Influence of protracted infusion of glucose and insulin on the composition and regeneration activity of liver after partial hepatectomy in rats

Radix Astragali decoction improves liver regeneration by upregulating hepatic expression of aquaporin-9

BACKGROUND

The therapeutic efficacy of liver injuries heavily relies on the liver's remarkable regenerative capacity, necessitating the maintenance of glycose/lipids homeostasis and oxidative eustasis during the recovery process. Astragali Radix, an herbal tonic widely used in China and many other countries, is believed to have many positive effects, including immune stimulation, nourishing, antioxidant, liver protection, diuresis, anti-diabetes, anti-cancer and expectorant. Astragali Radix is widely integrated into hepatoprotective formulas as it is believed to facilitate liver regeneration. Nevertheless, the precise molecular pharmacological mechanisms underlying this hepatoprotective effect remain elusive.

PURPOSE

To investigate the improving effects of Astragali Radix on liver regeneration and the underlying mechanisms.

METHODS

A mouse model of 70% partial hepatectomy (PHx) was employed to investigate the impact of Radix Astragali decoction (HQD) on liver regeneration. HQD was orally administered for 7 days before the PHx procedure and throughout the experiment. N-acetylcysteine (NAC) was used as a positive control for liver regeneration. Liver regeneration was assessed by evaluating the liver-to-body weight ratio (LW/BW) and the expression of representative cell proliferation marker proteins. Oxidative stress and glucose metabolism were analyzed using biochemical assays, Western blotting, dihydroethidium (DHE) fluorescence, and periodic acid-Schiff (PAS) staining methods. To understand the role of AQP9 as a potential molecular target of HQD in promoting liver regeneration, td-Tomato-tagged AQP9 transgenic mice (AQP9-RFP) were employed to determine the expression pattern of AQP9 protein. AQP9 knockout mice (AQP9-/-) were used to assess the specific targeting of AQP9 in the promotion of liver regeneration by HQD.

RESULTS

HQD significantly upregulated hepatic AQP9 expression, alleviated liver injury and promoted liver regeneration in wild-type (AQP9+/+) mice after 70% PHx. However, the beneficial impact of HQD on liver regeneration was absent in AQP9 gene knockout (AQP9-/-) mice. Moreover, HQD facilitated the uptake of glycerol by hepatocytes, enhanced gluconeogenesis, and concurrently reduced H2O2 content and oxidative stress levels in AQP9+/+ but not AQP9-/- mouse livers. Additionally, main active substance of Radix Astragali, astragaloside IV (AS-IV) and cycloastragenol (CAG), demonstrated substantial upregulation of AQP9 expression and promoted liver regeneration in AQP9+/+ but not AQP9-/- mice.

CONCLUSION

This study is the first to demonstrate that Radix Astragali and its main active constituents (AS-IV and CAG) improve liver regeneration by upregulating the expression of AQP9 in hepatocytes to increase gluconeogenesis and reduce oxidative stress. The study revealed novel molecular pharmacological mechanisms of Radix Astragali and provided a promising therapeutic target of liver diseases.

Aquaporin-9 facilitates liver regeneration following hepatectomy

Aquaporin-9 (AQP9) is an aquaglyceroporin strongly expressed in the basolateral membrane of hepatocytes facing the sinusoids. AQP9 is permeable to hydrogen peroxide (H₂O₂) and glycerol as well as to water. Here, we report impaired liver regeneration in AQP9^−/− mice which involves altered steady-state H₂O₂ concentration and glucose metabolism in hepatocytes. AQP9^−/− mice showed remarkably delayed liver regeneration and increased mortality following 70% or 90% partial hepatectomy. Compared to AQP9^+/+ littermates, AQP9^−/− mice showed significantly greater hepatic H₂O₂ concentration and more severe liver injury. Fluorescence measurements indicated impaired H₂O₂ transport across plasma membrane of primary cultured hepatocytes from AQP9^−/− mice, supporting the hypothesis that AQP9 deficiency results in H₂O₂ accumulation and oxidative injury in regenerating liver because of reduced export of intracellular H₂O₂ from hepatocytes. The H₂O₂ overload in AQP9^−/− hepatocytes reduced PI3K-Akt and insulin signaling, inhibited autophagy and promoted apoptosis, resulting in impaired proliferation and increased cell death. In addition, hepatocytes from AQP9^−/− mice had low liver glycerol and high blood glycerol levels, suggesting decreased glycerol uptake and gluconeogenesis in AQP9^−/− hepatocytes. Adeno-associated virus (AAV)-mediated expression of hepatic expression of aquaglyceroporins AQP9 and AQP3 in AQP9^−/− mice, but not water-selective channel AQP4, fully rescued the impaired liver regeneration phenotype as well as the oxidative injury and abnormal glucose metabolism. Our data revealed a pivotal role of AQP9 in liver regeneration by regulating hepatocyte H₂O₂ homeostasis and glucose metabolism, suggesting AQP9 as a novel target to enhance liver regeneration following injury, surgical resection or transplantation.

