Importance of Adipose Tissue NAD⁺ Biology in Regulating Metabolic Flexibility

Abstract

Nicotinamide adenine dinucleotide (NAD⁺) is an essential coenzyme that regulates cellular energy metabolism in many cell types. The major purpose of the present study was to test the hypothesis that NAD⁺ in white adipose tissue (WAT) is a regulator of whole-body metabolic flexibility in response to changes in insulin sensitivity and with respect to substrate availability and use during feeding and fasting conditions. To this end, we first evaluated the relationship between WAT NAD⁺ concentration and metabolic flexibility in mice and humans. We found that WAT NAD⁺ concentration was increased in mice after calorie restriction and exercise, 2 enhancers of metabolic flexibility. Bariatric surgery-induced 20% weight loss increased plasma adiponectin concentration, skeletal muscle insulin sensitivity, and WAT NAD⁺ concentration in people with obesity. We next analyzed adipocyte-specific nicotinamide phosphoribosyltransferase (Nampt) knockout (ANKO) mice, which have markedly decreased NAD⁺ concentrations in WAT. ANKO mice oxidized more glucose during the light period and after fasting than control mice. In contrast, the normal postprandial stimulation of glucose oxidation and suppression of fat oxidation were impaired in ANKO mice. Data obtained from RNA-sequencing of WAT suggest that loss of NAMPT increases inflammation, and impairs insulin sensitivity, glucose oxidation, lipolysis, branched-chain amino acid catabolism, and mitochondrial function in WAT, which are features of metabolic inflexibility. These results demonstrate a novel function of WAT NAMPT-mediated NAD⁺ biosynthesis in regulating whole-body metabolic flexibility, and provide new insights into the role of adipose tissue NAD⁺ biology in metabolic health.

NAD⁺, adipose tissue, obesity, metabolic flexibility, adiponectin, insulin sensitivity

Metabolic flexibility is a concept that describes the capability of an organism to adjust the oxidation of glucose and fatty acids in relation to glucose and fatty acid availability (1-5). It has been postulated that metabolic inflexibility is a common metabolic abnormality associated with obesity (4-6). Data obtained from studies conducted in rodents suggest that white adipose tissue (WAT) is involved in regulating metabolic flexibility by modulating whole-body substrate use and insulin sensitivity by modulating the availability of circulating free fatty acids (FFAs) and by producing adipokines that can increase (eg, adiponectin) or decrease (eg, interleukin-6, monocyte chemotactic protein-1) whole-body insulin sensitivity (1-5, 7). WAT stores energy as triglycerides (TGs) through lipogenesis in the fed state and hydrolyzes TGs that release FFA into the bloodstream in the fasted state (8, 9). Furthermore, the secretion of adiponectin, the major adipokine produced by WAT, and FFAs into the circulation affect both insulin sensitivity and metabolic flexibility (7, 10-12). However, the mechanism that links WAT and whole-body metabolic flexibility is not clear.

Nicotinamide adenine dinucleotide (NAD⁺) is an essential coenzyme that regulates redox metabolism in all species. In mammalian cells, the cytosolic NAD⁺/NADH ratio varies from 1 to 700 under normal physiological conditions (13, 14). Mitochondrial NAD⁺ content is higher than nuclear and cytoplasmic contents and accounts for 30% to 70% of the intracellular NAD⁺ pool (14, 15). Intracellular NAD⁺ biosynthetic enzymes play pivotal roles in maintaining NAD⁺ homeostasis because NAD⁺ cannot diffuse across membranes (16). In mammals, nicotinamide phosphoribosyltransferase (NAMPT) functions as a rate-limiting enzyme in the salvage NAD⁺ biosynthetic pathway (17). Data obtained from recent studies conducted in genetically engineered mouse models suggest NAMPT-mediated NAD⁺ biosynthesis regulates whole-body and tissue-specific metabolic function in response to changes in nutritional status (18-23). Diet-induced obesity and aging decrease NAMPT expression and NAD⁺ concentration in WAT (24-26). More recently, we and others have found that adipocyte-specific Nampt deletion causes metabolic abnormalities that are typically associated with obesity, including skeletal muscle and adipose tissue insulin resistance, hypoadiponectinemia, and blunted β ₃-adrenergic–mediated lipolytic activity (27-29). These findings suggest adipose tissue NAMPT-mediated NAD⁺ biosynthesis is involved in the pathophysiology associated with obesity. However, the role of adipose tissue NAD⁺ biology in whole-body metabolic flexibility is not clear.

