WHAT IS NAD+
Scientists have now confirmed a direct link between falling NAD+ levels and aging in both animal and in human subjects.
Read more about NAD+
NAD+ DECLINES WITH AGE
As we age, our bodies produce less NAD+ and the communication between the Mitochondria and cell nucleus is impaired. (5,8,10).
Familiarization to swimming
Rats individually swam in deep water tanks at a water temperature of 34oC. Rats were familiarized using a modified protocol of our group (Goutianos et al., 2016; Veskoukis et al., 2008). The familiarization protocol lasted 11 consecutive days (starting from day 11 of the experiment) and the daily swimming time was 5 min. All familiarization sessions for all animals took place during the same time period (i.e., 9:00-12:00). In day 11, the rats swam free of load in order to avoid exacerbation of the anticipated stress caused by their first contact with water. In day 12, load equal to 1% of the rats’ body weight was adjusted at the base of their tails and it was gradually incremented by 1% each day until day 21 (final load 10%). The rats rested in their cages for one day before the exercise test.
Pre-loaded incremental swimming test (time-to-exhaustion) One day after the last familiarization session and the last NR administration (day 22), the Ex and NR+Ex groups underwent a pre-loaded incremental swimming test between 9:00-12:00. Two hours before the beginning of the test, food supply was removed from the cages. The swimming test consisted of four consecutive incremental phases until exhaustion. The load was adjusted at the base of the tail. Specifically, a load equal to 2% of the rats’ body weight was used for the first 4 min and then loads equal to 3.5% and 5% were used for the next 8 min (4 min each). A final load equal to 10% was used and from that point on the time until exhaustion was recorded. The time required to change the load between each phase was under 10 s. A rat was considered to have reached exhaustion when it exhibited loss of coordinated movements and failure to return to the surface within 10 s three consecutive times.
We would like to clarify the fact that the times reported in Kourtzidis et al., (2016; 94-145 s) are the times to exhaustion starting to count after the 12th min. This means that the duration of the exercise performance test was 12 min plus the 94-145 s. Taking into account the duration of the first three phases (4 min each, 12 min in total) and the loads used (2%, 3.5% and 5% of body weight), the type of exercise was mainly aerobic. However, the load used during the last phase was 10% of body weight, and the duration from that point on was only 94-145 s. This indicates that that after the 12th min the type of exercise switched to mainly anaerobic, since the effort was maximal and until exhaustion. The swimming protocol that was used is based on Veskoukis et al., (2008), properly modified to reduce the inter-individual variability in performance times recorded.
Blood and muscle collection and homogenization
Rats were anesthetized with exposure to isoflurane mixture (30% v/v isoflurane in propylene glycol) in a bell jar. A cotton pad was wetted with 1.0 cc of mixture for every 500 cc volume of the jar, and it was placed under a perforated floor. Criterion for the depth of anesthesia was eye pupil contraction as well as simple sensory tests such as eyelid retraction and bending of the foot after pressure stimulation. Afterwards, animals were sacrificed by exsanguination. Thoracic cavity was opened and whole blood was collected via cardiac puncture in EDTA tubes. This process lasted 90 s. Then, blood was centrifuged immediately at 1370 g for 10 min at 4oC, and then plasma was collected. Gastrocnemius muscle and liver were quickly removed and snap-frozen in liquid nitrogen. Plasma and tissues were then stored at -80oC until analysis. The tissues were ground, using a mortar and pestle, under liquid nitrogen. One part (mg) of muscle or liver powder was then homogenized with 2 parts (mL) of 0.01 mol·L-1 PBS (138 mmol·L-1 NaCl, 2.7 mmol·L-1 KCl, 1 mmol·L-1 EDTA; pH 7.4), and a cocktail of protease inhibitors (1 μmol·L-1 aprotinin, 1 μg·mL-1 leupeptin, 1 mmol·L-1 phenylmethanesulfonyl fluoride). The homogenate was vigorously vortexed, centrifuged at 12,000 g for 3 min at 4oC and the supernatant was collected.