1.

AQP9 mediates H₂O₂ and glycerol transport across hepatocytes plasma membrane

2.

AQP9^−/− mice exhibit retained liver regeneration and higher mortality after PH

3.

Elevated H₂O₂ and reduced glucose levels appear in AQP9^−/− regenerating liver

4.

Replacement of aquaglyceroporin rescued impaired AQP9^−/− mouse liver regeneration

5.

AQP9 may become a novel target to improve liver regeneration

To divide or not to divide: revisiting liver regeneration

The liver has a remarkable capacity to regenerate. Even with surgical removal (partial hepatectomy) of 70% of liver mass, the remnant tissue grows to recover the original mass and functions. Liver regeneration after partial hepatectomy has been studied extensively since the 19th century, establishing the long-standing model that hepatocytes, which account for most of the liver weight, proliferate to recover the original mass of the liver. The basis of this model is the fact that almost all hepatocytes undergo S phase, as shown by the incorporation of radioactive nucleotides during liver regeneration. However, DNA replication does not necessarily indicate the execution of cell division, and a possible change in hepatocyte size is not considered in the model. In addition, as 15-30% of hepatocytes in adult liver are binuclear, the difference in nuclear number may affect the mode of cell division during regeneration. Thus, the traditional model seems to be oversimplified. Recently, we developed new techniques to investigate the process of liver regeneration, and revealed interesting features of hepatocytes. In this review, we first provide a historical overview of how the widely accepted model of liver regeneration was established and then discuss some overlooked observations together with our recent findings. Finally, we describe the revised model and perspectives on liver regeneration research.

Functional Relationships between Lipid Metabolism and Liver Regeneration

The regenerative capacity of the liver is well known, and the mechanisms that regulate this process have been extensively studied using experimental model systems including surgical resection and hepatotoxin exposure. The response to primary mitogens has also been used to investigate the regulation of hepatocellular proliferation. Such analyses have identified many specific cytokines and growth factors, intracellular signaling events, and transcription factors that are regulated during and necessary for normal liver regeneration. Nevertheless, the nature and identities of the most proximal events that initiate hepatic regeneration as well as those distal signals that terminate this process remain unknown. Here, we review the data implicating acute alterations in lipid metabolism as important determinants of experimental liver regeneration and propose a novel metabolic model of regeneration based on these data. We also discuss the association between chronic hepatic steatosis and impaired regeneration in animal models and humans and consider important areas for future research.

p21 is required for dextrose-mediated inhibition of mouse liver regeneration

The inhibitory effect of dextrose supplementation on liver regeneration was first described more than 4 decades ago. Nevertheless, the molecular mechanisms responsible for this observation have not been elucidated. We investigated these mechanisms using the partial hepatectomy model in mice given standard or 10% dextrose (D10)-supplemented drinking water. The results showed that D10-treated mice exhibited significantly reduced hepatic regeneration compared with controls, as assessed by hepatocellular bromodeoxyuridine (BrdU) incorporation and mitotic frequency. D10 supplementation did not suppress activation of hepatocyte growth factor (HGF), induction of transforming growth factor alpha (TGF-alpha) expression, or tumor necrosis factor alpha-interleukin-6 cytokine signaling, p42/44 extracellular signal-regulated kinase (ERK) activation, immediate early gene expression, or expression of CCAAT/enhancer binding protein beta (C/EBPbeta), but did augment expression of the mito-inhibitory factors C/EBPalpha, p21(Waf1/Cip1), and p27(Kip1). In addition, forkhead box M1 (FoxM1) expression, which is required for normal liver regeneration, was suppressed by D10 treatment. Finally, D10 did not suppress either FoxM1 expression or hepatocellular proliferation in p21 null mice subjected to partial hepatectomy, establishing the functional significance of these events in mediating the effects of D10 on liver regeneration.