The major purpose of the present study was to test the hypothesis that adipose tissue NAD⁺ is a physiological regulator of whole-body metabolic flexibility in the context of substrate use in response to altered nutritional status. To this end, we first evaluated the relationship between WAT NAD⁺ concentration and metabolic flexibility in mice and people. We next analyzed adipocyte-specific Nampt knockout (ANKO) mice, which have markedly decreased NAD⁺ concentrations in WAT (27, 30-32). We evaluated whole-body energy metabolism and substrate use in response to i) nocturnal and diurnal fasted and fed cycles, ii) prolonged fasting, and iii) change from fasting to insulin-stimulated postprandial conditions. We also conducted unbiased RNA-sequencing of WAT in an effort to identify the molecular events that link adipose tissue NAD⁺ biology with metabolic flexibility.

Materials and Methods

Animal Experimentation

ANKO mice were generated by using floxed-Nampt (flox/flox) mice (33) and adiponectin-Cre transgenic mice as described previously (27, 31, 32). C57BL/6J (B6) male mice (No. 380050) were purchased from the Jackson Laboratory. Mice were housed at normal room temperature (22 °C) with 12-hour light (6 am to 6 pm)/12-hour dark (6 pm to 6 am) cycles and were fed a standard diet (Nos. 5053; LabDiet) ad libitum with free access to water. Tissue samples were harvested, frozen in liquid nitrogen, and stored at –80 °C until analyses. All animal studies were approved by the Institutional Animal Care and Use Committee at Washington University.

Calorie Restriction and Exercise

For the calorie restriction (CR) study, B6 mice were maintained on 1-g pellets of a semisynthetic chow, AIN-93M (Bioserv). Ad libitum–fed control mice were given pellets equal to calculated average daily food intake (~4 pellets), and CR mice were given pellets equal to 60% of average daily food intake for 14 days as previously described (34). The exercise study was conducted in B6 mice at Washington University Mouse Cardiovascular Phenotyping Core. B6 mice were acclimated for 5 minutes to a 4-lane treadmill on a 0% incline that included an electric grid at the back of each treadmill, followed by a 10-minute warm-up period at 5 m/min. The treadmill speed was then increased by 5 m/min, reaching a maximum speed of 25 m/min for 30 minutes or until the mice reached exhaustion, after which tissues were collected. Exhaustion was defined as the inability of the mouse to resume running after 10 seconds on the electric grid.

Human studies

We analyzed WAT samples obtained from people with obesity and insulin resistance (age, 46 ± 6 years, body mass index: 51.9 ± 4.3) who participated in a previously published study that evaluated the metabolic effects of 20% weight loss induced by bariatric surgery (35, 36). All studies were conducted in the Clinical and Translational Research Unit at Washington University School of Medicine in St Louis, Missouri. Individuals completed a comprehensive medical screening before baseline testing; no participant had diabetes or were taking medications that affect insulin action. A hyperinsulinemic-euglycemic clamp procedure, in conjunction with stable isotopically labeled glucose tracer infusion, was performed to evaluate insulin sensitivity, as described previously (35-37). Skeletal muscle insulin sensitivity was assessed as the relative increase in glucose disposal rate during insulin infusion. Periumbilical abdominal subcutaneous WAT tissue biopsies were obtained during the basal stage of the clamp procedure as previously described (35-37). Tissue samples were rinsed with ice-cold saline and flash-frozen in liquid nitrogen until subsequent NAD⁺ measurement. Plasma adiponectin concentration was determined by using a commercially available enzyme-linked immunosorbent assay kit (No. DRP300; R&D Systems) (38). All participants provided written informed consent before their participation in the research protocols, which were approved by the institutional review board of Washington University School of Medicine, and registered in the ClinicalTrials.gov database (NCT00981500).

Nicotinamide Adenine Dinucleotide Measurement

NAD⁺ concentrations were determined in WAT samples by using a high-performance liquid chromatography system with a Supelco LC-18-T column (No. 58970-U; Sigma) as described previously (24, 31, 39). NAD⁺ concentrations were normalized to tissue weight.

