Isolation and characterization of rat-mouse somatic cell hybrids secreting growth hormone and prolactin

Experimental

Cell Research 162 (1986) 475-485

Isolation and Characterization Hybrids Secreting Growth

Department

SHLOMO of Medicine,

MELMED*

of Rat-Mouse Somatic Cell Hormone and Prolactin and JAMES

A. FAGIN

Cedars-Sinai Medical Center, UCLA Los Angeles, CA 90048, USA

School

of Medicine,

Interspecific somatic cell hybrid clones have been isolated and characterized in order to study growth hormone (GH) and prolactin (PRL) gene expression. Rat pituitary tumor cells (GHs, 69 chromosomes) secreting rat GH and PRL were grown for 48 h together with nonhormone secreting, aminopterin-sensitive murine tibroblast cells (LMTK-, 55 chromosomes) and fused using polyethylene glycol. Resultant heterokaryons were selected in hypoxanthineaminopterin-thymidine (HAT) medium and cloned. Five clones produced rat GH and PRL. Hormone-producing hybrids morphologically resembled the mouse parent tibroblast. Hybrids grew in monolayers and contained 80-142 chromosomes, and marker chromosomes for both rat (small submetacentric) and mouse (bi-armed and large true metacentric) were identified. The interspecific nature of the hybrids was further confiumed by the presence of both rat and mouse adenosine deaminase and superoxide dismutase isozymes. Using specific antisera and indirect immunoperoxidase staining, both hybrid clones and GHs rat parental cells stained positively for rat GH and PRL, white the murine tibroblast parental cells were negative. Hormone production by the hybrids has been sustained for over twenty subcultures; secretion rates were initially 150 ng PRL and 321 ng GH/lo6 cells/24 h and are currently 100 ng PRL and 90 ng GHllod cells/24 h. Parental GH3 rat cells secreted 720 ng PRL and 660 ng GWlod cells/24 h. Exposure of hybrids to KC1 (50 mM) resulted in acute stimulation of rat PRL, but not rat GH release, and long-term incubation with thyrotropin-releasing hormone (TRH, 80 nM) stimulated PRL secretion. Hormone-dependent modulation of PRL secretion was transferred to the hybrid cell thus enabling the model to be used in studying regulation of PRL gene eXpreSSiOn.

@ 1986 Academic

Ress,

Inc.

Somatic cell hybridization has been extensively used in many cell systems in order to study nuclear and cytoplasmic components of differentiated cell function [l, 21. The heterokaryon cell resulting from the fusion of plasma membranes of two cells contains nuclear material from both parent cells. Growth and division of heterokaryons results in hybrid cells containing two types of genetic material in a single nucleus. In order to isolate the hybrid cells from the parental cells, an effective selection procedure must be available. Furthermore, either one or both parental cells should exhibit proliferative capacity in order to ensure stable gene expression and allow study of gene-regulatory processes. Although somatic cell hybrids have been extensively used to study mechanisms of action of glucocorticoid hormones [31, application of the technique to hormone-secreting cells has * To whom offprint requests should be sent. Address: Division of Endocrinology, Cedars-Sinai Medical Center, 8700 Beverly Blvd, Los Angeles, CA 90048, USA. 31-868332

Room 1735,

Copyright @ 1986 by Academic Press, Inc. All rights of reproduction in any form reserved 0014-4827/86 $03.00