CONCLUSION

These data show that the inhibitory effect of dextrose supplementation on liver regeneration is associated with increased expression of C/EBPalpha, p21, and p27, and decreased expression of FoxM1, and that D10-mediated inhibition of liver regeneration is abrogated in p21-deficient animals. Our observations are consistent with a model in which hepatic sufficiency is defined by homeostasis between the energy-generating capacity of the liver and the energy demands of the body mass, with liver regeneration initiated when the functional liver mass is no longer sufficient to meet such demand.

Proliferation of hepatocytes

Hepatocyte proliferation may be controlled by reversible patterns of endocrine changes, monitored by the liver, involving known hormones and their receptors. A two-programme model of related interactions among nutrients, specific lipoproteins, and highly phosphorylated nucleotides is postulated. This hypothesis stems from in vitro studies of rat hepatocyte proliferation under chemically defined conditions and from in vivo studies using partially hepatectomized, hormone-infused, developing and lipotrope-deficient rats. Certain findings are discussed with regard to receptor systems which show negatively cooperative properties; to problems of proliferative specificity; and to novel approaches for defined studies of chemical hepatocarcinogenesis.

Hepatotrophic effects of pancreatic and gastrointestinal hormones in the rat in vivo and in vitro

The role of gastrointestinal and pancreatic hormones in regulating liver growth was evaluated by measuring their effect on DNA synthesis in the normal and regenerating liver of rats in vivo and in maintenance cultures of adult rat hepatocytes in vitro. After partial liver resection DNA synthesis reached peak levels after 24 hours while serum concentrations of immunoreactive insulin in portal and peripheral blood at this time were still suppressed. Increase of endogenous insulin levels by intravenous glucose infusion or portal infusion of insulin, glucagon or both together with glucose did not change DNA synthesis in normal or regenerating rat liver. After acute carbon tetrachloride poisoning of rats, survival rate and degree of liver necrosis was not changed by intraperitoneal infusion of glucagon and insulin with glucose. In vitro, insulin, glucagon and somatostatin synergistically stimulated the specific thymidine uptake in seven-day-old maintenance cultures of rat hepatocytes. The hormones did not cause cell multiplication but enhanced cell survival, probably by improving the uptake and utilization of nutrients. Gastrin G-17, secretin and cholecystokinin (contaminated with gastric inhibitory polypeptide) had no effect. It is concluded that the results do not support the contention that liver regeneration is regulated by the known pancreatic hormones. However, a trophic effect of pancreatic hormones on liver cells in vitro could be demonstrated. Gastrointestinal hormones had no such effect.

Mechanisms and outcomes of drug- and toxicant-induced liver toxicity in diabetes

Increase dincidences of hepatotoxicity have been observed in diabetic patients receiving drug therapies. Neither the mechanisms nor the predisposing factors underlying hepatotoxicity in diabetics are clearly understood. Animal studies designed to examine the mechanisms of diabetes-modulated hepatotoxicity have traditionally focused only on bioactivation/detoxification of drugs and toxicants. It is becoming clear that once injury is initiated, additional events determine the final outcome of liver injury. Foremost among them are two leading mechanisms: first, biochemical mechanisms that lead to progression or regression of injury; and second, whether or not timely and adequate liver tissue repair occurs to mitigate injury and restore liver function. The liver has a remarkable ability to repair and restore its structure and function after physical or chemical-induced damage. The dynamic interaction between biotransformation-based liver injury and compensatory tissue repair plays a pivotal role in determining the ultimate outcome of hepatotoxicity initiated by drugs or toxicants. In this review, mechanisms underlying altered hepatotoxicity in diabetes with emphasis on both altered bioactivation and liver tissue repair are discussed. Animal models of both marked sensitivity (diabetic rats) and equally marked protection (diabetic mice) from drug-induced hepatotoxicity are described. These examples represent a remarkable species difference. Availability of the rodent diabetic models offers a unique opportunity to uncover mechanisms of clinical interest in averting human diabetic sensitivity to drug-induced hepatotoxicities. While the rat diabetic models appear to be suitable, the diabetic mouse models might not be suitable in preclinical testing for potential hepatotoxic effects of drugs or toxicants, because regardless of type 1 or type2 diabetes, mice are resistant to acute drug-or toxicant-induced toxicities.

