Anti-Inflammatory Effects of Metformin Irrespective of Diabetes Status

Abstract

Rationale: The diabetes mellitus drug metformin is under investigation in cardiovascular disease, but the molecular mechanisms underlying possible benefits are poorly understood.

Objective: Here, we have studied anti-inflammatory effects of the drug and their relationship to antihyperglycemic properties.

Methods and results: In primary hepatocytes from healthy animals, metformin and the IKKβ (inhibitor of kappa B kinase) inhibitor BI605906 both inhibited tumor necrosis factor-α-dependent IκB degradation and expression of proinflammatory mediators interleukin-6, interleukin-1β, and CXCL1/2 (C-X-C motif ligand 1/2). Metformin suppressed IKKα/β activation, an effect that could be separated from some metabolic actions, in that BI605906 did not mimic effects of metformin on lipogenic gene expression, glucose production, and AMP-activated protein kinase activation. Equally AMP-activated protein kinase was not required either for mitochondrial suppression of IκB degradation. Consistent with discrete anti-inflammatory actions, in macrophages, metformin specifically blunted secretion of proinflammatory cytokines, without inhibiting M1/M2 differentiation or activation. In a large treatment naive diabetes mellitus population cohort, we observed differences in the systemic inflammation marker, neutrophil to lymphocyte ratio, after incident treatment with either metformin or sulfonylurea monotherapy. Compared with sulfonylurea exposure, metformin reduced the mean log-transformed neutrophil to lymphocyte ratio after 8 to 16 months by 0.09 U (95% confidence interval, 0.02-0.17; P=0.013) and increased the likelihood that neutrophil to lymphocyte ratio would be lower than baseline after 8 to 16 months (odds ratio, 1.83; 95% confidence interval, 1.22-2.75; P=0.00364). Following up these findings in a double-blind placebo controlled trial in nondiabetic heart failure (trial registration: NCT00473876), metformin suppressed plasma cytokines including the aging-associated cytokine CCL11 (C-C motif chemokine ligand 11).

Conclusion: We conclude that anti-inflammatory properties of metformin are exerted irrespective of diabetes mellitus status. This may accelerate investigation of drug utility in nondiabetic cardiovascular disease groups.

Clinical trial registration: Name of the trial registry: TAYSIDE trial (Metformin in Insulin Resistant Left Ventricular [LV] Dysfunction). URL: https://www.clinicaltrials.gov. Unique identifier: NCT00473876.

Keywords: NF-kappa B; cardiovascular diseases; diabetes mellitus; heart failure; inflammation; metabolism; metformin.

© 2016 The Authors.

Figures

Figure 1.

Effect of metformin on nuclear factor-κB (NF-κB) signaling and gene expression. A–C, Primary hepatocytes were incubated in serum-free medium overnight and then stimulated for 3 h with or without 0.5 to 5 mmol/L metformin. For the last 15 min, cells were treated with or without 10 ng/mL tumor necrosis factor (TNF)-α. Cells were lysed and prepared for immunoblotting using antibodies as described in the Methods section of this article. In this figure and elsewhere, each blot is representative of experiments carried out at least 3×. D–F, Primary hepatocytes were incubated as in A–C, before stimulation for 3 h with or without 2 mmol/L metformin and TNF-α. In addition, cells were incubated with/without 10 μmol/L BI605906 or 100 nmol/L rapamycin as shown, before lysis and immunoblotting as described in the Methods section of this article. G–J, Primary hepatocytes were treated with or without 10 ng/mL TNF-α, 2 mmol/L metformin, or 10 μmol/L BI605906 for 8 h followed by cell lysis, RNA extraction, and preparation of cDNA for real-time-polymerase chain reaction using primer sets for individual genes shown as described in the Methods section of this article. ACC indicates acetyl-CoA carboxylase; AMPK, AMP-activated protein kinase; p-ACC, phospho–acetyl-CoA carboxylase; and pAMPK, phospho–AMP-activated protein kinase.

Figure 2.

