Clonal variations in keratin

Experimental

Cell Research 156 (1985) 359-366

Clonal Variations Intermediate

LOREN

Filament

Expression

in Keratin

by Human

Somatic

Cell Hybrids

W. KNAPP,* W. MICHAEL O’GUIN, ROGER H. SAWYER, DIANE MITCHELL and CLIVE L. BUNN

Department

of Biology,

University

of South

Carolina,

Columbia,

SC 29208,

USA

The intermediate filament composition of differentiated vertebrate cells provides a stable phenotype which appears to be specifically regulated in each cell type. In order to analyse the regulation of intermediate filament expression we have constructed human somatic cell hybrids from the fusion of the HeLa-derived cell line HEB7A and a normal human diploid tibroblast, GM2291. These parental cells differ with respect to the presence or absence of keratin intermediate filaments. Isolation of independently arising clones produced two classes of hybrids. One class expresses keratin in a stable manner and the other class lacks keratin altogether. Indirect immunofluorescence of hybrid cells using antikeratin antiserum demonstrates that there are variations in the intensity and organization of cytoskeletal keratin staining. SDS-PAGE comparisons of cell extracts from these hybrids indicates that there are quantitative differences in the relative amounts of individual keratin polypeptides as well. These clonal variations have allowed us to begin assessing the consequences of genetic interactions between cell types that are normally capable of closely regulating different subsets of intermediate filament genes. 0 1985 Academic PKSS. IIIC.

Intermediate filaments (IF) comprise a major class of cytoskeletal elements found in vertebrate cells [ 11. Each of the live defined subclasses is developmentally regulated and establishes stable, restricted distributions in specific tissues and cell types of adults [14]. The phenotypic expression of these different, primary intermediate filaments in vivo is largely maintained in vitro. Thus, epithelial cells express keratin, neuronal cells express neurofilaments, glial cells express glial filaments and muscle cells express desmin [l, 51. The exception to this scheme is vimentin, which is restricted in vivo to cells in mesenchymal tissues, but is co-expressed in vitro with the primary intermediate filaments in most culture-adapted cell types, regardless of their origin [ 1, 61. The keratins are the most extensively studied of all intermediate filament types [l , &l 11. The presence of tissue-specific subsets of keratin polypeptides reflects differences in keratin gene expression in epithelial cells of intact tissue, cultured cells, and tissue recombinants [9, 12-161. One approach to the analysis of the control of intermediate filament expression and utilization is through the use of somatic cell hybrids in which each member of the pair of parental cells expresses different intermediate filaments reflective of their tissue of origin. Immunocytochemical localization of keratin in heterokar* TO whom offprint requests should be addressed. CopyrIght 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved 0014.482785 $03.00

360

Knupp et (11.

yons and hybrids has been performed on a limited number of epitheliai and nonepithelial cell combinations [ 17. IS]. Laurila et al. [ 171 have shown that keratin persists in heterokaryons made between human fibroblasts and extra-embryonic epithelial cells for up to 3 days after cell fusion. However, since immunocytochemical findings reported by Peehl & Stanbridge [18] for proliferating somatic cell hybrids formed between fibroblasts and HeLa cells suggested that the keratin expression normally observed in HeLa cells was suppressed by fusion with fibroblasts, further investigation of this phenomenon was needed. We now report that a detailed immunological and biochemical analysis of a number of independently isolated clones made between human diploid fibroblasts, which express vimentin but not keratin, and HeLa cells, which express keratin and vimentin, not only shows that keratin genes can be expressed in hybrids, but that stable quantitative and qualitative differences exist among the clones in which keratin is present. MATERIALS

AND

METHODS

Cell Lines and Culture Conditions The epithelial cell line HEB7A was used in these studies. HEB7A is a thymidine kinase-deficient derivative of the cervical carcinoma cell line HeLa [19]. The human diploid tibroblast cell line used was GM2291, obtained from the Human Genetic Mutant Cell Repository, Camden, N.J. These GM2291 cells were derived from a lung biopsy on a 19.week-old white male fetus with Lesch-Nyhan disease and contain the genetic marker of a deficiency in hypoxanthine phosphoribosyl transferase 1201. All conditions of cell culture and sub-culture, and chromosome counting methods, were as described ,by Bunn & Tat-rant [21].