Caveolin-1 is essential for liver regeneration

Liver regeneration is an orchestrated cellular response that coordinates cell activation, lipid metabolism, and cell division. We found that caveolin-1 gene-disrupted mice (cav1-/- mice) exhibited impaired liver regeneration and low survival after a partial hepatectomy. Hepatocytes showed dramatically reduced lipid droplet accumulation and did not advance through the cell division cycle. Treatment of cav1-/- mice with glucose (which is a predominant energy substrate when compared to lipids) drastically increased survival and reestablished progression of the cell cycle. Thus, caveolin-1 plays a crucial role in the mechanisms that coordinate lipid metabolism with the proliferative response occurring in the liver after cellular injury.

Kinetics and mechanisms of hepatic acute phase response to subtotal partial hepatectomy and cultural impact on environmental hepatic end-stage liver injury in the homeless

Intoxication and liver damage induced by carbon tetrachloride (CCl(4)), aflatoxin B1, diabetes, and subtotal partial hepatectomy (PH(90)) in rats in which approximately 90% of the total hepatic tissue mass is surgically removed produces an acute-phase response (APR) whose initial stage prior to regression closely mimics the APRs associated with the life-threatening hepatic failure seen in the homeless. Rats treated by PH(90)were either healthy, CCl(4)-intoxicated, diabetic, or alflatoxin B1 (AFB1) intoxicated to the point of 75% liver insufficiency. It is well documented that high rates of mortality following PH(90)in aseptic rats could be minimized by supplementing drinking water with 20% glucose, organic components of L-15 medium and housing animals in cages maintained at 33-35;C. Aseptic rats showed a mild 20-30% decrease in APR proteins during the first 4-5 days following PH(90), while a maximal APR was noted 9-12 days post PH(90)and lasted for ~30 days when it returned to values close to those of healthy controls. This delay in hepatic APR of the remnant caudate lobe favoured replacement of lost basophilic clumps and ribosomes. The newly synthesized ribosomes of the nascent hepatocytes quantitatively maintained the APR signals of the injured caudate hepatocytes, and biosynthesized and released a typical spectrum of APR proteins. We suggest that massively injured liver has decoded an already stored and irreversible DNA-biochemical sequence of events in which priority is given to recovery of lost tissues by delaying an APR response to injury. In PH(90)of diabetic and CCl(4)-intoxicated rats, the hepatic dual functions of regeneration and APR processes associated with intoxication-initiated catabolic signals, created a heavy metabolic burden on the remnant caudate lobe leading to higher rates of mortality. APR of healthy rats to AFB1 parallels that of alpha-amanitin-induced intoxication. Similarly, within shorter time scale proportional to the severity of surgery, livers undergoing 75% partially hepatectomy (PH(75)) delayed both the onset and regression of APR. We are therefore led to believe that approaches other than liver transplantation should be considered as viable alternatives in the treatment of various acute and chronic liver diseases to avoid rejection and retransplantation. Scarcity of cadaveric liver has forced the medical community to investigate xenotransplantation with its unknown risks. Concomitantly, it is suggested that in view of the incalculable risks of indifference, the homeless must receive much improved medical care as we have found that two-dimensional immunoelectrophoretic assay of their serum is indicative of acute and chronic liver injury. The scientific and moral interrelationships of related matters are illuminated.

Nutritional modulation of liver regeneration by carbohydrates, lipids, and amino acids: a review

The survival of patients after a life-threatening hepatic injury of varying etiology depends on the ability of the remaining hepatocytes to regenerate. Thus, the stimulation of hepatic regeneration can have tremendous therapeutic relevance. Experimental studies--performed mostly on a model of regenerating rat liver after partial hepatectomy--indicate that glucose administration inhibits, whereas infusion of a lipid emulsion can enhance, the rate of liver regeneration. However, the inhibitory effect of glucose on liver regeneration is not observed when glucose is administered together with other nutrients. The results further indicate that administration of a standard amino acid mixture without energy substrate has an inhibitory effect and that development of liver regeneration can be favorably influenced by branched-chain amino acids (valine, leucine, and isoleucine) and glutamine.

