effects of extracellular matrix on the expression of peroxisomes in primary rat hepatocyte cultures

Journal of Hepatology 2000; 32: 381-391 Printed in Denmark . All rights reserved Munksgaard Copenhagen

Copyright 6 European Association for the Study of the Liver 2000

Journal of Hepatology ISSN 0168-82 78

Effects of extracellular matrix on the expression of peroxisomes in primary rat hepatocyte cultures

Marianne Depreter*‘, Tom Tytgat *l, Sonja Beken*, Marc Espeel’, Karen De Smet*, Vera Rogiers2 and

Frank Roels’

‘University of Ghent, Department of Anatomy, Embryology and Histology, Ghent and 2Vrije Universiteit Brussels, Department of Toxicology, Brussels, Belgium

BackgrouncVAims: Peroxisomes in wild-type cells vary between tissues and developmental stages. In the liver of some peroxisomal deficiency disorder patients, rare parenchymal cells express normal peroxisomes (mosaics); the mechanism is unknown. Our aim was to find factors regulating peroxisome expression. Met/zods: Liver-specific as well as peroxisome characteristics were studied in three types of primary rat hepatocyte cultures. Results: Total glutathione S-transferase activity and albumin secretion both increased in the collagen I sandwich and immobilization gel cultures. In contrast, in monolayers cultured on plastic, total glutathione S- transferase activity decreased and albumin secretion was only 3040% compared to the collagen cultures Glycogen rosettes typical of liver parenchymal cells were always abundant. Laminin and collagen IV-producing stellate cells were numerous in the monolayer but almost absent in the sandwich cultures. In 6-day- monolayer cultures, the number of liver-specific peroxisomes had decreased while atypical small or elon-

ROXISOMES are single membrane-limited organelles involved in several anabolic and catabolic pro-

cesses (1) and responsible for a group of inherited metabolic diseases such as the Zellweger syndrome, adrenoleukodystrophy, infantile Refsum disease and rhi- zomelic chondrodysplasia punctata. Severe neurologi-

Received 7 January; revised 5 August; accepted 31 August 1999

Correspondence: Marianne Depreter, University of Ghent, Department of Anatomy, Embryology and Histology, Godshuizenlaan 4, B-9000 Ghent, Belgium. Tel: 329 2649225. Fax: 329 2259452. e-mail: [email protected]

* Both investigators have contributed equally to this study and both should be considered as first author.

gated peroxisomes appeared. Immunolabeling density for catalase and three B-oxidation enzymes was decreased compared to adult rat liver; catalase specific activity in homogenates had dropped to 15% and 4% in the sandwich and monolayer cultures, respectively. In 17-day-sandwich cultures, some peroxisomes showed a very weak catalase reaction; total activity was 5%. Supplementation of the collagen type I cultures with several extracellular matrix factors could not prevent peroxisome dedifferentiation. Conclusion: The presence of these extracellular matrix components is not sufficient for normal peroxisome expression. It is suggested that hepatocyte-spe- chic and peroxisomal features are regulated differently. The sandwich preserves hepatocyte differentiation better than the monolayer.

Key words: /l-oxidation enzymes; Catalase; Collagen IV; D-aminoacid oxidase; Fibronectiq Gene regulation; Glial fibriliary acidic protein; Heparan sulphate proteoglycans; Lamhrin; Polyamine oxidase.

cal impairment is a hallmark of these disorders (2). In dividing hepatocytes, or when induced by peroxisome proliferators, new peroxisomes originate by division of an existing “mother organelle” (for a review, see (3)). Peroxisomal matrix and membrane proteins are syn- thesized in the cytoplasm and are imported post-trans- lationally into peroxisomes. Several proteins (and their encoding genes) involved in peroxisome biogenesis (“peroxins” and pex-genes (4)) were identified and different knock-out models in mice were generated (5-7).

Peroxisomal P-oxidation enzymes can be induced by naturally occurring solutes (thyroid hormone (8), retinoic acid (9), and fatty acids (10) that possess their own receptors (11). Recently, peroxisomal disorder patients with a mosaic distribution of peroxisomes in the

381

M. Depreter et al.

liver parenchyma were described: small groups of parenchymal cells with normal peroxisomes are adjacent to parenchymal cells without peroxisomes (12-15). This small percentage of normal cells does not guaran- tee normal functional capacity of the liver: the bulk of liver bloodflow will miss these normal cells. However, most of the patients have a milder clinical history. The origin of peroxisome mosaicism remains unknown; regulation of peroxisome expression by signals from the extracellular environment was hypothetically considered as one possible explanation (13-l 5). In the liver of one patient, normal cells were grouped near veins

(15). When hepatocytes are cultured on plastic petri-dish-

es, liver-specific gene expression is rapidly lost. For several specific functions such as albumin secretion and transferrin synthesis, a modulating role in monolayer cultures by extracellular matrix proteins was shown (16-19). Dunn et al. (20) succeeded in maintaining differentiated hepatocytes in vitro for several weeks by culturing the cells between two layers of rat tail collagen type I (sandwich cultures). With respect to peroxisomal functions, published data indicate a loss in hepatocyte cultures (21-23). Recently, Ruiz et al. (24) no- ticed that the expression level of the peroxisomal adrenoleukodystrophy protein in cultured human skin fibroblasts varied according to the culture medium.

We have examined the expression of liver-specific peroxisomes in primary rat hepatocyte cultures and the effects of the presence of the extracellular matrix components laminin and collagen IV.

Hepatocytes were cultured either as a monolayer on a plastic petri dish or as a sandwich between rat tail collagen; monolayers were found to contain laminin and collagen IV-producing stellate cells, which are almost absent in sandwich cultures (14). The sandwich configuration better mimics liver structure, positively affects the maintenance of liver-specific functions (20) and sustains cellular polarity (25).

