Coexpression of periportal and perivenous enzymes in rat hepatocytes after experimental bile duct...

Histochem Cell Biol (1996) 105:319-329 �9 Springer-Verlag 1996

Lorraine R a c i n e - S a m s o n . Jean-Yves Scoazec Alain Moreau �9 Laurence Christa D o m i n i q u e Bernuau �9 G6rard Fe ldmann

Coexpression of periportal and perivenous enzymes in rat hepatocytes after experimental bile duct ligation: comparison with intrasplenically transplanted hepatocytes

Accepted: 18 December 1995

Abstract The coexpression of normally periportal and perivenous markers has been described in heterotopically transplanted hepatocytes. To determine whether such a coexpression might also occur in hepatocytes retaining their original intrahepatic location, we compared in bile- duct-ligated livers and intrasplenically transplanted hepatocytes, the expression and distribution of the predominantly periportal glucose-6-phosphatase, succinate dehydrogenase, and lactate dehydrogenase, the predominantly perivenous glutamate dehydrogenase, NADPH-dehydrogenase, and ~-hydroxybutyrate dehydrogenase, and the strictly perivenous glutamine synthetase. The coexpression of high levels of the two periportal markers glucose- 6-phosphatase and lactate dehydrogenase and of the perivenous marker NADPH dehydrogenase was observed in two situations: in clusters of hepatocytes isolated within the ductular proliferation in bile-duct-ligated livers and the majority of intrasplenically transplanted hepatocytes. The expression of glutamine synthetase was different according to the site. The protein was observed in certain intrasplenically transplanted hepatocytes bordering the splenic vessels but was never detected in hepatocyte clusters found in bile-duct-ligated livers. Our study therefore suggests that the coexpression of periportal and perivenous markers in the same hepatocytes is likely to be a non-specific consequence of the loss of the normal connections of hepatocytes with the normal liver microcirculation.

L. Racine-Samson - J.-Y. Scoazec (~) �9 A. Moreau �9 D. Bernuau G. Feldmann Laboratoire de Biologic Cellulaire and INSERM U327, Facult6 de M6decine Xavier Bichat, Universit6 Paris 7 Denis Diderot, BP 416, F-75870 Paris Cedex 18, France Tel. +39 1 44856190; Fax +39 1 44859279

L. Christa INSERM U370, Facult6 de M6decine Necker-Enfants Malades, Paris, France

Introduction

The mammalian liver is organized into functional units referred to as lobules or, according to Rappaport's con- cept, as acini (Gumucio 1989). Three main zones are usually recognized within the hepatic lobule: the periportal, the mediolobular, and the perivenous. According to their position in the various zones of the lobule, hepatocytes present distinctive phenotypes resulting from zonal differences in the expression of many enzyme activities (Gumucio 1989; Jungermann and Katz 1989; Gebhardt 1992) and of some metabolic products, such as plasma proteins (Feldmann et al. 1992). The functional heterogeneity of hepatocytes is maintained by a complex network of microenvironmental factors (Gumucio 1989; Gebhardt 1992), including: (a) blood=borne factors, such as the intralobular gradients of oxygen, metabolites, and hormones, determined by the unidirectional perfusion imposed by the vascular architecture of the liver; (b) pericellular factors, such as the cell-cell and cell-matrix interactions determined by the tissue organization of the liver and; possibly, (c) paracrine interactions between the hepatocytes of the different zones of the hepatic lobule.

The transplantation of rat hepatocytes in extrahepatic sites, such as the spleen or the interscapular fat pad, results in the loss of their position-dependent metabolic specializations. The most characteristic changes are the coexpression in the same hepatocytes of both normally periportal and perivenous markers (Gebhardt et al. 1989; Traber et al. 1989; Lamers et al. 1990; Maganto et al. 1990; Chen et al. 1994), such as, respectively, carbamoyl phosphate synthetase and glutamine synthetase (Geb- hardt et al. 1989; Lamers et al. 1990), and the induction in originally periportal hepatocytes of markers associated with the perivenous region of the intact hepatic lobule, such as members of the cytochrome P450 superfamily (Chen et al. 1994).

