a Scatter plot showing viability differences of vehicle and glucocorticoid pre-treated or co-treated screen arms with either 1 μM and 5 μM of the screening drugs. Compounds having significant differential viability are depicted in green (Padj < 0.05). Adjusted P values (Padj) were determined by two-sided t test with multiple testing correction (Benjamini–Hochberg method). b Normalised viability of H2122 cell lines in response to compounds found to diminish viability (<0.7) of vehicle arm of H1944 cell lines in the first screen at 5 μM concentration. P values were determined by two-sided Welch’s t test. c Compound Set Enrichment Analysis data computed using CSgator software of drugs with reduced efficacy upon GC treatment. Random controls of the same size were generated to compute background enrichment (n = 20). P values (FDR-adjusted) were determined by CSgator software. d PhosphoPath pathway analysis of phosphoproteomics data depicting signalling changes in glucocorticoid-treated cells (A549, H2122 and H1944) (n = 3). P values were determined by PhosphoPath software. e Representative images showing immunohistochemical stainings of xenograft cancer samples obtained after treatment of NOD-SCID-γ mice with either vehicle (Veh) or glucocorticoids (GC) (n = 4). Primary antibody was omitted as staining control. Scale bar, 100 μm. f Normalised tumour growth in xenograft models of H1944 cells in NOD-SCID-γ mice treated with either vehicle (Veh = blue), linsitinib (Lin = green), glucocorticoids (GC = red), or combination (GC + Lin = purple). Arrows indicate when treatment was started. Mean values ± SEM depicted (Dexa + Lin n = 4, Dexa n = 6, Veh n = 5, Lin n = 6 animals). P values were determined by mixed-model ANOVA (Tukey’s multiple comparison test).