Hepatic enzymic zonation: a reevaluation of the concept of the liver acinus

Three-dimensional reconstruction of parenchymal units in the liver of the rat

To investigate the parenchymal units in the liver of the rat three-dimensionally, 15 micrometer cryosections were used for the demonstration of glucose-6-phosphatase (G6Pase) activity to visualize the borders of the individual units. Together with the supplying and draining vessels, they were traced through a sequence of 146 sections and reconstructed. A cone-shaped secondary unit with a height of 2.1 mm and a volume of 3.3 mm3 was reconstructed. It was "covered" by a continuous vascular surface, consisting of portal tracts and vascular septa, connecting the portal venular branches. The secondary unit was subdivided by portal tracts and vascular septa, and by branches of a draining central venular tree into 14 primary units. Most of them were tri- to heptahedral in shape. The height varied between 330 and 840 micrometer, and the volume varied between 0.094 and 0.621 mm3. The branches of the portal venular tree, with diameters from 28 +/- 5 micrometer to 61 +/- 14 micrometer, were oriented preferentially along the vertical axis of the units. Most of the primary units were drained by single branches of the central venular tree, located in the center and oriented along the vertical axis of the units. Vessel diameters ranged from 62 +/- 14 micrometer to 216 +/- 9 micrometer. The average length of the sinusoids was 355 +/- 3 micrometer. From the results of this reconstruction study, it was concluded that the concept of the liver acinus cannot be applied to the liver of the rat.

Regulation of the spatiotemporal pattern of expression of the glutamine synthetase gene

Glutamine synthetase, the enzyme that catalyzes the ATP-dependent conversion of glutamate and ammonia into glutamine, is expressed in a tissue-specific and developmentally controlled manner. The first part of this review focuses on its spatiotemporal pattern of expression, the factors that regulate its levels under (patho)physiological conditions, and its role in glutamine, glutamate, and ammonia metabolism in mammals. Glutamine synthetase protein stability is more than 10-fold reduced by its product glutamine and by covalent modifications. During late fetal development, translational efficiency increases more than 10-fold. Glutamine synthetase mRNA stability is negatively affected by cAMP, whereas glucocorticoids, growth hormone, insulin (all positive), and cAMP (negative) regulate its rate of transcription. The signal transduction pathways by which these factors may regulate the expression of glutamine synthetase are briefly discussed. The second part of the review focuses on the evolution, structure, and transcriptional regulation of the glutamine synthetase gene in rat and chicken. Two enhancers (at -6.5 and -2.5 kb) were identified in the upstream region and two enhancers (between +156 and +857 bp) in the first intron of the rat glutamine synthetase gene. In addition, sequence analysis suggests a regulatory role for regions in the 3' untranslated region of the gene. The immediate-upstream region of the chicken glutamine synthetase gene is responsible for its cell-specific expression, whereas the glucocorticoid-induced developmental appearance in the neural retina is governed by its far-upstream region.

Effects of diosgenin, a plant-derived steroid, on bile secretion and hepatocellular cholestasis induced by estrogens in the rat

Increased biliary secretion of cholesterol and lipid vesicles (unilamellae and multilamellae) induced by diosgenin (D), a plant-derived steroid, has cytoprotective effects in the rat liver subjected to obstructive cholestasis. In this study, our aims were to investigate the following: 1) the effects of D on the bile secretory process and on the cholestasis induced by estradiol-17beta-(beta-D-glucuronide) (E17G) or 17 alpha-ethynylestradiol (E) administration; 2) whether the potentially protective effects of D are related to D-induced increase of biliary cholesterol and lipid lamellae; and 3) whether D has other effects capable of modifying specific bile secretory processes or preventing the cholestatic effects of estrogens. Rats were fed a standard ground chow (control group) or chow containing D for 6 days. E17G was administered i.v. to control and D-fed rats and bile flow, bile salt output, and alkaline phosphatase excretion were examined. 17alpha-E was administered from days 4 to 6 to rats fed standard chow or chow plus D for 6 days and different functional parameters of the bile secretory process as well as the ultrastructure of hepatocytes and histochemistry of alkaline phosphatase and Mg2+-adenosine triphosphatase (ATPase) were examined. D-treatment markedly increased cholesterol and lamellar structures in bile and attenuated the acute cholestatic effects of E17G. D-feeding prevented the decrease of taurocholate maximum secretory rate and the increase of biliary alkaline phosphatase and Ca2+,Mg2+-EctoATPase (EctoATPase) excretion, as well as the increase of cholesterol/ phospholipids ratio, alkaline phosphatase activity, and EctoATPase content in canalicular plasma membranes induced by E. D-feeding did not prevent E-induced decrease of basal bile flow, bile salt, cholesterol, and phospholipid secretory rates nor the decrease of Na+,K+-ATPase activity and Na+-taurocholate cotransporting polypeptide (Ntcp) content in isolated sinusoidal membranes. Cholestatic alterations of canalicular domain were apparent in E-treated rats. D administration was also associated with changes of ultrastructure and histochemistry of hepatocytes. E-induced alterations in ultrastructure and acinar distribution and intensity of histochemical reaction of both enzymes were partially prevented by D-feeding. We conclude that D administration, in addition to inducing a marked increase of biliary cholesterol and lipid lamellar structures output, was associated to changes in hepatocyte morphology and plasma membrane composition, enzymes activity, and histochemistry. D-feeding attenuated the acute cholestatic effects of E17G. D-induced increase of bile cholesterol and lipid lamellae content was not apparent when D-fed rats received E. Despite this fact, D administration prevented some cholestatic effects of E, probably through different metabolic effects and/or direct membrane effects, not related to increased lipid lamellae excretion.

Hepatic glutamine transport and metabolism

Although the liver was long known to play a major role in the uptake, synthesis, and disposition of glutamine, metabolite balance studies across the whole liver yielded apparently contradictory findings suggesting that little or no net turnover of glutamine occurred in this organ. Efforts to understand the unique regulatory properties of hepatic glutaminase culminated in the conceptual reformulation of the pathway for glutamine synthesis and turnover, especially as regards the role of sub-acinar distribution of glutamine synthetase and glutaminase. This chapter describes these processes as well as the role of glutamine in hepatocellular hydration, a process that is the consequence of cumulative, osmotically active uptake of glutamine into cells. This topic is also examined in terms of the effects of cell swelling on the selective stimulation or inhibition of other far-ranging cellular processes. The pathophysiology of the intercellular glutamine cycle in cirrhosis is also considered.

