Effects of Nicotinamide Riboside on Endocrine Pancreatic Function and Incretin Hormones in Nondiabetic Men With Obesity

Abstract

Objective

Augmenting nicotinamide adenine dinucleotide (NAD⁺) metabolism through dietary provision of NAD⁺ precursor vitamins translates to improved glucose handling in rodent models of obesity and diabetes. Preclinical evidence suggests that the NAD⁺/SIRT1 axis may be implicated in modulating important gut-related aspects of glucose regulation. We sought to test whether NAD⁺ precursor supplementation with nicotinamide riboside (NR) affects β-cell function, α-cell function, and incretin hormone secretion as well as circulating bile acid levels in humans.

Design

A 12-week randomized, double-blind, placebo-controlled, parallel-group trial in 40 males with obesity and insulin resistance allocated to NR at 1000 mg twice daily (n = 20) or placebo (n = 20). Two-hour 75-g oral glucose tolerance tests were performed before and after the intervention, and plasma concentrations of glucose, insulin, C-peptide, glucagon, glucagon-like peptide 1 (GLP-1), and glucose-dependent insulinotropic polypeptide (GIP) were determined. β-Cell function indices were calculated based on glucose, insulin, and C-peptide measurements. Fasting plasma concentrations of bile acids were determined.

Results

NR supplementation during 12 weeks did not affect fasting or postglucose challenge concentrations of glucose, insulin, C-peptide, glucagon, GLP-1, or GIP, and β-cell function did not respond to the intervention. Additionally, no changes in circulating adipsin or bile acids were observed following NR supplementation.

Conclusion

The current study does not provide evidence to support that dietary supplementation with the NAD⁺ precursor NR serves to impact glucose tolerance, β-cell secretory capacity, α-cell function, and incretin hormone secretion in nondiabetic males with obesity. Moreover, bile acid levels in plasma did not change in response to NR supplementation.

In the setting of obesity and insulin resistance, functional perturbation of enteroendocrine cells can disturb glucose homeostasis and may lead to overt type 2 diabetes (T2D). This involves insulin producing β-cells, glucagon-producing α-cells, and incretin hormone–producing L- and K-cells, and therefore these cells represent interesting targets to ameliorate obesity-related metabolic decline.

T2D arises as a product of increased insulin resistance and decreased β-cell function (1). In individuals with obesity and insulin resistance, the β-cell mass may initially expand to adjust insulin secretion to match increased insulin demand necessary to maintain normoglycemia (2). However, continuous stress on the β-cells may lead to progressive β-cell failure, resulting in a diminished capacity for insulin production and eventually hyperglycemia (3). Thus, preserving β-cell function in obesity is key to avoid development or progression of T2D.

Postprandial insulin responses are to a major extent mediated by the incretin hormones glucagon-like peptide 1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP). These hormones are produced by enteroendocrine L- and K-cells in the gut and act in concert to potentiate glucose-stimulated insulin secretion (GSIS) following orally delivered glucose (4). This action, termed the incretin effect, is diminished in obesity (5) and in patients with T2D (5, 6). Also, reduced incretin hormone secretion following oral glucose challenges has been reported in subjects with obesity and in individuals with so-called prediabetes (7, 8). Thus, body mass index (BMI) has been found to correlate inversely with postprandial GLP-1 responses (7, 9, 10). Hence, a blunted incretin hormone response in obesity might lead to weakened insulin release and contribute to glucose dysregulation.

Glucagon secretion is often disturbed in obesity and may become further aggravated during the transition from prediabetes to T2D (11). This is reflected in increasing fasting glucagon concentrations and weakened postprandial glucagon suppression, contributing to blood glucose elevations due to unrestrained glucagon-induced hepatic glucose production. Taken together, these important aspects of glucose regulation represent prime targets for antidiabetic therapy.

Recently, nicotinamide adenine dinucleotide (NAD⁺) “boosting” compounds have been proposed as a novel therapeutic route in preclinical studies (12, 13). Nicotinamide riboside (NR), an NAD⁺ precursor and member of the vitamin B₃ family, improves glucose tolerance and mitigates dysmetabolic traits in rodent models of obesity and diabetes (14–17). Mechanistically, NR-derived effects are mainly attributed to increased NAD⁺ availability, which augments activity of NAD⁺-dependent enzymes, including members of the sirtuin (SIRT) enzyme family (SIRT1–7) (18). SIRT1 is a nuclear-localized transcriptional regulator that uses NAD⁺ to deacetylate histones and various transcription factors, thereby modifying cellular transcriptional events and facilitating adaptive responses to nutritional and environmental cues (19). In this way, SIRT1 is involved in the regulation of multiple cellular processes that are dysregulated in T2D, including glucose and lipid metabolism, mitochondrial homeostasis, and β-cell function (20). Several lines of evidence interconnect SIRT1 activity, NAD⁺ metabolism, and β-cell function. SIRT1 is highly expressed in β-cells and is thought to positively promote insulin secretion (21). In SIRT1 knockout mouse models, GSIS is impaired, suggesting that SIRT1 might be necessary for normal β-cell function (22–24). Conversely, SIRT1 overexpression restored normal GSIS in an in vitro β-cell toxicity model in isolated rat islets (25). Moreover, in a mouse model overexpressing SIRT1 specifically in pancreatic β-cells, GSIS was improved and the animals showed improved glucose tolerance (26).