The effect of supplementing collagen I with several ECM (extracellular matrix) factors was studied in sandwich and immobilization cultures. Immobilization cultures, an easy to use sandwich model, were previously shown to be similar to sandwich cultures in maintaining liver-specific functions (26).

Liver peroxisome specificity was evaluated by catalase activity, by cytochemistry for catalase-, polyamine oxidase (PAOX)-, and D-aminoacid oxidase (DAAOX) activities, immunocytochemistry for three peroxisomal D-oxidation enzymes, ultrastructure (presence of nucleoid constituted by urate oxidase) and morphometry. Liver-specific features were evaluated by albumin secretion in the medium, glutathione-S-transferase (GST)

activity and the presence of glycogen rosettes specific for liver parenchymal cells.

Materials and Methods Hepatocyte cultures Hepatocytes (viability >80%) were isolated as previously described (27) from male outbred OFA Sprague Dawley rats (200-250 g, Iffa- Credo, Belgium). In the monolayer cultures (28) hepatocytes were seeded directly on plastic petri-dishes (Falcon, no. 3003). Collagen gel sandwich cultures were prepared according to Dunn et al. (20) and as described by Beken et al. (29).

Immobilization cultures were prepared according to Koebe et al. (30) and Beken et al. (29). Collagen gel was prepared from rat tail tendons as described by Koebe et al. (30). Cells were cultured in serum-free Dulbecco’s Modified Eagle’s Medium, supplemented with 0.5 U/ml bovine insulin, 0.007 @g/ml glucagon, 0.02 &ml EGE 7.5 &ml hydrocortisone sodium succinate and antibiotics (7.3 I.U.lml benzyl penicillin, 50 fig/ml streptomycin sulphate, 50 &ml kanamy- tin monosulphate and 10 &ml sodium ampicillin). 10% FBS was added as a supplement only for the lirst 4 h. Cultures were kept at 37°C in an atmosphere of 5% COz and 95% air with a relative hu- midity of 100%. The medium was changed daily.

In half of the sandwich cultures, 60 &ml L-proline was added to the medium.

In an additional experiment, collagen I of both sandwich and immobilization gel cultures was supplemented with collagen type IV (50 ng/cm2), heparan sulphate proteoglycans (6 ng/cmz), fibronectin (50 ng/cm2) with or without laminin (50 ng/cmz) (Beken et al. (31)). Pro- cedures for isolating and culturing rat hepatocytes were approved by the local ethics committee at the Vrije Universiteit Brussels.

Catalase- and GST-activity, albumin assay Catalase activity was measured according to Baudhuin (32) in hepatocytes harvested directly from the petri-dish (monolayer culture) or after collagenase treatment, which is necessary in the case of sandwich and immobilization cultures (33). Catalase was released from the peroxisomes by sonication. Protein was determined according to Peterson (34). Glutathione-S-transferase activity in cytosol samples, was determined as described previously (26,33), using l-chloro-2,4- dinitrobenzene as a substrate.

The culture medium was analyzed for its albumin content by enzyme-linked immunosorbent assay (ELISA) (26).

Histochemtstry Peroxisomes were visualized by staining for catalase activity with di- aminobenzidine (DAB) (35). DAAOX and PAOX were localized by a cerium method we have modified (36) from Stefanini (37) and Van den Munckhof (38). The cells were then postosmicated in 0~0, with KsFe(CN),. This method visualizes cerium hydroperoxide for light microscopy also.

Extracellular matrix and glial fibrillary acidic protein (GFAP) immunohistochemistry Laminin and collagen IV immunohistochemistry was performed as described (36). The laminin antibody was kindly provided by Prof. J. M. Foidart (36).

GFAP immunohistochemistry was performed on ethanol fixed cells (10 min, at -20°C); after permeabilization with Triton X-100, cells were incubated for 2 h with the monoclonal antibody directed against porcine GFAP, purchased from Sigma (Cat. No. G3893, clone GAS). The secondary antibody was rhodamine-coupled anti-mouse (Dako, Cat. No. R0270).

As a negative control the primary antibody was either omitted (laminin and collagen IV) or substituted with the normal IgG fraction (Dako, Cat. No. X0931) (GFAP).

Immunohistochemistry for peroxisomal proteins was performed as described (39). Acyl-CoA oxidase antibody was kindly provided by Dr. A. Volkl, Heidelberg, Germany, 3-ketoacyl-CoA tbiolase anti-

382

Peroxisomes in cultured hepatocytes

body by Dr. T. Hashimoto (Nagano, Japan), and multifunctional enzyme antibody by Dr. C. Causemt (Dijon, France) (36).

Labelling aknsity The number of gold particles per peroxisome was counted and the organelle area was measured with a semi-automatic image analysis system (Mini-Mop, Kontron Inc., Germany).

Rl?SUltS Comparison of monolayer versus sandwich

/

Glutathione-S-transferase (GST) activity (Table I) and albumin secretion (Table 2). In the sandwich cultures without L-proline, GST activity was 70% and 100% on days 7 and 14, respectively, when compared to the level observed for freshly isolated hepatocytes (FIH). In the sandwich cultures containing L-proline, GST activity on days 7 and 14 was similar to the level of FIH.

In the monolayer culture GST activity decreased significantly; on day 6 it was 60% of the activity found in FIH.

