Introduction

It is long been known that extracellular NAD+ can be rapidly hydrolyzed to ADP-ribose in isolated perfused rat livers1. The enzymatic system responsible for the transformation of extracellular NAD+ or NADP+ is usually referred to as glycohydrolase/ADP-ribosyl cyclase2. It has also been shown that rabbit and rat liver extracts are able to catalyze the formation of cyclic ADP-ribose3, 4, and that the hepatic ADP-ribosyl cyclase activity is similar to that of other tissues. There are claims that the NAD glycohydrolase/ADP-ribosyl cyclase activity corresponds to the CD38 protein, which is present in many cells2. The existence of CD38 in the plasma membrane of hepatocytes has indeed been demonstrated by Khoo and Chang5. In liver cells, this protein shows a 10-fold higher specific activity in the sinusoidal membrane fraction than in the bile canalicular membrane fraction. Recent investigations have shown that the main products in the effluent perfusate in the liver of Wistar rats after portal infusion of NAD+ are ADP-ribose, uric acid and inosine6.

Extracellular nucleotides such as NAD+ are considered to be ideally suited as extracellular signal transmitters because they can be rapidly mobilized from intracellular stores and the signal is rapidly terminated by degradation by nucleotide catabolizing enzymes. NAD+ is supposed to be preferentially released from intracellular stores in conditions of cell stress or inflammation7. There are many cells that respond in several ways to extracellular NAD+. In perfused livers from Wistar rats (Rattus novergicus), the infusion of NAD+ produces increases in portal and arterial pressures and the initial inhibition of oxygen consumption followed by stimulation, increase in glycogenolysis and inhibition of gluconeogenesis8, 9. These effects are Ca2+-dependent and seem to result from an interaction between parenchymal and non-parenchymal cells via eicosanoid production. Furthermore, recent evidence indicates that these metabolic effects of NAD+ in the liver are partly due to its transformation products, such as ADP-ribose, and that these effects are mediated by purinergic receptors6.

Unlike the rat liver, little is known about the action and transformation of NAD+ in the mouse liver. The NAD glycohydrolase/ADP-ribosyl cyclase is also present in mice10, 11, but no systematic study has been done in which transformation and the possible metabolic effects of extracellular NAD+ in the mouse liver have been investigated. To fill this gap, transformation of NAD+ was investigated using high-performance liquid chromatography (HPLC). In addition, gluconeogenesis from lactate and its associated changes in oxygen uptake were measured as typical NAD+-sensitive metabolic parameters. In order to accentuate these comparative characteristics, experiments were done with isolated perfused mouse livers and perfused livers from two rat strains, namely, Wistar and Holtzman. Since NAD+ has been proposed to be a universal extracellular signal transmitter and since there are many suggestions about its future use as a therapeutic drug2, 7, it is also of interest to characterize its transformation and effects in the greatest possible number of species.

Materials and methods

Materials

The liver perfusion apparatus was built in the workshops of the University of Maringá. NAD+ and all enzymes and coenzymes used in the enzymatic assays and chromatographic standards were purchased from Sigma Chemical Co (St Louis, USA). All standard chemicals were of the best available grade (98%−99.8% purity).

Animals and liver perfusion

Male Swiss mice (Mus musculus), weighing 30-40 g, were fed ad libitum with a standard laboratory diet (Nuvilab®). Male albino rats (Rattus novergicus; Wistar and Holtzman strains), weighing 200−250 g, were fed ad libitum with the same laboratory diet. Food was withdrawn 24 h prior to the perfusion experiments. All experiments were done in accordance with the worldwide accepted ethical guidelines for animal experimentation. For the surgical procedure, rats and mice were anesthetized by intraperitoneal injection of sodium thiopental (50 mg/kg). In situ hemoglobin-free, non-recirculating perfusion was performed12 with cannulation of both portal and cava veins. The perfusate flow was provided by a peristaltic pump (Minipuls 3, Gilson, France) and was around 12 mL·min−1 for the mouse liver and around 31 mL·min−1 for the rat liver. The perfusion fluid was Krebs/Henseleit-bicarbonate buffer (pH 7.4), saturated with a mixture of oxygen and carbon dioxide (95:5) by means of a membrane oxygenator with simultaneous temperature adjustment at 37 oC. The Krebs/Henseleit-bicarbonate buffer was composed of 115 mmol/L NaCl, 25 mmol/L NaHCO3, 5.8 mmol/L KCl, 1.2 mmol/L Na2SO4, 1.18 mmol/L MgCl2, 1.2 mmol/L NaH2PO4 and 2.5 mmol/L CaCl2.

