Nicotinic acid riboside maintains NAD+ homeostasis and ameliorates aging-associated NAD+ decline

Abstract

Liver-derived circulating nicotinamide from nicotinamide adenine dinucleotide (NAD⁺) catabolism primarily feeds systemic organs for NAD⁺ synthesis. We surprisingly found that, despite blunted hepatic NAD⁺ and nicotinamide production in liver-specific nicotinamide nucleotide adenylyltransferase 1 (NMNAT1) deletion mice (liver-specific knockout [LKO]), circulating nicotinamide and extra-hepatic organs' NAD⁺ are unaffected. Metabolomics reveals a massive accumulation of a novel molecule in the LKO liver, which we identify as nicotinic acid riboside (NaR). We further demonstrate cytosolic 5'-nucleotidase II (NT5C2) as the NaR-producing enzyme. The liver releases NaR to the bloodstream, and kidneys take up NaR to synthesize NAD⁺ through nicotinamide riboside kinase 1 (NRK1) and replenish circulating nicotinamide. Serum NaR levels decline with aging, whereas oral NaR supplementation in aged mice boosts serum nicotinamide and multi-organ NAD⁺, including kidneys, and reduces kidney inflammation and albuminuria. Thus, the liver-kidney axis maintains systemic NAD⁺ homeostasis via circulating NaR, and NaR supplement ameliorates aging-associated NAD⁺ decline and kidney dysfunction.

Keywords: NAD(+); aging; kidney; liver; nicotinic acid riboside.

Figure 1.. Intact systemic NAD⁺ homeostasis and metabolic fitness despite hepatic NMNAT1 deficiency

(A) Schematic of NAD⁺ and nicotinamide production pathways in the liver. (B) Schematic of liver-specific Nmnat1 knockout mice. (C) Gene expression of liver NAD⁺ synthesis enzymes. n = 5 mice/group. (D) Protein expression of liver NMNAT1. Asterisk, non-specific band. (E–I) Abundance of metabolites in control and LKO mouse livers. n = 4, 6 mice for Con and LKO. (J) Schematic of [U-¹³C]-tryptophan tracing in primary hepatocyte cultures. (K) Abundances of labeled metabolites in control and LKO primary hepatocytes. n = 3/group. (L) Schematic of in vivo [U-¹³C]-tryptophan tracing and time course tissue collection. (M–O) Abundance of labeled metabolites in control and LKO mouse livers. Except for the 15-min time point of Con liver (n = 2), n = 3–4 mice/time point. (P) Abundance of labeled nicotinamide in serum. ns, not significant. Except for the 15-min time point of Con liver (n = 2), n = 3–4 mice/time point. (Q) Abundance of nicotinamide in serum. n = 7 mice/group. (R) Tissue NAD⁺ levels in control and LKO mice. n = 10 mice for Con and n = 13, 14 mice for LKO. Data are means ± SE. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by a two-tailed unpaired Student’s t test (C–I and K) or two-way ANOVA with Šídák’s multiple comparisons test (M–P).

Figure 2.. NaR is produced by the liver and released into circulation

(A) Untargeted metabolomics in control and LKO mouse livers. n = 4, 6 mice for Con and LKO. (B) Abundance of m/z 256.08 in control and LKO mouse livers. n = 4, 6 mice for Con and LKO. (C) Untargeted metabolomics in control and LKO mouse sera. n = 7 mice/group. (D) Abundance of m/z 256.08 in control and LKO mouse sera. n = 7 mice/group. (E) LC retention time of m/z 256.08, NaMN, and NMN. (F) Prediction of m/z 256.08 as NaR based on MS/MS fragmentation patterns. (G) Comparison of MS/MS fragmentation between m/z 256.08 and NaR standard. (H) Comparison of LC retention time between m/z 256.08 and NaR standard. (I) Serum NaR levels in saline- or antibiotics-treated Con and LKO mice. n = 4–5 mice/group. (J) Labeled NaR in LKO mouse organs after [U-¹³C]-tryptophan gavage. n = 2–4 mice/time point. (K) The area under the curve (AUC) of labeled NaR in LKO mouse organs. n = 2–4/time point. (L) Labeled NaR levels in LKO mouse serum. n = 3–4 mice/time point. (M) Abundance of labeled NaR in control and LKO mouse serum following intravenous (i.v.) injection of [U-¹³C]-tryptophan. n = 3 mice/time point. (N and O) NaR levels in control and BKO mouse BAT (N) and serum (O). n = 5, 6 mice for Con and BKO. (P and Q) NaR levels in control and MKO mouse muscle (P) and serum (Q). n = 7, 5 mice for Con and MKO. Data are means ± SE. ***p < 0.001 and ****p < 0.0001 by a two-tailed unpaired Student’s t test (B and D), one-way ANOVA with Dunnett’s multiple comparisons (I), or two-way ANOVA with Šídák’s multiple comparisons test (M).

