Modulation of fetal and neonatal rat hepatocyte functional activity by glucocorticoids in co-culture

Cell Differentiation, 16 (1985) 259-268 Elsevier ScientificPublishers Ireland, Ltd.

259

CDF 00308

Modulation of fetal and neonatal rat hepatocyte functional activity by glucocorticoids in co-culture G. Lescoat, N. Theze, B. Clement, A. G u i l l o u z o a n d C. G u g u e n - G u i l l o u z o Unitb de Recherches Hbpatologiques U 49 de I'INSERM, Hbpital de Pontchaillou, 35011 Rennes Cedex. France

(Accepted 2 February 1985)

Fetal and neonatal rat hepatocytes were cultured alone or in association with another liver epithelial cell type, in a medium with or without hydrocortisone. Secretion of albumin and alpha-fetoprotein decreased in pure hepatocyte culture, whereas in co-culture it remained stable for several days. Furthermore, addition of hydrocortisone to the co-culture medium induced a rapid increase in albumin production which was maintained at a high level. In contrast, alpha-fetoprotein production was inhibited. At the same time, an abundant extracellular material was secreted between and around hepatocyte colonies. The results demonstrate that the reciprocal relation between albumin and alpha-fetoprotein production which occurs during in vivo perinatal hepatocyte maturation is also observed in vitro. Both cell-cell contacts and glucocorticoids play a key role in this process. It appears that fetal and neonatal hepatocytes can maturate when maintained in a co-culture system. protein secretion; perinatal rat hepatocyte; co-culture; glucocorticoids

Introduction

During in vivo hepatocyte maturation, synthesis of alpha-fetoprotein (AFP) gradually decreases whereas that of albumin increases (Abelev, 1971). Moreover, glucocorticoids have been demonstrated to influence AFP production. Pharmacological doses of dexamethasone injected into newborn rats cause an early decrease in serum AFP without any change of albumin level (B61anger et al., 1981). This reciprocal relationship between albumin and AFP secretion represents a useful model for studying the regulation of gene expression during liver maturation and for investigating glucocorticoid control of this process. Until now, there was no suitable in vitro system allowing prolonged cell survival of hepatocytes

with maintenance of their specific functions. Adult hepatocytes, as well as fetal hepatocytes from different developmental stages, lose their functional stability after a few days of culture (Bissell et al., 1973; Guguen-Guillouzo et al., 1980). This could explain why nobody has succeeded in reproducing in vitro the balance between AFP and albumin synthesis in fetal hepatocytes undergoing maturation. In recent studies (Baffet et al., 1982; GuguenGuillouzo et al., 1983), we have shown that hepatocytes from adult rats remained capable of producing high levels of specific proteins for several weeks when co-cultured with another liver cell type (RLEC). These cells currently named 'clear epithelial cells' represent a minor cell population after trypsin digestion of the liver. Under certain

0045-6039/85/$03.30 @1985 ElsevierScientificPublishers Ireland, Ltd.

260 conditions, particularly nutrient medium composition, they divide, form colonies and may be subcultivated (Williams and Gunn, 1974). These RLEC are thought to derive from endothelium (Sirica and Pitot, 1979) or more probably from primitive biliary ductular cells (Grisham, 1980). It may be questioned whether RLEC can also modulate functional activities of fetal and newborn rat hepatocytes; in this co-culture system the influence of hydrocortisone on hepatocyte maturation was studied. In the present work, we show that fetal and neonatal hepatocytes interacted with RLEC, since cell survival was greatly improved and secretion of albumin and AFP was stabilized. Furthermore, we demonstrate that, by adding hydrocortisone to the co-culture medium, albumin synthesis was greatly increased whereas AFP secretion drastically declined, as it does in vivo. In addition, an abundant extracellular material was rapidly produced, paralleling the amounts of secreted albumin.

