Distinct mechanisms of induction of hepatic growth hormone resistance by endogenous IL-6, TNF-α, and IL-1β

Abstract

During inflammation, the liver becomes resistant to growth hormone (GH) actions, leading to downregulation of the GH target gene IGF-I and activation of catabolism. Proinflammatory cytokines IL-6, TNF-α, and IL-1β are critically involved in the pathogenesis of hepatic GH resistance. However, the mechanisms used by endogenous IL-6, TNF-α, and IL-1β to inhibit the hepatic GH-IGF-I pathway during inflammation are not fully understood. Here, we show that TNF-α and IL-1β inhibited GH receptor (GHR) expression but had minor effects on the downstream suppressor of cytokine signaling (SOCS)3, while IL-6 induced SOCS3 expression but had no effect on GHR expression in Huh-7 cells. Consistent with the in vitro observations, neutralization of TNF-α and IL-1β in mouse models of inflammation did not significantly alter SOCS3 expression stimulated by inflammation but restored GHR and IGF-I expression suppressed by inflammation. Neutralization of IL-6 did not alter inflammation-suppressed GHR expression but drastically reduced the inflammation-stimulated SOCS3 expression and restored IGF-I expression. Interestingly, when the GH-IGF-I pathway was turned off by maximal inhibition of GHR expression, IL-6 and SOCS3 were no longer able to regulate IGF-I expression. Taken together, our results suggest that TNF-α/IL-1β and IL-6 use distinct mechanisms to induce hepatic GH resistance, with TNF-α and IL-1β acting primarily on GHR and IL-6 acting primarily on SOCS3. IL-6 action may be superseded by factors such as TNF-α and IL-1β that inhibit GHR expression.

growth hormone (GH) plays a pivotal role in activating anabolism in many organs, including the liver, skeletal muscle, and bone. GH binds to the preformed GH receptor (GHR) dimer and induces the formation of activated GHR-JAK2 complex, which recruits and phosphorylates signal transducer and activator of transcription (STAT)5b. Phosphorylated STAT5b translocates into the nucleus to regulate transcription of GH target genes including insulin-like growth factor (IGF)-I, a main mediator of the growth-promoting effects of GH (2). Most circulating IGF-I is produced by the liver, where IGF-I expression increases in response to GH stimulation (28). However, during systemic acute or chronic inflammation, the liver becomes resistant to GH action, with decreased hepatic IGF-I production despite normal or even elevated circulating levels of GH. The consequences of GH resistance include postnatal linear growth failure, muscle wasting, poor wound healing, and cachexia (2).

Proinflammatory cytokines IL-6, TNF-α, and IL-1β have been shown to be critical mediators of inflammation-induced GH resistance. Previous studies demonstrated that exogenous IL-6, TNF-α, and IL-1β inhibit GH signaling in vitro and in vivo (1, 3, 5, 6, 8–10, 13–15, 18, 19, 27, 32). However, the precise roles of each individual endogenous cytokine in the development of GH resistance when these cytokines are all present during inflammation are not clarified. A previous study demonstrated that deletion of IL-6 gene in mice prevented the inhibition of STAT5b phosphorylation by lipopolysaccharide (LPS) in livers, whereas in TNF receptor 1 knockout mice, LPS was able to inhibit STAT5b phosphorylation (10). These results suggest differential actions of endogenous IL-6 and TNF-α in regulating GH signaling.

The mechanisms used by individual endogenous cytokines to inhibit GH signaling remain incompletely characterized. Studies revealed that two main mechanisms are involved in the induction of hepatic GH resistance by IL-6, TNF-α, and IL-1β, i.e., downregulation of GHR and upregulation of the members of the suppressors of cytokine signaling (SOCS) family. Although it has been demonstrated that TNF-α inhibits hepatic GHR expression in vivo, there is controversy over whether IL-6 and IL-1β play a role in regulating hepatic GHR expression (1, 9–11, 25, 27).

Studies have demonstrated that SOCS family members including SOCS1–3 and cytokine-inducible SH2-containing protein-1 (Cis) can inhibit GH signaling by preventing GHR/JAK2-dependent STAT5b activation (16, 22). Among the SOCS family members, SOCS3 is found to be the major mediator of inflammation-induced GH resistance (4, 7). However, whether IL-6, TNF-α, and IL-1β differently regulate SOCS3 is an open question.

Dragon is one of the three repulsive guidance molecule (RGM) family members, RGMa, RGMb (Dragon), and RGMc. We recently discovered that Dragon functions as a coreceptor that enhances bone morphogenetic protein (BMP) signaling (23, 31). Our previous study identified Dragon KO mice as a model of chronic inflammation (31).

In the present study, we tested our hypothesis that endogenous IL-6, TNF-α, and IL-1β may use different mechanisms to induce GH resistance during inflammation. We investigated the role of IL-6, TNF-α, and IL-1β in the expression of GHR and SOCS3, the two key factors that mediate inflammation-induced hepatic GH resistance by using Huh-7 human hepatoma cells, LPS-induced inflammation, and the Dragon KO mouse model of chronic inflammation. We found that TNF-α/IL-1β and IL-6 differently regulated GHR and SOCS3 in Huh-7 cells. In line with these in vitro observations, we revealed that TNF-α/IL-1β and IL-6 use distinct mechanisms to inhibit IGF-I expression in inflammation-induced GH resistance in liver.

