Proc. Nati. Acad. Sci. USA
Vol. 76, No. 1, pp. 283-287, January 1979 Cell Biology
Fetal phenotypic expression by adult rat hepatocytes on collagen gel/nylon meshes (y-glutamyl transpeptidase/al-fetoprotein/aldolase/DNA synthesis/microfilaments) ALPHONSE E. SIRICA*, WILLIAM RICHARDS*, YUTAKA TSUKADAt, CAROL A. SATTLER*, AND HENRY C. PITOT*
*'McArdle Laboratory for Cancer Research, The Medical School, University of Wisconsin, Madison, Wisconsin 53706; and tDepartment of Biochemistry, Hokkaido University School of Medicine, Sapporo, Hokkaido, Japan
Communicated by Van R. Potter, September 5, 1978
ABSTRACT Hepatocytes from adult rats were maintained in primary culture for up to 10-13 days on nylon meshes coated with a thin layer of rat tail collagen gel. Their ultrastructu're closely resembled that of the liver parenchymal cell in vivo, but hepatocytes in late culture exhibited a pronounced buildup of microfilaments beneath their apical cell surface. Hepatocytes in earl and late cultures secreted albumin, transferrin, and al-acid glycoprotein into the medium; they exhibited a 7- to 10-fold induction of tyrosine aminotransferase activity by dexamethasone; and they expressed an alkaline phosphatase that was similar to that of normal rat liver with respect to its inhibition by the liver enzyme inhibitor L-homoarginine. In addition, the hepatocytes in culture demonstrated phenotypic changes characteristic of fetal liver parenchymal cells. These changes, which paralleled an increase in DNA synthesis, included the expression of and linear increase in the activity of the fetal liver cell enzyme y-glutamyl transpeptidase, an increased production of a1-fetoprotein, and a c ange in the substrate specificity of fructose-bisphosphate aldolase to that of the fetal liver isozyme.
Preparation of Collagen Gel/Nylon Mesh Substratum. Swiss nylon monofilament mesh fabric (Nitex HC3-253) was purchased from TETKO (Elmford, NY). Circular meshes were cut from the fabric to fit the bottoms of 100 X 20 mm tissue culture dishes (Falcon). Meshes were consecutively washed with 95% ethanol and distilled H20, dried, and autoclaved. To coat the mesh with collagen gel, 3.5 ml of a rat tail collagen solution (8) was evenly spread in a tissue culture dish. The mesh was then placed into the dish and gently tapped with a sterile forceps until its top surface became covered with the collagen solution. To this was added 0.86 ml of a 2:1 mixture of 10 times concentrated Hi/Wo/Ba medium and 0.34 M NaOH (8). The dish was then gently shaken until the mesh became evenly coated with a thin collagen gel matrix. Coated meshes were individually placed into culture dishes containing 7 ml of L-15 medium and stored in a 37°C incubator for 24 hr before use. Cell Isolation and Culture Conditions. Hepatocytes were isolated from normal adult male albino rats (Holtzman, 200-230 g) by the collagenase perfusion method of Berry and Friend (9) as modified by Bonney et al. (10). The cells were then sedimented and washed in L-15 medium as described (8). Hepatocytes were finally suspended in L-15 medium at pH 7.4 supplemented with Hepes (18 mM), albumin (2 mg/ml), penicillin (100 ,g/ml), streptomycin (100 jig/ml), glucose (1.5 mg/ml), insulin (0.5 ,ug/ml), and 5% fetal calf serum and plated at a density of 10-14 X 106 cells per mesh. The cultures were maintained at 37°C in air in an incubator. After incubation for 4 hr, the meshes with attached cells were transferred to culture dishes containing 7 ml of fresh medium. During this period, between 50 and 60% of the plated cells attached to the mesh/gel substratum. The medium was replaced with serum-free medium after the first 24 hr of incubation, and fresh medium changes were made every 24 hr thereafter. Cell Harvest Conditions. Meshes with attached cells were washed three times with phosphate-buffered saline (pH 7.4). The washes were discarded, and 7 ml of collagenase solution in phosphate-buffered saline (1 mg/ml) was added to each culture. After a 10-min incubation at 370C, the detached cells were collected with a pasteur pipette and placed in 15-ml Corex centrifuge tubes. This treatment was sufficient to release 100% of the cells from their substratum. The cells were then pelleted by centrifugation in the cold at 1000 rpm for 10 min in a Sorvall GLC-1 centrifuge. The resulting cell pellets were washed three times in 5 ml of phosphate-buffered saline and then resuspended in the appropriate buffers for subsequent biochemical
It is well-established that, in vivo, hepatocytes from adult rats express fetal liver cell characteristics during hyperplasia and hepatocarcinogenesis (1, 2). Examples include expression by the hepatocyte of the enzyme y-glutamyl transpeptidase (3-5), production of a-fetoprotein (6), and the synthesis of fetal isozymes of enzymes such as fructose-bisphosphate aldolase (7). The functional significance and biochemical regulation of these transitions as they relate to the proliferative state are for the most part unknown and are relatively difficult to study in vivo. We now report a system for maintaining adult rat hepatocytes in primary culture and provide evidence for their rapid transition to a more fetal-like state characterized by the production of fetal proteins and an increase in hepatic DNA synthesis.
MATERIALS AND METHODS Reagents. Leibovitz (L-15) tissue culture medium, penicillin/streptomycin mixture, and fetal calf serum were purchased from GIBCO. Hi/Wo/Ba medium (10 times concentrated) was obtained from International Scientific Industries (Cary, IL). Insulin, collagenase (type 1), L-y-glutamyl-p-nitroanilide, glycylglycine, p-nitrophenyl phosphate, D-fructose 1,6-bisphosphate, D-fructose 1-phosphate, and a-glycerophosphate dehydrogenase-triosephosphate isomerase were purchased from Sigma. Dexamethasone phosphate was purchased from Merck, Sharp & Dohme; NADH from Boehringer Mannheim; and L-(+)-homoarginine from Aldrich. y-Glutamyl-4-methoxy2-naphthylamide was obtained from Vega-Fox Biochemicals (Tucson, AZ), and [methyl-3H1thymidine (40-60 Ci/mmol) was purchased from New England Nuclear.
analysis.