Materials and Methods

Cell culture Fetal and neonatal hepatocytes were isolated from 19-day-old fetal and 1-day-old newborn Sprague-Dawley rat livers following a procedure previously described (Guguen-Guillouzo et al., 1980). The livers were excised, pooled and washed three times with Hepes buffer (N-2-hydroxyethylpiperazine N'-2-ethanesulfonic acid), p H 7.6. The tissue-dissociating solution was the same Hepesbuffered solution containing 0.025% collagenase (Worthington CLS) and 5 mM CaC12. Liver pieces were stirred gently at 37°C in 30 ml of dissociating medium for 10 min; two further stirrings, each lasting 10 min, were carried out. The supernatants were collected and centrifuged at low speed (50 × g) for 15 s. The hepatocytes were washed twice with Hepes buffer and once with the selective culture medium. The liver epithelial cell line used in this study was derived from 10 day-old Sprague-Dawley rat livers according to the method of Williams et al. (1971). These cells did not express detectable

hepatocytic functions and were used before they underwent transformation (Morel-Chany et al., 1978). Hepatocytes isolated from one liver were suspended in 3 ml of arginine-free H a m F 12 medium containing 200 # g / m l of bovine albumin, 10 /~g/ml of insulin and 10% fetal calf serum and plated in a 25 cm 2 polystyrene flask (Leffert and Paul, 1972). After 3 h, the medium was renewed. Co-cultures were obtained by adding 2.5 × 106 R L E C in 3 ml of medium per flask. In these conditions, cell confluency was obtained within 24 h. Every day thereafter, the medium either supplemented or not supplemented with hydrocortisone hemisuccinate (Roussel France) was renewed. The hormone concentration used was 7 x 10 6 M. In some experiments 1.4× 10 6 and 7 x 10 7 M hormone concentrations were also tested. Morphological studies Cell morphology. For electron microscopic studies, hepatocyte co-cultures were fixed in situ in 2.5% glutaraldehyde buffered with 0.1 M sodium cacodylate for 5 min, postfixed in 1% osmium tetroxide, dehydrated and embedded in Epon. DNA synthesis. D N A synthesis in pure and mixed cultures was studied by [methyl-3H]thymi dine incorporation (specific activity 25 C i / m M , Commissariat ~ l'Energie Atomique, Saclay, France). The cells were cultured in the presence of 1 ktCi/ml of isotope for 20 h. The samples were washed, fixed in situ in 2.5% glutaraldehyde buffered with 0.1 M sodium cacodylate for 5 min, and then processed for radioautography by using Eastman Kodak emulsion. Reticulin staining. After washing, the cells were fixed in a mixture of 4% p a r a f o r m a l d e h y d e / 0.2% glutaraldehyde buffered with 0.1 M sodium cacodylate for 15 rain at 4°C, and reticulin fibers were visualized using the silver impregnation method (Gordon and Sweets, 1936). Albumin localization. Intracellular albumin was demonstrated by immunoperoxidase according to a technique previously described (Guillouzo et al., 1982). Briefly, the cells were rinsed and fixed in a mixture of 4% paraformaldehyde/0.2% glutaraldehyde buffered with 0.1 M sodium

261 cacodylate for 3 min at 4°C, permeabilized with saponin and then incubated with peroxidaselabelled anti-albumin antibodies. Albumin and A F P assays

Culture media were collected daily and stored at - 2 0 ° C . Albumin and AFP were quantified by laser immunophelometry (Ritchie, 1975). Standard rat albumin and A F P solutions and incubation media were mixed with appropriate dilutions of anti-rat albumin and AFP antisera in 0.1 M phosphate buffer containing 2% polyethylene glycol. Light scatter was measured after 1 h incubation at room temperature. Standard rat proteins and antibodies prepared for immunophelometry use were from ATAB, Scarborough, ME, U.S.A. Each product is designated by a code-name which enables identification of its main characteristics, such as host animal origin, defined immunologic specificity, chemical purity, and conjugation form. The reliability of immunophelometry for the assay of albumin and AFP in the culture medium was confirmed by: (1) the absence of detectable cross reactivity with other proteins present in the medium, especially transferrin and bovine albumin; (2) the absence of detectable A F P and

albumin in the medium of pure RLEC maintained under the same culture conditions; (3) the linear curve established as reference by addition to the medium of increasing amounts of rat albumin or A F P ranging from 1 to 35 ~g. The sensitivity of assays was 0.5-1 /~g/ml. Each experimental point was the mean of triplicate cultures; the assays were repeated three times. Inter- and intra-assay variations did not exceed 10%.