MATERIALS AND METHODS

Animals.

Male C57BL/6 wild-type (WT) mice at 6–7 wk or 11–12 days of age were used for LPS injection and administration of IL-6, TNF-α, and IL-1β antibodies.

Homozygous Dragon KO mice (57/B6/129) die between 2 and 3 wk after birth, but heterozygous mice are normal (31). Heterozygous animals were bred to obtain homozygous Dragon KO mice. Genotyping was performed as previously described by us (31). WT and homozygous Dragon KO mice at 12 days of age were used for administration of IL-6, TNF-α, and IL-1β antibodies.

Mice were housed in groups under specific pathogen-free conditions with a 12:12-h light-dark cycle and ad libitum access to food (Extruded Global Rodent Diet, Harlan, USA) and water. China poplar sawdust was used for bedding. The room temperature and humidity were maintained at 18–23°C and 60–70%, respectively. All the procedures were performed in accordance with The Chinese University of Hong Kong animal care regulations.

Antibodies and LPS administration and tissue collection.

Male C57BL/6 mice at similar weights received an intraperitoneal injection of LPS (E. coli O127:B8, Sigma-Aldrich) in PBS at a dose of 2.5 μg/g body wt for 6- to 7-wk-old mice or 1 μg/g body wt for 11- to 12-day-old mice, and control mice were injected with PBS. Mice were euthanized by cervical dislocation 3–24 h after LPS injection. Livers and pituitaries were harvested for preparation of proteins and total RNA.

For anti-IL-6 antibody administration experiments, C57BL/6 mice at 6–7 wk of age received an intraperitoneal injection of anti-IL-6 antibody (R&D Systems, MAB406) at a dose of 5 μg/g body wt 1 h or 9.5 h after LPS injection. Control mice received an injection of nonspecific rat IgG (R&D Systems, 6-001-A) at the same dose. Four hours after antibody injection, mice were euthanized by cervical dislocation, and livers were harvested for preparation of proteins and total RNA.

For TNF-α antibody and IL-1β antibody administration experiments, male C57BL/6 mice received an intraperitoneal injection of a combination of TNF-α antibody (R&D Systems, AB-410-NA) and IL-1β antibody (R&D Systems, AB-401-NA) at a dose of 10 μg/g body wt for each antibody for 6- to 7-wk-old mice or 5 μg/g body wt for 11- to 12-day-old mice 9.5 h after LPS injection. Control mice received an injection of nonspecific goat IgG (R&D Systems, AB-108-C) at the same dose. Four hours after antibody injection, mice were euthanized by cervical dislocation, and livers were harvested for preparation of proteins and total RNA.

For experiments on Dragon KO mice, Dragon KO mice received an intraperitoneal injection of IL-6 antibody or a combination of TNF-α antibody and IL-1β antibody at a dose of 5 μg/g body wt. WT and Dragon KO littermates injected with nonspecific IgG at the same dose were used as controls; 15–16 h later, the mice received another injection of the antibodies or nonspecific IgG; and 5–6 h after the second injection, livers and pituitaries were harvested for preparation of proteins and total RNA.

The antibody doses were chosen based on previous studies where 100 μg of the same IL-6, TNF-α, or IL-1β antibody given (ip) to an adult mouse effectively blocked respective inflammatory responses in mice (17, 24, 26).

Cell culture and inflammatory cytokine treatment.

Huh-7 human hepatoma cells were cultured in DMEM supplemented with 10% FBS. Cells were seeded into 12-well plates and grown to 80% confluence before being starved overnight in DMEM supplemented with 0.1% BSA. Cells were treated with recombinant human IL-6, TNF-α, and IL-1β (R&D Systems) before they were collected for measurement of GHR and SOCS3 mRNA expression.

Transient transfection.

Flag-tagged mouse SOCS3, flag-tagged WT rat Stat5b (WT Stat5b), flag-tagged constitutively active rat Stat5b (N642H, CA Stat5b), or a combination of SOCS3 with WT Stat5b or CA Stat5b was transfected into Huh-7 cells using Lipofectamine 2000 (Invitrogen). Approximately 24 h after transfection, cells were incubated with or without GH (R&D Systems, 1067-GH) at 500 ng/ml in FBS-free DMEM supplemented with 0.1% BSA for another 24 h. Assays to measure mRNA levels of IGF-I and SOCS3 were performed 48 h after transfection.

Small interfering RNA knockdown.

Human GHR small interfering RNAs (siRNAs) were purchased from Shanghai GenePharma, (Shanghai, China). A mixture of the following four sense GHR siRNA sequences were used: 5′-GGACUCAAGAAUGGAAAGATT-3′, 5′-GCGUGUGAGAUCCAAACAATT-3′, 5′-GCUAACAGUGAUGCUAUUUTT-3′, and 5′-GCAGCCCAGUGUUAUCCAATT-3′. Scrambled control siRNA was purchased from Ambion. Huh-7 cells were transfected with siRNA duplexes (80 nM) in combination with or without SOCS3 cDNA using Lipofectamine 2000. Cells were then incubated with or without GH (500 ng/ml) in FBS-free DMEM supplemented with 0.1% BSA for 24 h. Assays to measure mRNA levels of IGF-I and GHR were performed 48 h after transfection.