Figure 3.. NT5C2 mediates hepatic NaR production

(A) Comparison of MS/MS fragmentation between unlabeled and M + 6-labeled NaR in LKO mouse liver. (B) NAD⁺ and nicotinamide production pathway from [U-¹³C]-tryptophan. (C) Similar structures and enzymatic reactions between cytidine and NaR production. (D) NaMN concentrations in LKO mouse organs. n = 13–15/organ. (E) Schematic of [U-¹³C]-tryptophan tracing in primary hepatocytes with knockdown of each Nt5c enzyme. (F) Labeled NaR/NaMN in primary hepatocytes with knockdown of each Nt5c enzyme. n = 3/group. (G) Labeled NaR levels in the media of primary hepatocytes with knockdown of each Nt5c enzyme. NS, no signal. n = 3/group. (H) Labeled NaR level in the media of primary hepatocytes treated with NT5C2 inhibitor, CRCD2. n = 3/group. (I) Schematic of [U-¹³C]-tryptophan tracing in LKO mice injected with AAV8-shCtrl or AAV8-shNt5c2. (J) Liver labeled NaR/NaMN in LKO mice injected with AAV8-shCtrl or AAV8-shNt5c2. n = 8 mice/group. (K) Serum-labeled NaR in LKO mice injected with AAV8-shCtrl or AAV8-shNt5c2. n = 8 mice/group. Data are means ± SE. **p < 0.01, ***p < 0.001, and ****p < 0.0001 by a two-tailed unpaired Student’s t test (J), one-way ANOVA with Dunnett’s multiple comparisons (F–H), or two-way ANOVA with Šídák’s multiple comparisons test (K).

Figure 4.. The kidney uses circulating NaR to produce NAD⁺

(A) Schematic of bolus ²H₄-NaR tracing and time course tissue collection. (B and C) Labeled NaAD and NAD⁺ levels in mice after ²H₄-NaR injection. Except for the 30-min time point of jejunum (n = 2), n = 3–4 mice/time point. (D) Schematic of ²H₄-NaR continuous infusion tracing in LKO mice and the calculation of circulating NaR contribution to tissue metabolites. (E) Contribution of circulating NaR to NaAD and NAD⁺ in LKO kidneys following ²H₄-NaR continuous infusion. n = 3/group. (F) Labeled NAD⁺ levels in control and LKO mouse organs from [U-¹³C]-tryptophan. n = 2–4 mice/time point. (G) NAD⁺ synthesis pathways from NaR or tryptophan in the kidney. (H–J) Abundances of metabolites in control and LKO kidneys. Except for the 15- and 60-min time points of Con kidneys (n = 2), n = 3–4 mice/time point. (K) Schematic of primary renal cell cultures treated with [²H₄]-NaR with Nmrk1 knockdown. (L and M) Labeled NaAD and NAD levels in primary renal cell cultures treated with [²H₄]-NaR. n = 6/group. (N) Schematic of [U-¹³C]-tryptophan tracing in LKO mice injected with AAV8-shCtrl or AAV8-shNt5c2. (O) Labeled NAD⁺ levels in LKO mice injected with AAV8-shCtrl or AAV8-shNt5c2. n = 8 mice/group. (P) Correlation between serum-labeled NaR levels and kidney-labeled NaMN levels. n = 8 mice/group. (Q) Correlation between serum-labeled NaR levels and kidney-labeled NaAD levels. n = 8 mice/group. Data are means ± SE. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by a two-tailed unpaired Student’s t test (L, M, and O) or two-way ANOVA with Šídákk’s multiple comparisons test (H–J). Simple linear regression is used to calculate R and p values (P and Q).

Figure 5.. The kidney generates circulating nicotinamide from NaR

(A and B) Labeling of nicotinamide in control and LKO mouse organs from [U-¹³C]-tryptophan. n = 2–4 mice/time point. (C) Labeling of nicotinamide in wild-type mouse organs after bolus ²H₄-NaR injection. Except for the 30-min time point of jejunum (n = 2), n = 3–4 mice/time point. (D) Contribution of circulating NaR to nicotinamide in LKO mouse organs and blood following continuous ²H₄-NaR infusion. n = 3 mice/group. (E) Schematic of arterio-venous (AV) metabolite gradient measurement across the kidney. (F and G) Log₂ V/A ratio of NaR (G) and nicotinamide (H) in control and LKO mouse kidneys. Positive values mean net release, while negative values mean net uptake. n = 7, 9 mice for Con and LKO. (H) Labeled nicotinamide in the media from primary renal cell cultures treated with [₂H₄]-NaR. n = 6/group. (I) Labeled nicotinamide in kidneys of LKO mice injected with AAV-shCtrl or AAV-shNt5c2. n = 8 mice/group. (J) Labeled nicotinamide in blood of LKO mice injected with AAV-shCtrl or AAV-shNt5c2. n = 8 mice/group. (K) Correlation between kidney-labeled NAD₊ and serum-labeled nicotinamide. n = 8 mice/group. (L) Labeled NAD₊ levels across organs in LKO mice injected with AAV-shCtrl or AAV-shNt5c2. n = 6–8 mice for LKO+shCtrl and n = 8 for LKO+shNt5c2. Data are means ± SE. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by a one-sample t test (F and G) or two-tailed unpaired Student’s t test (H, I, and L), two-way ANOVA with Šídák’s multiple comparisons test (J). Simple linear regression is used to calculate R and p values (K).