Enzyme and Chemical Assays. y-Glutamyl transpeptidase activity was assayed at 25°C according to a slight modification of the procedure of Tateishi et al. (5) in 1.0 ml of 0.1 M N,Nbis(2-hydroxyethyl)glycine (Bicine buffer), pH 8.3, containing 4.4 mM L--y-glutamyl-p-nitroanilide, 11 mM MgCl2, and 40 mM glycylglycine. Alkaline phosphatase activity was deter-
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
283
284
Cell Biology: Sirica et al.
mined as described (11) in 1.0 ml of 0.07 M 2-amino-2methyl-1,3-propanediol buffer, pH 10.0, containing 5 mM p-nitrophenyl phosphate and 5 mM MgCl2. Reactions were run at 370C after a 45-min preincubation of the enzyme sample at 00C in buffer or with 25 mM L-homoarginine, an inhibitor of liver-type alkaline phosphatase (12). The fructose-bisphosphate aldolase activities were determined. according to a slight modification of the procedure of Lebherz and Rutter (13). Activities were assayed at 250C in 1.0 ml of 0.1 M Tris buffer, pH 7.5, containing 0.3 mM NADH, 25 ,g of a-glycerophosphate dehydrogenase-triosephosphate isomerase, and either 2.5 mM fructose 1,6-bisphosphate or 10 mM fructose 1-phosphate. Tyrosine aminotransferase activity was assayed at pH 7.6 by the procedure of Diamondstone (14). Protein was determined by the method of Lowry et al. (15) with bovine serum albumin as a standard. DNA was measured according to Burton (16), as modified by Bonney et al. (10), with calf thymus DNA as a standard, or by a slight modification of the fluorometric procedure of Kissane and Robins (17). Histochemical Determination of -t-Glutamyl Transpeptidase. 'y-Glutamyl transpeptidase activity was determined histochemically by a slight modification of the procedure of Rutenburg et al. (18). Meshes with attached hepatocytes, cell smears, or cryostat sections of frozen cell pellets isolated from the meshes were fixed for 30 min in ice-cold acetone. The cells were then incubated for 30-40 min at 250C in 25 mM Tris buffer, pH 7.4, containing 0.4 mM 'y-glutamyl-4-methoxy2-naphthylamide, 3.0 mM glycylglycine, and a diazonium salt (fast blue B). After incubation, the resulting azo dye was chelated with cupric ion to yield an insoluble red dye at the site of enzyme activity. Immunochemical Assays. Monospecific rabbit or horse antibodies were prepared against rat al-fetoprotein, albumin, transferrin, and al-acid glycoprotein. These proteins were purified according to published procedures (19-22) and were homogeneous by immunoelectrophoresis or disc electrophoresis. al-Fetoprotein in cell-free culture medium and in hepatocytes collected at 24-hr intervals was determined by radioimmunoassay (23). The other serum proteins were measured in cell-free medium or cells by the single radial immunodiffusion method of Mancini et al. (24). Incorporation of [3HJThymidine into DNA. After various times in culture, hepatocytes were pulse-labeled for 18-20 hr
Proc. Natl. Acad. Sci. USA 76 (1979)
with [3H]thymidine at a concentration of 1 ACi/ml of medium. The cells were collected at the indicated times, and the incorporated radioactivity was measured either directly in DNA purified on CsCl gradients (25) or by autoradiography of cell smears and serial cryostat sections of frozen cell pellets. The cell smears and serial sections were stained with hematoxylin and eosin or for y-glutamyl transpeptidase activity, and the percentage of hepatocytes undergoing DNA synthesis was determined from random counts of [3H]thymidine-labeled nuclei per thousand cells. RESULTS The hepatocytes formed a confluent monolayer-by 24-48 hr after plating and could be maintained in culture for up to 10-13 days. Phase-contrast microscopy of the hepatocytes in monolayer culture demonstrated their square or polygonal shape and their large vesicular nuclei with pronounced nucleoli, a granular cytoplasm, and sharply delineated cell borders (Fig. 1 inset). Both mononucleate and binucleate hepatocytes were observed in the monolayer, but no attempt was made to quantitate their proportions in either early or late cultures. The ultrastructure of hepatocytes in early and late cultures closely resembled that seen in vivo (Fig. 1A). Structures resembling bile canaliculi occurred between adjacent cells, and desmosomes and tight junctions were observed at the plasma membranes that extended laterally from the canaliculi. The cells possessed a well-developed Golgi apparatus and numerous mitochondria; they contained variable amounts of glycogen; and their rough endoplasmic reticulum was preserved and, in some cells, was quite extensive. Lipid droplets and autophagic vacuoles were also seen, particularly in the late cultures. In sections cut perpendicularly to the mesh, the hepatocytes were found to have a cuboidal shape, and cells in late culture (days 7-10) frequently exhibited a progressive buildup of microfilaments at their apical cell surface but not at their lateral cell surfaces nor at their substratum surface (Fig. 1B). Table 1 shows the production of serum proteins by hepatocytes on collagen gel/nylon meshes. Albumin release into the medium decreased with time, but immunoreactive "albumin" in the hepatocytes steadily increased from 0.08 + 0.02 (SD) ng/,ug of DNA on day 1 to 1.14 + 0.3 ng/,ug of DNA on day 9. We have not determined, as yet, whether the increased amounts of "albumin" in the hepatocytes over time represents albumin,
FIG. 1. Adult rat hepatocytes on collagen gel/nylon meshes. (A) Electron micrograph of hepatocytes at 10 days in culture. (X3100.) Arrow, bile canaliculus; g, Golgi apparatus; L, lipid droplet. (B) Electron micrograph of a binucleate hepatocyte in a 10-day culture sectioned perpendicular to the mesh. (X3100.) M, microfilament network beneath cell surface facing the medium. (Inset) Phase-contrast micrograph of an 8-day-old culture. (X100.) NM, nylon mesh fiber.
Cell Biology: Sirica et al.
