AMPK Serves as a Therapeutic Target Against Anemia of Inflammation

Abstract

Aims:

Anemia of inflammation (AI), the second prevalent anemia, is associated with worse prognosis and increased mortality in numerous chronic diseases. We recently reported that the gasotransmitter hydrogen sulfide (H₂S) suppressed the inflammatory activation of signal transducer and activator of transcription 3 (STAT3) and hepcidin, the critical mediators of AI. Adenosine 5′-monophosphate-activated protein kinase (AMPK) is a novel inflammatory regulator and might be activated by H₂S. In this study, we determined whether AMPK played a role in H₂S-mediated anti-inflammatory response in AI and evaluated the therapeutic potential of AMPK against AI by pharmacological and clinical approaches.

Results:

We showed that AMPK mediated the inhibition of STAT3, hepcidin, and AI by H₂S during inflammation. Moreover, pharmacological and genetic activation of AMPK ameliorated hepcidin production, corrected iron dysregulation, and relieved hypoferremia and anemia in both acute and chronic inflammation models in mice. Mechanistic studies indicated that AMPK suppressed STAT3/hepcidin activation by promoting proteasome-mediated Janus kinase 2 (JAK2) degradation, which was dependent on the intact function of suppressor of cytokine signaling 1 (SOCS1) and increased interactions between SOCS1 and JAK2. Most importantly, the AMPK activator metformin was associated with decreased serum hepcidin content and anemia morbidity in Chinese type 2 diabetes mellitus patients.

Innovation:

This is the first study to demonstrate the inhibition of inflammatory hepcidin and AI by AMPK-induced JAK2 degradation. Our work uncovered AMPK as a novel therapeutic target, and metformin as a potential therapy against AI.

Conclusion:

The present work demonstrated that AMPK mediated the therapeutic effects of H₂S and relieved AI by promoting SOCS1-mediated JAK2 degradation. Antioxid. Redox Signal. 27, 251–268.

Introduction

Anemia of inflammation (AI) is the second most prevalent anemia in clinical conditions. Due to its immune-driven nature, AI usually occurs in patients suffering infections, cancer, and autoimmune diseases, with clinical prevalence ranging from 10% to 95% (71). Although AI is relatively mild, studies showed that lower hemoglobin, which was associated with worse outcomes, was an independent predictor for mortality in various diseases such as heart failure (21, 51), cancer (8), and chronic kidney disease (16, 70).

Innovation

Conventional remedies for anemia of inflammation (AI) have raised concerns regarding safety and effectiveness, calling for alternative therapies. Using both in vitro and in vivo models, we demonstrated that AMPK relieved iron dysregulation and anemic symptoms in AI mice by promoting Janus kinase 2 (JAK2) degradation and, subsequently, inhibiting hepatic signal transducer and activator of transcription 3/hepcidin activation. Moreover, metformin, a typical AMPK activator, was associated with lower serum hepcidin and reduced anemia morbidity in type 2 diabetes mellitus patients. These findings reveal AMPK as a novel therapeutic target, and metformin as a potential drug against AI.

AI is characterized by disturbance of iron homeostasis and dysfunction of erythroid system (47), in which iron dysregulation is one of the hallmarks. Suppressed dietary absorption and increased retention of iron within the reticuloendothelial system limit iron availability for erythroid progenitor cells, leading to restricted erythropoiesis (18). Therefore, restoration of the iron balance is proposed as an effective approach in AI treatment.

Hepcidin, a liver-derived iron-regulating peptide, is a well-established critical player in the development of AI (17). During inflammation, hepcidin is induced by interleukin-6 (IL-6)-mediated activation of the Janus kinase (JAK)/signal transducer and activator of transcription 3 (STAT3) signaling pathway. Hepcidin reduces intestinal iron absorption and promotes the diversion of circulating iron to the storage sites by promoting the degradation of ferroportin, the main cellular iron exporter (54, 75). Increased urinary hepcidin excretion is observed in AI patients (49), whereas inhibition of hepcidin by Il-6 or Stat3 knockout in mice protects against AI (35, 60). These findings suggest that inhibition of hepcidin in inflammation may be an effective strategy to resolve iron disorders and AI.

Adenosine 5′-monophosphate (AMP)-activated protein kinase (AMPK) is a conserved metabolic sensor that monitors intracellular AMP/adenosine triphosphate (ATP) ratio (27, 39). Due to its vital role in energy regulation, AMPK was extensively studied in metabolic disorders. Importantly, metformin, a potent AMPK activator, is widely prescribed as first-line therapy in type 2 diabetes mellitus (T2DM) (20). In addition, recent studies reported that AMPK might also act as a novel suppressor of inflammation (52). Notwithstanding these findings, there remains a knowledge gap on AMPK/metformin, hepcidin, and AI.

Hydrogen sulfide (H₂S), the third bio-active gasotransmitter after nitric oxide and carbon monoxide, is produced endogenously and plays a prominent role in inflammatory responses (37). We recently demonstrated that H₂S inhibited hepcidin expression and hypoferremia in an acute lipopolysaccharide (LPS) model (76), as previous studies showed that H₂S activated the AMPK signaling pathway in the central nervous system (45). These findings prompted us to assess the role of AMPK in H₂S-mediated inhibition of hepcidin and AI by using pharmacological and clinical approaches with the ultimate goal of evaluating the therapeutic potential of AMPK activation in AI.

In the present study, we found that AMPK contributed to the inhibition of hepcidin and alleviation of AI by H₂S by promoting suppressor of cytokine signaling 1 (SOCS1)-mediated JAK2 degradation. Moreover, in in vivo mouse models and by retrospective analysis of serum samples from T2DM patients, we identified AMPK as a novel therapeutic target and metformin as a potential remedy against AI. The graphic abstract is presented in Figure 1.

FIG. 1.

The graphic abstract of the present study. By performing pharmacological experiments and clinical analysis, the present work demonstrated that AMPK inhibited inflammatory hepcidin and relieved AI by promoting SOCS1-mediated JAK2 degradation. AI, anemia of inflammation; AMPK, adenosine 5′-monophosphate-activated protein kinase; SOCS1, suppressor of cytokine signaling 1. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Results

H₂S-mediated inhibition of STAT3 and hepcidin is dependent on AMPK activation

Based on our recent finding that H₂S suppressed the inflammatory activation of STAT3 and hepcidin, we sought to identify H₂S-mediated intracellular signal transduction pathways in primary mouse hepatocytes pretreated with sodium hydrosulfide (NaHS), an exogenous H₂S donor, followed by IL-6 treatment using an antibody array. As illustrated in Figure 2A, H₂S led to a significant increase in AMPK phosphorylation while inhibiting the activation of many transcriptional factors controlled by IL-6 included in the array.

FIG. 2.

AMPK mediates the inhibition of STAT3 and hepcidin by H₂S. (A) Antibody array indicated that H₂S activated AMPK. Primary mouse hepatocytes were pretreated with NaHS (50/100 μM) 1 h before murine IL-6 (25 ng/ml) challenge. Cells were collected for antibody array analysis 3 h later (n = 3). (B, C) Western blots confirmed the AMPK activation by H₂S. As in (A), HepG2 cells and primary mouse hepatocytes were incubated with NaHS (50–300 μM) followed by 10 ng/ml human IL-6 and 25 ng/ml murine IL-6, respectively. (D, E) Overexpression of endogenous H₂S synthases, CSE and CBS, activated AMPK as well. HepG2 cells were transfected with CSE construct or infected with CBS-overexpressing lentivirus. Cell lysates were analyzed 48 h later. (F) As in (B), NaHS (100 μM) and AICAR (2 mM) increased AMPK kinase activity that was impaired by IL-6 (10 ng/ml) in HepG2 cells (n = 3). AICAR, an AMPK activator, was used as a positive control. AMPK silencing by two specific siRNA (100 nM each) successfully abolished the suppression of STAT3 (H) and hepcidin promoter activity (G) by H₂S in Huh7 cells (n = 4). DN-AMPK reversed the effects of H₂S on hepcidin in primary mouse hepatocytes (I) and STAT3 in HepG2 cells (J) (n = 3). Representative immunoblots and images are presented. Molecular weight markers in kDa are at the right of the blots. Data are presented as the mean ± SEM. Statistical analysis was conducted with one-way ANOVA followed by Turkey's test. ^## p < 0.01 compared with the control group; **p < 0.01, ***p < 0.001 compared with the IL-6 group unless indicated; ^& p < 0.05, ^&& p < 0.01 compared with vector IL-6 + NaHS group unless indicated. See also Supplementary Figure S1 for densitometry analysis of immunoblots. Please refer to Supplementary Figure S10 for raw Western blot records shown in figure. AICAR, 5-aminoimidazole-4-carboxamide ribonucleotide; CBS, cystathionine β-synthase; CSE, cystathionine γ-lyase; DN-AMPK, dominant negative AMPK; H₂S, hydrogen sulfide; NaHS, sodium hydrosulfide; STAT3, signal transducer and activator of transcription 3. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

To investigate the role of AMPK in our system, we pretreated HepG2 hepatoma cells (Fig. 2B and Supplementary Fig. S1A; Supplementary Data are available online at www.liebertpub.com/ars) and mouse primary hepatocytes (Fig. 2C and Supplementary Fig. S1B) with 50–300 μM NaHS before IL-6 challenge. The NaHS doses range was based on previous studies (32, 64). AMPK was activated by NaHS in a dose-dependent manner, as supported by induced phosphorylation of acetyl-CoA carboxylase (ACC), a representative AMPK substrate. Results consistent with these findings were observed in HepG2 cells overexpressing cystathionine γ-lyase (CSE) or cystathionine β-synthase (CBS), two major endogenous H₂S synthases (Fig. 2D, E and Supplementary Fig. S1C, D). The regulation of AMPK by H₂S was further confirmed by increased AMPK activity determined by a kinase assay (Fig. 2F).

