NAD ⁺ precursor nutritional supplements sensitize the brain to future ischemic events

Abstract

Nicotinamide adenine dinucleotide (NAD⁺) is a redox cofactor critical for oxidative phosphorylation. Nicotinamide (NAM) and nicotinamide riboside (NR) are NAD⁺ precursors widely used as nutritional supplements to augment oxidative phosphorylation. Indeed, NAD⁺ precursors have been reported to improve outcomes in ischemic stroke when administered as a rescue therapy after stroke onset. However, we have also reported that enhanced reliance on oxidative phosphorylation before ischemia onset might worsen outcomes. To address the paradox, we examined how NAD⁺ precursors modulate the outcome of middle cerebral artery occlusion in mice, when administered either 20 minutes after reperfusion or daily for three days before ischemia onset. A single post-ischemic dose of NAM or NR indeed improved tissue and neurologic outcomes examined at 72 hours. In contrast, pre-ischemic treatment for three days enlarged the infarcts and worsened neurological deficits. As a possible explanation for the diametric outcomes, a single dose of NAM or NR augmented tissue AMPK, PGC1α, SIRT1, and ATP in both naïve and ischemic brains, while the multiple-dose paradigm failed to do so. Our data suggest that NAD⁺ precursor supplements may sensitize the brain to subsequent ischemic events, despite their neuroprotective effect when administered after ischemia onset.

Introduction

The brain is among the most metabolically demanding organs. The very large ATP demand to maintain membrane ion gradients is met by a heavy reliance on oxidative phosphorylation. Both cerebral blood flow (CBF) and hence oxygen delivery are higher than in most other organs. As a result, energy metabolism is rapidly compromised upon loss of CBF or oxygen delivery, and cell death and infarction ensue. The inability to quickly ramp up and rely on non-oxidative phosphorylation (i.e., glycolysis) may, in part, be responsible for the sensitivity of the brain to oxygen supply-demand mismatch. We have previously shown that preemptively shifting cell metabolism from oxidative phosphorylation to glycolysis using the metabolic toggle meclizine, an over-the-counter anti-vertigo drug, renders the brain more resistant to a subsequent focal ischemic challenge. ¹ Hence, reduced reliance on oxidative phosphorylation improved the outcome of an ischemic challenge, suggesting that shifting the metabolic reliance towards oxidative phosphorylation might conversely sensitize the brain tissue to a subsequent ischemic challenge and worsen stroke outcomes.

Nicotinamide adenine dinucleotide (NAD⁺) is a fundamental energy carrier present in all cells. Besides serving as a cofactor in mitochondrial oxidative phosphorylation, ² NAD⁺ is a substrate for various proteins that consume NAD⁺, such as class III histone deacetylases (sirtuins), poly-ADP-ribose-polymerases (PARPs), and ADP ribosyl-cyclases. NAD⁺ is synthesized via three major pathways: de novo biosynthesis from the amino acid tryptophan, the Preiss-Handler pathway from niacin, and the salvage pathway from nicotinamide (NAM). ³ Of these, the NAD⁺ salvage pathway is dominant in the mammalian brain. ⁴ The salvage pathway starts by converting NAM to nicotinamide mononucleotide (NMN) by nicotinamide phosphoribosyl transferase (NAMPT), which is then converted to NAD⁺. Nicotinamide riboside (NR) can also enter the salvage pathway via direct conversion to NMN. Hence, the salvage pathway is an efficient route for systemically administered NAM, NR and NMN supplements to boost NAD⁺. Unlike NR and NMN, however, NAM can also inhibit enzymes that consume NAD⁺, such as sirtuins. ⁵