Assays NADPH concentration was measured according to Wagner and Scott (1994). For the extraction of NADPH, 200 μL of the tissue homogenate (diluted 1/3) was added to 1800 μL of extraction buffer (containing 20 mM nicotinamide, 20 mM NaHCO3 and 100 mM Na2CO3). The samples were immediately frozen, quickly thawed in a water bath (22°C) and centrifuged (16,000 g, 30 s, 4°C). Finally, the supernatant (the tissue extract) was collected and kept on ice in the dark. 100 μL of the tissue extract was incubated in the dark (60°C for 30 min) in water bath. After heating, the tissue extract was added to 800 μL of ice-cold freshly made NADP cycling buffer (containing 100 mM Tris–HCl at pH 8.0, 0.5 mM thiazolyl blue tetrazolium bromide, 2 mM phenazine ethosulfate, 5 mM ethylenediaminetetraacetic acid tetrasodium salt and 1.3 μU/mL glucose-6-phosphate dehydrogenase). The mixture was incubated in the dark (37°C for 5 min), and following temperature equilibration, 100 μL of 10
mM glucose-6-phosphate was added. The change in absorbance was monitored at 570 nm for 2 min (Veskoukis et al., 2017). The NADPH concentration is calculated through the Beer–Lambert law using the millimolar extinction coefficient of formazan (13 l/mmol/cm).
Glutathione peroxidase activity was measured according to Flohé and Günzler (1984). 100 μL muscle homogenate (diluted 1/10), 100 μL glutathione reductase and 100 μL GSH were added in 500 μL of 100 mM phosphate buffer pH 7 (containing 200 mM KH2PO4, 200 mM K2HPO4 and 1 mM di-sodium EDTA). Then, 100 μL NADPH solution was added and the mixture was incubated for 3 min at room temperature. The hydroperoxide- dependent reaction was started by adding 100 μL of pre-warmed tert-butyl hydroperoxide. The change in absorbance at 340 nm was monitored for 5 min (Veskoukis et al., 2016). The non-enzymatic reaction rate is assessed by replacing the enzyme sample by buffer. The calculation of GPx activity is based on the molar extinction coefficient of NADPH (6200 l/mol/cm). Glutathione reductase activity was measured according to Smith and colleagues (1988). 250 μL DTNB and 50 μL β-NADPH were mixed with 700 μL of 200 mM phosphate buffer pH 7.8 (containing 400 mM KH2PO4, 400 mM K2HPO4 and 1 mM di-sodium EDTA). Then, 50 μL oxidized glutathione and, in rapid sequence, 25 μL muscle homogenate diluted 1/2 were added. The change in absorbance at 412 nm was monitored for 1 min (Veskoukis et al., 2016). The calculation of GR activity is based on the molar extinction coefficient of TNB (13,600 l/mol/cm). Catalase activity was measured according to Aebi (1984). 40 μL muscle homogenate (diluted 1/2) was added to 2955 μL of 67 mmol·L–1 sodium potassium phosphate (pH 7.4), and the samples were incubated at 37oC for 10 min. 5 μL 30% hydrogen peroxide was added to the samples. The change in absorbance at 240 nm was monitored for 1.5 min (Veskoukis et al., 2016). The determination of catalase activity is based on the molar extinction coefficient of H2O2 (40 l/mol/cm).
Liver and muscle glycogen were measured according to Lo and colleagues (1970). 50 mg powdered tissue, were mixed with 500 μL 30% KOH saturated with Na2SO4. The tubes were incubated in water bath at 100oC for 30 min. Then, the tubes were removed from the water bath and cooled on ice. Then, 600 μL 95% ethanol were added to precipitate the glycogen and the samples were kept on ice for 30 min and centrifuged (840g, 4°C, 25 min). The supernatant was removed and the glycogen pellet was resuspended with 600 μL distilled H2O. 150 μL of sample was collected and added into eppendorf tubes. Then, 100 μL phenol 5% and 850 μL H2SO4 (96-98%) were added in rapid sequence. The mixture was incubated for 15 min in water bath (25°C). Τhe change in absorbance at 490 nm was monitored for 1 min. Liver and muscle glycogen concentration was calculated via a standard curve of solutions with known glycogen concentrations (10-200 μg/ml for liver and 0.1-50 μg/ml for skeletal muscle). All the aforementioned measurements were performed using a Hitachi U-1900 spectrophotometer (Hitachi High-Technologies Corporation, Tokyo, Japan).
Measurement of total protein
The activities of the enzymes and the concentrations of the molecules in tissues are expressed in terms of the total protein concentration. The total protein concentration of tissue samples was measured using the Bradford method via a standard curve of solutions with known bovine serum albumin concentrations.
