Cholestasis In Mice Research Articles

The role of the hepatocyte cytokeratin network in bile formation and resistance to bile acid challenge and cholestasis in mice #

The intermediate filament cytoskeleton of hepatocytes is composed of keratin (K) 8 and K18 and has important mechanical and nonmechanical functions. However, the potential role of the K8/K18 network for proper membrane targeting of hepatocellular adenosine triphosphate-binding cassette transporters and bile formation is unknown. We therefore designed a comparative study in K8 and K18 knockout mice and respective wild-type controls to test the hypothesis that intermediate filaments of hepatocytes play a role in normal bile formation. In addition, we challenged mice either with a 1% cholic acid-supplemented diet or a diet containing the porphyrinogenic xenobiotic 3,5-diethoxycarbonyl-1,4-dihydrocollidine to determine the effect of K8/K18 loss on bile flow/composition and liver injury under different physiological and toxic stress stimuli. Protein expression levels and membrane localization of various transporters and anion exchangers were compared using western blotting and immunofluorescence microscopy, respectively, and bile flow and composition were determined under various experimental conditions. Our results demonstrate that loss of the intermediate filament network had no significant effect on bile formation and composition, as well as expression levels and membrane targeting of key hepatobiliary transporters under baseline and stress conditions. However, loss of K8 significantly increased liver injury in response to toxic stress. The intermediate filament network of hepatocytes is not specifically required for proper bile formation in mice.

In vivo dynamic metabolic imaging of obstructive cholestasis in mice

We tried to image obstructive cholestasis by using a newly developed imaging system to measure the alterations of hepatobiliary function in living mice with their bile ducts ligated. A hepatic imaging window was installed on the upper abdomen soon after the mice underwent ligation of the common bile duct. On the next day, the mice received intravenous injection of rhodamine B isothiocyanate-dextran and carboxyfluorescein diacetate. The later would be transformed into fluorogenic carboxyfluorescein (detected at approximately 500-550 nm) by hepatocytes and then excreted into bile canaliculi. The images were acquired by multiphoton microscopy. The fluorescence intensities at approximately 500-550 nm within hepatocytes or sinusoids were measured in time series. In mice with bile duct ligation, bile canaliculi failed to appear during the whole observation period over 100 min following carboxyfluorescein diacetate injection, whereas the fluorescence was retained much longer within sinusoids. Furthermore, the fluorescence intensities in sinusoids were persistently higher than in hepatocytes during the course. Bile duct ligation impedes hepatocytes to excrete carboxyfluorescein into bile canaliculi. The kinetics of fluorescence intensities in hepatocytes and sinusoids indicated there is an active machinery operating backflow of this fluorogenic bile solute from hepatocytes into sinusoids in the liver with obstructive cholestasis.

Dysregulation of glutathione synthesis during cholestasis in mice: molecular mechanisms and therapeutic implications.

Glutathione (GSH) provides important antioxidant defense and regulates multiple critical processes including fibrogenesis. There are conflicting literature studies regarding changes in GSH during cholestasis. Here we examined changes in the GSH synthetic enzymes during bile duct ligation (BDL) in mice and how treatment with ursodeoxycholic acid (UDCA) and/or S-adenosylmethionine (SAMe) affects the expression of these enzymes and liver injury. The hepatic expression of glutamate-cysteine ligase (GCL) subunits and GSH synthase (GS) increased transiently after BDL but fell to 50% of baseline by 2 weeks. Nuclear factor-erythroid 2-related factor 2 (Nrf2) trans-activates gene expression by way of the antioxidant response element (ARE), which controls the expression of all three genes. Despite increased Nrf2 nuclear levels, Nrf2 nuclear binding to ARE fell 2 weeks after BDL. Nuclear levels of c-Maf and MafG, which can negatively regulate ARE, were persistently induced during BDL and the dominant proteins bound to ARE on day 14. UDCA and SAMe induced the expression of GCL subunits and raised GSH levels. They increased nuclear Nrf2 levels, prevented c-Maf and MafG induction, and prevented the fall in Nrf2 nuclear binding to ARE. Combined treatment had additive effects, reduced liver cell death, and prevented fibrosis. GSH synthesis falls during later stages of BDL due to lower expression of GSH synthetic enzymes. UDCA and SAMe treatment prevented this fall and combined therapy was more effective on preserving GSH levels and preventing liver injury.