The mechanisms responsible for the coexpression of periportal and perivenous markers in heterotopically transplanted hepatocytes remain largely unknown. Sever- al hypotheses may be raised. For example, the coexpres-

320

sion of periportal and perivenous markers could be a non-specific response to heterotopic transplantation, which removes the hepatocytes from the normal regula- tory, position-dependent factors operating in the liver tissue. Alternatively, coexpression could be the product of inducing, tissue-dependent factors, operating only in extrahepatic sites. Interestingly, the coexpression of periportal and perivenous markers has never been reported for hepatocytes retaining their original intrahepatic location, even under conditions resulting in marked alterations of their pericellular microenvironment, such as in experimental cirrhosis (Van Noorden et al. 1987; Sokal et al. 1990). This might indicate that repressing factors act in the hepatic tissue to prevent the coexpression of normally periportal and perivenous markers in the same hepatocytes. However, this question is in need of reeval- uation, since the use of different sets of phenotypic markers in studies of experimental cirrhosis (Van Noor- den et al. 1987; Sokal et al. 1990, 1992) and of heterotopically transplanted hepatocytes (Gebhardt et al. 1989; Traber et al. 1989; Lamers et al. 1990; Maganto et al. 1990; Chen et al. 1994) makes it difficult to perform ac- curate comparisons between the two experimental situations.

We therefore decided to address the following two questions. First, is it possible to observe a coexpression of normally periportal and perivenous markers in hepatocytes retaining their original intrahepatic location and, particularly, is it possible to observe the induction of perivenous markers in originally periportal hepatocytes retaining their normal position? Second, is it possible to observe the coexpression of the corresponding markers in heterotopicatly transplanted hepatocytes? In order to address these questions, we analyzed the expression of a large set of periportal and perivenous markers in the rat liver after bile duct ligation. This widely used model of experimental cirrhosis is particularly relevant to our pur- poses since it makes it possible to obtain periportal hepatocyte clusters isolated from the remaining liver parenchyma (Van Noorden et al. 1987). We simultaneously analyzed the pattern of expression of the same markers in intrasplenically transplanted hepatocytes, which, in previous studies, have provided clear-cut evidence of the coexpression of normally periportal and perivenous markers.

The following representative periportal and perivenous markers were analyzed: (a) three enzymes distributed in gradients with periportal predominance, glucose-6- phosphatase (G6Pase), succinate dehydrogenase (SDH), and lactate dehydrogenase (LDH); (b) three enzymes distributed in gradients with perivenous predominance, glutamate dehydrogenase (GDH), NADPH dehydrogenase (ND), and [3-hydroxybutyrate dehydrogenase (HBDH); and (c) glutamine synthetase, restricted to a small ring of hepatocytes surrounding the centrilobular veins.

Materials and methods

Experimental models

Bile duct Iigation

Eight adult male Sprague-Dawley rats (Charles River, Saint Aubin l~s Elbeuf, France) weighing 200-250 g were submitted to bile duct ligation-section using the technique described by Kountouras et aI. (1984) and were killed after 3 weeks.

Hepatocyte transplantation into the spleen

Adult male syngenic Wistar Furth rats (Iffa Credo, L'Arbresle, France), initially weighing 100 to 150 g, were used as hepatocyte donors (n=8) or recipients (n=40) for hepatocyte transplantation. Isolated hepatocytes were prepared from the liver of a donor rat using the collagenase perfusion technique of Berry and Friend (1969). Hepatocytes were transplanted into the spleen according to the technique described by Mito et al. (1979), as modified by Nordlinger et al. (1987). Briefly, 5x106 viable hepatocytes, sus- pended in 0.4 ml Leibowitz medium (Gibco, Cergy-Pontoise, France) containing 15% rat serum, were injected into the splenic parenchyma of recipient rats using a 27-gauge intradermic nee- dle. The splenic vessels were clamped during the injection to reduce cell leakage into the portal vein, and a hemostatic sponge was applied at the site of injection to reduce hemorrhage and cell loss. Animals were killed and their liver and spleen were removed at 1o 3, 6, 10, 14, 18, and 20 months after hepatocyte transplantation.

Control rats

For each model, sham-operated animals served as controls.

Morphological techniques

Tissue samples were snap-frozen in isopentane prechilled in liquid nitrogen for histoenzymology and immunohistochemistry and stored at -80~ until use. Part of the specimens was fixed by im- mersion in 4% paraformaldehyde, 0.1 M phosphate buffer pH 7.4, and embedded in paraffin for histological examination.

Histological examination

For histological examination, 3-gm-thick sections were stained with hematoxylin and eosin. Bile duct cells were specifically stained with a monoclonal anti-cytokeratin 19 antibody (Amers- ham International, Little Chalfont, UK) revealed by an indirect immunoperoxidase technique.

HistoenzymologicaI analysis of the distribution of selected marker enzymes

The following enzymes were studied: G6Pase (EC 3.1.3.9), LDH (EC 1.1.1.27), SDH (EC 1.3.99.1), ND (EC 1.6.99.1), HBDH (EC 1.1.1.30), and GDH (EC 1.4.1.3).