Statistical analysis The normality of distribution of all dependent variables was examined and verified by the Kolmogorov–Smirnov test. A two-way (supplementation × exercise) analysis of variance (ANOVA) was used. Significant interaction effects were followed by Tuckey’s post hoc test to locate the significantly different means. The level of statistical significance was set at P<.05. The SPSS version 21.0 was used (SPSS Inc., Chicago, Ill.). Data are presented as mean±SD.
Plasma glucose and plasma lactate were measured in a Cobas Integra Plus 400 chemistry analyzer (Roche
Diagnostics, Mannheim, Germany). A competitive immunoassay was used for the quantitation of F2-isoprostanes
in plasma (Cayman Chemical, Charlotte, USA). Plasma was purified using the solid phase extraction cartridges.
and the subsequent ELISA assay were performed following the manufacturer’s
NR administration increased the levels of NADPH in liver (P=.05), whereas exercise decreased liver NADPH levels (P=.022) (Figure 2A). Conversely, NR administration did not affect the levels of NADPH in muscle (P=.370), whereas exercise increased muscle NADPH levels (P=.002) (Figure 2B). Both NR administration and exercise increased the levels of F2-isoprostanes in plasma (P=.047 and P=.027, respectively) (Figure 3). Likewise, both NR administration and exercise decreased the activity of glutathione peroxidase (P=.017 and P=.010, respectively) in muscle. NR administration decreased muscle glutathione reductase activity (P<.001), whereas exercise did not exert any significant effect (P=.228). Finally, a significant interaction (supplement × exercise) was found for muscle catalase activity (P=.024), with main effects of NR administration and exercise equal to P=.024 and P<.001, respectively (Figure 4A, 4B and 4C).
Regarding energy metabolism, NR administration markedly increased liver glycogen concentration (i.e., by 68% in the NR group compared to the Con group and by 163% in the NR+Ex group compared to the Ex group, P<.001), but did not exert any significant effect on muscle glycogen concentration (P=.425). On the other hand, exercise did not affect the concentration of glycogen both in liver and muscle (P=.864 and P=.397, respectively) (Figure 5A and 5B). Both NR administration and exercise decreased glucose concentration in plasma (P=.016 and P<.001, respectively) (Figure 5C). Finally, the NR+Ex group exhibited a trend (marginally non-significant) towards lower maximal lactate accumulation levels after exercise compared to the Ex group (P=.084) (Figure 5D).
To our knowledge, this is the first study to examine the effects of NAD(P)H precursors on redox homeostasis. NR administration for 21 days induced systemic oxidative stress at rest (as assessed by the reference biomarker F2- isoprostanes in plasma) and decreased the activity of the key antioxidant enzymes glutathione peroxidase, glutathione reductase and catalase in muscle. A plausible mechanism explaining the decreased activity of the antioxidant enzymes may be the increased resting level of reactive oxygen and nitrogen species. In particular, taking into account that the activity of an enzyme is partially controlled by its oxidation state (Day et al., 2012; Koufen & Stark, 2000), the more oxidized environment in the NR group could reasonably explain the decreased antioxidant enzyme activity observed. However, we acknowledge the fact that we did not measure any oxidative stress biomarker in skeletal muscle, and future studies are warranted to elucidate this issue. According to these data, and taking into account the Janus-faced redox role of NADPH as substrate for both antioxidant (i.e., peroxiredoxins and GSH) and oxidant sources (NADPH oxidases and NO• synthases), it seems that NR supplementation favored the “pro-oxidant” properties of NADPH, as demonstrated by the increased levels of oxidative stress and the decreased antioxidant enzyme activity. These findings could possibly clarify, at least in part, the impaired exercise performance found in the same animals and presented in our previous report (Kourtzidis et al., 2016). Certainly, other mechanisms that have not been evaluated in the context of the present study (e.g., activity of PARPs and sirtuins; see “Other potential mechanisms explaining the negative effects of NR”) could also provide sound explanations (Bonkowski & Sinclair, 2016). Nevertheless, we believe that the strong redox component of NADPH and the increasingly acknowledged interplay between redox biology processes and energy metabolism (Mailloux & Treberg, 2016) represent central regulators of the impaired exercise performance observed in the NR+Ex group compared to the Ex group.