Bid-mediated apoptosis does not contribute to cholestatic liver injury following bile duct ligation

Death receptor-mediated hepatocyte apoptosis has been implicated in bile acid induced apoptosis of hepatocytes in vitro and in vivo. The BH3-interacting domain death agonist Bid is a critical mediator for apoptosis induced by activation of Fas and TNF-R1 death receptors in hepatocytes. Multiple studies have shown that bile acids induce Fas-dependent hepatocyte apoptosis in vitro providing a cellular mechanism for bile acid induced liver injury in vivo. More recent studies suggested that the mechanism of hepatocyte death is almost exclusively oncotic and that the protective effect in lpr mice most strikingly correlates with a reduced inflammatory response. In addition to these controversies, it is still unclear which down-stream targets of death receptors mediate the effect in rodent livers following BDL The aim of this study was therefore to elucidate the role of Bid in cholestatic liver injury. Overall survival and various aspects of liver injury were analyzed in WT and Bid-/- mice following BDL, a commonly used model to study obstructive cholestasis in mice. Liver injury was examined 3, 7 and 14 days after BDL. Loss of Bid did not affect the number of bile infarcts, serum AST values and animal survival. Processing of procaspase-3 and -9 and caspase-3 enzyme activities were not detectable in either group. Importantly, Bid-/- mice displayed the same pattern of TUNEL positive hepatocytes as WT controls following BDL and Bid-/- hepatocytes displayed the same sensitivity against DCA-induced apoptosis in vitro. In contrast to Fas-receptor deficient lpr mice, hepatic fibrosis and the inflammatory response was not affected by loss of Bid. Together, our data clearly show that Bid-mediated signaling does not play a role for BDL induced liver injury. Our findings strongly argue against selecting Bid as a therapeutic target in cholestasis and should therefore have implication for the development of new treatment strategies during bile acid induced liver injury.

In Vivo Effect of Selective Macrophage Suppression on the Development of Intrahepatic Cholestasis in Mice

We studied the role of selective suppression of liver Kupffer cells (gadolinium chloride, 14 mg/kg intravenously) in the development of intrahepatic cholestasis in CBA/C57B1/6 mice after intraperitoneal injection of alpha-naphthylisothiocyanate in a single dose of 200 mg/kg. Pretreatment with gadolinium chloride increased the severity of cholestasis and signs of liver damage. Gadolinium accumulation in the liver peaked after 24 h and was accompanied by a decrease in activities of cathepsin D and cathepsin B and concentration of matrix metalloprotease-2. Our results confirm the hypothesis that normal function of Kupffer cells and extracellular matrix plays an important role in cholestasis. Administration of gadolinium chloride serves as a convenient model to study the side effects, toxicity, and safety of lanthanides as nanoparticles.

P28. HNF-6 Mediates Growth Hormone Effect on Hepatocyte Function During Cholestasis in Mice

Abstract not available

35. HNF-6 Mediates Growth Hormone Effect on Hepatocyte Function During Cholestasis in Mice

35. HNF-6 Mediates Growth Hormone Effect on Hepatocyte Function During Cholestasis in Mice

Seizure susceptibility alteration following reversible cholestasis in mice: Modulation by opioids and nitric oxide

There is an increasing body of evidence that the central nervous system is affected by cholestatic liver disorders. Cholestasis has been shown to result in a decreased seizure propensity which is believed to be mediated by an increased opioidergic tone and nitric oxide (NO) signaling pathway. In this study, we used a reversible chemically-induced cholestasis model in mice to investigate the changes in seizure susceptibility. The cholestasis was induced by intragastric administration of alpha-naphthylisothiocyanate (ANIT) (100 mg/kg) or vehicle (corn oil). The threshold to generalized clonic seizures induced by timed intravenous infusion of pentylenetetrazole (PTZ) was used as an index of seizure propensity. The role of opioid receptors and NO pathway in the changes of seizure threshold, and the responsiveness to the anticonvulsant effect of opioid agonist, morphine, during and after the resolution of cholestasis was studied in this reversible paradigm of cholestatic disease. A significant increase in cholestasis-related biochemical markers as well as in clonic seizure threshold was observed; it was maximal at day 3 after cholestasis induction and slowly decreased to normal thereafter. Seizure threshold rise was inhibited by chronic administration of the opioid antagonist naltrexone or acute administration of N-nitro-L-arginine methyl ester (L-NAME), an inhibitor of NO production. Co-administration of subeffective doses of L-NAME and naltrexone showed an additive effect. Injection of an anticonvulsant dose of morphine on day 7 after cholestasis induction did not increase seizure threshold, suggestive of a downregulation of receptors even after cholestasis resolution. These data shows that ANIT-induced cholestasis leads to a reversible increased resistance to PTZ-induced seizures through an opioid/NO-mediated pathway, and is probably accompanied by downregulation of opioid receptors.