The distribution of enzyme activities was assessed by using histoenzymological techniques applied to 12-gm-thick serial sections of frozen liver tissue. All reactions were performed at 37~ Detection of LDH, GDH, and HBDH was performed essentially as previously described (Sokal et al. 1989; Van Noorden and Frederiks 1992). Briefly, acetone-fixed frozen sections were incubated in 0.1 M phosphate buffer pH 7.6, 20% polyvinyl alcohol, 0.1 M substrate (respectively, lactate, glutamate or 13-hydroxybutyrate), 10 mM NAD +, 10 mM N-ethylmaleimide, 5 mM nitro blue tetrazolium. All reagents were from Sigma (St Louis, Mo.,

USA). After incubation, sections were washed in 0.1 M phosphate buffer pH 5.3 and postfixed in 4% paraformaldehyde, 2% CaC12, rinsed in distilled water and mounted in aqueous medium (Aquamount Gurr, BDH, Poole, UK). For detection of SDH and ND, acetone-fixed sections were processed according to Braak- man et al. (1991). The incubation medium for SDH contained 47 mM phosphate buffer pH 7.6, 47 mM sodium succinate, 0.14 mM CaC12, 0.19 mM A1C13, 28 mM NaHCO 3 and 5 mM nitro blue tetrazolium. For demonstration of ND activity, incubation was performed in 50 mM phosphate buffer pH 7.6, 5 mM nitro blue tetrazolium and 1.2 mM NADPH. After incubation, sections were washed in distilled water, postfixed in 4% paraformaldehyde, 2% CaC12 rinsed in distilled water then mounted in aqueous medium (A'quamouut Gurr). For detection of G6Pase, the technique used was that described by Chiquoine (1953). Briefly, unfixed frozen sections were incubated in a solution containing 2.5 mM glucose-6-phosphate, 0.1 M sodium acetate, 0.15% lead acetate pH 6, rinsed in running water, incubated in 2% ammoni- um sulfide, rinsed, and mounted in aqueous medium. All reagents were from Sigma (St. Louis, Mo., USA).

Negative controls were obtained by omitting the specific substrate and the corresponding cofactors. Positive controls were obtained by including sections of normal human liver in each histoenzymological experiment. Fragments of normal human liver used as controls were stored for the same duration as the fragments of diseased liver processed in the same experiment.

321

lar gradients, in accordance with data in the literature (Novikoff 1959; Shank et al. 1959; Jungermann and Sasse 1978). Moderate gradients with periportal predominance were observed for G6Pase (Fig. l a), LDH, and SDH activities. Moderate gradients with perivenous predominance were observed for ND (Fig. lb) and HBDH activities. A sharp gradient with perivenous predominance was observed for GDH activity.

Distribution of glutamine synthetase

Glutamine synthetase was expressed in a restricted compartment formed by one to three rings of hepatocytes surrounding centrilobular veins (Fig. lc) and large col- lecting vessels. No staining was detected in the hepatocytes of the rest of the lobule.

Experimental bile duct ligation

Immunohistochemical analysis of the distribution of glutamine synthetase

A cross-reacting rabbit polyclonal antibody raised against recom- binant human glutamine synthetase prepared as previously described (Christa et al. 1994), was used in an indirect immunoperoxidase technique applied to 6-gm-thick acetone-fixed, air-dried cryostat sections of frozen tissue. Rehydrated sections were incubated with the primary antibody for 90 min, then rinsed in phosphate-buffered saline (PBS) and further incubated for 90 min with goat species-specific peroxidase-labeled anti-rabbit immunoglobulin preadsorbed F(ab')2 antibody (Immunotech, Marseilles, France). The peroxidase reaction product was developed according to Graham and Karnovsky (1966). Sections were lightly counter- stained with Mayer's hematoxylin and mounted. Negative controls were incubated in the absence of primary antibody, substituted with PBS, or with isotypic mouse immunoglobulins.

In some experiments, a double immunolabeling, combining immmuno-alkaline phosphatase for detection of glutamine synthetase and immunoperoxidase for detection of cytokeratin 19, was performed in the same sections. Primary antibodies followed by species-specific anti-mouse immunoglobulin F(ab')2 antibody conjugated with peroxidase (Immunotech, Marseille-Luminy, France) and species-specific anti-rabbit immunoglobulin F(ab')2 antibody conjugated with alkaline phosphatase (Immunotech, Marseille-Lu- miny, France) were simultaneously incubated on the same section. Peroxidase activity was revealed with aminoethylcarbazole to obtain reddish brown deposits. Alkaline phosphatase activity was revealed with 10 mg Fast blue diazonium salt, diluted in 9.8 ml TRIS pH 8.2 containing 200 gl N,N-dimethylformamide, 2 mg naphtol AS-MX phosphate, and 2.4 mg levamisole, to obtain blue deposits.

