cancer growth rate over 6 days versus mean of IL-15 untreated monocultures (blue triangles:IL-15 = 0 ng/mL) per ribociclib treatment). GAMs characterized treatment impacts on growth rates per lineage and coculture composition (dashed/solid lines = mono/coculture expectations;shaded = 95% confidence intervals (+/-1.96*se)). Sample size = 288 spheroids, 4 cancer lineages (CAMA-1/MDA-MB-134 resistant/sensitive), 2 compositions (mono/coculture), 6 IL-15 concentrations (0/0.5/0.75/1/2.5/5 ng/mL), 2 ribociclib doses (0/1uM), 3 experimental replicates. D Representative florescent imaging (4x magnification) demonstrating IL-15 activation of an effective cytotoxic T-Cell (unlabeled: black) response in sensitive (YFP labeled:yellow) or ribociclib-resistant (CFP labeled:blue) CAMA-1 cancer cells. Cancer monoculture (top rows) and cancer-T cell coculture (bottom rows) spheroids following 6-day treatment with DMSO (0.1%;control), ribociclib (1 µM), IL-15 (1-5 ng/mL), or combination ribociclib (1 µM) + IL-15 (1-5 ng/mL). E Schematic: diverse pre-treatment cancer communications stimulate immune-suppressing M2-like myeloid polarization in ribociclib-resistant tumors. Reduced pro-immune M1-like myeloid differentiation diminishes interleukin/integrin signaling and subsequent CD8 + T cell activation/recruitment, preventing effective killing of quiescent cancer cells. All statistical tests two-sided. Source data provided as a Source Data file.