Compromised NAD⁺ biosynthesis has also been implicated in impaired β-cell function resulting in reduced GSIS in mice, correctable by providing nicotinamide mononucleotide (NMN) (27). NMN is another NAD⁺ precursor variant that is suggested to be converted to NR extracellularly before contributing to intracellular NAD⁺ biosynthesis (28). Also in cytokine-induced β-cell impairment (29), in diet-induced diabetes in mice (30), as well as in human islet cells (31), NMN has been demonstrated to promote GSIS, whereas no studies are currently available on the effect of NR in this respect. Interestingly, studies in mice have demonstrated a role for the adipokine adipsin in maintaining β-cell function, and patients with T2D characterized by β-cell failure exhibit reduced levels of circulating adipsin (32). Serum adipsin levels were also severely diminished in transgenic mice lacking the key NAD⁺ biosynthetic enzyme nicotinamide phosphoribosyltransferase in adipocytes, but this could be restored by supplementing NMN (33). Taken together, the above supports the hypothesis that NAD⁺ precursor supplementation impacts β-cell function.

As previously mentioned, incretin hormones and glucagon are central in glucose regulation. However, there is a scarcity of studies investigating whether interventions aimed at affecting the NAD⁺/SIRT1 axis affect glucose metabolism though alterations of these hormones. Resveratrol, a proposed SIRT1 activator (34), was reported to increase GLP-1 secretion in diabetic mice in parallel with enhanced insulin secretion (35), but this finding has not been confirmed in humans (36, 37). However, resveratrol treatment did result in enhanced suppression of glucagon following a meal test in nondiabetic males with obesity (36).

Other gut-related factors may be implicated in metabolic control. Bile acids (BAs) have emerged as metabolic regulators in addition to their well-described roles in lipid emulsification and absorption (38). By triggering specific receptors, such as the farnesoid X receptor and TGR5, BAs may exert endocrine effects impacting lipid and glucose metabolism, and these receptors have been suggested as targets for antidiabetic treatment (39). Studies in SIRT1 knockout mice have established a role for SIRT1 in BA metabolism and transport (40–42). Furthermore, farnesoid X receptor has been identified as a specific SIRT1 target (43). It could be speculated that NAD⁺ precursor supplementation may induce changes in circulating BAs via SIRT1-mediated mechanisms.

As outlined above, multiple gut-related components, intrinsically involved in metabolic regulation, are potentially modulated by activity in the NAD⁺/SIRT1 axis, but whether supplementation with NR affects these gut-related factors in humans remains to be investigated. Using samples obtained during a recently published randomized, placebo-controlled clinical trial in men with obesity and insulin resistance (44), we explored whether NR affects β-cell function, α-cell function, incretin hormone response, and circulating BAs.

Materials and Methods

This investigator-initiated study was part of a larger randomized, double-blinded, placebo-controlled, parallel-grouped clinical trial to test the safety and effects of 12 weeks of NR supplementation on insulin sensitivity, substrate metabolism, and body composition in nondiabetic males with obesity, as previously described in detail (44). The current study focuses on NR-derived effects on the enteroendocrine system and presents previously unpublished data obtained during oral glucose tolerance tests (OGTTs).

The study protocol was approved by the regional Research Ethics Committee (H-3-2014-130) and the Danish Data Protection Agency (1-16-02-714-14), and it was performed in compliance with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to inclusion. The trial was preregistered at clinicaltrials.gov (NCT02303483). Recruitment, data collection, and analysis were carried out at the Department of Clinical Medicine, Aarhus University Hospital (Aarhus, Denmark) in the period 2016 to 2018.

Study procedure and subjects

Forty male subjects were recruited who were obese (BMI >30 kg/m²), sedentary (<30 minutes of exercise per day), and middle-aged (40 to 70 years of age). Subjects were otherwise healthy nonsmokers and did not receive any prescription medication. A detailed clinical examination of all candidates, including a full medical history, blood testing, and electrocardiography, was conducted by a physician to evaluate admissibility to the study.

Subjects were randomly allocated into two groups to receive orally administered NR (Niagen™, ChromaDex, Irvine, CA,) at 1000 mg × 2 (n = 20) or placebo (n = 20) for 12 weeks and were examined on 3 separate days before and after the intervention. On day 1, a 2-hour OGTT was performed independently of the additional examinations. On days 2 and 3, whole-body dual-energy x-ray absorptiometry scans, MRI, and hyperinsulinemic euglycemic clamps were conducted, as previously reported (44). Ingestion of trial medication was commenced following baseline investigations and continued until the final examination, including administration on the morning of the examination days. During the intervention, subjects were instructed not to change dietary habits or level of physical activity, and to abstain from all dietary supplements and vitamins.

Aarhus University Hospital Pharmacy was responsible for all aspects of randomization and blinding, as well as packaging and labeling of the trial medication, as previously described (44).