Zonal and regional differences identified from precision mapping of vitamin A-storing lipid droplets of the hepatic stellate cells in pig liver: a novel concept of addressing the intralobular area of heterogeneity

Knowledge of hepatic heterogeneity has been strikingly increased, while an accurate means for addressing intralobular positions is still lacking. We examined pig liver preparations of the gold impregnation method for vitamin A-storing lipid droplets in hepatic stellate cells. Droplet morphometry was performed under oil immersion, and the calculated volumes plotted on computerized maps. The heterogeneous results were assessed with five concentric zones and five radial regions; the latter were determined based on midseptum visualized by portal injection. Zonation and regionation thus subdivided lobules into 5-zone/5-region (5Z/5R) compartmentalization. Distribution of values exhibited a distinct zonal gradient, heightened at peripheral zones 1 and 2, decreased over intermediate zone 3 toward centrilobular zones 4 and 5; peak was always found at zone 2. Within a single zone, variations were obvious, forming a regional gradient. Values were significantly higher at periportal than midseptal regions. Digitized mapping showed that low values filled up centrilobular zones, whereas high values concentrated in periportal regions. Along the periphery, inlet venules were quantified, revealing an occurrence rate of 60% at periportal, and 5% at midseptal regions, closely compatible with the regional gradient of vitamin A-storing capacity. The interweaving between zonal and regional gradients results in a vitamin A-low territory, a compound area composed of centrilobular zones plus extensions into midseptal regions. Because the results could account for physiological and pathological events, we regard the 5Z/5R compartmentalization a model worth routine adoption for a precise description of any morphofunctionally demonstrable heterogeneity of the liver lobules.

Cytochemical and biochemical studies on adenylate cyclase activity in preneoplastic and neoplastic liver tissue and cultured liver cells

The ATP-analogue adenylyl(beta,gamma-methylene)diphosphonate was chosen as substrate for the cytochemical localization of adenylate cyclase (AC) activity. The tissues investigated covered normal rat liver and liver from carcinogen-treated animals with preneoplastic lesions and hepatocellular neoplasms, as well as cultured liver cells. The AC reaction product methylene diphosphonate was precipitated with Pb2+ immediately at the place of production. This approach permitted a precise localization of AC activity by light and electron microscopy. The specificity of the AC reaction was demonstrated by control reactions, including inhibition of AC with 2'5'-dideoxyadenosine and activation with forskolin, glucagon, and cholera toxin. Endogenous phosphatases were inhibited with tetramisole and NAD. In normal liver, AC activity was mainly localized in the sinusoidal membrane of hepatocytes. A distinct gradient in activity was observed within the liver lobule. Hepatocytes localized around the terminal hepatic venule showed a significant higher AC activity compared to hepatocytes near the portal tract. AC was clearly decreased in focal preneoplastic liver lesions of the glycogenotic-basophilic cell lineage leading to hepatocellular carcinomas. Cytochemically detected intensity of AC activity corresponded to data obtained by microbiochemical assays in laser-dissected tissue samples. A remarkable interdependence of AC activity and degree of differentiation was also seen in epithelial rat liver cell lines: Highly differentiated cells show high enzyme activity and vice versa, as shown by both cytochemical and biochemical examinations. It is concluded that alterations in cellular signal transduction caused by alterations in AC activity play an important role in hepatocarcinogenesis.

A computer-aided 3-D geometry of acute and chronic zonal necrosis: three-D tangent counting applied in an attempt to re-examine the structure of the human liver

In an attempt to examine the feasibility of the acinar concept in the liver of humans, the spatial distribution of zonal necrosis and its relation with the blood vessels were studied in terms of 3-D tissue microstructure. The material was five autopsy livers, two from patients of acute cresol intoxication and three with long-standing heart failure. Examined were: 1) whether the surviving areas (zones 1, 2) in acute zonal necrosis have surfaces convex toward the necrotic zone 3 as illustrated in the acinar schema, and 2) whether the necrosis extends as slender "s eeves" so as to reach the portal tracts and create portal-central (P-C) bridging, leaving zones 1 and 2 as a spheroid parenchymal clump. These were examined on serial histological sections of the livers, which were subjected to 3-D reconstruction with the aid of a computer system. The geometric properties of the interfaces between the surviving and necrotic areas were analyzed by 3-D tangent counting technique of DeHoff modified for 3-D application. It was shown by reconstruction that necrosis was uniformly distributed around the hepatic venules. In acute cases, tangent counting showed that the surfaces of the surviving areas were mainly concave toward necrosis. In contrast, in chronic cases, the surfaces became mainly convex as a result of parenchymal regeneration. At no place, in acute or chronic cases, was necrosis shown to reach the portal tracts, and because of this, neither acini nor their agglomerates revealed themselves as a separate unit. All these findings contradict what one expects would be, so long as based on the assumption of the acinar model.

The portal lobule in rat liver fibrosis: a re-evaluation of the liver unit

We re-evaluated three schemes of liver organization: the classic lobule, the portal lobule, and Rappaport's liver acinus. The lobular angioarchitecture of normal rat liver and the three-dimensional structure of pseudolubules found in rat livers with fibrosis induced by swine serum were compared with the classic lobule of the pig. Normal and fibrotic rat livers and pig livers were perfused, injected with either India ink or 0.75% OsO4 through the portal and/or hepatic vein, and immersionfixed. Whole lobes and hand-cut thick sections were made transparent with a solution of benzyl benzoate and methyl salicylate. The angioarchitecture of normal rat liver differs from pig liver. In the former, terminal portal branches and central veins interdigitate, and in the latter, numerous terminal portal branches that arise from interlobular portal veins establish a vascular basket surrounding one central vein and forming classic lobule. The structure of liver acinus is never found in the pig liver. The terminal portal branch, together with the terminal hepatic artery and bile duct, are present inside each pseudolobule of fibrotic rat livers. Blood from the terminal portal branch flows through inlet venules into radiating sinusoids, and, at the periphery converges into newly formed septal and angular outlet venules; these venules terminate in fibrotic central veins located at each corner. Pseudolobules are not rugby ball-like as Rappaport's liver acini are but are polyhedron in shape. The rat pseudolobules are comparable with the portal lobule; its structure and microcirculation are the reverse of the pig classic lobule. Rat pseudolobules are different from liver acini, as shown by the following: 1) their three-dimensional shape is different; and 2) they have a reverse relationship to classic lobules while acini are defined to subdivide classic lobules. In normal and fibrotic rat livers, the liver unit is the portal lobule with a terminal portal branch as the axial branch and central veins at the periphery. The co-existence of liver acini and classic lobules is doubtful.