It is well known that ischemia depletes tissue NAD⁺ and, in this way, hinders mitochondrial ATP generation.^6,7 Consistent with this, pharmacologically enhancing tissue NAD⁺ levels have been neuroprotective when administered after focal cerebral ischemia. For example, treatment with the NAD⁺ precursor NAM, which rapidly penetrates the blood-brain barrier upon systemic administration and elevates brain NAD⁺ levels,^5,7 improved mitochondrial energetics, suppressed oxidative stress, and reduced infarct volumes when administered after focal cerebral ischemia.^5,7–13 Moreover, enhancing NAD⁺ generation via the salvage pathway protected against experimental neuronal injury. ¹⁴

NAD⁺ precursors have emerged as popular daily over-the-counter supplements to attenuate the effects of aging. However, our previous work suggested that pre-ischemic augmentation of cerebral oxidative phosphorylation could be pernicious for stroke outcomes. ¹ We, therefore, undertook experiments to compare the effects of pre-ischemic vs. post-ischemic administration of NAD⁺ precursors in a commonly used model of focal cerebral ischemia. Our findings suggest dichotomous outcomes that implicate a need to recruit different paths of energy metabolism depending on the context. Further, they raise concern that chronic high-dose NAD⁺ precursor supplementation could increase the risk of adverse outcomes among individuals at high risk for cerebral ischemia.

Material and methods

Experimental animals

Experiments were approved by the MGH Institutional Animal Care and Use Committee, carried out in accordance with the Guide for Care and Use of Laboratory Animals (NIH Publication No. 85-23, 1996) and reported in accordance with the ARRIVE guidelines 2.0. We studied C57BL/6J male mice (male, 2–3 months, 28 ± 3 grams; Charles River Laboratories, Wilmington, MA, USA; Hubei Laboratory Animal Service Center, Wuhan, China), and Pgc1α deficient (Pgc1α^−/−) mice (stock no. 008597; Jackson Laboratories; C57BL6J background) that were bred in-house and compared to littermate controls.

Middle cerebral artery occlusion (MCAO)

A single operator performed all transient intraluminal filament MCAO procedures. After a vertical ventral neck incision, the left common carotid artery (CCA), internal carotid artery (ICA) and external carotid artery (ECA) were exposed and dissected clean. The distal ECA was ligated with a 5-0 silk suture. A microvascular clamp was placed on the ICA to avoid retrograde bleeding, and proximal CCA was ligated. A coated filament (602212PK10, Doccol Corporation, USA) was retrogradely introduced into the ECA and advanced into the ICA until a drop in CBF was detected by transcranial laser Doppler flowmetry (Perimed, Sweden) over the left parietal area. The filament was then advanced for another 0.5 mm and secured in place by suture ligation. Forty-five minutes later, the filament was withdrawn, the ECA was permanently ligated, CCA opened and CBF monitored for another 5–10 minutes to confirm reperfusion. Anoxic depolarization latency upon MCAO and neurological deficits and infarct volumes at 72 hours were assessed as described in Supplemental Data 1.

In vivo and vitro treatments

Both NAM (Sigma) and NR (Carbosynth, Toronto Research Chemicals) were dissolved in saline (40 mg/ml) and administered intraperitoneally at 400 mg/kg either as a single dose or once a day for three days. ¹⁵ The dose has been adopted from previous studies where a dose range of 200–500 mg/kg has been used.^{7,8,10–12,16,17} The single-dose post-treatment paradigm has already been tested by others and shown efficacious; we tested a single injection to reproduce the effect. The decision to administer three days of pre-treatment was based on the published 12-hour half-life of NAM. We considered more than five half-lives to be an acceptable time to reach an elevated steady-state. After primary cortical neurons were cultured, oxygen-glucose deprivation (OGD) was used to mimic ischemia in vitro. For culture application, NAM and NR were dissolved in saline and applied at 1, 5 and 20 mM. Compounds were applied for 2 days before OGD for prolonged pretreatment, or at the onset of OGD for acute treatment. Treatment paradigms are specified in the figures.

Ex vivo and in vitro assays

Detailed methodology for high-performance liquid chromatography (HPLC), Western blot, Sirt-1 activation by deacetylation assay, primary neuronal cultures, and cell viability tests are described in the online-only Supplement Data 1.