OGTTs

OGTTs were performed before and after the intervention. Subjects reported to the Department of Clinical Medicine at approximately 8:00 am after an overnight fast and were studied in the supine position for 2 hours. A peripheral venous catheter was inserted in a dorsal hand vein, and the hand was placed in a heated pillow to enable sampling of arterialized blood. Baseline blood samples were collected (time 0 minutes) immediately, followed by oral ingestion of 75 g of glucose dissolved in water during a short period of time (maximally 5 minutes). Further blood samples were drawn at 15, 30, 60, and 120 minutes.

The World Health Organization criteria from 2006 were used to define (i) normal glucose tolerance as fasting glucose (FG) <6.1 mmol/L and 2-hour glucose <7.8 mmol/L, (ii) impaired FG (IFG) as FG ≥6.1 and ≤7.0 mmol/L and normal 2-hour glucose, and (iii) impaired glucose tolerance (IGT) as 2-hour glucose ≥7.8 mmol/L and <11.1 mmol/L and normal FG (45).

Blood hormone analysis

Plasma glucose was promptly analyzed after blood collection on the YSI 2300 Stat Plus glucose analyzer (YSI, Yellow Springs, OH). Serum insulin (insulin ELISA, Mercodia, Uppsala, Sweden) and C-peptide (C-peptide ELISA, Mercodia, Uppsala, Sweden) were analyzed in duplicates using commercially available immunoassays.

For measurements of GLP-1, GIP, and glucagon, blood was collected in precooled EDTA tubes, centrifuged at 4 °C, and stored at −20°C prior to analysis. Total GLP-1 was measured as described previously (46), with a radioimmunoassay using an antibody (code no. 89390) specific for the C-terminal end of the GLP-1 molecule and reacting equally with intact GLP-1 and the primary (N-terminally truncated) metabolite. Total GIP concentration was measured using an antibody (code no. 80867) directed toward the C-terminal end, which reacts fully with intact GIP and the N-terminally truncated forms as described previously (47). For glucagon measurements, a C-terminally directed antibody (code no. 4305) was used, measuring glucagon of pancreatic origin as described previously (48). All samples were extracted in a final concentration of 70% ethanol before measurements. Sensitivity for all assays was <1 pmol/L, and the intraassay coefficient of variation was <10%.

Blood BA and adipsin analysis

Blood for plasma BA and serum adipsin measurements was collected on a separate day (on the clamp study day), in the morning after an overnight fast. As determined using a diet journal, the dietary patterns of the individual subjects were similar during 3 days preceding sampling, before and after the intervention. Furthermore, subjects were asked to avoid physical activity and alcohol in the same period. Blood for plasma preparation was collected in precooled EDTA tubes and centrifuged at 4 °C. Blood for serum preparation was collected in precooled BD Vacutainer^® serum tubes, left at room temperature for 20 minutes, and centrifuged at 3600 rpm at 4 °C. Samples were immediately transferred into separate tubes and stored at −80 °C before analysis. Liquid chromatography coupled to mass spectrometry was applied for plasma BA measurements. Sample preparation and liquid chromatography coupled to mass spectrometry were performed as previously described (49). Briefly, plasma was extracted with methanol-containing labeled internal standards. Clarified extracts were dried via speed vacuum and then reconstituted in 50 µL of 5 mM ammonium acetate in water/acetonitrile (95%/5% v/v) containing 0.1% formic acid. Samples were injected (15 µL) and separated using reverse-phase liquid chromatography. Analytes were ionized using electrospray ionization attached to a Bruker Impact II quadrupole time-of-flight mass spectrometer operated in full-scan, negative-ion mode. Serum adipsin concentrations were determined with the use of a commercially available human ELISA kit [human factor D ELISA kit (ab99969), Abcam, Cambridge, UK], following the instructions of the manufacturer. The sensitivity of the assay is 4 pg/mL. Samples were measured in duplicates, and the mean value was used in the data analysis.

Calculations

Areas under the curve (AUCs) for glucose, insulin, C-peptide, glucagon, GLP-1, and GIP were calculated using the trapezoid rule. Multiple β-cell function indices were calculated using glucose, insulin, and C-peptide values obtained during fasting and during the OGTT to reflect different aspects of the β-cell secretory function. Homeostatic model assessment (HOMA) was used as a measure of β-cell function and insulin resistance (HOMA-IR) and calculated based on fasting levels of glucose and insulin using standard equations (50). Simple fasting ratios of insulin and C-peptide to glucose were calculated as insulin_{0 min}/glucose_{0 min} and C-peptide_{0 min}/glucose_{0 min} (51). The insulinogenic index and the corresponding C-peptidogenic index were determined as markers of first-phase insulin release (52) and calculated as the ratio of increment of insulin or C-peptide to glucose during the initial 30 minutes of the OGTT (Δinsulin_{30–0 min}/Δglucose_{30–0 min} and ΔC-peptide_{30–0 min}/Δglucose_{30–0 min}). The oral disposition index was determined as a marker of insulin release when accounting for the prevailing insulin sensitivity and calculated as the product between Δinsulin_{30–0 min}/Δglucose_{30–0 min} × 1/fasting insulin (53). Finally, composite β-cell indices were calculated based on the total AUC for insulin, C-peptide, and glucose during the 2-hour OGTT: insulin-AUC_0–120/glucose-AUC_0–120 and C-peptide-AUC_0–120/glucose-AUC_0–120 (54).

