Inhibition of glucose production during hepatic nerve stimulation in regenerating rat liver perfused in situ. Possible involvement of gap junctions in the action of sympathetic nerves

Hepatic nervous system and neurobiology of the liver

The liver has a nervous system containing both afferent and efferent neurons that are involved in a number of processes. The afferent arm includes the sensation of lipids, glucose, and metabolites (after eating and drinking) and triggers the nervous system to make appropriate physiological changes. The efferent arm is essential for metabolic regulation, modulation of fibrosis and biliary function and the control of a number of other processes. Experimental models have helped us to establish how: (i) the liver is innervated by the autonomic nervous system; and (ii) the cell types that are involved in these processes. Thus, the liver acts as both a sensor and effector that is influenced by neurological signals and ablation. Understanding these processes hold significant implications in disease processes such as diabetes and obesity, which are influenced by appetite and hormonal signals.

Dietary fat sensing via fatty acid oxidation in enterocytes: possible role in the control of eating

Various mechanisms detect the presence of dietary triacylglycerols (TAG) in the digestive tract and link TAG ingestion to the regulation of energy homeostasis. We here propose a novel sensing mechanism with the potential to encode dietary TAG-derived energy by translating enterocyte fatty acid oxidation (FAO) into vagal afferent signals controlling eating. Peripheral FAO has long been implicated in the control of eating ( 141 ). The prevailing view was that mercaptoacetate (MA) and other FAO inhibitors stimulate eating by modulating vagal afferent signaling from the liver. This concept has been challenged because hepatic parenchymal vagal afferent innervation is scarce and because experimentally induced changes in hepatic FAO often fail to affect eating. Nevertheless, intraperitoneally administered MA acts in the abdomen to stimulate eating because this effect was blocked by subdiaphragmatic vagal deafferentation ( 21 ), a surgical technique that eliminates all vagal afferents from the upper gut. These and other data support a role of the small intestine rather than the liver as a FAO sensor that can influence eating. After intrajejunal infusions, MA also stimulated eating in rats through vagal afferent signaling, and after infusion into the superior mesenteric artery, MA increased the activity of celiac vagal afferent fibers originating in the proximal small intestine. Also, pharmacological interference with TAG synthesis targeting the small intestine induced a metabolic profile indicative of increased FAO and inhibited eating in rats on a high-fat diet but not on chow. Finally, cell culture studies indicate that enterocytes oxidize fatty acids, which can be modified pharmacologically. Thus enterocytes may sense dietary TAG-derived fatty acids via FAO and influence eating through changes in intestinal vagal afferent activity. Further studies are necessary to identify the link between enterocyte FAO and vagal afferents and to examine the specificity and potential physiological relevance of such a mechanism.

The role of the autonomic nervous liver innervation in the control of energy metabolism

Despite a longstanding research interest ever since the early work by Claude Bernard, the functional significance of autonomic liver innervation, either sympathetic or parasympathetic, is still ill defined. This scarcity of information not only holds for the brain control of hepatic metabolism, but also for the metabolic sensing function of the liver and the way in which this metabolic information from the liver affects the brain. Clinical information from the bedside suggests that successful human liver transplantation (implying a complete autonomic liver denervation) causes no life threatening metabolic derangements, at least in the absence of severe metabolic challenges such as hypoglycemia. However, from the benchside, data are accumulating that interference with the neuronal brain-liver connection does cause pronounced changes in liver metabolism. This review provides an extensive overview on how metabolic information is sensed by the liver, and how this information is processed via neuronal pathways to the brain. With this information the brain controls liver metabolism and that of other organs and tissues. We will pay special attention to the hypothalamic pathways involved in these liver-brain-liver circuits. At this stage, we still do not know the final destination and processing of the metabolic information that is transferred from the liver to the brain. On the other hand, in recent years, there has been a considerable increase in the understanding which brain areas are involved in the control of liver metabolism via its autonomic innervation. However, in view of the ever rising prevalence of type 2 diabetes, this potentially highly relevant knowledge is still by far too limited. Thus the autonomic innervation of the liver and its role in the control of metabolism needs our continued and devoted attention.