Results Culture characteristics

After collagenase dissociation, about 90% of the fetal or neonatal hepatocytes were viable as indicated by their refringence under phase contrast microscopy. Within 3 h of culture, about 80% of these cells attached and spread; then they formed monolayers of granular epithelial cells. Hepatocytes were either cultured alone under conventional conditions or co-cultured with RLEC. In conventional culture conditions, by using a selective arginine-free medium, the purity of the cultures was strongly improved (Leffert and Paul, 1972). The addition of hydrocortisone to the selec-

Fig. 1. Phase-contrast micrographs of fetal rat hepatocytes maintained alone (a) or co-cultured with RLEC (b) for 1 day in the presence of 7 x 10 - 6 M hydrocortisone. RLEC were seeded 3 h after parenchymal cells. In places, refringent intercellular spaces corresponding to bile canaliculi are visible(arrow). Day 9 of co-culture, fetal hepatocytes exhibit a more globular shape (c). a c × 140.

a

//\

b

ib

IO0

/

,

Z

'i < L~

ii

I

1

2

3

4

5

O

7

DAYS

DAYS

d

/

A

\

/

\

/

\

!

\

/

/

~'--~

/

,<

/ /

z

/

7~" ~

-~,...

/

~ ' ~ ~ "-'-~v~

;..~. [MD.

\

'l

I

loo agl .<

~-

1

3

~

5

6

7 I)AYS

- ~-

-

-I.--

-I-

-

_-g-- -

--t-

-

~

8

DAYS

Fig. 2. Albumin and AFP secretions by fetal (a, b) and neonatal (c, d) rat hepatocytes cultured in conventional conditions (O) or with R L E C ( 0 ) in the absence ( - - ) or presence of 7 × 1 0 -6 M hydrocortisone hemisuccinate ( . . . . . . ). Fetal and neonatal hepatocytes were cultured in an arginine-free medium supplemented with serum. Albumin and AFP release are expressed in percent of the values obtained with l-day cultures (albumin: 15 # g / m l of medium; AFP: 20 ~tg/ml of medium), Each point represents the mean of triplicate cultures.

263 tive medium increased hepatocyte survival and induced structures similar to bile canaliculi, characteristic of parenchymal cells (Fig. la). In co-culture, RLEC actively divided, reached confluency with hepatocyte cords within 24 h after their addition and then stopped dividing (Fig. lb). In the presence of glucocorticoids, cell survival was prolonged for several weeks. After a few days of co-culture, instead of spreading, hepatocytes prepared from fetuses, as well as newborn rats, formed groups of packed cells (Fig. lc). Hepatocyte growth was limited to the first 5 6 days of culture as indicated by [methyl-3H]thymi dine incorporation and appeared to depend upon the culture conditions: first, the percentage of cells that entered into S phase was two times higher in pure than in mixed cultures; second, the percentage of labelled cells, and thus cell division, was

lowered about 35% when glucocorticoids were added to the medium of pure and mixed cultures. Albumin and A F P production

In the absence of hydrocortisone, secretion of albumin and AFP by pure populations of hepatocytes maintained in a selective arginine-free medium, decreased dramatically within the first three days of culture and then remained at a very low level (Fig. 2a-d). In the presence of hormone, AFP secretion by fetal and newborn rat hepatocytes also fell strongly (Fig. 2b, d). Albumin secretion, on the other hand, decreased only in fetal cultured hepatocytes and appeared more stable in neonatal hepatocytes (Fig. 2a, c). In co-culture, in the absence of hydrocortisone. albumin secretion by both fetal and newborn rat

Fig. 3. Localization by peroxidase-labelledantibodies of albumin in fetal rat hepatocytes cultiyated in association with RLEC. 6 days after hepatocyte seeding. All the hepatocytes, but not RLEC contain albumin ( × 160).

