Search

There is evidence to suggest that integrity of the neurovascular unit may be compromised in acute liver failure (ALF). In order to address this issue from a molecular standpoint, expression of an array of genes coding for key cerebrovascular endothelial cell and tight junction proteins were measured by reverse transcription-polymerase chain reaction in cerebral cortex of rats with ischemic liver failure resulting from hepatic devascularization (portacaval anastomosis followed 24h later by hepatic artery ligation) compared to appropriate sham-operated controls. Expression of P-glycoprotein, endothelin-1, von Willebrand factor, caveolin-1, occludin, and the endothelial nitric oxide synthase isoform (eNOS) were measured in brain extracts from rats with ALF at coma/edema stages of encephalopathy. The effects of mild hypothermia (35 degrees C) sufficient to prevent cerebral edema in ALF animals on the expression of these genes were also studied. Brain edema and hepatic coma in normothermic ALF rats was accompanied by selective increases in expression of eNOS. Expression of occludin and von Willebrand factor mRNAs were decreased at coma/edema stages of encephalopathy in ALF rats whereas, expression of other cerebrovascular endothelial cell markers endothelin-1, P-glycoprotein, and caveolin-1 were unaffected. Mild hypothermia led to normalization of brain water content and of eNOS mRNA. However, the correlation between increased eNOS expression and encephalopathy/edema grade was poor suggesting the existence of additional mechanisms. These findings underscore the multifactorial nature of brain edema/encephalopathy mechanisms in ALF and question the role of BBB breakdown as a major pathogenetic factor.

This review briefly summarizes what is known about the dynamic morphology of the hepatic microvascular system that includes all vessels in the liver with a diameter less than 300 microm and various morphological sites within these vessels that regulate the distribution of blood flow. The latter include the various segments of the afferent portal venules and hepatic arterioles, the sinusoids, and central and hepatic venules. Sinusoids are unique exchange vessels lined by fenestrated endothelial cells which have important endocytotic functions and phagocytic Kupffer cells which are important for host defense. These are encircled by extraluminal stellate cells that are specialized pericytes containing fat droplets that store vitamin A. The principle sites for regulating blood flow are in the sinusoidal network with stellate and endothelial cells playing a major role in regulating the diameters of sinusoids and the distribution of blood flow in individual sinusoids, lobules, or segments of lobules. The sinusoidal endothelial cells are a sensitive and early target for several toxicants. For example, as early as 30 minutes after the administration of acetaminophen, the endothelial cells become swollen and begin to lose the ability to endocytose ligands. Within 2 hr, gaps through the cytoplasm appear formed by the destruction and/or coalescence of fenestrae which permit red blood cells to penetrate into the space of Disse. Subsequently, the sinusoid may collapse or disintegrate reducing blood flow.

Catecholamines are often administered during and after liver transplantation (LTx) to support systemic perfusion and to increase organ oxygen supply. Some vasoactive agents can compromise visceral organ perfusion. We followed the hypothesis that the vasculature of transplanted livers presents with a higher sensitivity, which leads to an increased vulnerability for flow derangement after application of epinephrine (Epi) or norepinephrine (NorEpi). Hepatic macroperfusion and microperfusion during systemic Epi or NorEpi infusion were measured by Doppler flow and thermodiffusion probes in porcine native, denervated, and transplanted livers (n = 16 in each group). Epi or NorEpi were infused (n = 8 in each subgroup) in predefined dosages (low dose = 5 microg/kg/minute and high dose = 10 microg/kg/minute) over 240 minutes. Systemic cardiocirculatory parameters were monitored continuously. Hepatic perfusion data were compared between all groups at comparable time points and dosages. In all native, denervated, and transplanted liver groups, Epi and NorEpi induced an inconsistent rise of mean arterial pressure and heart rate shortly after onset of infusion in both dosages compared with baseline. No significant differences of cardiovascular parameters at comparable time points were observed. In native livers, Epi and NorEpi induced only temporary alterations of hepatic macrocirculation and microcirculation, which returned to baseline 2 hours after onset of infusion. No significant alterations of hepatic blood flow were detected after isolated surgical denervation of the liver. By contrast, transplanted livers showed a progressive decline of hepatic macrocirculation (33-75% reduction) and microcirculation (39-58% reduction) during catecholamine infusions in a dose-dependent fashion. Characteristics of liver blood flow impairment were comparable for both vasoactive agents. In conclusion, pronounced disturbances of hepatic macrocirculation and microcirculation were observed during systemic Epi and NorEpi infusion after LTx compared with native and denervated livers. Microcirculation disturbances after LTx might be explained by impairment of hepatic blood flow regulation caused by an increased sensitivity of hepatic vasculature after ischemia-reperfusion and by lengthening of vasopressor effects caused by reduced hepatocyte metabolism. Clinicians should be aware of this potentially hazardous effect. Therefore, application of catecholamines after clinical LTx should be indicated carefully.

The role of neural elements in regulating blood flow through the hepatic sinusoids, solute exchange, and parenchymal function is incompletely understood. This is due in part to limited investigation in only a few species whose hepatic innervation may differ significantly from humans. For example, most experimental studies have used rats and mice having livers with little or no intralobular innervation. In contrast, most other mammals, including humans, have aminergic and peptidergic nerves extending from perivascular plexus in the portal space into the lobule, where they course in Disse's space in close relationship to stellate cells (fat storing cells of Ito) and hepatic parenchymal cells. While these fibers extend throughout the lobule, they predominate in the periportal region. Cholinergic innervation, however, appears to be restricted to structures in the portal space and immediately adjacent hepatic parenchymal cells. Neuropeptides have been colocalized with neurotransmitters in both adrenergic and cholinergic nerves. Neuropeptide Y (NPY) has been colocalized in aminergic nerves supplying all segments of the hepatic-portal venous and the hepatic arterial and biliary systems. Nerve fibers immunoreactive for substance P and somatostatin follow a similar distribution. Intralobular distribution of all of these nerve fibers is species-dependent and similar to that reported for aminergic fibers. Vasoactive intestinal peptide and calcitonin gene-related peptide (CGRP) are reported to coexist in cholinergic and sensory afferent nerves innervating portal veins and hepatic arteries and their branches, but not the other vascular segments or the bile ducts. Nitrergic nerves immunoreactive for neuronal nitric oxide (nNOS) are located in the portal tract where nNOS colocalizes with both NPY- and CGRP-containing fibers. In summary, the liver is innervated by aminergic, cholinergic, peptidergic, and nitrergic nerves. While innervation of structures in the portal tract is relatively similar between species, the extent and distribution of intralobular innervation are highly variable as well as species-dependent and may be inversely related to the density of gap junctions between contiguous hepatic parenchymal cells., (Copyright 2004 Wiley-Liss, Inc.)