Proc. Natl. Acad. Sci. USA 76 (1979)
Table 1. Production of rat serum proteins by adult hepatocytes on collagen gel/nylon meshes
9
Days in culture
Albumin
Transferrin
1 2 3 4 5
147.3 ± 6.0 131.9+15 138.6 ± 10 64.6 + 9.4 42.9*
8.3 i 0.4 11.8 1.1 17.1 + 3.0 17.1 ± 2.3 15.5* (14.8-16.2) 12.8 ± 0.5 12.1 ± 1.3 7.7 + 0.6
(35-51) 6 7 9
25.9 + 2.5 22.9 + 0.8 14.9 + 1.4
a,-Acid glycoprotein
0.180.16_ 0.14 0.120.10 0.080.06 0.040.02-
90% of the hepatocytes in culture exhibited a strong histochemical reaction for y-glutamyl transpeptidase activity. The enzyme activity was confined to the cytoplasm of the vast majority of the hepatocytes, but we also observed some hepatocytes that exhibited a strong histochemical reaction at their cell surface. Furthermore, both mononucleate and binucleate hepatocytes were positive for the enzyme activity. No enzyme activity was detected when the -y-glutamyl substrate was omitted from the reaction mixture or when L-leucyl-4-methoxy-2-naphthylamide was used as substrate, and the histochemical reaction was markedly decreased when glycylglycine was not included in the reaction mixture. Fig. 3 shows the production of a1-fetoprotein by the hepatocyte cultures. a1-Fetoprotein could not be detected by radioimmunoassay in the culture medium or in the hepatocytes collected at 24-hr intervals during the first 4 days of culture. However, significant levels of a1-fetoprotein were measurable in the medium and in the cells during the last 5 days of culture. A preliminary cytoimmunofluorescence study has also shown that at least 8.0% of the total number of cells contained alfetoprotein on day 8 in culture. The specific activities of fructose-bisphosphate aldolase and alkaline phosphatase in the hepatocytes in culture decreased over time (Table 2). As with normal liver, the aldolase of the cultured hepatocytes exhibited a fructose 1,6-bisphosphate/ fructose 1-phosphate specific activity ratio of 1 during the first 6 days of culture. However, at day 8 in culture, the substrate specificity of the aldolase was more like that of the fetal and neoplastic livers (1, 7). On the other hand, the alkaline phosphatase of the cultured hepatocytes responded like the normal liver enzyme with respect to its inhibition by L-homoarginine.
DNA Synthesis in Cultured Hepatocytes. The incorporation of [3H]thymidine into DNA as a function of the age of the hepatocyte culture is shown in Fig. 4. Autoradiography of cryostat sections and cell smears prepared from a day 8 culture demonstrated that between 6 and 10% of the y-glutamyl transpeptidase-positive cells exhibited strong nuclear labeling with [3H]thymidine. Similar labeling index curves for [3H]thymidine were obtained in a separate experiment from cell sections and smears stained with hematoxylin and eosin (not shown). Hydroxyurea (25 mM) inhibited the incorporation of [3H]thymidine into DNA by 95% in a day-8 culture but did not inhibit the 7y-glutamyl transpeptidase activity of the hepatocytes. Also, exposure of the hepatocytes in culture to 10 1AM
dexamethasone for 6 days significantly decreased the [3H]thymidine labeling index but not the number of cells exhibiting -y-glutamyl transpeptidase activity. The percentage of y-glutamyl transpeptidase-positive cells on day 8 in culture was 90.8 ± 8.6 (SD) for the controls and 97.6 i 2.2 for the dexamethasone-treated cultures, but the [3H]thymidine labeling index for the y-glutamyl transpeptidase-positive cells was decreased from 9.5 ± 4.5% for the controls to 0.3 + 0.5% for the cells cultured with dexamethasone. DISCUSSION The collagen gel/nylon mesh support system represents an alternative to the floating collagen gel (8) and Millipore filter (27) substrata and, as such, provides a number of advantages over these two as well as the more conventional plastic (10, 28) and collagen-coated plastic surfaces (29) used to support rat hepatocytes in primary cultures.§ Unlike the floating collagen gel, the collagen coated/nylon mesh does not shrink with time in culture, it is easier to handle than the thin flexible gels, and it allows for the rapid removal and separation of the hepatocytes from their collagen substratum with collagenase. The mesh system also permits direct microscopic observation of the cells in culture, and, like the floating collagenf gel and Millipore filter support systems, it extends the functional longevity of the hepatocytes in culture over those maintained under comparable conditions on plastic and collagen-coated plastic surfaces. Hepatocytes in early and late mesh cultures possessed a complex ultrastructure that was identical to that reported for adult rat hepatocytes supported on floating collagen gels (31). In this respect, the hepatocytes on the mesh exhibited morphological features that resembled to varying degrees those of both adult and fetal hepatocytes in vivo. The most dramatic ultrastructural change observed was the extensive buildup of microfilaments beneath the apical surface of the hepatocytes in late culture. A similar buildup of microfilaments at the cell periphery has been described for rat hepatocytes cultured on plastic surfaces (32, 33) but the complexity of the microfilament network was generally more pronounced in hepatocytes on the mesh than in those on plastic. The functional significance of this polarized accumulation of microfilaments in the cultured hepatocytes is unknown at present, but it is interesting that an increase in microfilaments has also been shown to occur in vivo
in carcinogen-induced preneoplastic hepatocytes (34), in § While this manuscript was in preparation, Cereijido et al. (30) described the preparation of a collagen-coated nylon cloth disk used to maintain an epithelial cell line (MDCK) in monolayer culture. Their system differs from the collagen gel/nylon mesh primarily in that 4% glutaraldehyde was used to promote crosslinking of the collagen covering the cloth disk.