Statistical analysis

Results are presented as mean ± SEM for normally distributed data and as geometric mean ± SEM when skewed. Baseline data were compared between groups by unpaired two-sample t tests. Effects of NR supplementation were tested between groups by applying two-way repeated measurement mixed model analysis, including the two main effects, that is, group (NR/placebo) and visit (baseline/end-of-trial), and the interaction between group and visit. Post hoc linear pairwise comparisons were conducted in case of a significant interaction. Normal distributions of data were evaluated by inspection of diagnostic plots. In case of skewed data, appropriate logarithmic transformation was used before statistical testing. All available data were included in each analysis. A P value <0.05 was considered statistically significant. The power calculation to determine sample size was based on the primary endpoint measure of the study, that is, insulin sensitivity, which has previously been reported (44). Statistical analysis was made in Stata 14.1 (StataCorp, College Station, TX) and graphical presentations were made in SigmaPlot 11.0 (Systat Software, San Jose, CA).

Results

Baseline characteristics and internal validity

Key metabolic features of the NR group (n = 20) and placebo group (n = 20) are summarized in Table 1. At baseline, the groups were comparable on all parameters. Only fasting glucagon levels tended to differ between groups (P = 0.09). The study groups reflected a phenotype with obesity and insulin resistance. Compliance with the trial medication as well as documentation of absorption and excretion of orally administered NR have previously been published (44). All included participants completed the trial.

Table 1.

Clinical Characteristics of Study Participants at Baseline

	Placebo (n = 20)	NR (n = 20)	P Value
Age, y	60 ± 2.0	58 ± 1.6	0.42
Weight, kg	104.6 ± 2.0	104.8 ± 2.2	0.94
BMI, kg/m2	33.3 ± 0.6	32.4 ± 0.5	0.28
Waist/hip ratio	1.0 ± 0.01	1.0 ± 0.01	0.77
Lean mass, kg	67.2 ± 1.1	68.8 ± 1.6	0.41
Fat mass, kg	32.2 ± 1.4	31.1 ± 1.2	0.55
Fat percentage, %	31.3 ± 1.0	30.2 ± 0.9	0.41
Abdominal visceral adipose tissue, cm2	320.4 ± 17.0	295.5 ± 16.3	0.30
Abdominal subcutaneous adipose tissue, cm2	225.1 ± 13.6	229.6 ± 12.9	0.81
HbA1c, mmol/mol	39.8 ± 0.8	37.7 ± 1.0	0.10
HOMA-IR	2.9 ± 0.4	2.6 ± 0.3	0.52
Fasting glucose, mmol/L	5.7 ± 0.1	5.6 ± 0.1	0.45
Fasting insulin, pmol/L	77.0 ± 9.0	72.2 ± 7.5	0.69
Fasting C-peptide, pmol/L	866.2 ± 64.0	848.4 ± 60.3	0.84
Fasting glucagon, pmol/L	11.0 ± 0.9	8.8 ± 1.0	0.09
Fasting GLP-1, pmol/L	15.8 ± 0.8	14.7 ± 0.6	0.27
Fasting GIP, pmol/L	13.0 ± 1.3	12.9 ± 1.1	0.93
Adipsin, ng/mL	5294.7 ± 232.4	5160.7 ± 307.9	0.73

Data are mean ± SEM. Group comparisons were performed by unpaired two-sample t tests. Age, HbA_1c, and body composition measures have previously been published (44).

Glucose

Plasma glucose concentration curves during OGTTs before and after 12 weeks with NR supplementation or placebo are depicted in Fig. 1A. As specified in Table 2, no differences in fasting glucose, 2-hour glucose, peak glucose, or glucose tolerance (assessed as AUC_total for glucose) were apparent due to treatment.

Figure 1.

Time courses of (A) glucose, (B) insulin, (C) C-peptide, and (D) glucagon during a 75-g OGTT before and after 12 wk with NR supplementation (n = 20) or placebo (n = 20). Data are mean ± SEM.

Table 2.