Effects of Phalloidin on Hepatic Gene Expression in Mice

An attempt has been made to identify molecular markers of intrahepatic cholestasis in mice employing phalloidin as a cholestatic agent. Phalloidin was administered to BALB/c mice at three predetermined dose: 250 microg/kg, 500 microg/kg, and 1 mg/kg for 1, 3, and 7 days. Liver function was estimated to confirm cholestasis. Histopathological observations on liver were also made to confirm liver injury. Phalloidin at 1 mg/kg for 7 days was found to induce cholestasis. Therefore gene expression studies were confined to this group only. A total of 88 genes were found to be affected by phalloidin. These were the genes associated with cytoskeleton regulation as well as tight junction, focal adhesion, and ATP-binding cassette transporters. Such proteins obstruct the removal of bile components from hepatocytes to the bile canaliculus or blood. Phalloidin treatment did not affect the proteins responsible for cell maintenance or death. The authors show that phalloidin-induced intrahepatic cholestasis is manifested by disturbing the cytoskeleton. The set of genes up-regulated by phalloidin can be considered as molecular markers of intrahepatic cholestasis. The observations are further expected to be helpful in the management of cholestatic pharmaceuticals and associated problems of liver diseases in humans.

Role of plasma and liver cholesterol- and lipoprotein-metabolism determinants in LpX formation in the mouse

Cholestasis is characterized by hypercholesterolemia and the appearance of an abnormal lipoprotein, lipoprotein X (LpX), in plasma. The mechanisms responsible for this cholestatic plasma lipid phenotype are not fully understood. We used ATP-binding cassette A1 (ABCA1)(−/−) and scavenger receptor class B type I (SR-BI)(−/−) mice to test the hypothesis that hepatic sinusoidal cholesterol transporters contribute to LpX formation and hypercholesterolemia during cholestasis. Bile-duct ligation (BDL) of both ABCA1(−/−) and SR-BI(−/−) mice, as well as their respective controls, induced a dramatic increase in plasma cholesterol and phospholipid concentrations. Plasma fractionation revealed the presence of LpX in plasma of cholestatic mice, irrespective of their genetic background. We observed that the presence of HDL before cholestasis, a decrease in the activity of LCAT, and an increase in VLDL synthesis were not required for hypercholesterolemia and lipoprotein modifications induced by obstructive cholestasis in mice. In addition, murine cholestasis resulted in increased hepatic cholesterol synthesis that may contribute to the higher plasma free cholesterol levels found during the early hours after BDL. Together these findings indicate that hypercholesterolemia and LpX formation associated with obstructive cholestasis are correlated with an increase in hepatic cholesterol synthesis and are independent of plasma HDL levels, LCAT activity, VLDL synthesis, and ABCA1 and SR-BI expression.

Maintained expression of alternative bile salt excretion pathways in inflammatory cholestasis in mice by treatment with class II nuclear hormon receptor ligands

Expression of hepatic organic anion transporters including basolateral overflow systems such as Mrp3 and Mrp4 is decreased during inflammatory cholestasis by experimental treatment with endotoxin (LPS). This crucial event putatively occurs by a cytokine-induced nucleocytoplasmic shift of RXR, the obligate heterodimerization partner of transactivating class II nuclear hormon receptors (NHR), as a central mechanism. We hypothesized that treatment with specific ligands of RXR and CAR may induce a redistribution of these receptors and subsequently a maintained transporter gene expression.

