Epithelial NAD+ depletion drives mitochondrial dysfunction and contributes to intestinal inflammation

Abstract

Introduction: We have previously demonstrated that a pathologic downregulation of peroxisome proliferator-activated receptor-gamma coactivator 1-alpha (PGC1α) within the intestinal epithelium contributes to the pathogenesis of inflammatory bowel disease (IBD). However, the mechanism underlying downregulation of PGC1α expression and activity during IBD is not yet clear.

Methods: Mice (male; C57Bl/6, Villincre/+;Pgc1afl/fl mice, and Pgc1afl/fl) were subjected to experimental colitis and treated with nicotinamide riboside. Western blot, high-resolution respirometry, nicotinamide adenine dinucleotide (NAD+) quantification, and immunoprecipitation were used to in this study.

Results: We demonstrate a significant depletion in the NAD+ levels within the intestinal epithelium of mice undergoing experimental colitis, as well as humans with ulcerative colitis. While we found no decrease in the levels of NAD+-synthesizing enzymes within the intestinal epithelium of mice undergoing experimental colitis, we did find an increase in the mRNA level, as well as the enzymatic activity, of the NAD+-consuming enzyme poly(ADP-ribose) polymerase-1 (PARP1). Treatment of mice undergoing experimental colitis with an NAD+ precursor reduced the severity of colitis, restored mitochondrial function, and increased active PGC1α levels; however, NAD+ repletion did not benefit transgenic mice that lack PGC1α within the intestinal epithelium, suggesting that the therapeutic effects require an intact PGC1α axis.

Discussion: Our results emphasize the importance of PGC1α expression to both mitochondrial health and homeostasis within the intestinal epithelium and suggest a novel therapeutic approach for disease management. These findings also provide a mechanistic basis for clinical trials of nicotinamide riboside in IBD patients.

Keywords: PGC1α; colitis; nicotinamide adenine dinucleotide; nicotinamide riboside; poly(ADP) riboside polymers.

Figure 1

SIRT1 mRNA and protein levels were decreased during experimental colitis. Acute colitis was induced by giving mice 2% DSS in their water for 7 days. (A) Sirt1 mRNA levels (n=10/group) and (B, C) protein levels (n=4/group) were decreased in mice subjected to DSS as compared to control mice. (D) The deacetylase activity of SIRT1 was assessed in SIRT1 protein immunoprecipitated from the colonic mucosal scrapings of control and DSS mice (n=4/group). (E) NAD⁺ levels within the colonic mucosa of control and DSS mice were evaluated using a NAD⁺/NADH Quantification Kit (n=8/group). Results are shown as the mean ± SD. The data are representative of at least two independent experiments. Significance differences between groups (as determined by an unpaired two-tailed unpaired t test) are indicated on the graphs. * p < 0.05, ** p < 0.005, and *** p < 0.0005; ns, not significant.

Figure 2

The expression of NAD⁺-synthesizing and NAD⁺-consuming enzymes in the intestinal epithelium during experimental colitis. Acute colitis was induced by giving mice 2% DSS in their water for 7 days. (A) The mRNA (n=10/group) and (B, C) protein levels (n=4/group) of several NAD⁺-synthesizing enzymes were assessed. (D) The mRNA expression (n=10/group) of NAD⁺-consuming enzymes was assessed. (E, F) The protein levels (n=4/group) of the NAD⁺-consuming protein, PARP1, within the intestinal epithelium of DSS mice as compared to the control mice were determined. (G, H) Western blot analysis of poly(ADP)-ribose (pADPr)-modified proteins within the intestinal epithelium of mice subjected to DSS as compared to control mice (n=4/group). Results are shown as the mean ± SD. The data are representative of at least two independent experiments. Significance differences between groups (as determined by an unpaired two-tailed t test) are indicated on the graphs. * p < 0.05, ** p < 0.005, and *** p < 0.0005; ns, not significant.

Figure 3

NAD⁺ supplementation improved disease in mice subjected to DSS. Acute colitis was induced by giving mice 2% DSS in their water for 7 days. Mice were treated with nicotinamide riboside (NR; 500 mg/kg/day) or vehicle (sterilized water; same mg/kg volume as for NR) once a day during DSS exposure. (A) The percent weight loss of each mouse was calculated and averaged as a group. (B) Disease activity index (DAI) scores were calculated for each mouse daily and averaged as a group. For A and B, the data are shown as the mean ± SEM; n=12/group. Asterisks (*) indicate a significance difference between DSS+Vehicle vs. DSS+NR groups via a two-way ANOVA. Plus signs (+) indicate a significant difference between Vehicle Control vs. DSS+Vehicle groups via a two-way ANOVA. Dollar signs ($) indicate a significance difference between NR Control vs. DSS+NR groups via a two-way ANOVA. (C) Colon lengths were measured (n=12/group). (D, E) Colonic sections were stained via H&E and scored in a blinded manner by a pathologist (n=8 for Vehicle Control and NR Control groups; n=10 for DSS+Vehicle and DSS+NR groups); scale bar = 100 μM. (F) The mRNA levels of proinflammatory cytokines were assessed via qPCR (n=10 for Vehicle Control and NR Control groups; n=11 for DSS+Vehicle and DSS+NR groups). Unless indicated otherwise, the data are shown as the mean ± SD. The data are representative of at least two independent experiments. Significance differences between groups [as determined by a one-way ANOVA (C, D, F) or two-way ANOVA (A, B)] are indicated on the graphs. * p < 0.05, ** p < 0.005, and *** p < 0.0005; ns, not significant.