Cell Fusion and Hybrid

Selection

Cells (5x10’ for each parent) were fused in suspension with 40% (w/v) polyethylene glycol dissolved in 20 mM HEPES buffer, pH 7.4 [21]. Hybrids were selected from the fusion mixture with the HAT selection regime of Littlefield [22]. The fusion mixture was inoculated into wells of 24-place Linbro trays (Flow Laboratories, McLean, Va) at IO’-IO4 total cells per well. After 12-24 days of incubation, wells containing only one hybrid colony were selected for subculture and analysis. Control cultures of 5x10' parental cells were inoculated alone into selective medium. The hybrid clones were maintained continuously in HAT medium.

Electrophoresis Horizontal starch gel electrophoresis was carried out as described [23], except that gels consisted of 12% starch in 2: I proportions of hydrolysed starch (Sigma Chemical Co.. St. Louis. MO.) to Electrostarch (Electrostarch Co., Madison, Wise.). Electrophoresis was carried out at 5 V/cm for 22 h at 4°C. Gels were stained for glucose-6-phosphate dehydrogenase (G6PD, EC 1.1.1.49) activity by standard procedures [23]. For SDS-polyacrylamide gel electrophoresis (SDS-PAGE), cytoskeleton-enriched polypeptide fractions were extracted from parental cell types and hybrid clones using a modification of the Triton X-IO& 1.5 M KCI method [24] for HeLa cells [25]. The extracted proteins were dissolved in sample buffer and separated on 10% polyacrylamide gels at a pH of 8.3 [12]. Gels were stained with Coomassie brilliant blue.

indirect

irnmunojluorescence

Subconfluent cultures of hybrid acetone (1 : I) for 5 min at -20°C. Exp Cell Res 156 (19851

Microscopy and parental Cells were

cells were fixed and permeabilized in methanol and air-dried. then rehydrated in Sorensen’s phosphate

Keratin

expression

in somatic

cell hybrids

361

buffer (pH 7.2-7.4). Rabbit antikeratin antiserum [26] was used to localize cytokeratin [27, 281. Cells were incubated in a 1 : 30 dilution of the antiserum at 37°C for 20-30 min. They were rinsed extensively with phosphate buffer and subsequently reacted with fluorescein-conjugated goat antirabbit IgG (Miles-Yeda) at 37°C for 20-30 min. Cells were rinsed thoroughly, mounted in 10% glycerol and examined by epifluorescence microscopy.

RESULTS The complementary recessive genetic markers of HEB7A and GM2291 allowed selection of hybrid cells in HAT medium [22]. Each hybrid culture was an independent isolate as it was picked from a well containing only one colony. Five hybrid clones out of a total of 19 were examined. These were chosen as being representative of the range of morphologies observed among hybrids. They were designated H-8, H-12, H-13, H-18 and H-19. The growth of these cells in HAT medium indicates that they are hybrids. Each hybrid cell was larger in size and different in morphology compared with parental cells. No cells grew from cultures of parental lines inoculated into selective medium. To further confirm the hybrid nature of these cells, extracts of each cell line were assayed for their G6PD isozymes in starch gel electrophoresis. HeLa cells possess the A+ variant form of G6PD, while GM2291 cells possess the B variant G6PD. The hybrid cell lines H-8, H-13, and H-18 possess the G6PD heteropolymer of intermediate mobility (data not shown). The presumptive hybrid H-19 has apparently lost the B form of G6PD which can occur in hybrids of this kind [29]. The rapid onset of senescence prevented collection of sufficient material for G6PD analysis of H-12. Chromosome counts showed that H-19 had 66-t3 chromosomes, compared with 53+2 for HEB7A and 46+0 for GM2291. Other hybrids from this fusion possess 73-90 chromosomes [21]. No cells with parental chromosome numbers were observed in chromosome preparations of hybrids. Cell line GM2291 was fused at population doubling level (PDL) [ 171, and in this laboratory, this line has a lifespan of 45-49 PDL. The hybrid cultures were examined for the presence of keratin IF and other cytoskeletal elements (below) after they had undergone 26-30 population doublings (PD) from fusion. The limited lifespan hybrid H-12 died after 34 PD from fusion. Hybrids H-8, H-13, H18 and H-19 were still proliferating actively 50 PD from fusion, and had apparently escaped senescence. SDS-PAGE analysis of extracts from parental cells and hybrids stained with Coomassie brilliant blue (fig. 1) demonstrate the presence of characteristic human keratin polypeptides [25, 281. The major keratin polypeptides in HEB7A and hybrids are represented in three distinct bands (M, 52,48 and 46 kD) bounded on the high side by vimentin (M, 57 kD) and on the low side by actin (M, 43 kD). This is the distribution of polypeptides typical for HeLa cells previously described by Franke et al. [25]. GM2291 libroblast extracts also contain vimentin and actin but do not have the keratin polypeptides (fig. 1, B). H-8, H-13, H-18 and E-~/J Cell Rrs