Western blotting.

Liver tissues and Huh-7 cells were lysed as described previously (31). GHR protein in mouse livers was analyzed by Western blotting using a previously validated rabbit anti-GHR antibody (AL-47) (20, 21, 33). Membranes were reprobed with a GAPDH antibody from Santa Cruz Biotechnology (sc-25778). All the antibodies were diluted at 1:1,000. Detection was performed with the enhanced chemiluminescence method (Millipore). Densitometry was performed using Bio-Rad Quantity One software with signals lying in the linear range.

RNA isolation and real-time PCR analysis.

Total RNA was isolated from liver and pituitary tissues or Huh-7 cells using a Pure Link RNA mini kit (Ambion) according to the manufacturer's instructions. First-strand cDNA synthesis was performed using the PrimeScript RT reagent kit (TAKARA) and was amplified with the primers as shown in Table 1 (mouse GHR, human GHR, and IGF-I) or as previously described by us [mouse IL-6, TNF-α, IL-1β, and RPL19 (ribosomal protein L19); human RPL19 (31)] or by others [human/mouse SOCS3 and mouse Cis and IGF-I (10)]. RPL19, a housing keeping gene, was used as the internal control.

Table 1.

Sequences, expected product sizes, and GenBank accession numbers for primers used in RT-PCR

Genes	Species	Primers Forward (5′-3′)	Primers Reverse (5′-3′)	Size, bp	Accession No.
GHR	H	GTCTGCAAAGTGTTAATCCAGGC	CTCTCGCTCAGGTGAACGG	83	NM_001242462
IGF-I	H	GCTCTTCAGTTCGTGTGTGGA	CGACTGCTGGAGCCATACC	71	NM_001111285
GHR	M	ACAGTGCCTACTTTTGTGAGTC	GTAGTGGTAAGGCTTTCTGTGG	133	NM_010284

H, human; M, mouse; GHR, GH receptor.

Statistical analysis.

All data are represented as means ± SD of independent replicates (n ≥ 3). Comparisons between different time points and different cytokines in Fig. 1 were made by two-way ANOVA with Bonferroni's multiple comparison posttest. Comparisons between different time points after LPS treatment in Fig. 2 and between different treatments in Figs. 3 and 5–9 were made by one-way ANOVA with Tukey's multiple comparison posttest. Student's t-test was used for comparison of two normally distributed groups in Figs. 4 and 9. A P value of less than or equal to 0.05 was considered statistically significant.

Fig. 1.

IL-6, TNF-α, and IL-1β differently regulate GH receptor (GHR) and SOCS3 expression in Huh-7 cells. A and B: cells were incubated with IL-6, TNF-α, or IL-1β (0.5 nM) for 0, 4, 8, and 12 h. Expression levels of mRNAs for GHR and SOCS3 were determined by real-time PCR. C and D: cells were incubated with increasing amounts of IL-6, TNF-α, or IL-1β for 8 h before being harvested for analysis of GHR and SOCS3 mRNA levels. E and F: cells were incubated with IL-6, TNF-α, IL-1β, or a combination of TNF-α and IL-1β (0.5 nM for each cytokine) for 8 h. Means without a common letter are significantly different: P < 0.05; n = 3.

Fig. 2.

Effects of LPS treatment on IL-6, TNF-α, IL-1β, GHR, SOCS3, and IGF-I expression in livers of 6- to 7-wk-old mice. Male mice at 6–7 wk of age were injected with or without LPS (2.5 μg/g body wt ip). Liver samples were collected at 3, 6, 9, 12, and 24 h after LPS injection for real-time PCR analysis for IL-6 (A), TNF-α (B), IL-1β (C), GHR (D), SOCS3 (E), and IGF-I (F) mRNA expression. GHR protein levels in liver were examined by Western blotting (G) and quantified by densitometry (H). GHR for the 9- and 12-h time points was subjected to longer exposure to reveal the trace expression of GHR in LPS-treated mice. GAPDH was the loading control for Western blotting; ribosomal protein L19 (RPL19) was the internal control for real-time PCR analysis. Comparisons between LPS-treated groups: means without a common letter are significantly different, P < 0.05. Comparisons between LPS-treated and control groups: *P < 0.05, **P < 0.01, ***P < 0.001; n = 3.

Fig. 3.

Effects of neutralization of TNF-α and IL-1β on GHR, SOCS3 and IGF-I expression in the livers of LPS-treated mice. A: male mice at 6–7 wk or 11–12 days of age were injected with LPS at 2.5 or 1 μg/g body wt ip, respectively. Anti-TNF-α and anti-IL-1β antibodies (10 μg/g body wt ip for adult mice or 5 μg/g body wt ip for young mice) or normal IgG were injected 9.5 h after LPS injection, and mice were euthanized 4 h after antibody injection. B: relative hepatic GHR mRNA levels in mice at 6–7 wk of age. C: hepatic GHR protein levels in mice at 6–7 wk of age. GHR proteins were examined by Western blotting (top) and quantified by densitometry (bottom). D and E: relative hepatic SOCS3 and IGF-I mRNA levels in mice at 6–7 wk of age. F–H: relative hepatic GHR, SOCS3, and IGF-I mRNA levels in mice at 11–12 days of age (D11–12). GAPDH is the loading control for Western blotting; RPL19 is the internal control for real-time PCR analysis. Means without a common letter are significantly different: P < 0.05; n = 5 for B and D–H, n = 3 for C.