Indirect Calorimetry (Comprehensive Lab Animal Monitoring System)

The rates of oxygen consumption (VO₂) and carbon dioxide production (VCO₂), food intake, and physical activity were determined by using a Comprehensive Lab Animal Monitoring System (CLAMS, Columbus Instruments). Mice were individually placed into the sealed metabolic chambers with free access to food and water. The study was carried out in an experimental room with 12- to 12-hour (18:00-6:00) dark-light cycles. Respiratory quotient (RQ) was calculated as the ratio of VCO₂ to VO₂ (VCO₂/VO₂). The rates of glucose oxidation and fat oxidation were calculated as 4.57 × VCO₂ (mL/kg-lean mass/hour) – 3.23 × VO₂ (mL/kg-lean mass/hour) and 1.69 × VO₂ (mL/kg-lean mass/hour) – 1.69 × VCO₂ (mL/kg-lean mass/hour), respectively (40, 41). Energy expenditure was calculated as 3.91 × VO₂ (mL/kg-lean mass/hour) + 1.10 × VCO₂ (mL/kg-lean mass/hour). The amount of food of each animal was monitored through a precision balance attached below the metabolic chamber. The motor activity was recorded every second in X and Z dimensions.

Tissue Triglyceride and Glycogen Measurements

Tissue TG and glycogen contents were determined by the Infinity Triglyceride Reagent (Thermo Scientific) and Glycogen Assay Kit (No. MAK016, Sigma), respectively. Data were normalized by tissue weight.

RNA Sequencing

Total RNA was isolated from WAT by using the RNeasy Mini Kit (No. 74104; Qiagen). Ribosomal RNA was removed by a hybridization method and messenger RNA was reverse-transcribed to yield complementary DNA, fragments of which were sequenced on a NovaSeq S4 (2 × 150-bp reads). Differentially expressed genes (DEGs) (false discovery rate [FDR] < 0.05) between flox/flox and ANKO mice were determined by using the edgeR R Bioconductor package, version 3.6.2. Enrichment analyses were performed on DEGs using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) Bioinformatics Resources 6.8 (http://david.abcc.ncifcrf.gov/) (42) with Gene ontology biological process (GO BP FAT) and Cellular Component (GO CC Direct), and KEGG terms as previously described (27). Pathways with an FDR of less than 0.05 were considered to be significantly enriched.

Western Blot Analysis

Proteins were extracted from frozen tissue samples and loaded onto polyacrylamide gels, separated by SDS-PAGE (sodium dodecyl sulfate–polyacrylamide gel electrophoresis), and transferred to polyvinylidene fluoride membranes as described previously (27, 31, 43). The blotted membranes were incubated with the following antibodies: total OXPHOS antibody cocktail (No. ab110413; Abcam) (44), mouse monoclonal anti-α-tubulin antibody (No. T5168, Sigma) (45), and horseradish peroxidase–linked antimouse antibody (No. sc-2005, Santa Cruz Biotechnology) (46). Blots were developed by using the ECL Western Blotting Detection Reagent (GE Healthcare Life Sciences). Western blot densitometry was quantitated using ImageJ software (National Institutes of Health ImageJ 1.47; http://imagej.nih.gov/ij) (47).

Quantification of Mitochondrial DNA Content

DNA was isolated by using QIAamp DNA Mini Kit (No. 51034, Qiagen). Mitochondrial DNA contents were determined by quantitating expression of 16S rRNA and Hk2 as described previously (27, 43).

Statistical Analysis

Differences between 2 groups were assessed by using t tests. Differences in key continuous metabolic parameters were evaluated by using repeated-measures analyses of variance. Data are presented as means ± SEM. A P value less than 0.05 was considered statistically significant.

Results

Adipose Tissue Nicotinamide Adenine Dinucleotide Biology Is Associated With Metabolic Flexibility in Mice and Humans

We found caloric restriction and endurance exercise, 2 enhancers of metabolic flexibility and insulin sensitivity (1, 2, 4), significantly increased WAT NAD⁺ concentrations by approximately 50% in B6 mice (Fig. 1A and 1B). Bariatric surgery–induced 20% weight loss in our participants markedly improved skeletal muscle insulin sensitivity, assessed as insulin-stimulated increase in glucose disposal rate, and increased plasma concentration of adiponectin, a biomarker and regulator of metabolic flexibility and insulin sensitivity (7, 10, 11) (Fig. 1C). These weight loss-induced metabolic benefits were accompanied by an increase in WAT NAD⁺ concentration (Fig. 1D).