156 (1985)

362

Knapp et al

”

k a-

-

c

Fig. 1. SDS-PAGE pattern of cytoskeleton enriched cell cxt~-act~ stained with Coomassie brilliant blue. Triton X-100: 1.5 KCI insoluble polypeptide4 were prepared from cultures of A. HEB7A: B. GM2291; C, H-8; D, H-12; E, H-13; F. H-18: G. H-19 and run on a IO% gel. The major cytoskeletal polypeptides present are vimentin (u). keratins (k) and actin 14

H-19 all contain the characteristic keratin polypeptides (fig. 1, C, E-G). The H-12 clone does not (fig. 1, D). In those clones expressing keratin there are quantitative differences in the relative amounts of individual keratin polypeptides. There are also differences in the relative amounts of keratin and vimentin (fig. l), indicating variable, concurrent expression of both intermediate filament protein types. Indirect immunofluorescence microscopy of parental and hybrid cells demonstrates not only the presence or absence of keratin, but also variation in the intensity and distribution of cytoskeletal keratin staining in those cells producing it (fig. 2). The parental HEB7A cells, as well as hybrids H-8, H-13, H-18 and H-19 all clearly produce a keratin cytoskeleton (fig. 2). No cells in the parental GM2291 or in the H-12 hybrid cultures express keratin. Clonal variation of keratin in H-8, H-13, H-18 and H-19 was assessed visually on the basis of overall cytoskeletal organization and the intensity of the fluorescence. All cells in keratin-positive cultures demonstrated a keratin cytoskeleton. There is a range of morphological differences among the hybrid clones. H-19 cells closely resemble the parental HeLa cells (fig . 2A. B). In contrast H-12 cells (fig. 25) resemble the parental fibroblasts, though the hybrid cells are larger (fig. 2A, C, D, F, G, H). The other hybrids represent a range of intermediate Exp Cell Res 156 (1985)

Keratin

expression

in somatic cell hybrids

363

Fig. 2. Indirect immunofluorescence localization of keratin cytoskeleton in HEBj’A, GM2291 and hybrids derived from their fusion. Epifluorescence microscopy using anti-keratin antiserum was carried out on (A) H-19; (B), HEB7A; (C) H-13; (D) GM2291; (E) H-8; (G) H-18; (I) H-12. Phase microscopy of(F) H-8; (H) H-18; (J) H-12 is presented in conjunction with their immunofluorescently stained counterparts. Bar, 20 pm. (A,B,D,E, G,I) x480;(C) x75O;(F, H,J) x325.