Fig. 5.

Effects of neutralization of TNF-α and IL-1β on GHR, SOCS3, and IGF-I expression in livers of Dragon KO mice at 12 days of age. Hepatic GHR (A), SOCS3 (C), and IGF-I (D) mRNA levels were measured by real-time PCR in WT mice administered with normal IgG (WT/normal IgG) and in Dragon KO littermates administered with normal IgG (KO/normal IgG) or anti-TNF-α and anti-IL-1β antibodies (KO/anti-TNF-α + anti-IL-1β). GHR protein levels in liver (B) were examined by Western blotting (top) and quantified by densitometry (bottom). L1, L2, and L3, litters from 3 different parents. GAPDH is the loading control for Western blotting; RPL19 is the internal control for real-time PCR analysis. Means without a common letter are significantly different: P < 0.05; n = 6–7 for A, C, and D, n = 3 for B.

Fig. 9.

SOCS3 activity in inhibiting the GH-IGF-I pathway is GHR dependent in Huh-7 cells. A–D: SOCS3 acts on GH signaling upstream of STAT5b. A: Huh-7 cells were transfected with control or Flag-SOCS3 cDNA. Cells were then incubated in the absence or presence of 500 ng/ml GH for 24 h. Cell lysates were analyzed for IGF-I mRNA expression by real-time PCR. B: Huh-7 cells were transfected with control or Flag-SOCS3 cDNA in the presence or absence of WT Stat5b cDNA. Cells were then incubated in the absence or presence of 500 ng/ml GH for 24 h. Cell lysates were analyzed for IGF-I mRNA by real-time PCR. C and D: Huh-7 cells were transfected with control or Flag-SOCS3 cDNA in the presence or absence of constitutively active (CA) Stat5b cDNA. Cell lysates were analyzed for SOCS3 (C) and IGF-I (D) mRNA expression by real-time PCR. E and F: effects of inhibition of GHR expression on SOCS3 activity in regulating IGF-I expression. Huh-7 cells were transfected with control or Flag-SOCS3 cDNA in the presence of control siRNA or GHR siRNA. Cells were then incubated in the absence or presence of 500 ng/ml GH for 24 h. Cell lysates were analyzed for GHR (E) and IGF-I (F) mRNA expression by real-time PCR. RPL19 was the internal control for real-time PCR analysis. A–D and F: means without a common letter are significantly different: P < 0.05; E: ***P < 0.001, n = 3.

Fig. 4.

Inflammation and hepatic GH resistance in Dragon (repulsive guidance molecule b) knockout (KO) mice. A–C: expression of IL-6, TNF-α, and IL-1β in livers of Dragon KO mice at 12 days of age. Relative IL-6 (A), TNF-α (B), and IL-1β (C) mRNA levels in livers of wild-type (WT) and Dragon KO mice were measured by real-time PCR. D–I: hepatic GH resistance in Dragon KO mice at 12 days of age. WT and Dragon KO mice were weighed (D). mRNA levels for GH in pituitary (E), and GHR (F), SOCS3 (H), and IGF-I (I) in liver were measured by real-time PCR. GHR protein levels in liver (G) were examined by Western blotting (top) and quantified by densitometry (bottom). L1, L2, and L3, litters from 3 different parents. GAPDH is the loading control for Western blotting; RPL19 is the internal control for real-time PCR analysis. *P < 0.05, **P < 0.01, ***P < 0.001; n = 6–7 for A–F and H–J, n = 3 for G.

RESULTS

TNF-α and IL-1β inhibited GHR expression and IL-6 increased SOCS3 expression in Huh-7 cells.

Previous studies identified downregulation of GHR and upregulation of SOCS3 as two critical mechanisms that mediate inflammation-induced GH resistance. To determine whether IL-6, TNF-α, and IL-1β differently regulate GHR and SOCS3 expression, we performed head-to-head comparisons among the three cytokines in regulating GHR and SOCS3 expression in vitro. We used Huh-7 human hepatoma cells because these cells express 350-fold higher levels of GHR than Hep3B and HepG2 cells (data not shown). We first performed time courses for IL-6, TNF-α, or IL-1β treatments. Huh-7 cells were incubated with IL-6, TNF-α, or IL-1β (0.5 nM) for 0, 4, 8, and 12 h before the cells were harvested for analysis for GHR and SOCS3 mRNA levels (Fig. 1, A and B). GHR mRNA levels declined from 0 to 4 h before they leveled off from 4 to 12 h after TNF-α or IL-1β treatments, whereas GHR mRNA levels were not altered by IL-6 treatments (Fig. 1A). SOCS3 mRNA levels drastically increased from 0 to 4 h before they declined from 4 to 12 h after IL-6 treatments. SOCS3 mRNA levels only slightly increased from 0 to 4 h and then remained unchanged from 4 to12 h after TNF-α or IL-1β treatments (Fig. 1B). SOCS3 mRNA levels were much higher in the IL-6-treated cells than in the TNF-α- or IL-1β-treated cells at all the time points (Fig. 1B).

