ESPERIMENTAI.

CELL

HESEARCH

195,

1-7

(1991)

Immortalization of Human Fetal Sinusoidal Liver Cells by Polyoma Virus Large T Antigen S. HERING,*.’ ‘(‘c>nfra/

lmtitutez

of Molr~ular

Biology, Ikpartmcnt

B. E. GRIFFIN,~

Robert-Riisslr-Strassr 10, Rwlin, I 11.5, Grmon,s; Du Cane Road, London W 12 ONN.

of Virology,

Fetal sinusoidal liver cells were isolated from human liver explant cultures and transfected with pCMVLT, a plasmid containing the immediate early promotor of cytomegalovirus (CMV) and the large tumor antigen (LT) coding part of the polyoma virus (py) genome. Whereas nontransfected cells stopped proliferating after 4 weeks, the transfected sinusoidal cells were stimulated to divide more quickly without changes in their morphology. Up to now, cells have been permanently cultured for more than I8 months and passaged over 130 times, corresponding to around 400 generations. This allows them to be regarded as “immortalized” cells. The presence of LT protein in the cells has been documented by means of immunoprecipitation and immunofluorescence. Expression of the v. Willebrandt factor VIII was the main criterion for classifying the cell population as endot helial cells. The presence of cytokeratins 7, 8, and 18 in these cells underlines their close ontogenie and functional relationship to mesothelial cells. Sinusoidal endothelial cells (SECs) synthesize vimentin and the typical extracellular matrix components collagen IV and fibronectin, but are negative for laminin and entactin. We used immortalized SEC’s in co-culture experiments with fresh fetal human hepatocytes and adult mouse hepatocytes. They promoted survival of both types of hepatocytes over a period of S-10 weeks. Control human fetal liver explant culture cells survived for only 3-4 weeks, whereas control adult mouse liver cells c 199 I Academic retained vitality for 8-10 days only. Press.

AND M. STRAUSS*.’

Inc.

INTRODUCTION

A normal contribution of nonparenchymal liver cells (mainly sinusoidal endothelial cells and lipocytes) to the metabolical and morphological network of liver seems to play an essential part in the maintainence of

I Present address: Fresenius GmhH, Oherursel, Germany. ’ To whom correspondence and reprint requests should dressed at Zentralinst fur Molekularbiologie, Robert-Riissle-Str. Berlin-Ruth. II-11 15, F.R.G.

be ad10,

and tRoya Postgraduate 1 :nitcd Kingdom

Medical

School,

healthy liver function. Sinusoidal endothelial cells (SECs) provide contact between hepatocytes and portal blood in the smallest liver capillaries-the sinusoids. Together with lipocytes, the SEC produces the components of the extracellular matrix for embedding hepatocytes [l-4]. Recently, an hepatocyte growth factor (HGF) [5], isolated from liver tissue, was identified as a product not of hepatocytes, but of nonparenchymal liver cells, presumably SECs [6]. In addibion to humoral factors such as HGF, growth-promoting effects of SEC on hepatocytes are obviously exerted via direct celllcell contact, inasmuch as they could not be replaced by conditioned media in vitro [7]. Taken together, these properties of SECs designate them as essential partners for longterm hepatocyte culture: SECs provide the hepatospecifit extracellular matrix needed for hepatocytes to attach to culture vessels without loosing their diff’erentiated features, they secrete HGF, stimulating DNA synthesis and proliferation, and they probably support the generation of three-dimensional structures and vascularization during organoid formation in vitro [8]. Thus, to generate continuous SEC lines is an essential step in the development of normal liver cell cultures. Here we describe immortalization of human fetal SECs by means of polyoma virus large tumor antigen (py LT). The LT gene, when under control of a strong promotor, has been shown to induce an immortal phenotype not only in rodent but also in human cells [9, lo]. Unlike the situation with SV40 LT expression, the presence of py LT in human cells is virtually not accompanied by transformation and loss of differentiation. We regard this SEC line as a useful precondition for studying aspects of liver cell regeneration and differentiation in long-term culture. MATERIALS

AND

METHODS

Human fetal liter. For isolation of SEC, abortion material (20th week) was used. Abortion was performed as medically indicated in the Queen Charlotte’s Hospital, London. Further human fetal liver for coculture experiments was obtained from the Tissue Rank, Royal Marsdon Hospital, London. All experiments were carried out with permission of the Medical Ethics Committee. Hammersmith and Queen Charlotte’s Hospital.