Cell Biology: Sirica et al. hepatocytes during liver regeneration (35), and in fetal liver hepatocytes (36, 37). The most unique feature of the hepatocytes on the collagen gel/nylon mesh was their positive expression of fetal gene properties along with a concomitant increase in DNA synthesis. In this respect, the changes in y-glutamyl transpeptidase activity, a1-fetoprotein production, and aldolase substrate specificity simulated those exhibited by hepatocytes in vivo during various hepatoproliferative conditions. Recently, Leffert et al. (38) have independently described a liver cell culture system that, like the mesh system, displayed adult-to-fetal cell-like changes in apparent association with hepatoproliferative transitions. In comparison, the mesh/gel culture system exhibited kinetics for al-fetoprotein production similar to those found with this other system, and the alteration in the aldolase isozyme pattern of the hepatocytes on the mesh occurred at a time in culture similar to that shown by Leffert et al. for adult-to-fetal form changes in pyruvate kinase isozymes. However, in contrast to the findings of Leffert et al., we were not able to demonstrate "fetal-to-adult phenotypic reversal" or an increase in hepatocyte cell number and, despite the increased DNA synthesis, mitotic figures were observed infrequently in the mesh/gel cultures. Our results showing an inhibition of DNA synthesis by dexamethasone in culture without affecting the level of y-glutamyl transpeptidase further indicated that the expression of this fetal cell enzyme by the cultured hepatocytes may not be directly linked to an increased DNA synthesis but rather may be representative of an adaptive response by the hepatocytes to their artificial environment. Additional studies are needed to determine the necessary factors required to maintain an adult hepatocyte phenotype in culture as well as to define those that are needed to augment and maintain the proliferative activity of hepatocytes on their collagen gel/nylon substratum. We thank Mr. Gerald L. Sattler for his excellent technical assistance and Dr. James A. Gurr for performing the tyrosine aminotransferase assay. We also thank Dr. Ilse Riegel for her most helpful suggestions concerning the preparation of this manuscript. This study was supported by Grants CA-07175 and RO1-CA-17334-02 from the National Cancer Institute. A.E.S. is a recipient of a National Institutes of Health Postdoctoral Traineeship (T32-CA 09230-02) and W. R. is a recipient of a National Institutes of Health Postdoctoral Fellowship (F32-CA 05395-02). Y.T. is a visiting scientist supported by a grant from the Japan Society for the Promotion of Science. This work was presented in part at the 69th Annual Meeting of the American Association for Cancer Research and at the 62nd Annual Meeting of the Federation of American Societies for Experimental Biology. 1. Uriel, J. (1975) in Cancer, A Comprehensive Treatise, ed. Becker, F. F. (Plenum, New York), Vol. 3, pp. 21-55. 2. Walker, P. R. & Potter, V. R. (1972) Adv. Enzyme Regul. 10,
339-364. 3. Cheng, S., Nassar, K. & Levy, D. (1978) FEBS Lett. 85, 310312. 4. Fiala, S., Mohindru, A., Kettering, W. G., Fiala, A. E. & Morris, H. P. (1976) J. Natl. Cancer Inst. 57,591-598. 5. Tateishi, N., Higashi, T., Nomura, T., Naruse, A., Nakashima, K., Shiozaki, H. & Sakamoto, Y. (1976) Cann 67,215-222.
Proc. Nat!. Acad. Scd. USA 76 (1979)
287
6. Sell, S., Becker, F. F., Leffert, H. L. & Watabe, H. (1976) Cancer Res. 36,4239-4249. 7. Schapira, F., Reuber, M. D. & Hatzfeld, A. (1970) Biochem. Biophys. Res. Commun. 40,321-327. 8. Michalopoulos, G. & Pitot, H. C. (1975) Exp. Cell Res. 94, 7078. 9. Berry, N. M. & Friend, D. S. (1969) J. Cell Biol. 43,506-520. 10. Bonney, R. J., Becker, J. E., Walker, P. R. & Potter, V. R. (1974) In Vitro 9,399-413. 11. Sirica, A. E., Goldblatt, P. J. & McKelvy, J. F. (1975) J. Biol. Chem. 250, 6464-6468. 12. Rufo, M. B. & Fishman, W. 41972) J. Histochem. Cytochem. 20, 336-343. 13. Lebherz, H. G. & Rutter, W. J. (1975) Methods Enzymol. 42, 249-258. 14. Diamondstone, T. I. (1966) Anal. Biochem. 16,395-401. 15. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 16. Burton, K. (1956) Biochem. J. 62, 315-322. 17. Kissane, J. M. & Robins, E. (1959) J. Biol. Chem. 233, 184189. 18. Rutenburg, A. M., Kim, H., Fischbein, J. W., Hanker, J. S., Wasserkrug, H. L. & Seligman, A. M. (1969) J. Histochem. Cytochem. 17,517-526. 19. Watabe, H. (1974) Int. J. Cancer 13,377-388. 20. Schreiber, G., Rotermund, H. M., Maeno, H., Weigand, K. & Lesch, R. (1969) Eur. J. Biochem. 10, 355-361. 21. Boettcher, E. W., Kistler, P. & Nitschmann, Hs. (1958) Nature (London) 181, 490-491. 22. Kamiyama, S. & Schmid, K. (1962) Biochim. Biophys. Acta 58, 80-87. 23. Nishi, S. & Hirai, H. (1976) in Protides of the Biological Fluids, ed. Peeters, H. (Pergamon, Oxford), Vol. 23, pp. 303-307. 24. Mancini, G., Carbonara, A. 0. & Heremans, J. F. (1965) Immunochemistry 2,235-254. 25. Shaw, J. L., Blanco, J. & Mueller, G. C. (1975) Anal. Biochem. 65, 125-131. 26. Chung Lin, J., Lee, G., Sarma, D. S. R. & Farber, E. (1977) Fed. Proc. Fed. Am. Soc. Exp. Biol. 36, 1078. 27. Savage, C. R. & Bonney, R. J. (1978) Exp. Cell Res. 114,307315. 28. Bissel, D. M., Hammaker, L. E. & Meyer, U. A. (1973) J. Cell Biol. 59,722-734. 29. Kletzien, R. F., Pariza, M. W., Becker, J. E. & Potter, V. R. (1976) J. Biol. Chem. 251,3014-3020. 30. Cereijido, M., Robbins, E. S., Dolan, W. J., Rotunno, C. A. & Sabatini, D. D. (1978) J. Cell Biol. 77,853-880. 31. Sattler, C. A., Michalopoulos, G., Sattler, G. L. & Pitot, H. C. (1978) Cancer Res. 38, 1539-1549. 32. Bernaert, D., Wanson, J., Drochmans, P. & Popowski, A. (1977) J. Cell Biol. 74, 878-900. 33. Chapman, G. S., Jones, A. L., Meyer, U. A. & Bissell, D. M. (1973) J. Cell Biol. 59, 735-747. 34. Ogawa, K. & Medline, A. (1978) Proc. Am. Assoc. Cancer Res. 19,63. 35. Gabbiani, G. & Ryan, G. B. (1974) J. Submicrosc. Cytol. 6, 143-157. 36. Chedid, A. & Nair, V. (1974) Dev. Biol. 39,49-62. 37. Santerre, R. F. & Rich, A. (1976) Dev. Biol. 54, 1-12. 38. Leffert, H., Moran, T., Sell, S., Skelly, H. Ibsen, K., Mueller, M. & Arias, I. (1978) Proc. Nat!. Acad. Sci. USA 75, 1834-1838.