Results

Control rat livers

Distribution of zonal marker enzymes

In control rat livers, G6Pase, ND, SDH, LDH, GDH, and HBDH activities were distributed along continuous lobu-


In all cases, marked histological changes were observed. Normal liver architecture was disrupted by an extensive bile duct proliferation originating f rom the periportal areas and accompanied by massive extracellular matrix de- position. As previously described (Poo et al. 1992), the extent of ductular proliferation was variable. In five animals, it was restricted to the immediate periportal region (Fig. 2a). Large, interconnected zones of hepatic parenchyma, corresponding to the mediolobular and perivenous zones of the original lobules, were preserved. In three animals, the parenchymal architecture was marked- ly disrupted. A massive ductular proliferation completely surrounded small islands of residual parenchyma centered by a centrilobular vein (Fig. 2b). In all cases, small, isolated clusters formed by a few hepatocytes were trapped among the bile duct proliferation (Fig. 2c). These clusters were more numerous in cases of massive ductular proliferation.

Distribution of zonal enzyme markers

In the five animals with moderate ductular proliferation, the distribution of zonal enzyme markers retained a certain degree of heterogeneity along hepatocyte plates. The activities of the three periportal markers, G6Pase (Fig. 3a), LDH (Fig. 4a), and SDH, were readily detected all over the residual parenchyma and presented a slight predominance at the contact of enlarged portal tracts. The distribution of the three perivenous markers, ND (Fig. 3b), GDH (Fig. 4b), and HBDH, retained a more marked heterogeneity. As in the normal liver, ND and HBDH activities predominated around the centrilobular veins. In contrast to the normal liver, GDH activity was

322

Fig. la-c Normal rat liver: distribution of glucose-6-phosphatase (G6Phase; a), NADP dehydrogenase (ND; b), and glutamine synthetase (c). a The predominance of G6Pase activity, as detected by histoenzymology, in the periportal zone of the rat hepatic lobule, b The perivenous predominance of ND activity, as detected by histoenzymology, e The immunohistochemical identification of glutamine synthetase in a restricted compartment, formed by a narrow ring of hepatocytes surrounding sections of centrilobular vein. (PT Portal tract, CV centrilobular vein) Original magnifications: a, b x70, e xl00

more intense not only around the centrilobular veins, but also at the contact of enlarged portal tracts.

In the three animals with massive ductular proliferation, the expression of zonal enzyme markers was mark- edly altered in residual hepatocyte plates. The normally periportal enzyme, G6Pase (Fig. 3c), was barely detectable. The activities of the other periportal markers, LDH (Fig. 4c) and SDH, were retained and homogeneously distributed along the hepatocyte plates. The normally perivenous enzymes, ND (Fig. 3d), GDH (Fig. 4d), and HBDH, were homogeneously distributed.

The clusters of hepatocytes trapped among the ductular proliferation presented a constant and distinctive phenotype. They showed very high activities of the normally periportal enzymes G6Pase (Fig. 3c) and LDH (Fig. 4c), and of the normally perivenous enzyme ND (Fig. 3d). In contrast, they presented low activities of the

323

Fig. 2a-c Immunohistochemical localization of cytokeratin 19 in the livers of bile duct-ligated rats killed after 3 weeks, a An example of moderate ductular proliferation (arrowheads), limited to the periportal and mediolobular zones. Large areas of intact parenchyma are preserved, b An example of massive ductular proliferation resulting in a marked disruption of the liver parenchyma. A small island of parenchyma, centered by a centrilobular vein, is surrounded by numerous proliferating bile ducts (arrowheads), which trap isolated clusters of hepatocytes, c A small cluster of hepatocytes (H), surrounded by several bile ductules. (PT Portal tract, CV centrilobular vein) Immunoperoxidase with light nuclear coun- terstaining with Harris' hematoxylin. Original magnifications: a, b x40, c x240

normally periportal enzyme SDH and of the normally perivenous enzymes GDH (Fig. 4d) and HBDH.


Glutamine synthetase retained its restricted distribution. In cases of either moderate (Fig. 5a) or massive (Fig. 5b) ductular proliferation, the protein remained detectable in the few rings of hepatocytes surrounding the centrilobular veins. The protein was never detected in the hepatocyte clusters trapped among the bile duct proliferation.