Background: Lafutidine, (+/-)-2-(furfurylsulfinyl)-N-[4-[4-(piperidinomethyl)-2-pyridyl]oxy-(Z)-2 butenyl]acetamide, is a newly synthesized histamine H2 receptor antagonist and possesses a cytoprotective efficacy, which comprises mucin biosynthesis and stimulation of gastric blood flow mediated through capsaicin-sensitive sensory neurons and endogenous calcitonin gene-related peptide (CGRP). In the present study, an effect of lafutidine on hepatic blood flow was investigated in rats that received an intracisternal injection of a subthreshold dose of thyrotropin-releasing hormone (TRH) analog, RX 77368., Methods: Change in hepatic blood flow was determined by laser Doppler flowmetry. Male Wistar rats were anesthetized with urethane (1.5 g/kg, i.p.), and positioned on a stereotaxic apparatus. An abdominal incision was made, and a probe of laser Doppler flowmeter was placed on the surface of the liver. After a 60-min stabilization, basal hepatic blood flow was measured for 30 min, and lafutidine (0.5, 1, 3, 5 or 10 mg/kg) or vehicle was injected into the portal vein and a subthreshold dose (1.5 ng) of RX 77368 was injected intracisternally. Hepatic blood flow was monitored for 120 min postinjection. To investigate a role of capsaicin-sensitive sensory neurons and endogenous CGRP, systemic capsaicin treatment (125 mg/kg, s.c., 10-14 days before) and intravenous infusion of a CGRP receptor antagonist, human CGRP-(8-37) (15 micro g/kg as a bolus, followed by infusion at 3 micro g/kg/h) were performed, respectively., Results: Intracisternal injection of RX 77368 (1.5 ng) or intraportal lafutidine (10 mg/kg) by itself did not affect hepatic blood flow, but co-injection of intracisternal RX 77368 (1.5 ng) and intraportal lafutidine (5 mg/kg) increased it with peak response at 30 min postinjection. The effect of lafutidine on hepatic blood flow in rats given RX 77368 was dose-related over the range 1-5 mg/kg. By contrast, intracisternal injection of RX 77368 (1.5 ng) did not change hepatic blood flow in rats injected with another histamine H2 receptor antagonist, famotidine (5 mg/kg), intraportally. The stimulatory effect of co-injection of TRH analog and lafutidine was abolished by systemic capsaicin-treatment and CGRP antagonist., Conclusion: These data suggest that lafutidine increases hepatic blood flow by sensitizing the liver to the action of central TRH via both capsaicin-sensitive sensory neurons and endogenous CGRP in urethane-anesthetized rats.

Background: . The effects of dopamine (DA) on systemic hemodynamics are better understood than its effects on hepatic hemodynamics, especially after liver denervation occurring during liver transplantation. Therefore, a porcine model was used to study DA's effects on hemodynamics after hepatic denervation., Materials and Methods: Fifteen pigs underwent laparotomy for catheter and flow probe placement. The experimental group (n = 7) also underwent hepatic denervation. After 1 week, all pigs underwent DA infusion at increasing doses (3-30 mcg/kg/min) while measuring hepatic parameters [portal vein flow (PVF), hepatic artery flow (HAF), total hepatic blood flow (THBF = HAF + PVF), portal and hepatic vein pressures] and systemic parameters [heart rate (HR), mean arterial pressure (MAP)]., Results: There was a significant increase in HAF from baseline to the 30 mcg/kg/min DA infusion rate (within-subjects P < 0.01), but the differences between the two groups were not significant. PVF and THBF showed large effects (increases) with denervation, but the increase in flow with DA infusion was not present after denervation. Perihepatic pressures were unchanged by denervation or DA. Heart rate differed significantly between the control and denervated animals at baseline, 3, 6, 12 (all P < 0.05), and 30 mcg/kg/min DA (P = 0.10). Control vs denervation MAP at baseline was 100 +/- 4 vs 98 +/- 4 Torr and at 30 mcg/kg/min it was 110 +/- 3 vs 101 +/- 5 mm Hg., Conclusions: Hepatic flows tended to be higher after denervation. HAF showed similar increases with DA in both control and denervation groups. Increases in PVF and THBF with DA infusion were not present after denervation. HR was significantly decreased and MAP tended to be lower after denervation. The HR and MAP response to DA was similar in both groups. Therefore, both denervation and DA infusion have an effect on systemic and hepatic hemodynamics., (Copyright 2001 Academic Press.)

Central neuropeptides play a role in many physiological regulatory processes through the autonomic nervous system. Thyrotropin-releasing hormone (TRH) is distributed in the central nervous system and acts as a neurotransmitter to regulate gastric functions through vagal-muscarinic pathways. The central effect of the TRH analog on hepatic blood flow was investigated in urethan-anesthetized rats. Hepatic blood flow was determined by the hydrogen gas clearance technique. Intracisternal injection of the stable TRH analog RX-77368 (5-100 ng) dose dependently increased hepatic blood flow with peak response at 15 min after the peptide was administered (net change from basal for vehicle and 5, 10, 100, and 500 ng RX-77368 was 2.0 +/- 0.2, 8.9 +/- 0.8, 19.4 +/- 2.6, 32.6 +/- 3.3, and 28.5 +/- 6.8 ml.min-1.100 g-1, respectively), and this stimulatory effect returned to baseline at 90 min. The stimulation of hepatic blood flow by the intracisternally administered TRH analog was abolished by atropine methyl nitrate (0.15 mg/kg ip), indomethacin (5 mg/kg ip), NG-nitro-L-arginine methyl ester (10 mg/kg iv), and hepatic branch vagotomy but not by cervical spinal cord transection (C6 level). Intravenous injection of RX-77368 did not have any effect on hepatic blood flow. These results indicate that TRH acts in the central nervous system to stimulate hepatic blood flow through vagal-muscarinic and indomethacin- and nitric oxide-dependent pathways.