Fasting, Peak, and AUC Values for Glucose, Insulin, C-Peptide, Glucagon, GLP-1, GIP, and Adipsin During a 75-g OGTT Before and After 12 wk With NR Supplementation or Placebo

	Placebo (n = 20)		NR (n = 20)		P Valuea
	Baseline	End-of-trial	Baseline	End-of-trial
Glucose
Fasting, mmol/L	5.7 ± 0.1	5.8 ± 0.1	5.6 ± 0.1	5.6 ± 0.1	0.33
120 min, mmol/L	7.61 ± 0.6	8.06 ± 0.6	7.55 ± 0.4	7.49 ± 0.4	0.29
Peak, mmol/L	10.9 ± 0.6	11.1 ± 0.6	10.4 ± 0.3	10.3 ± 0.3	0.53
AUC0–120, mmol/L⋅min	1080.4 ± 50.2	1093.6 ± 53.8	1042.1 ± 28.4	1036.2 ± 28.9	0.60
Insulin
Fasting, pmol/Lb	77.0 ± 9.0	88.2 ± 9.7	72.2 ± 7.5	80.8 ± 6.8	0.70
Peak, pmol/L	592.6 ± 52.2	578.1 ± 30.1	583.4 ± 56.5	597.4 ± 70.5	0.69
AUC0–120, pmol/L⋅min	48,568.1 ± 4266.9	47,148.0 ± 3004.6	51,497.6 ± 5069.1	50,253.8 ± 5725.9	0.98
AUC0–15, pmol/L⋅min	2494.9 ± 194.1	2910.4 ± 213.9	2886.8 ± 235.1	2736.4 ± 235.5	0.09
AUC15–60, pmol/L⋅min	19,721.3 ± 1596.7	19,505.6 ± 1299.8	20,728.9 ± 1919.5	19,906.9 ± 2035.2	0.63
C-peptide
Fasting, pmol/Lb	866.2 ± 64.0	960.8 ± 73.6	848.4 ± 60.3	928.5 ± 51.5	0.75
Peak, pmol/L	3117.7 ± 167.0	2989.8 ± 137.1	3090.2 ± 171.3	3068.9 ± 200.6	0.61
AUC0–120, pmol/L⋅min	287,297.3 ± 13,606.0	279,238.1 ± 11,930.0	295,910.3 ± 16,653.9	288,537.4 ± 15,912.6	0.97
AUC0–15, pmol/L⋅min	17,745.4 ± 919.9	19,520.3 ± 1042.9	18,946.1 ± 1092.4	19,202.6 ± 918.5	0.13
AUC15–60, pmol/L⋅min	103,285.9 ± 4457.5	102,072.4 ± 3639.3	107,963.6 ± 5946.2	104,480.3 ± 5534.9	0.76
Glucagon
Fasting, pmol/L	11.0 ± 0.9	10.4 ± 0.9	8.8 ± 1.0	8.5 ± 1.1	0.60
Nadir, pmol/L	3.5 ± 0.5	4.4 ± 0.5	3.1 ± 0.6	2.8 ± 0.4	0.12
AUC0–120, pmol/L⋅minc	900.8 ± 78.6	904.5 ± 72.3	664.9 ± 92.2	689.6 ± 89.3	0.83
GLP-1
Fasting, pmol/Lb	15.8 ± 0.8	17.6 ± 1.0	14.7 ± 0.6	16.9 ± 0.9	0.69
Peak, pmol/L	34.8 ± 1.7	35.0 ± 2.5	29.7 ± 2.0	31.4 ± 1.5	0.59
AUC0–120, pmol/L⋅min	3317.6 ± 133.9	3242.3 ± 218.9	2926.9 ± 183.3	3114.8 ± 147.5	0.26
GIP
Fasting, pmol/L	13.0 ± 1.3	11.7 ± 0.9	12.9 ± 1.1	14.1 ± 1.8	0.18
Peak, pmol/L	63.9 ± 5.7	61.4 ± 5.6	62.8 ± 5.1	67.7 ± 7.0	0.31
AUC0–120, pmol/L⋅min	5595.4 ± 398.9	5507.6 ± 425.3	5700.4 ± 428.4	5859.4 ± 413.0	0.60
Adipsin
Fasting, ng/mL	5294.7 ± 232.4	5441.3 ± 282.0	5160.7 ± 307.9	5147.2 ± 289.1	0.63

Data are mean ± SEM. Group comparisons were made by two-way repeated measurement mixed model analysis.

^a

P value denotes interaction (group × visit).

^b

Main effect of visit (P < 0.05).

^c

Main effect of group (P < 0.05).

The mean fasting glucose and 2-hour glucose values for the groups approached the World Health Organization limits defining IFG and IGT. Within the NR group, five individuals had IFG at baseline and three at the end of the trial, whereas seven individuals had IGT at baseline and seven at the end of the trial. In the placebo group, five individuals had IFG at baseline and four at the end of the trial, whereas eight individuals had IGT at baseline and seven at the end of the trial.

Pancreatic hormones

Overall, insulin, C-peptide, and glucagon did not change in response to 12 weeks of NR supplementation (Fig. 1B–1D; Table 2).

Fasting levels of insulin and C-peptide increased slightly in both groups during the course of the study independently of treatment (main effect of visit: insulin, P < 0.01; C-peptide, P < 0.001). Also, HOMA-IR increased in the both groups independently of treatment, indicating a slight aggravation of mainly hepatic insulin resistance (placebo group, 2.87 ± 0.36 to 3.37 ± 0.42; NR group, 2.59 ± 0.26 to 2.86 ± 0.23; main effect of visit, P < 0.01).