Figure 1.

Adipose tissue nicotinamide adenine dinucleotide (NAD⁺) biology is associated with metabolic flexibility in mice and humans. A, NAD⁺ concentrations were determined in white adipose tissue (WAT) obtained from C57BL/6 mice after 14 days’ ad libitum (AL) and calorie-restriction (CR) C57BL/6 mice (n = 5 per group). B, NAD⁺ concentrations were determined in WAT obtained from control C57BL/6 mice rested in the cage and mice exercised on a motorized treadmill (EX) (n = 5 per group). Skeletal muscle insulin sensitivity (insulin-stimulated increase in glucose disposal rate [Rd]), C, plasma adiponectin concentration and D, WAT NAD⁺ concentration were determined in people with obesity and insulin resistance before and after bariatric surgery-induced 20% weight loss (n = 7). Values are means ± SEM. Data were analyzed by A and B, unpaired t test, and C and D, paired t test. *P less than .05.

Adipocyte-Specific Nampt Deletion Impairs Whole-Body Metabolic Flexibility

To determine the role of adipose tissue NAD⁺ biology in whole-body metabolic flexibility, we analyzed ANKO mice that have reduced NAD⁺ concentration in adipose tissue only (27, 31, 32). During normal ad libitum–feeding conditions, adipocyte-specific Nampt deletion markedly increased plasma insulin and FFA concentrations without changing blood glucose concentration during both light and dark phases (Fig. 2A), reflecting insulin resistance in ANKO mice (31). The RQ value, which indicates relative substrate use, was higher during the light period and had less variability in ANKO mice than in control (flox/flox) mice (Fig. 2B). Glucose oxidation rate was higher, whereas fat oxidation rate was lower during the light period in ANKO mice than flox/flox mice (Fig. 2C). There was no difference in energy expenditure, food intake, and physical activity between genotypes (data not shown). In addition, the RQ value and glucose oxidation rate were higher and fat oxidation rate was lower in ANKO mice after prolonged fasting than in flox/flox mice (see Fig. 2C). Consistent with our previous data (31), hepatic TG content was significantly higher in ANKO mice than flox/flox mice during fed conditions (Fig. 2D). However, during fasting conditions, ANKO mice had a decrease in hepatic TG content, compared with flox/flox mice (see Fig. 2D). In contrast, fasted ANKO mice stored more glycogen in the liver than flox/flox mice (Fig. 2E). Collectively, these results suggest adipocyte-specific Nampt deletion impairs adaptive fuel switching from glucose to lipids in response to reduced energy input.

Figure 2.

ANKO mice have metabolic inflexibility to lower nutritional input. A. Blood glucose and plasma concentrations of insulin and free fatty acids (FFA) were determined in 3- to 6-month-old flox/flox and adipocyte-specific nicotinamide phosphoribosyltransferase (Nampt) knockout (ANKO) mice (n = 4-5 per group). Blood was collected from mice during light (9 am) and dark (9 pm) phases. B, The respiratory quotient (RQ; carbon dioxide production [VCO₂]/oxygen consumption [VO₂], RQ) values, glucose oxidation rate, and fat oxidation rate were determined in 5- to 8-month-old flox/flox and ANKO mice by indirect metabolic calorimetry (CLAMS) (n = 7-8 per group). Repeated-measures analysis of variance was used to calculate P value for group × time interaction (Px) during this light period. C, The average of RQ values and glucose and fat oxidation rates after a 6-hour fasting period. C, Liver triglyceride and E, glycogen contents, were determined in 4- to 6-month-old flox/flox and ANKO mice after overnight fasting (n = 5 per group). Values are means ± SEM. Data were analyzed by unpaired t test. *P less than .05; ^**P less than .01; ^***P less than .001.