Albumin secretion in the sandwich without L-proline initially was low but showed a 3-fold increase at 7 days. This level was maintained until 14 days. When in sandwich cultures L-proline was added to the medium, an additional increase in albumin secretion was no-

TABLE 1

Glutathione s-transferase activity

Freshly isolated hepatocytes 1.07?0.18*

Monolayer cultures day 1 1.19kO.08 day 6 0.60-t0.04

Sandwich cultures day 7 - L-proline 0.71’0.12

7 + L-proline 1.262 0.24 day 14 - L-proline 1.4150.33

+ L-proline 1.4420.14

Glutathione S-transferase activity, expressed as pmokninlmg cytosol- ic protein (n=4). I-chloro-2,4-dinitrobenzene was used as a general substrate. *: 22 Standard Error of the Mean. In the monolayer, GST- activity decreased during culture, but not in the sandwich.

TABLE 2

Albumin secretion in monolaver and sandwich cultures

day 2 day 4 day 7 day 14

Monolayer (n=5) Sandwich (n=3) Sandwich (n=3) -proline -proline +proline

0.3720.17a 090~0.10~ 1.33kO.73 0.65kO.23 2.01~1.19c 4.72k0.77c 0.77+0.50b 2 74kl lob,* 6.OO-t1.21*

2:46+0:95’ 4.5711.05”

Albumin secretion, expressed as ~gib/1.6.106 cells?2 Standard Error of the Mean. Monolayers do not survive beyond 7 days. The difference in albumin secretion between sandwich cultures with and without proline and between monolayer and sandwich cultures without proline was evaluated (Student’s t-test). ap<O.Ol; bzqd*ep<0.05.

cultum typo

Fig. 1. Catalase specific activity in adult liver (n=4), freshly isolated hepatocytes (FIH) (n=9), monolayer culture on day 2 (M02) (n=5), monolayer culture on day 6 (MO6) (n=S), sandwich culture on day 6 without (SW6-P) (n= 10) and with L-proline (SW6+P) (n=12), sandwich culture on day 17 without (SW17-P) (n=5) and with L- proline (SW17+P) (n=2). Expressed as units Baudhuin per mg total cellular protein 22 SEM. The difference was significant between groups (FIH-liver), (FIH-MO2), (MO2-MO6), (MO6-S W6-P) and (S Wd? P-S WI 7t P) at the ~~0.05 level (Mann-Whitney U-test; groups were considered as independent).

ticed, reaching a near-physiological level on day 7, which as calculated from (40) and (41), corresponds to 8-9 ,ug/h/l.6. lo6 cells.

In contrast, in the monolayer culture, albumin secretion was only 40% to 30% of the secretion by the sandwich culture without L-proline.

Catalase specific activity (Fig. 1) in FIH was significantly higher than in adult liver tissue. In the monolayer cultures on day 2, it was decreased to about one third of the activity in FIH; on day 6 only 4% of the activity remained. However, after 6 days in sandwich culture, catalase specific activity was 3-fold higher than in monolayer cultures. In the sandwich cultures at 17 days, catalase activity had decreased to about 5% of the level in FIH. Addition of L-proline to the sandwich cultures did not have a significant effect on catalase specific activity at 6 or 17 days.

Microscopic features of peroxisomes. Monolayer: On day 2, two cell types were dis-

tinguished in the monolayer culture after staining for catalase activity and for glycogen. Parenchymal cells had a cuboid shape, they showed a strong reactivity with periodic acidSchilI staining; they contained numerous peroxisomes (Fig. 2A). The second type were flattened cells with thin cytoplasmic extensions (“stellate”), they showed no reaction for the periodic acid- Schiff staining and peroxisomes were not detected.

GFAP was strongly expressed in stellate cells and it

383

M. Depreter et al.

Fig. 2 A & C. Monolayer cultures stained for catalase activity: numerous peroxisomes are visualized as dark granules in all parenchymal cells on day 2 (A), while on day 6 (C), the number of peroxisomes is visibly decreased in most cells when compared to Fig. 2A; in addition, note heterogeneity between individual hepatocytes. BrightJield, methyl green counterstain, B. Monolayer culture at day 2 stained for GFAP: a stellate cell with many cytoplasmic extensions growing underneath the parenchymal cell layer and cytoplasmic extensions of another stellate cell. Immunofluorescence, rhodamine. D. Monolayer culture immunostainedfor laminin on day 6. Stellate cells, lying underneath the parenchymal monolayer, are laminin positive; arrows indicate immunoreactive cell extensions. Bright field, hematoxylin counterstain. Scale bar Fig. 2A-D=20 pm.

was clear that they grew underneath the parenchymal cell layer (Fig. 2B). They made up less than 5% of the total population. On day 6, many parenchymal cells had lost the cuboid shape and became elongated. The number of large liver-specific peroxisomes was visibly decreased and showed a distinct heterogeneity between cells (Fig. 2C). All parenchymal cells still had some large peroxisomes, but with a 100X phase contrast im- mersion objective also very small peroxisomes were observed. The number of stellate cells had strongly increased. GFAP staining which was fainter than in cultures of 2 days, showed that they were still growing underneath the parenchymal cell layer and that the open spaces between parenchymal cells were almost completely covered by stellate cells. The latter showed immunoreactivity for laminin (Fig. 2D) and collagen IV at their periphery; this reaction was absent in the parenchymal cells. Exceptionally, parenchymal cells

were growing as three-dimensional aggregates, which contained numerous peroxisomes showing strong catalase activity.