We next assessed the involvement of AMPK by gene silencing and pharmacological inhibition by using Huh7 cells that exhibited a higher response than HepG2 cells in the luciferase activity assay. As shown in Figure 2G, H and Supplementary Figure S1E, knockdown of AMPK by two specific siRNAs successfully abolished the suppression of STAT3 and hepcidin by H₂S. As hepcidin was previously shown to be suppressed by compound C, the classic AMPK inhibitor, by suppressing the bone morphogenetic protein pathway in an AMPK-independent manner (53), instead we introduced a lentiviral vector expressing dominant negative AMPK (DN-AMPK) and demonstrated the abrogation of H₂S-mediated STAT3 and hepcidin inhibition (Fig. 2I, J and Supplementary Fig. S1F). These data suggested that AMPK contributed to the inhibition of inflammatory STAT3 and hepcidin by H₂S.

AMPK mediated the treatment effects of H₂S on AI

Next, we examined the effects of H₂S on AI and the interplay between AMPK and H₂S. Constitutively active AMPK (CA-AMPK), DN-AMPK, or empty lentiviral vectors were i.v. injected to C57BL/6 mice, at 2 × 10⁷ lentiviral particles/mouse. One week later, mice were treated with weekly subcutaneous injections of turpentine (100 μl/20 g weight) for 4 weeks to induce a chronic non-infectious model of AI. Starting in the second week of turpentine treatment, NaHS (6 mg/kg) was i.p. administered twice a week, according to a previous report (38). Mice were sacrificed after anesthesia 1 week after the final injection of turpentine. The workflow of the chronic AI model is presented in Figure 3A.

FIG. 3.

H₂S corrects turpentine-induced AI through AMPK. (A) A diagram about the induction of AI by turpentine and the treatment of NaHS. (B, C) Both serum iron and transferrin (Tf) saturation were increased by CA-AMPK and NaHS (n = 5). (D) CA-AMPK improved erythrocyte number (RBC) and HGB levels, whereas DN-AMPK abolished the effects of NaHS (n = 5). (E) Consistent results were observed with representative images of Wright-Giemsa staining. Solid arrows indicate damaged erythrocytes. Tur, turpentine. Scale bars: 10 μm. (F, G) CA-AMPK and NaHS inhibited serum hepcidin levels and hepatic JAK2/STAT3 activation, whereas the effects of NaHS were diminished by DN-AMPK (n = 5). Representative immunoblots and images are presented. Molecular weight markers in kDa are at the right of the blots. Data are presented as the mean ± SEM. Statistical analysis was performed with one-way ANOVA followed by Turkey's test. ^## p < 0.01, ^### p < 0.001 compared with the control group; *p < 0.05, **p < 0.01, compared with the vector (Vec) model group; ^& p < 0.05 compared with the vector (Vec) NaHS group. See also Supplementary Table S1 for hematological and mouse indices. Please refer to Supplementary Figure S11 for raw Western blot records shown in figure. HGB, hemoglobin; RBC, red blood cell; TIBC, total iron binding capacity. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Serum iron levels, as well as transferrin (Tf) saturation, were restored in mice treated by NaHS and those infected with CA-AMPK, though no significant change was observed with DN-AMPK (Fig. 3B, C). To better evaluate the parameters of circulating erythrocytes, a complete blood count was conducted. As demonstrated in Supplementary Table S1, the chronic AI model using turpentine shared the normocytic normochromic nature observed in clinical AI (34, 56). Both NaHS treatment and CA-AMPK infection successfully improved erythrocyte numbers and hemoglobin content (Fig. 3D). Moreover, the effects of NaHS were reduced in DN-AMPK-infected mice. These results were confirmed by blood smear analysis with Wright-Giemsa staining, which demonstrated reduced erythrocyte membrane irregularity in NaHS-treated and CA-AMPK-infected mice (Fig. 3E). Immunoblots and enzyme-linked immunosorbent assay (ELISA) showed that CA-AMPK infection and NaHS treatment markedly ameliorated JAK2/STAT3 signaling and downstream hepcidin production, as opposed to DN-AMPK (Fig. 3F, G). Intriguingly, total protein levels of JAK2, the critical signal transducer within the IL-6/STAT3 pathway, were also decreased. These data not only demonstrated that H₂S relieved AI in an AMPK-dependent manner but also suggested that direct AMPK activation might elicit similar therapeutic effects.

AMPK activation ameliorates IL-6-induced hepcidin levels both in vitro and in vivo

We next examined the effects of pharmacological and genetic activation of AMPK on inflammatory hepcidin. As a potent AMPK activator, 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR, 0.5/2 mM) (73) dramatically suppressed hepcidin promoter activity and the transcriptional function of STAT3 as assessed by a dual luciferase reporter assay, in parallel with reduced phosphorylation and nuclear translocation of STAT3 (Fig. 4A–C). Similar results were obtained with the CA-AMPK construct (Fig. 4D and Supplementary Fig. S2A).

FIG. 4.

AMPK activation reduces IL-6-induced hepcidin levels both in vitro and in vivo. (A) AICAR (0.5/2 mM), a potent AMPK activator, inhibited IL-6 (10 ng/ml)-induced hepcidin (HAMP) promoter activity and STAT3 transcriptional function assessed by an SIE reporter in Huh7 cells (n = 3). (B, C) As in (A), AICAR (AIC) activated AMPK and suppressed STAT3 phosphorylation and nuclear translocation in primary mouse hepatocytes treated with 25 ng/ml IL-6. Scale bars: 20 μm. (D) CA-AMPK construct showed similar suppression on hepcidin mRNA levels in HepG2 cells (n = 3). AICAR (AIC, i.p., 250/500 mg/kg) and metformin (Met, i.g., 250 mg/kg) decreased serum hepcidin (E), splenic iron accumulation (G), and improved serum iron levels (F) in C57BL/6 mice injected with IL-6 (i.p., 3 μg/20 g weight, n = 5). (H) As in (E–G), representative immunoblots demonstrated ameliorated JAK2/STAT3 activation and reduced JAK2 protein levels. Separated blots of the same protein indicated two different membranes that were visualized at the same time. Representative immunoblots and images are presented. Molecular weight markers in kDa are at the right of the blots. Data are presented as the mean ± SEM. Statistical analysis was conducted with one-way ANOVA followed by Turkey's test. ^# p < 0.05, ^## p < 0.01, and ^### p < 0.001, all compared with the control group; *p < 0.05, **p < 0.01, and ***p < 0.001, compared with the IL-6/model group unless indicated. See also Supplementary Figure S2 for the effects of CA-AMPK on STAT3, and densitometry analysis of (H). Please refer to Supplementary Figure S12 for raw Western blot records shown in Figure 4. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Our hypothesis was then tested in vivo. C57BL/6 mice were pretreated with one of the two AMPK activators, AICAR (i.p., 250 & 500 mg/kg) (80) and metformin (i.g., 250 mg/kg) (22), 1 h before a 3-h challenge with IL-6 (i.p., 3 μg/20 g weight). The dose of metformin used here was converted from a human dose of 1500 mg/75 kg by following a translation formula previously described (57). The mice were sacrificed as stated earlier (54). As illustrated in Figure 4E and Supplementary Figure S2B, both AICAR and metformin suppressed hepatic hepcidin mRNA (Hamp) and serum hepcidin levels evoked by IL-6. As expected, serum iron levels were restored and splenic iron content was reduced in AICAR- and metformin-treated mice (Fig. 4F, G). Immunoblot assays indicated that both AICAR and metformin decreased hepatic phospho-STAT3 (p-STAT3) levels induced by IL-6 (Fig. 4H and Supplementary Fig. S2C–E). Taken together, these results suggested that AMPK activation ameliorated inflammatory STAT3 and hepcidin activation both in vitro and in vivo.

AMPK activators improve turpentine-induced AI in vivo

To support our hypothesis that AMPK was a therapeutic target against AI, the effects of AICAR (i.p., 500 mg/kg) and metformin (i.g., 250 mg/kg) were evaluated in turpentine-treated C57BL/6 mice as described in Figure 5A.

FIG. 5.