LPS signaling enhances hepatic fibrogenesis caused by experimental cholestasis in mice

Although it is clear that bile acid accumulation is the major initiator of fibrosis caused by cholestatic liver disease, endotoxemia is a common side effect. However, the depletion of hepatic macrophages with gadolinium chloride blunts hepatic fibrosis. Because endotoxin is a key activator of hepatic macrophages, this study was designed to test the hypothesis that LPS signaling through CD14 contributes to hepatic fibrosis caused by experimental cholestasis. Wild-type mice and CD14 knockout mice (CD14(-/-)) underwent sham operation or bile duct ligation and were killed 3 wk later. Measures of liver injury, such as focal necrosis, biliary cell proliferation, and inflammatory cell influx, were not significantly different among the strains 3 wk after bile duct ligation. Markers of liver fibrosis such as Sirius red staining, liver hydroxyproline, and alpha-smooth muscle actin expression were blunted in CD14(-/-) mice compared with wild-type mice after bile duct ligation. Despite no difference in lymphocyte infiltration, the macrophage/monocyte activation marker OX42 (CD11b) and the oxidative stress/lipid peroxidation marker 4-hydroxynonenal were significantly upregulated in wild-type mice after bile duct ligation but not in CD14(-/-) mice. Increased profibrogenic cytokine mRNA expression in the liver after bile duct ligation was significantly blunted in CD14(-/-) mice compared with the wild type. The hypothesis that LPS was involved in experimental cholestatic liver fibrosis was tested using mice deficient in LPS-binding protein (LBP(-/-)). LBP(-/-) mice had less liver injury and fibrosis (Siruis red staining and hydroxyproline content) compared with wild-type mice after bile duct ligation. In conclusion, these data demonstrate that endotoxin in a CD14-dependent manner exacerbates hepatic fibrogenesis and macrophage activation to produce oxidants and cytokines after bile duct ligation.

Endogenous opioid-mediated antinociception in cholestatic mice is peripherally, not centrally, mediated

Cholestasis is associated with naloxone reversible antinociception and opiate receptor antagonists are used clinically to treat pruritus. Pain and pruritus are closely interrelated and opioids modulate both sensations. Therefore, we undertook a series of experiments to characterize opioid-mediated antinociception in cholestasis and determine if it occurs inside or outside the CNS. Antinociception scores to both thermal and mechanical stimuli were determined in mice with cholestasis due to bile duct resection vs sham controls. Cholestatic mice demonstrated significant antinociception to both stimuli compared to controls, which was reversible by the opiate receptor antagonist naloxone. The experiments were repeated with a naloxone derivative, which does not cross the blood-brain-barrier (i.e. naloxone methiodide) with similar results, indicating an opioid antinociceptive effect mediated outside of the CNS. Experiments with intraplantar injections of low dose naloxone methiodide confirmed that cholestasis-associated antinociception occurs at the level of cutaneous nerve endings. These findings were supported by findings of increased dermal met-enkephalin expression in cholestatic mice. Cholestasis in mice is associated with antinociception due to local effects of endogenous opioids (i.e. met-enkephalin) at the level of sensory nerve endings. These findings may have direct implications in the management of cholestasis associated pruritus.

174 POSTER Pramanicin induces apoptosis in Jurkat leukemia cells: a role for JNK, p38 and caspase activation

Rifampicin is a well-known hepatotoxicant, but little is known about the mechanism of rifampicin-induced hepatotoxicity. The aim of this study was to characterize the expression and localization of hepatocyte tight junctions in rifampicin-induced cholestasis in mice. Cholestasis was induced by administration of rifampicin (200 mg/kg) for 7 consecutive days or treatment with a single dose of rifampicin (200 mg/kg) by gastric intubation. The expression of mRNA for hepatic zonula occludens (ZO)-1, ZO-2, ZO-3, occludin and claudin-1 was determined using RT-PCR. Localization of ZO-1 and occludin was detected using immunofluorescence. Results showed that there was an 82-fold increase in the conjugated bilirubin in serum in rifampicin-treated mice. In addition, an 8-fold increase in total bile acid in serum was observed after a seven-day administration of rifampicin. The integrity of hepatocyte ZO-1 and occludin was altered by a seven-day administration of rifampicin. Importantly, the integrity and intensity of hepatocyte tight junctions were altered as early as 30 min after a single dose of rifampicin. The expression of hepatic ZO-1 and ZO-2 mRNA was significantly decreased, beginning as early as 30 min and remaining a lower level 12 h after a single dose of rifampicin. Taken together, these results suggest that the altered integrity and internalization of hepatocyte tight junctions are associated with rifampicin-induced cholestasis.