The effects of hepatic denervation on liver blood flow and the influence of the denervation on the hemodynamic response to exogenously administered agents remain equivocal. This study examined the effects of the beta 2-adrenoceptor agonist, terbutaline, on liver blood flow and how these effects were influenced by surgical hepatic denervation in a conscious canine model. Two weeks after denervation, hepatic arterial flow, portal venous flow, and aortic blood flow were measured during intravenous infusion of terbutaline at 0.1, 0.2, 0.4, and 0.8 microgram/kg/min, and responses were compared between the hepatic denervated and sham-operated dogs. Implantable transit-time ultrasonic flowmeter and probes were used. Hepatic arterial flow increased dose-dependently in response to terbutaline in denervated dogs, while it remained unchanged in control dogs except at the highest dose of terbutaline. The increase in hepatic arterial flow of denervated dogs was significantly greater than that of controls at all doses of terbutaline. Aortic blood flow-indexed hepatic arterial flow in response to terbutaline also differed substantially between the two groups. Terbutaline produced a marked dose-dependent increase of portal venous flow similarly in both groups. These results suggest that the beta 2-adrenoceptor agonist terbutaline should be effective for increasing liver blood flow and could lessen postoperative liver ischemia. We speculate that hepatic denervation induces supersensitivity of the hepatic arterial system to intravenous terbutaline.

The effect of hepatic nerve stimulation and norepinephrine (NE) on the laser Doppler signal from the surface of the perfused rat liver was tested. The livers from male Wistar rats were perfused in situ via the portal vein with Krebs-Henseleit buffer containing 20% bovine erythrocytes (37 degrees C, pH 7.4) and total liver blood flow (TLBF) was by timed collection of effluent. Portal vascular resistance (PVR) was calculated from the pressure difference across the liver. Linearity of laser Doppler flowmetry (LDF) with TLBF was confirmed in all preparations. Stimulation of the hepatic nerves (2 ms, 20 V) was performed at frequencies between 0.5 and 20 Hz (n = 11). NE was added to the buffer at concentrations between 10(-10) and 10(-6) M (n = 8). A stimulus-dependent rise in PVR occurred during hepatic nerve stimulation (basal, 3.11 +/- 0.26 dyn s cm-5) and NE administration (basal, 2.62 +/- 0.29 dyn s cm-5), with a maximum effect at 20 Hz (311 +/- 45%) and 10(-6) M (591 +/- 72%), respectively. Both LDF and TLBF fell during nerve stimulation and NE. A linear relationship (r = 0.99; p < 0.001) between change in TLBF (%) and LDF flux (%) was found for NE (10(-10) to 10(-6) M). During nerve stimulation, the fall in TLBF and LDF flux was linear with the logarithm of stimulus frequency and reached a maximum at 10 and 20 Hz, respectively. At a stimulus frequency of 20 Hz, the change in LDF was significantly different from the change in TLBF (p < 0.001). We conclude from our findings that during high-frequency hepatic nerve stimulation, LDF underestimates TLBF.

Hepatic blood flow is regulated by the autonomic nervous system; however, blood flow through the denervated liver has been controversial. The purpose of this study is to evaluate the changes in hepatic blood flow caused by experimental hepatic denervation during the intravenous infusions of catecholamines (dopamine or dobutamine) with or without prior administration of alpha-adrenoceptor antagonist (phenoxybenzamine) or beta-adrenoceptor antagonist (propranolol). The liver blood flow was measured using transit time ultrasonic flow meter probes at the portal vein and hepatic artery in conscious dogs which (a) underwent hepatic denervation (denervation group, n = 9) and (b) were intact (control group, n = 10). Norepinephrine concentrations in the liver were determined to evaluate the effects of hepatic denervation and were decreased at 1 week and 4 weeks after hepatic denervation. In the control group, dopamine and dobutamine produced an increase of portal venous blood flow (PVF). Conversely, hepatic denervation reduced the increase in PVF by dopamine and dobutamine. Dopamine with prior administration of phenoxybenzamine produced a much larger increase in PVF in both groups. Pretreatment of propranolol in both groups abolished the increasing effects of dopamine and dobutamine in PVF. Dopamine reduced the hepatic arterial blood flow (HAF) regardless of hepatic denervation. During dobutamine infusion, HAF was decreased by hepatic denervation and prior administration of propranolol. These results suggest that hepatic denervation reduces the increasing effects of catecholamines on the hepatic blood flow through greater enhancement of alpha-adrenergic effects than of beta-adrenergic effects in the hepatic vascular autonomic nerve response.

In isolated in situ perfused rat livers, infusion of anaphylatoxins C3a and C5a, activation peptides of the complement system, as well as stimulation of sympathetic hepatic nerves have been shown to increase hepatic glucose output and to reduce hepatic flow. These effects were mediated via an at least partially prostanoid-dependent intercellular signalling chain between nonparenchymal cells and hepatocytes. Kupffer cells have been implicated as the source of prostanoids in the anaphylatoxin-dependent signalling chain and Ito cells in the nerve stimulation-dependent signalling chain, because anaphylatoxins and noradrenaline increased prostanoid synthesis in isolated Kupffer and Ito cells, respectively. To further corroborate this hypothesis, anaphylatoxins were infused and hepatic nerves were stimulated in perfused rat livers in which Kupffer cells had been largely depleted by treatment of the animals with gadolinium chloride. Native human anaphylatoxin C3a (nhC3a) and recombinant rat anaphylatoxin C5a (rrC5a) increased prostanoid formation as well as glucose output and reduced flow in perfused rat liver. In Kupffer cell-depleted rat livers, the nhC3a- and rrC5a-mediated prostanoid formation was nearly abolished, and the increase in glucose output and the reduction of flow were reduced to between 30% and 50% (area under the curve [AUC]) of control livers. In contrast, stimulation of hepatic nerves increased glucose output and reduced flow to a similar extent in Kupffer cell-depleted livers as in control livers. These results indicate that Kupffer cells were not involved in the prostanoid-mediated nerve stimulation-dependent increase in glucose output and reduction of flow. Kupffer cells seemed, however, to be at least one major source of the anaphylatoxin-mediated prostanoid formation and, consequently, stimulation of glucose release and flow reduction in perfused liver. Because the metabolic and hemodynamic anaphylatoxin effects were not completely blocked in livers of gadolinium-treated animals, either Kupffer cells may not have been entirely eliminated, or yet another nonparenchymal cell type and mediator might be involved in the anaphylatoxin-elicited intercellular communication between nonparenchymal cells and hepatocytes.