Hepatocytes transplanted into the spleen


In all cases, the intrasplenic graft of hepatocytes suc- ceeded. In the 23 animals killed 1-3 months after transplantation, transplanted hepatocytes were scattered all over the splenic red pulp and formed small clusters. In the 17 animals killed 3-20 months after transplantation, large nodules of hepatocytes occupied nearly the whole splenic red pulp. Intrasplenic hepatocytes never reorga- nized any organoid structure.

Distribution of zonal enzyme markers

No difference in the expression or distribution of enzyme activities was observed at the various times of sacrifice. Like normal hepatocytes, intrasplenically transplanted hepatocytes expressed detectable activities of G6Pase (Fig. 3e), ND (Fig. 3f), LDH (Fig. 4e), SDH, and GDH (Fig. 4f). In contrast, the activity after HBDH was unde- tectable at any time of transplantation. In all cases, intrasplenically transplanted hepatocytes presented a significant level of interindividual heterogeneity in the apparent level o f expression of the various enzymatic activities. However, the comparison of serial sections showed that most intrasplenic hepatocytes were characterized by the

coexpression of high levels of the two periportal markers G6Pase and LDH and of the perivenous marker ND (Figs. 3e, f, 6).


Glutamine synthetase was usually detected in the hepatocytes located within the connective tissue surrounding the splenic vessels (Fig. 5c). In intrasplenically transplanted hepatocytes located at a distance from splenic vessels, glutamine synthetase was only occasionally detected.

Discussion

Our study demonstrates for the first time that normally periportal and perivenous markers may be coexpressed in hepatocytes retaining their original intrahepatic location, and not only in heterotopically transplanted hepatocytes. In bile duct-ligated livers, such a coexpression is restricted to clusters of originally periportal hepatocytes isolated in the bile ductular proliferation. The coexpression of certain, but not all, of these markers may be observed in heterotopically transplanted hepatocytes.

In both bile duct-ligated rat livers and intrasplenically transplanted hepatocytes, we identified the presence of a population of hepatocytes characterized by the coexpression of high levels of the normally periportat markers G6Pase and LDH and of the normally perivenous marker ND. In intrasplenically transplanted hepatocytes, this phenotype was displayed by the majority of cells. In bile duct ligated-livers, it was restricted to the clusters of hepatocytes trapped within the ductular proliferation and was likely to correspond, on topographical grounds, to residual periportal hepatocytes. Our study therefore provides direct evidence that the coexpression of periportal and perivenous markers in the same hepa-

Fig. 3a-f Comparison of the expressions of G6Pase and ND activities in bile duct-ligated livers and in intrasplenically transplanted hepatocytes. Histoenzymological detection of G6Pase (a) and ND (b) activities in bile duct-ligated rat livers with moderate bile duct proliferation. G6Pase activity is readily detected all over the hepatic parenchyma and presents a slight predominance around an enlarged portal tract (PT). ND activity retained a well-marked predominance in the perivenous region, at a distance from an enlarged portal tract (PT). Histoenzymological detections of respectively G6Pase (e) and ND (d) activities in bile duct-ligated rat livers with massive bile duct proliferation. G6Pase activity is barely detected in the residual parenchyma (P) but is strongly expressed in clusters of hepatocytes trapped among the ductular proliferation (arrows). ND activity is detected both in the residual parenchyma (P) and in clusters of hepatocytes scattered amongst bile ductules (arrows). Histoenzymological detection of respectively G6Pase (e) and ND (f) activities in adjacent sections across a cluster of intrasplenically transplanted hepatocytes (arrows). The two enzymatic activities are strongly detected. Original magnifications: a, b x70, e x120, d x150, e x180, f x180

325

tocytes and the induction of perivenous markers in periportal hepatocytes are not specific for heterotopically transplanted hepatocytes but might also be observed in hepatocytes retaining their normal intrahepatic position. It is interesting to recall that earlier incidental observations have noted the expression of normally perivenous markers, such as glutamine synthetase or members of the cytochrome P450 family, in clusters of hepatocytes isolated at the periphery of cirrhotic nodules in experimental CC14-induced cirrhosis in the rat (Gebhardt and Reichen 1994) and in human liver disease (Sano et al. 1989).