Rigor and statistics

In the absence of prior experience with the ischemia duration and the surgeon, the sample size in initial MCAO experiments was selected empirically to achieve 90% power to detect a 25% effect size, assuming a standard deviation of 20% of the mean infarct volume (α = 0.05). Technical losses and premature mortality that precluded infarct volume assessment were replaced. The sample size calculation was repeated for each subsequent experiment based on the standard deviation of the prior dataset as well as technical losses and mortality. The AD latency was calculated only in the subset of experiments with reliable CBF recordings and a clearly discernible CBF response to AD, including the experiments that were later excluded from infarct and neurological outcome analyses due to other technical issues. Western blot experiments were repeated three to five times, neuronal cultures four times, and biochemical assays six times. Final sample sizes are indicated in the figures. All experiments and quantifications, as well as data analyses, were performed in a blinded fashion. Mice were randomly picked from the cages and assigned to treatment arms; no specific randomization tool was used. A priori exclusion criteria included surgical complications and technical failures (n = 2 each in NR postischemic and saline preischemic treatment arms). Normal distribution was tested by the Shapiro-Wilk test. Statistical comparisons were designed, a priori, to compare vehicle vs. drug treatment arms in preischemic and postischemic treatment paradigms using two-way ANOVA for repeated measures, followed by Holm-Šídák's multiple comparisons (Prism 8, GraphPad Software, Inc., CA, USA). Data are shown as mean ± SD.

Results

Focal ischemic outcomes

We first sought to confirm the previous reports of a beneficial effect of NAD⁺ enhancers when administered after stroke onset, using a widely accepted transient MCAO model (Figure 1(a)). Systemic administration of NAM or NR (400 mg/kg in saline, intraperitoneal) 20 minutes after reperfusion indeed reduced infarct volumes when measured 3 days later; NR appeared to be more effective and improved the neurological deficits and weight loss as well (Figure 1(b); representative TTC-stained coronal slices for each experimental group are shown in Supplementary data 2). Mortality also tended to be lower in NAM- or NR-treated animals compared with their controls. Importantly, CBF did not differ between drug and vehicle arms.

Figure 1.

The effect of postischemic or preischemic treatment with NAD⁺ precursors on middle cerebral artery occlusion outcomes. (a) After temporary ligation of the common carotid artery (CCA), the filament (red) with a silicone tip (blue) was inserted into the internal carotid artery (ICA) via the external carotid artery (ECA) to induce middle cerebral artery occlusion (MCAO) in mice (left panel). A typical cerebral blood flow (CBF) tracing (middle panel) shows perfusion drop upon CCA and MCAO, and Continued.recovery after removal of the filament for MCA reperfusion (MCAR) and opening of the CCA (CCAR). The sudden stepwise CBF drop at the onset of anoxic depolarization (AD) is also marked. Grey shaded segments indicate where average CBF was measured at baseline (B), after CCAO (C), after MCAO (M), and after reperfusion (R). The right panel shows a representative triphenyl tetrazolium chloride (TTC)-stained brain on day 3 after MCAO (infarct in white). (b) Mice were treated with vehicle (Veh), nicotinamide (NAM), or nicotinamide riboside (NR) 20 minutes after MCAR following 45 minutes MCAO (Treatment), and infarct volume and neuroscore were assessed on day 3. Mortality, weight loss over 3 days, and CBF changes during and after the MCAO procedure are also shown and (c) Mice were treated with either NAM or NR 2 days, 1 day, and immediately before 45 minutes of MCAO (Treatment), and infarct volume and neuroscore were assessed on day 3. Mortality, weight loss over 3 days, and CBF changes during and after the MCAO procedure are also shown. Data were analyzed using two-way ANOVA. Independent variables were treatment (vehicle vs. drug) and compound (NAM vs. NR). P values represent the treatment effect (vehicle vs. drug).