AMPK activators improve turpentine-induced AI in vivo. (A) A diagram about the induction of AI by turpentine and the compound treatment. (B, C) Serum iron levels and transferrin (Tf) saturation were restored by metformin and AICAR (n = 6). (D) Complete blood count suggested increased erythrocyte number, HGB levels, and HCT by AICAR and metformin (n = 6). Data are presented as the mean ± SD. (E) Representative images of blood smear with Wright-Giemsa staining showed improved erythrocyte morphology by metformin and AICAR. Solid arrows indicate damaged erythrocytes. Tur, turpentine. Scale bars: 10 μm. (F) Representative images indicated relieved splenomegaly. (G, H) Splenic iron accumulation was ameliorated by metformin and AICAR, as assessed by Perl's Prussian blue staining and non-heme iron analysis (n = 6). Scale bars: 100 μm. (I, J) Serum hepcidin levels were reduced, and hepatic JAK2/STAT3 was suppressed by two AMPK activators (n = 6). Representative immunoblots and images are presented. Molecular weight markers in kDa are at the right of the blots. Data are presented as the mean ± SEM, except (D). Statistical analysis was conducted with one-way ANOVA followed by Turkey's test. ^## p < 0.01, ^### p < 0.001 compared with the control group; *p < 0.05, **p < 0.01, and ***p < 0.001, compared with the model group. See also Supplementary Figure S3 for serum IL-6 and densitometry analysis of (J). Please refer to Supplementary Figure S13 for raw Western blot records shown in figure. HCT, hematocrit; MCH, mean corpuscular hemoglobin; MCV, mean corpuscular volume; PLT, platelet; RDW, red blood cell distribution width. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Both AMPK activators successfully increased serum iron levels and transferrin saturation that were compromised in turpentine-treated mice (Fig. 5B, C). Consistently, hematological indices and erythrocyte morphology were also improved (Fig. 5D, E). In addition, inflammatory splenomegaly induced by turpentine was relieved by AMPK activation (Fig. 5F). In agreement with Perl's Prussian blue staining, splenic iron accumulation was ameliorated as evidenced by non-heme iron analysis (Fig. 5G, H). Similar to that observed in the acute IL-6 model, serum hepcidin was reduced and hepatic JAK2/STAT3 was suppressed (Fig. 5I, J and Supplementary Fig. S3). Collectively, these results indicated that both AICAR and metformin might have therapeutic potential against AI.

H₂S and AMPK suppress inflammatory STAT3 and hepcidin by promoting proteasome-mediated JAK2 degradation

Janus family of tyrosine kinases is activated on ligand binding to associated cytokine receptors, and it modulates the phosphorylation of related signaling molecules (36). Among these kinases, JAK1 and JAK2 play key roles in the IL-6/STAT3 pathway (11, 46). Based on our observation of JAK2 regulation by AMPK in Figures 3G, 4H, and 5J, we sought to examine whether JAK2 contributed to the inhibition of STAT3 and hepcidin.

As demonstrated in Figure 6A and Supplementary Figure S4A, JAK1 and JAK2 phosphorylation levels were markedly suppressed by AICAR treatment. However, contrary to JAK2, there was minimal phosphorylation of JAK1 in response to IL-6 challenge, suggesting that JAK2 was the predominant JAK involved in signal transduction in this experimental paradigm. In addition, as was observed in vivo, AICAR decreased the protein levels of JAK2 but not JAK1 in mouse primary hepatocytes treated by IL-6. Moreover, lactate dehydrogenase (LDH) assay ruled out the potential possibility of contribution of AICAR-induced cytotoxicity to decreased JAK2 protein levels (Supplementary Fig. S5A). The involvement of JAK2 in STAT3 and hepcidin activation was further confirmed with specific JAK2 inhibitors (AG490, WP1066) and JAK2 overexpression (Fig. 6B and Supplementary Fig. S5B–D). These findings were further corroborated in experiments using CA-AMPK construct (Fig. 6C and Supplementary Fig. S4B) and in the absence of IL-6 (Supplementary Fig. S5E). In agreement with AMPK, NaHS exerted a similar effect on JAK2 (Fig. 6D and Supplementary Fig. S4C). These data showed that H₂S and AMPK inhibited STAT3/hepcidin by suppressing JAK2.

FIG. 6.

H₂S and AMPK decrease inflammatory STAT3 and hepcidin by promoting proteasome-mediated JAK2 degradation. (A) JAK1 and JAK2 phosphorylation, along with JAK2 protein levels were reduced by AICAR (0.5/2 mM) treatment. (B) Inhibition of JAK2 by AG490 (25 μM) and WP1066 (10 μM) suppressed hepcidin mRNA expression (n = 3). (C, D) CA-AMPK and NaHS inhibited JAK2 phosphorylation and protein levels. (E) Jak2 mRNA levels were not significantly changed by AICAR, in both the presence and absence of IL-6 (n = 3). (F) Cycloheximide chase assay demonstrated that JAK2 degradation was promoted in mouse primary hepatocytes that were treated with AICAR (n = 3). (G, H) MG132 (25 μM), a potent proteasome inhibitor, reversed the suppression of JAK2/STAT3 activation by NaHS and AICAR, in parallel with increased JAK2 levels. (I) AICAR and NaHS induced JAK2 ubiquitylation in primary mouse hepatocytes. Ub, ubiquitin. (J) The inhibition of inflammatory hepcidin by AICAR and NaHS was reversed by 25 μM MG132 (n = 3). Representative immunoblots and images are presented. Molecular weight markers in kDa are at the right of the blots. Data are presented as the mean ± SEM. Statistical analysis was conducted with one-way ANOVA followed by Turkey's test. ^## p < 0.01 compared with the control group; *p < 0.05, **p < 0.01, compared with the model group; ^& p < 0.05 compared with the NaHS or AICAR group without Mg132. See also Supplementary Figures S4 and S5 for more data on AMPK-induced JAK2 degradation and densitometry analysis of immunoblots. Please refer to Supplementary Figure S14 for raw Western blot records shown in figure.

To further explore the regulation of JAK2 by AMPK, we focused on the mechanisms within the decrease of JAK2 protein levels. As shown in Figure 6E, Jak2 mRNA levels were not significantly decreased by AICAR, independent of IL-6. Considering the rapid changes observed in JAK2 protein following AICAR application (1–3 h), these results indicated that the reduction of JAK2 was probably achieved at the posttranscriptional level. Indeed, cycloheximide (CHX) chase assay revealed that JAK2 degradation was promoted in mouse primary hepatocytes treated with AICAR, with a shift of its half-life from more than 6 h to about 3 h (Fig. 6F). The inhibition of STAT3 with JAK2 abrogation was confirmed in a time-course experiment of mouse primary hepatocytes that were pretreated with AICAR for 3 h followed by IL-6 treatment (Supplementary Fig. S5F).

These data prompted us to ask which pathway of protein degradation is involved. By applying proteasome inhibitor MG132, we found that the effects of AICAR, NaHS, and CA-AMPK were diminished, as opposed to ectopically expressed ubiquitin (Fig. 6G, H, Supplementary Figs. S4D, E and S5G, H). These observations were supported by in vivo JAK2 ubiquitylation assay and hepcidin levels in mouse primary hepatocytes (Fig. 6I, J). Taken together, our data indicated that both AMPK activation and H₂S reduced inflammatory hepcidin levels by promoting proteasome-mediated JAK2 degradation.

SOCS1 mediates the JAK2 degradation induced by AMPK activation

As a member of the SOCS family, SOCS1 plays a prominent role in the negative regulation of cytokine signal transduction (30). With the help of a conserved carboxy-terminal motif, SOCS box, SOCS1 interacts with elongin B/C and acts as an E3 ligase in the degradation of JAK2 (78). As demonstrated in Figure 7A and B, silencing of Socs1 reduced JAK2 ubiquitylation levels and prolonged JAK2 half-life in AICAR-treated mouse primary hepatocytes. As expected, suppression of the JAK2/STAT3 pathway was also reversed with Socs1 silencing (Fig. 7C and Supplementary Fig. S6A). Co-immunoprecipitation assay indicated that AICAR and NaHS significantly promoted JAK2-SOCS1 interactions (Fig. 7D, E and Supplementary Fig. S6B, C). These observations were further supported by transfection of constructs bearing wild-type SOCS1 and a truncated fragment with SOCS box deleted (SOCS1ΔBox, Fig. 7F, G and Supplementary Figs. S6D–F and S7).

FIG. 7.

SOCS1 mediates the JAK2 degradation and hepcidin reduction induced by AMPK. (A) JAK2 ubiquitylation induced by AICAR was diminished by Socs1 siRNA (100 nM). (B) Socs1 knockdown prolonged JAK2 half-life under AICAR (1 mM) treatment (n = 3). (C) Silencing of Socs1 abolished the inhibition of JAK2/STAT3 by AICAR. (D, E) SOCS1-JAK2 interactions were enhanced by AICAR and NaHS treatment in the presence of MG132 (25 μM). (F, G) Full-length SOCS1, but not SOCS1ΔBox dead mutation, suppressed JAK2/STAT3 and hepcidin promoter activity induced by IL-6. The effects of AICAR and NaHS were reversed by SOCS1ΔBox construct (n = 3). (H, I) Consistent results were observed in vivo with Socs1 siRNA (n = 5). Representative immunoblots and images are presented. Molecular weight markers in kDa are at the right of the blots. Data are presented as the mean ± SEM. Statistical analysis was conducted with one-way ANOVA followed by Turkey's test. ^### p < 0.001 compared with the control group; *p < 0.05, **p < 0.01, and ***p < 0.001, all compared with the model group unless indicated; ^& p < 0.05 compared with the siNC AICAR (AIC) group. See also Supplementary Figures S6 and S7 for data on JAK2 degradation promoted by SOCS1 and densitometry analysis of immunoblots. Please refer to Supplementary Figure S15 for raw Western blot records shown in the figure.