In situ measurement of glutamate concentrations in the periportal, intermediate, and pericentral zones of rat liver

We developed a quantitative histochemical assay for measurement of local glutamate concentrations in cryostat sections of rat liver. Deamination of glutamate by glutamate dehydrogenase (GDH) was coupled to the production of formazan and formazan precipitation was used for colorimetric visualization. The method was tested and validated with gelatin model sections with known glutamate concentrations. Calibration graphs showed linear relationships with high correlation coefficients (> 96%) between glutamate concentrations or section thickness and absorbance values. The method was reproducible, with a constant percentage of 60 +/- 5% of glutamate being converted in gelatin model sections containing glutamate concentrations of 2 mM and higher. Glutamate concentrations were estimated in periportal, intermediate, and pericentral zones of liver lobules that contain low, intermediate, and high GDH activity, respectively. In fed adult male rat livers, periportal zones contained the highest concentrations of glutamate (approximately 14 mM) and intermediate and pericentral zones approximately 13 and 9 mM, respectively. On starvation, glutamate concentrations increased only in the small rim of pericentral cells that express glutamine synthetase, to approximately 15 mM. In livers of fetal and newborn rats, glutamate was homogeneously distributed, with a concentration of approximately 5 mM. In suckling rat liver, distribution of glutamate was still homogeneous but the concentration was increased to approximately 8 mM. These glutamate distribution patterns were in agreement with those detected immunohistochemically.

A multicompartment geometric model of the liver in relation to regional induction of cytochrome P450s

A geometric, multicompartment model of the liver was developed to examine regional protein induction and to provide model output suitable for predicting the degree of induction in both the whole liver and in specific regions. The model was based on functional hexagonal arrays within the liver. A geometric representation was used to divide these functional units into five zones: a concentric periportal zone, a fenestrated periportal region that interconnects among multiple functional units, and three concentric centrilobular areas, referred to, respectively, as zones 1 through 5. The surface areas (and volumes for hexagonal cylinders) of these live zones were, respectively, 13.5, 25.2, 33.9, 20.3, and 6.8% of the total liver. The pharmacokinetic model for induction had dissociation constants (Kd) and Hill constants (n) for interactions of transcriptional activator-ligand complexes with response elements on DNA. Estimates of regional induction were converted to color intensities to "paint" the two-dimensional liver for a visual comparison with immunohistochemical observations. To obtain sharp moving boundaries of induced areas with increasing dose (as noted in various experiments), n values in each subcompartment must be large. To create realistic total induction curves that are relatively smooth, the differences in Kd values between adjacent subcompartments must be less than fivefold. Because of the high n values, the low-dose induction characteristics predicted with the multicompartment liver model differ significantly from those predicted with a model that considers the liver as a single homogeneous compartment.

Regional hepatic CYP1A1 and CYP1A2 induction with 2,3,7,8-tetrachlorodibenzo-p-dioxin evaluated with a multicompartment geometric model of hepatic zonation

A physiologically based pharmacokinetic (PBPK) model for 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) was combined with a five-compartment geometric model of hepatic zonation to predict both total and regional induction of CYP450 proteins within the liver. Three literature studies on TCDD pharmacokinetics and protein induction in female rats were analyzed. In simulating low-dose behavior for mRNA in whole liver and, particularly, in representing immunohistochemical observations, the five-compartment model was more successful than conventional homogeneous one-compartment liver models. The five-compartment liver model was used with the affinity of TCDD for the Ah receptor (AhR) held constant across all the liver (Kb = 0.2 nM). The presumed affinities of the AhR-TCDD complex for TCDD responsive elements in the CYP1A1 (Kd1) and CYP1A2 (Kd2) genes varied between adjacent compartments by a factor of 3. This parameterization leads to predicted 81-fold differences in affinities between the centrilobular and the periportal regions. The affinities used for AhR-TCDD complex binding to TCDD response elements for CYP1A2 in compartment 3 (the midzonal area) ranged from 0.08 to 1.0 nM in the three studies modeled. For CYP1A1 the corresponding dissociation constant in compartment 3 varied from 0.6 to 2.0 nM. In each compartment, the Hill coefficient for induction had to be 4 or greater to match the immunohistochemical results. This multi-compartment liver model is consistent with data on protein and mRNA induction throughout the liver and on the regional distribution of these proteins. No previous model has incorporated regional variations in induction. The PBPK analysis based on the multicompartment liver model suggests that the low-dose behavior for hepatic CYP1A1/CYP1A2 induction by TCDD is highly non-linear.

Peri- and postnatal changes in reduced nicotinamide adenine dinucleotide phosphate-cytochrome P-450 reductase content in hepatocytes of rats

To study the process of the expression of reduced nicotinamide adenine dinucleotide phosphate-cytochrome P-450 reductase (EC 1.6.2.4) in the liver during development, the amount of enzyme in the cytoplasm of periportal and perivenular hepatocytes in sections cut from livers of male rats was measured during peri- and postnatal growth by quantitative immunohistochemistry with a video image processor. In livers of 19-day-old foetuses, the reductase content in the cytoplasm of periportal and perivenular hepatocytes was 0.16 microM and 0.20 microM, respectively. From the 19th day of gestation to 5 days after birth, the enzyme content increased markedly in the cytoplasm of periportal (288%) and perivenular hepatocytes (301%). Subsequently, the content in the cytoplasm of periportal hepatocytes increased slightly (46%) from 5 to 20 days of age, remained unchanged from 20 to 45 days of age, and increased slightly (15%) from 45 to 90 days of age. However, the content in the cytoplasm of perivenular hepatocytes increased progressively (125%) between 5 and 90 days of age. Thus, the amount of cytochrome P-450 reductase increases markedly in periportal and perivenular hepatocytes during the perinatal period, and subsequently the enzyme content increases gradually in periportal hepatocytes and progressively in perivenular hepatocytes. The present results also suggest that the divergence between cytochrome P-450 expression and the cytochrome P-450-dependent drug metabolic activity in hepatocytes during the perinatal period, found in previous studies, can be attributed to a low cytochrome P-450 reductase density in the membrane of endoplasmic reticulum of periportal and perivenular hepatocytes.

