Age-associated changes in oxidative stress and NAD+ metabolism in human tissue

Abstract

Nicotinamide adenine dinucleotide (NAD(+)) is an essential electron transporter in mitochondrial respiration and oxidative phosphorylation. In genomic DNA, NAD(+) also represents the sole substrate for the nuclear repair enzyme, poly(ADP-ribose) polymerase (PARP) and the sirtuin family of NAD-dependent histone deacetylases. Age associated increases in oxidative nuclear damage have been associated with PARP-mediated NAD(+) depletion and loss of SIRT1 activity in rodents. In this study, we further investigated whether these same associations were present in aging human tissue. Human pelvic skin samples were obtained from consenting patients aged between 15-77 and newborn babies (0-1 year old) (n = 49) previously scheduled for an unrelated surgical procedure. DNA damage correlated strongly with age in both males (p = 0.029; r = 0.490) and females (p = 0.003; r = 0.600) whereas lipid oxidation (MDA) levels increased with age in males (p = 0.004; r = 0.623) but not females (p = 0.3734; r = 0.200). PARP activity significantly increased with age in males (p<0.0001; r = 0.768) and inversely correlated with tissue NAD(+) levels (p = 0.0003; r = -0.639). These associations were less evident in females. A strong negative correlation was observed between NAD(+) levels and age in both males (p = 0.001; r = -0.706) and females (p = 0.01; r = -0.537). SIRT1 activity also negatively correlated with age in males (p = 0.007; r = -0.612) but not in females. Strong positive correlations were also observed between lipid peroxidation and DNA damage (p<0.0001; r = 0.4962), and PARP activity and NAD(+) levels (p = 0.0213; r = 0.5241) in post pubescent males. This study provides quantitative evidence in support of the hypothesis that hyperactivation of PARP due to an accumulation of oxidative damage to DNA during aging may be responsible for increased NAD(+) catabolism in human tissue. The resulting NAD(+) depletion may play a major role in the aging process, by limiting energy production, DNA repair and genomic signalling.

Conflict of interest statement

Competing Interests: The authors have read the journal's policy and have the following conflicts: R. Grant and G. Guillemin are directors of a new non-listed start-up Biotech company called NADLife. This does not alter the authors' adherence to all the PLoS ONE policies on sharing data and materials.

Figures

Figure 1. Effect of age on tissue lipid peroxidation in (A) males (B) females.

(A) Lipid peroxidation increases with age in male subjects. MDA levels in human tissue increased significantly with age in males aged between 0–77 years (Line a; p = 0.0003; n = 27). Line b shows post-pubescent (males only) data (p = 0.0044; n = 19). Spearman's correlation coefficient for the non-normally distributed population was r = 0.643 and r = 0.623 for line a, and line b respectively. An exponential (first-order) least squares fit was used to generate the nonlinear trend lines (line a and b). (B) Changes in lipid peroxidation with age in female subjects. The apparent increase in MDA levels in post-pubescent females (36–76 years) is not statistically significant (p = 0.3734; n = 22). Spearman's Correlation coefficient for a non-normally distributed population was 0.200. An exponential (first-order) least squares fit was used to generate the nonlinear trend line.

Figure 2. Correlation between Phosphorylation of H2AX at Ser139 and age in both (A) Males (B) Females.

(A) Phosphorylated-H2AX as a marker for DNA damage in male subjects show a significant correlation with age. Line a represents the correlation for subjects aged between 0–77 years (p = 0.0095; n = 27). Line b shows post-pubescent data (males only) (p = 0.0286; n = 19). Pearson's Correlation coefficient for normally distributed data was r = 0.490 and r = 0.502 for line a, and line b respectively. An exponential (first-order) least squares fit was used to generate the nonlinear trend lines (line a and b). (B) Phosphorylated-H2AX as a marker for DNA damage in post-pubescent female subjects shows a significant positive correlation with age (36–76 years) (p = 0.003; n = 22). Pearson's Correlation coefficient for a normally distributed population was r = 0.600. An exponential (first-order) least squares fit was used to generate the nonlinear trend line.

Figure 3. Correlation between PARP activity and aging in (A) Males (B) Females.

(A) PARP activity increases with age in male subjects. PARP activity increases significantly in male subjects aged between 0–77 years (line a; p<0.0001; n = 27). The data including the post-pubescent subjects (males only) shows no significant change in PARP activity with age (line b; p = 0.0913; n = 19). Pearson's correlation coefficient was a normally distributed population was r = 0.768 and r = 0.399 for line a, and line b respectively. An exponential (first-order) least squares fit was used to generate the nonlinear trend lines (line a and b). (B) PARP activity with age in female subjects (n = 27). The apparent increase in PARP activity with age (36–76 years) is not statistically significant in post-pubescent female subjects (p = 0.4390; n = 22). Spearman's Correlation coefficient for a non-normally distributed population was r = 0.174. An exponential (first-order) least squares fit was used to generate the nonlinear trend line.

Figure 4. Correlation between NAD+ levels and Age in (A) Males (B) Females.

(A) NAD+ concentrations decline with age in males. NAD+ levels decreased significantly in males aged between 0–77 years (line a; p = 0.0007; n = 27). Pearson's correlation coefficient for a normally distributed population, r = −0.769. The post-pubescent data for male subjects also showed a decline in NAD+ levels with age (line b; r = −0.706; p = 0.0001; n = 19). An exponential (first-order) least squares fit was used to generate the nonlinear trend lines (line a and b). (B) NAD+ concentrations decreased significantly with age (36–76) in post-pubescent females (p = 0.01; n = 22). Pearson's correlation coefficient for a normally distributed population,r = −0.537. An exponential (first-order) least squares fit was used to generate the nonlinear trend line.

Figure 5. Correlation between SIRT1 Activity and Age in (A) Males (B) Females.

(A) SIRT1 activity declines with age in post-pubescent males. The apparent negative correlation in SIRT1 activity with age in males aged between 0–77 years was not statistically significant (line a; p = 0.0385; n = 27). Spearman's correlation coefficient for a non-normally distributed population, r = −0.400. Line b represents the post-pubescent data shows a significant negative correlation with age (r = −0.612; p = 0.007; n = 19). An exponential (first-order) least squares fit was used to generate the nonlinear trend lines (line a and b). (B) Changes in SIRT1 with age in females. The apparent increase in PARP activity with age (36–76 years) is not statistically significant (p = 0.3743; n = 22). Spearman's correlation coefficient for a non-normally distributed population,r = 0.194. An exponential (first-order) least squares fit was used to generate the nonlinear trend line.

Figure 6. (A) Lipid peroxidation increases significantly in parallel to increased DNA damage in post pubescent males (p<0.0001).

The subjects were aged between 15–77 years (n = 19). Spearman's correlation coefficient, r = 0.7209. An exponential (first-order) least squares fit was used to generate the nonlinear trend line. (B) PARP activity decreases in line with a decline in NAD+ levels in post pubescent male subjects (p = 0.0149). . The subjects were aged between 15–77 years (n = 19). Pearson's correlation coefficient,r = −0.5491 An exponential (first-order) least squares fit was used to generate the nonlinear trend line. (C) Correlation between NAD+ levels and SIRT1 activity in post pubescent male subjects. The subjects were aged between 15–77 years (n = 19). The correlation is not of statistical significance (p = 0.9829). Pearson's correlation coefficient,r = −0.005263.