Vol. 76, No. 1, pp. 283-287, January 1979 Cell Biology
Fetal phenotypic expression by adult rat hepatocytes on collagen gel/nylon meshes (y-glutamyl transpeptidase/al-fetoprotein/aldolase/DNA synthesis/microfilaments) ALPHONSE E. SIRICA*, WILLIAM RICHARDS*, YUTAKA TSUKADAt, CAROL A. SATTLER*, AND HENRY C. PITOT*
*'McArdle Laboratory for Cancer Research, The Medical School, University of Wisconsin, Madison, Wisconsin 53706; and tDepartment of Biochemistry, Hokkaido University School of Medicine, Sapporo, Hokkaido, Japan
Communicated by Van R. Potter, September 5, 1978
ABSTRACT Hepatocytes from adult rats were maintained in primary culture for up to 10-13 days on nylon meshes coated with a thin layer of rat tail collagen gel. Their ultrastructu're closely resembled that of the liver parenchymal cell in vivo, but hepatocytes in late culture exhibited a pronounced buildup of microfilaments beneath their apical cell surface. Hepatocytes in earl and late cultures secreted albumin, transferrin, and al-acid glycoprotein into the medium; they exhibited a 7- to 10-fold induction of tyrosine aminotransferase activity by dexamethasone; and they expressed an alkaline phosphatase that was similar to that of normal rat liver with respect to its inhibition by the liver enzyme inhibitor L-homoarginine. In addition, the hepatocytes in culture demonstrated phenotypic changes characteristic of fetal liver parenchymal cells. These changes, which paralleled an increase in DNA synthesis, included the expression of and linear increase in the activity of the fetal liver cell enzyme y-glutamyl transpeptidase, an increased production of a1-fetoprotein, and a c ange in the substrate specificity of fructose-bisphosphate aldolase to that of the fetal liver isozyme.
Preparation of Collagen Gel/Nylon Mesh Substratum. Swiss nylon monofilament mesh fabric (Nitex HC3-253) was purchased from TETKO (Elmford, NY). Circular meshes were cut from the fabric to fit the bottoms of 100 X 20 mm tissue culture dishes (Falcon). Meshes were consecutively washed with 95% ethanol and distilled H20, dried, and autoclaved. To coat the mesh with collagen gel, 3.5 ml of a rat tail collagen solution (8) was evenly spread in a tissue culture dish. The mesh was then placed into the dish and gently tapped with a sterile forceps until its top surface became covered with the collagen solution. To this was added 0.86 ml of a 2:1 mixture of 10 times concentrated Hi/Wo/Ba medium and 0.34 M NaOH (8). The dish was then gently shaken until the mesh became evenly coated with a thin collagen gel matrix. Coated meshes were individually placed into culture dishes containing 7 ml of L-15 medium and stored in a 37°C incubator for 24 hr before use. Cell Isolation and Culture Conditions. Hepatocytes were isolated from normal adult male albino rats (Holtzman, 200-230 g) by the collagenase perfusion method of Berry and Friend (9) as modified by Bonney et al. (10). The cells were then sedimented and washed in L-15 medium as described (8). Hepatocytes were finally suspended in L-15 medium at pH 7.4 supplemented with Hepes (18 mM), albumin (2 mg/ml), penicillin (100 ,g/ml), streptomycin (100 jig/ml), glucose (1.5 mg/ml), insulin (0.5 ,ug/ml), and 5% fetal calf serum and plated at a density of 10-14 X 106 cells per mesh. The cultures were maintained at 37°C in air in an incubator. After incubation for 4 hr, the meshes with attached cells were transferred to culture dishes containing 7 ml of fresh medium. During this period, between 50 and 60% of the plated cells attached to the mesh/gel substratum. The medium was replaced with serum-free medium after the first 24 hr of incubation, and fresh medium changes were made every 24 hr thereafter. Cell Harvest Conditions. Meshes with attached cells were washed three times with phosphate-buffered saline (pH 7.4). The washes were discarded, and 7 ml of collagenase solution in phosphate-buffered saline (1 mg/ml) was added to each culture. After a 10-min incubation at 370C, the detached cells were collected with a pasteur pipette and placed in 15-ml Corex centrifuge tubes. This treatment was sufficient to release 100% of the cells from their substratum. The cells were then pelleted by centrifugation in the cold at 1000 rpm for 10 min in a Sorvall GLC-1 centrifuge. The resulting cell pellets were washed three times in 5 ml of phosphate-buffered saline and then resuspended in the appropriate buffers for subsequent biochemical
It is well-established that, in vivo, hepatocytes from adult rats express fetal liver cell characteristics during hyperplasia and hepatocarcinogenesis (1, 2). Examples include expression by the hepatocyte of the enzyme y-glutamyl transpeptidase (3-5), production of a-fetoprotein (6), and the synthesis of fetal isozymes of enzymes such as fructose-bisphosphate aldolase (7). The functional significance and biochemical regulation of these transitions as they relate to the proliferative state are for the most part unknown and are relatively difficult to study in vivo. We now report a system for maintaining adult rat hepatocytes in primary culture and provide evidence for their rapid transition to a more fetal-like state characterized by the production of fetal proteins and an increase in hepatic DNA synthesis.
MATERIALS AND METHODS Reagents. Leibovitz (L-15) tissue culture medium, penicillin/streptomycin mixture, and fetal calf serum were purchased from GIBCO. Hi/Wo/Ba medium (10 times concentrated) was obtained from International Scientific Industries (Cary, IL). Insulin, collagenase (type 1), L-y-glutamyl-p-nitroanilide, glycylglycine, p-nitrophenyl phosphate, D-fructose 1,6-bisphosphate, D-fructose 1-phosphate, and a-glycerophosphate dehydrogenase-triosephosphate isomerase were purchased from Sigma. Dexamethasone phosphate was purchased from Merck, Sharp & Dohme; NADH from Boehringer Mannheim; and L-(+)-homoarginine from Aldrich. y-Glutamyl-4-methoxy2-naphthylamide was obtained from Vega-Fox Biochemicals (Tucson, AZ), and [methyl-3H1thymidine (40-60 Ci/mmol) was purchased from New England Nuclear.
analysis.