Electron microscopy after staining for catalase activity revealed that in addition to large liver-specific peroxisomes, numerous unusually small round peroxisomes (Fig. 3A) (profile diameters as low as 0.13 ,um) were present, both displaying a typical core (urate oxidase). Some peroxisomes had an elongated shape; a core could be discerned in the PAOX stained ones. Gly- cogen rosettes, specific for hepatic parenchyma were present in many cells. DAAOX activity was localized in the rounded peroxisomes containing a nucleoid, but not in the elongated and atypically shaped forms identified by catalase and PAOX staining. As in adult liver, staining intensity for DAAOX and PAOX differed between peroxisomes in the same cell. As a rule, both cerium methods better visualized the unstained

384


Fig. 3. Peroxisomes from different hepatocyte cultures stainedfor catalase activity (A, B, C), and DAAOX activity (D). Compare (A) small peroxisomes in monolayer after 6 days showing strong catalase activity, to (B) large liver-specific peroxisomes with an unstained nucleoid and strong catalase activity in sandwich after 6 days; a nucleoid is not visible in A. C. Largeperoxisomes with minimal DAB reaction as present in few parenchymal cells in sandwich after I7 days; the central electron dense area is the nucleoid which may be degrading. Other cells contained large peroxisomes clearly showing catalase activity. D. DAAOXmatrix activity differing between individual peroxisomes; sandwich, 17d. The striated nucleoid is clearly seen because unreactive

(arrowhead). A4: mitochondrion has dilated cristae as seen during stimulation of oxidative phosphorylation (42). Glycogen is abundant, rosettes as well as monoparticulate. Note different magnt$cations of A-D. Scale bar (A-D) =0.5 pm.

nucleoid than did staining for catalase activity In the DAB-stained small round peroxisomes, and in the elongated organelles, the presence of a nucleoid was at first overlooked.

In these small round peroxisomes the nucleoid visibly occupied a larger fraction of the matrix than in normal organelles.

Sandwich: On day 6, parenchymal cells had a cuboid

morphology, and dilated primitive bile canaliculi between adjacent cells were frequent. Stellate cells as revealed by GFAP staining, represented at most 5% of the population. All parenchymal cells contained DAB- stained peroxisomes, but the number of peroxisomes varied greatly between adjacent cells. It was always lower than in adult liver. This was confirmed by electron microscopy

385

M. Depreter et al.

TABLE 3

Relative labeling densities for catalase and peroxisomal /?-oxidation enzymes

Catalase acylCoA Multifunct. 3-ketoacylCoA oxidase enzyme thiolase

MO6 -proline 52.0a 50.8’ 29.9k 63.5’= SAN6 -proline 45.lb 52.4s 28.1’ 39.79 SAN6 +proline 41.5c 46.1h 53.5m 51.0’ SAN17 -proline 35.5d 58.3’ N.D. 86.5s SAN17 +prolme 31.7e 23.4 32.3” 66.9’

Relative labeling density of catalase and peroxisomal &oxidation enzymes expressed as percentage of the value in adult liver. (M06) monolayer cultures at day 6, (SAN6, SAN17) sandwich cultures at day 6 and at day 17. N.D.: not determined. ANOVA tests with least significant difference post hoc tests, after controlling the assumptions for ANOVA were done on absolute labeling densities (Statistica Soft- ware Package). p-values: “~~~~~~~ss~~p<O.OO1; “Jr**‘p<O.Ol; abJ’? p<o.o5; hp=O.19; d”p=O.13; s’p=o.36; rsp=oJ31; @Jp=o.30; ‘dp=o.77; q*p=O.O6. Number of peroxisomes measured in M06, SAN6-P, SAN6+P SAN17-P, SAN17+P, respectively: for catalase: 31, 18, 35, 23, 38; AcylCoA oxidase: 31, 26, 41, 29, 28; multifunctional enzyme: 35, 47, 88, -, 8; 3-ketoacylCoA thiolase: 31,25, 23, 12, 39.

At the ultrastructural level peroxisomes were large and contained a nucleoid (Fig. 3B). The small type of peroxisomes that characterized monolayer cultures was infrequent. Elongated peroxisomes were present in part of the hepatocytes. Staining intensity for DAAOX and PAOX activity was different from one peroxisome to another as is the case in adult liver. Glycogen rosettes were present in all cells.

On day 17, no change was obvious in peroxisome number per cell when compared to the results after 6 days. Most peroxisomes were large, contained a nucleoid, and showed a distinct DAB reaction. In some cells, however, peroxisomes showed a very weak catalase activity (Fig. 3C). Glycogen rosettes were still present in all cells. DAAOX- (Fig. 3D) and PAOX activity persisted.

Immunolocalization and labeling density of peroxiso- ma1 proteins (Fig. 4, Table 3). All three p-oxidation enzyme proteins were present in the matrix of the small as well as the large rounded peroxisomes (Fig. 4); cata-

Fig. 4 . Immunolabeling for peroxisomal matrix proteins: A. catalase in monolayer at 6 days; B. 3-ketoacylCoA sandn. lich at 6 days: C. multifunctional protein in sandwich at I7 days. Striations of nucleoid are seen in B and C (A-C ) =0..5 ,um.

thiolaj re in 1 Scale bar

386

lase and multifunctional enzyme were colocalized in the same organelle, when double-labelled with 15 nm

and 10 nm gold particles. On day 6, mean labeling density for catalase in the monolayer and sandwich cultures was about 50% of adult liver. On day 17 in sandwich culture, the mean labeling density had decreased