264

Fig. 4. Electron micrographs of fetal rat hepatocytes (H) co-cultured with RLEC. Day l of co-culture: (a) close contacts (arrow) are visible between the two cell types ( × 14000); (b) secretion of an amorphous extracellular material (ECM) was observed in intercellular spaces between the two types (×26000). Day 6 of co-culture: (c) an increased ECM formed of fibrillar material was observed ( × 26000). gl = glycogen particles; M = mitochondrion.

265 hepatocytes declined slightly (Fig. 2a, c). In contrast, AFP secretion was maintained at a high level for one week in fetal cell cultures, but was only partly maintained in newborn rat cultures (Fig. 2b, d). In the presence of hormone, albumin secretion was slightly increased or stabilized at a high level for more than one week in cultured hepatocytes froha fetal and newborn rats (Fig. 2a, c). By using the immunoperoxidase technique, albumin was localized only in most of the hepatocytes (Fig. 3). In contrast, AFP secretion completely disappeared within 24 h after glucocorticoid addition, whatever the developmental stage of the rat liver (Fig. 2b, d).

Reticulin staining. The insoluble extracellular material was analyzed by the silver impregnation method. Important differences were observed in the number and the distribution of the reticulin fibers according to the presence or absence of hydrocortisone. In the absence of hydrocortisone, very few fibers were observed between epithelial cells (Fig. 5a). In the presence of 7 × 10 6 M hydrocortisone, the number of fibers increased in epithelial cell areas (Fig. 5b). Whatever hydrocortisone concentration, reticulin fibers were never observed in pure hepatocyte cultures.

Discussion

Morphological studies Cell morphology.

In co-culture, within 24 h of RLEC addition, discontinuous close contacts were established between hepatocytes and these cells (Fig. 4a), and an amorphous extracellular material was observed in the intercellular spaces (Fig. 4b). The amount of extracellular material increased with age of the co-culture. Typical striated fibrils were observed after a few days (Fig. 4c). During the same time, contacts between hepatocytes and RLEC tended to disappear (Fig. 4c). Moreover, in these culture conditions, hepatocytes exhibited a typical fine structure for several days; they contained an abundant rough endoplasmic reticulum and numerous glycogen particles (Fig. 4c).

The present study shows that fetal and newborn rat hepatocytes maintain specific activities for a prolonged period when co-cultured with RLEC, while in pure culture they are rapidly lost (Leffert and Paul, 1972; Armato et al., 1975; Acosta et al., 1978; Guguen-Guillouzo et al., 1980). The differences between pure and mixed cultures could be related to a different hepatocyte proliferative rate or to establishment of intercellular contacts between the two cell types. In co-cultures, less hepatocytes entered into S phase and mitosis as indicated by the lower incorporation of labelled thymidine. Therefore, the number of parenchymal cells per flask was lower than in pure culture after a few days. It appeared that the better maintenance of specific liver func-

Fig. 5. Silver impregnation of cultures of fetal rat hepatocytes mixed with RLEC on day 9. The cells were incubated in the absence (a) or presence of 7x10 -6 M hydrocortisone (b). Reticulin fibers (arrow) are observed between the two cell types as well as in hepatocytes and RLEC areas ( x 120).