We investigated the effects of liver denervation on hemodynamic circulation in seven anesthetized pigs. Simultaneous measurements of the hepatic artery and portal vein were performed with an ultrasound Doppler flow meter before and after liver denervation. Neither resting systemic nor hepatic hemodynamics changed following liver denervation. However, temporary occlusion of the portal vein resulted in a significant increase in hepatic artery flow in the innervated liver (from 123 +/- 15 ml/min to 177 +/- 17 ml/min, p < .01), whereas, in the denervated liver, a significant decrease was observed (from 128 +/- 11 ml/min to 106 +/- 19 ml/min, p < .05). Thus, the reciprocity between the hepatic artery and portal vein in the hepatic artery flow during portal vein occlusion might intensify symptoms of portal vein thrombosis in liver transplantation. In the denervated liver, a significant decrease also occurred in systolic blood pressure and central venous pressure from 1 to 3 min after portal vein occlusion. Since the liver plays a crucial role in the maintenance of cardiovascular homeostasis during blood loss, it is likely that denervation at the porta hepatis induced a lack of vasoconstriction in the portal territory. Liver denervation might further exacerbate this response to hypotension. The current study confirms that the hepatic nerves play an important role in hepatic arterial and portal venous interactions aimed at maintaining a constant blood flow through the liver. We also suggest that the hepatic nerves are important for cardiovascular homeostasis.

Ito cells, located in the space of Disse, extend their numerous long cytoplasmic processes to surround the sinusoidal wall. Conventional and immune electron microscopy demonstrates abundant microfilaments and a great amount of actin and myosin in these cytoplasmic processes. These morphological findings suggest the possibility of control of the sinusoidal circulation by their contraction. Two mechanisms may be involved in their contraction: namely nervous and humoral factors. Ito cells, often in close contact with the nerve endings, which contain many granules of substance P (SP), have numerous receptors for this peptide. Ito cells also have receptors for endothelin-I (ET-1), more commonly found in the periportal area of the hepatic lobule. Experimental studies using cultured Ito cells showed their contraction following treatment with ET-1 or SP. These results suggest that Ito cells play an important role in the regulation of hepatic sinusoidal circulation.

Cats anesthetized with pentobarbital sodium were hemorrhaged (1 ml.min-1.kg body wt-1) until arterial pressure declined to 55 mmHg. Hepatic volume was recorded by plethysmography. In controls, 20 +/- 5 (SD) ml/kg was removed and hepatic volume decreased 3.6 +/- 0.8 ml/kg. Splanchnic nerve section or administration of hexamethonium-atropine reduced hemorrhage volumes (to 10.3 +/- 2.3 and 4.8 +/- 1.8 ml/kg, respectively) and liver volume changes (to 1.5 +/- 0.5 and 0.7 +/- 0.3 ml/kg, respectively). Section of only the hepatic nerves had no effect on hemorrhage volume or the decrease in hepatic volume. Section of the hepatic nerves after removal of the adrenals and kidneys also had no significant effect on hemorrhage volume or the decrease in liver volume. We conclude that the hepatic capacitance response to hemorrhage does not require direct sympathetic venoconstriction of the hepatic capacitance vessels. Changes in inferior vena caval pressure play a significant but small role. The splanchnic nerves (excluding hepatic nerves) play a major role, possibly by splanchnic arteriolar constriction. Whereas the liver comprises only 2.5% of the body weight, it contributed 18% of the hemorrhage volume.

To determine whether the liver or gut releases neuropeptide Y (NPY) from their sympathetic nerves, we performed bilateral thoracic sympathetic nerve stimulation (BTSNS) in halothane-anesthetized dogs and calculated gut and liver NPY spillover. BTSNS markedly increased hepatic NPY spillover (delta = +32 +/- 8 ng/min) and arterial NPY concentration (delta = +220 +/- 56 pg/ml), despite no effect on gut NPY spillover (delta = +8 +/- 7 ng/min). To determine the liver's contribution to this increase of circulating NPY, hepatic nerves were selectively stimulated (HNS). Liver NPY spillover increased markedly (delta = +114 +/- 42 ng/min, P < 0.025) during HNS, causing a large increase of arterial NPY (delta = +586 +/- 237 pg/ml, P < 0.025). Using this ratio of liver spillover to arterial increments of NPY, we calculated that the liver makes a major contribution (70%) to circulating NPY levels during BTSNS. The predominant form of canine NPY coeluted with synthetic [Met17]NPY and the minor form of canine NPY coeluted with the oxidized form of [Met17]NPY on high-performance liquid chromatography. We therefore conclude that dog NPY is likely [Met17]NPY and that the liver, rather than the gut, is a major source of circulating NPY during sympathetic nerve stimulation and perhaps stress.

The involvement of the sympathetic nervous system in the control of basal hepatic hemodynamics was investigated. Hepatic denervation was achieved by orthotopic transplantation or chemical denervation of the organ. In male Lewis rats, transplantation with rearterialization of the graft was performed. Chemical denervation was achieved by intraportal injection of 6-hydroxydopamine (75 mg/kg). Normal liver physiology was confirmed by histology and liver function tests. Four weeks post-transplantation and 7 days post-denervation, histological examination revealed no differences between transplanted, denervated and untreated or sham-operated control animals. Liver function measured by standard tests (e.g., plasma SGOT, bilirubin) was normal in all groups. The rate constants for aminopyrine breakdown in transplanted (0.015 +/- 0.005 min-1), denervated (0.015 +/- 0.0012 min-1) and control rats (0.015 +/- 0.001 min-1) were not significantly different. No significant difference in the rate of galactose breakdown was found. Total liver blood flow (measured by the 133Xe clearance technique in the anesthetized animal) was unaffected by transplantation (rate constant, 0.245 +/- 0.062 min-1; control 0.279 +/- 0.011 min-1). The interlobular distribution of portal blood flow was tested by intraportal injection of 51Cr-labelled microspheres. A linear relationship between flow to lobe and lobe size was confirmed in control (r = 0.95), denervated, (r = 0.99) and transplanted rats (r = 0.97) and the 'relative' flow to each lobe was not significantly different in the 3 groups. No significant differences in the 'core' to 'periphery' distribution of portal blood flow were found in the 3 groups. A small but significant portal systemic shunt was found in transplanted but not denervated or control animals.(ABSTRACT TRUNCATED AT 250 WORDS)

Hepatic capacitance responses were compared in sham-operated and 14-day bile duct-ligated cats under pentobarbital anesthesia. Both groups were subjected to splenectomy and had the anterior hepatic nerve plexus sectioned to allow stimulation; the posterior plexus was intact. Blood volume compensation for hemorrhage was reduced in the bile duct ligation group compared with the control group: The liver compensated for 20.1% and 10.6% of blood loss, respectively. Portal hypertension did not exist and hepatic compliance was unaltered despite the presence of severe biliary hyperplasia, portal tract distortion and fibrosis. The capacitance responses to hemorrhage in the bile duct ligation group were accounted for entirely by a passive compliant response to reduced portal pressure, whereas the sham surgery group showed an additional active component. Responses of liver volume and venous resistance to norepinephrine were normal, but responses to nerve stimulation were reduced. The nerve dysfunction was not universal; reflex arterial blood pressure response to carotid occlusion was normal. The data suggest that hepatic blood volume responses to hemorrhage is mediated by passive compliant responses to reduced portal pressure (stressed volume) and active responses (unstressed volume) to nerve stimulation but that long-term bile duct ligation produces selective hepatic neuropathy resulting in loss of the active component.