Effect of NR supplementation on tissue NADPH levels and redox homeostasis
Our results show that NADPH concentration increased by 49% in liver at rest after chronic administration of NR (NR group compared to Con group). In contrast, NADPH concentration in skeletal muscle did not significantly change after NR administration. The only relevant study available has reported an 115% increase in the oxidized form of NADPH (i.e., NADP+) in liver, but no significant change was observed in heart tissue in mice (Trammell et al., 2016). This probably indicates that the main NADPH producer tissue is liver and that other tissues lack this capacity (Ramos-Martinez, 2017). In addition, NR compared to other more frequently used supplements, such as nicotinamide and nicotinic acid, has been proved to be the most bioavailable precursor of NADPH after oral
administration (Trammell et al., 2016). We emphasize that mode of delivery, dose and timing could all have affected the impact of NR on redox status, metabolism and performance (Liu et al. 2018). We did not measure the concentration of NR because of its extremely short half-life, which is approximately 3 min after intravenous administration in mice (Liu et al. 2018). However, we measured the target molecule of NR supplementation NADPH.
Potential mechanisms driving the detrimental effects of NR supplementation on performance
Exercise performance is a complex physiological process not determined by the function of one tissue only. Instead, it depends on an integrated set of pulmonary, cardiovascular, neurological, erythrocyte and muscle metabolic functions (Hawley et al., 2014; Lundby et al., 2017). Redox reactions regulate, at least in part, these processes, thus, a disturbed redox homeostasis may lead to aberrant physiological outcomes (Hoppeler, 2016; Margaritelis et al., 2016). For instance, non-physiological levels of oxidative stress (as was the case in the NR group) may impair pulmonary diffusing capacity (Garcia-Rio et al., 2011), vasodilation and blood flow (Trinity et al., 2016), deformability of erythrocytes and oxygen delivery (Mohanty et al., 2014) and mitochondrial bioenergetics (Anderson et al., 2009). Some or, most probably, all these oxidative stress-induced impairments may have contributed to the reduced performance detected in the NR group.
Trying to link the redox alterations with the impaired exercise performance in the NR+Ex group compared to Ex group, we measured a set of metabolic biomarkers related to carbohydrate metabolism. Our results showed that NR supplementation excessively increased glycogen concentration in liver. This finding becomes intriguing taking also into account the concomitant decrease in blood glucose levels and the lower maximal lactate accumulation after exercise in the NR+Ex group compared to Ex group. Connecting the dots, these findings indicate that glycogenolysis in liver was probably inhibited by NR administration, leading to slower release of glucose into circulation and, consequently, hampered anaerobic glycolysis in muscle during exercise. Of course, this is a speculative mechanistic explanation of the decreased exercise performance observed after NR administration in the same animals (Kourtzidis et al., 2016). Unfortunately, we did not measure lactate concentration between each stage in order to gain insights to the different metabolic demands imposed by each exercise stage. In support to our finding, a recent review by Hill and Williams (2017), highlighted the significance of maintaining an optimal level of vitamin B3, since shifts towards both directions, either to hypo- (due to malnutrition) or hyper-vitaminosis (due to nicotinamide overload), may lead to adverse clinical/physiological outcomes. Exercise alone did not manage to decrease the levels of liver and muscle glycogen in either NR- administered or control rats and, as a result, differences in glycogen utilization during exercise cannot explain the differences observed between the groups. This is probably because the first 12 min of the swimming protocol were mainly of low intensity and as a result were predominantly supported by fat utilization thus sparing liver and muscle glycogen content. We hypothesize that the last 90 s of high intensity swimming were possibly not enough to significantly decrease glycogen content in liver and muscle.
This “hormetic” impact of NR concentration on physiology seems analogous to the effects of other nutritional and pharmacological agents with redox properties, such as antioxidants. More specifically, although antioxidants are generally considered essential for maintaining redox homeostasis, deficiency or overload with antioxidants lead the biological system towards a non-optimal redox state (more oxidized or reductive, respectively) leading to adverse effects (Gomez-Cabrera et al., 2015; Merry & Ristow, 2016). This concept was recently termed by Sies and colleagues (2017), as oxidative “eustress”, where a physiological level of oxidant challenge is essential for regulating cellular processes through redox signaling, whereas oxidative “distress” characterizes an aberrant (non-optimal) redox homeostasis leading to damage (as was the case in our study after NR supplementation).
