Fig 2: Kitl protein expression is stimulated by IGF1 via multiple IGF1R-activated pathways. A, Time-dependent stimulation of Kitl expression in the gastric tunica muscularis of 14-16-day-old BALB/c mice by the potent IGF1 analog LR3-rhIGF1 administered in a single 150 µg/kg dose i.p.; n=5 mice/group. B, Time-dependent stimulation of Kitl expression by 100 ng/mL rhIGF1 in gastric corpus+antrum tunica muscularis organotypic cultures from 14-16-day-old BALB/c mice; n=4-5/group. C, Concentration-dependent stimulation of Kitl protein expression by 18-h rhIGF1 treatment; n=7 organotypic cultures/group. D, Blockade of the rhIGF1-induced Kitl expression in organotypic cultures by AG1024 (1 µM), a specific inhibitor of IGF1/insulin receptor tyrosine kinase activity; n=4/group. E-F, Effects of 1-h (E) and 3-h (F) rhIGF1treatment on AKT (S473/S474/S472), GSK3β (S9), p70S6K (T389) and ERK1/2 (T202/Y204 and T185/Y187) phosphorylation; n=5 organotypic cultures/group. G, Blockade of rhIGF1-induced Kitl expression in organotypic cultures by specific inhibitors of AKT (AKT Inhibitor X; 150 µM), MTOR–p70S6K (rapamycin; 1.5 nM) and MEK–ERK (PD98059; 50 µM); n=5-16/group. The GSK-3 inhibitor SB415286 (30 µM), which is expected to mimic, rather than inhibit, the effect of the IGF1-induced inhibitory phosphorylation of GSK-3, had no significant effect. Kitl and Gapdh or total (T) and phosphorylated proteins (P) were simultaneously detected in the same samples by two-color immunofluorescence (Figure S1). Representative immunoblots show identical areas of the blots imaged at different wavelengths. Box plots show medians and interquartile ranges; bar graphs indicate means±SEM. Data were normalized to the control groups indicated in the panels (dashed lines). Groups marked by asterisk are different from the control group, and groups not sharing the same superscript letter are different from each other (P<0.05 by post-hoc multiple comparisons). IGF1 stimulated Kitl expression in gastric smooth muscles in vivo and in vitro in a time- and concentration-dependent manner by activating IGFR1 and the AKT–GSK3, MTOR–p70S6K and ERK MAPK pathways.

Fig 3: Role of histone acetylation in the activation of Kitl/KITLG transcription by IGF1 and GSK3i. A, Increased occupancy of the Kitl core promoter region by H4ac in response to 6-h rhIGF1 (100 ng/mL) and SB415286 (30 µM) treatment in murine gastric smooth muscles. Only SB415286 increased occupancy by H3K9ac (B). Representatives of two independent experiments, each performed in triplicates, are shown. C, Dose-dependent stimulation of KITLG expression in LX-2 cells by 24-h treatment with the class I-II HDAC inhibitor SAHA (n=3/group). D, Dose-dependent inhibition of the SB415286-induced stimulation of KITLG expression by the p300/PCAF HAT inhibitor garcinol (24 h) in LX-2 cells (n=3/group). Drugs were applied following 24-h serum deprivation. Groups marked by asterisk are different from the control group, and groups not sharing the same superscript letter are different from each other (P<0.05 by post-hoc multiple comparisons).

Fig 4: Contribution of KITLG-activated KIT signaling to baseline proliferation of GIST and LX-2 cells. A, KIT Y721 phosphorylation was activated by exogenous rhKITLG (100 ng/mL for 10 min following 2 h serum deprivation [13,35]) in GIST-T1 cells containing a heterozygous activating KIT mutation (n=4/group) and in GIST-T1-5R cells containing an additional, imatinib-resistant KIT mutation (n=6/group), but not in GIST882 cells lacking a WT KIT allele (n=5/group). B, Culturing with anti-human KITLG neutralizing antibody for 4 days inhibited baseline proliferation of GIST-T1 cells (P<0.001; n=3; regression and 95% confidence band are shown as solid and dashed lines, respectively) and GIST-T1-5R cells (P<0.001; n=6). The effect of KITLG immunoneutralization on GIST-T1 cells was blocked by preabsorbing the antibody with 10-fold molar excess of rhKITLG (see open symbols in the left panel, second row; n=3/group). KITLG immunoneutralization did not inhibit the proliferation of GIST882 or GIST48B cells lacking a WT KIT allele (note that GIST48B cells also express very low to undetectable levels of KIT protein, see Figure S1E) and of LX-2 cells lacking KIT protein (Figure S1E) (n=3/cell line). C, Inhibition of the proliferation of GIST-T1 and GIST-T1-5R cells by siRNA-mediated knock-down of KITLG (n=4/cell line/group; P GIST-T1: day 2: 0.008, days 4, 6 and 8: <0.001; P GIST-T1-5R: day 2: <0.02, day 4: 0.004, days 6 and 8: <0.001).

Fig 5: Role of reduced H3K9 methylation in the activation of Kitl/KITLG transcription by IGF1 and GSK3i. A-B, Reduced occupancy of the Kitl core promoter by the repressive H3K9me2 (A) and H3K9me3 (B) histone modifications in response to 6-h rhIGF1 (100 ng/mL) and SB415286 (30 µM) treatment in murine gastric smooth muscles. C-D, Probing the role of HP1 isoforms in transcriptional repression of KITLG in LX-2 cells by siRNA- (CBX5: HP1a; 25 nM, 72 h; C) or shRNA-mediated knock-down (CBX1: HP1ß; CBX3: HP1?; 30 µg plasmid, 72 h; D). Note activation of KITLG expression by shRNA-mediated knock-down of HP1? (n=3/group). E-F, Stimulation of KITLG expression by the G9A/GLP H3K9me1/2 HKMT inhibitor BIX-01294 in LX-2 cells (E) and GIST-T1 cells (F) (n=3/group). G, Stimulation of KITLG expression by the H3K9 HKMT inhibitor chaetocin in LX-2 cells (n=3/group). Drugs were applied for 24 h following 24-h serum deprivation. See further details in the legend to Figure 5.