Analytical methods

NAD+ and its metabolites in the outflowing perfusate (Krebs/Henseleit-bicarbonate buffer, pH 7.4) were separated by means of high-performance liquid chromatography (HPLC). The HPLC system (Shimadzu, Japan) consisted of an SCL-10AVP system controller, two model LC10ADVP pumps, a model CTO-10AVP column oven and a model SPD-10AVP UV-VIS detector. Standards were dissolved in Krebs/Henseleit-bicarbonate buffer, pH 7.4. Solutions were filtered through disposable 0.2 μm filters directly before injection. A reversed-phase C18 HRC-ODS column (5 μm; 150 × 6 mmol/L ID; Shimadzu, Japan) protected with a GHRC-ODS pre-column (5 μm; 10×4 mm ID; Shimadzu, Japan) was used with a gradient from reversed-phase 0.044 mol/L phosphate buffer solution, pH 6.0, to 0.044 mol/L phosphate buffer solution plus methanol (1.1), pH 7.0, at 0.8 mL·min−1. The gradient consisted of the following steps (in % of methanol): 0 min, 0%; 2.5 min, 0.5%; 5 min, 3%; 7 min, 5%; 8 min, 12%; 10 min, 15%; 12 min, 20%; 20 min, 30%. The temperature was kept at 35 °C, and the injection volume was constant at 20 μL. The UV-absorbance detector was autozeroed at the start of each chromatogram, and absorbance was measured at 254 nm.

For the identification of the components in the perfusate, a retention time window of ±2% was adopted. The concentrations of NAD+ and its transformation products in each perfusate sample were calculated by means of the regression parameters obtained from calibration curves. The calibration curves were constructed by chromatographically separating standard Krebs/Henseleit-bicarbonate buffer solutions of the various compounds. Linear relationships were obtained between the concentration and the areas under the absorbance curves. The mean parameters obtained for the calibration curves were the following: (a) NAD+, linear regression coefficient=17214±98, correlation coefficient=0.999; (b) nicotinamide, linear regression coefficient=2922±82, correlation coefficient=0.999; (c) ADP-ribose, linear regression coefficient=14350±196, correlation coefficient=0.999; (d) AMP, linear regression coefficient=12222±347, correlation coefficient=0.999; (e) inosine, linear regression coefficient=14904±94, correlation coefficient=0.999; (f) uric acid, linear regression coefficient=4400±11, correlation coefficient=0.999.

During metabolic analyses of NAD+, the oxygen concentration in the effluent perfusate was monitored continuously using a teflon-shielded platinum electrode12. Samples of the effluent perfusion fluid were collected according to the experimental protocol and analyzed for glucose content13. The metabolic rates were expressed as μmol per minute per gram liver wet weight (μmol·min−1·g−1).

Data analysis

The metabolic rates were expressed as μmol per minute per gram liver wet weight (μmol·min−1·g−1). Statistical analysis of the data was done by means of the StatisticaTM program (Statsoft®, 1998). Student's t test was applied for evaluating the statistical difference between two means; P<0.05 was adopted as a criterion of significance. Peak times and times for half-maximal recovery of glucose production and oxygen uptake during NAD+ infusion of livers from Wistar and Holtzman rats were determined using numerical interpolation (Stineman's interpolation) by means of the Scientist® software from MicroMath Scientific Software (Salt Lake City, UT, USA).