Figure 6.. Circulating NaR is elevated upon oral nicotinic acid supplementation

(A and B) Liver NaMN and NaR levels in wild-type mice after oral administration of nicotinic acid. n = 4–5 mice/group. (C) Serum NaR levels in wild-type mice after oral administration of nicotinic acid. n = 4 mice/time point. (D–F) Kidney NaMN, NaAD, and NAD⁺ levels in wild-type mice after oral administration of nicotinic acid. n = 4–5 mice/group. (G) Liver NaMN and NaR labeling 6 h after oral administration of [²H₄]-nicotinic acid. n = 5 mice. (H) Serum ²H₄-NaR levels following oral administration of [²H₄]-nicotinic acid. n = 5 mice/time point. (I) Kidney NaAD, NAD⁺, and nicotinamide labeling 6 h after oral administration of [²H₄]-nicotinic acid. n = 5 mice. (J–L) Labeled NaAD, NAD⁺, and nicotinamide levels in primary renal cells treated with 100 μM [²H₄]-nicotinic acid or [²H₄]-NaR. n = 4/group. (M and N) Abundance of labeled NaR and NaMN in primary hepatocytes treated with [²H₄]-nicotinic acid. n = 3/group. (O) Labeled NaR-to-NaMN ratio in primary hepatocytes treated with [²H₄]-nicotinic acid. n = 3/group. (P) Labeled NaR levels in the media of primary hepatocytes with Nt5c2 knockdown. n = 3/group. (Q) Schematic of liver-kidney crosstalk via NaR after oral nicotinic acid administration. Data are means ± SE. *p < 0.05, **p < 0.01, ***p < 0.001, and ***p < 0.0001 by a two-tailed unpaired Student’s t test (A, B, D–F, and J–O), or two-way ANOVA with Šídák’s multiple comparisons test (P).

Figure 7.. Aging reduces NaR production, while NaR supplement improves organ NAD⁺ levels and kidney function in aged mice

(A) Serum NaR levels in young (n = 9) and aged (n = 7) C57BL6/J bred in UC Irvine. (B) Schematic of [U-¹³C]-tryptophan tracing in young and aged mice. (C) Labeled serum NaR levels in young (n = 5) and aged (n = 4) mice. (D) Tryptophan catabolism gene expression in young (n = 9) and aged (n = 7) mouse livers. (E) Labeled NaMN/QA ratio in young (n = 5) and aged (n = 4) mouse livers. (F and G) NAD⁺ and nicotinamide levels in young (n = 9) and aged (n = 7) mouse kidneys. (H–J) Labeled NaMN, NaAD, and NAD⁺ levels in young (n = 5) and aged (n = 4) mouse kidneys. (K) Schematic of NaR feeding in drinking water in young and aged mice. (L) Serum NaR level in young and aged mice after 2-week NaR feeding. n = 9–14 mice/group. (M–O) Kidney metabolite levels in young and aged mice after 2-week NaR feeding. n = 9–14 mice/group. (P) Serum nicotinamide level in young and aged mice after 2-week NaR feeding. n = 9–14 mice/group. (Q) Organ NAD⁺ levels in young and aged mice after 2-week NaR feeding. n = 9–14 mice/group for liver and n = 8–14 mice/group for heart and pancreas. (R) Kidney and heart ATP levels in young and aged mice after 2-week NaR feeding. n = 9–14 mice/group for kidney and n = 8–14 mice/group for heart. (S and T) Kidney gene expression of Kim1 and Fgl2 in young and aged mice after NaR feeding. n = 9–14 mice/group. (U) Representative immunofluorescent staining for F4/80 (green) and nuclei (DAPI; blue) in young and aged kidneys after NaR feeding. n = 3 images/group. (V) Urine ACR in young and aged mice after NaR feeding. n = 7–12 mice/group. Data are means ± SE. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by a two-tailed unpaired Student’s t test (A and C–J) or one-way ANOVA with Dunnett’s multiple comparisons (L–T and V).