INTRODUCTION

      Parkinson’s disease (PD) affects 1%–2% of the population above the age of 65 and is a major cause of death and disability, with a rapidly growing global socioeconomic impact (Gooch et al., 2017; de Rijk et al., 2000). Current treatments for PD can provide partial symptomatic relief, mainly for motor symptoms but make no substantial impact on disease progression (Athauda and Foltynie, 2015; Bloem et al., 2021; Kalia and Lang, 2015). Despite several candidate neuroprotective therapies showing encouraging preclinical results, these have failed to show disease-modifying effects in clinical trials (Athauda and Foltynie, 2015; Espay et al., 2017).

      A growing body of evidence supports that boosting cellular levels of nicotinamide adenine dinucleotide (NAD) may confer neuroprotective effects in both healthy aging and neurodegeneration (Lautrup et al., 2019). NAD is reversibly interconverted between its oxidized (NAD+ ) and reduced (NADH) state and thereby constitutes an essential cofactor for metabolic redox reactions, including mitochondrial respiration. Furthermore, NAD+ is substrate to vital signaling reactions involved in DNA repair, histone and other protein deacylation reactions, and second messenger generation (Katsyuba et al., 2020). These reactions consume NAD+ at high rates, requiring constant replenishment via NAD biosynthesis. NAD levels have been shown to decline with age and this is believed to contribute to age-related diseases (Johnson and Imai, 2018; Katsyuba et al., 2020). Increasing the NADreplenishment rate (e.g., via supplementation of precursors) and/ or enhancing the NAD+ /NADH ratio (e.g., via caloric restriction) have shown beneficial effects on lifespan and healthspan in multiple model systems and provided evidence of neuroprotection in animal models of neurodegeneration and other agerelated diseases (Braidy and Liu, 2020; Johnson and Imai, 2018; Katsyuba et al., 2020)

      Enhancing NAD replenishment could potentially help ameliorate several major processes implicated in the pathogenesis of PD, including mitochondrial respiratory dysfunction (Flønes et al., 2018; Hattori et al., 1991; Schapira et al., 1989), neuroin-flammation (Hirsch and Hunot, 2009), epigenomic dysregulation (Park et al., 2016; Toker et al., 2021), and increased neuronal DNA damage (Gonzalez-Hunt and Sanders, 2021).

      NAD can be replenished via supplementation of nicotinamide riboside (NR), a vitamin B3 molecule and biosynthetic precursor of NAD (Bieganowski and Brenner, 2004; Katsyuba et al., 2020). NR has undergone extensive preclinical testing (Conze et al., 2016) and is well tolerated by adult humans, showing no evidence of toxicity with doses up to at least 2,000 mg daily (Dollerup et al., 2018). Trials in healthy individuals have shown that an oral intake of 1,000 mg NR daily substantially elevates total levels of NAD and related metabolites in blood and muscle, boosts mitochondrial bioenergetics, and decreases circulating inflammatory cytokines (Conze et al., 2019; Elhassan et al., 2019; Martens et al., 2018; Trammell et al., 2016). Moreover, evidence from cell and animal studies suggests that NR supplementation promotes health span and has neuroprotective effects in models of Cockayne syndrome (Okur et al., 2020), noiseinduced injury (Brown et al., 2014; Han et al., 2020), amyotrophic lateral sclerosis (Harlan et al., 2020), Alzheimer’s disease (Sorrentino et al., 2017; Xie et al., 2019), and PD (Scho¨ndorf et al., 2018).

      Taken together, this evidence suggests that NR may hold promise as a potential neuroprotective agent for PD. However, several critical knowledge gaps have yet to be addressed. Specifically, it needs to be established whether NR is well tolerated, augments cerebral NAD levels, and affects cerebral metabolism in patients with PD.