2

HERING,

GRIFFIN,

Explant cultures of fetal liuer; isolation of SEC9 and other hoer c&. Liver fragments were stored on ice until use, for periods not extending 3 h. Tissue was washed twice in Versene, transferred into a fresh petri dish with Dulbecco’s modified Eagles medium (DMEM), and cut with two scalpels into very small fragments (max l-2 mm”). These were washed with DMEM, placed in petri dishes, and covered slightly with liver cell media (see below). After 2-3 days, outgrowth of hepatocytes, lipocytes, and endothelial cells was observed. Transfection with pCMVLT2 was performed, as described, on cells that had been in culture for 3 weeks. Different cell types could be isolated thereafter by means of locally restricted trypsination. This procedure allowed pure populations of hepatocytes, lipocytes, and SECs to be passaged. Liver cell medium. Liver cell medium was a 1:l mixture of DMEM (Flow) and Ham’s F12 (GIBCO) complemented with 6 or 10% FCS (Flow or GIBCO), 1 pg/ml insulin, 10m6 M dexamethasone, 5 pg/ml hydrocortisone, 1 rig/ml EGF, without antibiotics. In some cases, serum-free chemically defined medium, MX 83, was used. This medium was designed for hepatocytes and specifically suppresses growth of other contaminating cells when no arginine is added 1111, Indirect immunofluorescence (IE’). This assay was performed as described earlier 1121. Briefly, cells were grown on Flow multiwell slides, washed in PBS at 37”, fixed either in ice-cold acetone for 20 s (for detection of cytokeratins) or in 3.8% formaldehyde for 5 min at room temperature (for all other antigens). The acetone-fixed slides could be stored in dry atmosphere until covered with the first antibody. Formalin-fixed slides were kept in a moist atmosphere until use. They were permeabilized with l?;, Triton for 5 min, washed, preabsorbed with 0.25%’ BSA/O.25% gelatin in PBS for 1 h. In both variants, the first antibody was left on the samples for 1 h at 37”; slides were washed four times in PBS; then the corresponding FITClabeled second antibody was added and incubated for 1 h at 4°C; and slides were washed another four times in PBS, covered with PBS:glycerin, 1:l. mounted, and examined under a Leitz fluorescence microscope. Antibodies. For detection of v.Willebrandt factor VIII, polyclonal sheep anti-human antibodies (Serotec) were used, and for collagen I\ a monoclonal mouse antibody (PHM 12, Serotec) was employed. Fibronectin was visualized by means of the mouse monoclonal antibody HFN-llF6 (Serotec). Monoclonal mouse antibodies against cytokeratins 7,8, 18, 19, and vimentin were kindly provided by Dr. J. Bartek, ICRF, London [13]. Rat monoclonals against polyoma virus large T were obtained from Dr. S. Dilworth, R.P.M.S., London [14]. FITC-labeled antibodies were correspondingly rabbit anti-sheep antibodies, or Fab-fragments of sheep anti-mouse antibodies or rabbit anti-rat antibodies (all from Sigma). 7’ransfection of SE(‘s ubith pCMVL7’. The recombinant DNA clone, pCMVLT has been described earlier [lo]. Transfer of plasmid DNA into SECs was performed by calcium phosphate precipitation according to a standard procedure [15]. Briefly, 6 g’g DNA was mixed with 250 ~1 2~ HEBS buffer and water to make up 475 ~1, thereafter 25 ~1 2.5 M CaCl, was added slowly. The mixture was let stand for lo-20 min and evenly distributed over the cell monolayer. Medium was changed after 4 h. Immunoprecipitation. Immunoprecipitation was performed as an optimized procedure for the precipitation of polyoma virus tumor antigens, as described earlier [16].

