In the absence of carbohydrates in the perfusate, the apparent rate of glycogenolysis evaluated by the changes in the area of glycogen C-1 NMR resonance was 0.90 ± 0.08%/min (n = 4) in the control KHB group (100% being considered as the initial content = 73 ± 8.50 μmol.g^-1 glycosyl units, n = 8). After the initial 30 min KHB perfusion, the remaining amount of glycogen (60–75%) when the ethanol was added could potentially ensure glucose supply for glucose excretion and/or glycolytic activity. In order to compare the apparent rate of glycogenolysis, 10 mM ethanol or 2-min IAA (0.5 mM) were added to KHB; glycogenolysis occurred in the same extent [0.78 ± 0.13%/min (n = 4; NS vs KHB) and 0.85 ± 0.08%/min (n = 5; NS vs KHB), respectively]. Glucose excretion was 1.71 ± 0.40 μmol/min.g ww during the equilibration period, and was not significantly changed throughout the perfusion of KHB or ethanol, nor when 2 min- IAA (0.5 mM) was added. Since glycogenolysis and glucose excretion were unchanged when glycolysis was inhibited, the question arises as other pathways for glucose utilization.

The change of the signal of phosphomonoesters (PME; 3.83 ppm) is mainly representative of that of sn-glycerol-3-phosphate, by-product of glycolysis (provided from dihydroxyacetone phosphate [DHAP]). As shown in Figure 1, the suppression of glycolysis with 2-min IAA (0.5 mM) addition in the KHB group (n = 3) led to a moderate 5–10% increase in the PME peak. A 20–30% increase occurred under ethanol perfusion (n = 3) without any significant dose-dependent effect in the range of 10–100 mM. The conjunction of ethanol and 2 min-IAA perfusion allowing the total inhibition of glycolysis (n = 3) amplified this increase by a maximum of 40–50% (Fig 1 and 2). In all groups, the maxima of these transient increases were reached after 10–20 min perfusion.

In order to stimulate the formation of DHAP, the oxidative phosphorylation is inhibited by addition of KCN (2.5 mM; inhibitor of the cytochrome oxidase). The glycolysis was enhanced to compensate the ATP supply; the by-product PME was then increased, as observed in KHB group (30 ± 3%; n = 5). The subsequent addition of ethanol led in 30 min to a dramatic and transient increase of PME (227 ± 45%; n = 3; P < 0.02 vs KHB).

Representative ³¹P NMR spectra obtained during KHB perfusion are presented in Figure 2A. The addition of ethanol after the 30 min initial KHB perfusion enhanced the initial rate of net total ATP consumption (Vi = 100 ± 12 nmol/min.g ww, P < 0.001) as compared to KHB (Vi = 6.0 ± 0.12 nmol/min.g ww, n = 20), whatever the range of concentration studied (10–100 mM). The ATP content stabilized at 75 ± 3% (n = 16) of the initial ATP level after about 20–30 min. The new steady state of ATP content-mainly representative of cytosolic ATP-observed was independent of the concentration of ethanol used, and was characterized by a slight net consumption of the total ATP similar to that of the KHB control. On the other hand, addition of ethanol induced an increase in the fraction of the form of ATP bound to divalent ion (bound ATP), F=bound ATP/total ATP, as shown in the upper graph of Figure 3 (typical experiment performed with 100 mM). Under the initial KHB condition, the difference between total and bound ATP was 11% and was reduced about 15 min after alcohol-addition to reach 6% at the end of experiment. The increase in the bound ATP/total ATP ratio observed under ethanol perfusion was unexpected and an explanation such as Mg²⁺ leakage from mitochondria remains to be validated in further investigations.

The suppression of glycolysis by 2 min-IAA (0.5 mM) in KHB group decreased the liver ATP content to reach a new steady state [77.2 ± 3.1% (n = 6) of the initial ATP level]; the ΔATP was about 20% between the two steady states (KHB-IAA). As previously demonstrated 1, addition of 2 min- IAA (0.5 mM) in ethanol-perfused livers (n = 5) slightly decreased the liver ATP content, leading to a ΔATP of about 4% between the two steady states (ethanol-IAA). This evidenced that ethanol has strongly reduced glycolytic ATP production before the IAA addition.