      To address these questions, we conducted a double-blinded, randomized, placebo-controlled phase I study of NR in newly diagnosed PD patients, naive to dopaminergic therapy (the NADPARK study, Clinicaltrials.gov: NCT03816020). Primary outcomes were the detection of cerebral penetration and metabolic response, measured by 31phospohorous-magnetic resonance spectroscopy (31P-MRS), cerebrospinal fluid (CSF) metabolomics, and 18fluoro-deoxyglucose positron emission tomography (FDG-PET). Secondary outcomes were the safety and tolerability of NR in the study population (measured by the frequency and severity of adverse events and changes in vital signs and clinical laboratory values), the impact of NR on clinical symptoms (measured using the Movement Disorder Society unified Parkinson’s disease rating scale [MDS-UPDRS]), and the impact of NR on the NAD metabolome in peripheral tissues of PD patients. We found that NR supplementation was well tolerated with no adverse effects. Treatment increased brain NAD levels, as well as related metabolites in the CSF and peripheral tissues. At the individual level, the increase in cerebral NAD was not universal but evident in 10/13 NR recipients from whom data were available. The NR recipients who showed increased brain NAD levels exhibited altered cerebral metabolism, inducing a specific treatment-related metabolic network that overlapped with key elements of the PD-related spatial covariance pattern (PDRP) (Schindlbeck and Eidelberg, 2018) and were associated with mild clinical improvement.

      Our findings suggest that NR may be useful as a neuroprotective therapy in PD, pending further investigation in a phase II study.

RESULTS

      In total, 36 patients were screened, and 30 eligible patients were enrolled, all of whom completed the study (Figure S1). There were no significant demographic differences between the NR and placebo group (Table 1). Seven subjects in the NR group and three subjects in the placebo group reported adverse events, all of which were minor and considered to be unrelated to NR (Table S1). Based on the number of remaining capsules in returned vials, we estimated a mean study drug adherence in the NR and placebo group to be 98%. The average time on study drug was 32.5 (±2.7) days for the NR group and 32.4 (±2.53) days for the placebo group (Table S2). Four participants (two in the NR and two in the placebo group) did not return any capsules but reported they were drug compliant.

NR increases cerebral NAD levels

      31P-MRS was employed to assess phosphorylated metabolites in the patients’ brain. At the end of the study, data were available from 27 individuals, including 13 in the NR group and 14 in the placebo group (Table S2). 31P-MRS data analyses allowed for identification and quantitation of multiple phosphorylated compounds (Figures 1A–1C; Table S3). NAD levels, which were normalized to ATP-a, showed a significant increase in the NR group (paired t test, p = 0.016), but not in the placebo group (Figure 1D). Direct comparison between the groups showed that the between visit change (visit 2/visit 1) in the NAD/ATP-a ratio was significantly higher in the NR group compared with the placebo group (t test, p = 0.025). No other detected cerebral metabolites exhibited significant change in either the NR or placebo group between visits. At the individual level, the cerebral NAD response was heterogeneous, with 10/13 patients showing an increase, of whom 9 showed a change exceeding 10% of baseline levels. We will henceforth refer to the subgroup of 10 NR recipients who showed an increase in cerebral NAD levels as MRS responders. Since this variability raised the possibility of heterogeneous treatment effects, we chose to stratify downstream neurometabolic and clinical analyses.

NR induces a novel treatment-related metabolic network

      We next interrogated the FDG-PET data of the NR recipients to determine whether treatment-related increases in cerebral NAD levels were associated with a significant metabolic brain network. To this end, we applied a supervised principal component analysis (PCA) algorithm (ordinal trends/canonical variates analysis [OrT/CVA]) to paired scan data from the NR participants (n = 10) for whom brain NAD levels were simultaneously increased with treatment. The analysis revealed a significant ordinal trend pattern, which was represented by the first principal component (PC1), accounting for 20.6% of the variance in the paired data. This NR-related metabolic pattern (NRRP; Figures 2A and S2) was characterized by bilateral metabolic reductions in the caudate and putamen, extending into the adjacent globus pallidus, and in the thalamus (Table S4). As part of the network, these changes were also associated with localized cortical reductions along the medial wall of the hemisphere involving the precuneus (Brodmann area [BA] 7), medial frontal cortex (BA 9 and 10), anterior cingulate area (BA 24 and 32), and in the posterior cingulate gyrus (BA 31). The reliability of these regions was demonstrated by bootstrap iteration (inverse coefficient of variation [ICV] z =
3.83,
2.58, p < 0.005; 1,000 iterations). Areas with relative metabolic increases did not contribute significantly to this topography.