We next determined the potential in vivo role of SOCS1 by hydrodynamic transfection of Socs1 siRNA in C57BL/6 mice. Knockdown of Socs1 abrogated the AMPK-mediated inhibition of JAK2/STAT3 and hepatic hepcidin (Fig. 7H, I and Supplementary Fig. S6G). In conclusion, our data suggested that SOCS1 mediated the JAK2 degradation and hepcidin inhibition induced by AMPK activation.

Metformin treatment is associated with lower serum hepcidin levels and reduced anemia morbidity in patients with T2DM

Metformin is a first-line medication for T2DM, which is closely associated with chronic inflammation (72), as evidenced by elevated IL-6 content, and increased anemia risk (67). To assess whether metformin elicits similar clinical effects on anemia and hepcidin, we retrospectively analyzed the hematological indices of 137 T2DM patients, from whom 83 serum samples were collected. The remaining patients were classified by gender and metformin therapy.

No significant difference was observed in baseline characteristics of the entire cohort (Fig. 8A). In addition, baseline characteristics except for the mean age of female patients were similar among patients with serum samples (Supplementary Fig. S8A).

FIG. 8.

Metformin is associated with reduced serum hepcidin and anemia morbidity in T2DM patients. (A) Baseline characteristics of all enrolled patients. T-test was performed unless indicated. ^#Mann–Whitney U test; ^##Fisher's exact probability test. Details of combined antidiabetic medications can be found in Supplementary Figure S8. (B) Serum hepcidin levels were markedly lower in male patients taking metformin, compared with those not taking it. Higher basal serum hepcidin content was observed in male patients than in female ones. One-way ANOVA followed by Turkey's test was used for statistical analysis. *p < 0.05. (C, D) Metformin application was associated with increased RBC number and HGB levels in male T2DM patients. Please refer to Supplementary Figure S8 for data of female patients. (E) Serum H₂S content tended to increase in male patients taking metformin. (F) Anemia prevalence was significantly lower in the metformin group for all patients and male ones. A borderline significant association was seen between anemia and high hepcidin levels in male patients. Statistical analysis was performed with chi-square test. *OR and 95% CI could not be calculated, since no patient had anemia in the metformin and hepcidin group. (G) Scheme figure illustrating the inhibition of hepcidin and AI by AMPK through promoting SOCS1-mediated JAK2 degradation. Data are presented as the mean ± SD or mean (%). T-test was performed for statistical analysis unless indicated. CI, confidence interval; OR, odds ratio; T2DM, type 2 diabetes mellitus. To see this illustration in color, the reader is referred to the web version of this article at www.liebertpub.com/ars

Importantly, as demonstrated in Figure 8B, metformin therapy was associated with decreased serum hepcidin in male patients. Accordingly, erythrocyte number and hemoglobin content were significantly higher in male patients taking metformin (Fig. 8C, D). An increasing trend was also observed in hematocrit of the male patients on metformin, and in erythrocyte number and hemoglobin content of the female patients taking metformin (Supplementary Fig. S8B–E). Moreover, an insignificant increase in the serum H₂S content was observed in male patients taking metformin (Fig. 8E).

We next examined the association of anemia with metformin therapy and serum hepcidin. As shown in Figure 8F, among all patients, anemia was significantly lower in the metformin group than in the non-metformin group (odds ratio = 0.081, 95% confidence interval = 0.010–0.639; p = 0.004). Similarly, the rate of anemia was significantly lower in male patients treated with metformin (p = 0.010) and it was borderline significantly higher in male patients with high serum hepcidin levels (p = 0.050). The odds ratio and 95% confidence interval could not be calculated, as no male patient had anemia in the metformin and hepcidin groups. Overall, our results indicated that metformin was, at least for men, associated with reduced serum hepcidin content and anemia prevalence in Chinese T2DM patients.

Discussion

In the present study, we demonstrated that AMPK mediated the treatment effects of H₂S on inflammatory hepcidin and AI. In addition, we elucidated the therapeutic potential of AMPK in the resolution of AI and the restoration of iron balance, and we demonstrated that AMPK induced JAK2 degradation via the ubiquitin-proteasome pathway (Fig. 8G). These results provided new insights into the anti-inflammatory roles of AMPK activators, including metformin. Notably, evaluation of clinical samples showed that metformin treatment was associated with decreased serum hepcidin levels and anemia in T2DM patients. On the whole, our work uncovered AMPK as a novel therapeutic target, and metformin as a potential drug against AI.

Conventional remedies for AI include blood transfusion, iron supplementation, and erythropoietic stimulating agents (ESAs). However, blood transfusion is associated with a high risk of multi-organ failure and immunomodulation, such as allergic reaction, infection transmission, and immunosuppression (69). Conversely, in addition to well-known issues such as limited oral absorption and potential cardiovascular events (10), iron therapy was shown to be associated with increased infection risk in hemodialysis patients (65). Although the beneficial effects of ESAs are well documented, a meta-analysis examining 13,933 patients with cancer in 53 trials demonstrating an association of ESA use with increased mortality and worse overall survival raises concerns about their use (3). These studies highlight the urgent need for alternative therapies and illustrate the significance and implications of our findings in clinical management of AI.

AI is a complex disease arising from a confluence of inflammatory responses by multiple organs; therefore, a variety of infectious and non-infectious mouse models have been established to date. Compared with the currently used infectious models, non/pseudo-infectious AI models achieve improved reproducibility and are safer for both experimenters and other experimental animals. LPS and recombinant IL-6 are commonly used to induce inflammatory hepcidin and hypoferremia not only in animal models but also in human trials (29, 48); however, there still remains a gap between acute inflammation insult and AI. Turpentine has been extensively studied in inflammatory animal models for decades. A number of studies previously used turpentine to induce normocytic normochromic AI (34, 50). We previously utilized LPS to induce hepcidin levels. In the present study, both IL-6 and turpentine were used to investigate the therapeutic potential of AMPK by several methods. Our results demonstrated that AMPK not only reversed acute inflammatory hepcidin and hypoferremia but also improved AI symptoms in a chronic model.

A large number of studies explored the role of H₂S in diseases of the cardiovascular, central nervous systems as well as in inflammation (38). We herein reported that H₂S successfully improved AI symptoms and corrected iron disturbance in animal models via AMPK activation. In line with our observations, other groups also reported the activation of AMPK by H₂S in cerebral cortex (45) and microglial cells (81). However, few studies sought to assess the role of AMPK in the H₂S-mediated anti-inflammatory processes. Zhou et al. (81) suggested that AMPK was responsible for the suppression of neuroinflammation by H₂S. In accordance with their findings, our data indicated that AMPK contributed to the suppression of hepatic inflammation by H₂S both in vitro and in vivo, suggesting multiple physiological functions of AMPK.

As shown in Figure 2, the inhibition of hepcidin by H₂S was not completely abolished by AMPK siRNA or DN-AMPK, which might be attributed to limited knockdown or inhibition efficiency and potential compensation by alternative H₂S-mediated signaling pathways (38). We recently demonstrated that H₂S inhibited STAT3 phosphorylation by activating Sirtuin 1 (SIRT1) (65). Since AMPK was reported to regulate SIRT1 activity in skeletal muscle (6), we tested whether a similar pattern was also in the present model. As shown in Supplementary Figure S9, the suppression of STAT3 and hepcidin activation by AICAR was not rescued by EX-527, a specific SIRT1 inhibitor. These data indicated that AMPK suppressed STAT3/hepcidin in an SIRT1-independent manner, at least in the current cell model, and that H₂S might activate AMPK and SIRT1 by distinct mechanisms.

AMPK is a metabolic sensor that is highly conserved from yeast to mammals. As a heterotrimeric complex consisted of three subunits, AMPK is activated by phosphorylation at Thr172, which leads to stimulation of glucose utilization and inhibition of gluconeogenesis; therefore, AMPK is regarded as an ideal therapeutic target for T2DM. In addition to its role in energy homeostasis, AMPK is also involved in cancer, autophagy, and inflammation (15, 44). By suppressing the mammalian target of rapamycin signaling and ACC activity, AMPK and metformin impair protein and fatty acid synthesis, and, consequently, inhibit proliferation of breast cancer cells (13). Similar signaling patterns were also observed in induction of autophagy by AMPK (31, 59).

Previous studies investigating the crosstalk between inflammation and AMPK suggested that IL-6, a typical inflammatory cytokine, activated AMPK in a tissue-specific manner. Specifically, IL-6 increased AMPK phosphorylation in both skeletal muscle and adipose tissue, whereas exercise-mediated AMPK activation was diminished in Il-6 knockout mice (28). However, in the same study, the authors found that hepatic AMPK activation was IL-6 independent. In agreement with these earlier observations, we did not find any significant change in AMPK phosphorylation during IL-6 and turpentine; however, we observed a decrease in AMPK kinase activity.

On the contrary, the mechanisms underlying the anti-inflammatory effects of AMPK are not clear. According to a recent review by O'Neill (52), AMPK-mediated pseudo-starvation leads to increased recruitment of anti-inflammatory M2 macrophages and regulatory T cells. In the present study, one of the most notable findings was AMPK-induced JAK2 degradation via the ubiquitin-proteasome pathway. By promoting the interaction between JAK2 and SOCS1, as an endogenous suppressor of cytokine signaling, AMPK blocked IL-6/STAT3 signaling, thus ameliorating downstream inflammatory response. A similar strategy was used in the development of antitumor drugs, such as the targeted degradation of androgen receptor (74) and the cellular bromodomain protein BRD4 (42). Conversely, JAK family members are attractive pharmacological targets based on their roles in immune-driven diseases. Several JAK inhibitors, such as tofacitinib and ruxolitinib, have been approved for treatment of rheumatoid arthritis and myelofibrosis, respectively. Our findings might provide new perspectives and therapeutic targets for AI.