RESULTS

Isolation and immortalization of SEC’s. Three weeks after starting explant cultures from fetal human liver, a massive outgrowth of different cell types around the liver tissue could be observed (Fig. la). At this stage,

AND

STRAUSS

FIG. 1. Cultured human liver cells. (a) Explant culture (Day 4) from fetal human liver fragments (x40). (b) Sinusoidal endothelial cells (SECs) derived from fetal human liver cells after transfection (X40).

transfection with pCMVLT was performed. After a further 2 weeks, the cells were vitally growing as individual islands, allowing a clear differentiation between hepatocytes, lipocytes, and SECs to be made by morphological criteria. All three cell types were separated from each other by local trypsination and grown further. The results observed were as follows: Proliferation of hepatocytes ceased between 10 and 12 weeks after starting the culture, but intact cell structure and cytokeratin-expression were retained for a further 4 weeks. Lipocytes proliferated very quickly during the first five to six passages, simultaneously loosing their incorporated fat droplets. After 2 weeks, lipocytes gradually acquired a slower growth rate requiring passaging once in 2 weeks only. Lipocytes have to date been cloned and cultured for a period of more than 18 months, reaching a passage number of 57. Compared to lipocytes, SECs had a much higher proliferative capacity; they have been passaged more than 130 times over a period of 18 month. Nontransfected control cultures from all cell types initially

IMMORTALIZATION

TABLE

Cell

type

ransfected: liver cells, explant culture pCMVI,T transfected Hepatocytes SEC’ Lipocytes

Survival culture (weeks)

OF

HUMAN

1

in Number passages

of

Number generations

of Cloning

Nont

45

nune

n.d.

no

16 90 90

2 149 56

n.d. 450” 150”

IlO no yes

’ Continued.

outgrowing from liver explants stopped growing after 4-5 weeks. The data are summarized in Table 1. Characterization of SEC. By light microscopic investigation, the appearance of SECs was variable, depending on cell culture media and on presence or absence of hepatocytes in culture. When the usual liver cell media was used, cells obtained a spindle-like, seemingly less-differentiated phenotype (Fig. lb). In the chemically defined media, MX 83 [ 111, and in coculture with hepatocytes, cells became round or oval-shaped with very thin long and branched cytoplasmic extensions, sometimes forming complete or incomplete “rings,” resembling cross-sections of capillaries. Indirect immunofluorescence assays (IF) were used for detection of characteristic patterns of endothelial cells as well as of py LT expression in nuclei of immortalized SECs (ISEC). The endothelial nature of SECs was verified by a homogeneous v.Willebrandt factor VIII expression in cytoplasma (Fig. 2a). iSEC were positive for collagen IV (Fig. 2b), fibronectin (Fig. 2c), vimentin (Fig. 2d), and cytokeratins 7, 8 (Fig. 2e) and 18 (Fig. 2f). They were negative for laminin and entactin expression (not shown). Appearance of actin cables was more faint than in nonimmortalized fibroblastoid cells (Figs. 3a and 3b). Due to poor growth of sinusoidal cells in the primary explant culture no direct comparison with the nonimmortalized cells could be done. Expression of py LT in SEC was revealed by means of IF and immunoprecipitation (IP), using a mixture of three anti-LT monoclonal antibodies, LTl, LT4, and LT7 [ 141. py LT could be visualized in the nuclei of 100% of the iSEC even after being cultured for longer than 1 year (Fig. 4a). In immunoprecipitation experiments, the LT protein was observed at 98 kDa as one of the major bands brought down with the anti-LT antibodies (Fig. 4b). Cloning of iSEC by limiting dilution and growth in low serum. Four experiments aimed at cloning iSEC by limiting dilution were unsuccessful; it did not prove pos-