Intracellular inactivation of the hepatitis B virus by cytotoxic T lymphocytes

It is widely believed that viral clearance is mediated principally by the destruction of infected cells by CTLs. In this report, we use a transgenic mouse model of HBV replication to demonstrate that this assumption may not be true for all viruses. We find that adoptively transferred virus-specific CTLs can abolish HBV gene expression and replication in the liver without killing the hepatocytes. This antiviral function is mediated by IFN gamma and TNF alpha secreted by the CTL or by the antigen-nonspecific macrophages and T cells that they activate following antigen recognition. These cytokines activate two independent virocidal pathways: the first pathway eliminates HBV nucleocapsid particles and their cargo of replicating viral genomes, while the second pathway destabilizes the viral RNA. Intracellular viral inactivation mechanisms such as these could greatly amplify the protective effects of the immune response, while failure of such mechanisms could lead to viral persistence or to the death of the host.

The establishment of the hepatic architecture is a prerequisite for the development of a lobular pattern of gene expression

We have studied the expression patterns of ammonia-metabolising enzymes and serum proteins in intrasplenically transplanted embryonic rat hepatocytes by in situ hybridisation and immunohistochemical analysis. The enzymic phenotype of individually settled hepatocytes was compared with that of hepatocytes being organised into a three-dimensional hepatic structure. Our results demonstrate that development towards the terminally differentiated state with zonal differences in enzyme content requires the incorporation of hepatocytes into lobular structures. Outside such an architectural context, phenotypic maturation becomes arrested and hepatocytes linger in the protodifferentiated state. These features identify the foetal period as a crucial time for normal liver development and show that the establishment of the terminally differentiated hepatocellular phenotype, beginning with the differentiation of hepatocytes from the embryonic foregut, is realised via a multistep process.

Negative selection in hepatic tumor promotion in relation to cancer risk assessment

Mechanistic studies with phenobarbital (PB), 2,3,7,8,-tetrachlorodibenzo-p-dioxin (TCDD) and other liver tumor promoters support a general model of promotion involving negative selection where specifically-mutated cells derive a growth advantage in the presence of persistent mitosuppression. Exposure to these liver tumor promoters appears to transiently enhance hepatocyte replication, presumably via transcriptional activation of growth regulatory genes, leading to a homeostatic increase in mitoinhibitory growth factors in the liver to constrain proliferation. Transforming growth factor beta 1 (TGF-beta), a potent mitoinhibitory growth factor for hepatocytes, has been associated with the mitosuppression caused by PB and certain peroxisomal proliferators. Escape from TGF-beta mitosuppression may involve loss or alteration of function of the mannose 6-phosphate/insulin-like growth factor II (M6P/IGFII) receptor, which is required for TGF-beta 1 activation, or alterations of the TGF-beta types I, II and III signal transduction receptors. A risk assessment based on a negative selection mechanism could be conducted for tumor promotion endpoints with TCDD and compared with current approaches that implicitly regard TCDD as an initiator. Benchmark dose calculation using centrilobular induction of cytochromes P450 1A1 and 1A2 as a surrogate for periportal growth stimulation would provide a rational starting point for application of conventional safety factor approaches, similar to those used with non-cancer effects. In the future, tissue and plasma concentrations of specific growth factors, e.g. TGF-beta or hepatocyte growth factor, HGF, might be considered as more direct dose surrogates for tumor-promoting effects of xenobiotics. Uncertainty factor adjustments to a TCDD benchmark dose calculation should eventually rely on direct knowledge of regulation of specific growth regulatory genes and their receptors in relevant species and on species differences in TCDD pharmacokinetics, instead of application of default animal-to-human and interindividual uncertainty factors.

Persistent expression of genes transferred in the fetal rat liver via retroviruses

The transfer of genes into the fetal liver is a promising approach for correction of inborn errors in metabolism identified in prenatal life. In this study, we demonstrate that gene transfer to the fetal rat liver resulted in the stable expression of the gene in the hepatocytes of the adult animals. This was achieved by a combination of gene transfer via ecotropic retroviruses in the fetal liver with subsequent partial hepatectomy of the offspring. Replication incompetent, ecotropic and amphotropic retroviruses were used to transfer the bovine growth hormone gene (bGH) linked to the promoter (-450 to +73) for the P-enolpyruvate carboxykinase (PEPCK) gene into the fetal liver in the last trimester of gestation. Amphotropic retroviruses were unable to infect the fetal liver due to the lack of expression of their receptors. The fetal liver was infected by the ecotropic retroviruses and partial hepatectomy of the offspring at one month of age stimulated expression of the PEPCK/bGH gene in the liver over ten fold. Expression of the gene persisted for as long as one year. A heterogeneous pattern of expression of the chimeric gene throughout the liver parenchymal cells was identified with higher expression in the pericentral region of the liver. This zonation of expression was not expected, since the endogenous PEPCK gene is expressed in periportal hepatocytes. We suggest that, following partial hepatectomy DNA replication activates expression of the proviral PEPCK/bGH gene, mainly in midzonal and pericentral hepatocytes. Proviral sequences may influence the expression of the PEPCK/bGH gene in parenchymal cells in which the PEPCK promoter is not normally active.

Interpretation of liver biopsies

Liver biopsies help define the disease process and can assist the clinician in forming a prognosis. The small size of liver biopsy specimens and limited responses of the liver to a wide spectrum of injuries often do not allow for a specific etiologic diagnosis. Determination of the distribution and morphology of the lesion often provides useful diagnostic and prognostic information. Maximum value can be attained from liver biopsies when the histologic findings are correlated with the results of other diagnostic tests and there is good communication between the clinician and pathologist.