Previous studies indicated diabetes as an independent risk factor for anemia (14, 23), whereas anemia was shown as a risk factor for cardiovascular disease and all-cause mortality in diabetes (70). It is generally accepted that kidney failure contributes to anemia in diabetic patients. However, Grossman et al. reported that anemia was more common even in diabetic patients with normal renal function compared with nondiabetic subjects (10.8% vs. 2.7%, p < 0.001) (19). In addition, anemia may develop before renal impairment (12) and has a two- to threefold higher prevalence rate (23%) in diabetic patients than in subjects in the general population with comparable renal impairment and iron stores (67). These observations suggest additional causes for diabetic anemia.

Indeed, chronic inflammation, indicated by elevated serum IL-6, is intimately involved in the pathogenesis of T2DM (55). Furthermore, the presence of AI is evidenced by both iron dysregulation (25) and suppressed erythropoiesis despite normal or increased erythropoietin (9). Accordingly, several groups reported increased hepcidin levels in T2DM patients (1, 24), whereas others claim conflicting results (61). In the present work, T2DM patients with renal impairment were excluded to rule out the potential confounding effect of renal anemia. Consequently, our analysis determined that the metformin-treated group had lower serum hepcidin levels and anemia prevalence, supporting our hypothesis that hepcidin was a mediator in diabetic anemia. Future investigation is necessary to further elucidate this outcome.

Metformin, a biguanide antidiabetic drug, has been applied clinically for more than half a century. Recent studies revealed that metformin's beneficial effects extended beyond diabetes. Several groups reported that metformin was related to a reduced risk of cancer in diabetic patients (40). Other reports suggested that metformin retarded aging in Caenorhabditis elegans (4) and improved lifespan in mice (43). However, the relationship between metformin and anemia is not clear. In the only large study published to date, although anemia prevalence was higher in patients taking metformin, the results were debatable due to the lack of proper controls and statistically significant difference (2).

Conversely, few studies explored the relationship between metformin and hepcidin, with conflicting results. Here in the present study, we demonstrated for the first time that metformin was associated with reduced serum hepcidin levels in Chinese male T2DM patients. What is more, the inverse correlation of serum hemoglobin with cardiovascular outcomes (70), and cardiac and renal benefits and improved quality of life with correction of hemoglobin levels are well established (58, 63). Our results showing the association of metformin with increased hemoglobin in Chinese T2DM patients, therefore, warrant further investigation into the role of hemoglobin and hepcidin in the improved clinical outcomes by metformin in diabetic patients. It is also worth noting that anemia morbidity is decreased in the metformin group, suggesting metformin as a potential therapeutic agent for anemia not only in diabetes but also in other inflammatory diseases.

Materials and Methods

Reagents and antibodies

NaHS, turpentine, metformin, and 1,10-phenanthroline monohydrate were purchased from Sigma-Aldrich. AICAR, AG490, WP1066, and MG132 were obtained from Selleck Chemicals. CHX was purchased from Melonepharma.

Recombinant human and murine IL-6 were obtained from Peprotech. Antibodies to total JAK2 (catalog 3230), phospho-ACC (catalog 3661), total (catalog 9139) and phospho-STAT3 (Tyr705, catalog 9145), and total (catalog 5832) and phospho-AMPK (Thr172, catalog 2535) were purchased from Cell Signaling Technology. Antibody to phospho-JAK2 (Tyr221, catalog BS4837) was acquired from Bioworld Technology. Antibodies to ubiquitin (catalog 10201-2-AP), HA-tag (catalog 66006-1-Ig), and GAPDH (catalog 60004-1-Ig) were obtained from Proteintech Group. Antibodies to SOCS1 (catalog RLT4362), total (catalog RLP0154) and phospho-JAK1 (catalog RLP2424) were purchased from Ruiyingbio. In addition, antibodies to CSE (catalog sc-374249) and CBS (catalog sc-67154) were obtained from Santa Cruz Biotechnology.

Animals and induction of chronic turpentine model

All animal experimental protocols complied with the Animal Management Rules of the Ministry of Health of the People's Republic of China, and they were approved by the Animal Care Committee of Fudan University.

Eight-week-old male C57BL/6 mice were purchased from Sippr-bk Experimental Animal Center. Mice were housed in specific pathogen-free rooms at 25°C and maintained under a 12-h/12-h light/dark cycle with ad libitum access to food and water.

For the turpentine model, C57BL/6 mice were subcutaneously injected with turpentine (100 μl/20 g weight) once a week for 4 weeks to induce chronic AI as described earlier (50). Compound treatment was performed twice a week, starting from the second injection of turpentine. Mice were sacrificed 1 week after the last injection, after anesthesia with pentobarbital sodium.

Cell culture and isolation of mouse primary hepatocytes

Two human hepatoma cell lines, Huh7 (Japanese Collection of Research Bioresources) and HepG2 (American Type Culture Collection), were cultured in high-glucose DMEM that was supplemented with 10% fetal bovine serum (ExCell Bio) and 1% penicillin and streptomycin (Gibco).

Mouse primary hepatocytes were isolated from 8-week-old C57BL/6 mice by using a modified two-step collagenase perfusion protocol as previously described (76).

Antibody array

Antibody arrays were conducted with the PathScan Intracellular Signaling Array Kit (#7744; Cell Signaling Technology) by following the manufacturer's instructions. The array slides were visualized and analyzed with the LICOR Odyssey Sa Infrared Imaging System. The intensity of each array was balanced by internal negative controls and then normalized to internal positive controls. The MeV software (www.tm4.org/mev.html) was used to plot the heat maps.

Plasmid, lentivirus, and siRNA preparation

Human CBS lentiviral vector was purchased from Obio Tech. pcDNA3.1 CSE construct was from Dr. Yaqi Shen as previously described (62). Human JAK2 overexpression plasmid was kindly provided by Dr. Shenmeng Gao (77). HA-ubiquitin construct was donated by Dr. Qunying Lei (41). Murine CA-AMPKα2 (1–312, T172D) construct was kindly provided by Dr. Xiaolong Liu (7); dominant negative AMPKα2 was obtained by K45R mutation of the wide type as previously described (79). The 2 AMPK plasmids were then subcloned to lentiviral vectors. SOCS1 construct and the SOCS Box deleted form were kindly provided by Dr. Gerardo Ferbeyre (5). All plasmids were purified by using an EndoFree Plasmid Maxiprep Kit (Biotool). Transient transfection was performed by using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions.

Lentivirus packaging was carried out with the psPAX2/VSVG system and HEK 293T cells by using FuGENE HD transfection reagent (Promega Biotech). For in vivo application, the lentivirus was concentrated with ultra-centrifugation to reach a titer of 5–10 × 10⁸ TU/ml.

For silencing experiments in vitro, Huh7 cells and mouse primary hepatocytes were transfected with siRNA (siAMPK-1: 5′-GAGGAGAGCUAUUUGAUUATT-3′; siAMPK-2: 5′-GCGUGUACGAAGGAAGAAU-3′; siSocs1: 5′-CTACCT GAGTTCCTTCCCCTT-3′) by using Lipofectamine 2000, as previously described (76). For murine RNAi in vivo, 2′OMe-modified siRNA of Socs1 was rapidly injected into mouse tail vein (50 μg in 1 ml normal saline). Analysis was performed 24 h later.

AMPK kinase assay

After the indicated treatment, mouse primary hepatocytes were lysed in immunoprecipitation buffer and immunoprecipitated with AMPK antibody. The precipitates were combined with 0.2 mM SAMS peptide (HMRSAMSGLHLVKRR; Biotool). The kinase reaction was conducted at 37°C for 30 min by using ADP-Glo™ Kinase Assay system (Promega).

Immunofluorescence staining

Immunofluorescence staining was conducted as previously described (76). Briefly, mouse primary hepatocytes were fixed by 4% formaldehyde and treated with cold methanol for 10 min at −20°C. The cells were incubated with specific primary antibodies overnight after blocking for 1 h. The slides were observed by confocal laser scanning microscopy (Zeiss LSM 710).

RNA isolation and real-time qRT-PCR

RNA extraction and real-time qRT-PCR were carried out as previously described (76) with Bio-Rad CFX Connect Real-Time PCR System and primers as follows: human β-ACTIN forward 5′-AGGATGCAGAAGGAGATCACTG-3′, reverse 5′-GGGTGTAACGCAACTAAGTCATAG-3′; human HAMP forward 5′-CTGCAACCCCAGGACAGAG-3′, reverse 5′-GGAATAAATAAGGAAGGGAGGGG-3′; murine β-actin forward 5′-TGTTACCAACTGGGACGACA-3′, reverse 5′-GGTGTTGAAGGTCTCAAA-3′; murine Hamp forward 5′-AGAGCTGCAGCCTTTGCAC-3′, reverse 5′-GAAGATGCAGATGGGGAAGT-3′; murine Il-6 forward 5′-TGTGCAATGGCAATTCTGAT-3′, reverse 5′-CCAGA GGAAATTTTCAATAGGC-3′; murine Jak2 forward 5′-TTGTGGTATTACGCCTGTGTATC-3′, reverse 5′-ATGC CTGGTTGACTCGTCTAT-3′; specificity of all PCR products was confirmed by melting curve analysis.