175 POSTER MDGA1, a novel human protein with a functional role related to cellular adhesion

Rifampicin is a well-known hepatotoxicant, but little is known about the mechanism of rifampicin-induced hepatotoxicity. The aim of this study was to characterize the expression and localization of hepatocyte tight junctions in rifampicin-induced cholestasis in mice. Cholestasis was induced by administration of rifampicin (200 mg/kg) for 7 consecutive days or treatment with a single dose of rifampicin (200 mg/kg) by gastric intubation. The expression of mRNA for hepatic zonula occludens (ZO)-1, ZO-2, ZO-3, occludin and claudin-1 was determined using RT-PCR. Localization of ZO-1 and occludin was detected using immunofluorescence. Results showed that there was an 82-fold increase in the conjugated bilirubin in serum in rifampicin-treated mice. In addition, an 8-fold increase in total bile acid in serum was observed after a seven-day administration of rifampicin. The integrity of hepatocyte ZO-1 and occludin was altered by a seven-day administration of rifampicin. Importantly, the integrity and intensity of hepatocyte tight junctions were altered as early as 30 min after a single dose of rifampicin. The expression of hepatic ZO-1 and ZO-2 mRNA was significantly decreased, beginning as early as 30 min and remaining a lower level 12 h after a single dose of rifampicin. Taken together, these results suggest that the altered integrity and internalization of hepatocyte tight junctions are associated with rifampicin-induced cholestasis.

Lack of gp130 Expression Results in More Bacterial Infection and Higher Mortality During Chronic Cholestasis in Mice *

Chronic cholestasis is associated with increased bacterial infections and sepsis resulting in higher mortality in humans. In the current study, we investigated the relevance of gp130-dependent pathways after bile duct ligation (BDL). BDL was performed in conditional gp130 knockout (loxP/Cre system) mice and respective controls. Liver injury, regulation of the acute phase response, and the impact on survival and bacterial infections were determined. Acute BDL resulted in increased IL-6 levels, Stat3 activation, and an increase in acute-phase proteins (serum-amyloid-A [SAA]), which was blocked in gp130-deleted animals. In addition, the antimicrobial gene hepcidin was regulated in a gp130-dependent manner after BDL. During chronic cholestasis Stat3 activation was dramatically reduced, while high SAA levels were maintained via gp130-dependent signaling. Inhibition of gp130-dependent pathways resulted in higher mortality and liver damage, which was associated with higher infiltration of immune-activated cells and increased germ number in the liver. In conclusion, during acute and chronic cholestasis, the gp130 system is essential for controlling the acute-phase response. Lack of gp130 expression results in pronounced bacterial growth in bile and liver after BDL, which is associated with higher mortality. Activation of gp130-dependent pathways after BDL is essential and appears to be a therapeutic target during cholestasis.

Oncosis represents the main type of cell death in mouse models of cholestasis

Since the mechanisms leading to hepatocyte death in cholestasis are not well defined, we aimed to obtain closer insights into the related pathogenetic principles. Cell death was assessed in common bile duct ligated (CBDL) and cholic acid (CA)-fed mice, and compared to Fas agonist Jo2-injected mice by studying H and E-stained tissue sections, DNA ladder analysis, caspase-3-like activity assay, immunohistochemistry, double immunofluorescence microscopy for activated caspase-3 and cytokeratin (CK) 18, the TUNEL method, and electron microscopy. Jo2-treated mice showed activation of caspase-3, breakdown of the CK intermediate filament network, and classical morphological features of apoptosis. In contrast, in CA-fed and CBDL mice, oncosis characterized by cell swelling and ruptured cell membranes was the predominant type of cell death, whereas in both experimental conditions significant activation of caspase-3 was absent and typical CK alterations were rare despite frequent positivity of the TUNEL assay. (i) Oncosis represents the main type of hepatocyte death in acute cholestasis in mice. (ii) The importance of apoptosis in cholestasis may be overestimated if non-specific detection systems (e.g. TUNEL assay) are used.

