Nicotinamide N-methyltransferase knockdown protects against diet-induced obesity

Abstract

In obesity and type 2 diabetes, Glut4 glucose transporter expression is decreased selectively in adipocytes. Adipose-specific knockout or overexpression of Glut4 alters systemic insulin sensitivity. Here we show, using DNA array analyses, that nicotinamide N-methyltransferase (Nnmt) is the most strongly reciprocally regulated gene when comparing gene expression in white adipose tissue (WAT) from adipose-specific Glut4-knockout or adipose-specific Glut4-overexpressing mice with their respective controls. NNMT methylates nicotinamide (vitamin B3) using S-adenosylmethionine (SAM) as a methyl donor. Nicotinamide is a precursor of NAD(+), an important cofactor linking cellular redox states with energy metabolism. SAM provides propylamine for polyamine biosynthesis and donates a methyl group for histone methylation. Polyamine flux including synthesis, catabolism and excretion, is controlled by the rate-limiting enzymes ornithine decarboxylase (ODC) and spermidine-spermine N(1)-acetyltransferase (SSAT; encoded by Sat1) and by polyamine oxidase (PAO), and has a major role in energy metabolism. We report that NNMT expression is increased in WAT and liver of obese and diabetic mice. Nnmt knockdown in WAT and liver protects against diet-induced obesity by augmenting cellular energy expenditure. NNMT inhibition increases adipose SAM and NAD(+) levels and upregulates ODC and SSAT activity as well as expression, owing to the effects of NNMT on histone H3 lysine 4 methylation in adipose tissue. Direct evidence for increased polyamine flux resulting from NNMT inhibition includes elevated urinary excretion and adipocyte secretion of diacetylspermine, a product of polyamine metabolism. NNMT inhibition in adipocytes increases oxygen consumption in an ODC-, SSAT- and PAO-dependent manner. Thus, NNMT is a novel regulator of histone methylation, polyamine flux and NAD(+)-dependent SIRT1 signalling, and is a unique and attractive target for treating obesity and type 2 diabetes.

Figure 1. NNMT expression is increased in WAT and liver of obese and insulin-resistant mice

a, b, Nnmt mRNA expression normalized by cyclophilin in WAT of adipose-specific Glut4 knockout (AG4KO) mice and aP2-Cre controls (n =4 per group) (a) and adipose-specific Glut4 overexpressing (AG4OX) and wild-type littermate controls (n =6 per group) (b). c–e, NNMT protein levels in WAT of ob/ob mice (n =8) and lean controls (n =4) (c); db/db mice and lean controls (n =7 per group) (d), and high-fat diet (HFD)-fed (n =6) and chow-fed mice (n =7) (e). f–h, NNMT protein levels in liver of ob/ob mice (n =9) and lean controls (n =6) (f); db/db mice and lean controls (n =7 per group) (g); and HFD-fed and chow-fed mice (n =6 per group) (h). Actin was used as a control for western blot analysis and the levels were not different between lean and obese mice. AU, arbitrary units. Error bars, ±s.e.m; *P <0.05.

Figure 2. Nnmt knockdown prevents diet-induced obesity and insulin resistance

a–f, Knockdown efficiency of Nnmt-ASO. mRNA expression was normalized by cyclophilin and protein levels were corrected with actin levels: Nnmt mRNA (a) and NNMT protein in WAT (b); Nnmt mRNA (c) and NNMT protein in liver (d); NNMT protein in brown adipose tissue (BAT) (e) and kidney (f). g, Body weights of C57BL/6 mice fed a high-fat diet and treated with Nnmt ASO, control ASO, or vehicle (saline) for 8 weeks. h, Fat mass as a percentage of body weight. i, Lean mass as a percentage of body weight. j, Subcutaneous WAT (SQWAT) fat-pad weights. k, Epididymal WAT (EWAT) fat-pad weights. l, Epididymal adipocyte cross-sectional area and haematoxylin and eosin (H&E)-stained sections of SQWAT. m, Serum insulin levels. n, Glucose × insulin product (ng ml⁻¹ × mg dl⁻¹) in the fed state. o, Intraperitoneal glucose tolerance test. p, Area under the curve (AUC) of the glucose tolerance. q, H&E stain of liver sections of HFD-fed Nnmt-ASO-and control-ASO-treated mice, and of chow-fed mice. r, Hepatic triglyceride levels in Nnmt- and control-ASO-treated mice. The scale bars in l and q represent 100 μm; n =8 per group for a–p, n =13 per group for r. AU, arbitrary units. Error bars, ±s.e.m.; *P <0.05.