Results

Transformation of extracellular NAD+ in mouse and rat livers

Single-pass transformation of extracellular NAD+ was measured. After a short pre-perfusion period in which oxygen uptake by the liver was allowed to stabilize, NAD+ was infused. Samples of the outflowing perfusate were collected for chromatographic analyses at different times after starting the infusion. In the mouse liver perfusion, two different NAD+ concentrations were infused, 100 and 500 μmol/L. In the mouse liver, the single-pass extraction of NAD+ was practically complete when the compound was infused into the portal vein at the concentration of 100 μmol/L (Figure 1, curve A). At 500 μmol/L, however, some NAD+ was recovered (Figure 1, curve A). In addition to the non-transformed NAD+, two transformation products were found, ADP-ribose and nicotinamide, as can be deduced from their retention times, 9.54±0.06 and 17.37±0.11 min, respectively. Very small amounts of uric acid (5.51±0.05 min) and AMP (12.98±0.01 min) were also produced and were clearly visible in the chromatogram of the 100 μmol/L NAD+ infusion (Figure 1, curve A). Livers from both rat strains, Wistar (curve A) and Holtzman (curve B), show similar transformation products (Figure 2). At least six components could be identified: uric acid (5.51±0.05 min), ADP-ribose (9.54±0.06 min), AMP (12.98±0.01 min), NAD+ (16.70±0.06 min), nicotinamide (17.37±0.11) and inosine (17.91±0.12 min). AMP was also clearly visible for the Holtzman strain. The first peak, with a retention time of around 3.5 min, could not be identified. No transformation product with a retention time longer than 20 min was found, an observation that excludes free adenosine, which had a retention time of 23.64±0.07 min. Furthermore, cyclic-ADP-ribose (cADPR) and β-nicotinic acid adenine dinucleotide phosphate (β-NAADP) were expected to appear with retention times between 7 and 8 min6. This region, however, was practically empty in all chromatograms obtained in the present work (Figures 1 and 2).

Figure 1
figure 1

Typical chromatographic profiles of standards and outflowing per-fusate of the mouse liver. Perfusate samples were collected at 1.3 min after initiation of NAD+ infusion. HPLC fractionation with spectrophotometric detection at 254 nm was done as described in the Materials and methods section. Legends: (A) outflowing perfusate with 100 μmol/L NAD+ infusion; (B) outflowing perfusate with 500 μmol/L NAD+ infusion; (C) standards. The absorbance scale was normalized as a fraction of the highest absorbance peak in each run.

Full size image
Figure 2
figure 2

Typical chromatographic profiles of standards and outflowing perfusate of rat livers from the Wistar and Holzman strains. Perfusate samples were collected at 1.3 min after initiation of 100 μmol/L NAD+ infusion. HPLC fractionation with spectrophotometric detection at 254 nm was done as described in the Materials and methods section. Legends: (A) Holzman strain; (B) Wistar strain; (C) standards. The absorbance scale was normalized as a fraction of the highest absorbance peak in each run.

Full size image

The concentrations of all identified compounds in the effluent perfusate were calculated from the areas under the chromatographic curves using the regression coefficients obtained from appropriate calibration curves, as described in the Materials and Methods section. These concentrations were represented as a function of the NAD+ infusion times. For the mouse liver, the portal 100 μmol/L NAD+ infusion corresponds to an infusion rate of 0.8 μmol· min−1·g−1. Since metabolism of NAD+ was complete under these conditions, 0.8 μmol·min−1·g−1also corresponds to the transformation rate. The concentration of ADP-ribose in the outflowing perfusate increased during the first 3 min of NAD+ infusion and stabilized thereafter (Figure 3). The concentration of nicotinamide also increased during the first 3 min and reached approximately 90% of the NAD+ input (Figure 3). ADP-ribose and nicotinamide, however, were not released in equimolar amounts, suggesting a more accentuated intracellular processing of the ADP-ribose moiety of NAD+. The productions of AMP and uric acid were minimal. In the mouse liver, the portal 500 μmol/L NAD+ infusion corresponds to an infusion rate of 4 μmol·min−1·g−1. Under these conditions, ADP-ribose and nicotinamide were released in almost equimolar amounts (Figure 4). A decline in the concentration of both ADP-ribose and nicotinamide was evident at 14 min of infusion. This decline was accompanied by an increase in NAD+ recovery, which reached 28.5% at 14 min. The rate of NAD+ transformation at this time was 2.86 μmol·min−1·g−1. It seems, therefore, that the capacity of NAD+ transformation decreased significantly with the infusion time, at least at this high portal concentration of 500 μmol/L. Here again, the productions of AMP and uric acid were very small.