Enzyme and Chemical Assays. y-Glutamyl transpeptidase activity was assayed at 25°C according to a slight modification of the procedure of Tateishi et al. (5) in 1.0 ml of 0.1 M N,Nbis(2-hydroxyethyl)glycine (Bicine buffer), pH 8.3, containing 4.4 mM L--y-glutamyl-p-nitroanilide, 11 mM MgCl2, and 40 mM glycylglycine. Alkaline phosphatase activity was deter-
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U. S. C. §1734 solely to indicate this fact.
283
284
Cell Biology: Sirica et al.
mined as described (11) in 1.0 ml of 0.07 M 2-amino-2methyl-1,3-propanediol buffer, pH 10.0, containing 5 mM p-nitrophenyl phosphate and 5 mM MgCl2. Reactions were run at 370C after a 45-min preincubation of the enzyme sample at 00C in buffer or with 25 mM L-homoarginine, an inhibitor of liver-type alkaline phosphatase (12). The fructose-bisphosphate aldolase activities were determined. according to a slight modification of the procedure of Lebherz and Rutter (13). Activities were assayed at 250C in 1.0 ml of 0.1 M Tris buffer, pH 7.5, containing 0.3 mM NADH, 25 ,g of a-glycerophosphate dehydrogenase-triosephosphate isomerase, and either 2.5 mM fructose 1,6-bisphosphate or 10 mM fructose 1-phosphate. Tyrosine aminotransferase activity was assayed at pH 7.6 by the procedure of Diamondstone (14). Protein was determined by the method of Lowry et al. (15) with bovine serum albumin as a standard. DNA was measured according to Burton (16), as modified by Bonney et al. (10), with calf thymus DNA as a standard, or by a slight modification of the fluorometric procedure of Kissane and Robins (17). Histochemical Determination of -t-Glutamyl Transpeptidase. 'y-Glutamyl transpeptidase activity was determined histochemically by a slight modification of the procedure of Rutenburg et al. (18). Meshes with attached hepatocytes, cell smears, or cryostat sections of frozen cell pellets isolated from the meshes were fixed for 30 min in ice-cold acetone. The cells were then incubated for 30-40 min at 250C in 25 mM Tris buffer, pH 7.4, containing 0.4 mM 'y-glutamyl-4-methoxy2-naphthylamide, 3.0 mM glycylglycine, and a diazonium salt (fast blue B). After incubation, the resulting azo dye was chelated with cupric ion to yield an insoluble red dye at the site of enzyme activity. Immunochemical Assays. Monospecific rabbit or horse antibodies were prepared against rat al-fetoprotein, albumin, transferrin, and al-acid glycoprotein. These proteins were purified according to published procedures (19-22) and were homogeneous by immunoelectrophoresis or disc electrophoresis. al-Fetoprotein in cell-free culture medium and in hepatocytes collected at 24-hr intervals was determined by radioimmunoassay (23). The other serum proteins were measured in cell-free medium or cells by the single radial immunodiffusion method of Mancini et al. (24). Incorporation of [3HJThymidine into DNA. After various times in culture, hepatocytes were pulse-labeled for 18-20 hr
Proc. Natl. Acad. Sci. USA 76 (1979)
with [3H]thymidine at a concentration of 1 ACi/ml of medium. The cells were collected at the indicated times, and the incorporated radioactivity was measured either directly in DNA purified on CsCl gradients (25) or by autoradiography of cell smears and serial cryostat sections of frozen cell pellets. The cell smears and serial sections were stained with hematoxylin and eosin or for y-glutamyl transpeptidase activity, and the percentage of hepatocytes undergoing DNA synthesis was determined from random counts of [3H]thymidine-labeled nuclei per thousand cells. RESULTS The hepatocytes formed a confluent monolayer-by 24-48 hr after plating and could be maintained in culture for up to 10-13 days. Phase-contrast microscopy of the hepatocytes in monolayer culture demonstrated their square or polygonal shape and their large vesicular nuclei with pronounced nucleoli, a granular cytoplasm, and sharply delineated cell borders (Fig. 1 inset). Both mononucleate and binucleate hepatocytes were observed in the monolayer, but no attempt was made to quantitate their proportions in either early or late cultures. The ultrastructure of hepatocytes in early and late cultures closely resembled that seen in vivo (Fig. 1A). Structures resembling bile canaliculi occurred between adjacent cells, and desmosomes and tight junctions were observed at the plasma membranes that extended laterally from the canaliculi. The cells possessed a well-developed Golgi apparatus and numerous mitochondria; they contained variable amounts of glycogen; and their rough endoplasmic reticulum was preserved and, in some cells, was quite extensive. Lipid droplets and autophagic vacuoles were also seen, particularly in the late cultures. In sections cut perpendicularly to the mesh, the hepatocytes were found to have a cuboidal shape, and cells in late culture (days 7-10) frequently exhibited a progressive buildup of microfilaments at their apical cell surface but not at their lateral cell surfaces nor at their substratum surface (Fig. 1B). Table 1 shows the production of serum proteins by hepatocytes on collagen gel/nylon meshes. Albumin release into the medium decreased with time, but immunoreactive "albumin" in the hepatocytes steadily increased from 0.08 + 0.02 (SD) ng/,ug of DNA on day 1 to 1.14 + 0.3 ng/,ug of DNA on day 9. We have not determined, as yet, whether the increased amounts of "albumin" in the hepatocytes over time represents albumin,
FIG. 1. Adult rat hepatocytes on collagen gel/nylon meshes. (A) Electron micrograph of hepatocytes at 10 days in culture. (X3100.) Arrow, bile canaliculus; g, Golgi apparatus; L, lipid droplet. (B) Electron micrograph of a binucleate hepatocyte in a 10-day culture sectioned perpendicular to the mesh. (X3100.) M, microfilament network beneath cell surface facing the medium. (Inset) Phase-contrast micrograph of an 8-day-old culture. (X100.) NM, nylon mesh fiber.
Cell Biology: Sirica et al.