Control of hepatocyte metabolism by sympathetic and parasympathetic hepatic nerves

More than any other organ, the liver contributes to maintaining metabolic equilibrium of the body, most importantly of glucose homeostasis. It can store or release large quantities of glucose according to changing demands. This homeostasis is controlled by circulating hormones and direct innervation of the liver by autonomous hepatic nerves. Sympathetic hepatic nerves can increase hepatic glucose output; they appear, however, to contribute little to the stimulation of hepatic glucose output under physiological conditions. Parasympathetic hepatic nerves potentiate the insulin-dependent hepatic glucose extraction when a portal glucose sensor detects prandial glucose delivery from the gut. In addition, they might coordinate the hepatic and extrahepatic glucose utilization to prevent hypoglycemia and, at the same time, warrant efficient disposal of excess glucose.

Functional and chemical anatomy of the afferent vagal system

The results of neural tracing studies suggest that vagal afferent fibers in cervical and thoracic branches innervate the esophagus, lower airways, heart, aorta, and possibly the thymus, and via abdominal branches the entire gastrointestinal tract, liver, portal vein, billiary system, pancreas, but not the spleen. In addition, vagal afferents innervate numerous thoracic and abdominal paraganglia associated with the vagus nerves. Specific terminal structures such as flower basket terminals, intraganglionic laminar endings and intramuscular arrays have been identified in the various organs and organ compartments, suggesting functional specializations. Electrophysiological recording studies have identified mechano- and chemo-receptors, as well as temperature- and osmo-sensors. In the rat and several other species, mostly polymodal units, while in the cat more specialized units have been reported. Few details of the peripheral transduction cascades and the transmitters for signal propagation in the CNS are known. Glutamate and its various receptors are likely to play an important role at the level of primary afferent signaling to the solitary nucleus. The vagal afferent system is thus in an excellent position to detect immune-related events in the periphery and generate appropriate autonomic, endocrine, and behavioral responses via central reflex pathways. There is also good evidence for a role of vagal afferents in nociception, as manifested by affective-emotional responses such as increased blood pressure and tachycardia, typically associated with the perception of pain, and mediated via central reflex pathways involving the amygdala and other parts of the limbic system. The massive central projections are likely to be responsible for the antiepileptic properties of afferent vagal stimulation in humans. Furthermore, these functions are in line with a general defensive character ascribed to the vagal afferent, paraventricular system in lower vertebrates.

Vagus nerve is involved in lack of blood reflow into sinusoids after rat hepatic ischemia

Although recovery of microcirculation is an important determinant for ischemia-reperfusion injury, little information is available about hepatic blood flow after ischemia. To examine regulatory mechanisms of postischemic hepatic microcirculation, we studied the sinusoidal blood flow after portal triad clamping of rat livers for 5, 15, or 30 min. Hepatic tissue blood flow and erythrocyte blood flow in sinusoids were measured using a laser-Doppler flowmeter and an intravital microspectroscope, respectively. There was a time of no blood flow (lag time) in sinusoids after declamping, dependent on the ischemic time. Cholinergic blockade agents eliminated the lag time, whereas nerve stimulation at the hiatus esophagus or on the hepatoduodenal ligament during reperfusion prolonged it. Chemical denervation with 10% phenol or surgical denervation on the hepatoduodenal ligament eliminated the lag time. The prolongation of lag time by nerve stimulation was completely abrogated by truncal vagotomy. These results suggest that the cholinergic vagus nerve is involved in causing the lag time of sinusoidal blood flow in hepatic ischemia-reperfusion.

Isolated hepatic cholinergic denervation impairs glucose and glycogen metabolism

BACKGROUND

Hepatic innervation plays an essential role in insulin extraction and glucose production, but the specific role of hepatic cholinergic innervation remains unclear. We sought to establish a model of isolated hepatic cholinergic denervation (IHCD), and to assess whether glycogen storage or the control of net hepatic glucose production (HGP) was altered by IHCD.

MATERIALS AND METHODS

Sprague-Dawley rats underwent either hepatic vagotomy or sham operation. Liver tissue was stained for vesicular acetylcholine transporter (VAChT) and (nonspecific neural) protein gene product 9. 5 (PGP) for verification of IHCD. Liver glycogen content was quantified in fed and fasted IHCD or sham-operated animals. HGP was determined after single-pass isolated liver perfusion, during which a 30-min 12 ng/ml glucagon infusion was begun after equilibration, and after 10 min, a 200 microU/ml insulin infusion was added.