The β-cell secretory response during the OGTTs (assessed as AUC_total for insulin and C-peptide) as well as first- and second-phase insulin release (assessed as AUC_0–15 and AUC_15–60) were unaffected by the intervention (Table 2). Likewise, NR did not alter fasting levels of glucagon, nor the ability to suppress glucagon during the OGTT (glucagon nadir) (Table 2). On both visits the placebo group displayed a significantly higher overall glucagon level (AUC_total) than did the NR group.

Incretin hormones and adipsin

Secretion of the incretin hormones GLP-1 and GIP was evaluated during the OGTT as illustrated in Fig. 2, and as indicated in Table 2, there were no effects of NR on fasting levels, peak levels, or the total integrated AUC of either hormone. Fasting GLP-1 increased slightly in both groups independently of treatment. Fasting levels of serum adipsin did not change after 12 weeks of supplementation with NR or placebo (Table 2).

Figure 2.

Time courses of (A) GLP-1 and (B) GIP during a 75-g OGTT before and after 12 wk with NR supplementation (n = 20) or placebo (n = 20). Data are mean ± SEM.

β-Cell function indices

Effects of NR supplementation on β-cell function were assessed based on calculations of various β-cell function indices, both in the basal, fasted state and dynamically during the OGTT (Table 3). No significant differences in β-cell function due to treatment were observed for any calculated index. β-Cell indices derived from basal state data increased in both groups during the study period, but this was independent of treatment.

Table 3.

β-Cell Function Indices Assessed in Basal Conditions and During a 75-g OGTT Before and After 12 wk With NR Supplementation or Placebo

	Placebo (n = 20)		NR (n = 20)
	Baseline	End-of-Trial	Baseline	End-of-Trial	P Valuea
Basal assessments
HOMA of β-cell function, %b,c	91.8 ± 10.7	99.5 ± 10.0	93.2 ± 14.3	109.0 ± 14.7	0.40
Fasting insulin/fasting glucose, pmol/mmolb	13.3 ± 1.5	14.9 ± 1.5	13.0 ± 1.4	14.7 ± 1.4	0.95
Fasting C-peptide/fasting glucose, pmol/mmolb	150.6 ± 10.4	163.9 ± 11.1	152.0 ± 11.2	168.3 ± 10.4	0.71
Dynamic assessments
Insulinogenic index, pmol/mmol	91.1 ± 9.3	97.3 ± 9.8	100.2 ± 9.3	98.0 ± 10.6	0.48
C-peptidogenic index, pmol/mmol	345.2 ± 29.1	342.6 ± 33.3	389.6 ± 30.7	355.7 ± 28.4	0.30
Oral disposition index	1.4 ± 0.1	1.5 ± 0.3	1.5 ± 0.1	1.3 ± 0.1	0.11
Composite β-cell indices
Insulin-AUC0–120/glucose-AUC0–120, pmol/mmol	45.3 ± 4.0	44.0 ± 3.0	48.9 ± 4.3	48.6 ± 5.4	0.83
C-peptide-AUC0–120/glucose-AUC0–120, pmol/mmol	271.1 ± 13.3	260.9 ± 11.8	283.8 ± 13.6	280.9 ± 16.0	0.57

Data are mean ± SEM. Group comparisons were made by two-way repeated measurement mixed model analysis.

^a

P value denotes interaction (group x visit).

^b

Main effect of visit (P < 0.05).

^c

Geometric mean ± SEM.

Circulating BAs

Fifteen unconjugated, glycine-conjugated, and taurine-conjugated BAs were measured semiquantitatively in plasma before and after 12 weeks of NR supplementation or placebo (Fig. 3). Overall, there were no significant differences between groups, although glycoursodeoxycholic acid tended to increase following NR supplementation (interaction, P = 0.07). Several BAs differed significantly between visits, including cholic acid, muricholic acid γ, and tauroursodeoxycholic acid, whereas glycocholic acid tended to increase between visits (main effect of visit, P = 0.09). However, these differences occurred independently of the intervention.

Figure 3.

Circulating BAs response to 12 wk with NR supplementation (n = 20) or placebo (n = 20). Data are presented as fold change from baseline to the end of the trial within each group. A dotted, horizontal line represents no change between baseline and end-of-trial values. Comparisons between groups were made by two-way repeated measurement mixed model analysis. No significant treatment effects were observed. Data are mean ± SEM. *P < 0.05, main effect of visit (baseline vs end-of-trial). CA, cholic acid; DCA, deoxycholic acid; GCA, glycocholic acid; GCDCA, glycochenodeoxycholic acid; GDCA, glycodeoxycholic acid; GUDCA, glycoursodeoxycholic acid; MCA alpha, muricholic acid α; MCA gamma, muricholic acid γ; TCA, taurocholic acid; TCDCA, taurochenodeoxycholic acid; TDCA, taurodeoxycholic acid; TLCA, taurolithocholic acid; TMCA alpha, tauromuricholic acid α; TUDCA, tauroursodeoxycholic acid; UDCA, ursodeoxycholic acid.