476 Melmed

and Fagin

been limited. When rat pituitary tumor cells (GH3, 69 chromosomes) were fused with Sendai virus [4] or allowed to fuse spontaneously with mouse tibroblasts [5], the resultant hybrids did not produce GH, while only a single subclone secreted prolactin (PRL) [4]. Successful fusion of hCG-producing JEG-3 choriocarcinoma cells with Cl-ld mouse cells resulted in hybrid cells which produced hCG [6] and facilitated identification of chromosomes 10 and 18 as necessary for hCG expression [7]. Human-murine cell hybrids producing human immunoreactive insulin further confirmed the application of the technique to endocrine cells [8]. Recently, the dominant expression of human PRL secretion has been reported in human-mouse cell hybrids isolated by us [9]. In this report, interspecific somatic cell fusion is described between a functional GH3 and mouse tibroblast cell line. The aminopterin-sensitive murine fibroblast (LMTK-) was chosen, as it is deficient in enzyme thymidine kinase. These cells die in the presence of aminopterin, which inhibits the de novo synthesis of both purines and pyrimidines [lo]. Hybrid cells, having obtained the thymidine kinase gene from a thymidine kinase-positive parent will survive in the presence of aminopterin and thymidine [2, lo]. GH3 cells [ll] secreting rat growth hormone (GH) and PRL were used as the wild-type rat parent cells. This relatively stable well differentiated pituitary tumor cell line has been shown to lack hypoxanthine-guanine-phosphoribosyltransferase (HGPRT) activity [ 121and therefore also to be aminopterin-sensitive and thus used here to study transfer of a differentiated cell function into an inter-specific hybrid cell. The two parental cells were fused using polyethylene glycol (PEG) and the resultant hybrids have sustained secretion of rat GH and PRL and have also retained hormone-dependent modulation of PRL secretion. MATERIALS

AND METHODS

Cells GHs cells, a cloned line of rat pituitary tumor cells secreting rat GH and PKL [13] were obtained from the American Type Culture Colkction, Bethesda, Md. Cells were grown in monolayer culture in a humidified atmosphere of 95% air+5% CC&, using Ham’s F 10 medium, horse serum (15 %) and fetal calf serum (FCS) (2.5%), as previously described [14]. Cells used for these experiments have undergone approx. 30 subcultures in this laboratory. These cells are deficient in hypoxanthine phosphoribosyl transferase as evidenced by their resistance to dthioguanine and their sensitivity to hypoxanthine aminopterin and thymidine (HAT) medium [12]. Mutant mouse fibrobhust cells wre used as the second parent. The thymidine kinasedeticient murine fibroblast cell line (LMTK-) was hindly provided by Dr R. E. K. Foumier, Department of Microbiology and the Comprehensive Cancer Center, University of Southern California School of Medicine, Los Aqples, Calif. These ceils were maintained as monolayers in Dulbecco’s Modified Eagle Medium (DME) supplemented with FCS (10%).

Cell Fusion Fusion of the two ceU types was achieved using polyethykneglycol (PEG, MW 1000, Batter Co.) [14]. GHs cells and LMTKcells growing in mondaytr culture were trypsinizad and approx. Id ceils of each parent cell type were seeded in 25 cm2 tissue cukure Baahs (Falcon, Gxnard. C&f.). The two cell types grew rapidly in DME suppkmented with FCS (10%) and confhmnce was achieved Exp Cell Res 162 (1986)

Hormone-secreting

cell hybrids

477

within 48 h. The medium was aspirated and the mixed monolayer of cells washed twice with warm (37°C) Hanks balanced salt solution. Two ml of warmed (37’C) 50% (w : w) polyethylene glycol [IS] dissolved in Hanks buffer @H 7.6) was then gently pipetted over the monolayer. After 60 set PEG was aspirated and the cells washed twice with 2 ml of Hanks buffer (37°C). The cells were then incubated at 37°C in 5 ml selection medium: Ham’s FlO medium containing hypoxanthine (0.1 mM), aminopterin (0.4 l&f), thymidine (16 @I), FCS (10%) and antibiotics (HAT medium) [lo]. Forty-eight hours later, medium and floating cells were aspirated and fused cells were noted. Cells were incubated with HAT medium for a further 14 days.

Cloning Procedure Large fused cells (HAT-resistant) were easily distinguished morphologically. As the mouse TKcells and rat GHs cells were HAT-sensitive, they did not survive in the selection medium. The location of the hybrid clones was marked. The tops of the flasks were removed and cells isolated with a cloning cylinder [17]. Isolated cells were trypsinized and transferred to 48-well multiwell tissue culture plates at a density of O-2 cells/well. HAT medium (1 ml) was added to each well and cells were incubated for a further 2 weeks. Every 5 days, 0.3 ml medium was aspirated for radio-immunoassay (RIA) screening and freshly replenished. RIA for rat GH and rat PRL were performed on the media obtained from all the wells. Cells growing in wells which yielded positive hormone immunoreactivity were subsequently recloned in multiwells at a density of O-3 cells/well and the medium screened by RIA. The cells in the wells yielding positive hormone immunoreactivity were either subcultured as separate clones or frozen.