We next examined effects of different doses of IL-6, TNF-α, or IL-1β on GHR and SOCS3 expression. Huh-7 cells were incubated with increasing amounts of IL-6, TNF-α, and IL-1β (0–7.5 nM) for 8 h. As shown in Fig. 1, C and D, TNF-α and IL-1β suppressed GHR mRNA expression in a dose-dependent manner, whereas IL-6 had no effect (Fig. 1C). In contrast, IL-6 stimulated SOCS3 expression in a dose-dependent manner, but TNF-α and IL-1β had only small effects (Fig. 1D).

We also examined whether TNF-α and IL-1β synergize in their effects on GHR and SOCS3 expression. Huh-7 cells were incubated with IL-6, TNF-α, IL-1β, or a combination of TNF-α and IL-1β (0.5 nM for each cytokine) for 8 h. As shown in Fig. 1E, TNF-α and IL-1β each inhibited GHR mRNA expression, and a combination of the two cytokines further reduced GHR mRNA levels. TNF-α and IL-1β each slightly increased SOCS3 expression, but a combination of the two cytokines did not further increase SOCS3 expression. These results suggest that TNF-α and IL-1β, but not IL-6, inhibit GHR expression and that TNF-α and IL-1β may synergistically inhibit GHR expression. IL-6 is much more potent in stimulating SOCS3 expression than TNF-α and IL-1β in Huh-7 cells.

TNF-α and IL-1β inhibited hepatic GHR expression but had no significant effects on SOCS3 expression in mice.

To investigate the effects of endogenous TNF-α and IL-1β on hepatic GHR and SOCS3 expression during inflammation, we first used the LPS-induced inflammation mouse model. Male mice at 6–7 wk of age were injected with or without LPS (2.5 μg/g body wt ip), and liver samples were collected at 3, 6, 9, 12, and 24 h after the LPS injection. As expected, LPS treatment induced rapid increases in expression of hepatic IL-6, TNF-α, and IL-1β (Fig. 2, A–C), followed by hepatic GH resistance as shown by a decrease in IGF-I mRNA expression, which was associated with a decrease in GHR mRNA expression and an increase in SOCS3 mRNA expression (Fig. 2, D–F). GHR protein expression was not altered at 3 h after LPS treatment, and it was inhibited by 37.1% at 6 h. GHR protein was maximally inhibited by 77.4% at 12 h after LPS treatment (Fig. 2, G and H).

Since TNF-α and IL-1β showed similar effects on GHR and SOCS3 expression in Huh-7 cells, we administered a combination of anti-TNF-α antibody and anti-IL-1β antibody to 6- to 7-wk-old mice 9.5 h after LPS treatment. The mice were euthanized 4 h after the antibody injection (Fig. 3A). LPS treatment inhibited GHR mRNA and protein expression by 85.8 and 85.7%, respectively (compare bars 2 and 1, Fig. 3, B and C). Neutralization of TNF-α and IL-1β increased GHR mRNA and protein levels (compare bars 3 and 2, Fig. 3, B and C). LPS dramatically induced SOCS3 expression (compare bars 2 and 1, Fig. 3D), but neutralization of TNF-α and IL-1β had no statistically significant effects on SOCS3 expression in LPS-treated mice (p = 0.086, compare bars 3 and 2, Fig. 3D). In association with the increased GHR expression, IGF-I expression in the liver was increased by neutralization of TNF-α and IL-1β in LPS-treated mice (Fig. 3E). Similar results were observed when young mice at 11–12 days of age were used (Fig. 3, F–H).

We also utilized our previously established Dragon KO mouse model of chronic inflammation (31). These mice die between 2 and 3 wk after birth. They demonstrated increased levels of IL-6, TNF-α, and IL-1β in the lung (31) and in the liver (Fig. 4, A–C) compared with WT mice, confirming that there is chronic inflammation in multiple organs including the liver and lung in Dragon KO mice. Interestingly, despite the increased GH expression by the pituitary (Fig. 4E), Dragon KO mice at postnatal day 12 showed GH resistance as indicated by decreased IGF-I expression and body weight, which were associated with decreased GHR expression and increased SOCS3 expression in the liver (Fig. 4, D–J) and muscle (data not shown). Of note, hepatic GHR protein levels were reduced by 79.6% in Dragon KO mice compared with WT mice (Fig. 4G).

We injected Dragon KO mice at postnatal day 12 with anti-TNF-α antibody and anti-IL-1β antibody. WT and Dragon KO littermates injected with normal IgG were used as controls (Fig. 5). Neutralization of TNF-α and IL-1β increased GHR mRNA and protein expression (Fig. 5, A and B) but had no significant effect on SOCS3 expression in Dragon KO mice (Fig. 5C). Neutralization of TNF-α and IL-1β also increased IGF-I expression (Fig. 5D).

Collectively, all these results from LPS-treated and Dragon KO mice demonstrate that endogenous TNF-α and IL-1β inhibit hepatic GHR but do not play a major role in regulating SOCS3 expression during inflammation.