Moan F%?roxisomol Aroo 0.25

0.2

0.15

s 0.1

0.05

0 1 l- T

t

+ + L LNm km SW-P SW&P SWlFP SW17+P

A

25

20

I

15

10

5

0

culture type

q MOS

8sw-P

4

Fig. 5A. Mean peroxisomal area in adult rat liver (LIVER) (number of peroxisomes, n=201), monolayer cultures on day 6 (Mod) (n=l28), sandwich cultures on day 6 without (SW6-P) (n=115) or with L-proline (SW6+P) (n= 187) and on day I7 without (SW1 7-P) (n=64) or with L- proline (SWl7tP) (n=113). Expressed in ,am2?2 SEM. The mean peroxisomal area was signt&antly different between all groups (~~0.05) except for SW6+PISWl7+P and adult -PIS WI 7-P; ANOVA (Statistica Soft Ware Package) with least signt&?cant difference post hoc tests (after controlling the assumptions for ANOVA). B. Fre- quency histogram of peroxisomal area (pm2) in the monolayer cultures at 6 days (M06) (number of peroxisomes, n=128) versus sandwich cultures at 6 days without L-proline (SW6-P) (n=IIZ). Frequency expressed as the percentage of the total number of peroxisomes measured in the respective groups.


to about 35%. Labeling density for the B-oxidation enzymes was. also lower than in adult liver peroxisomes. Thiolase labeling density remained highest, and multifunctional enzyme showed the strongest decline compared to adult liver (Table 3). Labeling of distinct enzymes showed some differences between sandwich and monolayer, and between sandwich cultures with and without proline (Table 3), but a general trend was not obvious.

Morphometrv. On day 6 in the monolayer cultures, the mean peroxisomal area of DAB-stained peroxisomes (0.07 pm’) was smaller than that in adult liver (0.12 pm2) and in the g-day-sandwich cultures without L-proline (0.21 ,um’) (~~0.05) (Fig. 5A). The monolayer data were highly skewed to the higher area values with the highest frequency (about 15%) in the smallest size class (Fig. 5B).

Fig. 6. Localization of catalase activity in supplemented collagen immobilization gel culture: rat collagen supplemented with fibronectin, heparan sulphate proteoglycan, collagen type IV and laminin. An atypical, very elongated peroxisome, transverse diameter 0.12 pm, adjacent to a normal-sized peroxisome; small granules are ribosomes either free or on endoplasmic reticulum. Scale bar=0.5 pm.

387

M. Depreter et al.

Comparison of simple collagen I gel cultures versus ECM-supplemented cultures Glutathione-S-transferase activity and albumin secretion. Glutathione-S-transferase and albumin secretion followed the same trend as in sandwich cultures when hepatocytes were cultured in immobilization gels (26). Supplementation with other ECM factors of both had no effect on glutathione-S-transferase activity and albumin secretion after 7 days (Beken et al. (31)).

Localization of catalase activity and catalase immunolabeling. ECM supplementation of the collagen I gel did not prevent dedifferentiation of peroxisomes in the sandwich nor in the immobilization cultures. Elon- gated and small type peroxisomes, atypical in normal liver, which in a very small percentage are present in simple sandwich and immobilization cultures, were more frequently observed in hepatocytes cultured in the supplemented collagen I matrices (Fig. 6). A few peroxisomes with only weak catalase activity staining were observed already after 7 days.

Catalase labeling density had decreased in the supplemented matrix cultures as well, when compared to adult liver.

Discussion In human inherited peroxisomal disorders, alterations of peroxisomes are explained by gene mutations. The mechanism underlying the peroxisome mosaics found in the liver of some peroxisomal disorder patients is unknown. In normal wild-type cells, peroxisomal enzyme expression, size and number of peroxisomes vary between tissues, developmental stages and in cell cultures (14,15,35,36,43,44). The main purpose of the present study was the search for extracellular factors regulating peroxisome expression.

Specific catalase activity in primary hepatocyte cultures Catalase specific activity showed a strong decline in both the monolayer and the sandwich cultures when compared to freshly isolated hepatocytes. In the latter, catalase activity is higher than in total homogenate of adult liver. This is explained by the purification of parenchymal cells. A decrease in the number of liver-specific large-sized peroxisomes with a nucleoid was evident in sandwich as well as monolayer cultures at 6 days, compared to the monolayers of 2 days and adult liver. This was confirmed by electron microscopy. Catalase labeling density on day 6 was similar in both culture types and significantly lower than in adult liver peroxisomes. Both these factors could explain the decreased total catalase activity.

The larger peroxisome size in our sandwich culture

at 6 days combined with a catalase labeling density similar to the monolayer, as demonstrated by morphometry, agrees with higher catalase activity found in the sandwich. Catalase activities in the monolayer culture correspond to values expected in the absence of catalase synthesis and reflecting spontaneous turnover. In- deed, in adult rat liver, catalase has a half-life of 1.34 days (45); if this also holds for primary cultures, only 4.5% activity must be preserved after 6 days. This logic implies that in sandwich cultures either some catalase synthesis is taking place, or that the turnover is sloiver than in liver.

Catalase and peroxisomal P-oxidation activities in hepatocyte cultures reported by others also were strongly decreased: catalase to 4% (21), and according to our own calculations, B-oxidation capacity to 3.7-9.0% (based on published data (22,23,46), assuming that wet liver con- tains 12.5% of protein and taking into account a 1.6-fold enrichment of peroxisomal activity at isolation of parenchymal cells) compared to that of FIH.

Ultrastructure of peroxisomes in primary cultures - link with development and disease The elongated peroxisomes (mostly containing a core but negative for DAAOX activity), reflect peroxisomal dedifferentiation; elongated peroxisomes are abnormal in rat liver but are seen in cultured hepatoma cells (47), in duodenal epithelium (43), human kidney (48), and quail oocytes (49) as well as in the liver of rare peroxisomal disorder patients (50).

The very small peroxisomes (diameter 0.13 pm) that are observed most frequently in the monolayer cultures, resemble the immature peroxisomes in early stages of fetal rat liver (36). They differ, however, by the frequent presence of a prominent core, of three p- oxidation enzymes, and of DAAOX activity This image of small rounded peroxisomes with a huge core suggests to us that they might result from shrinking due to a quantitative loss of matrix components. The possibility that these small peroxisomes might repre- sent sections through elongated peroxisomes is not supported by the frequent observation of a core.