      At the individual subject level, a significant ordinal trend in NRRP expression was observed (p < 0.018; permutation test, 1,000 iterations) in the 10 NR recipients in the pattern identification group (Figure 2B, black lines; Figure 2C). Although analogous NR-mediated increases in pattern expression were not seen in the remaining 4 NR recipients who either did not exhibit NAD responses on MRS or for whom MRS data were not available (Figure 2B, gray lines), the ordinal trend remained significant for the NR group as a whole (p = 0.027). That said, consistent changes in NRRP expression were not significant in the placebo group (p = 0.497). Significant correlations were not observed between treatment-related changes in cerebral NAD levels and NRRP expression in either of the trial groups (p > 0.10).

      Nevertheless, changes in NRRP expression in the NR group correlated significantly (r =
0.59, p = 0.026) with the changes in UPDRS motor ratings recorded at the time of PET (Figure 2D; Table S2). Thus, the largest NRRP increases were observed in the NR subjects with the greatest improvement in motor ratings. An analogous correlation was not seen in the placebo group (p = 0.79). We noted that certain NRRP regions were shared with the previously validated PDRP topography (Figure 2E, top; Figure S2, middle). In particular, areas of reduced activity in the caudate and putamen, extending into the adjacent globus pallidus, overlapped with corresponding areas of overactivity in the PDRP (Figure 2E, bottom, hot voxels; Figure S2). Likewise, spatial overlap was observed between underactive regions in the frontal and parietal cortex for the two patterns (Figure 2E, bottom, purple voxels; Figure S2, bottom). Indeed, a significant correlation was observed between baseline PDRP and NRRP expression values across the trial population (r = 0.56, p = 0.002; Figure 2F). As typically seen in PD populations, PDRP expression values at baseline correlated with UPDRS motor ratings (r = 0.374, p = 0.046 for the two groups; Figure 2G). In contrast to NRRP in which the changes in pattern expression with NR correlated with clinical improvement (Figure 2D), no correlation with outcome was observed for the PDRP changes (p = 0.73).

NR-induced increase in cerebral NAD is associated with clinical improvement of PD

      No significant change was found in the MDS-UPDRS (total or subsections I–III) in the NR or placebo groups. However, a trend for decreased MDS-UPDRS was seen in the MRS-responder subgroup (mean decrease 1.9 ± 2.78; paired t test, p = 0.071), and this reached statistical significance when only the 9 individuals showing >10% increase in cerebral NAD levels were considered (mean decrease 2.33 ± 2.35; paired t test, p = 0.017). Upon closer inspection, this effect appeared to be mainly driven by subsections I and III of the MDS-UPDRS. The clinical scores of the participants at each visit are shown in Tables 2 and S2.

NR supplementation augments the NAD metabolome in the CSF and peripheral tissues of PD patients

      Metabolomic analyses were carried out in CSF, skeletal muscle, and peripheral blood mononuclear cells (PBMCs) to confirm NR intake and investigate potential changes in related metabolic pathways. In CSF, we detected a substantial increase of the nicotinamide (Nam) degradation product N-methyl-2-pyridone- 5-carboxamide (Me-2-PY), an established indicator of NR supplementation (Elhassan et al., 2019; Trammell et al., 2016) (Figure 3) in all NR recipients. Other related metabolites, including NAD+ , NADH, nicotinamide mononucleotide (NMN), nicotinic acid adenine dinucleotide (NAAD), and methyl nicotinamide (Me-Nam), were below the detection limit in the CSF. This finding further corroborated that oral NR therapy increases brain NAD levels across the blood brain barrier, as indicated by the 31PMRS data. In muscle tissue, several NAD-related metabolites, including Nam degradation products nicotinamide N-oxide (Nam N-oxide), Me-Nam, and the methyl pyridones (Me-2-PY) and N-methyl-4-pyridone-5-carboxamide (Me-4-PY), as well as the acid form of NAD, NAAD, were strongly elevated after treatment with NR (Figure 3). The steady-state levels of NAD itself (both oxidized NAD+ and reduced NADH), and NAD precursors and intermediates, were not significantly changed by NR treatment (Figure S3), similar to reports from previous trials in humans (Elhassan et al., 2019). PBMCs showed less extensive changes, recapitulating the increase in NAAD and Me-Nam (Figures 3 and S4). Additional metabolites that could potentially be affected by changes in the NAD metabolome were investigated, including acetyl-CoA and CoA, the methyl donor S-adenosylmethionine (SAM) and its metabolites, and energy compounds, such as ATP. These showed no significant change in PBMCs or muscle (Figures S3 and S4). Importantly, substantial metabolic changes were seen in the entire NR group in all tissues, irrespective of the cerebral NAD response in 31P-MRS.