Heterogeneous expression of cholesterol 7 alpha-hydroxylase and sterol 27-hydroxylase genes in the rat liver lobulus

We investigated the lobular localization and molecular level of expression of cholesterol 7 alpha-hydroxylase and sterol 27-hydroxylase, two key enzymes in bile acid synthesis, in isolated periportal and pericentral hepatocytes and by in situ hybridization of rat liver. Enzyme activity, mRNA, and gene transcription of cholesterol 7 alpha-hydroxylase were predominant in pericentral hepatocytes of control rats, being 7.9-, 9.9-, and 4.4-fold higher than in periportal hepatocytes, respectively. Similar localization was found for sterol 27-hydroxylase: 2.9-, 2.5-, and 1.7-fold higher enzyme activity, mRNA, and gene transcription, respectively, was found in pericentral hepatocytes. Interruption of the enterohepatic circulation with colestid resulted in upregulation of these parameters for both enzymes, as a consequence of stimulated gene expression mainly in the periportal zone. In contrast, mRNA levels and gene transcription of 3-hydroxy-3-methylglutaryl CoA reductase showed opposite lobular distribution. Selective periportal expression for the latter was enhanced, but remained local, after colestid treatment. In situ hybridization showed unambiguously that cholesterol 7 alpha-hydroxylase mRNA is localized exclusively in the pericentral zone and that sterol 27-hydroxylase mRNA is expressed preferentially in the pericentral region, though less pronounced. Administration of colestid led to expression of both genes within a larger area of the liver lobulus. In conclusion, we suggest that cholesterol 7 alpha-hydroxylase and sterol 27-hydroxylase are coordinately regulated by the bile acid gradient over the lobulus, resulting in predominant expression in the pericentral zone. Opposite lobular localization of cholesterol and bile acid synthesis provides an alternative view to interregulation of these metabolic pathways.

Heterogeneity of kinetic parameters of enzymes in situ in rat liver lobules

In the present review, metabolic compartmentation in liver lobules is discussed as being dynamic and more complex than thus far assumed on the basis of numbers of mRNA or protein molecules or the capacity (zero-order activity) of enzymes. Isoenzyme distribution patterns and local kinetic parameters of enzymes may vary over the different zones of liver lobules. As a consequence, metabolic fluxes in vivo at physiological substrate concentrations may be completely different from those that are assumed on the basis of the number of molecules or the capacity of enzymes present in zones of liver lobules. For a more correct estimation of the levels of metabolic processes in the different compartments of liver tissue, local kinetic parameters and substrate concentrations have to be determined to calculate local metabolic fluxes. Direct measurements of metabolic fluxes in vivo with the use of noninvasive techniques is a promising alternative and the techniques will become increasingly important in future metabolic research.

Analysis of enzyme reactions in situ

Estimations of metabolic rates in cells and tissues and their regulation on the basis of kinetic properties of enzymes in diluted solutions may not be applicable to intact living cells or tissues. Enzymes often behave differently in living cells because of the high cellular protein content that can lead to homologous and heterologous associations of protein molecules. These associations often change the kinetics of enzymes as part of post-translational regulation mechanisms. An overview is given of these interactions between enzyme molecules or between enzyme molecules and structural elements in the cell, such as the cytoskeleton. Biochemical and histochemical methods are discussed that have been developed for in vivo and in situ analyses of enzyme reactions, particularly for the study of effects of molecular interactions. Quantitative (histochemical) analysis of local enzyme reactions or fluxes of metabolites has become increasingly important. At present, it is possible to calculate local concentrations of substrates in cells or tissue compartments and to express local kinetic parameters in units that are directly comparable with those obtained by biochemical assays of enzymes in suspensions. In situ analysis of the activities of a number of enzymes have revealed variations in their kinetic properties (Km and Vmax) in different tissue compartments. This stresses the importance of in vivo or in situ analyses of cellular metabolism. Finally, histochemical determinations of enzyme activity in parallel with immunohistochemistry for the detection of the total number of enzyme molecules and in situ hybridization of its messenger RNA allow the analysis of regulation mechanisms at all levels between transcription of the gene and post-translational activity modulation.

Hepatic glutaminase mRNA is confined to part of the urea cycle domain in the adult rodent liver lobule

This in situ hybridization study describes the developmental appearance of the lobular distribution of the mRNA encoding hepatic glutaminase in normal rat liver. Glutaminase has been proposed to provide the urea cycle with ammonia [Häussinger and Gerok (1983) Eur. J. Biochem. 133, 269-275]. Hence, the (developmental) pattern of expression of the mRNA would be expected to be closely linked to that of the urea cycle enzymes. From embryonic day 20 onward, hepatic glutaminase mRNA can be detected along the entire porto-central axis, with predominant expression in the portal area. In the adult phenotype, which is acquired at the end of the first postnatal week, glutaminase mRNA is no longer present along the entire porto-central distance but has become confined to a relatively small periportal domain in which the expression decreases in a porto-central direction. Thus, in contrast to the large periportal domain, in which the urea cycle enzymes are expressed, the glutaminase mRNA-expressing domain is much smaller and not contiguous with the glutamine synthase mRNA-expressing pericentral domain, leaving a midlobular area that is devoid of glutaminase mRNA. A similar pattern of distribution was found in adult mouse liver. The significance of these observations is that, within the liver lobules, there is an area in which glutaminase is not expressed and, hence, glutamine can not be the substrate for urea synthesis.

Vascular branching pattern and zonation of gene expression in the mammalian liver. A comparative study in rat, mouse, cynomolgus monkey, and pig

BACKGROUND

A significant part of the liver volume consists of regions in which hepatocytes are in close contact with large branches of the afferent (portal vein) or efferent (hepatic vein) vessels. As most studies have addressed zonation of gene expression around the parenchymal branches of the portal and hepatic vein only, the patterns of gene expression in hepatocytes surrounding larger vessels are largely unknown.

METHODS

For that reason, we studied the patterns of expression of the mRNAs and proteins of the pericentral marker enzymes glutamine synthase, ornithine aminotransferase, and glutamate dehydrogenase and the periportal marker enzymes phosphoenolpyruvate carboxykinase and carbamoylphosphate synthase in the rat liver, in relation to the branching pattern of the afferent and efferent hepatic veins with immuno and hybridocytochemical techniques. These patterns of expression were compared with those seen in mouse, monkey, and pig liver.