Radio-immunoassay

(RIA)

RIA for rat GH and PRL were performed using materials supplied by the National Hormone L Pituitary Agency, NIADDK. Sensitivity of these RIA’s (80% BISo) was 0.2 and 0.1 &tube for rat GH and PRL respectively.

Immunocytochemical

Staining

Both parent cells and hybrid cells were harvested and fixed to glass slides using a cyto-centrifuge. An immunoperoxidase indirect sandwich technique [18] was used to identify clones producing GH and PRL. Endogenous peroxidase activity was blocked by exposure of the slides to methanolic hydrogen peroxide (3 % aqueous hydrogen peroxide : 5 vol of methanol) for 30 min. Slides were then washed with water and incubated with either rabbit anti-rat GH serum (1 : 2 OflO), rabbit anti-rat PRL serum (1: 2 000), or non-immune rabbit serum (1: 2 000) for 30 min. These antisera were obtained from the National Hormone 8r Pituitary Agency, NIADDK. After washing with phosphate-buffer saline, slides were sequentially incubated with swine anti-rabbit serum (1 : 20 dilution, Dakopatts A/S, Copenhagen, Denmark) and horseradish peroxidase-rabbit anti-horse radish peroxidase (PAP, 1: 100 dilution). Specific sites of antibody binding were determined by adding diaminobenzidine tetrahydrochloride (Sigma, 0.6 mg/ml Tlis buffer) and two drops of 3 % hydrogen peroxide for 5 min followed by washing of slides with water. The reaction yields a deep brown color in the cells containing the specific antigen. In order to further confhm the specificity of the immunoperoxidase reaction, rabbit anti-rat PRL or GH serum was neutralized with serum obtained from rats bearing somato-mammotropic tumors [14]. This serum contained >l 000 &ml of rat GH and rat PRL respectively. After 6 h incubation at 37°C the serum-antiserum mixture was centrifuged at 1000 g for 2 h and the supematant (1: 2 000 dilution) used for immunostaining.

Karyotyping Chromosome analysis of both parent (GH3 and LMTK-) cells and hybrid cells were performed as follows: Metaphase arrest was induced by addition of colcemid (0.5 pg/ml) to the medium for 5 h. Medium was aspirated, cells washed with Hanks balanced salt solution and trypsinized. Trypsinized cells were resuspended in 0.075 M KC1 for 20 min. After fixing, cells were dropped on to glass slides, stained and banded with Giemsa (4%), and air-dried as previously described [19]. Exp Cell

Res 162 (1986)

478 Melmed

and Fagin

1. Hybrid cells growing in HAT medium 12 days after fusion of rat pituitary tumor cells and mouse LMTK- fibroblasts. x400. Fig. 2. Monolayer culture of hybrid clone (A3) growing in HAT medium. Cells secreted rGH and rPRL and doubling time was 40 h. (a) x200; (b) x400. Fig.

Isozyme Analysis Assays for isozymes were kindly performed by Dr Mary Ellen Sparkes, Cytogenetics Laboratory, UCLA School of Medicine. Two enzyme markers, adenosine deaminase and superoxide dismutase were used to differentiate rat and murine species specificity. Equal aliquots of trypsinized cell sonicates of rat parent cells, murine parent cells and hybrid cells respectively were subjected to starch gel electrophoresis followed by specific staining as described [20].