IL-6 stimulated hepatic SOCS3 but had no effect on GHR expression in mice.

To examine the role of endogenous IL-6 in regulation of GHR and SOCS3 expression in vivo, mice at 6–7 wk of age were injected with anti-IL-6 antibody or normal IgG 9.5 h after LPS injection, and the mice were euthanized 4 h after anti-IL-6 antibody injection (Fig. 6A). LPS reduced GHR mRNA and protein levels, and neutralization of IL-6 did not alter GHR mRNA and protein levels in LPS-treated mice (Fig. 6, B and C). LPS dramatically increased SOCS3 expression, and neutralization of IL-6 reduced SOCS3 expression in LPS-treated mice toward the levels in the control mice without LPS treatment (Fig. 6D). Of note, IGF-I was not altered by neutralization of IL-6 in LPS-treated mice (Fig. 6E).

Fig. 6.

Effects of neutralization of IL-6 on GHR, SOCS3, and IGF-I expression in livers of 6- to 7-wk-old mice during late phase of LPS treatment. Male mice at 6–7 wk of age were injected with LPS (2.5 μg/g body wt ip). A: anti-IL-6 antibody or normal IgG was injected (5 μg/g body wt ip) 9.5 h after LPS injection, and mice were euthanized 4 h after anti-IL-6 antibody injection. Relative hepatic GHR (B), SOCS3 (D), and IGF-I (E) mRNA levels were measured by real-time PCR. GHR protein levels in liver (C) were examined by Western blotting (top) and quantified by densitometry (bottom). GAPDH was the loading control for Western blotting; RPL19 was the internal control for real-time PCR analysis. Means without a common letter are significantly different: P < 0.05; n = 5 for B, D, and E, n = 3 for C.

We also administered anti-IL-6 antibody to 12-day-old Dragon KO mice (Fig. 7). WT and Dragon KO littermates injected with normal IgG were used as controls. Neutralization of IL-6 did not alter hepatic GHR mRNA expression in Dragon KO mice (Fig. 7A), but it reduced SOCS3 mRNA in Dragon KO mice to the same levels seen in WT mice (Fig. 7B). Again, IGF-I expression was not altered by neutralization of IL-6 in Dragon KO mice (Fig. 7C).

Fig. 7.

Effects of neutralization of IL-6 on GHR, SOCS3, and IGF-I expression in livers of Dragon KO mice at 12 days of age. Relative hepatic GHR (A), SOCS3 (B), and IGF-I (C) mRNA levels were measured by real-time PCR in WT mice administered normal IgG (WT/normal IgG) and in Dragon KO littermates administered normal IgG (KO/normal IgG) or anti-IL-6 antibody (KO/anti-IL-6). RPL19 was the internal control for real-time PCR analysis. Means without a common letter are significantly different: P < 0.05; n = 3–4.

Taken together, our results suggest that endogenous IL-6 plays a major role in mediating inflammation-induced SOCS3 expression but does not mediate the inhibition of GHR expression by inflammation.

Activity of the IL-6-SOCS3 pathway in inhibiting hepatic IGF-I expression was dependent on the active GH pathway.

It is unexpected that neutralization of IL-6 during 9.5–13.5 h after LPS treatment did not alter the suppression of IGF-I expression by LPS despite the dramatic reduction in SOCS3 expression (Fig. 6, D and E). In Dragon KO mice, neutralization of IL-6 completely restored SOCS3 expression but failed to recover IGF-I expression (Fig. 7, B and C). These observations suggest that IL-6 and its effector SOCS3 do not play a role in regulating IGF-I expression in these mouse models.

A previous study showed that deletion of IL-6 abolished the suppression of IGF-I by LPS, suggesting that LPS-induced inhibition of the GH-IGF-I pathway requires IL-6 (11). This finding appears to be contradictory to our findings here. However, in this previous study, the liver samples were collected 3 h after LPS treatment, when GHR mRNA and protein levels were reduced by only 30% (10). In mice at 13.5 h after LPS injection or in Dragon KO mice in the present study, GHR protein levels were already maximally inhibited by over 70% (Figs. 2, G and H, 3C, 4G, 5B, and 6C). It is possible that the ineffectiveness of the IL-6-SOCS3 axis in affecting IGF-I expression in our mouse models was attributable to the loss of GHR expression. We therefore examined whether the IL-6-SOCS3 pathway regulates hepatic IGF-I expression during the initial phase of LPS treatment when GHR protein was not so much inhibited by LPS. Mice at 6–7 wk of age were injected with either the anti-IL-6 antibody or normal IgG 1 h after LPS treatment. The mice were euthanized 4 h after the anti-IL-6 antibody injection (Fig. 8A). As expected, neutralization of IL-6 did not alter GHR mRNA and protein levels in LPS-treated mice (Fig. 8, B and C). Neutralization of IL-6 reduced SOCS3 mRNA expression induced by LPS (Fig. 8D). Interestingly, in association with the reduction in SOCS3 expression, IGF-I mRNA was now restored to the levels seen in mice without LPS treatment (Fig. 8E) by removal of IL-6 in LPS-treated mice. All these results suggest that IL-6 or its effector SOCS3 inhibits IGF-I expression when GHR expression is high, but they lose the ability to affect IGF-I expression when GHR is maximally inhibited in mice.