This effect was also demonstrated when glycolytic ATP production was previously stimulated by inhibition of the oxidative phosphorylation (2.5 mM KCN). In the KHB group (n = 3), KCN addition rapidly led to a new steady state of the ATP content (around 45–65% of the initial content). ATP was supplied until glycogenolysis and glycolysis remained active. The subsequent addition of IAA (0.5 mM during 2 min) to KCN (n = 3) led to the total inhibition of the residual glycolytic ATP synthesis, as expected. The subsequent addition of 10 mM ethanol (instead of IAA) to KCN led to a similar decrease of the ATP content, thus confirming the inhibition of glycolysis by ethanol. Representative experiments are shown in Figure 4.

The average pHi value was 7.20 ± 0.06 (n = 15) under KHB perfusion. Ethanol led to a slight dose-dependent decrease in pHi, with a maximum of -0.069 pH units obtained from 50 mM ethanol, the ethanol concentration for the half-maximum effect on pHi being around 13 mM (Figure 5). A plateau of the pHi was obtained within 15–20 min. Ethanol induced also an acidification of the effluent from the beginning of perfusion with variation from -0.05 to -0.10 pH units.

The total ATP hepatic content (75% of the initial level) reached about 30 min after ethanol perfusion characterized a new energy steady state. In our experimental conditions in the absence of glucose in the perfusate, glycolysis depended only on the glycogenolytic activity, this latter indeed occurring in the presence of ethanol, as also demonstrated in isolated hepatocytes 3. Glycolysis might then supply some ATP until the whole carbohydrate store is oxidized, since no increase in glucose excretion was observed. The inhibition of glycolysis with IAA indeed induced a significant decrease (-20%) in the ATP content in the control (KHB) perfused livers. Moreover, the glycolysis inhibition with IAA in ethanol-perfused livers induced only a slight decrease (-4%) in the ATP content. Beside of previous works observing the ethanol-induced decrease of lactate and/or pyruvate excretion which demonstrated decrease in the glycolytic activity in ethanol-perfused liver isolated from fed rats 4 and in isolated hepatocytes from ethanol-fed rats 56, the present work clearly evidenced that ethanol highly diminished the glycolytic ATP production. Hence, ATP production is mainly mitochondrial in the presence of ethanol.

Since glycolysis was inhibited and no glucose excretion occurred whereas glycogenolysis remained active, the question arises as other pathways for glucose utilization.

The interference of ethanol with the liver metabolism is to a large extent due to the NADH+H⁺ produced in the oxidation catalyzed by alcohol dehydrogenase in the cytosol and aldehyde dehydrogenase in the mitochondria. In the latter, the electron transfer system allows the reoxidation of NADH+H⁺ to NAD⁺. The reoxidation of cytosolic NADH involves the malate/aspartate and/or irreversible DHAP/sn-glycerol-3-phosphate shuttles, since the mitochondrial membrane is known to be impermeable to NADH. An alternative pathway for regeneration of cytosolic NAD, without the transfer of reducing equivalents to the mitochondria, is the reduction of pyruvate to lactate. It has already been demonstrated in perfused rat liver that 10 mM ethanol largely increase the lactate/pyruvate ratio, whereas glycolysis was inhibited as evidenced by a decrease in the release of lactate+pyruvate 9. In the present work, immediately after ethanol perfusion, PME content increased reflecting the large accumulation of sn-glycerol-3-phosphate as previously demonstrated 891213. This ethanol-stimulated PME production evidenced, as when we have used IAA to inhibit glycolysis, the deviation of glycolysis to the DHAP/sn-glycerol-3-phosphate shuttle. However in all our experimental conditions, the PME content increase was transient. The subsequent decrease in the PME content could be due to (i) stimulation of the use of glycerol-3 phosphate in the synthesis of hepatic triacylglycerol specially in the presence of ethanol 14 and/or (ii) the decrease in the glycogen content. As consequence, it could be postulated that the metabolic effects of ethanol must be enhanced in absence of carbohydrate store (i.e., fast).

As demonstrated in this work, ATP production in the presence of ethanol is mainly mitochondrial and the observed increase in its rate 1 at the new steady state might reflect an increase in ATP consumption in the whole liver, since at a steady state the net ATP consumption rate equals the net ATP synthesis. These findings raise the issue of the involvement of the ethanol-induced ATP-consuming processes.