FETAL

3

SINIJSOIDAL

sible to grow clones derived from a single cell. At low density, e.g., 312 cells per ml, proliferation of iSEC also ceased. Cells in these circumstances underwent an unusual expansion in size but did not divide. A comparable situation was observed when cells were passaged in media containing 0.5% FCS (unless chemically defined medium MX 83 was used). The lowest density at which normal growth and morphology of iSEC could be detected was about 5 X l@‘/ml (Fig. 5) in DMEM/HAM’s F12 with a serum content of 4%. Coculture of freshly isolated hepatocytes with immortalized SECs. Human fetal liver tissue was treated as described above. Tissue explants were set out either on naked TC petri dishes (Nunc) or on top of semiconfluent iSEC cultures previously seeded onto the dishes. In both cases, cell outgrowth from explants occurred within 2-3 days. In the pure explant culture, cells started to die after 4 weeks and explants detached from the petri dish. In cocultures, hepatocytes were growing island-like between iSEC for a minimum of 12 weeks (Fig. 6a). It was possible to passage hepatocytes together with iSEC two times. Analogous experiments were performed with adult mouse hepatocytes as partners for coculture with SEC. Following isolation via in situ perfusion with collagenase [17], mouse hepatocytes were seeded on semiconfluent iSEC or, alternatively, as usual on petri dishes. Whereas hepatocytes in the first variant created organized structures in intimate contact with iSEC (Fig. 6b), the control hepatocytes attached poorly and did not form compact structures. Cocultured mouse hepatocytes were kept in culture for 8 weeks (until overgrowth from iSEC occurred). Control hepatocytes, on the contrary, retained their healthy appearance for 10 days only. It was of interest to analyze the growth-stimulating potential of conditioned medium from either iSEC or other cells immortalized by polyoma viral LT like HEL cells [lo]. Experiments to test this resulted in negative results or nonsignificant effects even with IO-fold-concentrated supernatants (data not shown). Also, experiments to use immortalized HEL cells as feeder layers failed, pointing at a specific role of SEC in supporting hepatocyte growth. Experiments are in progress using iSEC as a specific “feeder layer” for continuous growth of hepatocytes following growth arrest of iSEC by mitomycin C.

DISCUSSION

Fetal SECs have been isolated from human liver explant cultures. They could be passaged continuously over a period of more than 18 months after being transfected with a plasmid containing the polyoma virus

4

FIG. 2. Expression of’ markers in SEC Cc) fihronectin; Cd) vimentin; (e) cytokeratins

HERING.

GRIFFIN.

AND

revealed by indirect immunolluorescence 7, 8; and (f) cytokeratin 18.

large T antigen coding part. The prolonged life span of the SEC coincides with expression of LT protein in these cells, as revealed by immunoprecipitation and nuclear immunofluorescence. As the cells have been passaged more than 130 times or over at least 400 generations, we regard them as an established immortalized cell line. The line (iSEC) is not of monoclonal origin, however, as so far all experiments aimed at the generation of clones from single cells have been unsuccessful.

STRAlJSS

assays

f’or (a) v. Willehrandt

f’actor

VIII;

(h) collagen

IV;

We interpret this behavior as a marker for the nontransformed, differentiated nature of these cells and the fact that they are strongly dependent on autocrine growth factor(s), as, for example, bFGF (basic fibroblastic growth factor). bFGF is a candidate for autocrine growth stimulation of SECs because it has been shown that endothelial cells both produce [18] and respond [19] to it. Another explanation for the difficulties in growing separate colonies might be the influence of a “scatter”

IMMORTALIZATION

OF

HIJMAN

FETAL

5

SINUSOIDAL

t,wo types: vimentin--typical for rnesenchymal elements-and cytokeratins-a distinctive feature of epithelial cells. This pattern of markers, to our knowledge,

has not been described before for SECs. Cytokeratin expression is not unusual, however, for cells of mesenchymal origin. Mesothelial cells are known to express cytokeratins 7,8,18,19. In our tests, iSEC bound monoclonal antibodies against cytokeratins 7,8, and 18. Since

FIG. 3. Fluorescence patterns of actin cables (stained TRITC-conjugated phalloidin) (a) iSEC; (h) nonimmortalized hroblasts.