Fig. 8.

Effects of neutralization of IL-6 on GHR, SOCS3, and IGF-I expression in livers of 6- to 7-wk-old mice during early phase of LPS administration. Male mice at 6–7 wk of age were injected with LPS (2.5 μg/g body wt ip). A: anti-IL-6 antibody or normal IgG was injected (5 μg/g body wt ip) 1 h after LPS injection, and mice were euthanized 4 h after anti-IL-6 antibody injection. Relative hepatic GHR (B), SOCS3 (D), and IGF-I (E) mRNA levels were measured by real-time PCR. GHR protein levels in liver (C) were examined by Western blotting (top) and quantified by densitometry (bottom). GAPDH was the loading control for Western blotting; RPL19 was the internal control for real-time PCR analysis. Means without a common letter are significantly different: P < 0.05; n = 5 for B, D, and E; n = 3 for C.

SOCS3 lost its ability to regulate IGF-I when the GH pathway was switched off by inhibition of GHR expression in Huh-7 cells.

We then used Huh-7 cells to further determine the effect of GHR expression on the activity of SOCS3, the effector of IL-6, in inhibiting the GH-IGF-I pathway. Overexpression of SOCS3 blocked the induction of IGF-I expression by GH (Fig. 9A, compare bars 4 and 3) or by GH combined with WT Stat5b (Fig. 9B, compare bars 4 and 3) but failed to inhibit IGF-I expression induced by CA Stat5b (Fig. 9, C and D). These results verify that SOCS3 acts upstream of STAT5b. Interestingly, when GHR mRNA was reduced by 76% by siRNA targeting (Fig. 9E), GH treatment was no longer able to induce IGF-I expression, and overexpression of SOCS3 had no effects on IGF-I expression (Fig. 9F). These results indicate that SOCS3 does not regulate IGF-I expression when the GH-IGF-I pathway is switched off by the maximal inhibition of GHR expression.

DISCUSSION

Previous studies have demonstrated that exogenous IL-6, TNF-α, and IL-1β can induce GH resistance (1, 5, 6, 8–10, 13–15, 18, 19, 25, 27). However, the relative importance of endogenous IL-6, TNF-α, and IL-1β in meditating inflammation-induced GH resistance is not yet defined. Clarification of the roles of these individual cytokines will be a crucial step in designing therapies to rescue the GH-IGF-I pathway in inflammatory conditions.

Our results showed that TNF-α and IL-1β had minor effects on SOCS3 expression in Huh-7 cells. Consistently, in LPS-induced inflammation in the mice, neutralization of TNF-α and IL-1β did not appear to have a major effect on hepatic SOCS3 expression. Similar results were seen with the Dragon KO mouse model of chronic inflammation. The failure of neutralization of TNF-α and IL-1β in significantly modulating SOCS3 expression did not seem to be due to the ineffectiveness of the antibodies, because the antibodies significantly increased GHR and IGF-I expression in the two mouse models of inflammation. Interestingly, a previous study had revealed that TNF receptor (TNFR) deletion did not affect LPS-induced SOCS3 expression (10). All these observations suggest that TNF-α and IL-1β appear to be of limited importance in regulating SOCS3 expression during inflammation.

Previous studies had demonstrated that TNF-α and IL-1β inhibited expression of GHR in primary rat hepatocytes (12, 27). Furthermore, in TNFR1-deficient mice, the ability of LPS to inhibit hepatic GHR expression was diminished (11). In addition, administration of TNF-α binding protein ameliorated sepsis-induced decreased in hepatic GHR mRNA in rats (32), and TNF-α neutralization upregulated liver GHR abundance and restored GH signaling in colitic mice (30). Consistent with these previous observations, we found that TNF-α inhibited GHR expression in Huh-7 cells and that neutralization of TNF-α and IL-1β increased GHR expression in both LPS-treated mice and Dragon KO mice. All these results suggest that TNF-α plays an important role in inhibiting GHR expression in vivo.

In Huh-7 cells, IL-1β appears to have a higher efficiency than TNF-α in inhibiting GHR expression. However, the importance of IL-1β in inhibiting GHR in vivo cannot be addressed by the present study. To our knowledge, it has not been studied whether IL-1β inhibits hepatic GHR expression in vivo. Further studies will be needed to define the roles of IL-1β in GHR expression and GH resistance.

Coupled with the increased GHR expression following injection of TNF-α and IL-1β antibodies in LPS-treated mice and in Dragon KO mice, IGF-I expression was increased. Combined with previous studies showing that removal of TNF-α prevented the decreases in GHR expression and GH signaling by sepsis and colitis (30, 32), our results reinforce the notion that GHR mediates the inhibition of the GH-IGF-I pathway by TNF-α and IL-1β during inflammation.

In contrast to TNF-α and IL-1β, IL-6 did not have any effect on GHR expression in Huh-7 cells. Neutralization of IL-6 failed to alter GHR expression both in LPS-treated mice and in Dragon KO mice, although it drastically inhibited inflammation-induced SOCS3 levels. Therefore, our in vitro and in vivo studies clearly demonstrate that IL-6 does not regulate hepatic GHR expression. This observation is consistent with a previous study showing that IL-6 treatment did not inhibit GHR expression in CWSV-1 hepatocytes (25).