In bile duct-ligated livers, the coexpression of normally periportal and perivenous hepatocytes was specific for isolated hepatocyte clusters. No comparable alteration was observed in the adjacent cirrhotic nodules, in which the overall gradients observed in intact liver tissue are usually retained (Van Noorden et al. 1987; Sokal et al. 1992). It must be noted that, in our study, the expression of glutamine synthetase was constantly retained around residual centrilobular veins. Our findings, therefore, contrast with the observations performed in CC14-induced rat liver cirrhosis (Gebhardt and Reichen 1994), in which no glutamine synthetase expression was detected in cirrhotic nodules. This apparent discrepancy might be due to differences in histogenesis between the various experimental models of rat cirrhosis. In CC14-induced cirrhosis, the initial site of injury is perivenous. The perivenous glutamine synthetase-expressing hepatocytes are, therefore, selectively affected. In contrast, in bile duct ligation, the initial site of injury is periportal and the perivenous glutamine synthetase-expressing hepatocytes are not directly affected. Our observations reinforce the con- cept that the distribution of zonal markers in experimental cirrhosis in the rat depends on the site of initial injury (Sokal et al. 1992).

The emergence of comparable phenotypes in both intrahepatic periportal hepatocyte clusters and intrasplenically transplanted hepatocytes is likely to result from comparable alterations in the respective hepatocyte mi- croenvironments. One of the most significant of these alterations may be the loss of the normal anatomical connections of hepatocytes with the liver microcirculation. Because of their anatomical position, intrasplenically transplanted hepatocytes are removed from the influence of the portal blood circulating in hepatic sinusoids. In the same way, reconstruction studies strongly suggest that periportal intrahepatic hepatocyte clusters have lost any direct connection with the sinusoidal microcirculation (Van Noorden et al. 1987; Dinges et al. 1992). It might therefore be hypothesized that some of the phenotypic changes common to intrahepatic clusters and to intrasplenically transplanted hepatocytes are a non-specific response to the removal of hepatocytes from the influence of the regulating factors transported by the sinusoidal blood.

It must be noted that, despite their similarities, the phenotypes of hepatocyte clusters in bile duct-ligated livers and of heterotopically transplanted hepatocytes also

Fig. 4a-f Comparison of the expressions of lactate dehydrogenase (LDH) and glutamate dehydrogenase (GDH) activities in bile duct-ligated livers and in intrasplenically transplanted hepatocytes. Histoenzymological detections of respectively LDH (a) and GDH (b) activities in bile duct-ligated rat livers with moderate bile duct proliferation. LDH activity is readily detected all over the hepatic parenchyma and presents a slight predominance around an enlarged portal tract (PiO. GDH activity retained a slight predominance in the perivenous region, around a centriIobular vein (CV), but is also detected at the contact of an enlarged portal tract (PT). Histoenzymological detection of LDH (c) and GDH (d) activities in bile duct-ligated rat livers with massive bile duct proliferation. LDH activity is readily detected in the residual parenchyma (P) as well as in the clusters of hepatocytes trapped among the ductular proliferation (arrows). GDH activity is homogeneously distributed in the residual parenchyma (P) but is faintly detected in clusters of hepatocytes scattered amongst bile ductules (arrow). Histoen- zymological detection of LDH (e) and GDH (f) activities in adjacent sections across a cluster of intrasplenically transplanted hepatocytes (arrows). The two enzymatic activities are strongly detected but present a significant level of interindividual heterogeneity. Original magnifications: a, b x70, c x100, d xl00, e x180, f x180

327

present some significant differences. We observed that g lutamine synthetase was never detected in intrahepatic periportal hepatocyte clusters but was present in the ma- jor i ty of intrasplenical ly transplanted hepatocytes , particularly in cells located at the contact o f intrasplenic vessels, as previously descr ibed (Gebhardt et al. 1989; La- mers et al. 1990). In contrast, H B D H was detected in in- t rahepatic hepatocyte clusters but not in intrasplenically t ransplanted hepatocytes . Our observat ions are reminis- cent o f the phenotypic differences observed be tween het- erotopical ly transplanted hepatocytes , according to their site of transplantation. For instance, intrasplenical ly transplanted hepatocytes predominant ly expressed glutamine synthetase (Lamers et al. 1990), while hepatocytes t ransplanted in the interscapular fat pad lack any detectable express ion of this protein (Gebhardt et al. 1989). It is therefore l ikely that the phenotypic characterist ics o f hepatocytes r emoved f rom the normal liver microcirculation are modula ted by local, t issue-specific, pericel lular factors. It is interesting to note that previous studies f rom our laboratory have demonst ra ted the existence of important differences in the composi t ion of the extracellular matr ix surrounding intrahepatic clusters and intrasplenic-