with rat ti-

factor, known to cause migration and spreading of endothelial and epithelial cells [20]. This factor has been described to be a product of mesenchymal cells. Thus, SECs themselves or small contaminations of another sinusoidal cell type, Ito-cells (lipocytesl, could be a source of scatter factor, preventing growth of compact colonies in cloning experiments. Even if not clonal, nearly 100% of the iSEC express a characteristic marker of endothelial cells, the v.Willebrandt factor VIII [al], earlier described to be expressed in cultured rat SEC [22]. Whereas in the former cells the factor is localized in Weibel-Pallade bodies it appears difIused in the cytoplasm which might be characteristic for the embryonal human SEC. With respect to earlier findings [4], we confirm the synthesis of basement membrane components collagen IV and fibronectin by SECs. The iSEC do not synthesize laminin and entactin, which are other typical components of hepatic extracellular matrix. These results are in agreement with several reports characterizing Itocells as the source of laminin and entactin [23]. The intermediate filaments of iSEC are composed of

- 9 2.4

-

co

S8C

anti-

LT

45

4b

FIG. 4. Detection of polyorna large T antigen (pyLT) in iSIX. (a) Immunofluorescence. (1~) Immunoprecipitation from ““S-methionine labeled cells. (Arrow r 9%kDa band corresponding to LT protein; co, population of nontransfected liver cells as control: set, I,T immortalized sinusoidal endothelial cells).

20000

FIG.

5.

Growth

kinetics

of iSEC

depending

on the numbers

of’ cells seeded

the described iSEC are of fetal origin (20th week), it is necessary to question, whether this cytokeratin status marks a specific stage in liver differentiation, whether the differentiation process has been interrupted by cultivation, or whether pyLT expression and immortalization interfere with this process. The expression of the markers was consistent throughout the period of cultivation, which is further evidence for the stable maintenance of a nontransformed phenotype. Moreover, no focus formation or disorientated growth was observed. These data are in agreement with the immortalizing but nontransforming properties of polyoma viral large T antigen [lo, 251 as opposed to the effects of SV40 large T antigen. Permanently growing, functionally active but nontransformed SECs provide a crucial component in longterm culture of freshly isolated hepatocytes. The absence of a growth-stimulating effect of the conditioned medium of iSEC is in accordance with previous data with primary SEC [7] and together with the inability of immortalized fibroblasts to support hepatocyte growth it suggests that cell-to-cell contacts between SEC and hepatocytes are essential. It seems also very likely that immortalization of SEC by LT does not exert a specific effect on hepatocyte growth. However, it allows the SEC to grow continuously. Cocultures such as those described herein allow one to study the interrelationship between different hepatic cell populations and provide a necessary precondition for growing liver organoids in L&-O. Our findings further extend earlier studies [lo] which demonstrate the use-

FIG.

Fetal trast,

(0, 312; 0, 625; A, 1250;

. 2500;

t. moo;

x, 10,000)

6. &cultures of heterologous hepatocytes with SEC (a) human hepatocytes; (b) adult mouse hepatocytes; (phase con40x; S, SEC: H, hepatocytes).

IMMORTALIZATION

OF

HUMAN

fulness of the polyoma virus large T antigen in prolonging the life span of human cells in culture.

1%

REFERENCES 1. “.

Martinez-Hernandez, A. (1984) Rojkind, M., and Ponce-Noyola, 151-175.

3.

De Leeuw,

A. M., McCarthy,

14.

Lab. Inuest. 51, 57-74. P. (1982) CollagenRelat.

S. P., Geerto,

Res. 2,

A., and Knook,

D. L.

(1984)Ifeppatology 4, X92-403. 4. (‘lement, (:uillouzo, M.,

T., Nishizawa, Sugimura, A., Tashiro,

342,440-44:1. 6. Kinashita, T., Tashiro, Biophy.

16.

B., Grimaud, J. A., Champion, J.-P., A. (1986) Hepatology 8, 225-234.

5. Nakamura,

Rrs. Commun.

Deugnier,

Y., and

T., Hagiya, M., Seki, T., Shimonishi, Ii., and Shim& S. (1989) Nature

K., and Nakamura, 165, 1229-1234.