We next investigated postprandial metabolic flexibility by evaluating the metabolic response to refeeding a normal diet after a 24-hour fast. Adipocyte-specific Nampt deletion increased plasma concentrations of insulin, FFAs, and TGs without changing blood glucose concentration after refeeding (Fig. 3A). During the postprandial period, RQ values and glucose oxidation rate were lower, whereas the fat oxidation rate was much higher in ANKO mice than flox/flox mice (Fig. 3B). ANKO mice had impaired increases in RQ (ΔRQ) and glucose oxidation (Δglucose oxidation rate), and lower suppression of fat oxidation (Δfat oxidation rate) during the postprandial period compared with flox/flox mice (Fig. 3C). These findings demonstrate the ability to alter substrate oxidation during insulin-stimulated postprandial conditions is impaired in ANKO mice.

Figure 3.

Adipocyte-specific nicotinamide phosphoribosyltransferase (Nampt) knockout (ANKO) mice have metabolic inflexibility during the postprandial period. A, Blood glucose and plasma concentrations of insulin, free fatty acids (FFA), and triglyceride (TG) were determined in 2- to 6-month-old flox/flox and ANKO mice (n = 9 per group). Blood was collected from mice that were fasted for 24 hours and then refed for 6 hours. B, Respiratory quotient (RQ) and glucose and fat oxidation rates were determined in 5- to 8-month-old flox/flox and ANKO mice (n = 7-8 per group). Mice received a normal standard diet after a 24-hour fasting period. Repeated-measures analysis of variance was used to calculate P value for group × time interaction (Px) during this light period. C, Average ΔRQ values, Δglucose oxidation rate, and Δfat oxidation rate are shown. Values are means ± SEM. Data were analyzed by unpaired t test. *P less than .05; ^**P less than .01; ^***P less than .001.

Loss of Nicotinamide Phosphoribosyltransferase Impairs Metabolic Pathways Involved in Inflammatory and Oxidative Stress Responses, Insulin Sensitivity, Substrate Catabolism, and Mitochondrial Function in White Adipose Tissue

We conducted RNA-sequencing of WAT and determined transcriptional changes induced by Nampt deletion. There were 443 DEGs between flox/flox and ANKO mice (Fig. 4A). Upregulated DEGs involved markers of inflammation and oxidative stress, such as Gdf15, Hmox1, Cd68 (Fig. 4B), and DEGs were significantly enriched with the biological pathways involved in inflammatory and oxidative stress responses (eg, acute inflammatory response, response to oxidative stress, reactive oxygen species metabolic process) (Table 1). Consistent with our previous findings (31), downregulated DEGs involved specific targets of peroxisome proliferator-activated receptor γ (PPARG) serine-273 (S273) phosphorylation (48, 49), such as Adiponectin and Adipsin (Fig. 4B), and DEGs were enriched with “adiponectin-activated signaling pathway” and “PPAR signaling pathway” (see Table 1). In addition, DEGs were enriched with many biological pathways involved in regulating insulin sensitivity and substrate catabolism (see Table 1 and Fig. 4C), such as pathways involved in glucose uptake and oxidation (eg, positive regulation of glucose import, pyruvate metabolic process) (see Fig. 4C). In addition, loss of NAMPT reduced WAT gene expression of key regulators of lipolysis, such as Hsl and Adrb3, and DEGs were enriched in lipolytic pathways (eg, lipid oxidation, regulation of lipolysis in adipocyte, lipid particle) (see Fig. 4B and 4C). These results are consistent with data obtained from our recent studies that found ANKO mice have insulin resistance and blunted adrenergic lipolysis in WAT (27, 31). Additionally, DEGs were highly enriched with branched-chain amino acid (BCAA) degradation pathways (eg, leucine metabolic process, valine, leucine, and isoleucine degradation) and loss of NAMPT markedly downregulated key mitochondrial BCAA degradation enzymes implicated in obesity and insulin resistance, including Bcat2 and Hibch (50, 51) (see Fig. 4B and 4C). Finally, ANKO mice had decreases in gene expression of proteins involved in mitochondrial oxidative metabolism, such as Nduf3 and Cox8b (Fig. 4B), and DEGs were enriched with “mitochondrial matrix” and “oxidation-reduction process” (see Fig. 4C). Taken together, these findings suggest loss of NAMPT causes inflammation, oxidative stress, insulin resistance, impaired glucose oxidation, defective lipolysis, and BCAA catabolism, and mitochondrial dysfunction in WAT, which are hallmarks of obesity and metabolic inflexibility (3, 11, 51). Evidence of mitochondrial dysfunction in WAT of ANKO mice was further supported by decreases in protein expression of UQCRC2, a key component of electron transport chain complex III and mitochondrial DNA content (Fig. 4D and 4E).