In contrast, preischemic treatment with NAM or NR 2 days, 1 day, and immediately before MCAO onset resulted in larger infarct volumes and worse neurological deficits; mortality was also higher in the NAM arm (Figure 1(c)). Once again, CBF was not affected. Interestingly, hippocampal infarction rates also reflected the opposite effect of postischemic vs. preischemic treatment on stroke outcome (Supplementary data 3). These data revealed a paradoxical worsening of ischemic outcomes upon 2-day preischemic use of NAD⁺ enhancers known to augment mitochondrial oxidative phosphorylation.

One mechanism by which increased reliance on oxidative phosphorylation worsens ischemic outcomes might be faster energetic failure upon loss of oxygenation, which could lead to more rapid AD upon MCAO. To test this, we measured the AD latency using the characteristic CBF surrogate (Figure 2), as previously described. ¹⁸ We found that pretreatment with either NAD⁺ enhancers indeed hastened AD onset after MCAO. Not surprisingly, postischemic treatment cohorts did not significantly differ from their respective controls (p = 0.8884 for treatment effect, two-way ANOVA; not shown), since treatments were administered 20 minutes after reperfusion, i.e., long after AD onset. These data were consistent with worse stroke outcomes upon preischemic augmentation of mitochondrial oxidative phosphorylation.

Figure 2.

The effect of preischemic treatment with NAD⁺ precursors on anoxic depolarization (AD) latency. The timeline shows the experimental protocol. AD latency was quantified after preischemic treatment by measuring the duration between MCAO to the onset of the characteristic CBF drop associated with AD onset. Pretreatment with either NAM or NR hastened the AD onset after MCAO. Two-way ANOVA (independent variables: vehicle vs. drug and NAM vs. NR). P value represents the treatment effect (vehicle vs. drug). AD: anoxic depolarization; MCAO: middle cerebral artery occlusion; CCAO: common carotid artery occlusion; NAM: nicotinamide; NR: nicotinamide riboside; CBF: cerebral blood flow.

The peroxisome proliferator-activated receptor-gamma coactivator 1-alpha (PGC1α) orchestrates mitochondrial biogenesis and is involved in regulating the mitochondrial NAD⁺ pool. ¹⁹ We have previously shown that genetic knockout of PGC1α (PGC1α^−/−) leads to worse injury upon kidney ischemia, in part because of NAM and consequently NAD⁺ depletion. ²⁰ We, therefore, reasoned that PGC1α^−/− mice might also be more sensitive to cerebral ischemia. On the contrary, PGC1α^−/− mice appeared to be more resistant to injury after MCAO compared with wild-type (WT) littermates; mortality tended to be lower in PGC1α^−/− mice, and AD was significantly delayed (Figure 3). The high mortality in WT mice (50%) precluded infarct quantification in the most severely affected animals and likely masked group differences in infarct volume (i.e., mortality bias). Nevertheless, the results suggested that, in sharp contrast to their sensitivity to renal ischemic injury, PGC1α^−/− are resistant to cerebral ischemia, consistent with the overall hypothesis that reduced reliance on mitochondrial oxidative phosphorylation at rest renders the brain resistant to subsequent ischemic insults.

Figure 3.

The outcome of middle cerebral artery occlusion in PGC1α^−/− mice. The timeline shows the experimental protocol. PGC1α^−/− mice showed lower mortality and prolonged AD latencies compared with WT. Tissue outcome was likely confounded by high mortality in the WT. P values represent (PGC1α^−/− vs. WT; t-test). PGC1α^−/−: peroxisome proliferator-activated receptor-gamma coactivator 1-alpha knockout mice; WT: wild type; MCAO: middle cerebral artery occlusion; CCAO: common carotid artery occlusion; M/CCAR: reperfusion.

