Growth Hormone Induces Colon DNA Damage Independent of IGF-1

Abstract

DNA damage occurs as a result of environmental insults and aging and, if unrepaired, may lead to chromosomal instability and tumorigenesis. Because GH suppresses ataxia-telangiectasia mutated kinase phosphorylation, decreases DNA repair, and increases DNA damage accumulation, we elucidated whether GH effects on DNA damage are mediated through induced IGF-1. In nontumorous human colon cells, GH, but not IGF-1, increased DNA damage. Stably disrupted IGF-1 receptor (IGF-1R) by lentivirus-expressing short hairpin RNA in vitro or treatment with the IGF-1R phosphorylation inhibitor picropodophyllotoxin (PPP) in vitro and in vivo led to markedly induced GH receptor (GHR) abundance, rendering cells more responsive to GH actions. Suppressing IGF-1R triggered DNA damage in both normal human colon cells and three-dimensional human intestinal organoids. DNA damage was further increased when cells with disrupted IGF-1R were treated with GH. Because GH induction of DNA damage accumulation appeared to be mediated not by IGF-1R but probably by more abundant GH receptor expression, we injected athymic mice with GH-secreting xenografts and then treated them with PPP. In these mice, high circulating GH levels were associated with increased colon DNA damage despite disrupted IGF-1R activity (P < 0.01), whereas GHR levels were also induced. Further confirming that GH effects on DNA damage are directly mediated by GHR signaling, GHR-/- mice injected with PPP did not show increased DNA damage, whereas wild-type mice with intact GHR exhibited increased colon DNA damage in the face of IGF-1 signaling suppression. The results indicate that GH directly induces DNA damage independent of IGF-1.

Figure 1.

GH induces DNA damage in hNCCs. Olive tail moment in hNCCs (A) treated with 500 ng/mL GH and harvested 24 h later, (B) infected with lentiGH or lentiV and harvested 7 d after infection, and (C) treated with 100 and 500 ng/mL human IGF-1, respectively, and harvested 24 h later. Results shown are mean ± SEM of three independent experiments. Differences were assessed with Tukey-adjusted mixed model regression. *P < 0.05, **P < 0.01 vs control. Data are graphed as percentage of control, and statistical testing was performed on individual values.

Figure 2.

Suppression of IGF-1R induces GHR expression and increases DNA damage in hNCCs. (A) Representative Western blot of hNCCs infected with lentivirus expressing either scramble shRNA (shScr) as control or shGHR or shIGF-1R RNAi and analyzed 24 h after infection and then assessed for GHR and IGF-1R expression. (B) ImageJ quantification of protein expression in (A) normalized to loading controls. Results shown are mean ± SEM of three independent experiments. Differences were assessed with Tukey-adjusted mixed model regression. (C) Olive tail moment of hNCCs stably infected with shScr (control), shGHR, or shIGF-1 RNAi. (D) Olive tail moment of hNCCs stably infected with shScr (control) or shIGF-1R, or infected with shIGF-1R RNAi and treated with 500 ng/mL hGH (shScr + GH and shIGF-1R + GH, respectively) and harvested 24 h later. (E) Olive tail moment of hNCCs nucleofected with either pcDNA (control) or pcDNA-GHR and harvested 48 h after nucleofection. Results shown are mean ± SEM of three independent experiments. Differences were assessed with Tukey-adjusted mixed model regression. *P < 0.05, **P < 0.01 vs control. Data are graphed as percentage of control, and statistical testing was performed on individual values.

Figure 3.

Suppression of IGF-1R activity induces GHR expression and increases DNA damage in hNCCs. (A) Representative Western blot of hNCCs treated with 300 and 600 nM PPP and harvested 24 h later. (B) ImageJ quantification of protein expression in (A) normalized to loading control. Results shown are mean ± SEM of three independent experiments. (C) Olive tail moment in hNCCs treated with DMSO control or with 300 and 600 nM PPP and harvested 24 h later. Results shown are mean ± SEM of three independent experiments. Differences were assessed with Tukey-adjusted mixed model regression. *P < 0.05, **P < 0.01 vs control. Data are graphed as percentage of control, and statistical testing was performed on individual values. C, control (DMSO).

Figure 4.

Suppression of IGF-1R activity induces GHR expression and increases DNA damage in human 3D intestinal organoids. (A) Representative Western blot of organoids treated with 100, 300, and 600 nM PPP and harvested 24 h later. (B) ImageJ quantification of protein expression in (A) normalized to loading control. Results shown are mean ± SEM of three independent experiments. (C) Olive tail moment in organoids treated with 500 ng/mL GH (GH), 300 nM PPP (PPP), or 300 nM PPP and 500 ng/mL GH (PPP + GH) and harvested 24 h later. Results shown are mean ± SEM of three independent experiments. Differences were assessed with Tukey-adjusted mixed model regression. *P < 0.05, **P < 0.01 vs control. Data are graphed as percentage of control, and statistical testing was performed on individual values.

Figure 5.

High circulating GH increases colon DNA damage in mice independent of IGF-1R activity. Nude male mice were injected with xenograft bearing HCT116 lentiGH or lentiV and treated with PPP or DMSO (control). (A) Concentration of circulating GH. Results shown are mean ± SEM of 8 to 10 mice per group. Differences are assessed by two-tail Student t test. (B) Olive tail moment. Results shown are mean ± SEM. Four mice per group were analyzed. (C) Representative Western blot of colon tissue. (D) ImageJ quantification of protein expression in (C) normalized to loading controls. Results shown are mean ± SEM of 8 to 10 mice per group. In (B) and (D), differences were assessed with Tukey-adjusted mixed model regression. *P < 0.05, **P < 0.01 vs control. Data are graphed as percentage of control, and statistical testing was performed on individual values.

Figure 6.

GH signaling promotes DNA damage independent of IGF-1R. (A, B) Olive tail moment in the colon of GHR^−/− and WT (A) paired according to age and sex and (B) after treatment with 300 nM PPP or DMSO (control), n = 4 (two males + two females) per group. Data were tested by two-way ANOVA, followed by Tukey test to adjust for multiple group comparisons showing significant interaction of genotype (P = 0.032) and PPP treatments (P = 0.0023) with DNA damage. Pairwise differences were assessed with Tukey-adjusted mixed model regression. (C) Representative Western blot of the colon tissue of WT and GHR^−/− male mice treated with 300 nM PPP. (D) ImageJ quantification of protein expression in (C) normalized to loading controls. Results shown are mean ± SEM of five to six mice per group. Differences were assessed with Tukey-adjusted mixed model regression. *P < 0.05, **P < 0.01 vs control. Data are graphed as percentage of control, and statistical testing was performed on individual values. F, female; M, male.