Hormone

Secretion

After hybrid cells were cloned and subcultured, they were maintained as monolayer cultures in Ham’s F10 medium supplemented with FCS (10%) and antibiotics. For growth curves and hormone secretory studies approx. ld cells were seeded into 35mm multiwell (Falcon) plates and medium was aspirated for RIA at the designated time points. At each time point, cells were washed with serumfree Ham’s FlO medium, trypsinized and an aliquot of the cell suspension counted in an automatic cell counter (Coulter Electronics, Hialeah, Fla.). For acute release experiments, medium was aspirated from wells containing hybrid cells in the midlogarithmic phase of growth. Monolayers were carefully washed with warm (37°C) serum-free Ham’s FlO medium. One ml of serum-free Ham’s medium containing HEPES buffer (5 mM) and the appropriate drug or vehicle was then added to the well for 45 min and aspirated for RIA. All incubations were performed at 37°C in a humiditled atmosphere of 95 % air +5 % COZ.

Reagents Thyrotropin-releasing hormone (TRH) was obtained from Abbott Laboratories, Chicago, Ill. Unless otherwise stated, all cell culture materials were purchased from Irvine Scientific Co., Santa Ana, Cdif.

Hormone-secreting

Fig.

40

60 60 CHROMOSOME

100

120 NUMBER

140

cell hybrids

419

3. Histogram of chromosome number in three

160

Statistics Differences were assessed by unpaired t-testing.

RESULTS Five hybrid clones secreting rGH and rPRL were derived from two fusion experiments. Hybrid cells have been subcultured for over 20 passages and both rGH and rPRL secretion have been sustained. The original cloned hybrids secreted 150 ng rPRL and 321 ng rGH/106 cells/24 h in the first subculture. Hormone secretion rates are currently 100 ng PRL and 90 ng rGW106 cells/24 h (subculture 22). These secretion rates are lower than those of the parent GH3 tumor cells which secreted 720 ng rPRL and 660 ng rGH/106 cells/24 h at the time of fusion. All functional hybrid clones identified and thus far studied have secreted both rGH and rPRL. No immunoactive rGH or rPRL was detected in the medium of the mouse parental (LMTK-) cells. Morphology

and Growth

Fig. 1 shows typical hybrid cells seen during the first 14 days after fusion. The resultant subcultured cells, grown in HAT medium, resembled the HAT-sensitive mouse fibroblast parent. Hybrid clones (fig. 2), grew rapidly in monolayer and doubling time ranged between 36 and 48 h. Cells contained multiple nucleoli and were fibroblastic in appearance. The hybrid cells were flat with long processes and adhered very readily to the plastic surface. The hybrids bore no morphological resemblance to the hexagonal epithelial nature of the GHs pituitary tumor cell parent. The chromosomal distribution of three hybrid clones is depicted in fig. 3. The modal distribution of chromosomes in the parent cell lines was 68 in the rat GHJ Exp Cell

Res 162 (1986)

480 Melmed

and Fagin

Fig. 4. G-banded karyogram of a rat-mouse somatic cell hybrid. Large true metacentric mouse chromosome markers (top line) and small submetacentric rat chromosomes (bottom line) are evident.

and 55 in the mouse LMTKcell. As shown in a typical karyogram in fig. 4, the hybrid cells had chromosomal markers typical of both mouse and rat parents. The karyotype included large true metacentric and telocentric chromosomes characteristic of the mouse parent [21] and many small metacentric and sub-metacentric chromosomes seen in the rat parent [22].

A

B

C

D

Fig. 5. Starch gel electrophoresis of adenosine deaminase. Band patterns shown are A, mouse (LMTK-); B, hybrid; C, hybrid; and D, rat (GH,). Protein concentration of cell extracts applied to the gel was similar for all four samples. The patterns of the hybrid cells correspond to those of both the mouse and rat iso-enzyme. Exp

Cell

Res 162 (1986)

Hormone-secreting

I

500

900

400

700

= 4

F

~~

cell hybrids

481

a

2

300

500

2-

2c

200

300

%

s 9 ‘0 C : -

ao6040-

4 2 6 I,...' 0

0 12

3

4

DAY

Fig. 6. Growth curve and hormone production by A3 hybrid during a 12th subculture. 5x104 cells were plated on day 0 in 35-mm wells containing 2 ml of medium. Medium was aspirated and cells counted from triplicate wells at each time-point. SE was less than 10% of the means shown. Fig. 7. Incubation of B6 hybrid cells (15th subculture) C--O, without or O--4, with added TRH 80 nM. Medium was aspirated for BIA and cells harvested and counted at each point. (Mean of triplicate wells + SD.)