Implications for precision redox nutrition
Many studies conducted during the last decade have shown that aged and diseased populations, who generally suffer from low NAD(P) levels, experience great biochemical and physiological benefits after NR supplementation (Vaur et al., 2017; Frederick et al., 2016; Trammell et al., 2016; Cantó et al., 2012; Zhang et al., 2016). In contrast, our study demonstrated that NR supplementation yielded negative effects on exercise performance and redox homeostasis in healthy young rats. This “contradiction” clearly emphasizes the importance of the emerging concept of precision (also called personalized) medicine/nutrition in order to identify the individuals who are in need of a specific nutritional treatment. In line with this notion, and as regards to redox biology, we have recently
shown that a precise antioxidant supplementation protocol (i.e., vitamin C or N-acetylcysteine) elicits great benefits to individuals with a corresponding antioxidant deficiency (i.e., vitamin C or glutathione, respectively), whereas the antioxidant-sufficient counterparts either do not experience any benefit or, even worse, sporadic detrimental effects may also been observed (i.e., worse exercise performance) (Paschalis et al., 2018; Paschalis et al., 2016).
This concept could also explain in part the current consensus about antioxidant supplementation during exercise, which states that antioxidants impair exercise performance and adaptations (Gomez-Cabrera et al., 2015; Nikolaidis et al., 2012). In particular, we believe that the participants in most studies may not have been the appropriate to experience any benefit from the antioxidant supplement (healthy individuals with a normal redox profile). On the contrary, antioxidant-deficient individuals (e.g., due to malnutrition or environmental factors, such as passive smoking) typically exhibit high resting oxidative stress levels (Theodorou et al., 2014). Hence, antioxidant supplementation reverses this aberrant redox homeostasis (i.e., reduces resting oxidative stress) and subsequently provides greater potential for exercise-induced oxidative stress, which has been demonstrated to be an essential signal for exercise adaptations (Margaritelis et al., 2018). In the same vein, in our study, NR (which can be considered a redox agent due to its NADPH-increasing capacity) elicited negative effects to young healthy rats. Collectively, our findings strongly support the idea of more personalized and targeted approached in the field of redox nutrition.
Other potential mechanisms explaining the negative effects of NR
Nicotinamide riboside is associated with PGC-1a function and PGC-1a expression, which consequently affects mitochondrial metabolism (Gong et al., 2013). In addition, NR has been demonstrated to contribute to the replenishment of cellular NAD+ and the promotion of sirtuin (SIRT1 and SIRT3) gene activation (Canto et al., 2012; Frye, 1999; Grube & Burkle, 1992; Tanner et al., 2000). Glycoprotein CD38, as well as PARP activity may be the major sources of intracellular NAD+ consumption (Bongan et al., 2008; Escande et al., 2013). CD38 is a multifunctional ectoenzyme that catalyzes the synthesis and hydrolysis of cyclic ADP-ribose (cADPR) from NAD+ to ADP-ribose. These reaction products are essential for the regulation of intracellular Ca2+ (Malavasi et al., 1994). PARPs use NAD+ as a co-substrate to modify target proteins releasing NAM (Bai and Canto, 2012). Some of these NAD(P)-controlled signaling mechanisms might have been negatively affected by chronic NR administration leading to impaired metabolism and/or muscle function leading to reduced performance.
Based on our findings, we report that chronic administration of NR: i) increased NADPH levels in liver, but not in muscle; ii) increased systemic oxidative stress as assessed by plasma F2-isoprostanes; iii) decreased the activity of major antioxidant enzymes in muscle; iv) excessively increased glycogen in liver, but not in muscle; v) decreased glucose levels in blood and vi) decreased maximal lactate production during exercise. Nicotinamide riboside is currently regarded the best available nutritional supplement to increase NAD(P) levels and alleviates major diseases. Most of the evidence presented in this study, and in our previous report (Kourtzidis et al., 2016), indicates that a 21-day NR supplementation protocol leads to dysregulation in redox and energy metabolism and impaired exercise performance in healthy young rats. We believe that our findings are timely and add to the expanding literature showing that altering metabolic and redox homeostasis via exogenously administered agents in heathy populations may lead to adverse and not necessarily to beneficial or neutral effects.