NR therapy induces the expression of mitochondrial, antioxidant, and proteostatic processes

      To investigate the effects of NR therapy on gene expression, we carried out RNA-seq analysis in PBMCs and muscle tissue from all study participants and assessed the between visit differences in the NR group compared with the placebo group. In muscle tissue, NR supplementation was significantly associated with differential expression of 58 genes (FDR < 0.05; Table S5). These included substantial upregulation of KLF2, which is associated with decreased adipogenesis and induction of the nuclear factor erythroid 2-related factor 2 (Nrf2), a transcription factor with a key role in protection against oxidative damage, which has been linked to PD (Petrillo et al., 2020). We also noted upregulation of genes linked to NAD-degradation, such as ADPribosylation (PARP15), and NAD-dependent redox processes, such as the glycine cleavage system (AMT), as well as mitochondrial translation and respiratory complex assembly (FARS2, TMEM242). Gene set enrichment analyses revealed NR-induced upregulation of biological processes, including proteasomal function, and RNA transport and stability. In PBMCs, a total of 13 genes were significantly associated with NR supplementation, including upregulation of BLOC1S2, a component of the BLOC-1 complex involved in lysosomal biogenesis and trafficking (Starcevic and Dell’Angelica, 2004). Functional enrichment revealed highly significant upregulation of multiple biological processes, including ribosomal, proteasomal, lysosomal, and mitochondrial (oxidative phosphorylation) pathways. A complete account of NR-regulated pathways is provided in Table S6. No significant changes were found in genes encoding known NAD biosynthetic enzymes, including NMRK1, encoding nicotinamide riboside kinase 1, which converts NR to NMN. Finally, NR treatment did not show a significant association with estimated cell type proportions in the PBMC samples (p > 0.5, visit:treatment interaction term, linear mixed model, Welch-Satterthwaite’s t test, p < 0.05; Figure S5).

NR modulates biomarkers of inflammation and mitochondrial dysfunction

      Since mitochondrial dysfunction (Flønes et al., 2018; Hattori et al., 1991; Schapira et al., 1989) and inflammation (Hirsch and Hunot, 2009) have been associated with the pathophysiology of PD, we sought to identify NR-induced changes in relevant biomarkers in the serum and CSF of our patients. We assessed the levels of growth factors FGF21 and GDF15, which have been associated with mitochondrial dysfunction (Boenzi and Diodato, 2018; Lehtonen et al., 2016), and a panel of 35 inflammatory cytokines. Growth factor analysis revealed a mild but significant decrease of GDF15 levels in the serum, but not in the CSF. FGF21 was unchanged in the serum and was below detection limit in CSF. Several inflammatory cytokines decreased in the serum and/or CSF of the NR-treated group (Figures 4 and S6). Notably, several of the serum cytokines also showed a significant decrease in the placebo group. Comparing the fold change of those cytokines between the groups revealed no significant difference (Figure S6A). In addition, we assessed the levels of neurofilament light chain (Nf-L), which are indicative of neuronal damage. These were unchanged by NR treatment in both the serum and CSF.

DISCUSSION

      We present the results of the first NR trial in PD. Our study met its primary outcome, which was to assess penetration and metabolic responses of the brain in patients with PD. We established that orally administered NR is safe, leads to an increase in cerebral NAD levels, and alters cerebral metabolism in patients with PD. Cerebral penetration was further validated by detecting an increase of the metabolite Me-2-PY in the CSF of participants receiving NR, but not in the placebo. Furthermore, NR augmented the NAD metabolome and increased the expression of genes related to mitochondrial, lysosomal, and proteasomal function in blood and muscle and decreased the levels of inflammatory cytokines in serum and CSF.