Figure 3
figure 3

Time course of 100 μmol/L NAD+ transformation in the perfused mouse liver. Livers of fasted mice were perfused with Krebs/Henseileit-bicarbonate buffer (pH 7.4) as described in Materials and Methods. The concentrations of NAD+ and its metabolites in the outflowing perfusate at different times after initiation of NAD+ infusion were determined by HPLC. Data are the means±mean standard errors of four liver perfusion experiments.

Full size image
Figure 4
figure 4

Time course of 500 μmol/L NAD+ transformation in the perfused mouse liver. Livers of fasted mice were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) as described in Materials and Methods. The concentrations of NAD+ and its metabolites in the outflowing perfusate at different times after initiation of NAD+ infusion were determined by HPLC. Data are the means±mean standard errors of four liver perfusion experiments.

Full size image

In the rat liver, a portal NAD+ concentration of 100 μmol/L corresponds to an infusion rate of 0.32 μmol·min−1·g−1. The single-pass transformation of NAD+ was 97% at all infusion times, except at 1.5 min, where it was 94% (Figure 5). An extraction of 97% corresponds to a transformation rate of 0.31 μmol·min−1·g−1. A similar, but considerably higher, recovery peak was also found in livers from rats of the Wistar strain, and it coincided with the time of maximal increase in perfusion pressure caused by NAD+6. Nicotinamide tended to approach 90% of the infused NAD+at the end of the infusion period (Figure 5). AMP and inosine comprised very small fractions of the transformation products, but the contributions of ADP-ribose and uric acid were highly significant (Figure 5). This is again partly similar to what was found in livers from the Wistar strain, but in those rats, ADP-ribose was considerably more abundant than uric acid, especially at the end of the infusion period6.

Figure 5
figure 5

Time course of 100 μmol/L NAD+ transformation in the perfused liver of Holtzman rats. Livers of fasted rats were perfused with Krebs/Henseleit-bicarbonate buffer (pH 7.4) as described in Materials and Methods. The concentrations of NAD+ and its metabolites in the outflowing perfusate at different times after initiation of NAD+ infusion were determined by HPLC. Data are the means±mean standard errors of four liver perfusion experiments.

Full size image

Metabolic actions of extracellular NAD+ in mouse and rat livers

In the liver of the Wistar rats, NAD+ and its major transformation product, ADP-ribose, are able to induce transient decreases in both oxygen uptake and gluconeogenesis from lactate (Figure 6)6, 9. Livers from rats of the Holtzman strain also respond in this way (Figure 7). In the experiments, livers from 24-h fasted rats were perfused in order to minimize interference by glycogen catabolism9. Lactate (2 mmol/L) was infused for 44 min before a 100 μmol/L NAD+ infusion (Figures 6 and 7). At 46 min, both oxygen uptake and gluconeogenesis had almost stabilized. The NAD+ induced rapid decreases in both oxygen uptake and gluconeogenesis, followed by recoveries that culminated in transient stimulations. The peak decreases and stimulations found in livers of Wistar and Holtzman rats were evaluated. It is apparent that the peak inhibitions of gluconeogenesis and oxygen uptake were more pronounced (by factors of 2.4 and 2.8, respectively) in the Wistar strain (Table 1). The opposite occurred with the transient stimulations, which were more pronounced in the livers of the Holtzman strain (by factors of 6.0 and 4.3, respectively, for gluconeogenesis and oxygen uptake). The metabolic effects of NAD+ were further characterized by evaluating the times for maximal inhibition (peak times) of glucose production and oxygen uptake and the times for half-maximal recoveries (t½). These parameters have been considered measures of the persistency of the metabolic effects of extracellular NAD+14. In livers from Holtzman rats, the peak times are approximately half those of livers from Wistar rats (Table 2). Additionally, the recovery is 2.95 to 4.59 times faster in livers from rats of the Holtzman strain (Table 2).

Figure 6
figure 6

Gluconeogenesis and oxygen uptake in the liver of Wistar rats: influence of lactate and NAD+. Livers from 24-h fasted rats were perfused as described in Materials and Methods. After oxygen uptake stabilization, lactate and NAD+ (at the given concentrations) were infused at the indicated times. Samples of the effluent perfusate were collected for glucose assay. Oxygen was monitored polarographically. Data represent the means±SEM of five liver perfusion experiments.