Discussion

The coordinated actions of β-cells, incretin hormone–producing cells, and α-cells are major determinants of healthy glucose homeostasis. In this study, we sought to test whether NAD⁺ precursor supplementation would affect the secretory capacity of these cells in humans. In this randomized placebo-controlled trial, we demonstrate that 12 weeks of supplementation with NR had no impact on fasting or post-OGTT levels of glucose, insulin, C-peptide, glucagon, GLP-1, or GIP in nondiabetic males with obesity and insulin resistance. Furthermore, our data indicate that NR does not change the level of circulating adipsin or BAs. As previously published (44), we have also conducted hyperinsulinemic euglycemic clamps in this study group and observed no changes in insulin sensitivity. Collectively, our data do not support that NR impacts glucose homeostasis in men with obesity and insulin resistance, either through mechanisms of ameliorating insulin resistance or augmenting insulin secretion.

Several preclinical studies have shown beneficial effects of NR on glucose metabolism and other aspects of metabolic regulation [reviewed in (12, 13)]. In animal models of diet-induced obesity and diabetes, NR improves glucose tolerance primarily by requiring less insulin to counter glucose levels following glucose challenges, a finding indicative of increased insulin sensitivity (14–16). None of these studies focused on aspects of β-cell function or measured incretin hormones. Only the NAD⁺ precursor NMN has specifically been shown to enhance GSIS in rodent and in in vitro models (27, 29–31). No clinical studies examining NMN are available, but one study is currently recruiting and seeking to test NMN effects on insulin sensitivity and β-cell function in humans (NCT03151239). Differences in the cellular uptake in distinct tissues may aid in explaining dissimilarities between various forms of NAD⁺ precursors, and recently Slc12a8 was identified as a specific NMN transporter in mice, with high levels of gene expression in the murine pancreas (55). Although a specific NR transporter has been identified in yeast cells (56), NR is thought to enter mammalian cells via equilibrative nucleoside transporters (57), which are considered to be ubiquitously expressed in most human tissues, including the pancreas (58). Significantly, a requirement for extracellular dephosphorylation of NMN to NR prior to tissue uptake has been suggested (28, 57). To which extent NR augments NAD⁺ metabolism in specific tissues in humans remain unknown and requires further exploration, for example, by using isotope-labeled NR tracers.

The B₃ vitamin nicotinamide has been tested in humans for β-cell protective effects. In the large-scale, randomized, placebo-controlled type 1 diabetes (T1D) prevention trial (ENDIT trial), 552 first-degree relatives of T1D patients were subjected to high-dose nicotinamide (1 to 2 g/m²) for 5 years (59). IV glucose tolerance tests (IVGTTs) to determine first-phase insulin response were performed repeatedly during the lengthy intervention period, but differences in first-phase insulin response were not detected between groups, and nicotinamide did not reduce T1D occurrence. Similar to NR, nicotinamide is an NAD⁺ precursor; however, in contrast to NR, it is a direct inhibitor of SIRT1 activity (60), which may impede NAD⁺/SIRT1-derived effects on metabolism.

Only a few human clinical studies investigating NR have been published, although several (>20) are recruiting (clinicaltrials.gov). It has been established that NR can provide increases in NAD⁺ levels in blood, is well-tolerated, and appears safe in humans (44, 61–63). A single clinical trial investigated potential NR-derived effects on a wide range of outcomes in 24 healthy, middle-aged, and normal-weight individuals during a 6-week randomized, placebo-controlled crossover trial (63). Although results pointed to effects of NR on blood pressure and arterial stiffness, no metabolic parameters changed following NR supplementation, including fasting insulin, glucose, HOMA-IR, and HOMA of β-cell function. Additionally, frequently sampled IVGTTs were performed to assess insulin sensitivity, but only in a subset of subjects (n = 13), who did not display changes following treatment. However, the lack of changes in metabolic endpoints in the above-described study is perhaps not surprising considering the healthy, lean phenotype of the study participants.

We chose to investigate men with obesity and insulin resistance who were otherwise healthy and nonmedicated. We sought to obtain homogeneous groups and reduce confounding factors, which would limit detection of NR treatment effects. Furthermore, this design offered an opportunity to test NR as a potential preventive strategy for T2D. However, although our study participants were insulin resistant according to an increased HOMA-IR >2.5, these middle-aged men with obesity did not have overt diabetes, suggesting an increased resilience toward obesity-derived metabolic perturbations. Thus, our groups had normal fasting glucose (<6.1 mmol/L) and 2-hour glucose following an OGTT (<7.8 mmol/L) on average, although approaching the cutoff values for IFG and IGT. Retaining normoglycemia in the face of insulin resistance expectedly required insulin levels to rise above the normal range. Hence, an appropriate β-cell compensatory response was already present in these subjects, which possibly would make it difficult to detect further improvements.

Raised fasting and postprandial levels of glucagon has detrimental effects on glucose homeostasis. To our knowledge, no studies have assessed glucagon in relation to NAD⁺ precursor supplementation. In our study, we did not observe effects of NR on prevailing glucagon levels. The placebo group displayed higher glucagon levels than did the NR group, suggesting a slightly more unhealthy metabolic profile. An initial paradoxical rise in glucagon levels following oral ingestion of glucose was observed in both groups (Fig. 1D). This phenomenon has previously been described among both individuals with insulin resistance and individuals with T2D (64). We considered whether NR might alter the circulating levels of adipsin, as this adipokine has been associated with NAD⁺ metabolism in mice (33). The adipsin concentrations measured in this study were comparable to those found in patients with impaired glucose tolerance (65); however, adipsin did not change in response to NR.