Table 1.

List of white adipose tissue biological pathways enriched with differentially expressed genes between flox/flox and ANKO mice

Pathway	Fold enrichment	FDR
Gene Ontology
Adiponectin-activated signaling pathway	31.01	7.80E-03
Leucine metabolic process	26.58	1.11E-02
Neurotransmitter catabolic process	16.92	3.19E-02
Response to folic acid	14.31	4.46E-02
Brown fat cell differentiation	11.02	1.29E-04
L-amino acid transport	10.18	2.84E-03
Regulation of glycogen biosynthetic process	8.62	4.55E-02
Regulation of systemic arterial blood pressure mediated by a chemical signal	6.89	6.72E-03
Lipid oxidation	6.84	6.38E-06
Regulation of lipid catabolic process	6.53	8.28E-03
Positive regulation of glucose import	6.34	4.49E-02
Fat cell differentiation	5.95	2.49E-10
Organonitrogen compound catabolic process	5.09	1.27E-09
Pyruvate metabolic process	5.00	7.36E-03
Iron ion homeostasis	4.38	4.38E-02
Positive regulation of angiogenesis	4.14	7.01E-03
Negative regulation of developmental growth	4.03	3.73E-02
Acute inflammatory response	3.94	1.58E-02
Regulation of smooth muscle cell proliferation	3.90	9.58E-03
Muscle cell proliferation	3.88	1.28E-03
Cellular response to hypoxia	3.88	2.74E-02
Sulfur compound metabolic process	3.65	5.24E-05
Response to toxic substance	3.51	2.86E-02
Cellular modified amino acid metabolic process	3.47	4.49E-02
Oxidation-reduction process	3.45	2.62E-16
Negative regulation of secretion by cell	3.28	8.36E-03
Negative regulation of protein kinase activity	3.26	8.53E-03
Reactive oxygen species metabolic process	3.19	7.01E-03
Response to extracellular stimulus	3.02	3.59E-05
Lipid particle	5.81	1.53E-02
Mitochondrial matrix	3.32	1.87E-02
KEGG
Propanoate metabolism	9.73	2.37E-03
Valine, leucine, and isoleucine degradation	8.87	1.39E-06
β-Alanine metabolism	6.82	3.40E-02
Tyrosine metabolism	6.73	1.67E-02
Regulation of lipolysis in adipocytes	5.27	2.01E-02
Glycolysis/Gluconeogenesis	4.55	3.40E-02
Carbon metabolism	4.53	6.70E-04
PPAR signaling pathway	4.22	2.96E-02
Biosynthesis of antibiotics	3.33	6.70E-04

The DAVID Bioinformatics Resources identified significantly (FDR < 0.05) enriched pathways of DEGs (flox/flox vs after ANKO mice, n = 3) with a fold enrichment greater than 3.

Abbreviations: ANKO, adipocyte-specific nicotinamide phosphoribosyltransferase (Nampt) knockout; FDR, false discovery rate; DAVID, Database for Annotation, Visualization, and Integrated Discovery; DEG, differentially expressed gene; PPAR, peroxisome proliferator-activated receptor.

Figure 4.

