Non‐pituitary growth hormone enables colon cell senescence evasion

Abstract DNA damage‐induced senescence is initially sustained by p53. Senescent cells produce a senescence‐associated secretory phenotype (SASP) that impacts the aging microenvironment, often promoting cell transformation. Employing normal non‐tumorous human colon cells (hNCC) derived from surgical biopsies and three‐dimensional human intestinal organoids, we show that local non‐pituitary growth hormone (npGH) induced in senescent cells is a SASP component acting to suppress p53. npGH autocrine/paracrine suppression of p53 results in senescence evasion and cell‐cycle reentry, as evidenced by increased Ki67 and BrdU incorporation. Post‐senescent cells exhibit activated epithelial‐to‐mesenchymal transition (EMT), and increased cell motility. Nu/J mice harboring GH‐secreting HCT116 xenografts with resultant high GH levels and injected intrasplenic with post‐senescent hNCC developed fourfold more metastases than did mice harboring control xenografts, suggesting that paracrine npGH enables post‐senescent cell transformation. By contrast, senescent cells with suppressed npGH exhibit downregulated Ki67 and decreased soft agar colony formation. Mechanisms underlying these observations include npGH induction by the SASP chemokine CXCL1, which attracts immune effectors to eliminate senescent cells; GH, in turn, suppresses CXCL1, likely by inhibiting phospho‐NFκB, resulting in SASP cytokine downregulation. Consistent with these findings, GH‐receptor knockout mice exhibited increased colon phospho‐NFκB and CXCL1, while GH excess decreased colon CXCL1. The results elucidate mechanisms for local hormonal regulation of microenvironmental changes in DNA‐damaged non‐tumorous epithelial cells and portray a heretofore unappreciated GH action favoring age‐associated epithelial cell transformation.


Figure S2 .
Figure S2.GH is induced in senescent cells and is a SASP component.(A) ImageJ quantification of senescence markers in hNCC treated with 50 μM etoposide (Etop) or DMSO for 48h, and analyzed 7 days after beginning treatment.(B) Western blot of GH, p16 and PCNA expression and (C) Real time PCR results of GH and IGF1 in hNCC stably infected with lenti shScr (as control) or lenti shGH and treated with 50 μM

Figure S3 .
Figure S3.GH triggers proliferation in senescent cells.(A) Representative images of non-senescent parental hNCC not treated (NT) and treated with GH only.Ki67 (brown) expression in senescent (blue) SA-β-gal-positive cells.Cells treated with etoposide only or with etoposide and GH are depicted in Figure 2A.(B) Senescence markers of hNCC treated with 50 µM etoposide for 48 hours and analyzed 6 days after beginning treatment.(C) BrdU incorporation.hNCC that reached replicative

Figure
Figure S4.GH suppresses p53 in senescent cells and affects colony formation.(A-B) ImageJ quantification of p53 in (A) hNCC infected with either lenti-shScr as control or lenti-shGH, treated with 50 μM etoposide (Etop) or left untreated (Control) for 48h, and grown for an additional 4 days; and (B) hNCC pre-treated with 500 ng/mL GH and indicated doses of etoposide, then analyzed 24h later.Results are shown as mean ± SEM of at least 3 replicate measurements normalized to loading control.Differences were analyzed in A by one-way ANOVA and in B by two-way ANOVA followed by ad-hoc Tukey's test to adjust for multiple comparisons.(C) Western blot of GH and 53 in colon tissue of hypophysectomized rats after 5 injection of either Oxaliplatin (Oxa) (i.p. 4 mg/kg) alone or oxaliplatin with pegvisomant (PEG) (s.q.0.57 mg/kg).Lower panel: ImageJ quantification of GH and p53.For (A-C) C, control.(D-F) hNCC were pretreated with 50 µM etoposide for 48h, sorted for senescence, and cultured in the presence of 500 ng/mL GH 10 days after plating.(D) Number of colonies per well.(E) Western blot of cleaved caspase 3. (F) Colony size.One dot represents one experiment.In C and E, results are shown as mean ± SEM.In C, D and F results were analyzed by two tailed Student's t test.Results are graphed as percent of control, but statistical testing was performed on raw numbers.*, p<0.05; **, p<0.01 vs control.