Isozyme Analysis In order to further confirm unambiguously the inter-specific nature of the hybrid cells, aliquots of cell sonicates were subjected to starch gel electrophoresis and adenosine deaminase (ADA) and superoxide dismutase activities were visualized by immunochemical staining. Fig. 5 shows the ADA isozyme pattern. Fig. 5, lanes A, D, shows the isozyme pattern for mouse and rat ADA respectively. Two hybrid cell clones (fig. 5, B, C) both showed mixed inter-specific migratory patterns. The hybrid pattern for superoxide dismutase (not shown) also confirmed that the cells expressed both mouse and rat forms of the enzyme. Zmmunocytochemistry Hybrid cells and GH3 rat parental cells stained positively for rGH and rPRL using an immunoperoxidase histochemical technique with specific rPRL and rGH antiserum. Mouse LMTKparental cells did not exhibit immunoperoxidase activity. Incubation of either GHs or hybrid cells with non-immune rabbit serum (1: 2 000) yielded no immunoperoxidase color reaction. Peroxidase stain of hybrid cells was blocked by prior neutralization of the specific rat antiserum with excess rGH or rPRL. Hormone

Secretion

Fig. 6 shows the growth curve and hormone-secretory pattern of a subclone (A3) during its 12th subculture. Doubling time of the cells (approx. 48 h) paralExp Cell

Res

162 (1986)

482 Melmed

and Fagin

leled the secretion of both rGH and rPRL. In this clone, the secretion of rPRL (500 ng/4 days) was about 55% of rGH (880 ng/4 days) on a weight basis. During later subcultures, as the secretion rates of both hormones declined, the proportion of rGH declined to less than 50% of rPRL. Gel filtration of the hybrid cell culture medium on Sephadex G-75 showed that both the immunoreactive PRL and GH co-chromatographed with both rGH and rPRL secreted by GH3 parent cells and 1251-labelled hormone standard (rGH I-4 and rPRL-I-5, National Pituitary Agency, NIADDKD) . Hormone

Modulation

Exposure of hybrid cells (subculture 14) to KC1 (50 mM) for 45 min in serumfree medium resulted in stimulation of rPRL release from 34+2 to 56+5 ng/106 cells (mean + SD, ~<0.05). No significant increase in rGH release was seen after acute (45 min) KC1 exposure. rPRL release was not significantly increased during acute (45 min) exposure of the cells to 80 nM TRH (51 f 10 ng/106 cells). The same doses of KC1 and TRH elicited over 100% increases in both hormone release by the parent GH3 cells. Fig. 7 shows the time course of rPRL secretion by the B6 hybrid (subculture 15). Simultaneous continuous incubation of the cells with TRH (80 nM) resulted in significant stimulation of hormone release from 24 h of exposure (p~O.02). As seen from fig. 76, TRH did not alter cell replication of doubling time. Secretion of rGH by this subculture after 72 h was 39f9 rig/well, and was not altered by TRH. DISCUSSION This paper describes the isolation of inter-specific functional hormone-producing cell hybrids. Somatic cell fusion of a rat pituitary tumor cell (GH3) and murine fibroblast cell (LMTK-) resulted in the initiation of hybrid cell lines secreting rat GH and PRL. The unambiguous hybrid nature of the cell line was demonstrated by morphology clearly resembling the murine parent, HAT resistance conferred by intergenic complementation, isozyme markers specific for both rat and mouse, and hybrid rat and mouse karyotypes. As the mouse LMTKcell is sensitive to aminopterin, the fact that the hybrid cells grew in the HAT medium is evidence of intergenic complementation of the thymidine kinase gene from the rat parent, allowing the hybrid cells to selectively thrive in HAT medium [ 101. Specific gene expression for both rat GH and PRL has been retained. Hybrid cells exhibited specific immunoperoxidase staining for rat GH and PRL. Hormone secretion, measured by RIA of the culture medium, has been sustained for 22 subcultures. The karyotype of the hybrid cells was heterogeneous, which is similar to findings by others [5-7, 231. Some clones (A3) contained more than the expected (1 s+ 1s) complement of chromosomes, probably reflecting fusion of more than two hyperdiploid cells. Hormone secretion could not be correlated with chromosome number or morphology of the different hybrid clones. This may be due to Exp Cell