The activation of free fatty acyl CoA required for triglycerides synthesis could be one of them, but its contribution is probably minor. On the other hand, a concentration of 10 mM ethanol is known to produce a maximum rate of alcohol oxidation averaging 2–4 μmol/g.min^-1 71516. Hence it leads to 4–8 μmol/g.min^-1 of NADH and H⁺ in the cytosol, the latter potentially representing a large acid load when respiration remains unchanged. Cytosolic ethanol oxidation leads to an excess of protons whose cellular extrusion stimulates mechanisms leading to ATP consumption via the activation of the Na⁺-K⁺-ATPase pump, which in turn extrudes the Na⁺. Indeed, a Na⁺ overload induced by ethanol observed in hepatocytes 17 results from the activity of membrane ionic carriers. In physiological conditions, in hepatocytes the active carriers are the the Na⁺-HCO₃^- symport and the Na⁺/H⁺ antiport, this latter being activated below 7.2–7.3 18, whereas in the liver the activity of the Na⁺/H⁺ antiport is negligible above its set point of pHi 7.0 1011; other carriers such as Na/K/Cl cotransporter may be implicated. The Na⁺-K⁺-ATPase pump requires about 25% of the cellular ATP turn over and oxygen consumption in basal conditions 1920. The pHi regulation from the beginning of ethanol perfusion was here demonstrated by (i) the slight dose-dependent decrease in pHi, with a maximum of -0.069 pH units and (ii) the expulsion of protons observed in the effluent. It has already been demonstrated by our team in the isolated liver 21 that acute ethanol administration under hypoxia dramatically accentuates the intracellular acidosis and the simultaneous ATP decrease; from this result, we can point out the contribution of ATP to pHi homeostasis in this condition. However, no detailed study concerning the involved mechanisms has been previously performed in an isolated perfused liver. In the present work, the cytosolic acidification induced either by SITS or ouabaïn evidenced for the first time in presence of ethanol a contribution of Na⁺-HCO₃^- symport to a rapid H⁺ elimination and Na⁺-K⁺-ATPase, respectively. It probably accounts for the ATP consumption observed at the onset of the alcohol perfusion.

Beside of ATP-consuming processes, we clearly demonstrate that a cellular way eliminating protons saves energy. A concentration of 10 mM ethanol is known to produce a maximum rate of alcohol oxidation averaging 2–4 μmol/g.min^-1 71516. Hence it leads to 2–4 μmol/g.min^-1 of acetate and to 4–8 μmol/g.min^-1 of NADH and H⁺. It has been demonstrated that acetate ions arising from ethanol oxidation are barely metabolized in the liver, with 75–85% recovered in the effluent 16. Theoretically, the elimination of electronegative acetate ions from the liver during ethanol metabolism, mainly through acetic acid passive excretion, must necessarily result in the net output of intracellular H⁺ 22. Beside the energy cost relative to the contribution of Na⁺:HCO₃^- symport to H⁺ extrusion, our study of pHi regulation evidences that acetate excretion largely participates in the elimination of the proton excess arising from ethanol oxidation. The passive transport of acetate along its concentration gradient thus avoids a dramatic energy expenditure.

Male Wistar rats (80–120 g) fed ad libitum were anesthetized by intraperitoneal injection of pentobarbital sodium (50 mg/kg of rat). The technique of non-recirculating liver perfusion through the portal vein (antegrade perfusion) was described previously 23. The liver (4–6 g) perfused with Krebs-Heinseleit (KHB) buffer at 37°C was excised from the abdomen and transferred into a 20 mm-diameter NMR cell. KHB was composed of (in mM) 120 NaCl, 4.70 KCl, 1.20 MgSO₄, 25 NaHCO₃, 1.20 KH₂PO₄-K₂HPO₄, 1.30 CaCl₂, 0.30 Na-pyruvate and 2.10 Na-lactate (pH = 7.35 at 37°C). When any substrate or inhibitor was added, the perfusate pH was always adjusted.

The perfusate was pumped through a Silastic home-made oxygenator (95% O₂ + 5% CO₂). Constant perfusion flow (5 ml/min.g liver wet weight) ensured good oxygenation of the liver as judged from the effluent oxygen concentration. Perfusate temperature and pH were monitored both before entering and after leaving the liver by continuous-flow pH and temperature electrodes.

The standard sequence protocol at 37°C was as follows: (i) initial KHB perfusion (30 min), (ii) addition of ethanol (70–90 min) in a 1–100 mM concentration range.

The contribution of the glycolytic pathway to ATP production was estimated in real-time by using 0.5 mM iodacetate (IAA), an inhibitor of glyceraldehyde-3 phosphate dehydrogenase activity. IAA was added within 2 min in the following conditions: (i) after a 30-min perfusion of KHB and (ii) after a 30-min perfusion of ethanol.