In the sandwich cultures, significantly larger peroxisomes were found, and this should be compared to the over-sized peroxisomes found in the liver of patients with many peroxisomal disorders lacking one or more enzymes (51), and in the liver of knock-out mice defec- tive in peroxisomal membrane protein PMP 70 (7). If the final peroxisomal size results from the balance between growth of the organelle and a need to divide, data indicate that this equilibrium is disturbed in the sandwich. Recently, the gene PEX 11 was found to be related to such equilibrium (52).

388


Influence of culture type; influence of extracellular matrix

Maher et al. (53) described desmin-positive stellate cells in primary hepatocyte cultures and demonstrated that these are co-isolated Ito cells; they form the main source of collagen and laminin. In our monolayer cultures at 6 days, stellate cells were abundant and they showed a distinct immunopositivity for laminin and collagen IV GFAP staining was faint in comparison to 2 days, as described by Buniatian (54). However, it was clear that the percentage of stellate cells had strongly increased from 2 to 6 days of culture. Despite the avail- ability of laminin and collagen IV in the micro-environment of the hepatocytes, their peroxisome population showed an important number (Fig. 5B) of both atypical small and elongated peroxisomes. Catalase activity showed a dramatic decline. Concomitant with these changes in the peroxisome population, the hepatocyte-specific functions of monolayers showed an obvious regression.

By culturing hepatocytes between two layers of rat tail collagen Dunn et al. (20) succeeded in maintaining the cells differentiated for several weeks in vitro.

In our sandwich model, the higher catalase activity after 6 days and the presence of large peroxisomes containing a nucleoid even after 17 days constitute a better preservation of the differentiated condition compared to monolayers. However, specific catalase activity and immunolabeling of 4 peroxisomal proteins had strongly decreased compared to FIH and adult rat liver and a small percentage of small and elongated peroxisomes were observed in the sandwich cultures as well.

In contrast to peroxisome characteristics (catalase and DAAOX activity, labeling densities of catalase and multifunctional enzyme, atypical shapes), GST-activity and albumin secretion in our sandwich cultures almost recovered to physiological levels (better in the cultures with than without L-proline) and were much higher than in the monolayers. On the other hand, addition of L-proline to sandwich cultures had no consistent effect on peroxisome characteristics. Despite peroxisome dedifferentiation, glycogen rosettes were (always) preserved in parenchymal cells.

Taken together, the data indicate that peroxisomal features are regulated differently from albumin secretion, GST activity and the storage type of glycogen.

From the experiment with the supplementation of the sandwich and immobilization cultures with collagen type IV, fibronectin, heparan sulphate proteoglycans and laminin, it was clear that these components did not prevent the dedifferentiation of peroxisomes, which was even more pronounced. GST activity

or albumin secretion remained unchanged, suggesting again different ways of gene regulation.

Conclusions The presence of extracellular matrix components laminin, collagen I\! fibronectin or heparan sulphate proteoglycans is not sufficient to preserve normal peroxisome expression. This is comparable to development of early liver where immature peroxisomes coexist with laminin and collagen (36). Hepatocyte-specific and peroxisomal features are regulated in different ways; The sandwich culture better preserves hepatocyte differentiation with respect to albumin secretion, GST activity, preponderance of large peroxisomes and higher catalase activity Whether the beneficial effect of the sandwich culture

(large peroxisomes with nucleoid, higher catalase activity compared to monolayer) results from the contact at two sides of parenchymal cells with collagen I or, indirectly, from the influence of this culture type on cell polarity (25) cannot be concluded from our study. Two arguments favor the latter hypothesis: a) hepatocytes cultured on a single layer of collagen I had the same catalase activity at 6 days as the monolayer cultures on plastic; b) when hepatocytes in the monolayer culture exceptionally grew as three-dimensional aggregates, peroxisomes showed a strong catalase activity and their number was much higher than in adjacent hepatocytes growing as flattened cells.

Acknowledgements We thank Betty De Prest, Simonne Van Hulle, Guido De Pestel, Guido Van Limbergen, No&l Verweire and Dominique Jacobus for excellent technical assistance. We thank Prof. Magda Vincx for help with the statisti- cal analysis.

The study was supported by the University of Gent (Verkennend Europees Onderzoek 0 11 V0495; Bijzond- er Onderzoeksfonds 01106797 and 01 lB6197) and the European Biomed Concerted Action “Peroxisomal leukodystrophy” (BMH4 CT961621; DGXII).

References 1. Van Den Bosch H, Schutgens RBH, Wanders RJA, Tager JM.

Biochemistry of peroxisomes. Ann Rev Biochem 1992; 61: 157- 97.

2. Lazarow PB, Moser HW. Disorders in peroxisome biogenesis. In: Striver CR, Beaudet AL, Sly WS, Valle D, editors. The Metabolic Basis of Inherited Disease, 6th ed. New York: McGraw-Hill; 1995. p. 1479-509.

3. Roels E Peroxisomes: a personal account, 1st ed. Brussels: WB- Press; 1991.

4. Distel B, Erdmann R, Gould SJ, Blobel G, Crane DI, Cregg JM, et al. A unified nomenclature for peroxisome biogenesis factors. J Cell Biol 1996: 135: l-3.

389

M. Depreter et al.

5. Baes M, Gressens E Baumgart E, Carmeliet E Casteels M, Fransen M, et al. A mouse model for Zellweger syndrome. Nat Genet 1997; 17: 49-57.

6. Forss-Petter S, Werner H, Berger J, Lassmann H, Molzer B, Schwab MH, et al. Targeted inactivation of the X-linked adrenoleukodystrophy gene in mice. J Neurosci Res 1997; SO: 82-3.