Full size image
Figure 7
figure 7

Gluconeogenesis and oxygen uptake in the liver of Holtzman rats: influence of lactate and NAD+. Livers from 24-h fasted rats were perfused as described in Materials and Methods. After oxygen uptake stabilization, lactate and NAD+ (at the given concentrations) were infused at the indicated times. Samples of the effluent perfusate were collected for glucose assay. Oxygen was monitored polarographically. Data represent the means±SEM of five liver perfusion experiments.

Full size image
Table 1 Peak decreases and peak stimulations in glucose production and oxygen uptake during NAD+ infusion in livers from Holtzman rats, Wistar rats and Swiss mice. The data were obtained from perfusion experiments of the kind illustrated by Figures 6 and 7. Peak inhibition represents the difference between the rate just before the onset of NAD+ infusion and the minimal rate immediately after initiation of the infusion. Similarly, peak stimulation represents the difference between the rate just before the onset of NAD+ infusion and the maximal rate after the recovery from inhibition. The P values refer to Student's t test in the com-parison between the Holtzman and Wistar strain. aP=0.001; bP=0.002; cP=0.004; dP=0.01.
Full size table
Table 2 Peak times and times for half-maximal recovery in glucose production, and oxygen uptake during NAD+ infusion in livers from rats of the Holtzman and Wistar strains. The data were obtained from perfusion experiments of the kind illustrated by Figure 6. The P values refer to Student's t test in the comparison between the Holtzman and Wistar strain. aP=0.003; bP<0.001; cP=0.001; dP=0.004.
Full size table

To test how the mouse liver behaves during NAD+ infusion, similar perfusion experiments were done with livers from fasted mice (Figure 8). The infusion of lactate increased both glucose production and oxygen uptake to extents similar to the rat liver. The infusion of 100 μmol/L NAD+, on the other hand, produced only small alterations in both glucose production and oxygen uptake (Figure 7 and Table 1). The standard errors are quite high, denoting poor significance. Repetition of the experiments with 500 μmol/L NAD+ was unsuccessful: no significant enhancements of the metabolic responses were found with this concentration (data not shown). The mouse liver also did not respond to a 100 μmol/L ADP-ribose infusion (data not shown), an expected phenomenon if one considers that a high concentration of this metabolite is generated during a 500 μmol/L NAD+ infusion.

Figure 8
figure 8

Gluconeogenesis and oxygen uptake in the mouse liver: influence of lactate and NAD+. Livers from 24-h fasted mice were perfused as described in Materials and Methods. After oxygen uptake stabilization, lactate and NAD+ (at the given concentrations) were infused at the indicated times. Samples of the effluent perfusate were collected for glucose assay. Oxygen was monitored polarographically. Data represent the means±mean standard errors of five liver perfusion experiments.

Full size image

Discussion

The results of the present work extend previous knowledge about the transformation and effects of extracellular NAD+ in the livers of Wistar rats as compared to the livers of Holtzman rats and mice. The mouse liver possesses a large capacity for metabolizing extracellular NAD+ into basically two products: ADP-ribose and nicotinamide. Actually, the mouse liver is more efficient than the rat liver when the transformation rates per unit liver weight are compared. The weight of the mouse livers used in the present work corresponded to approximately 15% of the weight of the rat livers, and metabolism of the 100 μmol/L NAD+ portal infusion was complete in the former but not necessarily in the latter6. It should be remarked that the lower metabolism of 500 μmol/L NAD+ when compared to 100 μmol/L NAD+ reflects, most likely, the fact that the transformation is a saturable process like all enzyme-catalyzed processes. The fact that only two products could be detected in significant amounts in the outflowing perfusate in the mouse liver perfusion presents a noteworthy contrast to the rat liver perfusion, in which several additional transformation products were detected, especially uric acid and inosine. This is true for both the Wistar6 and Holtzman strains. ADP-ribose and nicotinamide can be generated extracellularly, but uric acid and inosine result from reactions within the cell. The virtual absence of uric acid and inosine in the effluent perfusate in the mouse liver perfusion could thus reflect differences in intracellular transformation of the extracellular NAD+ products between mice and rats.