266

tions in co-culture cannot be attributed to a higher proliferative rate of hepatocytes. However, the need for interactions between hepatic endoderm and mesenchyme has been well documented during liver organogenesis (Houssaint, 1980), but not during the perinatal period. We have demonstrated that non-proliferating adult hepatocytes similarly express specific functions at high levels when cocultured with RLEC (Guguen-Guillouzo et al., 1983). It can be presumed that a similar signal provided by contacts between hepatocytes and RLEC is concerned in maintenance of specific functions in fetal and neonatal rat liver. During perinatal development, AFP secretion declines while albumin production increases (Abelev, 1971: Sell and Becker, 1978; Sala-Tr~pat et al., 1979). Glucocorticoids dramatically inhibit AFP synthesis without any effect on albumin when administered to newborn rats (Commer et al., 1979; Brlanger et al., 1981). They also inhibit production of AFP by cultured hepatoma cells while that of albumin is increased (Chou et al., 1982). These results contrast with those obtained by Freeman et al. (1981) with explant cultures of mouse fetal liver. Moreover, in newborn rats, dexamethasone has been shown to decrease the number of AFP mRNA sequences~ which indicates that this hormone could operate at the transcriptional level (Brlanger et al., 1981: Chou et al., 1982). Co-cultured hepatocytes, in the absence of glucocorticoids, remained capable of secreting high levels of both proteins, suggesting that they retained their initial degree of differentiation. Indeed as shown for albumin by immunoperoxidase~ RLEC do not participate in the production of the protein. In co-cultures, incubated in the presence of hydrocortisone, AFP secretion was inhibited while albumin synthesis remained high or increased. This reciprocal relationship between AFP and albumin suggests that as in vivo, hepatocyte maturation is possible in vitro. However, the mechanism by which glucocorticoids act on AFP gene transcription remains unclear, since it has been reported that these hormones inhibit cell proliferation and that the rate of AFP secretion is related to hepatocyte growth (Leffert and Sell, 1974; Guillouzo et al., 1979). It must be pointed out that in co-culture, whether glucocorticoids were

present or not, cell division was strongly reduced, whereas AFP secretion rapidly declined when hydrocortisone was added to the culture medium. This suggests that this hormone could play a direct role on AFP gene transcription. Work is in progress to demonstrate a transcriptional control of albumin and AFP genes by glucocorticoids in vitro. To express in vitro increased specific.functions, i.e. to maturate, fetal and newborn hepatocytes required the presence of glucocorticoids in the nutritional medium and establishment of contacts with another liver cell type. Under these conditions, the deposition of an abundant extracellular material primarily between the two cell types was demonstrated by conventional histological staining. Preliminary studies indicate that like the liver biomatrix, this material contains various components, including different collagen types and fibronectin (B. C16ment et al., unpublished observations); these previous results agree with the observations of Stenman and Vaheri (1978). Moreover, important differences were observed in the number and distribution of the reticulin fibers depending on the concentration of the hormone. Indeed, glucocorticoids are known to increase the biosynthesis of fibronectin in human fibrosarcoma cells and normal fibroblasts (Oliver et al., 1983) and to s t i m u l a t e the p r o d u c t i o n of a fibronectin-containing extracellular matrix in SV40 transformed human fibroblasts (Furcht et al., 1979a, b) and newborn rat hepatocytes (Marceau et al., 1980). The production of the extracellular material could represent a parallel maturation of hepatocytes and their pericellular environment. However, it remains to be determined whether the deposition of the extracellular material is an important requirement or just a consequence of the high functional activity of cultured hepatocytes. It must be kept in mind that although hepatocytes cultured on various components of extracellular matrices survive longer, they do not maintain a normal phenotype (Michalopoulos and Pitot, 1975; Sirica et al., 1979; Rojkind et al., 1980). In conclusion our previous observations with cultured adult hepatocytes and the present results with fetal rat hepatocytes emphasize the role played by specific cell-cell interactions. The fetal co-culture represents a good model for studying the

267

different mechanisms involved in the sequential events which occur during liver maturation.