The role of mdr2 in manganese–bilirubin induced cholestasis in mice

The pathogenesis of manganese–bilirubin (Mn–BR) induced cholestasis has only been studied in rats and is associated with alteration in the hepatic homeostasis of cholesterol and phospholipids. Multidrug resistance-2 (mdr2) transporter, which mediates excretion of these lipids, is suggested to be involved in this phenomenon. The present study was undertaken to examine if Mn–BR induced cholestasis is reproducible in mice, then to clarify the role of mdr2 in its pathogenesis, using mice with disrupted mdr2 gene (mdr2 (−/−)). Results showed that Mn–BR combination decreased bile flow in mice. This reduction in bile flow was similar in mdr2 (−/−) and the wild type mdr2 (+/+). Furthermore, the change in biliary lipid excretion was comparable in both genotypes. These data indicate that Mn–BR induced cholestasis is reproducible in mice and provide evidence that mdr2 alteration is not a primary event in this form of cholestasis.

Effect of mdr2 mutation with combined tandem disruption of canalicular glycoprotein transporters by cyclosporine A on bile formation in mice

The inhibition of canalicular glycoprotein transporters has been suggested as the cause of familial intrahepatic cholestasis. Mutations in multidrug resistance 3-glycoprotein (MDR3) gene induce progressive familial intrahepatic cholestasis type 3 (PFIC3). Phenotypically, mutation in mdr2 in mice resembles the disruption of MDR3 in human. Secondly, mutation in the bile salt exporting pump (BSEP)/sister of P-glycoprotein (spgp) gene causes progressive familial intrahepatic cholestasis type 2 in human. However, in spgp knock-out mice only a mild persistent cholestasis occurs. The aim of this study is to evaluate the effects of various P-glycoprotein (Pgp) transporters on bile formation and the canalicular transport of taurocholic acid (TCA) in an attempt to understand the combined role of these transporters in the pathogenesis of familial intrahepatic cholestasis. Total bile acid (TBA) and cholic acid secretion rate were decreased in the mdr2 knock-out mice. However, bile flow (BF) and the secretion of muricholic acids were increased. Secretion of cholesterol was negligible and no phospholipids were detected in bile of mdr2 knock-out mice. Treatment with cyclosporine A (CsA) decreased the BF, and the biliary secretion of bile salts (BS) and phospholipids as compared to wild type mice, but after the injection of TCA+CsA, the BF, and the biliary secretion of BS and lipids were increased as compared to the wild type mice treated with CsA alone. In the mdr2 knock-out mice, CsA treatment decreased the BF and the secretion of BS but after the injection of TCA+CsA, the BF and the biliary secretion of BS were increased and the phospholipid secretion was slightly stimulated as compared to the mdr2 knock-out mice treated with CsA alone. Conclusion: Disruption of the mdr2 gene and the inhibition of glycoprotein transporters by CsA induce cholestasis in mice which is characterized by reduced BF, BS and biliary lipid secretion. However, CsA treatment did not significantly increase the cholestatic effect in the mdr2 knock-out mice. The injection of TCA decreased the cholestatic effect in the mdr2 knock-out mice as well as the inhibition of glycoproteins transporters by CsA. These data suggest that mutation in the canalicular mdr2 is an important factor during the development of progressive familial cholestasis.

Inhibition by lithium of opioid withdrawal-like syndrome and physical dependency in a model of acute cholestasis in mice

Because of the claim that chronic cholestatic liver disease is associated with increased opioidergic tone and that lithium (Li+) inhibits physical dependency in morphine-dependent mice, the effects of chronic Li+ treatment on naloxone-precipitated withdrawal-like syndrome and antinociception were evaluated in an animal model of cholestasis. For this purpose, acute cholestasis was induced by bile duct ligation in mice. The treated group of mice was given lithium chloride (300 mg/l) as drinking fluid for 10–12 days before surgery and 5 days after surgery. Physical dependency was observed by precipitating an abstinence syndrome with naloxone (4 mg/kg, s.c.) 5 days after induction of cholestasis. Antinociception was assessed by tail-flick latency test in control and treated groups before administration of naloxone. Results of the present study revealed that chronic Li+ administration significantly reduced the withdrawal-like signs, whereas the antinociception assessed by tail-flick test was not observed in the control group of cholestatic mice. Serum Li+ level in this study was much lower than the commonly accepted therapeutic range. Lithium and opioid agonist ligands have diverse effects on transmembrane signal control systems. So the interaction of Li+ and opioid agonists observed in this study might also be through these systems. © 1998 John Wiley & Sons, Ltd.