Res 162 (1986)

Hormone-secreting

cell hybrids

483

the large number of small submetacentric rat chromosomes which are similar in structure and appearance using the staining methods described. Further chromosoma1 studies utilizing centromeric fluorescent staining will be required to map the rat chromosome (or chromosomes) containing the specific hormone genes. The karyograms were characteristic of rat-mouse hybrids. The LMTKcells have seven large marker chromosomes [19], characteristically large true metacentric, or bi-armed, apparently arising from fusion of two telocentric chromosomes. The GH3 cells (rat) have characteristic small metacentric chromosomes, with only one pair of large metacentrics [20]. All the hybrid clones studied exhibited karyotypes containing both mouse and rat marker chromosomes. The highly differentiated function of polypeptide hormone synthesis by GH3 tumor cells was transferred to the hybrid cell. Although hormone secretion has been sustained for over a year of subculturing, the secretion rates have declined, with rGH secretion declining more markedly than rPRL. This may indicate segregation of rat chromosomes responsible for regulating hormone production or may reflect repression of GH and PRL gene expression by the mouse fibroblast recipient cell [2, 41. Alternatively, as the hybrids are derived from tumor cells, a non-secreting mutant cell may have arisen. As rat GH gene loss could not be demonstrated in rat-mouse somatic cell hybrids with absent GH secretion [23], and neither translatable nor structural GH mRNA was present, it is possible that the mechanism of extinction of GH function may be mediated by repressive gene regulation involving defective transcription in hybrid cells [23]. The GHJ parent cloned cell line secretes both GH and PRL, strongly suggesting that the two hormones are produced by the same cell [12].The GH3 cell may represent a primitive pituitary stem cell and the ability to transfer the genetic expression of both hormones to the hybrid cell may suggest either close linkage of the two gene sites or even a common chromosomal site. No hybrid clones or subcultured clones have yet been isolated by us which do not secrete both rat GH and PRL. Both rat PRL and GH are probably derived from a common ancestral gene [24, 251. The similar structural organization of the coding regions of both genes (e.g., both contain four introns) as well as their similar length is evidence of a common evolutionary origin [24]. More detailed gene mapping of the hybrid cells is required in order to localize these genes to specific rat chromosomal sites. The sensitivity of the rat GH and PRL RIA is such that early hybrid clones were able to be screened effectively for specific gene expression. As a nonhormone secreting hybrid clone (Cl) also thrived in HAT medium and exhibited iso-enzyme banding characteristic of an inter-specific rat-mouse hybrid, it is possible that these cells could also be used to study other specific non-secretory pituitary cell functions, such as enzyme synthesis. Previously described pituitary cell hybrids showed extinction of hormone expression [4, 5,22, 231. The methodology used in the present study differs from that used in those studies. Their somatic cell fusion was achieved either spontaneously [5] or with inactivated Sendai virus [4]. In the present studies PEG was Exp Cell

Res

162 (1986)