T. (1989)

Hiochem.

7. Shimaoka, S.. Nakamura. T., and Ichihara, A. (1987) E:xp. Cc11 Rcs. 172, '2"s%%4%. 8. Thompson, J. A., Haudenschild, C. C., Anderson, K. D., DiPietro, J. M., Anderson, Acad. Sci. 1 TkyA 86,

W. F.. and 7928-7932.

Maciag,

T. (1989)

Proc.

Natl.

9. Strauss,

M., Liibbe, L., Kiessling, II., Platzer, M., and Griffin, R. E. (7989) in Current Topics in Microbiology and Immunology, Vol. 44, pp. 129-134, Springer-Verlag, Berlin.

10. Strauss, M., IHering, S., Lubbe, c0,qrnc 5, 122'3-1129. Il.

Hoffmann,

B., Piasecki,

L., and Griffin,

B. E. (1990)

On-

D. (1989)

J. Cell. Physiol.

Dilworth, Streuli,

Received Revised

S. M., Hansson, H.-A., Darnfors, C. S.. and Griffin, B. E. (1986) EMBOJ.

October 31, 1990 version received January

24, 1991

C., Bjursell, 5, 491-499.

G.,

7

SINUSOIDAL

Bartek, J., Kovarik, J., Burchell, ,J., Taylor-Papadimitriou, J.. Voitesek, B., Rejthar, J., Schneider, ,J., Petrek, M., Straskova, Z., and Millis, R. (1987) in Molecular Biology and Differentiation of’ Cancer Cells (Lapis, K., and Eckhardt. S., Eds.), p. 123. Akademiai Kiado, Budapest. Dilworth, S. M., and Griffin, B. E. (1982) Proc. Natl. Acnd. Sci. L’SA 79 , 1059-1063. c Graham, F. L., Bacchetti, S., McKinnon. R., Stanners, C., Cordell, B., and Goodman, H. M. (1980) in Introduction of Macromolecules into Viable Cells, pp. 3-25, A. R. Liss, New York. Harlow, E., Franza, B. R., and Schley. C. 11985) J. Virol. 55,

533-546. 17. Berry, M. N., and Friend. 18.

Rosenbaum, J., Mavier, (1989) Biochcm. Biophys.

19.

Lampugnani, A., Dejana,

384. 20. Rosen, berg,

D. S. (1969)

J. Cell Hiol. 43, 506-520.

P., Preaux, A.-M., and Dhumeaux, Kes. Commun. 164, 1099-I 104.

D.

M. G., Polentarutti, N., Pedenovi, M., Mantovani, E., and Colotta, F. (1990) Exp. (‘ell Res. 186, 381L

E. M., Meromski, I. D. (1990) I&p.

l,., Setter, E., Vinter, Ceil Res. 186, 22-31.

D. W., and Gold-

21. ,Jaffe, E. A., Hoyer. L. W., and Nachman. R. L. (1973) J. C’lin. Inumt. 52 .- ‘1757-2764. t 22. Hellman, L., Smedsrbd, B., Sandberg, H.. and Peterson, IT. (1989)

23. Maher, (1988)

24. Quinlan,

A.. and Paul,

139 . f 6S4-662. 1%.

15.

FETAL

Moll, (1985)

25. Strauss, Review

Br. J. Haematol.

73,

J. J., Friedman, (:a.strf~~~~ter~/l~~~~

5. L., Roll, 94, 1053-1062.

348-355. F. ,J., and

R. A., Schiller, D. L., Hatzfeld, R., .Jorcano, J. L., Margin, T. M., Ann. NY Acad. Sci 455, 282-306. M., and Griffin, 2, :360-3&T,.

1~. E. (1990)

Bissell,

D. M.

M., Achtstitter, T., and Franke, W. W.

Cancer

Cells.

A Monthly

Immortalization of human fetal sinusoidal liver cells by polyoma virus large T antigen.

Fetal sinusoidal liver cells were isolated from human liver explant cultures and transfected with pCMVLT, a plasmid containing the immediate early pro...
8MB Sizes 0 Downloads 0 Views