      Although a significant increase in cerebral NAD levels was detected by 31P-MRS in the NR group, this effect was not universal. Three patients showed no evidence of cerebral NAD increase, despite a clear metabolic response in CSF, blood, and muscle, thereby confirming treatment compliance and an impact on the NAD metabolome. The fact that we detected an increase of Me-2-PY in the CSF of all NR recipients suggests that central nervous system penetration was universal. Thus, the variable cerebral NAD response observed by 31P-MRS is likely to reflect interindividual variability in cerebral NAD metabolism and, possibly, limited sensitivity of the NAD measurement by MRS (i.e., MRS non-responders may have a positive, but mild cerebral NAD increase, which is below the detection limits of the method). Irrespective of the mechanisms underlying this variable response, our findings indicate that assessment of cerebral NAD levels may be an important monitoring parameter in clinical trials evaluating NR supplementation for brain health and disease. Variability in the achieved NAD increase in the patient brain raises the possibility of a heterogeneous biological and, potentially, clinical response to NR therapy. Increase in cerebral NAD levels was indeed associated with both neurometabolic and clinical response in our patients. Notably, a trend for clinical improvement, in the form of decreased MDS-UPDRS score, was seen in the subgroup of NR recipients who had an increase in cerebral NAD levels by 31P-MRS.

      FDG-PET analyses identified the NRRP—a novel network topography that was induced in PD trial participants receiving NR, but not placebo, under blinded conditions. Apart from its defining ordinal trend, changes in pattern expression with NR correlated with clinical improvement as measured by reductions in MDS-UPDRS ratings. The NRRP consisted mainly of decreased glucose uptake in the basal ganglia and neocortical areas. This may reflect a more efficient bioenergetic state, requiring less glucose consumption to maintain required ATP production. Increased NAD+ levels would promote fatty acid beta-oxidation (Akie et al., 2015; Guarino and Dufour, 2019; Mukherjee et al., 2017), thereby providing an increased supply of reducing equivalents (i.e., NADH, FADH2) to the respiratory chain independent of glycolysis. While the organization of the NRRP remains unknown at the systems level, the network shares a number of topographic features with the PDRP, which support a potential role for NR in PD therapeutics. The overlap between metabolically active PDRP regions in the posterior putamen and globus pallidus and areas of treatment-related metabolic reduction in the NRRP is particularly relevant. In recent graph theoretic studies of PDRP organization, we found that these regions comprise a discrete core zone, which drives overlap network activity. Metabolic reductions in core nodes have been observed in symptomatic PD treatments, such as levodopa administration and subthalamic deep brain stimulation (Asanuma et al., 2006; Niethammer and Eidelberg, 2012). The presence of similar changes with NR suggests that these regions may also be modulated by this treatment. In any event, clinical outcomes in the NR group relate more to the induction of NRRP than to the modulation of the PDRP activity—a situation similar to that encountered with subthalamic gene therapy (Niethammer et al., 2018).

      Metabolomic analyses in PBMCs and muscle confirmed NR intake in the treatment group and excluded this in the placebo group. Notably, a highly significant increase in established markers of NR-mediated NAD biosynthesis, such as NAAD and Me-Nam, was seen in all patients in the NR group, corroborating previous studies in healthy subjects (Elhassan et al., 2019). Measured NAD+ values were substantially lower than previously reported concentrations in these tissues (Elhassan et al., 2019; Lamb et al., 2020; Martens et al., 2018; Trammell et al., 2016); however, we did not detect an increase in NAD+ levels (Figures S3 and S4). It is conceivable that a new steady state of increased NAD flux may be established over time in muscle and blood cells without detectable changes in the measured NAD concentrations at any given time point. This could be due to increased NAD consumption combined with increased NAD biosynthesis, facilitated via continuous NR supply. Evidence of increased flux in the NAD metabolome was indeed detected in both PBMCs and muscle of our patients, indicating higher NAD availability in the NR group. Notably, unchanged NAD levels in muscle were also seen in a recent study of NR in healthy adults (Elhassan et al., 2019), whereas another study detected elevated NAD+levels in PBMCs (Trammell et al., 2016).

      In addition to the NAD metabolome, we investigated other metabolites, which are functionally linked to NAD metabolism and signaling, and may be affected by NR treatment. Synthesis of Me-Nam, which was greatly increased by NR supplementation, requires the methyl donor SAM. This, in turn, could limit SAM availability for other essential methylation reactions, such as DNA and histone methylation, and neurotransmitter synthesis, including dopamine (Jadavji et al., 2015). We found no significant changes in SAM or its related metabolites SAH and homocysteine, indicating that NR supplementation does not limit SAM availability for other vital reactions.