Treatment of cirrhotic rats with epidermal growth factor and insulin accelerates liver DNA synthesis after partial hepatectomy

Prevention of postoperative hepatic failure is important after hepatic resection. In patients with cirrhosis, impaired liver function and regenerative capacity after major hepatic resection are associated with increased morbidity and mortality. In this study, a combination of epidermal growth factor (EGF) and insulin were used as hepatotrophic factors in an attempt to stimulate DNA synthesis after 70% hepatectomy (HTX). Regenerative capacity was evaluated in normal and cirrhotic rat liver by measuring DNA synthesis in vivo. Micronodular liver cirrhosis was established by the simultaneous oral administration of CCl4 and phenobarbital. Epidermal growth factor plus insulin was injected subcutaneously immediately after and 12 h after HTX or sham operation was performed. Rats were killed 24 h after the operation and liver regeneration was estimated by [3H]-thymidine incorporation into DNA as well as an autoradiographic nuclear labelling index. Hepatectomy increased [3H]-thymidine incorporation significantly in both normal and cirrhotic rats. In cirrhotic rats, [3H]-thymidine incorporation after HTX was significantly lower than in normal rats and administration of a combination of EGF and insulin after HTX enhanced [3H]-thymidine incorporation. In conclusion, DNA synthesis 24 h after HTX is decreased in cirrhotic rats compared with normal rats and EGF supplementation with insulin accelerates DNA synthesis in hepatectomized cirrhotic rats. The data suggest that administration of combinations of exogenous hepatotrophic factors may play a useful role in the treatment of cirrhotic patients undergoing major hepatic resection.

Review: nutritional support for patients with cirrhosis

Nutritional support is indicated when cirrhotic patients undergo surgery because they are malnourished, hypercatabolic and immunocompromised. However, the choice of nutrient may be problematic as the liver itself is the central organ of protein, fat and glucose metabolism. Branched chain amino acid-enriched solution may be the choice of protein source, as it is anticatabolic and it stimulates liver regeneration. Excessive glucose is undesirable as it may suppress endogenous fat utilization, which may be the preferred pathway of metabolism after hepatectomy. Medium chain triglycerides are preferred to long chain triglycerides as they are readily utilized and are not deposited in the liver; however, the tendency of cirrhotic patients to accumulate free fatty acids and glycerol after infusion of triglycerides dictates their use intermittently. Clinical studies have shown that perioperative nutritional support is beneficial in cirrhotic patients undergoing major hepatectomy or liver transplantation. The judicious choice of nutrient, care of the catheter and a limitation of the fluid infused are all prerequisites for the efficient use of perioperative nutritional support, which is complementary to a technically perfect operation.

Transcriptional and posttranscriptional regulation of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase during liver regeneration

The control of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFK-2/FBPase-2; EC 2.7.1.105/3.1.3.46) gene expression during liver regeneration was studied. The level of PFK-2/FBPase-2 mRNA decreased to about 5% of the control value 6 hr after partial hepatectomy. Thereafter the mRNA increased to a maximum at 48 hr and returned to normal levels by 96 hr. In sham-operated animals, only a small increase was observed during the first 4 hr. The mRNA was recognized by a 299-base-pair liver-specific cDNA probe but not by a muscle-specific probe. The time course of mRNA modulation was well correlated with PFK-2/FBPase-2 activity and with the amount of bifunctional enzyme protein determined by immunoblotting with an antibody raised against the N-terminal decapeptide of liver PFK-2/FBPase-2. No alteration in the degradation rate of PFK-2/FBPase-2 mRNA was noted after partial hepatectomy. The modulation of PFK-2/FBPase-2 gene expression during liver regeneration involved changes in the transcription rate. The rate decreased by 50% at 6 hr after liver resection. The rate increased thereafter to a maximum at 72 hr and then returned to control values by 96 hr. The transcription rate of albumin did not change, whereas that of phosphoenolpyruvate carboxykinase increased 12-fold at 6 hr. These results show that PFK-2/FBPase-2 gene transcription is specifically regulated and that this regulation is in part responsible for the alterations in hepatic metabolism seen in regenerating liver.