Figure 4

NAD⁺ supplementation preserved PGC1α activity and mitochondrial function in mice subjected to DSS. Mice subjected to DSS were treated with nicotinamide riboside (NR; 500 mg/kg/day) or vehicle (sterilized water; same mg/kg volume as for NR) once a day during DSS exposure. (A) The mRNA levels of genes were assessed via qPCR (n=10 for Vehicle Control and NR Control; n=11 for DSS+Vehicle and DSS+NR). (B, C) Protein levels of PGC1α, TFAM, and SIRT1 in the intestinal epithelium of DSS+Vehicle vs. DSS+NR mice (n=7/group). (D, E) Levels of acetylated PGC1α were assessed via immunoprecipitation (IP) and immunoblot (IB); DV, DSS+Vehicle; DN, DSS+NR. (F) Complex I and II activity within the colonic mucosa from control (n=10/group) vs. experimental mice (n=12/group). (G) NAD⁺ levels were evaluated using a NAD⁺/NADH Quantification Kit. Unless indicated otherwise, the data are shown as the mean ± SD. The data are representative of at least two independent experiments. Significance differences between groups [as determined by a one-way ANOVA (A, F, G) or unpaired two-tailed t test (C, E)] are indicated on the graphs. * p < 0.05, ** p < 0.005, and *** p < 0.0005; ns, not significant.

Figure 5

Villin-cre;Pgc1α^fl/fl mice (Pgc1α^ΔIEC ) mice do not benefit from NR treatment during DSS colitis. Acute colitis was induced by giving mice 2% DSS in their water for 7 days. Mice were treated with nicotinamide riboside (NR; 500 mg/kg/day) or vehicle (sterilized water; same mg/kg volume as for NR) once a day during DSS exposure. (A) The percent weight loss of each mouse was calculated and averaged as a group (n=10/group). (B) Disease activity index (DAI) scores were calculated for each mouse daily and averaged as a group (n=10/group). For A and B, the data are shown as the mean ± SEM. Asterisks (*) indicate a significance difference between Pgc1α^fl/fl DSS+Vehicle vs. Pgc1α^fl/fl DSS+NR via a two-way ANOVA. Plus signs (+) indicate a significance difference between Pgc1α^fl/fl DSS+NR vs. Pgc1α^ΔIEC DSS+NR via a two-way ANOVA. Dollar signs ($) indicate a significance difference between Pgc1α^fl/fl DSS+Vehicle vs. Pgc1α^ΔIEC DSS+Vehicle. (C) Colon lengths were measured (n=10/group; mean ± SD). The data are representative of at least two independent experiments. Significance differences between groups [as determined by a one-way ANOVA (C) or two-way ANOVA (A, B)] are indicated on the graphs. * p < 0.05, ** p < 0.005, and *** p < 0.0005; ns, not significant.

Figure 6

The SIRT1-PGC1α axis is disrupted during infectious colitis. Infectious colitis was induced by infecting mice with 1×10⁹ colony-forming units (CFUs) of Citrobacter rodentium. (A, B) The protein levels of SIRT1 and (C, D) PGC1α and TFAM in the intestinal epithelium of Sham- and Citrobacter-infected mice (n=4/group). (E, F) Levels of acetylated PGC1α were assessed via immunoprecipitation (IP) and immunoblot (IB); S, Sham-infected; C, Citrobacter-infected. (G) NAD⁺ levels were evaluated using a NAD⁺/NADH Quantification Kit. (H, I) Western blot analysis of poly(ADP)-ribose (pADPr)-modified proteins within the intestinal epithelium of mice subjected to DSS as compared to control mice (n=3/group). Unless indicated otherwise, the data are shown as the mean ± SD. The data are representative of at least two independent experiments. Significance differences between groups (as determined by a two-tailed t test) are indicated on the graphs. * p < 0.05 and *** p < 0.0005; ns, not significant.