7. Espeel M, Jimenez-Sanchez G, Moser A, Roels F, Valle D. Liver peroxisomes in knock-out mice lacking the 70 kDa peroxisomal membrane protein, are enlarged and contain PT.%I and PTS-II targetted proteins. J Inher Metab Dis 1998; 21 Suppl 2: 99.

8. Just Ww, Hart1 F-U. Rat liver peroxisomes, II. Stimulation of peroxisomal fatty-acid /J-oxidation by thyroid hormones. Hoppe Seyler’s Z Physion Chem 1983; 364: 1541-7.

9. Hertz R, Bar-Tana J Induction of peroxisomal b-oxidation genes by retinoic acid in cultured rat hepatocytes. Biochem J 1992; 281: 41-3.

10. Dreyer C, Keller H, Mahfoudi A, Laudet V Krey G, Wahli W. Positive regulation of the peroxisomal b-oxidation pathway by fatty acids through activation of peroxisome proliferator acti- vated receptors (PPAR) Biol Cell 1993; 77: 67-76.

11. Mangelsdorf DJ, Evans RM. The RXR heterodimers and orphan receptors. Cell 1995; 83: 841-50.

12. Giros M, Roels F, Prats J, Ruiz M, Ribes A, Espeel M, et al. Long survival in a case of peroxisomal biogenesis disorder with peroxisome mosaicism in the liver. Ann NY Acad Sci 1996; 804: 747-9.

13. Espeel M, Mandel H, Poggi F, Smeitink JAM, Wanders RJA, Kerckaert I, et al. Peroxisome mosaicism in the livers of peroxisomal deficiency patients. Hepatology 1995; 22: 497-504.

14. Roels F, Tytgat T, Beken S, Giros M, Espeel M, De Prest B, et al. Peroxisome mosaics in the liver of patients and the regulation of peroxisome expression in rat hepatocyte cultures. Ann NY Acad Sci 1996; 804: 502-15.

15. Roels F, Espeel M, Mandel H, Poggi F, Smeitink JAM, Toorman J, et al. Cell and tissue heterogeneity in peroxisomal patients. In: Wanders RJA, Schutgens RBH, Tabak HF, editors. Functions and Biogenesis of Peroxisomes in Relation to Human Disease. Amsterdam: Royal Acad of Arts and Sciences, North-Holland; 1995. p. 271-94.

16. Bissell DM, Arenson DM, Maher JJ, Roll FJ. Support of cultured hepatocytes by a laminin-rich gel. Evidence for a functional significant subendothelial matrix in normal rat liver. J Clin Invest 1987; 79: 801-12.

17. Caron JM. Induction of albumin gene transcription in hepatocytes by extracellular matrix proteins. Mol Cell Biol 1986; 10: 123943.

18. Ben-Ze’ev A, Robinson GS, Bucher NLR, Farmer SR. Cell-cell and cell-matrix interactions differentially regulate the expression of hepatic and cytoskeletal genes in primary cultures of rat hepatocytes. Proc Nat1 Acad Sci USA 1988; 85: 2161-5.

19. DiPersio CM, Jackson DA, Zaret KS. The extracellular matrix coordinately modulates liver transcription factors and hepatocyte morphology. Mol Cell Biol 1991; 11: 4405-14.

20. Dunn JCY, Yarmush ML, Koebe H-G, Tompkins RG. Hepato- cyte function and extracellular matrix geometry: long-term culture in a sandwich configuration. FASEB J 1989; 3: 1747.

21. Sun Y, Colburn NH, Oberley LW. Depression of catalase gene expression after immortalization and transformation of mouse liver cells. Carcinogenesis 1993; 14: 1505-10.

22. Mitchell AM, Bridges Jw, Elcombe CR. Factors influencing peroxisome proliferation in cultured rat hepatocytes. Arch Toxic01 1984; 55: 23946.

23. Blaauboer BJ, van Holsteijn CW, Bleumink R, Mennes WC, van Pelt FN, Yap SH, et al. The effect of beclobric acid and clofibric acid on peroxisomal beta-oxidation and peroxisome proliferation in primary cultures of rat, monkey and human hepatocytes. Bio- them Pharmacol 1990; 40: 521-8.

24. Ruiz M, Co11 MJ, Pampols T, Giros M. ALDP expression in fetal cells and its application in prenatal diagnosis of X-linked adrenoleukodystrophy. Prenatal Diagn 1997; 17: 6516.

25. LeCluyse EL, Andus KL, Ho&man JH. Formation of extensive canalicular networks by rat hepatocytes cultured in collagen- sandwich contiguration. Am J Physiol 1994; 266: Cl76474.

26. Beken S, Tytgat T, Pahernik S, Koebe H-G, Vercruysse A, Rogi- ers V Cell morphology, albumin secretion and glutathione S- transferase expression in collagen gel sandwich and immobilization cultures of rat hepatocytes. Toxic in Vitro 1997; 11: 405-16.

27. De Smet K, Beken S, Vanhaecke T, Pauwels M, Vercruysse A, Rogiers V. Isolation of rat hepatocytes. In: Phillips IR, Shephard EA, editors. Cytochrome P450 Protocols. In series: Methods in Molecular Biology, Vol. 107. Totowa: Humana Press; 1998. p. 295-301.

28. Vanhaecke T, De Smet K, Beken S, Pauwels M, Vercruysse A, Rogiers V. Primary cultures and co-cultures of rat hepatocytes In Phillips E, Shephard E., editors. Cytochrome P450 Protocols. In series Methods in Molecular Biology, Vol. 107. Totowa: Hu- mana Press; 1998. p. 311-7.