364

Knapp et al.

morphologies. Based on immunocytochemical and biochemical criteria, there are three general categories of hybrid cells. In the first, cells have an abundant well organized keratin cytoskeleton and produce the major keratin polypeptides. This category includes H-8, H-18 and H-19. The morphology and polypeptide pattern of H-19 and HEB7A are remarkably similar. In the second category. cells have keratin, but it is generally less abundant and not as well organized. In addition, there are quantitative differences in individual polypeptides but all the typical keratins are present. This category includes H-13. In the third, the cells have no demonstrable keratin cytoskeleton by indirect immunofluorescence and the SDS-PAGE polypeptide pattern does not show the characteristic keratin bands. H-12 is representative of this category. DISCUSSION We have demonstrated by biochemical and immunocytochemical analysis that there is a wide range of expression and organization of keratin cytoskeletal elements in proliferating hybrids formed from the fusion of human fibroblasts (keratin-negative) with a transformed human cell of epithelial origin (keratinpositive). One source of this variation may be differential chromosome loss among these hybrids. Intraspecific hybrids made from one diploid and one aneuploid parent show slow (relative to interspecific hybrids) and variable chromosome loss over periods of several weeks to three months in culture. The extent of the loss over this period from a theoretical complement of 100-I 10 chromosomes (the sum of the two parental chromosome complements) has been reported as O-10 chromosomes [30], O-25 chromosomes [21] or O-20 chromosomes [31]. Variation may also occur as a result of gene rearrangements at the time of nuclear fusion or during subsequent hybrid cell proliferation. The wide variation in keratin expression from clone to clone suggests that several genes are responsible for the synthesis, organization and regulation of keratin cytoskeletal elements. The heterogeneity of keratin content and organization among hybrid clones may reflect cellular variation in keratin filament assembly. While the individual keratin polypeptides produced by these hybrids are always expressed together, there is considerable quantitative variation among individual polypeptides, as analysed by SDS-PAGE, which suggests that there may be coupled expression without coordinated regulation of these genes. Differences were also observed in the relative abundance of keratin and vimentin, both of which are constituents of intermediate filaments and coexist in all the keratinpositive hybrids examined. The continued expression of vimentin by all hybrid clones is not unexpected, since most cells in continuous culture (including the parental lines GM2291 and HEB7A) express this protein constitutively. Our results contrast with those of Peehl & Stanbridge [I81 who reported the loss of keratin filament expression in hybrid cells constructed from fibroblasts and HeLa (D98) cells, as detected by indirect immunofluorescence microscopy. Exs Cdl

Rrs 156 (1985)

Keratin

expression in somatic cell hybrids

365

The HeLa cells used have different modal chromosome numbers (53 for HEB7A and 61 for D98) which may account for this discrepancy. However, since D98 cells normally produce a keratin cytoskeleton, examination of a larger number of hybrids formed using these cells may result in clones that express keratin. Laurila et al. [17] demonstrated the retention of both vimentin and keratin filaments for 3 days in heterokaryons produced from the fusion of primary human epithelial cells and fibroblasts. In this case it is not clear whether genes are active, since changes in gene expression can be initially masked by low turnover rates of affected proteins, particularly in the relatively short existence of heterokaryons. In most of the uninuclear hybrids analysed here, the expression of keratin persists after nuclear fusion and subsequent cell proliferation, and may prove to be a valuable indication of long-term HeLa cell influences in the proliferation of these hybrids, which can show either limited lifespan or transformed growth patterns [21]. Although these hybrids were stable for keratin expression or extinction, reexpression of inactive genes can occur in somatic ceil hybrids [32]. These epithelialxfibroblast cell fusions provide a useful model system for studying intermediate filament gene regulation and the influence of one genome on another, which may be applicable to somatic cell hybrids of cells expressing other intermediate filament types (glial filaments, desmin filaments, neurofilaments). Examination of factors involved in regulating these classes of genes will provide insight into their specification during development, the stability of their expression in differentiated cells and their role(s) in the structure and function of cells and tissues. This work was supported by grant no. PCM83-09068 from NSF, grant no. 1 ROI HD18129-01 from NIH and grant no. AGO 2664 from NIA. We wish to thank Debra Chavis for typing this manuscript.