Dogs were anesthetized with pentobarbital sodium and placed on positive-pressure ventilation. The right phrenic nerve and/or its C5 branch were prepared for afferent recording. The hepatic veins, hepatic parenchyma, diaphragm, and inferior vena cava were studied for mechanoreceptors using light pressure and stroking as the stimuli. Mechanosensitive areas were found in the hepatic veins, hepatic parenchyma of the right medial lobe, and inferior vena cava. The hepatic vein and inferior vena caval receptors are located in the same 1- to 2-cm region as the sphincters that are found in these vessels. This study presents the first experimental evidence for the existence of hepatic vein receptors, hepatic parenchymal receptors, and inferior vena caval mechanoreceptors with phrenic afferents in the dog. These sensory areas of the circulation may be involved in the neural control of venous return as well as mediating changes in intrahepatic and portal venous blood pressure during normal respiration.

We earlier reported the existence of vagal arginine, alanine, and leucine sensors in the liver which modulate amino acid-induced pancreatic hormone secretion. To determine the possible existence of glycine sensors in the liver and their reflex effects, afferent discharges from the hepatic branch and efferent discharges from the pancreatic branch of the vagus nerve were recorded. Intraportal administration of 0.1, 1, and 10 mM of L-glycine solution (0.1 ml) depressed the afferent discharge rate of the hepatic branch of the nerve in a dose-dependent manner, and increased the efferent discharge rate of the pancreatic vagal branch. The results suggest the existence of glycine sensors in the hepato-portal system that exert reflex regulation on the pancreatic vagus nerve activity.

In perfused rat liver perivascular nerve stimulation (7.5 Hz, 20 V, 2 ms, 5 min) at the liver hilus caused an increase in glucose release, a shift of lactate uptake to output and a decrease in arterial, portal and total flow. The influence of the alpha 1-receptor blocker prazosin and the beta-antagonist propranolol on these nerve effects were studied in the isolated rat liver perfused via both the hepatic artery and the portal vein in three experimental series at 9 a.m., at 2 p.m. and at 8 p.m. 1) The nerve stimulation-dependent increase in glucose output was maximal at 9 a.m. (100%), halfmaximal at 2 p.m. (50%) and very low at 8 p.m. (15%). The alterations in arterial, portal and total flow were similar in all three series. 2) At 9 a.m., at 2 p.m. and at 8 p.m. 5 microM arterial plus portal prazosin nearly completely inhibited the metabolic alterations and blocked the reduction of arterial, portal and total flow by 80%. 3) At 9 a.m. 10 microM arterial, portal and arterial plus portal propranolol inhibited the increase in glucose release to less than 25% and the shift of lactate uptake to output to about 50%. At 2 p.m. this inhibitory effect was not observed, neither by selective arterial or portal, nor by simultaneous arterial plus portal addition of 10 microM propranolol. At 8 p.m. the influence of propranolol could not be studied because of the too small control values. The nerve stimulation-dependent reduction of arterial, portal and total flow was not influenced by propranolol, neither at 9 a.m. nor at 2 p.m.(ABSTRACT TRUNCATED AT 250 WORDS)

In perfused rat liver stimulation of the hepatic nerve plexuses increased via alpha 1-receptors glucose and lactate output decreased flow and caused an overflow of noradrenaline into the hepatic vein. Infusion of noradrenaline and adrenaline also elicited similar metabolic and hemodynamic alterations via alpha 1-receptors, whereas infusion of isoproterenol via beta 2-receptors enhanced glucose output and slightly reduced lactate release without affecting flow. The influence of circulating catecholamines on the nerve stimulation-dependent changes was investigated. Noradrenaline (100 nmol/L) or adrenaline (40 nmol/L) but not isoproterenol (1 mumol/L), which themselves caused about half-maximal alterations, strongly inhibited the nerve stimulation-induced increase in glucose and lactate output and decrease in flow but had no effect on noradrenaline overflow. The protein kinase C activator (4 beta)phorbol 12-myristate, 13-acetate (100 nmol/L) but not its analog (4 alpha)phorbol 12,13-didecanoate (100 nmol/L) strongly inhibited the metabolic and hemodynamic changes caused by nerve stimulation or noradrenaline infusion. The protein kinase C inhibitor H7 (20 mumol/L) partially prevented the inhibition of the nerve actions by noradrenaline. The results lead us to conclude that noradrenaline and adrenaline inhibited the metabolic and hemodynamic nerve actions by means of a mechanism involving protein kinase C rather than presynaptic alpha-receptors or beta-receptors. The catecholamines apparently increased via alpha 1-receptors inositol 1,4,5-trisphosphate, which in turn enhanced cytosolic Ca2+ and thus altered metabolism and in part hemodynamics, and diacylglycerol, which in turn activated protein kinase C and thus feedback inhibited the signal chain from alpha 1-receptors via G proteins to phospholipase C.

This study tested the hypothesis that the denervated liver is more susceptible to hypovolemic shock than the normal liver. Fourteen swine, seven nondenervated and seven after liver denervation, were studied during hypovolemic shock to 50% of baseline blood pressure. Hepatic artery and portal vein flows were measured using transonic flow probes, and cardiac output and central venous pressure were measured using Swan-Ganz catheters. Hepatic artery flow fell equivalently in the two groups, from 132 +/- 71 to 94 +/- 17 ml/min in the nondenervated group compared with 149 +/- 56 to 91 +/- 55 ml/min in the denervated group. In contrast, portal flow in the denervated group (276 +/- 71 to 119 +/- 53 ml/min) fell significantly (p less than 0.001) more than in the nondenervated group (289 +/- 135 to 194 +/- 70 ml/min). The 58% reduction from baseline in portal flow in the denervated group compared with the 30% reduction in the nondenervated group suggests that the normal compensatory mechanism to maintain portal flow during hypovolemic shock is neurally mediated. It can be hypothesized that sensory afferent fibers might initiate a feedback to splanchnic vasodilatation in response to reduced portal flow. This study supports the hypothesis that the denervated liver is more susceptible to hypovolemic shock.