Immunoprecipitation and immunoblot analysis

Immunoprecipitation and immunoblot were performed as previously described (76). In brief, cells were lysed in IP-specific buffer (20 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, pH 7.5) and immunoprecipitated with antibody and Protein A/G Plus-Agarose (Santa Cruz Biotechnology). SDS-PAGE and transferring procedure were carried out with the standard Western blot system (Bio-Rad).

In vivo ubiquitylation assay

In vivo ubiquitylation assay was performed as previously described (26). Briefly, mouse primary hepatocytes were first lysed in 1% SDS buffer (50 mM Tris, 0.5 mM EDTA, 1 mM dithiothreitol, pH 7.5) with MG132 and boiled for 10 min. The lysates were 10-fold diluted in Tris-HCl buffer before analysis of ubiquitylation levels by immunoprecipitation and Western blot.

Dual-luciferase reporter assay

Dual-luciferase reporter assay was conducted by using the pGL3 vector containing human hepcidin promoter (−1000/+71) and the pGL4.47 [luc2P/SIE/Hygro] vector (Promega Biotech). Relative luciferase activities were determined with the Dual Luciferase System (Promega Biotech) as previously described (76).

LDH assay

The LDH leakage assay was conducted as described by the manufacturers (Beyotime). The LDH leakage (%) was calculated by using the following equation: [(A − C)/(P − C)] × 100%, where A is the absorbance of the tested samples, C is the medium control, and P is the positive control (Triton X-100).

Serum iron parameters, hepcidin, and IL-6 analysis

Mouse serum iron levels and total iron-binding capacity (TIBC) were determined with a colorimetric method (Jiancheng Bioengineering Institute). Transferrin saturation was calculated as serum iron/TIBC × 100%.

Serum hepcidin and IL-6 were quantified by ELISA following protocols described by the manufacturers (both murine and human hepcidin kit from Uscn) (66, 68) (IL-6 kit from Boatman).

Tissue non-heme iron analysis

Tissue non-heme iron was measured by 1,10-phenanthroline monohydrate method as described in our previous work (76). Generally, mouse spleen samples were digested in acid solution after being dried at 65°C. The supernatant was collected and mixed with 1,10-phenanthroline monohydrate. The absorbance at 510 nm was measured for determination of Fe (II) concentration.

Perl's Prussian blue staining of paraffin-embedded sections

Perl's Prussian blue staining of mouse spleen paraffin sections was conducted as described and captured with the Zeiss Axio Scope A1 system (76).

Hematological analysis and Wright-Giemsa staining

After anesthesia with pentobarbital sodium, whole blood samples were obtained from the retro-orbital sinus and collected in tubes that were pretreated with EDTA. Complete blood count was performed with SYSMEX XT-2000i automatic analyzer.

Peripheral blood smears were carried out by using 3 μl fresh whole blood that was treated with EDTA, and prepared for Wright-Giemsa staining (Yeasen Biotech). The slides were visualized with the Zeiss Axio Scope A1 system.

Collection and analysis of serum samples from type 2 diabetic patients

Serum samples that were frozen between September 2015 and March 2016 were collected from the hospital tissue bank, and the results of complete blood count from T2DM were acquired from the Department of Endocrinology, Huashan Hospital, Fudan University, Shanghai, China. Written informed consent was obtained from all patients before inclusion. All procedures related were approved by the Ethics Committee of Huashan Hospital, Fudan University, and in accordance with the Declaration of Helsinki. Anemia was determined by the reference hemoglobin values in Huashan Hospital: <115 g/L for women and <130 g/L for men.

Exclusion criteria include infections during the previous week, tumor-related diseases, chronic kidney diseases, auto-immune diseases, history of organ transplantation and on anti-rejection drugs, and participation in clinical trials during the past 6 months. Patients taking metformin for less than 3 months were also excluded.

Measurement of serum H₂S levels

The H₂S levels of serum samples from T2DM patients were measured as previously described (33) with minor revision. Briefly, 30 μl serum samples were mixed with 10 μl Tris-HCl buffer (100 mM, pH 8.5) and 70 μl monobromobimane reagent. After vortexing for 1 h in the dark, the samples were centrifuged at 13,000 g, 4°C for 20 min after the addition of 10 μl 20% formic acid. The supernatant was subjected to reverse-phase HPLC for H₂S determination.

Statistics

Data are expressed as the mean ± SEM. Statistical analysis of continuous variables was performed with t-test, Mann–Whitney U test, or one-way ANOVA followed by Turkey's test. For counting variables, chi-square or Fisher's exact probability test was conducted. A two-tailed p < 0.05 was considered statistically significant.

Footnotes

Acknowledgments

The authors would like to thank the following individuals for providing plasmids as stated in the methods: Dr. Shenmeng Gao from the First Affiliated Hospital of Wenzhou Medical College, Dr. Qunying Lei and Dr. Yaqi Shen from Fudan University, Dr. Xiaolong Liu from Shanghai Institutes for Biological Sciences, and Dr. Gerardo Ferbeyre from Université de Montréal. This work was supported by grants from the National Natural Science Foundation of China (No. 81330080; 81573421; 81402917) and the Shanghai Committee of Science and Technology of China (No. 14JC1401100).

Authors' Contributions

M. W. and H. X. designed and conducted the experiments. M. W. and W. T. analyzed the data and wrote the article. Y. L., Z. Z., and L. F. collected clinical samples. L. M. and X. W. performed animal experiments. B. T. analyzed serum H₂S levels. H. X. and Y.Z.Z. reviewed and edited the article.

Author Disclosure Statement

No competing financial interests exist.

Abbreviations Used

References

Andrews

, Soto

, and Arredondo-Olguín

. Association between ferritin and hepcidin levels and inflammatory status in patients with type 2 diabetes mellitus and obesity. Nutrition, 31: 51–57, 2015.

Aroda

, Edelstein

, Goldberg

, Knowler

, Marcovina

, Orchard

, Bray

, Schade

, Temprosa

, and White

. Long-term metformin use and vitamin B12 deficiency in the Diabetes Prevention Program Outcomes Study. J Clin Endocrinol Metab, 101: 1754–1761, 2016.

Bohlius

, Schmidlin

, Brillant

, Schwarzer

, Trelle

, Seidenfeld

, Zwahlen

, Clarke

, Weingart

, and Kluge

. Recombinant human erythropoiesis-stimulating agents and mortality in patients with cancer: a meta-analysis of randomised trials. Lancet, 373: 1532–1542, 2009.

Cabreiro

, Au

, Leung

K-Y

, Vergara-Irigaray

, Cochemé

, Noori

, Weinkove

, Schuster

, Greene

, and Gems

. Metformin retards aging in C. elegans by altering microbial folate and methionine metabolism. Cell, 153: 228–239, 2013.

Calabrese

, Mallette

, Deschênes-Simard

, Ramanathan

, Gagnon

, Moores

, Ilangumaran

, and Ferbeyre

. SOCS1 links cytokine signaling to p53 and senescence. Mol Cell, 36: 754–767, 2009.

Cantó

, Gerhart-Hines

, Feige

, Lagouge

, Noriega

, Milne

, Elliott

, Puigserver

, and Auwerx

. AMPK regulates energy expenditure by modulating NAD+ metabolism and SIRT1 activity. Nature, 458: 1056–1060, 2009.

Cao

, Li

, Liu

, Zheng

, Ji

, and Liu

. The serine/threonine kinase LKB1 controls thymocyte survival through regulation of AMPK activation and Bcl-XL expression. Cell Res, 20: 99–108, 2010.

Caro

, Salas

, Ward

, and Goss

. Anemia as an independent prognostic factor for survival in patients with cancer. Cancer, 91: 2214–2221, 2001.

Craig

, Williams

, Riley

, Smith

, Owens

, Worthing

, Cavill

, and Phillips

. Anemia and diabetes in the absence of nephropathy. Diabetes Care, 28: 1118–1123, 2005.

10.

Day

, Duquaine

, Mundada

, Menon

, Khan

, Rajagopalan

, and Fay

. Chronic iron administration increases vascular oxidative stress and accelerates arterial thrombosis. Circulation, 107: 2601–2606, 2003.

11.

de Jonge

, van der Zanden

, The

, Bijlsma

, van Westerloo

, Bennink

, Berthoud

H-R

, Uematsu

, Akira

, and van den Wijngaard

. Stimulation of the vagus nerve attenuates macrophage activation by activating the Jak2-STAT3 signaling pathway. Nat Immunol, 6: 844–851, 2005.

12.

Dikow

, Schwenger

, Schömig

, and Ritz

. How should we manage anaemia in patients with diabetes?. Nephrol Dial Transplant, 17: 67–72, 2002.

13.

Dowling

, Zakikhani

, Fantus

, Pollak

, and Sonenberg

. Metformin inhibits mammalian target of rapamycin–dependent translation initiation in breast cancer cells. Cancer Res, 67: 10804–10812, 2007.

14.