Acknowledgements We acknowledge the technical assistance of Mrs. N . P a s d e l o u p a n d M r s . A. S a u n i e r ; w e t h a n k M r s . M . R i s s e l a n d M r s . Y, M a z u r i e r f o r p r e p a r i n g t h e m i c r o g r a p h s a n d Mrs. A. V a n n i e r for typing the manuscript. This work was supported by the Institut National

d e la S a n t 6 et d e la R e c h e r c h e

M r d i c a l e a n d b y t h e C e n t r e N a t i o n a l d e la R e cherche Scientifique.

References Abelev, G.I.: Alpha-fetoprotein in oncogenesis and its association with malignant tumors. Adv. Cancer Res. 14, 295-358 (1971). Acosta, D., D.C, Anuforo and R.V. Smith: Primary monolayer cultures of postnatal rat liver cells with extended differentiated functions. In Vitro 14, 428-435 (1978). Armato, U., E. Draghi and P.G. Andreis: Studies on the persistence of differentiated functions in rat hepatocytes set into primary tissue cultures. Specific binding of [3 H]bilirubin after eight days of staying in vitro. Cell Differ, 4, 147-153 (1975). Baffet, G., B. Clement, D. Glaise, A. Guillouzo and C. Guguen-Guillouzo: Hydrocortisone modulates the production of extracellular material and albumin in long-term co-cultures of adult rat hepatocytes with other liver epithelial cells. Biochem. Biophys. Res. Commun. 109, 507-512 (1982). Brlanger, L., M. Frain, P. Baril, M.C. Gingras, J. Bartkowiak and J.M. Sala-Trrpat: Glucocorticosteroid suppression of alpha 1-fetoprotein synthesis in developing rat liver. Evidence for selective gene repression at the transcriptional level. Biochemistry 20, 6665-6672 (1981). Bissell, D.M., L.E. Hammaker and U.A. Meyer: Parenchymal cells from adult rat liver in non-proliferating monolayer culture. I. Functional studies. J. Cell Biol. 59, 722-734 (1973). Chou, J.Y., M. Toshiaki and M. Feldman: Inhibition of synthesis of alpha-fetoprotein by glucocorticoids in cultured hepatoma cells. J. Cell Biol. 93, 314-317 (1982). Commer, P., C. Schwartz, S. Tracy, T. Tamaoki and J.F. Chiu: Dexamethasone inhibits alpha-fetoprotein gene expression in developing mouse liver. Biochem. Biophys. Res. Commun. 89, 1294-1299 (1979). Freeman, A.E., E. Engvall, K. Hirita, Y. Yoshida, R.H. Kottek V. Hilborn and E. Ruoslahti: Differentiation of fetal liver cells in vitro. Proc. Natl. Acad. Sci. USA 78, 3659-3663 (1981).