Fig. 5a-c Distribution of glutamine synthetase in bile duct-ligated livers and in intrasplenically transplanted hepatocytes, a The distribution of glutamine synthetase in bile duct-figated rat livers with moderate bile duct proliferation. Glutamine synthetase is restricted to a narrow ring of hepatocytes (arrow) around a centrilobular vein (CV), at a distance from proliferating bile ductules revealed by an anti-cytokeratin 19 antibody (arrowhead). b The distribution of glutamine synthetase in bile duct-ligated rat livers with massive bile duct proliferation. Glutamine synthetase is faintly detected in a narrow ring of hepatocytes (arrow) around a centrilobular vein (CV), at a distance from proliferating bile ductules revealed by an anti-cytokeratin 19 antibody (arrowhead). c The distribution of glutamine synthetase in intrasplenically transplanted hepatocytes. Glutamine synthetase is detected in some hepatocytes, mainly located at the contact of a large splenic vessel (11). (PT Portal tract, CV centrilobular vein) Original magnifications: a, b xl00, c xl60

328

Fig. 6a, b Histochemical detection of LDH (a) and ND (b) activities in two consecutive sections across a cluster of intrasplenically transplanted hepatocytes observed 6 months after transplantation. High levels of both LDH and ND activities are present in most hepatocytes (arrows). (V Vessel) Original magnifications: a, b x200

Acknowledgements We are particularly grateful to S. Caimal and D. Lebrec (INSERM U24, H6pital Beaujon, Clichy) for their as- sistance in bile duct ligation experiments. We thank R. Delelo, R Mariani, and B. Nordlinger (INSERM U402, CHU Saint-Antoine, Paris) for their help with intrasplenic transplantation of hepatocytes.

ally transplanted hepatocytes. Intrahepatic clusters are lined by typical basement membranes containing laminin, while intrasplenically transplanted hepatocytes are surrounded by an abundant extracellular matrix rich in collagen III and fibronectin but lacking laminin (Racine et al. 1994). Laminin is known to play an important reg- ulatory role in the differentiation and metabolic activity of epithelial cells, including hepatocytes. Our observa- tion therefore underlines the potential functional rele- vance of tissue-specific differences in the pericellular microenvironment of hepatocytes.

In conclusion, our study shows that the coexpression of periportal and perivenous markers in the same hepatocytes and the induction of perivenous markers in periportal hepatocytes is likely to be a non-specific consequence of the loss of the normal connections of hepatocytes with the normal liver microcirculation. It also confirms that several functional characteristics of hepatocytes, such as the expression of glutamine synthetase, are modulated by local, tissue-specific, pericellular factors.

References

Berry MN, Friend DS (1969) High-yield preparation of isolated rat liver parenchymal ceils. A biochemical and fine structural study. J Cell Bio143:506-520

Braakman I, Keij J, Meijer DKF, Groothuis GMM (1991) Separa- tion of periportal and perivenous hepatocytes by fluorescence- activated cell-sorting: confirmation with colloidal gold as an exogenous marker. Hepatology 13:73-82

Chen L, Davis GJ, Crabb DW, Lumeng L (1994) Intrasplenic transplantation of isolated periportal and perivenous hepatocytes as a long-term system for study of liver-specific gene expression. Hepatology 19:989-998

Chiquoine AD (1953) The distribution of glucose-6-phosphatase in the liver and in the kidney of the mouse. J Histochem Cyto- chem 1:429-435

Christa L, Simon MT, Flinois JR Gebhardt R, Br~chot C, Lasserre C (1994) Overexpression of glutamine synthetase in human primary liver cancer. Gastroenterology 106:1312-1320

Dinges HP, Zatloukal K, Denk H, Smotle J, Mair S (1992) Alco- holic liver disease. Parenchyma to stroma relationship in fibro- sis and cirrhosis as revealed by three-dimensional reconstruction and immunohistochemistry. Am J Pathol 141:69-83

Feldmann G, Scoazec JY, Racine L, Bernuau D (1992) Functional hepatocellular heterogeneity for the production of plasma proteins. Enzyme 46:139-154

Gebhardt R (1992) Metabolic zonation of the liver: regulation and implications for liver function. Pharmacol Ther 53:275-354

Gebhardt R, Reichen J (1994) Changes in distribution and activity of glutamine synthetase in carbon tetrachloride-induced cirrhosis in the rat: potential role in hyperammonemia. Hepatolo- gy 20:684-691

Gebhardt R, Jirtle R, Moorman AFM, Lamers WH, Michalopou- los G (1989) Induction of glutamine synthetase and transient co-expression with carbamoylphosphate synthetase in hepatocytes transplanted into fat pads of syngenic hosts. Histo- chemistry 92:337-342