El-Achkar

, Ohmit

, Mccullough PA

D C

, ROOK

, Brown

, Grimm

, Bakris

, Keane

, and Flack

. Higher prevalence of anemia with diabetes mellitus in moderate kidney insufficiency: the Kidney Early Evaluation Program. Kidney Int, 67: 1483–1488, 2005.

15.

Faubert

, Boily

, Izreig

, Griss

, Samborska

, Dong

, Dupuy

, Chambers

, Fuerth

, and Viollet

. AMPK is a negative regulator of the Warburg effect and suppresses tumor growth in vivo. Cell Metab, 17: 113–124, 2013.

16.

Foley

, Parfrey

, Harnett

, Kent

, Murray

, and Barre

. The impact of anemia on cardiomyopathy, morbidity, and mortality in end-stage renal disease. Am J Kidney Dis, 28: 53–61, 1996.

17.

Ganz

. Hepcidin, a key regulator of iron metabolism and mediator of anemia of inflammation. Blood, 102: 783–788, 2003.

18.

Ganz

and Nemeth

. Iron sequestration and anemia of inflammation. Semin Hematol, 46: 387–393, 2009.

19.

Grossman

, Dovrish

, Koren-Morag

, Bornstein

, and Leibowitz

. Diabetes mellitus with normal renal function is associated with anaemia. Diabetes Metab Res Rev, 30: 291–296, 2014.

20.

Group

DPPR

. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med, 346: 393, 2002.

21.

Horwich

, Fonarow

, Hamilton

, MacLellan

, and Borenstein

. Anemia is associated with worse symptoms, greater impairment in functional capacity and a significant increase in mortality in patients with advanced heart failure. J Am Coll Cardiol, 39: 1780–1786, 2002.

22.

Ikeda

, Iwata

, and Murakami

. Inhibitory effect of metformin on intestinal glucose absorption in the perfused rat intestine. Biochem Pharmacol, 59: 887–890, 2000.

23.

Ishimura

, Nishizawa

, Okuno

, Matsumoto

, Emoto

, Inaba

, Kawagishi

, Kim

C-W

, and Morii

. Diabetes mellitus increases the severity of anemia in non-dialyzed patients with renal failure. J Nephrol, 11: 83–86, 1997.

24.

Jiang

, Sun

Z-Z

, Tang

Y-T

, Xu

, and Jiao

X-Y

. Hepcidin expression and iron parameters change in Type 2 diabetic patients. Diabetes Res Clin Pract, 93: 43–48, 2011.

25.

Jiang

, Manson

, Meigs

, Ma

, Rifai

, and Hu

. Body iron stores in relation to risk of type 2 diabetes in apparently healthy women. JAMA, 291: 711–717, 2004.

26.

Jiang

, Wang

, Xiao

, Lin

, Zhou

, Lei

, Xiong

, Guan

K-L

, and Zhao

. Acetylation regulates gluconeogenesis by promoting PEPCK1 degradation via recruiting the UBR5 ubiquitin ligase. Mol Cell, 43: 33–44, 2011.

27.

Kahn

, Alquier

, Carling

, and Hardie

. AMP-activated protein kinase: ancient energy gauge provides clues to modern understanding of metabolism. Cell Metab, 1: 15–25, 2005.

28.

Kelly

, Keller

, Avilucea

, Keller

, Luo

, Xiang

, Giralt

, Hidalgo

, Saha

, and Pedersen

. AMPK activity is diminished in tissues of IL-6 knockout mice: the effect of exercise. Biochem Biophys Res Commun, 320: 449–454, 2004.

29.

Kemna

, Pickkers

, Nemeth

, van der Hoeven

, and Swinkels

. Time-course analysis of hepcidin, serum iron, and plasma cytokine levels in humans injected with LPS. Blood, 106: 1864–1866, 2005.

30.

Kile

, Schulman

, Alexander

, Nicola

, Martin

, and Hilton

. The SOCS box: a tale of destruction and degradation. Trends Biochem Sci, 27: 235–241, 2002.

31.

Kim

, Kundu

, Viollet

, and Guan

K-L

. AMPK and mTOR regulate autophagy through direct phosphorylation of Ulk1. Nat Cell Biol, 13: 132–141, 2011.

32.

Kimura

, Dargusch

, Schubert

, and Kimura

. Hydrogen sulfide protects HT22 neuronal cells from oxidative stress. Antioxid Redox Signal, 8: 661–670, 2006.

33.

Kimura

, Goto

Y-I

, and Kimura

. Hydrogen sulfide increases glutathione production and suppresses oxidative stress in mitochondria. Antioxid Redox Signal, 12: 1–13, 2010.

34.

Langdon

, Yates

, Femnou

, McCranor

, Cheadle

, Xue

, Vaulont

, Civin

, Walston

, and Roy

. Hepcidin-dependent and hepcidin-independent regulation of erythropoiesis in a mouse model of anemia of chronic inflammation. Am J Hematol, 89: 470–479, 2014.

35.

Lee

, Peng

, Gelbart

, and Beutler

. The IL-6-and lipopolysaccharide-induced transcription of hepcidin in HFE-, transferrin receptor 2-, and β2-microglobulin-deficient hepatocytes. Proc Natl Acad Sci U S A, 101: 9263–9265, 2004.

36.

Leonard

and O'Shea

. Jaks and STATs: biological implications. Annu Rev Immunol, 16: 293–322, 1998.

37.

, Bhatia

, and Moore

. Hydrogen sulphide—a novel mediator of inflammation?. Curr Opin Pharmacol, 6: 125–129, 2006.

38.

, Rose

, and Moore

. Hydrogen sulfide and cell signaling. Annu Rev Pharmacol Toxicol, 51: 169–187, 2011.

39.

, Wang

, Zhou

, Ke

, de Waal

, Gu

, Tan

, Wang

, Wu

, and Xu

. Structural basis of AMPK regulation by adenine nucleotides and glycogen. Cell Res, 25: 50–66, 2015.

40.

Libby

, Donnelly

, Donnan

, Alessi

, Morris

, and Evans

. New users of metformin are at low risk of incident cancer: a cohort study among people with type 2 diabetes. Diabetes Care, 32: 1620–1625, 2009.

41.

Lin

, Tao

, Gao

, Li

, Zhou

, Guan

K-L

, Xiong

, and Lei

Q-Y

. Acetylation stabilizes ATP-citrate lyase to promote lipid biosynthesis and tumor growth. Mol Cell, 51: 506–518, 2013.

42.

, Qian

, Altieri

, Dong

, Wang

, Raina

, Hines

, Winkler

, Crew

, and Coleman

. Hijacking the E3 ubiquitin ligase cereblon to efficiently target BRD4. Chem Biol, 22: 755–763, 2015.

43.

Martin-Montalvo

, Mercken

, Mitchell

, Palacios

, Mote

, Scheibye-Knudsen

, Gomes

, Ward

, Minor

, and Blouin

M-J

. Metformin improves healthspan and lifespan in mice. Nat Commun, 4: 2192, 2013.

44.

Mihaylova

and Shaw

. The AMPK signalling pathway coordinates cell growth, autophagy and metabolism. Nat Cell Biol, 13: 1016–1023, 2011.

45.

Minamishima

, Bougaki

, Sips

, De Yu

, Minamishima

, Elrod

, Lefer

, Bloch

, and Ichinose

. Hydrogen sulfide improves survival after cardiac arrest and cardiopulmonary resuscitation via a nitric oxide synthase 3–dependent mechanism in mice. Circulation, 120: 888–896, 2009.

46.

Murray

. The JAK-STAT signaling pathway: input and output integration. J Immunol, 178: 2623–2629, 2007.

47.

Nemeth

and Ganz

. Anemia of inflammation. Hematol Oncol Clin North Am, 28: 671–681, 2014.

48.

Nemeth

, Rivera

, Gabayan

, Keller

, Taudorf

, Pedersen

, and Ganz

. IL-6 mediates hypoferremia of inflammation by inducing the synthesis of the iron regulatory hormone hepcidin. J Clin Invest, 113: 1271–1276, 2004.

49.

Nemeth

, Valore

, Territo

, Schiller

, Lichtenstein

, and Ganz

. Hepcidin, a putative mediator of anemia of inflammation, is a type II acute-phase protein. Blood, 101: 2461–2463, 2003.

50.

Nicolas

, Chauvet

, Viatte

, Danan

, Bigard

, Devaux

, Beaumont

, Kahn

, and Vaulont

. The gene encoding the iron regulatory peptide hepcidin is regulated by anemia, hypoxia, and inflammation. J Clin Invest, 110: 1037–1044, 2002.

51.

O'Meara

, Clayton

, McEntegart

, McMurray

, Lang

, Roger

, Young

, Solomon

, Granger

, and Östergren

. Clinical correlates and consequences of anemia in a broad spectrum of patients with heart failure results of the Candesartan in Heart Failure: Assessment of Reduction in Mortality and Morbidity (CHARM) Program. Circulation, 113: 986–994, 2006.

52.

O'Neill

and Hardie

. Metabolism of inflammation limited by AMPK and pseudo-starvation. Nature, 493: 346–355, 2013.

53.

Paul

, Hong

, Sachidanandan

, Babitt

, Deng

, Hoyng

, Lin

, Bloch

, and Peterson

. Dorsomorphin inhibits BMP signals required for embryogenesis and iron metabolism. Nat Chem Biol, 4: 33–41, 2008.