However, other sequences of IAA addition have been applied in this study according to the goal. These sequences will be specified with the results. For instance, in order to better analyze the contribution of the glycolytic pathway in presence of ethanol, the oxidative phosphorylation was inhibited in a first step by addition of the specific inhibitor of the cytochrome oxidase (KCN, 2.5 mM) leading to an increase of the contribution of glycolysis alone to the ATP supply. When the new steady state was obtained, ethanol 10 mM was added and compared to an addition of 0.5 mM IAA (2 min). When necessary, the other sequences will be detailed.

The contribution of acetic acid and Na⁺-HCO₃^- symport to pHi regulation in the presence of ethanol (50 mM) was respectively studied by (i) reducing the intra-extracellular acetate gradient with the addition of 20 mM acetate in the perfusate, and (ii) using 1 mM 4-acetamido-4'-isothiocyanostilbene-2,2'-disulfonic acid (SITS), an inhibitor of the symport.

The indirect contribution of Na⁺-K⁺-ATPase pump to pHi regulation in the presence of ethanol (50 mM) was evaluated using its specific inhibitor ouabaïn (1 mM).

For each group, the number of experiments is specified in the results section.

The study is complied with the Guide for the Care and Use of Laboratory Animals, U.S. Department of Health and Humans Services, N.I.H. Publication n° 85-23, revised (1985).

³¹P and ¹³C NMR spectra of liver were recorded on a Bruker DPX-400 spectrometer operating at 9.4T (161.9 and 100.6 MHz respectively). For ³¹P NMR spectra without proton decoupling, 240 free induction decays (70° flip angle, 0.2 s acquisition time, 0.5 s total recycling time, 10000 Hz spectral width and 4 K data points) were routinely acquired within 2 min, except during the periods corresponding to the addition of the ATP synthesis inhibitor where the number of scans was reduced to 70 FIDs to increase the accuracy of the rate determination. ¹³C NMR spectra (500 FID, 66° radiofrequency pulse repeated every second, 25000 Hz spectral width) were proton-decoupled using a gated bilevel mode, with a high level of ¹H irradiation used during acquisition and a low level during relaxation delay. A Lorentzian line broadening of 15 Hz was applied prior to Fourier transformation by the Bruker software for both ³¹P and ¹³C NMR spectra. Peak assignment was done as previously described 23 using methylene diphosphonic acid (MDPA set at 18.40 ppm) as an external standard and glycerophosphorylcholine as an internal standard set at 0.47 ppm.¹³C chemical shifts were expressed from an external silicone reference (at 1.45 ppm). Phosphorylated metabolite content [ATP, phosphomonoesters (PME; namely sn-glycerol-3-phosphate)] amount were estimated from their peak areas and expressed as a percentage of the initial value obtained at the end of the 30 min-equilibration period in KHB. In physiological conditions, cellular ATP is mainly in the bound form, especially Mg²⁺-ATP, corresponding to about 90% of the total ATP 24. It can be thus interesting to distinguish the two cellular forms (free and bound) of ATP, by the ³¹P NMR chemical shifts between the α and βATP resonances, which resulted from the binding of divalent cations (Mg²⁺) to ATP 24.

The time course of ATP content during the first 10–15 min of ethanol perfusion could be fitted approximately as a linear regression (validated by an r2 value > 0.91) of the form:

ATP (t) =-kt+b, where k is the apparent velocity of the ATP decrease and b the liver ATP content at the beginning of the substrate perfusion. In order to compare all experimental groups, the values at the beginning of ATP depletion were arbitrarily fixed at 100%.

¹³C NMR quantification of glycogen content at the beginning of the KHB perfusion was performed from a calibration curve established with oyster glycogen (0 to 185 mM glycosyl units). The change of glycogen content was estimated from the C-1 peak area and expressed as a percentage of this initial value.

All chemical compounds are provided by Sigma.

Changes in intracellular pH (pHi) were determined from the chemical shift of the intracellular inorganic phosphate (Pi) signal using a titration curve constructed at 37°C according to the Henderson-Hasselbach equation 23. The accuracy of the NMR pHi determination for the same liver was <0.015 pH unit.

Liver effluent (3 ml) was collected in order to enzymatically determine glucose (glucose oxidase, Sigma France) excretion at each phase of the protocol. Glucose excretion was expressed in μmoles/min.g liver ww.

All results are expressed as means ± SEM. Since the factor analyzed was the nature of the perfusate, one-way analysis of variance (ANOVA) was applied before a t-test (software package statworks, Cricket Software, London). A P value lower than 0.05 was considered to be significant.