Glucose overload and hepatic energy metabolism after resection of the cirrhotic liver in rats

The effect of glucose hyperalimentation on energy metabolism in the cirrhotic rat liver after 70% hepatectomy was studied. After resection, rats received either 30 kcal/kg per day (group I) or 200 kcal/kg per day (group II) of glucose for 48 h. In both groups, hepatic mitochondrial ATP synthesis was accelerated when palmitic acid was used as substrate and suppressed when pyruvate was used. This suggests that the energy substrate of the remnant liver was principally fatty acids rather than glucose. Hepatic energy charge was within normal limits in group I, but decreased significantly in group II after hepatectomy. An abundance of glucose in the early postoperative period, therefore, caused a hepatic energy derangement by suppressing fatty acids utilization; this suppression was corroborated by the findings of lower immunoreactive glucagon and non-esterified fatty-acid concentrations in group II. To determine optimal glucose administration, the predicted value of glucose disposal rate (GDR) was calculated after an intravenous glucose tolerance test. GDR decreased significantly after hepatectomy and did not increase appreciably even with a large dose of insulin administration. These results suggest that glucose administration should be tailored to the GDR values after resection of the cirrhotic liver.

Fructose 2,6-bisphosphate and 6-phosphofructo-2-kinase during liver regeneration

Glycogen and fructose 2,6-bisphosphate levels in rat liver decreased quickly after partial hepatectomy. After 7 days the glycogen level was normalized and fructose 2,6-bisphosphate concentration still remained low. The 'active' (non-phosphorylated) form of 6-phosphofructo-2-kinase varied in parallel with fructose 2,6-bisphosphate levels, whereas the 'total' activity of the enzyme decreased only after 24 h, similarly to glucokinase. The response of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase from hepatectomized rats (96 h) to sn-glycerol 3-phosphate and to cyclic AMP-dependent protein kinase was different from that of the enzyme from control animals and similar to that of the foetal isoenzyme.

Hepatic regeneration and metabolism after partial hepatectomy in normal rats: effects of insulin therapy

The effect of insulin therapy on liver regeneration has been studied in normal fed rats 12, 24 and 48 h after partial hepatectomy. Dry weight of regenerating liver increased between 12 and 48 h after partial hepatectomy and was unaffected by insulin therapy. [6-3H] Thymidine uptake peaked at 24-h (24.7 +/- 2.4% of total liver cells) and insulin treatment had no additional effect. At 12-h after partial hepatectomy, hepatic [ATP] was decreased 15%, while [ADP] and [AMP] were increased 47% and 83% respectively compared with sham-operated animals. Partial hepatectomy also caused an increase in hepatic [triglyceride], a decrease in hepatic [glycogen] and an increase in the levels of glucose and several glycolytic intermediates. The hepatic redox ratios, [lactate]:[pyruvate] and [3-hydroxybutyrate]:[acetoacetate], were elevated. Insulin therapy had only minor effects on hepatic adenine nucleotide levels, intermediary metabolite concentrations or intrahepatic redox ratios after partial hepatectomy. These findings suggest a decreased hepatic intracellular energy state in regenerating liver; insulin therapy in normal rats does not influence this metabolic change nor the regenerative response.

Liver regeneration in normal and alloxan-induced diabetic rats

The effects of alloxan-induced diabetes on liver regeneration were investigated. Normal and diabetic rats were sacrificed at eight time periods between 16 hours and 4 weeks following two-thirds partial hepatectomy or sham operation. The results indicate that alloxan-induced diabetes delays but does not prevent liver regeneration following partial hepatectomy. This delay is indicated by a depressed synthesis of RNA, DNA and protein during the first post-operative day and a lack of mitotic figures in the 24-hour sample. In addition, the synthesis of these three cellular constituents did not return to control levels as rapidly in the diabetics. Compared with the sham operated animals, the concentrations of total serum protein remained depressed longer in the diabetic hepatectomized animals. The data indicate that the metabolic alterations associated with alloxan diabetes delay the onset of the regenerative process and prolong the recovery period.

Metabolic changes following varying degrees of partial hepatectomy in the rat

The metabolic and functional changes that follow hepatectomy in the rat are related to the extent of liver resection. Regenerating liver cells show an impaired ability to excrete bilirubin, bromsulphophthalein, and 125I Rose Bengal during the early postoperative period.