RESULTS

The distribution patterns of the genes studied appear to reflect the "intensity" of the pericentral and periportal environment, glutamine synthase and phosphoenolypyruvate carboxykinase requiring the most pronounced environment, respectively. The patterns of gene expression around the large branches of the portal and hepatic vein were found to be related to the parenchymal branches in the neighbourhood of these large blood vessels. Only the cells of the limiting plate retain their periportal and pericentral phenotype for those marker enzymes that do not require a pronounced periportal or pericentral environment to be expressed. GS-negative areas in the pericentral limiting plate appear to correlate with a local absence of draining central veins, and become more frequent and extensive around the larger branches of the hepatic vein.

CONCLUSIONS

The similarity of the observed patterns of gene expression of the genes studied in mouse, rat, monkey, pig, and man suggests that they reflect a general feature of gene expression in the mammalian liver. A comparison of mouse, rat, pig, and human liver suggests that the presence of glutamine synthase-negative areas reflects the branching order of the efferent hepatic blood vessel.

Heterogeneous (positional) expression of hepatic glutamine synthetase: features, regulation and implications for hepatocarcinogenesis

Glutamine synthetase expression in liver parenchyma is restricted to a small population of pericentral hepatocytes surrounding the central veins. Studies on the development of this heterogeneous (positional) gene expression and of the changes observed in response to experimental alterations of liver physiology or manipulations of hepatocytes in culture have revealed that it is dependent on cell-cell and cell-matrix interactions rather than on the levels of hormones and other modulating factors. The considerable stability of GS expression may point to further events leading to a defined differentiated GS+ phenotype. Observations during experimental hepatocarcinogenesis indicate that strong GS expression may be used for tracing hepatocellular lineages during preneoplastic and early neoplastic stages. Furthermore, these studies suggest a relationship between the GS+ phenotype and enhanced growth of these lesions. Future studies should help to define the diagnostic value of GS and its significance for new chemotherapeutic strategies.

Intralobular heterogeneity of perisinusoidal stellate cells in porcine liver

The aim of the present investigation was to elucidate the intralobular heterogeneity of the perisinusoidal stellate cells (fat-storing cells, lipocytes) in the porcine liver. Their three-dimensional structure, desmin immunoreactivity and vitamin-A storage were studied by use of the Golgi silver, immunocytochemical and gold chloride methods. In order to locate the stellate cells, the hepatic lobules were divided into 10 zones. The stellate cells were readily identified in Golgi preparations by their striking dendritic appearance with branching processes encompassing the sinusoids. The stellate cells in the centrolobular zones were conspicuously dendritic with longer processes in comparison to those emitted by periportal elements. Such arborizations were studded with numerous thorn-like microprojections. Desmin immunoreaction in the periportal zones was stronger than that in the centrolobular zones. Vitamin-A storage in the stellate cells was well developed in zones 2-4, but reduced gradually toward the central region. The perisinusoidal stellate cells display marked heterogeneity in morphology and function based on their zonal location in the hepatic lobule.

Predominant periportal expression of the phosphoenolpyruvate carboxykinase gene in liver of fed and fasted mice, hamsters and rats studied by in situ hybridization

Zonal expression of phosphoenolpyruvate carboxykinase (PCK) mRNA in mouse, hamster and rat liver was studied by in situ hybridization with a radiolabelled rat antisense RNA probe. The abundance of PCK mRNA was determined by Northern blot analysis of total RNA with a digoxigenin-labelled probe. Livers were taken from animals that were sacrificed during the normal day/night cycle and after 29 h fasting. In situ hybridization revealed a heterogeneous distribution pattern of PCK mRNA in the liver of all three species throughout the whole day/night cycle. At the end of the dark period, i.e. at the end of feeding, with rats and mice but at a point of continuous feeding with hamsters, low amounts of PCK mRNA were restricted mainly to the periportal area. At the end of the light period, i.e. at the end of fasting with rats and mice but at a point of continuous feeding with hamsters, PCK mRNA levels were increased to a maximum and extended from the periportal to the intermediate zone. In mouse liver prolonged fasting caused a significant increase in PCK mRNA abundance with a nearly homogeneous distribution within the parenchyma. In hamster and rat liver, however, PCK mRNA levels slightly declined or remained constant, respectively, and the predominant localization of PCK mRNA in the periportal and intermediate zone was preserved. The present data suggest that the heterogeneous zonal activation of the PCK gene was essentially very similar in mouse, hamster and rat liver.

Distribution and activity of glutamine synthase and carbamoylphosphate synthase upon enlargement of the liver lobule by repeated partial hepatectomies

Glutamine synthase and carbamoylphosphate synthase show a strikingly heterogeneous and fully complementary distribution in the rat liver. In the human liver, however, there is a midlobular zone where both enzymes are absent. The diameter of the human liver lobule is approximately twice the size of the rat lobule. To investigate whether lobule size is a major determinant for the expression patterns of glutamine synthase and carbamoylphosphate synthase, Wistar strain rats were partially hepatectomized 3 times, at weekly or monthly intervals. Due to hepatic regeneration the cross-sectional area of the liver lobules increased twofold. However, a midlobular zone which lacked expression of both glutamine synthase and carbamoylphosphate synthase did not develop in these livers, thus showing that lobular size is not a major determinant for the distribution patterns of glutamine and carbamoylphosphate synthase. The twofold increase in the cross-sectional area of the liver lobule was associated with a similar reduction in the relative number of glutamine synthase-positive cells and in the enzyme content of the liver, indicating that the regeneration process does not affect the pericentral pattern of glutamine synthase expression. After regeneration was complete, the glutamine synthase content in the liver was restored to its original value, demonstrating a twofold increase in the cellular concentration of glutamine synthase-positive hepatocytes. An increase in the diameter of the liver lobule was only seen after the first partial hepatectomy. Liver growth following subsequent partial hepatectomies can be explained by an increase in the length of the liver lobule and/or by splitting of liver lobules. The zonal distribution of DNA replication, which is characteristic of the first partial hepatectomy, is lost after repeated partial hepatectomies. Furthermore, evidence was obtained that the signal for inducing DNA synthesis may originate at the level of single liver units.