Loss of nicotinamide phosphoribosyltransferase (NAMPT) impairs metabolic pathways involved in inflammatory and oxidative stress responses, insulin sensitivity, substrate catabolism, and mitochondrial function in white adipose tissue (WAT). RNA-sequencing (RNA-seq) was conducted to determine the effects of Nampt deletion on the WAT transcriptome (n = 3 per group). A, A volcano plot of RNA-seq data of WAT with log₂-fold change (FC) (x axis) and -log₁₀-false discovery rate (FDR) value (y axis). Upregulated and downregulated differentially expressed genes (DEGs; FDR < 0.05, log₂- FC > 0.5) between flox/flox and adipocyte-specific nicotinamide phosphoribosyltransferase (Nampt) knockout (ANKO) mice are shown as red and blue dots, respectively. B, Heat map of representative DEGs involved in inflammation, oxidative stress, peroxisome proliferator-activated receptor γ (PPARG; serine 273), insulin sensitivity, substrate catabolism, and mitochondrial function. C, The Database for Annotation, Visualization, and Integrated Discovery (DAVID) Bioinformatics Resource identified Gene Ontology Biological Process (GO BP FAT) and Cellular Component (GO CC), and KEGG pathways significantly (FDR < 0.05) enriched with DEGs. Key metabolic pathways involved in substrate catabolism and mitochondrial function are shown. D, Western blot analysis of UQCRC2, a key component of electron transport chain (ETC) complex-III in WAT (n = 4-5 per group). Densitometric analysis of UQCRC2 normalized to TUBULIN is shown. E, WAT mitochondrial DNA (mtDNA) contents normalized to nuclear DNA (nucDNA) contents (n = 3-4 per group). Values are means ± SEM. Data were analyzed by unpaired t test. ^**P less than .01; ^***P less than .001.

Discussion

In the present study, we evaluated the importance of adipose tissue NAD⁺ biology in regulating whole-body metabolic flexibility by evaluating WAT NAD⁺ concentrations in insulin-sensitive and insulin-resistant mice and people and by evaluating the effect of manipulating WAT NAD⁺ metabolism on whole-body energy metabolism and substrate use in response to altered nutritional status in mice. We found 1) interventions that enhance insulin sensitivity increased WAT NAD⁺ concentrations both in mice and people; 2) ANKO mice, which have an impairment in NAD⁺ biosynthesis in WAT, oxidized more glucose during the light period and after prolonged fasting than flox/flox mice; 3) the normal postprandial stimulation of glucose oxidation and suppression of fat oxidation were impaired in ANKO mice; and 4) loss of NAMPT increased gene expression of markers of inflammation and oxidative stress, whereas it markedly decreased gene expression of proteins involved in regulating insulin sensitivity, glucose oxidation, lipolysis, BCAA catabolism, and mitochondrial function in WAT, which is consistent with the alterations associated with obesity and metabolic inflexibility in humans and rodents (2, 3, 11, 51, 52). Taken together, these findings suggest that adipose tissue NAMPT-mediated NAD⁺ biosynthesis is an important regulator of whole-body metabolic flexibility.

Data obtained from our previous studies conducted in ANKO mice showed that adipocyte-specific Nampt knockout impairs catecholamine-mediated stimulation of lipolysis and insulin-mediated suppression of lipolysis (27, 31). Consistent with these in vivo findings, we found loss of NAMPT reduces gene expression of proteins involved in insulin signaling and lipolysis in WAT. Furthermore, ANKO mice have a large reduction in plasma concentration of adiponectin, which is an insulin-sensitizing adipokine that enhances metabolic flexibility (10). These results suggest that blunted lipolytic responses of WAT to fasting and insulin and hypoadiponectinemia could limit the adaptive capability of mice to alter substrate oxidation in response to the alterations in nutritional status, leading to whole-body metabolic inflexibility in ANKO mice. The cellular mechanism responsible for these metabolic derangements is unclear but could involve deactivation of 2 interrelated metabolic regulators, namely NAD⁺-dependent protein deacetylase SIRT1 and caveolin-1 (CAV1). We recently found loss of NAMPT reduces expression of CAV1 and its downstream targets, such as ADRB3 and IRS1, through inactivation of SIRT1 and impairs adrenergic lipolytic activity in a CAV1-dependent manner (27). A series of elegant studies by Asterholm and colleagues found that Cav1-deficient mice have metabolic inflexibility and adipose tissue dysfunction that is similar to ANKO mice, including insulin resistance, hypoadiponectinemia, impaired adrenergic lipolytic and thermogenic responses, adipose tissue inflammation, and mitochondrial dysfunction (11, 27, 28, 31, 53, 54). In addition, data obtained from previous studies conducted in rodents and people found obesity is associated with decreases in WAT expression and activity of NAMPT and SIRT1 (24-26, 55). Collectively, these findings suggest that impairment of the adipose NAMPT-NAD⁺-SIRT1-CAV1 axis could be involved in the development of obesity-induced metabolic inflexibility.