We reported the existence of vagal arginine sensors in the liver which modulate arginine-induced pancreatic hormone secretion. The present study was carried out to examine the possible existence of other amino acid sensors such as L-alanine and L-leucine in the hepato-portal system in rats using an electrophysiological approach. Afferent discharges were recorded from fine filaments dissected from the peripheral cut end of the hepatic branch of the vagus nerve. Administration of 0.1, 1, and 10 mM L-alanine and L-leucine solution (0.1 ml) into the portal vein caused an increase in the discharge rate of hepatic vagal afferents in a dose-dependent manner. The results suggest the existence of vagal amino acid sensors which are sensitive to alanine and leucine in the hepato-portal system.

In perfused rat liver hepatic nerve stimulation (10 Hz, 2 ms) caused an increase in glucose and lactate output, a decrease in flow and an overflow of noradrenaline into the hepatic vein. Noradrenaline (1 microM) (NA) and prostaglandin F2 alpha (5 microM) (PGF2 alpha), which are implicated as mediators of nerve action, elicited similar effects. 1) All actions of nerve stimulation and the hemodynamic but not the metabolic effects of noradrenaline and PGF2 alpha were largely dependent on extracellular calcium. 2) The dihydropyridine type calcium antagonist nifedipine (5 microM) inhibited the hemodynamic but not the metabolic actions of nerve stimulation, NA and PGF2 alpha, while the phenylalkylamine type calcium antagonist verapamil (5 microM) had no effect. These findings allow the following conclusions: Calcium influx into I nerve endings, necessary for the release of neurotransmitter, II parenchymal cells, for the display of metabolic effects induced by nerve stimulation, and III the actions of NA and PGF2 alpha, do not appear to be mediated by the normal affinity nifedipine- or the verapamil-sensitive channels. Calcium influx into vascular smooth muscle and/or endothelial cells for the display of hemodynamic action induced by nerve stimulation and the NA and PGF2 alpha effects, appear to occur through nifedipine-sensitive but verapamil-insensitive channels.

The effect of mixed denervation on the hepatic microcirculatory bed was studied in dogs and rats. Seven days after denervation, dilated sinusoid capillaries and spaces around the sinuses (Disse's space), disorders in structure of the sinusoid reticular carcass were revealed in the liver of the dogs and rats. The labelled nuclei index in the stellate endotheliocytes is statistically increasing; one month later, in the sinusoid wall the number of cells containing glicosaminoglicanes in the cytoplasm is increasing, outgrowth of the reticular fibres in the perisinusoid spaces and reorganization of the reticular type of the hepatic stroma from the radial into the alveolar one is noted. All the signs mentioned above demonstrate certain adaptation of the microcirculatory bed in the denervated liver to the altered conditions of hemodynamics.

1. The terminal microcirculation in the transilluminated ventral margin of the rat liver was observed and recorded by a video-microscope system. The volumetric flow rate in a liver sinusoid was calculated from the observed diameter of the sinusoid and the intra-sinusoid erythrocyte flow velocity. 2. The topographic distribution of liver sinusoids within an arbitrary boundary of a microscopic field of terminal liver microcirculation was observed and the total inflow and outflow in the field were determined. 3. Both vagus nerves at the lower end of the oesophagus were stimulated at supramaximal voltage. Vagal stimulation dilated the calibre of liver sinusoids and paradoxically diminished the erythrocyte flow velocity in each individual liver sinusoid, but the total volumetric flows in a microscopic field remained unchanged. 4. Vagal stimulation also increased the number of liver sinusoids in a microscopic field by opening previously closed liver sinusoids. This recruitment contributed two-thirds of the total increase of the sinusoidal capacity while the other one third was the result of distension of existing liver sinusoids.

Studies were made on the hemodynamics in the liver and kidney during hemorrhage and retransfusion, using 20 dogs with or without hepatic periarterial nerve plexus, anaesthetized with sodium pentobarbital. The results were as follows. In 14 dogs in which the hepatic periarterial nerve plexus was left intact (Group A), the averages of the hepatic arterial flow (HAF), portal venous pressure (PVP), portal venous flow (PVF), and flow in the right renal artery (RAF) decreased significantly to 38, 64, 35, and 25%, respectively of the control levels when the abdominal aortic pressure (ABP) fell to 40 mm Hg (36% of the control level) during hemorrhage. At the same time the portal venous resistance (PVR) and the right renal vascular resistance (RVR) increased greatly and the hepatic arterial resistance (HAR) either increased or decreased in different dogs. During retransfusion the ABP and vascular resistance returned to nearly the control levels, the PVF increased beyond the control level. In 6 dogs in which the nerve plexus was resected (Group B), the decreases in the HAF and RAF during hemorrhage were not significantly different from those in Group A. These results may suggest the following. The decrease in the HAF during hemorrhage seems to be mostly a passive change resulting from decrease in the systemic blood pressure. Neurogenic regulation may be a little. Further, the existence of the humoral regulation and autoregulation in the HAF were suggested in some cases. While, the responsibilities to hemorrhage and retransfusion appear to be different among the examined three vessels, i.e. the hepatic artery, portal vein, and renal artery. No definite interrelation was seen between the hepatic and renal circulations.

The autonomic nervous system in rats has been assessed by means of indirect immunofluorescence using monospecific antibodies to neuron-specific enolase, neurofilaments, glial fibrillary acidic protein and S-100 protein (10 days after partial (70%) hepatectomy). Different groups of rats were studied: group A: 70% resection and normal dual blood supply (n = 5); group B: 70% resection with only portal blood to the liver remnant (n = 5); group C: 70% resection with only arterial blood to the liver remnant (n = 5); group D: sham operated controls (n = 5). All rats of groups A and D showed normal liver/body weight ratios after 10 days in contrast to groups B and C where liver weights were 50-60% of the preresection weight. In group A the regeneration process was histologically normal and associated with a remarkable increase of autonomic innervation patterns in the portal triad. In contrast, livers of animals in groups B and C showed under the light microscope features of hepatocyte degeneration associated with a decreased autonomic innervation compared to the controls. The changes are identical in groups B and C, and are therefore irrespective of the type of blood deprivation (arterial or portal). These results support the importance of dual blood supply for an optimal regenerative response in liver remnants after liver resection. We suggest that the autonomic nerve supply of the portal triad plays at least an important permissive role in liver regeneration.