IL-6 increased SOCS3 expression in a dose dependent manner in Huh-7 cells. In chronic inflammation of Dragon KO mice or during late phase of LPS treatment, neutralization of IL-6 reduced SOCS3 expression toward the levels seen in the control mice without inflammation. These results suggest that, among all the inflammatory factors, endogenous IL-6 is the predominant inducer of SOCS3 expression in liver during inflammation.

SOCS3 is the major postreceptor inhibitor among the SOCS family members that mediate the inhibitory effects of inflammation on hepatic GH signaling. Unexpectedly, removal of SOCS3 inhibitory activity by neutralization of IL-6 in Dragon KO mice or in WT mice for 9.5–13.5 h after LPS treatment did not alter IGF-I expression. In line with the failure of the IL-6-SOCS3 pathway in regulating IGF-I expression, GHR expression was maximally inhibited by over 70% in these mouse models. In Huh-7 cells, SOCS3 lost its ability to inhibit IGF-I expression when the GH-IGF-I pathway was inactivated due to inhibition of GHR by over 70%. Therefore, our results suggest that the IL-6-SOCS3 axis may lose the inhibitory effect on the GH-IGF-I pathway in inflammatory diseases with maximal inhibition or loss of GHR expression.

The inability of IL-6 to inhibit IGF-I expression in liver cells with maximal inhibition of GHR is consistent with our observation that IL-6 stimulates SOCS3 expression but does not regulate GHR expression. Previous studies and our present study demonstrate that SOCS3 acts downstream of GHR. Therefore, the IL-6-SOCS3 axis is likely to have nothing in the GH-IGF-I pathway to act on when GHR is maximally inhibited or even lost.

During the initial phase of LPS action, when GHR expression was not so much altered yet, neutralization of IL-6 inhibited LPS-induced SOCS3 expression and completely restored IGF-I expression. These results are consistent with a previous study (10) demonstrating that deletion of IL-6 blocked the LPS activity in inducing SOCS3 expression and in inhibiting GH-induced Stat5b phosphorylation in livers 3 h after LPS injection, when GHR proteins were only slightly inhibited. Therefore, during the initial phase of LPS-induced inflammation and possibly other inflammatory diseases, the IL-6-SOCS3 axis is critically involved in the development of GH resistance. However, in some systems, other factors (e.g., inducible GHR proteolytic reduction in GHR abundance) may also affect this early LPS-induced GH resistance (29).

Taken together, the present study demonstrates that endogenous TNF-α and IL-1β inhibit GHR expression, while endogenous IL-6 stimulates SOCS3 in liver during inflammation. On the basis of our results in combination with previous observations, we propose our working hypothesis (Fig. 10): TNF-α/IL-1β and IL-6 function at different levels of the GH-IGF-I pathway to induce hepatic GH resistance, with TNF-α and IL-1β primarily inhibiting GHR expression and IL-6 mainly stimulating SOCS3. The differential actions of IL-6 and TNF-α/IL-1β on GHR expression may lead to different roles of IL-6 and TNF-α/IL-1β in mediating inflammation-induced GH resistance. TNF-α and IL-1β exert a tonic inhibition on hepatic GH signaling by their constant inhibition of GHR expression. IL-6 inhibits the GH-IGF-I pathway when the GH pathway is not yet affected or only partially inhibited by inflammation. In inflammatory diseases with maximal inhibition or loss of GHR expression/activity and subsequent shutdown of the GH pathway, IL-6 and its effector SOCS3 lose their regulation of IGF-I expression. In the inflammatory diseases with maximal inhibition or loss of GHR, removal of TNF-α and IL-1β increases GHR expression, thus restoring IGF-I expression, while removal of IL-6 may not restore IGF-I expression although it increases SOCS3 expression.

Fig. 10.

Working hypothesis on the distinct mechanisms by which TNF-α/IL-1β and IL-6 inhibit the hepatic GH-IGF-I pathway during inflammation. TNF-α and IL-1β inhibit GH signaling via inhibition of GHR expression, while IL-6 inhibits GH signaling via stimulation of SOCS3 (left). When GHR expression is maximally inhibited by TNF-α and IL-1β and perhaps some other factors, the IL-6-SOCS3 pathway no longer plays a role in regulating the hepatic GH-IGF-I pathway (right).

GRANTS

Y. Xia was supported in part by a startup fund offered by The Chinese University of Hong Kong, an RGC/GRF grant from the Hong Kong Research Grant Council (CUHK477311), and Shenzhen Science and Technology Research and Development Funding (JC201105201069A). S. J. Frank was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R01 DK-58259 and R01 DK-46395 and a Veterans Affairs Merit Review award.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: Y.Z., X.X., H.Y.L., and Y.X. conception and design of research; Y.Z. and Y.X. performed experiments; Y.Z. and Y.X. analyzed data; Y.Z., S.J.F., and Y.X. interpreted results of experiments; Y.Z. and Y.X. prepared figures; Y.Z., X.X., H.Y.L., and Y.X. drafted manuscript; S.J.F. and Y.X. edited and revised manuscript.