(A) Dispersed mouse islet cells cultured in islet complete medium (15 mM glucose, 10% FBS, control) or starve medium (2 mM glucose, 0% FBS) for 16 hours were fixed, immunostained for FOXO1 (red), insulin (green), and Dapi (blue), and imaged by confocal microscopy. (B–G) Dispersed mouse islet cells transduced with the indicated adenoviruses were cultured as in A, then lysed for qPCR for known FOXO1 target genes (n ≥ 12) (B), proliferation markers (n ≥ 12) (C), and β cell differentiation genes (n ≥ 12) (D). (E–G) Transduction with CDK4 activators (Ad-h-CDK4, Ad-m-cyclinD2, and Ad-h-Cdk4+Ad-m-cyclinD2), or CDK4 inhibitor (Ad-m-p16) showed that activating CDK4 rescued abundance of proliferation markers (n ≥ 7) (E), rescued 2 of 3 FOXO1 targets (n ≥ 8) (F), and rescued Pdx1 (n ≥ 8) and Neurod1 (n ≥ 8) expression but not Mafa (n ≥ 8). Nkx6.1 (n ≥ 5) and Ngn3 (n ≥ 5) were not changed by starvation or CDK4 overexpression (G). Data in B–D are the controls from E–G, presented separately for clarity. (H and I) (n ≥ 4) Ad-h-CDK4 or Ad-m-cyclinD2 increased phosphorylation of FOXO1 in dispersed mouse islet cells exposed to starve conditions (H), but Ad-m-p16 did not, quantified in (I). Confocal microscopy (J) with quantification of the percentage β cells with nuclear FOXO1 showed nuclear FOXO1 was variably reduced by CDK4 activation. Top and bottom panels show 2 different biological replicates illustrating variability of nuclear FOXO1 abundance. Grayscale panels represent red-channel (FOXO1) immunofluorescence. Dashed lines represent mean starve control condition. Images in A and J are representative of n ≥ 4 experiments. Original magnification, ×400 (A and J). Statistics by t test (B-D) or 1-way ANOVA (E–I) with Tukey’s posthoc test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.