Bilateral carotid arterial occlusion was used as a means of reflexly activating hepatic sympathetic nerves in cats anesthetized with pentobarbital. The resultant hepatic arterial constriction was used as an index of effective surgical denervation of the liver in situations where the hepatic arterial blood pressure was held constant or was allowed to rise parallel with the systemic blood pressure. Cutting the anterior nerve plexus (around the common hepatic artery) produced a highly variable but incomplete reduction in the constrictor response to carotid occlusion. The remainder of the constrictor effect was abolished only when the posterior plexus and hepatic ligament were also cut. Thus, vascular resistance in the hepatic artery is controlled by both the anterior and posterior hepatic plexuses but the contribution of each plexus varies considerably from animal to animal. This must be considered in any experiments that use, or have used, denervation or stimulation of the hepatic nerves.

Intravenous administration of hypotensive doses (30-200 micrograms/kg) of nifedipine to cats anesthetized with pentobarbital caused an increase in cardiac output accompanied by hepatic venoconstriction. The hepatic venoconstriction and the increase in cardiac output were abolished in animals in which the hepatic sympathetic nerves were cut, the adrenal glands were excluded, and the kidneys were removed. This contrasts with the indirect hepatic venoconstrictor action of isoproterenol which was shown previously not to be abolished by these procedures. Further experiments showed that the hepatic venoconstrictor effect of nifedipine was blocked by removal of the kidneys, but not by removal of the hepatic sympathetic nerves and adrenals. These results support the hypothesis that venoconstriction plays an important role when drugs produce increased cardiac output. In nephrectomized animals, nifedipine had no direct effects on hepatic blood volume and it did not alter the effects of infusions of norepinephrine on hepatic blood volume, which have previously been shown to be mediated through alpha 2-adrenoceptors. However, it did reduce the hepatic venous responses to hepatic sympathetic nerve stimulation by 30%.

In the perfused rat liver stimulation of the hepatic nerves around the portal vein and the hepatic artery was previously shown to increase glucose output, to shift lactate uptake to output, to decrease and re-distribute intrahepatic perfusion flow and to cause an overflow of noradrenaline into the hepatic vein. The metabolic effects could be caused directly via nerve hepatocyte contacts or indirectly by the hemodynamic changes and/or by noradrenaline overflow from the afferent vasculature into the sinusoids. Evidence against the indirect modes of nerve action is presented. Reduction of perfusion flow by lowering the perfusion pressure from 2 to 1 ml X min-1 X g-1--as after nerve stimulation--or to 0.35 ml X min-1 X g-1--far beyond the nerve stimulation-dependent effect--did not change glucose output and lowered lactate uptake only slightly. Only re-increase of flow to 2 ml X min-1 X g-1 enhanced glucose and lactate release transiently due to washout of glucose and lactate accumulated in parenchymal areas not perfused during low perfusion flow. In chemically sympathectomized livers nerve stimulation decreased perfusion flow almost normally but without changing the intrahepatic microcirculation; yet it enhanced glucose and lactate output only insignificantly and caused noradrenaline overflow of less than 10% of normal. Conversely, in the presence of nitroprussiate (III) nerve stimulation reduced overall flow only slightly without intrahepatic redistribution but still increased glucose and lactate output strongly and caused normal noradrenaline overflow.(ABSTRACT TRUNCATED AT 250 WORDS)

In isolated rat liver perfused at constant flow with erythrocyte-free Krebs-Henseleit bicarbonate buffer containing 5 mM glucose and 2 mM lactate, perivascular stimulation of the hepatic nerves caused a rapid decrease of oxygen uptake, a decrease of the periportal and, after a transient rise, of the perivenous tissue po2 of surface acini, an increase of portal pressure, and an enhancement of glucose output. Furthermore, nerve stimulation changed the intrahepatic distribution of the perfusate drastically. Infusion of trypan blue 20 s after nerve stimulation resulted in a heterogeneous staining of the liver both at the surface and in cross-sections, while it led to a homogeneous distribution in non-stimulated controls. It is concluded that the major component in the mechanism of the nerve-dependent decrease of oxygen uptake is the microcirculatory change rather than a metabolic effect.

Sympathetic electric stimulation of the liver nerves increased hydrostatic pressure in sinusoids by 8-10% in cats. The sinusoidal filtration coefficient decreased, particularly at the stimulation frequency 10-20 Hz. No significant increase in the lymph outflow from the liver was revealed which suggested an absence of the fluid transfer from the sinusoids to extravascular space. The above changes occurred against the background of the constriction of arterial, portal and venous vessels, as well as a reduction of the liver capacity. Comparison of these shifts with analogous values in small intestine revealed that the microcirculation in excitation of sympathetic-adrenal system is more stable in liver than in small intestine.

Arterial pressures, portal pressures, and hepatic blood volumes were recorded after hepatic denervation in cats anesthetized with pentobarbital. Bromocryptine (50 micrograms/kg) lowered arterial pressure but did not significantly change portal pressure or hepatic blood volume. However, both portal pressure and hepatic blood volume responses to hepatic nerve stimulation were significantly depressed after bromocryptine especially at low frequencies of stimulation. Responses to intraportal infusions of norepinephrine were significantly impaired only at the highest dose. The inhibitory effect of bromocryptine on the neural responses may, therefore, involve a presynaptic inhibition of norepinephrine release, but the mechanism requires further study. These data provide further support for the hypothesis that drugs which impair hepatic venous responses to sympathetic stimuli cause significant impairment of postural reflexes and orthostatic hypotension during clinical use.

The mutual changes in hepatic and small intestinal blood flow on selective nervous stimulation of the periarterial vasoconstriction fibres, were studied in anaesthetized cats. Occlusion of the hepatic artery did not change portal blood flow, whereas occlusion of superior mesenteric blood flow caused a significant increase in hepatic arterial flow. Stimulation of the hepatic sympathetic nervous supply caused a phasic blood flow response with a marked transient peak flow reduction of hepatic arterial blood flow. The magnitude of the peak response varied with the frequency of the stimulation. Despite continuous stimulation the first phase went over into a second phase of less pronounced vasoconstriction. This "steady state" blood flow, was maintained at about 20% below the control level, irrespective of stimulation frequency. Corresponding in time with the peak vasoconstriction there was a transient increase of portal pressure. Sympathetic nerve stimulation increased portal pressure even on occlusion of the hepatic artery. Stimulation of the mesenteric sympathetic nerves evoked the characteristic transient peak vasoconstrictor response consisting of two phases, a brief intense peak resistance response followed by a second phase of less pronounced but generally well maintained constriction. Simultaneously a slight reduction of portal pressure and hepatic arterial vascular resistance was regularly seen. In contrast to the observations in the hepatic arterial circuit the magnitude of this "steady state" blood flow in the small intestine was dependent on the rate of the stimulation, however. On simultaneous stimulation of the hepatic and mesenteric sympathetic nerves the hemodynamic responses were largely the same as when these nerves were stimulated separately. The portal pressure affecting mean capillary pressure in the intestine differed, however. Small and variable pressure changes were followed by rapid return towards control and during steady state it did not differ from the prestimulatory level.