Energy metabolites

Next, we sought to link stroke outcomes with changes in energy metabolites measured in brain homogenates. In naïve (i.e., nonischemic) brains, systemic administration of a single or three doses of NAM, mimicking postischemic and preischemic treatment paradigms, respectively, elevated brain NAM levels by approximately 2-fold, while treatment with NR did not (Figure 4(a)). In contrast, both NAM and NR elevated tissue NAD⁺ levels by up to 50%. In general, the three-dose regimen appeared more effective than a single dose, and NR appeared more effective than NAM. Interestingly, however, tissue ATP levels were raised by the single-dose regimen only, where NR tended to have a stronger effect (∼27% increase). In fact, the three-dose NAM regimen decreased tissue ATP levels by almost 30%.

Figure 4.

The effect of NAD⁺ precursors on brain energy metabolites. (a) Naïve/nonischemic brains: The timeline shows the experimental protocol. The effect of single or multiple dosing with NAD⁺ precursors (mimicking postischemic and preischemic treatment paradigms, respectively) on brain tissue NAM, NAD⁺ and ATP levels are shown. One-way ANOVA followed by Holm-Šídák's multiple comparisons. (b) Ischemic brains: The timeline shows the experimental protocol. The effect of postischemic and preischemic dosing with NAD⁺ precursors on brain tissue NAM, NAD⁺ and ATP levels are shown longitudinally over time. Two-way ANOVA followed by Tukey’s multiple comparisons. P values represent the treatment effect. * to ****, p < 0.05 to p < 0.0001 vs. sham; † to ††††, p < 0.05 to p < 0.0001 vs. MCAO + Veh; § to §§§§, p < 0.05 to p < 0.0001 vs. MCAO + NAM Pre; ‡ to ‡‡‡‡, p < 0.05 to p < 0.0001 vs. MCAO + NR Pre; # to ####, p < 0.05 to p < 0.0001 vs. MCAO + NAM Post. NAM: nicotinamide; NR: nicotinamide riboside; Veh: vehicle; HPLC: high-performance liquid chromatography; MCAO: middle cerebral artery occlusion; NAD⁺: Nicotinamide adenine dinucleotide; ATP: adenosine triphosphate.

Focal ischemia significantly reduced both NAD⁺ and ATP levels compared with sham controls when measured 3, 6, and 24 hours after 45-minute MCAO followed by reperfusion (p < 0.0001; Figure 4(b), red vs. blue circles). All treatments significantly raised NAD⁺ levels compared with vehicle after MCAO, and there was no difference among the treatments (p = 0.4750–0.9996; Figure 4(b), left panel). Treatment effects on ATP, however, were not congruent with the NAD⁺ response (Figure 4(b), right panel). Overall, the postischemic treatment improved tissue ATP, and once again NR had a stronger and longer-lasting effect on ATP compared with NAM. In contrast, preischemic treatment was completely ineffective. In fact, preischemic NAM resulted in worse tissue ATP levels than in the vehicle arm. Altogether, NAM and NR treatment effects on tissue ATP after transient MCAO mirrored the MCAO outcomes (Figure 1) and provided a metabolic explanation for the discordance between preischemic and postischemic treatment regimens.