Graham RC Jr, Karnovsky MJ (1966) The early stages of absorp- tion of injected horseradish peroxidase in the proximal tubules of mouse kidney: ultrastructural cytochemistry by a new technique. J Histochem Cytochem 14:291-298

Gumucio JJ (1989) Hepatocyte heterogeneity: the coming of age from the description of a biological curiosity to a partial un- derstanding of its physiological meaning and regulation. Hepa- tology 9:154-160

Jungermann K, Katz N (1989) Functional specialization of different hepatocyte populations. Physiol Rev 69:708-764

Jungermann K, Sasse D (1978) Heterogeneity of liver parenchy- real cells. Trends Biochem Sci 3:198-200

Kountouras J, Billing BH, Scheuer PJ (1984) Prolonged bile duct obstruction: a new experimental model for cirrhosis in the rat. Br J Exp Pathol 65:305-311

Lamers WH, Been W, Charles R, Moorman AFM (1990) Hepa- tocytes explanted in the spleen preferentially express carbamoylphosphate synthetase rather than glutamine synthetase. He- patology 12:701-709

Maganto R Traber PG, Rusnell C, Dobbins WO III, Keren D, Gumucio JJ (1990) Long-term maintenance of the adult pattern of liver-specific expression for P-450b, P-450e, albumin and c~-fetoprotein genes in intrasplenically transplanted hepatocytes. Hepatology 11:585-593

Mito M, Ebata H, Kusano M, Onishi T, Saito T, Sakamoto S (1979) Morphology and function of isolated hepatocytes transplanted into rat spleen. Transplantation 28:499-505

Nordlinger B, Wang SR, Bouma ME, Verthier N, Hillan K, Delelo R, Infante R (1987) Can hepatocytes proliferate when trans-

329

planted into the spleen? Demonstration by autoradiography in the rat. Eur Surg Res 19:381-387

Novikoff AB (1959) Cell heterogeneity within the hepatic lobule of the rat (staining reactions). J Histochem Cytochem 7:240-246

Poo JL, Feldmann G, Erlinger S, Braillon A, Gaudin C, Dumont M, Lebrec D (1992) Ursodeoxycholic acid limits liver histo- logic alterations and portal hypertension induced by bile duct ligation in the rat. Gastroenterology 102:1752-1759

Racine L, Scoazec J-Y, Verthier N, Bemuau D, Feldmann G (1994) Morphological aspects of hepatocytes transplanted into the spleen. Transplantology 5:129-135

Sano K, Fujioka Y, Nagashima K, Hashizume T, Komori M, Kit- ada M, Kamataki T, Miyazaki T (1989) Distributional varia- tion of P-450 immunoreactive hepatocytes in human liver dis- orders. Human Pathol 20:1015-1020

Shank RE, Morrison G, Cheng CH, Kari I, Schwartz R (1959) Cell heterogeneity within the liver lobule (quantitative histochemistry). J Histochem Cytochem 7:237-239

Sokal EM, Trivedi R Cheeseman R Portmann B, Mowat AP (1989) The application of quantitative cytochemistry to study the acinar distribution of enzymatic activities in human liver biopsy sections. J Hepatol 9:42-48

Sokal EM, Trivedi R Portmann B, Mowat AP (1990) Adaptive changes of metabolic zonation during the development of cirrhosis in growing rats. Gastroenterology 99:785-792

Sokal EM, Mostin J, Buts J-P (1992) Liver metabolic zonation in rat biliary cirrhosis: distribution is reverse of that in toxic cirrhosis. Hepatology 15:904-908

Traber PG, Maganto P, Wojcik E, Keren D, Gumucio JJ (1989) In- duction of P-450IIB genes within the rat liver acinns is not dependent on the chemical inducer or on the acinar organization. J Biol Chem 264:10292-10298

Van Noorden CJF, Frederiks WM (1992) Enzyme histochemistry: a laboratory manual of current methods. Oxford University Press, Oxford

Van Noorden CJF, Frederiks WM, Aronson DC, Marx F, Bosch K, Jonges GN, Vogels IMC, James J (1987) Changes in the acinar distribution of some enzymes involved in carbohydrate metab- olism in rat liver parenchyma after experimentally induced cholestasis. Virchows Arch 52:501-511

Coexpression of periportal and perivenous enzymes in rat hepatocytes after experimental bile duct...

Documents

Transcript of Coexpression of periportal and perivenous enzymes in rat hepatocytes after experimental bile duct...