Portal venous pressure was controlled by resistance localized to specific sites in hepatic lobar veins in cats. All of the pressure drop from the portal vein to the vena cava occurred across postsinusoidal vessels; portal pressure, lobar venous pressure, and, therefore, sinusoidal pressure were not significantly different. Norepinephrine and angiotensin infusions (intraportal) caused elevation in portal pressure due to constriction of hepatic venous resistance sites as well as some constriction of presinusoidal (portal or sinusoidal) resistance sites. At low doses of norepinephrine presinusoidal constriction dominated whereas at higher doses the postsinusoidal constriction increased proportionately more. Hepatic nerve stimulation produced a similar response measured at an early time (1 min), but by 3 min the presinusoidal constriction showed complete escape so that elevated portal pressure was entirely due to hepatic venous constriction. The same site that provided basal vascular resistance also provided the increased hepatic venous resistance with nerve stimulation and infusion of angiotensin and norepinephrine. Rapid elevation of central venous pressure (CVP) caused elevated sinusoidal pressure. At high CVP (16 mm Hg), 75% of a rise in CVP was transmitted whereas at normal CVP (less than 4.5 mm Hg) less than 20% transmission occurred. The presence of a high resistance in the hepatic veins protected intrahepatic pressure from the effects of normal fluctuation of CVP.

This study measured the liver blood flow-dependent index of galactose clearance in patients after liver transplantation, to test the hypothesis that liver blood flow is increased in the denervated liver. Eight normal subjects and 16 patients 1 to 8 months after liver transplant were studied. All patients were stable with no evidence of severe rejection at the time of study. Galactose clearance was measured at steady state during continuous infusion of 75 mg per min of 5% galactose. The results show a statistically significant (p less than 0.01) higher average galactose clearance in the transplant patients (1,187 +/- 316 ml per min per m2) compared to the control group (709 +/- 121 ml per min per m2). The major limiting factor in galactose clearance at low concentrations is liver blood flow, and we interpret these data as evidence for increased blood flow in the transplanted liver. Possible mechanisms for the increased galactose clearance are (i) loss of normal vasomotor tone in the denervated liver, or (ii) persistence of abnormal systemic hemodynamics after transplantation. Elucidation of these mechanisms awaits further study.

Intrahepatic blood volume-pressure relationships were studied using plethysmography to measure hepatic blood volume and a hepatic venous long-circuit to control intrahepatic pressure. In cats anesthetized with pentobarbital or with ketamine-chloralose, hemorrhage (to reduce hepatic blood flow to 60% of control) caused marked reductions in hepatic blood volume and intrahepatic pressure but did not significantly change hepatic blood volume-pressure relationships. We were unable to demonstrate an active reflex venous response to hemorrhage in these preparations, although a large passive response occurred. The volume-pressure relationships in innervated livers were different from those in denervated livers: apparent venous compliance was much greater and apparent unstressed volume was zero or negative. Hepatic nerve stimulation in denervated livers caused a marked decrease in hepatic blood volume at low intrahepatic pressures but failed to alter hepatic blood volumes at high intrahepatic pressures (15 mmHg) (1 mmHg = 133.3 Pa). This resulted in large apparent compliances and apparently negative unstressed volumes, as seen in the innervated livers. Thus blood volume-pressure relationships in innervated livers may not give valid measurements of compliance and unstressed volume. A remarkable feature in all these experiments was the linearity of the relationship between hepatic blood volume and intrahepatic pressure. Exudation of fluid begins at higher intrahepatic pressures in innervated compared with denervated livers.

The relative effects of changes in hepatic blood flow and in levels of sympathetic activity on liver volumes were ascertained by use of multiple-indicator dilution studies. Portal vein-hepatic vein dilution patterns were obtained following injection of a mixture containing 51Cr-labeled red blood cells (a vascular reference) and 14C-labeled sucrose (a reference extracellular substance). Hepatic blood volume was calculated as the product of liver blood flow and red cell transit times, and interstitial space as the product of plasma flow and the difference between labeled sucrose and red cell transit times. Concentrations of norepinephrine and epinephrine were measured in aorta, portal vein, and hepatic vein. Hepatic blood volume and interstitial space increased with increased blood flow; and elevated plasma norepinephrine levels, especially in portal vein (following either exogenous infusion or activation of peripheral sympathetic fibers) selectively reduced the distending effect of hepatic inflow on the vascular volume. The data provide a description of the in vivo variation in canine liver vascular and interstitial volumes with blood flow and sympathetic activation.

A mean decrease of 60% in liver blood volume was recorded by a plethysmographic technique during electrical stimulation of the hepatic nerves in anesthetized, vagotomized dogs. A decrease in pressure in the vascularly isolated carotid sinus to 40 mm Hg, from a mean control of 144 mm Hg, decreased liver blood volume by a mean of 16%; arterial blood pressure increased by a mean of 77 mm Hg. Carotid sinus hypotension was accompanied by respective mean increases of 16% and 1.4% in hepatic arterial and portal venous blood flows, and of 45% and 22% in arterial and portal resistances. Increase in sinus pressure to 240 mm Hg increased liver blood volume by a mean of 20%; arterial blood pressure decreased by 90 mm Hg. Sinus hypertension was accompanied by respective mean decreased of 10% and 1.5% in hepatic arterial and portal venous blood flows, and of 44% and 18% in arterial and portal resistances. Interruption of afferent vagal traffic from cardiopulmonary receptors was maximally effective in decreasing liver blood volume at a carotid sinus pressure of 40 mm Hg and was ineffective at carotid sinus pressures greater than 160 mm Hg. Combined withdrawal of carotid and cardiopulmonary vasomotor inhibition decreased liver blood volume by 42%; of this 37% was due to the cardiopulmonary and 63% to the carotid baroreflex. The study showed the canine liver to function as a blood reservoir by active mobilization of a portion of its blood volume.