Metabolic regulators

Sirtuin-1, PGC1α, and AMP-activated protein kinase (AMPK) are energy sensors and posttranslational regulators of the metabolic response to hypoxic and ischemic challenges.^21–27 Sirtuin-1 is a critical regulator of multiple cell processes via deacetylation and has been implicated in neuroprotection.^28–31 Sirtuin-1 consumes and converts NAD⁺ to generate NAM, which may, in turn, exert feedback inhibition on sirtuin-1. ³² In contrast, NR does not inhibit sirtuin-1. To test whether a complex modulation of sirtuin-1 by NAM or NR could explain the divergent stroke outcomes with postischemic versus preischemic NAD⁺ enhancers, and the relative efficacies of NAM and NR, we measured sirtuin-1 expression and activity in the brain. In naïve brains, we found that a single dose of NAM or NR induced sirtuin-1 protein expression by ∼3-fold when measured 6 hours later (Figure 5(a)). In contrast, a three-dose regimen did not. Despite increased expression, sirtuin-1 activity was not elevated by a single dose of NAM and was, in fact, suppressed after three doses. This effect of NAM on sirtuin-1 activity was also confirmed using acetyl-lysine as a reporter (Supplementary data 4). Three once-daily doses of NR also suppressed sirtuin-1 activity. Focal ischemia by itself did not change sirtuin-1 expression, although the activity was suppressed (Figure 5(b)). A single postischemic dose of NAM and NR, and three once-daily doses of NR, but not NAM, increased sirtuin-1 expression. Once again, sirtuin-1 activity did not mirror the expression levels. Overall, a single dose yielded higher sirtuin-1 expression and activity than three once-daily doses. These data suggested that, unlike NR, NAM exerts an inhibitory effect on sirtuin-1, in vivo, and that longer exposure to NAD⁺ enhancers diminishes the impact on sirtuin-1, consistent with a cytoprotective role for this enzyme in ischemic stroke.^29,30,33 Of note, neither NAM nor NR altered the expression or activation of poly-ADP-ribose-polymerase (PARP)-1 (Supplementary data 5), an enzyme that cleaves NAD⁺ to add mono- or poly-ADP ribose chains to nuclear proteins in response to cellular stress and worsens focal cerebral ischemic outcomes. ³⁴

Figure 5.

The effect of NAD⁺ precursors on sirtuin-1. (a) Naïve/nonischemic brains: The timeline shows the experimental protocol. The effect of single or multiple dosing with NAD⁺ precursors (mimicking postischemic and preischemic treatment paradigms, respectively) on brain tissue sirtuin-1 protein expression and activity are shown and (b) Ischemic brains: The timeline shows the experimental protocol. The effect of postischemic and preischemic dosing with NAD⁺ precursors on brain tissue sirtuin-1 protein expression and activity are shown. One-way ANOVA followed by Dunnett’s multiple comparisons. NAM: nicotinamide; NR: nicotinamide riboside; MCAO: middle cerebral artery occlusion; Sirt1: sirtuin-1.

Another key modulator of cell survival under hypoxic-ischemic conditions, PGC1α, may not only be an upstream inducer of NAD⁺ biosynthesis but also a downstream effector of NAD⁺. ²⁵ We, therefore, examined how NAD⁺ enhancers altered PGC1α expression. In naïve brains, NAD⁺ enhancers strongly tended to increase PGC1α expression measured 6 hours after the last dose; a single dose appeared more efficacious than three once-daily doses (Figure 6(a)). Transient focal cerebral ischemia did not significantly change PGC1α expression (Figure 6(b)). A single dose of NAM or NR administered 20 minutes after reperfusion potently induced PGC1α expression measured 6 hours after MCAO onset. In contrast, three once-daily doses were completely ineffective. These data were congruent with the efficacy of single-dose postischemic over daily preischemic treatment on sirtuin-1 as well as stroke outcomes.

Figure 6.

The effect of NAD⁺ precursors on PGC1α. (a) Naïve/nonischemic brains: The timeline shows the experimental protocol. The effect of single or multiple dosing with NAD⁺ precursors (mimicking postischemic and preischemic treatment paradigms, respectively) on brain tissue PGC1α protein expression is shown and (b) Ischemic brains: The timeline shows the experimental protocol. The effect of postischemic and preischemic dosing with NAD⁺ precursors on brain tissue PGC1α protein expression is shown. One-way ANOVA followed by Dunnett’s multiple comparisons. NAM: nicotinamide; NR: nicotinamide riboside; MCAO: middle cerebral artery occlusion; PGC1α: peroxisome proliferator-activated receptor-gamma coactivator 1-alpha.