Furcht, L.T., D.F. Mosher, G. Wendelschafer-Crabb and J.M. Foidart: Reversal by glucocorticoid hormones of the loss of a fibronectin and procollagen matrix around transformed human cells. Cancer Res. 39, 2077-2083 (1979a). Furcht, L.T., D.F. Mosher, G. Wendelschafer-Crabb. P.A. Woodbridge and J.M. Foidart: Dexamethasone-induced accumulation of a fibronectin and collagen extracellular matrix in transformed human cells. Nature 277, 393-395 (1979b). Gordon, H. and H.J. Sweets, Jr.: A simple method for the silver impregnation of reticulin. Am. J. Pathol. 12, 545-551 (1936). Grisham, J.W.: Cell types in long-term propagable cultures of rat liver. Ann. N.Y. Acad. Sci. 349, 128-137 (1980). Guguen-Guillouzo, C., L. Tichonicky, M.F. Szajnert and J. Kruh: Changes in some chromatin and cytoplasmic enzymes of perinatal rat hepatocytes in culture. In Vitro 16, I 10 (1980). Guguen-Guillouzo, C., B. Clrment, G. Baffet, C. Beaumont, E. Morel-Chany, D. Glaise and A. Guillouzo: Maintenance and reversibility of active albumin secretion by adult rat hepatocytes co-cultured with another liver epithelial cell type. Exp. Cell Res. 143, 47 54 (1983). Guillouzo, A., M. Boisnard-Rissel, L. Belanger and M. Bourel: Alpha 1-fetoprotein production during the hepatocyte growth cycle of developing rat liver. Biochem. Biophys. Res. Commun. 91, 327-331 (1979). Guillouzo, A., C. Beaumont, E. Le Rumeur, M. Rissel, M.F. Latinier, C. Guguen-Guillouzo and M. Bourel: New findings on immunolocalization of albumin in rat hepatocytes. Biol. Cell 43, 163-172 (1982). Houssaint, E.: Differentiation of the mouse hepatic primordium. 1. An analysis of tissue interactions in hepatocyte differentiation. Cell Differ. 9, 269-279 (1980). Leffert, H.L. and D. Paul: Studies on primary cultures of differentiated fetal liver cells. J. Cell Biol. 52, 559-568 (1972). Leffert, H.L. and S. Sell: Alpha 1-fetoprotein biosynthesis during the growth cycle of differentiated fetal rat hepatocytes in primary cultures. J. Cell Biol. 61, 823 829 (1974). Marceau, N., R. Goyette, J.P. Valet and J. Deschenes: The effect of dexamethasone on formation of a fibronectin extracellular matrix by rat hepatocytes in vitro. Exp. Cell Res. 125,497-502 (1980). Michalopoulos, G. and H.C. Pitot: Primary culture of parenchymal ceils on collagen membranes: morphological and biochemical observations. Exp. Cell Res. 94, 70-78 (1975). Morel-Chany, E., C. Guillouzo, G. Tringal and M.F. Szajnert: 'Spontaneous' neoplastic transformation in vitro of epithelial cell strains of rat liver: cytology, growth and enzymatic activities. Eur. J. Cancer 14, 1341-1352 (1978). Oliver, N., R.F. Newby, L.T. Furcht and S. Bourgeois: Regulation of fibronectin biosynthesis by glucocorticoids in human fibrosarcoma cells and normal fibroblasts. Cell 33, 287-296 (1983). Ritchie, R.F.: Automated immunoprecipitation analysis of serum proteins. In: The Plasma Proteins: Structure, Func-

268 tion and Genetic Control, Vol, II, ed. F.W. Putman (Academic Press, New York) pp. 375 425 (1975). Rojkind, M., S. Mackensen, M.A. Giambrone, P. Ponce and E.M. Reid: Connective tissue biomatrix: its isolation and utilization for long-term cultures of normal rat hepatocytes. J. Cell Biol. 87, 255-283 (1980). Sala-Tr6pat, J.M., J. Dever, T.D. Sargent, K. Thomas, S. Sell and J. Bonnet: Changes in expression of albumin and alpha-fetoprotein genes during rat liver development and neoplasia. Biochemistry 18, 2167-2178 (1979). Sell, S. and F.F. Becker: Alpha-fetoprotein. J. Natl. Cancer lnst. 60, 19 26 (1978). Sirica, A.E. and H.C. Pitot: Drug metabolism and effects of

carcinogens in cultured hepatic cells. Pharmacol. Rev. 31, 205-228 (1979). Sirica, A.E., W, Richards, Y. Tsukada, C.A. Sattler and C. Pitot: Fetal phenotypic expression of adult rat hepatocytes on collagen gel/nylon meshes. Proc. Natl. Acad. Sci. USA 76, 283-287 (1979). Stenman, S. and A. Vaheri: Distribution of a major connective tissue protein, fibronectin, in normal human tissues. J. Exp. Med. 147, 1054-1064 (1978). Williams, G.M. and J.M. Gunn: Long-term cell culture of adult rat liver epithelial cells. Exp. Cell Res. 89, 139-142 (1974). Williams, G.M., E.K. Weisburger and J.H. Weisburger: Isolation and long-term cell culture of epithelial like cells from rat liver. Exp. Cell Res. 69, 106-112 (1971).