It should be noted that cyclic-ADP-ribose, an important signaling molecule that originates from extracellular NAD+ transformation, was not found in the outflowing perfusate in both mouse and rat liver perfusions. This observation, however, does not exclude its production. In both mouse and rat livers, cyclic-ADP-ribose is probably produced in very small amounts compared to ADP-ribose – only one molecule of cADP-ribose is generated intracellularly for every 100 molecules of NAD+ that are hydrolyzed15. Furthermore, cADP-ribose probably does not permeate the cell membrane, so that it cannot be expected to appear in the outflowing perfusate. These cells probably comprise a minor fraction of the total hepatic cell population. It has been reported that, at least in the rat liver, the NAD+-glycohydrolase activity is 65-fold higher in the plasma membrane of Kupffer cells than in the plasma membrane of hepatocytes16. This indeed makes it difficult to detect cyclic-ADPR in the whole liver, because Kupffer cells comprise a relatively small fraction of the total cell population in the liver.

The metabolic responses to NAD+ in livers from the Wistar and Holtzman strains clearly indicate stronger inhibition in the former and a more pronounced stimulation in the latter. These observations, however, do not necessarily mean that the signals generating inhibition and stimulation are different in the two strains. The times for maximal inhibition and the times for half-maximal recovery are both considerably shorter in the Holtzman strain. This observation could be indicating that only the stimulation is actually faster in the Holtzman strain. At least in theory, a faster stimulation with a similar rate of inhibition in the Holtzman strain would explain the different responses found in the present work.

The rapid transformation of NAD+ in the mouse liver contrasts with the absence of effect on oxygen uptake and gluconeogenesis. This does not necessarily mean that NAD+ is metabolically inert in the mouse liver. Gluconeogenesis and the associated oxygen uptake are only two among many metabolic pathways that could be affected by the compound. However, the fact that gluconeogenesis and oxygen uptake in the rat liver respond intensely to NAD+ 6, 9 allows at least the conclusion that observations in one species are only valid for that species and that generalizations and interspecies extrapolations should be done very carefully. Actually, differences can occur even between different strains of the same species, which is evident from the comparison of the results obtained in the present work with livers from rats of the Holtzman strain with those of the Wistar strain that were reported previously6, 8. The differences include both the proportions of transformation products and the metabolic effects. The transformation products are equally active in metabolism, e.g., ADP-ribose and inosine6, and the differences in the metabolic effects could be partly due to this phenomenon. One should remember that the functions of extracellular NAD+ or its main transformation products depend on eicosanoid synthesis6, 9. In this respect, there are a few reports in the literature about different responses of Wistar and Holtzman rats to agents whose action also depends on eicosanoid synthesis. It is well known, for example, that, in the induction of arthritis, Holtzman rats respond in a stable way to Freund's adjuvant17, whereas the response of Wistar rats varies considerably among individuals18. Another example is the action of bradykinin, which induces dose-dependent contraction of relaxed mesenteric arterial rings from Holtzman rats, but not from Wistar rats, and is mediated by eicosanoids19. Experiments on platelet aggregation induced by arachidonic acid, a precursor of eicosanoid synthesis, can also be regarded as an indicator that Holtzman rats may respond differently when eicosanoids are involved. Platelets from the Holtzman strain were more sensitive to arachidonic acid than were those from other strains20. Concerning the very pronounced difference in the response to extracellular NAD+ between mice and rats, there are no hints in the literature that this could be used to generate a working hypothesis, so this is a question that remains entirely open to future investigations. It is appropriate to remember, on the other hand, that there is currently intense investigation worldwide aiming to elucidate the precise roles of extracellular NAD+ in mammals, but that the ultimate scope is to define its role in humans2, 7, 21. In this respect, the accentuated differences found in the present work between rats and mice, and even between different rat strains, emphasize the necessity to perform more studies on the action of extracellular NAD+ in humans – or at least in cells derived from humans.

Author contribution

Dr Adelar BRACHT and Dr Ana Maria KELMER-BRACHT designed the research, analyzed the data and wrote the article; Dr Ana Carla BROETTO-BIAZON and M Sc Livia BRACHT performed the liver perfusion experiments and analyzed the data; Fabricio BRACHT was responsible for the chromatographic analyses.