RESULTS

Uniform staining of PGP and absence of VAChT staining in hepatic vagotomized rats demonstrated the validity of our model. Glycogen content of sham-operated livers (n = 8) increased from 6.0 +/- 1.7 in the fasting state to 10.6 +/- 1.8 mg/g liver, after feeding (P < 0.05). IHCD livers (n = 8) showed no comparable increase (3.5 +/- 0.6 to 4.0 +/- 0.7 mg/g liver). Perfusion with glucagon alone resulted in less HGP in IHCD livers (n = 12) compared with sham-operated livers (n = 10) (integrated HGP 3.3 +/- 0.3 mg/g liver min(-1) vs 5.1 +/- 0.5 mg/g liver min(-1), P < 0.05). Insulin infusion revealed impaired responsiveness to insulin after IHCD; the ratio of HGP in the final 10 min of perfusion (glucagon and insulin) to HGP in the initial 10 min (glucagon alone) was 90.3 +/- 2.4% for IHCD livers versus 68.1 +/- 4.4% for sham-operated controls, respectively (P = 0.0002).

CONCLUSIONS

Our study shows that IHCD results in significant impairment in liver glycogen storage and impaired hepatic sensitivity to glucagon and, possibly, to insulin. We conclude that hepatic cholinergic integrity is essential to normal hepatic glucose metabolism.

Receptor subtype mediating the action of circulating endothelin on glucose metabolism and hemodynamics in perfused rat liver

The subtype of endothelin receptor that mediates metabolic and hemodynamic effects of circulating endothelin was explored using perfused rat liver. Infusion of endothelin (ET)-1 or ET-3 into the portal vein at a concentration of 0.3 nM increased glucose and lactate output and decreased perfusion flow, although ET-3 was less effective than ET-1. The metabolic effects of ET-1 were observed even under costant-flow perfusion. Infusion of either sarafotoxin S6b or S6c, an ET(A)- or ET(B)-receptor agonist, mimicked the actions of ET-1 to an equal extent. The flow reduction and glucose production induced by ET-1 were partly attenuated by the ET(A)-receptor antagonist BQ485. By contrast, ET(B)-receptor antagonist BQ788 enhanced glucose production caused by ET-1 and ET-3 without affecting the hemodynamic change. The effects of ET-1 and ET-3 were almost totally inhibited by the combination of BQ485 and BQ788. These results suggest that both ET(A) and ET(B) receptors are involved in the metabolic and hemodynamic effects of circulating endothelin in rat liver, while the ET(A)-receptor-mediated action appears to be dominant.

Gap junctional communication and regulation of the glycogenic response to insulin by cell density and glucocorticoids in cultured fetal rat hepatocytes

Cell culture studies have revealed that metabolic functions of the adult hepatocyte are related to cell density. Development of the glycogenic response to insulin under glucocorticoid control was investigated in 15- and 18-day-old fetal rat hepatocytes plated at different cell densities. After culturing for 48 hours with glucocorticoids, the stimulatory effect of insulin on [14C]glucose incorporation into glycogen after 3 hours progressed from weak response (less than 1.4-fold) in sparse cultures to a maximal response in dense ones (3.0- to 4.5-fold), depending on the fetal stage. The response was always no more than 2.0-fold in the absence of glucocorticoids, even with dense cultures. Such a dual regulation pattern was not found for the glycogenolytic effect of glucagon similarly expressed regardless of culture conditions. When cells were clustered in limited circular regions of the dish, the insulin response was higher than for sparse cultures for a similar number of cells per culture. Using the scrape-loading technique with Lucifer Yellow CH, a positive dye transfer was obtained in clustered cultures providing that they were grown in the presence of glucocorticoids; insulin as well as glucagon stimulated twofold intercellular communication. Connexin32 (Cx32) and connexin26 (Cx26) protein levels were assayed by Western immunoblotting and developed according to age and exposure to glucocorticoids. Thus, glucocorticoids through development of gap junctions enabled establishment of intercellular communication that could be stimulated by insulin and glucagon in cultured fetal hepatocytes. Gap junction functioning and the biologic effect of insulin correlated closely.