Figure S5 .
Figure S5.Paracrine GH triggers proliferation and EMT and suppresses DNA damage pathway.(A) ImageJ quantification of proliferation, stemness, and EMT markers in hNCC line #1 treated with 50 µM etoposide, sorted for senescence 7 days later, and cultured for 10 days in the presence of or absence of GH; cells were then sorted again for SA-β-gal positivity.Only SA-β-gal negative (post-senescent cells) were analyzed.(B) Organoids were infected with lentiV or lentiGH (both expressing GFP) and cultured for 5 weeks; organoid cells were then sorted for GFP-positive and GFPnegative expression and only GFP-negative cells co-cultured with either lentiV or lentiGH organoid cells were analyzed.ImageJ quantification of DNA damage and repair proteins shown.In A and C, results are shown as mean ± SEM of triplicate measurements normalized to loading control.In A, results were analyzed by one-way ANOVA followed by ad-hoc Tukey's test to adjust for multiple comparisons.In C, results were analyzed by two-tailed Student's t-test.Results are graphed as percent of control, but statistical testing was performed on raw numbers.*, p<0.05; **, p<0.01 vs comparator.

Figure S6 .
Figure S6.CXCL1 is induced in senescent cells.(A-C) Image J quantification of CXCL1 in different models of senescence.(A) Oncogene-induced senescence.Cells were infected with lentivirus expressing constitutively activated HRAS oncogene (lentiV12HRAS) or empty vector (lentiV-CMV) and analyzed 7 days later.(B) DNA damage induced senescence.Cells were exposed to indicated doses of UVC light or left untreated (NT) and analyzed 6 days later.(C) Replicative senescence.Cells were passaged until proliferation was significantly reduced.Cells from passage 8 (p.8) and passage 61 (p.61) were compared.In all experiments, CXCL1 expression was normalized

Figure S7 .
Figure S7.CXCL1 induces GH expression.(A) ImageJ quantification of proteins from organoids treated with 100 ng/mL CXCL1 for 96h.(B) Real-time PCR of senescent organoids treated with CXCL1 for 96h.Experiment was repeated 3 times.In A and B, differences with control (NT) were analyzed with two-tailed Student's t-test.In B, results are shown as mean ± SEM of triplicate measurements, expressed as fold-change vs control taken as 1. (C) ImageJ quantification of hNCC treated with indicated doses of CXCL1 for 96h.(D) ImageJ quantification of hNCC infected with lenti-shCXCL1 or lenti-shScr and analyzed 3 days after infection.In C and D, results were analyzed by one-and two-way ANOVA, respectively, followed with ad-hoc Tukey's test to adjust for multiple comparisons.Results are shown as mean ± SEM of at least 3 replicate measurements normalized to loading controls.Results are graphed as percent of

Figure S8 .
Figure S8.GH suppresses CXCL1 expression and secretion in senescent cells.(A-C) hNCC were infected with either lenti-shGH or lenti-shScr, treated with 50 µM etoposide (Etop) to induce senescence for 48h or left untreated (C) and analyzed 6 days later.(A) ImageJ quantification of CXCL1 in cells.(B) ImageJ quantification of CXCL1 in culture medium.(C) Real-time PCR of CXCL1 mRNA.In A-C, differences with respective controls were analyzed by two-way ANOVA followed by ad-hoc Tukey's test to adjust for multiple comparisons.Results are shown as mean ± SEM of triplicate measurements and in C are depicted as fold-change vs control taken as 1. (D) ImageJ quantification of proteins in organoids 1 week after infection with lenti-shGH or lenti-shScr.(E) ImageJ quantification of proteins in organoids 1 week after infection with lentiGH or lentiV.(F) Image J quantification of proteins in hNCC treated with 50 µM etoposide for 2 days, sorted for senescence on day 3, and treated with GH for an additional 24h.Results are shown as mean ± SEM of triplicate measurements normalized to loading control.In D-F, results were analyzed by two-tailed Student's ttest.Results are graphed as percent of control, but statistical testing was performed on raw numbers.*, p<0.05; **, p<0.01 vs comparator.