Mammals, including mice and humans, possess 2 distinct forms of NAMPT, intracellular NAMPT (iNAMPT) and extracellular NAMPT (eNAMPT) (18, 56). In the present study, we identified adipose tissue iNAMPT as a novel regulator of whole-body metabolic flexibility. However, the relationship between adipose tissue-derived eNAMPT and metabolic flexibility remains unclear. Data obtained from recent studies conducted in ANKO mice and adipocyte-specific Nampt knockin (ANKI) mice have shown that adipose tissue is a major source of circulating eNAMPT, and adipose tissue-derived eNAMPT regulates hypothalamic NAD⁺ biosynthesis and function (32, 57). Given the importance of hypothalamic neurons in regulating WAT lipolysis and whole-body metabolic flexibility with respect to substrate use in response to dietary changes (58-60), reduced plasma eNAMPT availability could be involved in the mechanism responsible for metabolic inflexibility in ANKO mice, independent of defective adipose tissue NAD⁺ biosynthesis.

In conclusion, we demonstrate a new role of adipose tissue NAMPT-mediated NAD⁺ biosynthesis in regulating metabolic flexibility in response to altered nutritional status. Our findings are consistent with data obtained from previous studies conducted in rodents that found increasing NAD⁺ content by administering NAD⁺ intermediates, such as nicotinamide riboside and nicotinamide mononucleotide, or inhibiting a major NAD⁺-degrading enzyme CD38, improves glucose metabolism and markers of metabolic flexibility (18-23, 61-63). Additional studies are needed to determine whether enhancing NAD⁺ metabolism leads to improvements of adipose tissue function and whole-body metabolic flexibility in people.

Abbreviations

Abbreviations
 
• ANKO

  adipocyte-specific nicotinamide phosphoribosyltransferase (Nampt) knockout

 
• BCAA

  branched-chain amino acid

 
• CAV1

  caveolin-1

 
• CLAMS

  Comprehensive Lab Animal Monitoring System

 
• CR

  calorie restriction

 
• DEG

  differentially expressed gene

 
• eNAMPT

  extracellular NAMPT

 
• FDR

  false discovery rate

 
• FFAs

  free fatty acids

 
• NAD⁺

  nicotinamide adenine dinucleotide

 
• PPAR

  peroxisome proliferator-activated receptor

 
• PPARG

  proliferator-activated receptor γ

 
• RQ

  respiratory quotient

 
• TGs

  triglycerides

 
• VCO₂

  carbon dioxide production

 
• VO₂

  oxygen consumption

 
• WAT

  white adipose tissue

Acknowledgments

The authors thank Dr Shin-ichiro Imai (Washington University) for providing WAT samples obtained in his CR studies. We also thank Dr Sangeeta Adak (Washington University Diabetes Research Center [DRC]) and Melanie Schmitt and Jiane Feng (University of Michigan Mouse Metabolic Phenotyping Center [MMPC]) for technical assistance. J.Y. is a guest associate professor at Keio University Graduates School of Medicine and is supported by the Top Global University Project by MEXT.

Financial Support: This work was supported by the National Institutes of Health (grant Nos. DK104995; DK115764; DK117657; DK056341 to the Washington University Nutrition Obesity Research Center [NORC]; DK020579 to the Washington University DRC; DK052574 to the Washington University Digestive Diseases Research Core Center; CA91842 and UL1TR002345 to the Washington University Genome Technology Access Center; DK089503 to the University of Michigan NORC; DK020572 to the University of Michigan DRC; and U2CDK110678-01 to the University of Michigan MMPC) and the Longer Life Foundation.

Author Contributions: J.Y. conceptualized and designed the project, supervised all the experiments, and provided funding. M.P.F., N.Q., K.L.S., C.L, S.Y., Y.S., R.T.B., and J.Y. performed the experiments and analyzed the data. M.Y. and S.K. provided human adipose tissue biopsy samples and data related to human participants. H.I., B.N.F., B.J.D., and S.K. provided intellectual input for the study design and data interpretation and contributed to discussion. All authors reviewed and edited the manuscript. J.Y. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Additional Information

Disclosure Summary: J.Y. is an inventor of a patent application related to NAD and adiponectin (#20180228824).

Data Availability

All RNA-sequencing data used in this study have been deposited into the NCBI GEO database under accession number GSE159397.

Author notes

These authors contributed equally to this work.