The AMPK is a critical sensor and regulator of energy metabolism and a downstream effector of sirtuin-1. Activating AMPK via phosphorylation shifts energy metabolism to a catabolic state both acutely and longer-term via gene expression. Inhibition of AMPK activity has been shown to improve and stimulation of it worsens focal ischemic outcomes in mice. ³⁵ We first confirmed the latter by showing that AMPK activator 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR, 50 µg intracerebroventricular; Supplementary data 6) worsens stroke outcome when administered 1 hour before MCAO (Supplementary data 7), and then examined whether NAM or NR altered AMPK expression or phosphorylation. In naïve brains, a single dose of NAM or NR strongly increased AMPK phosphorylation (pAMPK) measured 6 hours after the treatment (Supplementary data 8 A). Immediate postischemic administration of NAM or NR also increased pAMPK (Supplementary data 8B). In both naïve and ischemic brains, longer treatment (i.e., three doses over two days) appeared less effective. Altogether, the results implicated AMPK activation in the detrimental effect of preischemic NAM and NR treatment. Because, by definition, postischemic treatment was administered after reperfusion was achieved and oxygen delivery restored, AMPK activation likely did not have a detrimental effect.

Neuronal cultures

Lastly, we tested the effect of NAD⁺ enhancers on neuronal cultures (Supplementary data 9). In naïve cultures, prolonged exposure (2 days) to either NAM or NR caused loss of viability; short exposure (6 hours) was ineffective except for the highest NR concentration (20 mM). Prolonged exposure to NAM or NR for two days before OGD did not rescue the viability, whereas short exposure (6 hours) starting at OGD onset was neuroprotective. These data were congruent with the MCAO outcomes after prolonged preischemic versus short postischemic treatment and highlighted the potentially detrimental effect of prolonged exposure to NAD⁺ enhancers, and suggested that the effect is cell-autonomous.

Discussion

Our data show that NAD⁺ augmentation by NAM and NR can be protective in ischemic stroke when administered after reperfusion is established, whereas longer preischemic exposure conditions the energy metabolism unfavorably for a subsequent cerebral ischemic challenge. Data suggest that treatment with NAD⁺ precursors known to enhance mitochondrial oxidative phosphorylation may diminish the resilience of brain tissue to hypoxia and ischemia. Such diametrically opposed outcomes with preischemic versus postischemic treatment paradigms underscore the need for context-dependent recruitment of metabolic pathways.

The improved outcomes upon postreperfusion treatment with NAD⁺ precursors confirm previous work^5,8,9,14 and might reflect more efficient utilization of restored O₂ supply by still-functioning mitochondria, as evident in improved tissue ATP levels. The effect of these precursors persisted in the OGD cellular model, suggesting that the beneficial effects of NAD⁺ augmentation are likely autonomous to neurons. Enhanced PGC1α expression, sirtuin-1 expression and activity, and AMPK expression and activation all aligned well with the protective effect of postreperfusion treatment with NAD⁺ precursors. Decreased reperfusion injury via diminished free radical production and oxidative stress by dysfunctional mitochondria could also have contributed to favorable outcomes. ³⁶ Our data suggest that short-term post-ischemic treatment with NAD⁺ precursors remains a promising candidate for acute ischemic stroke.

The worse outcome after only two days of preischemic treatment was unanticipated. Faster AD onset upon MCAO in this cohort likely reflected a rapid failure of Na⁺/K⁺ ATPase in the setting of lower ATP levels. However, the mechanism for the latter is less clear and likely complex. Sirtuin-1 activity was diminished by preischemic NAM but not NR, PGC1α, and PARP-1 expression did not differ from vehicle controls in either NAM or NR arm, and AMPK expression and activation did not appear to differ between preischemic and postischemic treatment cohorts. Therefore, tissue and neurological outcomes did not align well with metabolic markers of mitochondrial and energy homeostasis. Notably, several studies have linked excess activation of neuronal AMPK to neurodegenerative conditions,^37–39 raising the possibility of pathogenic effects of AMPK induction obscuring a positive effect on ATP. Indeed, NAD⁺ can activate the AMPK, ²³ which, in turn, can promote NAD⁺ homeostasis. ²¹ Any effect on other organs (e.g., liver) might also indirectly contribute to stroke outcomes. Of note, although we used previously published rodent doses of NAD⁺ enhancers, they are significantly higher than the daily recommended human doses, as is commonly the case due to species differences in pharmacokinetics. Moreover, it is well recognized that the public use of over-the-counter nutritional supplements often exceeds recommended daily doses. Clinically, a wide range of doses have been tested, and doses as high as 1–5 g/day have been safe upon long-term use in humans.^40–43 Therefore, we believe the doses tested here are clinically relevant. It should also be noted that we have not comprehensively examined other therapeutic paradigms (e.g., pre-ischemic treatment onset with continued treatment after MCAO, treatment onset during MCAO, prolonged post-ischemic treatment) that may yield different outcomes.