Finally, we speculated whether NR would affect BA levels in plasma. This might occur directly through a direct effect of NR on hepatic NAD phosphate (16) or indirectly by NR-derived changes in the gut microbiome composition, which is known to modulate the BA pool (66). In the current study, we did not detect significant NR-derived changes in circulating BAs. Several BAs differed between visits, but independently of treatment. We noticed large variability in BA levels between study participants and visits, even though participants were carefully selected to match phenotypically, and we had controlled for diet 3 days preceding fasting measurements. A recent study suggested positive effects of NR in rejuvenating intestinal stem cells from aging mice (67). Future clinical studies should address whether NR directly impacts gut health in humans, for example, by changing the microbiota profile.

Strengths of the current study include the randomized, parallel-group design, a fair sample size, and good compliance toward the investigational compounds. We chose to administer NR at 2000 mg daily, >100-fold higher than the daily requirement of B₃ vitamin (68), to ensure that effects were not missed due to subtherapeutic dosing. Additionally, the study duration of 12 weeks is a reasonable timeframe to expect changes in glucose regulation.

A certain degree of metabolic abnormalities might be a prerequisite to benefit from NR supplementation, and our study groups could be too metabolically healthy. We cannot assume that NR supplementation will not have an effect in patients with more severely deranged glucose metabolism and β-cell function, for example, as in T2D. Furthermore, this study was powered to detect a difference in the primary endpoint insulin sensitivity, reported previously (44), and the possibility of type II errors cannot be excluded. However, the absence of changes of clinically relevant effect sizes observed in this study suggests adequate statistical power.

Although the OGTT is a validated tool to assess β-cell function (69), this method is not considered a gold standard. Other more specific methods to examine β-cell function, such as the IVGTT and hyperglycemic clamp, would potentially enhance the possibility of detecting more subtle effects on β-cell function. However, considering the complete lack of response to NR observed in the current study, these methods would probably not provide additional value. Moreover, in comparison with these methods, the OGTT has the advantage of being a more physiologically relevant test, capturing the interplay between glucose, insulin, glucagon, and the incretin axis following orally delivered glucose.

In conclusion, the growing body of evidence established preclinically suggests benefits of NAD⁺ precursor supplementation to counter metabolic disease. In the present study, we designed and implemented an early proof-of-concept study in men with obesity and insulin resistance to test for translational value of NR in humans. Our data do not suggest effects of NR on modulating β-cell function, incretin hormones, glucagon, adipsin, or BAs in nondiabetics. Future studies should examine whether NR has effects in individuals with “prediabetic” states (e.g., IFG, IGT) or overt T2D. Additionally, other types of NAD⁺ precursors, such as NMN and NRPT (NR and SIRT1 activator pterostilbene) (70), should also be investigated, before making definitive conclusions on the lack of effects of NAD⁺-based therapies on glucose metabolism.

Acknowledgments

The authors thank the medical laboratory technicians Susanne Sørensen and Lone Kvist from the Medical Research Laboratory, Aarhus University Hospital (Aarhus, Denmark). The authors also thank the dedicated study participants and ChromaDex for the provision of Niagen™ and placebo capsules.

Financial Support: This work was supported by grants obtained from the Novo Nordisk Foundation (Excellence Project Award NNF14OC0009315 to J.T.T. and NNF13OC0003882 to N.J.), the Danish Council for Independent Research (DFF 4004-00235 to J.T.T.), and the Danish Diabetes Academy to O.L.D. Additionally, support for this study was provided by the Novo Nordisk Foundation Center for Basic Metabolic Research (NNF-CBMR). NNF-CBMR is an independent Research Center at the University of Copenhagen and is partially funded by an unrestricted donation from the Novo Nordisk Foundation.

Clinical Trial Information: ClinicalTrials.gov no. NCT02303483 (registered 1 December 2014).

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability: Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.

Abbreviations:

Abbreviations:
 
• AUC

  area under the curve

 
• BA

  bile acid

 
• BMI

  body mass index

 
• FG

  fasting glucose

 
• GIP

  glucose-dependent insulinotropic polypeptide

 
• GLP-1

  glucagon-like peptide 1

 
• GSIS

  glucose-stimulated insulin secretion

 
• HOMA

  homeostatic model assessment

 
• HOMA-IR

  homeostatic model assessment of insulin resistance

 
• IFG

  impaired fasting glucose

 
• IGT

  impaired glucose tolerance

 
• IVGTT

  IV glucose tolerance test

 
• NAD⁺

  nicotinamide adenine dinucleotide

 
• NMN

  nicotinamide mononucleotide

 
• NR

  nicotinamide riboside

 
• OGTT

  oral glucose tolerance test

 
• SIRT

  sirtuin

 
• T1D

  type 1 diabetes

 
• T2D

  type 2 diabetes