Connexin 32 gap junctions enhance stimulation of glucose output by glucagon and noradrenaline in mouse liver

Gap junctions connect neighboring cells via intercellular channels composed of connexins (Cx). Connexin 32 (Cx32) is the main connexin in hepatocytes. Gap junctions propagate a signal from periportal to perivenous hepatocytes generated by electrical stimulation of sympathetic liver nerves. Therefore, it was the aim of this study to examine the involvement of hepatocellular gap junctions in hormonal regulation. In perfused livers from wild-type mice and Cx32-deficient mice, the stimulation of glucose release by varying noradrenaline and glucagon concentrations was investigated. At saturating hormone concentrations, glucose release was the same in wild-type and Cx32-deficient livers. However, glucose output was significantly smaller in Cx32-deficient than wild-type livers at half-maximally effective hormone concentrations. Because the two hormones circulate at less than half-saturating concentrations and because they are degraded during passage of blood through the liver, they lose efficiency from the periportal to the perivenous zone. In wild-type livers, this decrease in efficiency can be partially compensated by intercellular signal propagation through gap junctions, resulting in higher hormone actions than in Cx32-deficient livers. It is concluded that gap junctions are not only involved in intercellular propagation of nervous, but also of hormonal signals from periportal to perivenous hepatocytes.

Prolonged decrease in hepatic connexin32 in chronic liver injury induced by carbon tetrachloride in rats

BACKGROUND/AIMS

Our previous study indicated that the amount of connexin32, the major gap-junctional protein of rat liver, is transiently reduced in acute liver injury after single administration of hepatotoxic chemicals. This study was designed to examine alteration in the expression of connexin32 in chronic liver injury, unassociated with hepatocyte proliferation.

METHODS

Rats were injected with carbon tetrachloride (CCI4, 0.5 ml/kg) twice a week for 12 weeks. After cessation of CCI4 injection, hepatic contents of connexin32 and its mRNA levels were measured by immunoblotting as well as immunohistochemical examination and by Northern-blot analysis.

RESULTS

The plasma alanine-aminotransferase activity was increased from 30 U/I to about 1000 U/l after 12 weeks of CCI4 injections, but recovered nearly to normal level in 7 days after cessation of the injection. Liver specimens 12 days after the last CCI4 injection appeared cirrhotic with a marked increase in fibrosis. Connexin32 contents in these livers decreased to about 37% of controls. The significant decrease in connexin32 content was sustained for at least 30 days and recovered to the control level by 60 days. The alteration of connexin32 content in chronically injured liver was confirmed immunohistochemically. The level of connexin32-mRNA, however, was not reduced, but rather increased by chronic injection of CCI4.

CONCLUSION

The results suggest that intercellular communication is disturbed in chronic liver injury, lasting even after recovery from the acute phase of injury. Since the mRNA levels of connexin32 were sustained, the prolonged decrease in connexin32 contents in these livers might be due to a post-transcriptional change that causes decrease in protein synthesis or a derangement of post-translational controls.

Defective propagation of signals generated by sympathetic nerve stimulation in the liver of connexin32-deficient mice

The gap junctional protein connexin32 is expressed in hepatocytes, exocrine pancreatic cells, Schwann cells, and other cell types. We have inactivated the connexin32 gene by homologous recombination in the mouse genome and have generated homozygous connexin32-deficient mice that were viable and fertile but weighed on the average approximately 17% less than wild-type controls. Electrical stimulation of sympathetic nerves in connexin32-deficient liver triggered a 78% lower amount of glucose mobilization from glycogen stores, when compared with wild-type liver. Thus, connexin32-containing gap junctions are essential in mouse liver for maximal intercellular propagation of the noradrenaline signal from the periportal (upstream) area, where it is received from sympathetic nerve endings, to perivenous (downstream) hepatocytes. In connexin32-defective liver, the amount of connexin26 protein expressed was found to be lower than in wild-type liver, and the total area of gap junction plaques was approximately 1000-fold smaller than in wild-type liver. In contrast to patients with connexin32 defects suffering from X chromosome-linked Charcot-Marie-Tooth disease (CMTX) due to demyelination in Schwann cells of peripheral nerves, connexin32-deficient mice did not show neurological abnormalities when analyzed at 3 months of age. It is possible, however, that they may develop neurodegenerative symptoms at older age.

Innervation of the liver: morphology and function

Although it has been known for many years that the liver receives a nerve supply, it is only with the advent of immunohistochemistry that this innervation has been analysed in depth. It is now appreciated not only that many different nerve types are present, but also that there are significant differences between species, especially in the degree of parenchymal innervation. This has stimulated more detailed investigation of the innervation of the human liver in both health and disease. At the same time, functional studies have been underlining the important roles that these nerves play in processes as diverse as osmoreception and liver regeneration. This article briefly reviews current understanding of the morphology and functions of the hepatic nerve supply.