      Another metabolite of interest was acetyl-CoA, donor of acetyl-groups for acetylation reactions, including histone acetylation (Pietrocola et al., 2015). Higher NAD availability could lead to increased acetyl-CoA synthesis by promoting glucose and fatty acid catabolism. This could, in turn, exacerbate the histone hyperacetylation state observed in the PD brain (Park et al., 2016; Toker et al., 2021), further dysregulating gene expression. Our analyses detected no significant changes in acetyl-CoA or CoA levels upon NR supplementation. Finally, we monitored energy metabolites, such as ATP, ADP, AMP, as well as GTP and GDP. None of these showed a significant change upon NR supplementation.

      The downstream metabolic impact of NR treatment was supported by the transcriptomic analyses in PBMC and muscle, which revealed effects in multiple disease-relevant pathways. Notably, NR was associated with upregulation of genes involved in mitochondrial respiration, antioxidant response, and protein degradation, including the proteasome and lysosome. Quantitative and functional respiratory deficiency (Flønes et al., 2018; Schapira et al., 1989), increased oxidative damage (Dias et al., 2013; Jenner, 2003), and impaired proteasomal and lysosomal function (Lehtonen et al., 2019) have all been strongly implicated in the pathophysiology of PD. Therefore, these findings are encouraging, and they support a potential neuroprotective effect of NR in PD. Furthermore, our results indicate that NR may have anti-inflammatory action by decreasing the levels of inflammatory cytokines not only peripherally as previously shown in healthy aged individuals (Elhassan et al., 2019) but also in the central nervous system. Interestingly, almost all serum cytokines showing a significant decrease in the NR group also decreased in the placebo group, making the significance of these findings uncertain. In contrast, several CSF cytokines showed a significant decrease exclusively in the NR-treated group, suggesting a potential anti-inflammatory effect in the nervous system. Neuroinflammation has been implicated in the pathogenesis of PD, and it is considered a potential target for neuroprotection (Rocha et al., 2015).

      Mitochondrial biomarker analyses revealed a mild, but significant, decrease of GDF15 in the serum, but not CSF, of NR recipients. This is in line with the observed transcriptomic upregulation of mitochondrial genes and could be indicative of improved mitochondrial function in our patients. Nf-L levels were not significantly affected by NR treatment in serum or CSF. However, Nf-L is only mildly elevated in PD (especially in early stages of the disease) compared with healthy controls and is better suited as a biomarker for monitoring disease progression over time (Mollenhauer et al., 2020). Therefore, also given the small sample size and short duration of our trial, the lack of Nf-L change is not surprising.

Conclusions

      In conclusion, the study meets its primary outcome and shows that oral NR therapy increases brain NAD levels and impacts cerebral metabolism in PD. Furthermore, our findings suggest that NR may target multiple processes implicated in the pathophysiology of the disease by upregulating the expression of genes involved in mitochondrial respiration, oxidative damage response, lysosomal and proteasomal function, and downregulating inflammatory cytokines in the central nervous system. In addition, it is possible that NR may mitigate epigenomic dysregulation in PD by regulating histone acetylation. Genome-wide histone hyperacetylation and altered transcriptional regulation occur in the brain of individuals with PD (Toker et al., 2021). Increasing neuronal NAD levels would boost the activity of the NAD-dependent histone deacetylases of the sirtuin family, potentially ameliorating histone hyperacetylation in PD. Taken together, our findings nominate NR as a potential neuroprotective agent against PD, which warrants further investigation in a larger trial. To this end, we are currently conducting the phase II trial NOPARK (Clinicaltrials.gov: NCT03568968), aiming to assess whether NR can delay nigrostriatal degeneration and clinical disease progression in patients with early PD.