Proc. Natl. Acad. Sci. USA 76 (1979)
Table 1. Production of rat serum proteins by adult hepatocytes on collagen gel/nylon meshes
9
Days in culture
Albumin
Transferrin
1 2 3 4 5
147.3 ± 6.0 131.9+15 138.6 ± 10 64.6 + 9.4 42.9*
8.3 i 0.4 11.8 1.1 17.1 + 3.0 17.1 ± 2.3 15.5* (14.8-16.2) 12.8 ± 0.5 12.1 ± 1.3 7.7 + 0.6
(35-51) 6 7 9
25.9 + 2.5 22.9 + 0.8 14.9 + 1.4
a,-Acid glycoprotein
0.180.16_ 0.14 0.120.10 0.080.06 0.040.02-
90% of the hepatocytes in culture exhibited a strong histochemical reaction for y-glutamyl transpeptidase activity. The enzyme activity was confined to the cytoplasm of the vast majority of the hepatocytes, but we also observed some hepatocytes that exhibited a strong histochemical reaction at their cell surface. Furthermore, both mononucleate and binucleate hepatocytes were positive for the enzyme activity. No enzyme activity was detected when the -y-glutamyl substrate was omitted from the reaction mixture or when L-leucyl-4-methoxy-2-naphthylamide was used as substrate, and the histochemical reaction was markedly decreased when glycylglycine was not included in the reaction mixture. Fig. 3 shows the production of a1-fetoprotein by the hepatocyte cultures. a1-Fetoprotein could not be detected by radioimmunoassay in the culture medium or in the hepatocytes collected at 24-hr intervals during the first 4 days of culture. However, significant levels of a1-fetoprotein were measurable in the medium and in the cells during the last 5 days of culture. A preliminary cytoimmunofluorescence study has also shown that at least 8.0% of the total number of cells contained alfetoprotein on day 8 in culture. The specific activities of fructose-bisphosphate aldolase and alkaline phosphatase in the hepatocytes in culture decreased over time (Table 2). As with normal liver, the aldolase of the cultured hepatocytes exhibited a fructose 1,6-bisphosphate/ fructose 1-phosphate specific activity ratio of 1 during the first 6 days of culture. However, at day 8 in culture, the substrate specificity of the aldolase was more like that of the fetal and neoplastic livers (1, 7). On the other hand, the alkaline phosphatase of the cultured hepatocytes responded like the normal liver enzyme with respect to its inhibition by L-homoarginine.
DNA Synthesis in Cultured Hepatocytes. The incorporation of [3H]thymidine into DNA as a function of the age of the hepatocyte culture is shown in Fig. 4. Autoradiography of cryostat sections and cell smears prepared from a day 8 culture demonstrated that between 6 and 10% of the y-glutamyl transpeptidase-positive cells exhibited strong nuclear labeling with [3H]thymidine. Similar labeling index curves for [3H]thymidine were obtained in a separate experiment from cell sections and smears stained with hematoxylin and eosin (not shown). Hydroxyurea (25 mM) inhibited the incorporation of [3H]thymidine into DNA by 95% in a day-8 culture but did not inhibit the 7y-glutamyl transpeptidase activity of the hepatocytes. Also, exposure of the hepatocytes in culture to 10 1AM
dexamethasone for 6 days significantly decreased the [3H]thymidine labeling index but not the number of cells exhibiting -y-glutamyl transpeptidase activity. The percentage of y-glutamyl transpeptidase-positive cells on day 8 in culture was 90.8 ± 8.6 (SD) for the controls and 97.6 i 2.2 for the dexamethasone-treated cultures, but the [3H]thymidine labeling index for the y-glutamyl transpeptidase-positive cells was decreased from 9.5 ± 4.5% for the controls to 0.3 + 0.5% for the cells cultured with dexamethasone. DISCUSSION The collagen gel/nylon mesh support system represents an alternative to the floating collagen gel (8) and Millipore filter (27) substrata and, as such, provides a number of advantages over these two as well as the more conventional plastic (10, 28) and collagen-coated plastic surfaces (29) used to support rat hepatocytes in primary cultures.§ Unlike the floating collagen gel, the collagen coated/nylon mesh does not shrink with time in culture, it is easier to handle than the thin flexible gels, and it allows for the rapid removal and separation of the hepatocytes from their collagen substratum with collagenase. The mesh system also permits direct microscopic observation of the cells in culture, and, like the floating collagenf gel and Millipore filter support systems, it extends the functional longevity of the hepatocytes in culture over those maintained under comparable conditions on plastic and collagen-coated plastic surfaces. Hepatocytes in early and late mesh cultures possessed a complex ultrastructure that was identical to that reported for adult rat hepatocytes supported on floating collagen gels (31). In this respect, the hepatocytes on the mesh exhibited morphological features that resembled to varying degrees those of both adult and fetal hepatocytes in vivo. The most dramatic ultrastructural change observed was the extensive buildup of microfilaments beneath the apical surface of the hepatocytes in late culture. A similar buildup of microfilaments at the cell periphery has been described for rat hepatocytes cultured on plastic surfaces (32, 33) but the complexity of the microfilament network was generally more pronounced in hepatocytes on the mesh than in those on plastic. The functional significance of this polarized accumulation of microfilaments in the cultured hepatocytes is unknown at present, but it is interesting that an increase in microfilaments has also been shown to occur in vivo
in carcinogen-induced preneoplastic hepatocytes (34), in § While this manuscript was in preparation, Cereijido et al. (30) described the preparation of a collagen-coated nylon cloth disk used to maintain an epithelial cell line (MDCK) in monolayer culture. Their system differs from the collagen gel/nylon mesh primarily in that 4% glutaraldehyde was used to promote crosslinking of the collagen covering the cloth disk.