29. Beken S, Vanhaecke T, De Smet K, Pauwels M, Vercruysse A, Rogiers V Collagen-gel sandwich and immobilization cultures. In: Phillips IR, Shephard EA, editors. Cytochrome P450 Proto- cols. In series Methods in Molecular Biology, Vol. 107. Totowa: Humana Press; 1998. p. 303-9

30. Koebe H-G, Pahernik S, Eyer P Schildberg FW. Collagen gel immobilization: a useful cell culture technique for long-term metabolic studies on human hepatocytes. Xenobiotica 1994; 24: 95-107.

31. Beken S, Slaus K, De Smet K, Depreter M, Roels F, Vercruysse A, et al. Effect of extracellular matrix composition on the expression of glutathione S-transferase isoenzymes in organotypical hepatocyte cultures. Toxic in Vitro 1999; 13: 571-77.

32. Baudhuin I? Isolation of rat liver peroxisomes. Methods Enzymol 1974; 31: 358-68.

33. Beken S, Pauwels M, Pahernik S, Koebe H-G, Vercruysse A, Ro- giers V Collagen gel sandwich and immobilization cultures of rat hepatocytes: problems encountered in expressing glutathione S- transferase activities. Toxic in Vitro 1997; 11: 741-52.

34. Peterson GL. Determination of total protein. Methods Enzymol 1983; 91: 95-105.

35. Roels E De Prest B, De Pestel G. Liver and chorion cytochemistry. J Inher Metab Dis 1995; 18 Suppl.1: 155-71.

36. Depreter M, Nardacci R, Tytgat T, Espeel M, Stefanini S, Roels E Maturation of the liver-specific peroxisome versus laminin, collagen IV and integrin expression. Biol Cell 1998; 90: 641-52.

37. Stefanini S. Differentiation of liver peroxisomes in the foetal and newborn rat. Cytochemistry of catalase and D-Amino acid oxidase. J Embryo1 Exp Morph01 1985; 88: 15163.

38. Van den Munckhof RJM. In situ heterogeneity of peroxisomal oxidase activities: an update. Histochem J 1996; 28: 401-29.

39. Espeel M, Van Limbergen G. Immunocytochemical localization of peroxisomal proteins in human liver and kidney. J Inher Metab Dis 1995; 18 Suppl 1: 135-54.

40. Kimball SR, Horetsky RL, Jefferson LS. Hormonal regulation of albumin gene expression in primary cultures of rat hepatocytes. Am J Physiol 1995; 268: E6-14.

41. Hsu C-J, Kimball SR, Antonetti DA, Jefferson LS. Effects of insulin on total RNA, poly(A)+ RNA, and mRNA in primary cultures of rat hepatocytes. Am J Physiol 1993; 263: E1106-12.

42. Hackenbrock CR, Rehn TG, Weinbach EC, Lemasters J. Oxida- tive phosphorylation and ultrastructural transformation in mito- chondria in the intact ascites tumor cell. J Cell Biol 1971; 51: 123-37.

43. Roels F, Espeel M, Pauwels M, De Craemer D, Egberts HJA, van der Spek I? Different types of peroxisomes in human duodenal epithelium. Gut 1991; 32: 858-65.

44. Espeel M, Briere N, De Craemer D, Jauniaux E, Roels I? Cata- lase-negative peroxisomes in human embryonic liver. Cell Tissue Res 1993; 272: 89-92.

45. Price VE, Sterling WR, Tarantola VA, Hartley RW Jr, Rechcigl M Jr. The kinetics of catalase synthesis and destruction in vivo. J Biol Chem 1962; 237: 3468-75.

390


46. Lazarow PB. Three hypolipidemic drugs increase hepatic pahni- tyl-coenzyme A oxidation in the rat. Science 1977; 197: 580-l.

47. Schrader M, Baumgart E, Valk A, Fahimi HD. Heterogeneity of peroxisomes in human hepatoblastoma cell line HepG2. Evidence of distinct subpopulations. Eur J Cell Biol1994; 64: 281-94.

48. Roels E Goldfischer S. Cytochemistry of human catalase: the demonstration of hepatic and renal peroxisomes by a high tem- perature procedure. J Histochem Cytochem 1979; 27: 1471-7.

49. Espeel M, Depreter M, Nardacci R, D’Herde K, Kerckaert I, Stefanini S, et al. Biogenesis of peroxisomes in fetal liver. Microsc Res Techn 1997; 39: 453-66.

50. Roels E Pauwels M, Poll-The BT, Scotto J, Ogier H, Aubourg E et al. Hepatic peroxisomes in adrenoleukodystrophy and related syndromes. Cytochemical and morphometric data. Virchows Arch A Path01 Anat 1988; 413: 275-85.

51. Roels F, Espeel M, Poggi E Mandel H, Van Maldergem L, Sau- dubray JM. Human liver pathology in peroxisomal disease: a review including novel data. Biochimie 1993; 75: 281-92.

52. Kunau WI-I, Erdmann R. Peroxisome biogenesis: back to the en- doplasmatic reticulum? Curr Biol 1998; 8: R299-302.

53. Maher JJ, Bissell DM, Friedman SL, Roll FJ. Collagen measured in primary cultures of normal rat hepatocytes derives from lipo- cytes within the monolayer. J Clin Invest 1988; 82: 450-9.

54. Buniatian B, Hamprecht B, Gebhardt R. Glial fibrillary acidic protein as a marker of perisinusoidal stellate cells that can distin- guish between the normal and myofibroblast-like phenotypes. Biol Cell 1996; 87: 65-73.

391

effects of extracellular matrix on the expression of peroxisomes in primary rat hepatocyte cultures

Documents