484 Melmed

and Fagin

effectively utilized to achieve somatic cell fusion and intergenic complementation. There is, however, no evidence that either of these fusion methods result in differing gene expression in resultant hybrids. As mentioned earlier, Strobl et al. have postulated that a transcriptional deficit may account for the lack of hormone secretion encountered in their pituitary hybrid model [23]. The fact that we were able to isolate five GH- and PRL-secreting hybrid cells implies that inhibition of GH transcription is not invariably present. Whether the gradual decline in secretion rate over time found in our hybrids is due to inhibition of GH transcription or other pre- or post-transcriptional events, requires further study. Only one hybrid subclone, GL4-CS, described by Thomson et al [4], was shown to secrete PRL in significant amounts. This subclone had no demonstrable mouse genetic information [4], raising the possibility that the clone may be a complete segregant or even a variant rat GH3 mutant cell. Although this specific cell had estrogen receptors and lacked TRH receptors, PRL production was not stimulated by either of these two agents. In the present study, PRL production was stimulated by TRH, implying that this regulatory mechanism, present in the parental GH3 cell [26, 271, is also operative in the hybrid cell line tested. These studies show that the highly differentiated function of polypeptide hormone secretion can be both expressed and modulated in an interspecific hybrid cell. The authors thank MS Marilyn Leung for outstanding technical assistance, Dr Jonathan Said for assistance with immunocytochemical staining, and MS Helene Zauderer for typing the manuscript. The authors are grateful to Drs Glenn Braunstein, Keith Foumier and Robert Sparkes for their constructive advice and encouragement. This work was supported in part by a NIH Grant AM 33802 from the National Institutes of Arthritis, Diabetes, Digestive and Kidney Diseases.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.

Davis, FM & Adelberg, E A, Bact rev 37 (1973) 197. Shows, R W & Sakaguchi, A Y, In vitro 16 (1980) 55. Thompson, E B, Norman, M R & Lippman, M E, Ret prog horm res 33 (1973) 571. Thompson, E B, Dandies, P S, Buckler, C E & Tashjian, A H Jr, J steroid biochem 12 (1980) 193. Sonnenschein, C, Richardson, V I & Tashjian, A H Jr, Exp cell res 69 (1971) 336. Bordelon, M R, Coon, H G & Kohler, P 0, Exp cell res 103 (1976) 303. Bordelon Riser, M E, Siciliano, M J & Kohler, P 0, Somat cell genet 5 (1979) 597. Boyd, A E III, Bordelon-Riser, M, Bolton, W, Siciliano, M J & Jordan, P, Diabetes 31 (1982) 1035. Mebned, S, J clin endocrinol metab 56 (1983) 1145. Littlefield, J W, Science 145 (1964) 709. Tashjian, A H Jr, Yasumura, Y, Levine, L, Sato, G H & Parker, M L, Endocrinology 82 (1968) 342. Melmed, S, Biochem biphys res commun 108 (1982) 1460. Tashjian, A H Jr, Bancroft, F C & Levine, L, J cell bio147 (1970) 61. Mebned, S, Carlson, H E, Briggs, J & Her&man, J M, Endocrinology 107 (1980) 789. Davidson, R L & Gerald, P S, Somat cell genet 2 (1976) 165. Davidson, R L, O’Mahey, K A & Wheeler, T B, Somat cell genet 2 (1976) 271. Puck, T, Marcus, P & Ciecura, S, J exp med 103 (1956) 273.

Exp CellRes

162 (1986)

Hormone-secreting

cell hybrids

485

18. Pinkus, G S & Said, J W, Am j path01 87 (1977) 47. 19. Seabright, M, Lancet 2 (1971) 971. 20. Harris, H & Hopkinson, D A, Handbook of enzyme electrophoresis in human genetics. Elsevier/ North-Holland, New York (1978). 21. Rushton, A R, Can j genet cytol 15 (1973) 791. 22. Sonnenschein, C, Richardson, U I & Tashjian, A H Jr, Exp cell res 61 (1970) 121. 23. Strobl, J S, Dannies, P S & Thompson, E B, J biol them 257 (1982) 6588. 24. Chien, Y H & Thompson, E B, Proc natl acad sci US 77 (1980) 4583. 25. Barta, A, Richards, R I, Baxter, J D & Shine, J, Proc natl acad sci US 78 (1981) 4867. 26. Tashjian, A H Jr, Barowsky, N H & Jensen, D K, Biochem biophys res commun 43 (1971) 516. 27. Hinkle, P M & Tashjian, A H Jr, J biol them 248 (1973) 6180. Received May 23, 1985 Revised version received August 13, 1985

Printed

in Sweden

Erp

Cell

Res 162 (1986)