Cell Biology: Sirica et al. hepatocytes during liver regeneration (35), and in fetal liver hepatocytes (36, 37). The most unique feature of the hepatocytes on the collagen gel/nylon mesh was their positive expression of fetal gene properties along with a concomitant increase in DNA synthesis. In this respect, the changes in y-glutamyl transpeptidase activity, a1-fetoprotein production, and aldolase substrate specificity simulated those exhibited by hepatocytes in vivo during various hepatoproliferative conditions. Recently, Leffert et al. (38) have independently described a liver cell culture system that, like the mesh system, displayed adult-to-fetal cell-like changes in apparent association with hepatoproliferative transitions. In comparison, the mesh/gel culture system exhibited kinetics for al-fetoprotein production similar to those found with this other system, and the alteration in the aldolase isozyme pattern of the hepatocytes on the mesh occurred at a time in culture similar to that shown by Leffert et al. for adult-to-fetal form changes in pyruvate kinase isozymes. However, in contrast to the findings of Leffert et al., we were not able to demonstrate "fetal-to-adult phenotypic reversal" or an increase in hepatocyte cell number and, despite the increased DNA synthesis, mitotic figures were observed infrequently in the mesh/gel cultures. Our results showing an inhibition of DNA synthesis by dexamethasone in culture without affecting the level of y-glutamyl transpeptidase further indicated that the expression of this fetal cell enzyme by the cultured hepatocytes may not be directly linked to an increased DNA synthesis but rather may be representative of an adaptive response by the hepatocytes to their artificial environment. Additional studies are needed to determine the necessary factors required to maintain an adult hepatocyte phenotype in culture as well as to define those that are needed to augment and maintain the proliferative activity of hepatocytes on their collagen gel/nylon substratum. We thank Mr. Gerald L. Sattler for his excellent technical assistance and Dr. James A. Gurr for performing the tyrosine aminotransferase assay. We also thank Dr. Ilse Riegel for her most helpful suggestions concerning the preparation of this manuscript. This study was supported by Grants CA-07175 and RO1-CA-17334-02 from the National Cancer Institute. A.E.S. is a recipient of a National Institutes of Health Postdoctoral Traineeship (T32-CA 09230-02) and W. R. is a recipient of a National Institutes of Health Postdoctoral Fellowship (F32-CA 05395-02). Y.T. is a visiting scientist supported by a grant from the Japan Society for the Promotion of Science. This work was presented in part at the 69th Annual Meeting of the American Association for Cancer Research and at the 62nd Annual Meeting of the Federation of American Societies for Experimental Biology. 1. Uriel, J. (1975) in Cancer, A Comprehensive Treatise, ed. Becker, F. F. (Plenum, New York), Vol. 3, pp. 21-55. 2. Walker, P. R. & Potter, V. R. (1972) Adv. Enzyme Regul. 10,
339-364. 3. Cheng, S., Nassar, K. & Levy, D. (1978) FEBS Lett. 85, 310312. 4. Fiala, S., Mohindru, A., Kettering, W. G., Fiala, A. E. & Morris, H. P. (1976) J. Natl. Cancer Inst. 57,591-598. 5. Tateishi, N., Higashi, T., Nomura, T., Naruse, A., Nakashima, K., Shiozaki, H. & Sakamoto, Y. (1976) Cann 67,215-222.
Proc. Nat!. Acad. Scd. USA 76 (1979)
287
6. Sell, S., Becker, F. F., Leffert, H. L. & Watabe, H. (1976) Cancer Res. 36,4239-4249. 7. Schapira, F., Reuber, M. D. & Hatzfeld, A. (1970) Biochem. Biophys. Res. Commun. 40,321-327. 8. Michalopoulos, G. & Pitot, H. C. (1975) Exp. Cell Res. 94, 7078. 9. Berry, N. M. & Friend, D. S. (1969) J. Cell Biol. 43,506-520. 10. Bonney, R. J., Becker, J. E., Walker, P. R. & Potter, V. R. (1974) In Vitro 9,399-413. 11. Sirica, A. E., Goldblatt, P. J. & McKelvy, J. F. (1975) J. Biol. Chem. 250, 6464-6468. 12. Rufo, M. B. & Fishman, W. 41972) J. Histochem. Cytochem. 20, 336-343. 13. Lebherz, H. G. & Rutter, W. J. (1975) Methods Enzymol. 42, 249-258. 14. Diamondstone, T. I. (1966) Anal. Biochem. 16,395-401. 15. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. (1951) J. Biol. Chem. 193,265-275. 16. Burton, K. (1956) Biochem. J. 62, 315-322. 17. Kissane, J. M. & Robins, E. (1959) J. Biol. Chem. 233, 184189. 18. Rutenburg, A. M., Kim, H., Fischbein, J. W., Hanker, J. S., Wasserkrug, H. L. & Seligman, A. M. (1969) J. Histochem. Cytochem. 17,517-526. 19. Watabe, H. (1974) Int. J. Cancer 13,377-388. 20. Schreiber, G., Rotermund, H. M., Maeno, H., Weigand, K. & Lesch, R. (1969) Eur. J. Biochem. 10, 355-361. 21. Boettcher, E. W., Kistler, P. & Nitschmann, Hs. (1958) Nature (London) 181, 490-491. 22. Kamiyama, S. & Schmid, K. (1962) Biochim. Biophys. Acta 58, 80-87. 23. Nishi, S. & Hirai, H. (1976) in Protides of the Biological Fluids, ed. Peeters, H. (Pergamon, Oxford), Vol. 23, pp. 303-307. 24. Mancini, G., Carbonara, A. 0. & Heremans, J. F. (1965) Immunochemistry 2,235-254. 25. Shaw, J. L., Blanco, J. & Mueller, G. C. (1975) Anal. Biochem. 65, 125-131. 26. Chung Lin, J., Lee, G., Sarma, D. S. R. & Farber, E. (1977) Fed. Proc. Fed. Am. Soc. Exp. Biol. 36, 1078. 27. Savage, C. R. & Bonney, R. J. (1978) Exp. Cell Res. 114,307315. 28. Bissel, D. M., Hammaker, L. E. & Meyer, U. A. (1973) J. Cell Biol. 59,722-734. 29. Kletzien, R. F., Pariza, M. W., Becker, J. E. & Potter, V. R. (1976) J. Biol. Chem. 251,3014-3020. 30. Cereijido, M., Robbins, E. S., Dolan, W. J., Rotunno, C. A. & Sabatini, D. D. (1978) J. Cell Biol. 77,853-880. 31. Sattler, C. A., Michalopoulos, G., Sattler, G. L. & Pitot, H. C. (1978) Cancer Res. 38, 1539-1549. 32. Bernaert, D., Wanson, J., Drochmans, P. & Popowski, A. (1977) J. Cell Biol. 74, 878-900. 33. Chapman, G. S., Jones, A. L., Meyer, U. A. & Bissell, D. M. (1973) J. Cell Biol. 59, 735-747. 34. Ogawa, K. & Medline, A. (1978) Proc. Am. Assoc. Cancer Res. 19,63. 35. Gabbiani, G. & Ryan, G. B. (1974) J. Submicrosc. Cytol. 6, 143-157. 36. Chedid, A. & Nair, V. (1974) Dev. Biol. 39,49-62. 37. Santerre, R. F. & Rich, A. (1976) Dev. Biol. 54, 1-12. 38. Leffert, H., Moran, T., Sell, S., Skelly, H. Ibsen, K., Mueller, M. & Arias, I. (1978) Proc. Nat!. Acad. Sci. USA 75, 1834-1838.