Posttranscriptional regulation of connexin 32 expression in liver during acute inflammation

Gap junctions mediate the communication between adjacent cells in tissues. In the liver, connexin 32 (Cx32) subunits make up the predominating gap junctions. The expression of Cx32 gene has been observed to be down-regulated in response to inflammatory states and during liver regeneration. In the present study we attempt to elucidate the molecular mechanisms underlying the down-regulation of the Cx32 expression during acute inflammation. A decrease in the level of Cx32 mRNA in rat liver occurred between 3 and 6 h after intravenous administration of bacterial lipopolysaccharide (LPS), simultaneously with the induction of an acute inflammatory response characterized by an increase in the level for beta-fibrinogen and a reduction of phosphoenolpyruvate carboxykinase mRNA. The reduction in Cx32 steady-state mRNA levels appears to occur at the posttranscriptional level, since the rate of degradation of this message seems to be higher than the rate of transcription of the gene. Degradation of Cx32 mRNA was blocked by the administration of actinomycin D, but not by cycloheximide, prior to injection of LPS. The stabilization of Cx32 message by actinomycin D correlated with the preservation of Cx32 on the cell surface, which otherwise disappears after administration of LPS alone. These results suggest that cellular communication via gap junctions could be regulated at the level of gene expression, by a posttranscriptional mechanism, during acute inflammatory states.

The mechanism of action of hepatic sympathetic nerves on ketone-body output from perfused rat liver. The effect of the interaction of noradrenaline with ATP on the release of beta-hydroxybutyrate

The regulatory mechanism of ketone-body output by the hepatic sympathetic nerves was studied in rat liver perfused in situ. Enrichment of the perfusion medium with 1 mM octanoate increased the basal ketone-body output from the liver up to 1.5 mumol.min-1.g liver-1. Under these conditions, electrical stimulation of the hepatic nerves (20 V, 20 Hz, 2 ms) decreased the output of both acetoacetate and beta-hydroxybutyrate, and was accompanied by an accumulation of beta-hydroxybutyrate in the liver. The effects of nerve stimulation were inhibited by the alpha 1-antagonist bunazosin (10 microM). However, noradrenaline, a typical sympathetic neurotransmitter, at a concentration of 1 microM decreased the output of acetoacetate but did not affect beta-hydroxybutyrate output. Prostaglandin F2 alpha at a concentration of 10 microM produced an effect similar to treatment with noradrenaline, without a decrease in beta-hydroxybutyrate output. ATP at 50 microM, however, decreased the output of both acetoacetate and beta-hydroxybutyrate and increased the tissue concentration of beta-hydroxybutyrate, mimicking the effect of nerve stimulation. Moreover, in the presence of 0.2 microM ATP, a concentration that produced neither metabolic nor hemodynamic changes, noradrenaline (1 microM) was shown to decrease the beta-hydroxybutyrate output. These results indicate the possible involvement of ATP in the action of hepatic sympathetic nerves on beta-hydroxybutyrate output from the liver, presumably through its interaction with noradrenaline.

Comparative study of the mammalian liver innervation: an immunohistochemical study of protein gene product 9.5, dopamine beta-hydroxylase and tyrosine hydroxylase

The liver innervation of eight different mammalian species was examined by immunohistochemical localization of protein gene product (PGP) 9.5 to visualize the general innervation for autonomic nerve fibres. In addition, dopamine beta-hydroxylase (DBH) and tyrosine hydroxylase (TH), two enzymes involved in catecholamine synthesis, were localized immunohistochemically to delineate hepatic sympathetic nerve fibres. We found that: (1) Within the interlobular region of each species, PGP 9.5, DBH and TH-positive nerve fibres were all seen in close association with branches of hepatic arteries, portal veins and bile ducts. (2) Within the parenchyma of the guinea-pig, cat, dog, pig, monkey and human liver, the presence of the three immuno-positive nerve fibres could be unequivocally identified, although the density of these intralobular fibres showed marked species variation. Moreover, immunoelectron microscopic study confirmed that PGP 9.5-positive nerve terminals of the human liver are in close apposition to hepatocytes. (3) In mouse and rat, no parenchymal nerve fibres immunoreactive for PGP 9.5, TH or DBH could be demonstrated.

