Humanin: a novel central regulator of peripheral insulin action

Abstract

Background: Decline in insulin action is a metabolic feature of aging and is involved in the development of age-related diseases including Type 2 Diabetes Mellitus (T2DM) and Alzheimer's disease (AD). A novel mitochondria-associated peptide, Humanin (HN), has a neuroprotective role against AD-related neurotoxicity. Considering the association between insulin resistance and AD, we investigated if HN influences insulin sensitivity.

Methods and findings: Using state of the art clamp technology, we examined the role of central and peripheral HN on insulin action. Continuous infusion of HN intra-cerebro-ventricularly significantly improved overall insulin sensitivity. The central effects of HN on insulin action were associated with activation of hypothalamic STAT-3 signaling; effects that were negated by co-inhibition of hypothalamic STAT-3. Peripheral intravenous infusions of novel and potent HN derivatives reproduced the insulin-sensitizing effects of central HN. Inhibition of hypothalamic STAT-3 completely negated the effects of IV HN analog on liver, suggesting that the hepatic actions of HN are centrally mediated. This is consistent with the lack of a direct effect of HN on primary hepatocytes. Furthermore, single treatment with a highly-potent HN analog significantly lowered blood glucose in Zucker diabetic fatty rats. Based upon the link of HN with two age-related diseases, we examined if there were age associated changes in HN levels. Indeed, the amount of detectable HN in hypothalamus, skeletal muscle, and cortex was decreased with age in rodents, and circulating levels of HN were decreased with age in humans and mice.

Conclusions: We conclude that the decline in HN with age could play a role in the pathogenesis of age-related diseases including AD and T2DM. HN represents a novel link between T2DM and neurodegeneration and along with its analogues offers a potential therapeutic tool to improve insulin action and treat T2DM.

Figure 1. Protein level of the rat homolog to HN, termed rattin (RN), in metabolically-relevant tissues in male S–D rats.

A total of 30 µg of rat skeletal muscle (lane 1), liver (lane 2), epididymal fat (lane 3) and hypothalamic protein (lane 4) and 1.5 ng of synthetic RN peptide (lane 5) were loaded. RN protein was detected in muscle, liver and hypothalamus but not in epididymal fat (lane 3).

Figure 2. ICV HN increases peripheral insulin sensitivity.

A) Schematic representation of the experimental design for the ICV studies; the upper panel demonstrates the time line for the surgical procedures and the lower panel demonstrates the protocol on the day of the clamp. B) Glucose infusion rate (GIR) and C) Hepatic glucose production (HGP) and degree of suppression of HGP by artificial cerebrospinal fluid (aCSF) and 0.16 µg/kg/min HN infusions (n = 5 each) during a basal pancreatic clamp. D) GIR, E) HGP and F) degree of HGP suppression during a hyperinsulinemic clamp with ICV infusion of aCSF, HN and a DDHN peptide (n = 7 each). G) Time course and H) last hour average of effects of ICV aCSF, HN and DDHN on peripheral glucose uptake during a hyperinsulinemic clamp. Effect of aCSF (black bars) and HN (white bars) on I) glycolysis, J) glycogen synthesis, and K) suppression of free fatty acids (FFA) in response to insulin. L) pAKT and acetyl CoA carboxylase (ACC) in skeletal muscle during insulin clamp in response to ICV aCSF and HN (n = 7 each). Values are means±SE. *Significantly different from other experimental groups, P<0.05.

Figure 3. Central effects of HN on glucose metabolism involve hypothalamic STAT-3 activation.

A) Effects of ICV HN or aCSF on pSTAT-3^Tyr705 levels in hypothalamic protein extracts. B) Effects of ICV HN and F6AHN on the levels of totalSTAT-3, pAKT^S473, insulin receptor (IR) and pAMPK^Thr172 in the hypothalamus (aCSF -black bars, HN -white bars, F6AHN-grey, n = 5 each). Effects of a STAT-3 inhibitor co-infused ICV with HN or aCSF (n = 6 each) on C) GIR, D) peripheral glucose uptake, and E) HGP and degree of suppression of HGP. Values are means±SE. *Significantly different from other experimental groups, P<0.05.

Figure 4. IGFBP-3 modulates the effects of HN on insulin sensitivity.

A) IGFBP-3 binding of HN and HN analogues measured by densitometry of dot blots probed with radiolabeled IGFBP-3 (Ins = Insulin). B) Mouse embryonic fibroblasts generated from wild-type or IGFBP-3 knockout mice were incubated in serum-free media for 24 h followed by incubation with 100 nM HNG or HNGF6A for 24 h. Apoptosis was assessed by ELISA for fragmentation of histone-associated DNA. (n = 4). Effects of ICV IGFBP-3, HN or F6AHN during a hyperinsulinemic clamp (n = 6 each)on C) GIR, D) HGP and E) degree of suppression of HGP. *Significantly different from other experimental groups, P<0.05.

Figure 5. Effects of peripheral administration of the potent HN analog, HNGF6A, on glucose metabolism.

Effect of IV saline, F6AHN and HNGF6A (n = 6 each) on A) GIR, B) HGP, and C) glucose uptake. Values are means±SE. *Significantly different from other experimental groups, P<0.05.

Figure 6. Effects of IV HNGF6A in the presence of hypothalamic STAT-3 inhibitor on A) GIR, B) HGP and C) peripheral glucose uptake.

D) Effect of HN on glucose production from primary isolated hepatocytes treated with (+), or without (−) insulin. Values are means±SE. *Significantly different from other experimental groups, P<0.05.

Figure 7. Effect of a single IV dose of HNGF6A on blood glucose levels in chronically catheterized, unstressed Zucker diabetic fatty (ZDF) rats over 4 hours.

*Significantly different than saline controls, P<0.05.

Figure 8. A decline in both plasma and tissue HN levels with aging is observed in rodents and humans.

A) HN levels as assessed by ELISA, in young and old mice and B) across age in humans. C) Expression of RN in the hypothalamus and skeletal muscle of young and old rats. *Significantly different from other experimental groups, P<0.05.