Cancer-cell traffic in the liver. II. Arrest, transit and death of B16F10 and M5076 cells in the sinusoids

Fluorescent probes were used to detect BUdR-labelled B16F10 and M5076 cancer cells delivered to the livers of mice via intrasplenic injection. In liver sections stained for succinic dehydrogenase, which permits the periportal, acinar zone 1 to be distinguished from the pericentral zone 3, counts were made of the zonal distribution of fluorescent, intact cancer cells and, by default, the numbers of "lost" cells. Very few intact cancer cells leave the liver from the single bolus of the intrasplenic injection, and even fewer of these generate pulmonary lesions; therefore, within the time limits of these experiments, the liver is virtually a closed system. A dynamic view of intrahepatic cancer-cell traffic with respect to zones 1 (periportal) and 3 (pericentral) was obtained from static measurements of cell densities at different times after intrasplenic injection, by means of Markov chain probability analysis. This indicated that, during the first hour after arrival in zone 1 of the liver sinusoids, there is a 10% probability of a B16F10 cell remaining intact in zone 1, an 89% probability of cell death in zone 1 and only a 1% probability of the cell passing into zone 3. During the same period, there is a 77% probability of an M5076 cell remaining intact in zone 1, a 21% probability of death, and a 2% probability of relocation to zone 3. In both cell types, very few cells were lost from zone 3. Further proportional death in zone 1 diminished over the next 23 hr, concomitant with an increased proportion of cell death in zone 3. Our results indicate that, although there is considerable variation between the 2 cell types studied here, most (B16) or many (M5076) of these cancer cells entering the liver via the portal vein die within 1 hr in zone 1 of liver lobules. In addition, very few of the cells delivered to zone 1 travel along the sinusoids to zone 3, and few of these reach the lungs in a viable state.

Histological evaluation of the zonation of colloidal gold uptake by the rat liver

The distribution and endocytotic function of Kupffer cells in the rat liver were studied after administration of fibrinogen stabilized colloidal gold suspensions either by injection directly into the circulatory system of anaesthetized rats or by application to the isolated perfused liver. After exposure to gold particles the livers were perfused with fixative and studied using several microscopic techniques. Gold was predominantly endocytosed by a highly active population of Kupffer cells surrounding the portal spaces resulting in distinct dark patterns around the terminal portal veins. In cross-sections of lobules the pattern appeared as incomplete networks composed of dark triangular areas with distinct borderlines towards light areas concentric with the terminal hepatic veins (central veins). The light areas contained few and relatively inactive small Kupffer cells. A wide variation of conditions gave essentially the same uptake pattern compatible with the concept of microcirculatory zones concentric with the terminal hepatic veins (Lamers et al., 1989; Quistorff and Rømert, 1989), but contradicting the traditional view of microcirculatory zones advanced by Rappaport et al. (1954). Since the same pattern developed during conditions of anoxia, it seems that oxygen is not the stimulus for the developmental distribution of Kupffer cells with high endocytotic activity. In vivo and perfusion experiments gave identical patterns, but a higher endocytotic activity of endothelial cells was found in perfused isolated livers.

Nodular regenerative hyperplasia in the rat induced by a selenium-enriched diet: study of a model

Weaned male rats were fed a 4 ppm selenium diet. Compared after 2 mo with a control group fed a 0.4 ppm diet, the rats' body weights had not significantly decreased and liver function was normal, but portal pressure was 1.8 times higher (p less than 0.05). Liver weight was slightly increased (10.3%; p less than 0.05). All livers had an abnormal appearance. In the less severe cases the surface was only slightly irregular, but in the more severe cases, atrophic micronodular lobes and hypertrophic lobes, with mildly irregular surfaces, were present. On light microscopy, atrophic lobes displayed a peripheral nodular zone with micronodules separated by rows of atrophic hepatocytes without fibrosis, characteristic of nodular regenerative hyperplasia, and a central atrophic zone that was sometimes peliotic. Hypertrophic lobes and livers in the less severe cases had only minor and relatively localized evidence of nodular regenerative hyperplasia; occasional peliosis was seen. In all cases portal veins, hepatic veins and hepatic arteries were normal. By electron microscopy, in nonnodular zones with no obvious evidence of parenchymal atrophy, the endothelial wall showed signs of complete or incomplete capillarization with frequent enlargement of the Disse space. The selenium-enriched diet is a reproducible model of liver nodular regenerative hyperplasia. In this model, damage to the sinusoidal wall could represent the primum movens of microcirculatory disturbances.

Metabolic zonation of the liver: regulation and implications for liver function

Liver parenchyma shows a remarkable heterogeneity of the hepatocytes along the porto-central axis with respect to ultrastructure and enzyme activities resulting in different cellular functions within different zones of the liver lobuli. According to the concept of metabolic zonation, the spatial organization of the various metabolic pathways and functions forms the basis for the efficient adaptation of liver metabolism to the different nutritional requirements of the whole organism in different metabolic states. The present review summarizes current knowledge about this heterogeneity, its development and determination, as well as about its significance for the understanding of all aspects of liver function and pathology, especially of intermediary metabolism, biotransformation of drugs and zonal toxicity of hepatotoxins.

Liver units in three dimensions: I. Organization of argyrophilic connective tissue skeleton in porcine liver with particular reference to the "compound hepatic lobule"

Liver units were investigated in pig livers by means of histologic serial tracing, physical model building, and computer-aided three-dimensional imaging. Observations of the argyrophilic connective tissue skeleton were based mainly on the celloidin-embedded serial sections treated with silver impregnation. The parenchymal mass that clothed the initial segments of hepatic venous radicles was demarcated by fibrous septa which formed isolable units with two basic patterns: the simple hepatic lobule (SHL) and the compound hepatic lobule (CHL). Both lobule types presented regular limiting structures circumscribing each unit. Three-dimensional studies revealed that 25% of the lobules in a section belonged to the SHL type and 75% to the CHL type, the latter being predominant among the surface lobules. When considered in only two dimensions, however, the SHL-like lobules constituted the majority. Polygonal analysis disclosed that the pentagonal lobule was the most typical, instead of the "hexagonal" or "classic" lobule. The CHLs represented a multiaxial unit containing a system of venous tributaries in accordance with intralobular septation, whereas the SHLs were found with one axial vessel having a dendritic tendency at the incipient end; some SHLs were drained eccentrically by separate vessels into a sublobular vein. It was observed that, in dividing CHLs, whereas particular sinosoids were transformed into portal twigs, other sinusoids were changed into central venous tributaries. Fibrous deposition occurred along the septal-line sinusoids, bringing into view the septum-initiating plane. Fibroconnective tissue was supplied from the portal area and central (sublobular) adventitia, where portal triad structures and adventitial arterioles, respectively, were included. The findings of the present study facilitate the understanding of several characters of the lobules that have been reported previously, or occasionally postulated, such as the portal-central bridging tendency, the intralobular arterioles or ductules, the translobular artery or portal vein, the "portal-portal" or "portal-central" anastomoses, and the apposition of pericentral zone close to periportal zone. Based on differences in argyrophilia of sinusoidal reticulum, in proportion of lobule types, and in vasculature, the anatomic heterogeneity of liver unit was demonstrable in zonality, regionality, and locality.