Figure 7

NAD⁺ supplementation improved disease in Citrobacter-infected mice. Mice were treated with nicotinamide riboside (NR; 500 mg/kg/day) or vehicle (sterilized water; same mg/kg volume as for NR) once a day starting one day post-infection. (A) A schematic outlining the infectious colitis model used in this study. Image was created with BioRender.com. (B) Disease activity index (DAI) scores were calculated for each mouse daily and averaged as a group. (C) The percent weight loss of each mouse was calculated and averaged as a group. For B and C, the data are shown as the mean ± SEM; n=10/group. Asterisks (*) indicate a significance difference between Sham+Vehicle vs. Citrobacter+Vehicle groups via a two-way ANOVA. Plus signs (+) indicate a significant difference between Sham+NR vs. Citrobacter+NR groups via a two-way ANOVA. Dollar signs ($) indicate a significance difference between Citrobacter+Vehicle vs. Citrobacter+NR groups via a two-way ANOVA. (D) Colon lengths were measured (n=10/group). (E) The fecal stool burden (CFU/g of stool) of C rodentium was determined by plating (n=10/group). (F, G) Colonic sections were stained via H&E and scored in a blinded manner by a pathologist (n=6/group); scale bar = 100 μM. Unless indicated otherwise, the data are shown as the mean ± SD. The data are representative of at least two independent experiments. Significance differences between groups [as determined by a one-way ANOVA (D, G) or two-way ANOVA (B, C, E)] are indicated on the graphs. * p < 0.05, ** p < 0.005, and *** p < 0.0005; ns, not significant.

Figure 8

NAD⁺ repletion improved disease in a model of infectious colitis. Infectious colitis was induced by infecting mice with 1×10⁹ colony-forming units (CFUs) of Citrobacter rodentium. Mice were treated with nicotinamide riboside (NR; 500 mg/kg/day) or vehicle (sterilized water; same mg/kg volume as for NR) once a day starting one day post-infection. (A, B) The protein levels of PGC1α, TFAM, and SIRT1 in the intestinal epithelium of Citrobacter-infected mice treated with Vehicle or NR (n=7/group). (B, C) Levels of acetylated PGC1α were assessed via immunoprecipitation (IP) and immunoblot (IB); CV, Citrobacter+Vehicle; CN, Citrobacter+NR. (D) Complex I and II activity within the colonic mucosa from control vs. experimental mice (n=10/group for Sham+Vehicle and Sham+NR; n=12/group for Citrobacter+Vehicle and Citrobacter+NR). (E, F) Western blot analysis of poly(ADP)-ribose (pADPr)-modified proteins within the intestinal epithelium of Citrobacter+Vehicle mice as compared to Citrobacter+NR mice (n=7/group). Unless indicated otherwise, the data are shown as the mean ± SD. The data are representative of at least two independent experiments. Significance differences between groups [as determined by an unpaired two-tailed t test (B, F) or one-way ANOVA (D)] are indicated on the graphs. * p < 0.05, ** p < 0.005, and *** p < 0.0005; ns, not significant.

Figure 9

The SIRT1-PGC1α axis is disrupted during human colitis. Human intestinal tissue was collected from healthy control pediatric patients or pediatric patients with ulcerative colitis (UC). (A) The mRNA levels of genes were assessed via qPCR (n=18 for Control Patients; n=20 for Inflamed UC and Uninflamed UC). (B, C) Western blot analysis of poly(ADP)-ribose (pADPr)-modified proteins within the intestinal epithelium of Healthy Controls vs. UC patients (n=4/group). (D) NAD⁺ levels were evaluated using a NAD⁺/NADH Quantification Kit. (E, F) Levels of acetylated PGC1α were assessed via immunoprecipitation (IP) and immunoblot (IB); C, Control; UC, ulcerative colitis. Unless indicated otherwise, the data are shown as the mean ± SD. The data are representative of at least two independent experiments. Significance differences between groups [as determined by an unpaired two-tailed t test (C, E) or one-way ANOVA (A, D)] are indicated on the graphs. * p < 0.05 and *** p < 0.0005; ns, not significant.

Figure 10

The SIRT1-PGC1α axis during intestinal inflammation. At homeostasis, NAD⁺ levels are adequate such that SIRT1 is able to deacetylate and activate PGC1α, resulting in mitochondrial biogenesis and normal mitochondrial function. However, during intestinal inflammation (e.g., experimental colitis or ulcerative colitis), NAD⁺ levels are depleted within the intestinal epithelium by the activation of PARP1, which renders SIRT1 unable to activate PGC1α. This results in decreased mitochondrial biogenesis and subsequent mitochondrial dysfunction. Mitochondrial dysfunction within the intestinal epithelial cells has detrimental effects on the barrier integrity of the intestinal epithelium. Our preclinical studies have shown that oral supplementation with the NAD⁺ precursor, nicotinamide riboside, increases the activation of PGC1α and overall mitochondrial health and function. Thus, future therapeutic approaches targeting the activity of PGC1α, and in turn mitochondrial health, may complement the treatment for IBD and improve outcomes in patients.