Figure 3. NNMT regulates energy expenditure

a, Energy intake of HFD-fed control-ASO- and Nnmt-ASO-treated mice. b, Feed efficiency (body-weight gain per kilocalorie (kcal) consumed) of HFD-fed control-ASO- and Nnmt–ASO-treated mice. c, Fat mass gain per kcal consumed in HFD-fed mice treated with Nnmt ASO or control ASO. n =12 per group for a–c. d, Faecal lipid excretion (n =10 per group). e, Oxygen consumption (Vol. O₂) measured by CLAMS (n =7 per group). f–h, Effects of a 16-h fast on body weight (f), fat mass (g), and lean body mass (n =8 per group) (h). i–k, Oxygen consumption in 3T3-L1 adipocytes transfected with control or Nnmt ASO (n =6 per group) (i), treated with 10 mM N¹-methylnicotinamide (me-Nam) (n =6 per group) (j), or transfected with Nnmt cDNA (n =10 per group) (k). Error bars, ±s.e.m.; *P <0.05.

Figure 4. NNMT regulates SAM and NAD⁺ pathways in adipose tissue

a, Adipose S-adenosylmethionine (SAM), S-adenosylhomocysteine (SAH) and SAM:SAH ratio measured by LC–MS/MS (n =8 for control-ASO-treated mice; n =12 for Nnmt-ASO-treated mice). b, c, ODC (b) and SSAT (c) activity (n =10 per group). d, Amd1, Odc and Ssat mRNA expression in adipose tissue of Nnmt-ASO- and control-ASO-treated mice (n =12 per group). e, Urinary diacetylspermine:creatinine ratio (n =22 for control-ASO-treated mice; n =29 for Nnmt-ASO-treated mice). f, g, Odc and Ssat mRNA levels in 3T3-L1 adipocytes with Nnmt knockdown (n =9 per group) (f) and N¹-methylnicotinamide (me-Nam) (g) treatment (n =6 per group). h, Diacetylspermine secretion from 3T3-L1 adipocytes treated with me-Nam (n =10 per group). i, Expression of mono-, di- and tri-methylation of lysine histone 3 (H3K4) normalized to total H3 levels in adipose tissue (n =8 per group). j, H3K4me2 occupancy on Odc, Ssat genes and an open reading frame free region (Igx1a) as a negative control in adipocytes measured by ChIP-qPCR (n =12 per group). k–m, Oxygen consumption in adipocytes transfected with control or Ssat siRNA, and treated with or without 10 mM me-Nam (n =6 per group) (k); treated with 10 mM me-Nam with or without 5 mM difluoromethylornithine (DFMO), a specific ODC inhibitor (n =6 per group) (l); or treated with 10 mM me-Nam with or without MDL72527, a specific PAO inhibitor (n =6 per group) (m). n, NAD⁺ levels (n =7 Con-ASO; n =12 Nnmt ASO). o, mRNA levels of nicotinamide phosphoribosyltransferase (Nampt) (n =11 per group). p, mRNA levels of SIRT1 target genes: Cd36, catalase (Cat), succinate dehydrogenase B (SdhB) and growth arrest and DNA-damage-inducible protein (Gadd45a) in adipose tissue of NNMT-ASO- and control-ASO treated mice (n =11 per group). Error bars, ± s.e.m., *P <0.05, †P =0.06. q, Model of NNMT-regulated energy expenditure in adipocytes. NNMT methylates nicotinamide (Nam), a precursor of NAD⁺, using SAM as a methyl donor. SAM regulates polyamine flux by providing substrates and modulating histone methylation. Polyamine flux utilizes acetyl-CoA to generate acetyl-polyamines, which are oxidized or excreted in the urine. Taken together, this results in adipose metabolic substrate consumption and loss coupled with systemic alteration of energy expenditure.