Limitations of study

      The sample size of this trial was relatively small, although we had adequate power to robustly support the primary and most of the secondary and tertiary outcomes. The observed trend for clinical improvement among the NR recipients with increased cerebral NAD levels should be interpreted with caution due to the low number of subjects, short observation time, and high interindividual variability in MDS-UPDRS scores. While these observations are encouraging, only a phase II trial will provide conclusive clinical evidence on any disease-modifying effects of NR in PD. Although we were able to confidently detect and measure total cerebral NAD levels by 31P-MRS, our analyses were limited by the strength of the magnet used in this study, as well as adhering to normal-mode specific absorption rate (SAR) levels, which did not allow us to confidently discriminate between the oxidized (NAD+ ) and reduced (NADH) forms. A higher field strength or more effective decoupling by increasing SAR deposition limit may make discrimination possible (Lu et al., 2014; Peeters et al., 2019). More detailed analysis of the NAD+ /NADH redox ratio in the brain upon NR treatment would improve our knowledge about the metabolic impacts of NR in the brain. The transcriptomic analyses showed an NR-associated upregulation of genes involved in mitochondrial respiration. Establishing whether this leads to improved mitochondrial respiration would require the assessment of respiratory enzyme activities in the patients’ tissue. Performing these analyses post hoc was not feasible in this trial due to the limited tissue sample availability.

STAR+METHODS

Detailed methods are provided in the online version of this paper

and include the following:

● KEY RESOURCES TABLE

● RESOURCE AVAILABILITY

 Lead contact

 Materials availability

 Data and code availability

● EXPERIMENTAL MODEL AND SUBJECT DETAILS

 Participants and study design

 Approval, registration, and patient consent

 Randomization and masking

● METHOD DETAILS

 Outcomes

 Procedures

 RNA sequencing

 Metabolomic analysis

 Nf-L, GDF15, FGF21, and Cytokine detection

● QUANTIFICATION AND STATISTICAL ANALYSIS

 Sample size

 Statistical analyses

● ADDITIONAL RESOURCES

SUPPLEMENTAL INFORMATION

Supplemental information can be found online at https://doi.org/10.1016/j.cmet.2022.02.001.

ACKNOWLEDGMENTS

We are grateful to our study nurse, Marit Rensa˚, and technical laboratory staff, Hanne Linda Nakkestad, Gry Hilde Nilsen, Dagny-Ann Sandnes, Kibret Yimer Mazengia, Wenche Hauge Eilifsen, and Kristin Paulsen Rye, for excellent technical assistance. We thank Prof. Kjell-Morten Myhr for the fruitful discussions and ChromaDex (Irvine, California) for providing the NR and placebo capsules. We also thank Anne Mathilde Kvammen, Research and Development Department, Haukeland University Hospital; Dr Ann Cathrine Kroksveen, Biobank Haukeland, Haukeland University Hospital; Dr Cecilie B. Rygh, Department of Radiology, Haukeland University Hospital; and Prof. Martin Biermann, Department of Nuclear Medicine, Haukeland University Hospital. We are grateful to the staff at the Department of Neurology, Department of Radiology and the PET-Center, Haukeland University Hospital, for providing essential personnel and infrastructure for the study. This work was supported by grants from The Research Council of Norway (288164), Bergen Research Foundation (BFS2017REK05), and the Western Norway Regional Health Authority (F-11470). The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

AUTHOR CONTRIBUTIONS

Study conceptualization, C.T. and C.D.; design, C.T., B.B., and C.D.; clinical trial execution, B.B., K.H., G.O.S., O.-B.T., and C.T.; data collection, B.B., C.D., K.H., G.O.S., O.-B.T., T.S., N.B., V.S., K.V., L.S., J.D., R.G., and C.T.; data analyses and interpretation, B.B., C.D., F.R., Y.M., G.S.N., A.R.C., T.S., N.B., V.S., K.V., L.S., J.D., S.P., R.G., D.E., M.Z., and C.T.; drafting the manuscript, B.B., C.D., F.R., Y.M., G.S.N., A.R.C., T.S., S.P., D.E., M.Z., and C.T. All authors have read and approved the manuscript.

DECLARATION OF INTERESTS

C.T. and C.D. have filed a patent application relating to the use of NR in PD. All other authors declare no competing interests.

Received: July 15, 2021

Revised: November 17, 2021

Accepted: January 31, 2022

Published: March 1, 2022

This article is excerpted from Cell Metabolism 34, 396–407, March 1, 2022  by Wound World.