Diet- and hormone-induced reversal of the carbamoylphosphate synthetase mRNA gradient in the rat liver lobulus

A hybridocytochemical analysis of adult liver from normal control and from hormonally and dietary-treated rats was carried out, using radioactively-labelled probes for the mRNAs of glutamine synthetase (GS), carbamoylphosphate synthetase (CPS) and phosphoenolpyruvate carboxykinase (PEPCK). In line with previous findings, GS mRNA is exclusively expressed in a small pericentral compartment, CPS mRNA exclusively in a contiguous large periportal compartment and PEPCK mRNA across the entire porto-central distance. The density of labelling in CPS and PEPCK mRNA-positive hepatocytes decreases in a porto-central direction. Starvation resulted in a reversal of the gradient of CPS mRNA within its periportal compartment; glucose refeeding counteracted this effect. Livers of glucocorticosteroid-treated, starved or diabetic rats also revealed a reversal of the normal gradient of CPS mRNA, but now across the entire porto-central distance. The patterns of expression of GS and PEPCK mRNA remained essentially unchanged, notwithstanding substantial changes in the levels of expression. It is concluded that blood-borne factors constitute the major determinants for the expression patterns of CPS mRNA within the context of the architecture of the liver lobulus.

Expression patterns of mRNAs for alpha-fetoprotein and albumin in the developing rat: the ontogenesis of hepatocyte heterogeneity

In developing and normal adult rat liver the expression patterns of the mRNAs for alpha-fetoprotein (AFP) and albumin (ALB) were analysed by in situ hybridization using specific 35S-labelled complementary DNA probes. In the developing liver AFP and ALB mRNA are found from embryonic day (ED) 11 and 12, respectively, onward. At ED 20 the first signs of a zonal distribution of these mRNAs across the liver lobule can be observed, AFP mRNA concentration being higher in the pericentral area and ALB mRNA concentration higher in the periportal area. This distribution pattern of reciprocal, overlapping gradients of mRNA can be clearly recognized in the neonatal period. In the adult liver AFP mRNA can no longer be detected and similar to the neonatal situation, ALB mRNA is expressed across the entire porto-central distance decreasing in concentration going from the portal to the central area. Transient extra-hepatic expression of AFP mRNA is found in the embryonic heart and in the epithelial lining of intestine and lung; furthermore, AFP and ALB mRNA are found to be transiently expressed in the developing renal tubules. Similar expression patterns have been observed for other liver-characteristic mRNAs (Moorman et al., 1990), suggesting that common regulatory factors are operative during development.

Hepatocytes explanted in the spleen preferentially express carbamoylphosphate synthetase rather than glutamine synthetase

Urea cycle enzymes and glutamine synthetase are essential for NH3 detoxification and systemic pH homeostasis in mammals. Carbamoylphosphate synthetase, the first and flux-determining enzyme of the cycle, is found only in a large periportal compartment, and glutamine synthetase is found only in a small, complementary pericentral compartment. Because it is not possible to manipulate experimentally the intrahepatic distribution of carbamoylphosphate synthetase and glutamine synthetase, we looked for conditions in which explanted hepatocytes would exhibit either the carbamoylphosphate synthetase phenotype or glutamine synthetase phenotype. In the spleen hepatocytes either settle as individual cells or in small agglomerates. The dispersed cells only express the carbamoylphosphate synthetase phenotype. Within the agglomerates, sinusoids that drain on venules develop. Hepatocytes surrounding the venules stain only weakly for carbamoylphosphate synthetase but are strongly positive for glutamine synthetase. These observations were made for explanted embryonic hepatocytes (no prior expression of either carbamoylphosphate synthetase or glutamine synthetase), neonatal hepatocytes (compartments of gene expression not yet established) and adult periportal and pericentral hepatocytes.

Transferrin receptor expression in rat liver: immunohistochemical and biochemical analysis of the effect of age and iron storage

Hepatic transferrin receptors were studied in normal male rats at 1 to 59 wk after weaning, using immunohistochemical and biochemical techniques. The number of transferrin receptors measured and the intensity of the staining in situ decreased rapidly during the first 10 wk of life and more slowly thereafter. Immunohistochemistry further demonstrated changes in the topographical and (sub)cellular localization of the transferrin receptor. In the young rat livers, staining was almost exclusively present on hepatocytes in acinar zone 2 + 3 in a honeycomb to sinusoidal pattern. With aging, a panacinar heterogeneous and mainly sinusoidal staining of hepatocytes was more frequent. Kupffer cell positivity was more obvious as compared with the young rat livers. The observed changes in transferrin receptor expression may partly be explained by age-dependent alterations in DNA synthesis and proliferative potential of the liver cells. A series of rats were iron loaded with carbonyl iron up to 39 wk and "unloaded" by administration of a normal diet during 20 wk. In these animals, serial histochemical studies showed predominantly parenchymal (7 to 14 wk), mixed parenchymal and reticuloendothelial (39 wk) and almost exclusive reticuloendothelial siderosis (59 wk). In the siderotic livers transferrin receptor numbers tended to be lower than in the controls with significant differences after 14 and 39 wk. Immunohistochemistry showed decreased parenchymal but increased reticuloendothelial transferrin receptor expression with iron load. After the period of unloading, parenchymal transferrin receptors were virtually absent despite the negligible siderosis of these cells. In contrast, siderotic reticuloendothelial cells were intensely positive. These findings support down-regulation of parenchymal transferrin receptor resulting from iron storage. However, the positivity of siderotic reticuloendothelial cells and the absence of re-emergence of parenchymal receptors in conditions of minimal parenchymal and prominent reticuloendothelial siderosis need further elucidation.