REFERENCES I. 2. 3. 4. 5. 6. 7. 8. 9. 10. II. 12. 13. 14. 15.

Lazarides, E, Ann rev biochem 51 (1982) 219. Sun. T-T, Shih, C & Green, H, Proc natl acad sci US 76(6) (1979) 2813. Paulin, D, Babinet, C, Weber, K & Osborn, M, Exp cell res 130 (1980) 297. Jackson, B B, Grund. C, Schmid, E, Borki, K, Franke. W & Illmensee, K, Differentiation 17 (1980) 161. Sun, T-T & Green, H, Cell 14 (1978) 469. Osborn, M, Franke, W & Weber, K, Exp cell res 125 (1980) 37. Fuchs, E & Green, H, Cell 15 (1979) 887. Steinert, P M, Idler, W W & Zimmerman, S B, J mol biol 108 (1976) 541. Sun, T-T, Eichner, R, Nelson, W G, Tseng, S C G, Weiss, R A, Jarvinen, M & WoodcockMitchell, J, J invest dermatol 81 (1983) 109. Winter, H, Schweizer, 3 & Goerttler, K, Carcinogenesis 1 (1980) 391. Franke, W, Weber, K, Osborn, M, Schmid, E & Freudenstein, C, Exp cell res 116 (1978) 429. O’Guin, W M & Sawyer, R H, Dev biol 89 (1982) 485. Sawyer, R H, O’Guin, W M & Knapp, L W. Dev biol 101 (1984) 8. Mall, R, Franke, W W, Schiller, D L, Geiger, B & Krepler, R, Cell 32 (1982) 11. Wu, Y-J, Parker, L M, Binder, N E, Beckett, M A, Sinard, J H, Grifftths. C T & Rheinwald, J G, Cell 31 (1982) 693. Exp Cell Res 156 (1985)

366 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

Knapp et ul. Tseng. S C G, Jarvinen. M J, Nelson. W G. Huang. J. Mitchell, J W & Sun. ‘I‘. Cell 30 (19%) 361. Laurila, P, V&men. I. Lehto. V-P, Vartio, T & Stenman. S. J cell biol 94 (1982) 308. Peehl, D M & Stanbridge, E J. Int j of cancer 17 (1981) 625. Wallace. D C. Bunn, C L & Eisenstadt. J M, J cell biol 67 (1975) 174. Tischfield. J. Schafer, I A. Dickerman, L H. Trill. J. Mulivor, R A. Greene. A E & Cornell. L L. Cytogenet cell genet 24 (1979) 199. Bunn. C L & Tarrant. G M. Exp cell res 127 (1980) 38.5. Littlefield, J W, Science I45 (1964) 709. Harris, H & Hopkinson. D A. Handbook of enzyme electrophoresis in human genetics. NorthHolland Publ. Co.. Amsterdam (1976). Franke, W, Schmid, E, Osborn. M & Weber, K. Proc natl acad sci US 7% IO) (1978) 5034. Franke, W, Schmid, E. Weber, K & Osborn, M, Exp cell res II8 (1979) 95. O’Guin, W M, Knapp, L W & Sawyer, R H, J exp zoo1 220 (1982) 371. Knapp, L W, O’Guin, W M & Sawyer, R H, Science 219 (1983) 501. - J cell biol 97 (1983) 1788. Migeon, B R, Norum, R A & Cosaro, C M, Proc natl acad sci US 71 (1974) 937. Bengtsson. B 0, Nabholz, M. Kennett, R, Bodmer. W F. Povey, S & Swallow. D, Somat cell genet I (197.5) 41. Stanbridge. E J, Der, C J. Doersen. C-J, Nishimi. R Y. Peehl, D M. Weissman, B E&Wilkinson. J E, Science 215 (1982) 252. MCvel-Ninio. M & Weiss, M C, J cell biol 90 (1981) 339.

Received Revised

April version

26. 1984 received

Exa Cd

Res 156 i/Y851

August

13. 1984