Estrogen inhibits GH signaling by suppressing GH-induced JAK2 phosphorylation, an effect mediated by SOCS-2

Abstract

Oral estrogen administration attenuates the metabolic action of growth hormone (GH) in humans. To investigate the mechanism involved, we studied the effects of estrogen on GH signaling through Janus kinase (JAK)2 and the signal transducers and activators of transcription (STATs) in HEK293 cells stably expressing the GH receptor (293GHR), HuH7 (hepatoma) and T-47D (breast cancer) cells. 293GHR cells were transiently transfected with an estrogen receptor-alpha expression plasmid and luciferase reporters with binding elements for STAT3 and STAT5 or the beta-casein promoter. GH stimulated the reporter activities by four- to sixfold. Cotreatment with 17beta-estradiol (E(2)) resulted in a dose-dependent reduction in the response of all three reporters to GH to a maximum of 49-66% of control at 100 nM (P < 0.05). No reduction was seen when E(2) was added 1-2 h after GH treatment. Similar inhibitory effects were observed in HuH7 and T-47D cells. E(2) suppressed GH-induced JAK2 phosphorylation, an effect attenuated by actinomycin D, suggesting a requirement for gene expression. Next, we investigated the role of the suppressors of cytokine signaling (SOCS) in E(2) inhibition. E(2) increased the mRNA abundance of SOCS-2 but not SOCS-1 and SOCS-3 in HEK293 cells. The inhibitory effect of E(2) was absent in cells lacking SOCS-2 but not in those lacking SOCS-1 and SOCS-3. In conclusion, estrogen inhibits GH signaling, an action mediated by SOCS-2. This paper provides evidence for regulatory interaction between a sex steroid and the GHJAKSTAT pathway, in which SOCS-2 plays a central mechanistic role.

Figure 1

Effects of E₂ on GH transcriptional activity in 293GHR cells. (a) Cells were transiently transfected with an ERα expression plasmid and STAT5 reporter and treated with 500 ng/ml GH and 250 nM Dex and E₂ at indicated concentrations for 6 h. *, P < 0.05; †, P < 0.01. (b) Effect of anti-estrogen. After pretreatment for 30 min with 1 μM ICI182780, the cells were incubated with 500 ng/ml GH and 250 nM Dex and E₂ for 6 h; vs. control with E₂ at corresponding concentrations, *, P < 0.05. (c) Time course of E₂ effect. The cells were treated with a 1-h pulse of 500 ng/ml GH, washed, and incubated in fresh medium for a further 5 h before luciferase activity measurement. E₂ at 100 nM was added at indicated time before (minus) or after (plus) the GH pulse. For −2, −1, and 0 h, E₂ was replaced in fresh medium during the 5-h incubation without GH. Control cultures without E₂ were set up for each time point. Luciferase activity of E₂-treated samples is presented as percentage of non-E₂-treated control of the same time point; vs. controls at the same time point: *, P < 0.01; †, P < 0.05. (d) STAT3 reporter. Cells expressing exogenous STAT3 and ERα were transiently transfected with the STAT3 reporter and treated with 500 ng/ml GH and E₂ at indicated concentrations for 6 h. *, P < 0.05. (e) β-Casein promoter reporter. Cells expressing exogenous STAT5a and ERα were transiently transfected with the β-casein promoter reporter and treated with 500 ng/ml GH and 250 nM Dex and E₂ for 24 h. *, P < 0.05.

Figure 2

Effects of E₂ on transcriptional activities of prolactin and IL6. (a) T-47D cells transfected with the STAT5 reporter were treated for 24 h with 500 ng/ml prolactin and 250 nM Dex and E₂. *, P < 0.05. (b) 293GHR cells transfected with the STAT3 reporter and expression plasmids for ERα and STAT3 were treated for 24 h with 500 ng/ml IL6 and E₂. *, P < 0.05.

Figure 3

Western blotting of tyrosine phosphorylation (PY) of STAT3 (a) and STAT5 (b) in 293GHR cells. The cells were treated for 2.5 h with 100 nM E₂ and then for 1 h with 500 ng/ml GH (also with 250 nM Dex in the case of STAT5), as indicated. Cell lysates were immunoprecipitated with antibodies for STAT3 or STAT5 and Western blotted with an anti-phosphotyrosine antibody (Upper) and the anti-STAT antibodies (Lower).

Figure 4

Coimmunoprecipitation of STAT5 with GR and ERα in 293GHR cells. Western blots of STAT5 (a), GR (b), and ERα (c) in 293GHR cells treated for 1 h with 250 nM Dex, without or with 500 ng/ml GH and 100 nM E₂, and precipitated (IP) with antibodies against STAT5, GR, or ERα, as indicated. The bottom bands (Ig) were the heavy chains of the antibodies used for immunoprecipitation.

Figure 5

Regulation of GH-induced phosphorylation of JAK2 by E₂ in 293GHR (a) and HuH7 (b) cells, or by Dex in 293GHR cells (c). The cells were treated for 2.5 h with 100 nM E₂ or 250 nM Dex and then for 2 min with 500 ng/ml GH as indicated, followed by immunoprecipitation for JAK2 and Western blotting for phosphorylated (PY) and total JAK2. Arrowheads indicate nonspecific bands.

Figure 6

Effects of actinomycin D and vanadate in 293GHR cells. Western blots of GH-induced tyrosine phosphorylation (PY) of JAK2 in 293GHR cells treated for 1 h with 5 μg/ml actinomycin D (a; Act D) or 1 mM vanadate (b), followed by treatment for 2.5 h with 100 nM E₂ and for 2 min with 500 ng/ml GH, as indicated. Arrowheads indicate nonspecific bands.

Figure 7

SOCS. (a) Effects of E₂ and GH on SOCS-2 expression in HEK293 cells expressing ERα and GHR. The cells were treated with 100 nM E₂ or 500 ng/ml GH for time as indicated. SOCS-2 mRNA abundance was quantified by reverse transcription and real-time PCR; vs. respective control at time 0: *, P < 0.01; †, P < 0.05. (b) Effects of E₂ (1 nM) on STAT5 activation by GH (500 ng/ml) in wild-type (WT) and SOCS-deficient fibroblasts expressing ERα and GHR. Results are expressed as percentages of GH-treated control for the respective cell lines. *, P < 0.01. (c) Western blots of GH-induced JAK2 phosphorylation (PY) in the fibroblasts treated for 2.5 h with 1 nM E₂ and then for 2 min with 500 ng/ml GH, as indicated. Arrowheads indicate nonspecific bands. (d) Effects of SOCS-3 siRNA transfection on GH activation of STAT5 reporter in HEK293 cells expressing SOCS-3. The cells were transiently transfected with a GHR expression plasmid, a STAT5 reporter, with or without SOCS-3 siRNA (8 nM) and a SOCS-3 expression plasmid (50 ng), as indicated, and then treated for 3 h with 500 ng/ml GH and 250 nM Dex. Data shown are the mean ± SE of triplicate measurements in a representative experiment, which was repeated twice; vs. respective control without SOCS-3: *, P < 0.01, †, P < 0.05; vs. sample without siRNA and with SOCS-3: ‡, P < 0.01. (e) Effects of SOCS-3 siRNA on E₂ inhibition of GH-induced STAT5 activity in HEK293 cells. Cells expressing ERα and GHR were cotransfected with or without SOCS-3 siRNA (8 nM) and treated for 3 h with 500 ng/ml GH and 250 nM Dex and E₂ at indicated concentrations. Data shown are the mean ± SE of triplicate measurements in a representative experiment, which was repeated twice; vs. respective control: *, P < 0.01; †, P < 0.05.