Ischemia-reperfusion injury to metabolically active tissues has been attributed in part to derangements in oxidative phosphorylation, leading to the generation of mitochondrially derived reactive oxygen species. ⁴⁴ Hence, enhanced reliance on oxidative phosphorylation at rest might also predispose to worse ischemia-reperfusion injury. Nevertheless, worse outcomes with preischemic treatment with NAD⁺ precursors that augment oxidative phosphorylation were consistent with our previous work showing that a preischemic shift away from oxidative phosphorylation improved the outcomes. ¹ In the latter study, we similarly pretreated the animals with the metabolic toggle meclizine for three days before MCAO and found smaller infarcts. Indeed, data herein showing favorable outcomes in the constitutional PGC1α knockout may reflect reduced reliance on mitochondrial oxidative phosphorylation and thus be consistent with the overall model.

The unique roles of PGC1α in the brain merit further investigation. First, it is noteworthy that ischemia-reperfusion injury applied elsewhere, such as the kidneys, suggest that PGC1α^−/− deletion exacerbates rather than ameliorates physiological outcomes relative to control littermates. ²⁰ This difference may reflect the kidney’s preference for fatty acid fuel substrates, the ongoing uptake of which in low mitochondrial settings may contribute to the toxic accumulation of intracellular triglycerides. ⁴⁵ Second, neuronal expression of PGC1α has been linked to protection from oxidant-induced neurodegeneration through a mechanism involving the induction of antioxidant enzymes. ²⁶ Lastly, repression of PGC1α appears to be a pathogenic contributor to experimental models of Parkinson’s disease.^46,47 In this context, the present results suggest that while a metabolic shift away from oxidative phosphorylation may be effective immediately pre-ischemia, PGC1α may signal long-term neuroprotective pathways as well.

We also noted that NR appeared to have a superior therapeutic profile than NAM (e.g., neurological score, tissue NAD⁺ and ATP after MCAO). There are a few potential explanations.^48,49 First, NAM is a known inhibitor of sirtuins, which regulate broad aspects of cellular signaling and homeostasis and are cytoprotective. In contrast, NR does not inhibit sirtuin. Therefore, we speculate that NR’s greater efficacy reflects a “pure” NAD⁺ augmenting effect, whereas NAM’s NAD⁺ augmentation is partially offset by sirtuin inhibition. Second, the conversion of NR to NAD⁺ is more efficient than NAM, and NR has superior pharmacokinetic properties compared with other NAD⁺ precursors. Third, NR may have additional mechanisms of neuroprotection over NAM.

We show that NAD⁺ precursor nutritional supplements known to augment oxidative phosphorylation in the brain may sensitize the tissue to subsequent ischemic challenges, akin to chemical preconditioning, albeit one that is unfavorable for ischemic stroke. Regardless of the mechanisms, these data have previously unrecognized implications for individuals who take chronic high doses of the over-the-counter NAD⁺ precursor nutritional supplements who are also at high risk of coincident superimposed cerebral ischemic events (e.g., atrial fibrillation, atherosclerosis, small vessel disease, hypercoagulable state). Further investigation of the metabolic responses to NAD⁺ augmentation in the healthy brain and following other kinds of acute cerebral insults (e.g., hemorrhage, trauma) are clearly warranted.