Effect of 5-aminoimidazole-4-carboxamide riboside (AICAR) and A769662 on nuclear factor-κB (NF-κB) signaling. A–D, Primary wild-type (WT) hepatocytes (A and B) and those taken from double-knockout (KO) AMPK animals or matched controls (WT) (C) were incubated in serum-free medium overnight, before stimulation for 3 h with or without doses of A769662 and AICAR as shown. For the last 15 min, cells were treated with or without 10 ng/mL tumor necrosis factor (TNF)-α. D, Hepatocytes from KO or WT animals treated with and without doses of metformin for 3 h. For the last 15 min, cells were treated with 10 ng/mL TNF-α. Cells were then lysed, and immunoblots were prepared as described in the Methods section of this article and in Figure 1. ACC indicates acetyl-CoA carboxylase; AMPK, AMP-activated protein kinase; IKK, inhibitor of kappa B kinase; p-ACC, phospho–acetyl-CoA carboxylase; and pAMPK, phospho–AMP-activated protein kinase.

Figure 3.

Effects of cytokines on glucose production and lipogenic gene expression in primary hepatocytes. A and B, Primary hepatocytes were treated with/without metformin (2 mmol/L), interleukin (IL)-6 (5 ng/mL), IL-1β (10 ng/mL), chemokine (C-X-C motif) ligand (CXCL) 1 (100 ng/mL) and tumor necrosis factor (TNF)-α (10 ng/mL) for 12 h, and glucose production was measured by GAGO (glucose [glucose oxidase]) assay as described in the Methods section of this article. C–E, Primary hepatocytes were treated with or without 10 ng/mL tumor necrosis factor (TNF)-α, 2 mmol/L metformin, and 10 μmol/L BI605906 for 8 h followed by cell lysis, RNA extraction, and preparation of cDNA for real-time polymerase chain reaction using primer sets for individual genes shown as described in the Methods section of this article. FASN indicates fatty acid synthase; PPAR, peroxisome proliferator–activated receptor; and SREBP, sterol regulatory element-binding protein.

Figure 4.

Effect of metformin and its analogue biguanide on bone marrow–derived macrophages: phenotypic markers and cytokine secretion. A, Macrophages were treated with/without metformin (2 mmol/L) or biguanide (BIG 2 mmol/L) to determine the effect on the M1 and M2 phenotypes of macrophages, which was measured by flow cytometry for CD11c and CD206 expression. The colors denote the following: red, undifferentiated; blue, differentiated, untreated; orange, differentiated, metformin; green, differentiated, BIG. B, Macrophages were treated with/without metformin (2 mmol/L) or biguanide (BIG, 2 mmol/L) to determine the effect on activation in response to 100 ng/mL lipopolysaccharide (LPS), which was measured by studying CD69 and CD40 expression. Histograms are representative of n=4. The colors denote the following: red, unactivated; blue, activated, untreated; orange, activated, metformin; green, activated, BIG. C–E, Macrophages were treated with/without metformin (Met) or BIG (2 mmol/L) to determine the effect of these drugs on IL-6 (C), IL-12p40 (D), and IL-10 (E) production (n=4).

Figure 5.

Effect of long-term metformin treatment on nuclear factor-κB (NF-κB) signaling responses in hepatocytes. A and B, Primary hepatocytes were treated as in Figure 1 with metformin or propanediimidamide (PDI) at the doses indicated except that the treatment time was 24 h. For the last 15 min, cells were treated with 10 ng/mL tumor necrosis factor (TNF)-α. In addition to antibodies used elsewhere, phosphorylation of inhibitor of kappa B kinase (IKK)α/β was investigated using the phosphospecific antibodies indicated. After cell lysis, SDS-PAGE and immunoblotting were performed as in Figure 1. C and D, Hepatocytes from wild-type (WT) and AMPK double-knockout (KO) livers treated as in A or with doses of rotenone for 45 min before cell lysis, SDS-PAGE and immunoblotting. E, Primary hepatocytes were treated in the presence or absence of the agents shown. Cells were treated with/without metformin (2 mmol/L), C-C motif chemokine ligand (CCL)-11 (5 ng/mL), interleukin (IL)-2, IL-4, stromal cell–derived factor (SDF), and CCL22 (10 ng/mL) for 12 h, and glucose production was measured by GAGO (glucose [glucose oxidase]) assay as described in the Methods section of this article. ACC indicates acetyl-CoA carboxylase; AMPK, AMP-activated protein kinase; p-ACC, phospho–acetyl-CoA carboxylase; and pAMPK, phospho–AMP-activated protein kinase.