54.

Pietrangelo

, Dierssen

, Valli

, Garuti

, Rump

, Corradini

, Ernst

, Klein

, and Trautwein

. STAT3 is required for IL-6-gp130–dependent activation of hepcidin in vivo. Gastroenterology, 132: 294–300, 2007.

55.

Pradhan

, Manson

, Rifai

, Buring

, and Ridker

. C-reactive protein, interleukin 6, and risk of developing type 2 diabetes mellitus. JAMA, 286: 327–334, 2001.

56.

Prince

, Langdon

, Layman

, Prince

, Sabogal

, Mak

, Berger

, Cheadle

, Chrest

, and Yu

. Late stage erythroid precursor production is impaired in mice with chronic inflammation. Haematologica, 97: 1648–1656, 2012.

57.

Reagan-Shaw

, Nihal

, and Ahmad

. Dose translation from animal to human studies revisited. FASEB J, 22: 659–661, 2008.

58.

Ritz

, Laville

, Bilous

, O'Donoghue

, Scherhag

, Burger

, De Alvaro

, and Investigators

. Target level for hemoglobin correction in patients with diabetes and CKD: primary results of the Anemia Correction in Diabetes (ACORD) Study. Am J Kidney Dis, 49: 194–207, 2007.

59.

Russell

, Yuan

H-X

, and Guan

K-L

. Autophagy regulation by nutrient signaling. Cell Res, 24: 42–57, 2014.

60.

Sakamori

, Takehara

, Tatsumi

, Shigekawa

, Hikita

, Hiramatsu

, Kanto

, and Hayashi

. STAT3 signaling within hepatocytes is required for anemia of inflammation in vivo. J Gastroenterol, 45: 244–248, 2010.

61.

Sam

, Busbridge

, Amin

, Webber

, White

, Franks

, Martin

, Sleeth

, Ismail

, and Daud

. Hepcidin levels in diabetes mellitus and polycystic ovary syndrome. Diabet Med, 30: 1495–1499, 2013.

62.

Shen

, Shen

, Miao

, Xin

, Lin

, Zhu

, Guo

, and Zhu

. miRNA-30 family inhibition protects against cardiac ischemic injury by regulating cystathionine-γ-lyase expression. Antioxid Redox Signal, 22: 224–240, 2015.

63.

Silverberg

, Wexler

, and Iaina

. The importance of anemia and its correction in the management of severe congestive heart failure. Eur J Heart Fail, 4: 681–686, 2002.

64.

Suzuki

, Olah

, Modis

, Coletta

, Kulp

, Gerö

, Szoleczky

, Chang

, Zhou

, and Wu

. Hydrogen sulfide replacement therapy protects the vascular endothelium in hyperglycemia by preserving mitochondrial function. Proc Natl Acad Sci U S A, 108: 13829–13834, 2011.

65.

Teehan

, Bahdouch

, Ruthazer

, Balakrishnan

, Snydman

, and Jaber

. Iron storage indices: novel predictors of bacteremia in hemodialysis patients initiating intravenous iron therapy. Clin Infect Dis, 38: 1090–1094, 2004.

66.

Tesfay

, Clausen

, Kim

, Hegde

, Wang

, Miller

, Deng

, Blanchette

, Arvedson

, and Miranti

. Hepcidin regulation in prostate and its disruption in prostate cancer. Cancer Res, 75: 2254–2263, 2015.

67.

Thomas

, MacIsaac

, Tsalamandris

, Power

, and Jerums

. Unrecognized anemia in patients with diabetes: a cross-sectional survey. Diabetes Care, 26: 1164–1169, 2003.

68.

Vecchi

, Montosi

, Garuti

, Corradini

, Sabelli

, Canali

, and Pietrangelo

. Gluconeogenic signals regulate iron homeostasis via hepcidin in mice. Gastroenterology, 146: 1060–1069. e3, 2014.

69.

Vincent

, Baron

J-F

, Reinhart

, Gattinoni

, Thijs

, Webb

, Meier-Hellmann

, Nollet

, Peres-Bota

, and Investigators

. Anemia and blood transfusion in critically ill patients. JAMA, 288: 1499–1507, 2002.

70.

Vlagopoulos

, Tighiouart

, Weiner

, Griffith

, Pettitt

, Salem

, Levey

, and Sarnak

. Anemia as a risk factor for cardiovascular disease and all-cause mortality in diabetes: the impact of chronic kidney disease. J Am Soc Nephrol, 16: 3403–3410, 2005.

71.

Weiss

and Goodnough

. Anemia of chronic disease. N Engl J Med, 352: 1011–1023, 2005.

72.

Wellen

and Hotamisligil

. Inflammation, stress, and diabetes. J Clin Invest, 115: 1111–1119, 2005.

73.

Williamson

, Bolster

, Kimball

, and Jefferson

. Time course changes in signaling pathways and protein synthesis in C2C12 myotubes following AMPK activation by AICAR. Am J Physiol Endocrinol Metab, 291: E80–E89, 2006.

74.

Winkler

, Jin

, Crew

, Rossi

, Willard

, Dong

, Siu

, Wang

, Gordon

, and Chen

. Abstract PR08: ARV-330:an androgen receptor PROTAC degrader for prostate cancer. Mol Cancer Ther, 14: PR08, 2015.

75.

Wrighting

and Andrews

. Interleukin-6 induces hepcidin expression through STAT3. Blood, 108: 3204–3209, 2006.

76.

Xin

, Wang

, Tang

, Shen

, Miao

, Wu

, Li

, Wang

, Xin

, and Zhu

. Hydrogen sulfide attenuates inflammatory hepcidin by reducing IL-6 secretion and promoting SIRT1-mediated STAT3 deacetylation. Antioxid Redox Signal, 24: 70–83, 2016.

77.

Yin

L-H

, Zheng

X-Q

, Li

H-Y

, Bi

L-X

, Shi

Y-F

, Ye

A-F

, Wu

J-B

, and Gao

S-M

. Epigenetic deregulated miR-375 contributes to the constitutive activation of JAK2/STAT signaling in myeloproliferative neoplasm. Leuk Res, 39: 471–478, 2015.

78.

Yoshimura

and Yasukawa

. JAK's SOCS: a mechanism of inhibition. Immunity, 36: 157–159, 2012.

79.

Zang

, Xu

, Maitland-Toolan

, Zuccollo

, Hou

, Jiang

, Wierzbicki

, Verbeuren

, and Cohen

. Polyphenols stimulate AMP-activated protein kinase, lower lipids, and inhibit accelerated atherosclerosis in diabetic LDL receptor–deficient mice. Diabetes, 55: 2180–2191, 2006.

80.

Zhou

, Myers

, Li

, Chen

, Shen

, Fenyk-Melody

, Wu

, Ventre

, Doebber

, and Fujii

. Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest, 108: 1167–1174, 2001.

81.

Zhou

, Cao

, Ao

, Hu

, Liu

, Wu

, Wang

, Jin

, Zheng

, and Zhen

. CaMKKβ-dependent activation of AMP-activated protein kinase is critical to suppressive effects of hydrogen sulfide on neuroinflammation. Antioxid Redox Signal, 21: 1741–1758, 2014.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.03 MB

0.07 MB

0.12 MB

0.05 MB

0.09 MB

0.05 MB

0.10 MB

0.09 MB

0.18 MB

0.12 MB

0.11 MB

0.17 MB

0.19 MB

0.18 MB

0.10 MB

0.11 MB

AMPK Serves as a Therapeutic Target Against Anemia of Inflammation

Abstract

Aims:

Results:

Innovation:

Conclusion:

Introduction

Results

H2S-mediated inhibition of STAT3 and hepcidin is dependent on AMPK activation

AMPK mediated the treatment effects of H2S on AI

AMPK activation ameliorates IL-6-induced hepcidin levels both in vitro and in vivo

AMPK activators improve turpentine-induced AI in vivo

H2S and AMPK suppress inflammatory STAT3 and hepcidin by promoting proteasome-mediated JAK2 degradation

SOCS1 mediates the JAK2 degradation induced by AMPK activation

Metformin treatment is associated with lower serum hepcidin levels and reduced anemia morbidity in patients with T2DM

Discussion

Materials and Methods

Reagents and antibodies

Animals and induction of chronic turpentine model

Cell culture and isolation of mouse primary hepatocytes

Antibody array

Plasmid, lentivirus, and siRNA preparation

AMPK kinase assay

Immunofluorescence staining

RNA isolation and real-time qRT-PCR

Immunoprecipitation and immunoblot analysis

In vivo ubiquitylation assay

Dual-luciferase reporter assay

LDH assay

Serum iron parameters, hepcidin, and IL-6 analysis

Tissue non-heme iron analysis

Perl's Prussian blue staining of paraffin-embedded sections

Hematological analysis and Wright-Giemsa staining

Collection and analysis of serum samples from type 2 diabetic patients

Measurement of serum H2S levels

Statistics

Footnotes

Acknowledgments

Authors' Contributions

Author Disclosure Statement

Abbreviations Used

References

Supplementary Material

H₂S-mediated inhibition of STAT3 and hepcidin is dependent on AMPK activation

AMPK mediated the treatment effects of H₂S on AI

H₂S and AMPK suppress inflammatory STAT3 and hepcidin by promoting proteasome-mediated JAK2 degradation

Measurement of serum H₂S levels