Differential regulation of multiple gap junction transcripts and proteins during rat liver regeneration

The mRNA and protein expression of alpha 1 (connexin 43), beta 1 (connexin 32), and beta 2 (connexin 26) gap junction genes were examined in the regenerating rat liver after 70% partial hepatectomy (PH). Expression of beta 1 and beta 2 steady-state mRNA levels changed minimally until 12 h after PH when both transcripts decreased to approximately 15% of baseline values. A similar decrease in assembled connexin levels was detected by immunoblot and indirect immunofluorescence at 18 h after PH. Both transcripts simultaneously increased between 24 and 42 h and again rapidly decreased by 48 h post-PH. beta 1 and beta 2 assembled gap junction protein expression increased at 48 h post-PH and rapidly decreased by 56 h. By 72 to 84 h post-PH, beta 1 and beta 2 mRNA and assembled protein expression returned to near baseline levels and were maintained. Interestingly, inhibition of protein synthesis with cycloheximide completely inhibited disappearance of the beta 2 transcript, in contrast to beta 1 mRNA which was unaffected. Nuclear run-on assays showed no change in transcriptional rates for either gene during the regenerative period. However, both beta 1 and beta 2 transcripts exhibited significantly decreased mRNA half-lives at 12 h post-PH (3.8 and 3.7 h, respectively) relative to those at 0 h (10.9 and 6.1 h, respectively). Surprisingly, although the transcriptional rate for alpha 1 was similar to that observed for beta 2, no alpha 1 transcripts were detectable by northern or RNase protection analysis. The results suggest that in the regenerating rat liver, beta 1 and beta 2 gap junction genes are not regulated at the transcriptional level. Rather, the cyclical modulation of their steady-state transcripts is regulated primarily by posttranscriptional events of which mRNA stability is at least one critical factor in the control process.

Alteration in sympathetic nerve activity during liver regeneration in rats after partial hepatectomy

To determine if the sympathetic nerve has a role in liver regeneration, the alterations of tissue norepinephrine (NE) content and its turnover were measured in rats after partial hepatectomy. NE content per unit liver mass decreased progressively to about one-fourth of controls by the ninth day after partial hepatectomy. Since the total amount of NE in the whole liver did not change during nine days of liver regeneration, it was supposed that sympathetic innervation could make slower progress than proliferation of hepatocytes. Fractional turnover rate of NE was reduced transiently in regenerating liver 24-48 h after partial hepatectomy and recovered to normal 8-9 days after the operation. Such a transient reduction of NE turnover was observed specifically in the liver. These results suggest that sympathetic nerve activity of the liver is suppressed at the early stage of regeneration.

Signal propagation via gap junctions, a key step in the regulation of liver metabolism by the sympathetic hepatic nerves

Cell-to-cell communication via gap junctions has been proposed to be involved in the metabolic actions of sympathetic liver nerves in the rat. The effects of hepatic nerve stimulation and noradrenaline-, PGF2 alpha- and glucagon infusion on glucose metabolism and perfusion flow were studied in perfused rat liver in the absence and presence of the gap junctional inhibitors, heptanol, carbenoxolone and (4 beta)phorbol 12-myristate 13-acetate (4 beta PMA). (i) Stimulation of the hepatic nerve plexus increased glucose output, decreased flow and caused an overflow of noradrenaline into the hepatic vein. (ii) Heptanol completely inhibited not only the nerve stimulation-dependent metabolic and hemodynamic alterations but also the noradrenaline overflow. Thus the heptanol-dependent inhibitions were caused primarily by a strong impairment of transmitter release. (iii) Carbenoxolone inhibited the effects of neurostimulation on glucose metabolism partially by about 50%, whereas it left perfusion flow and noradrenaline overflow essentially unaltered. (iv) 4 beta PMA reduced the nerve stimulation-dependent enhancement of glucose release by about 80% but the noradrenaline-dependent increase in glucose output only by about 30%; the increase in glucose release by PGF2 alpha and by glucagon remained essentially unaltered. 4 beta PMA reduced the nerve stimulation-dependent decrease in portal flow by about 35% but did not affect the noradrenaline-and PGF2 alpha-elicited alterations, nor did it alter noradrenaline overflow. The results allow the conclusion that gap junctional communication plays a major role in the regulation of hepatic carbohydrate metabolism by sympathetic liver nerves, but not by circulating noradrenaline, PGF2 alpha or glucagon.