- Email: [email protected]
insulin-like growth factor I production by the human lymphoid cell line, IM-9
167
Molecular and Cellular Endocrinology, 63 (1989) 167-173
Elsevier Scientific Publishers Ireland, Ltd. MCE 02049
Human growth hormone stimulates somatomedin C/insulin-like growth factor I production by the human lymphoid cell line, WI-9 J. Michael Palmer and Michael Wallis Biochemistry Laboratory, School of Biological Sciences, University of Sussex, Falmer, Brighton BNI 9QG, U.K.
(Received 13 January 1989; accepted 16 January 1989)
Key words: Growth hormone; Somatomedin C/insulin-like
growth factor; Insulin-like growth factor I; IM-9 cell; Growth hormone
receptor
The human lymphoid cell line, IM-9, is known to possess receptors for human growth hormone (hGH), but the only biological response that has been shown to follow binding of this hormone to the cells is receptor down-regulation. We have studied the actions of hGH on production of insulin-like growth factor I (IGF-I) by IM-9 cells. In order to demonstrate effects cells had to be transferred to a serum-free medium in which cell multip~~tion almost ceased, and cell viability fell to 50-60%. hGH stimulated IGF-I production by up to 400%. The effect was dose-related, but the dose-response curve was bimodal, with peaks of activity at approximately 15 ng/ml and 1000 ng/ml hGH. The effect of hGH was of slow onset, becoming significant only after about 24 h, and approaching a maximum after 2-5 days of treatment. hGH had a much greater stimulatory effect than non-primate growth hormones. The physiological significance of the effect observed is not yet clear, but it is apparent that the IM-9 line is a potentially useful model for study of the actions of growth hormone.
Introduction The IM-9 cell line is one of several human lymphoid cell lines established in culture (Fahey et al., 1972). IM-9 cells bind “251-labelled insulin (Gavin et al., 1974) growth hormone (GH) (Lesniak et al., 1974, 1977; Lesniak and Roth, 1976), somatomedin C/insulin-like growth factor I (IGF-I) (Rosenfeld and Hintz, 1980a; Jacobs et al., 1986) and insulin-like growth factor II (IGF-II) (Misra et al., 1986). The cell line has been widely used for studies on the nature of the receptors for
Address for correspondence: M. Wallis, Biochemistry Laboratory, School of Biological Sciences, University of Sussex, Falmer, Brighton BNl 9QG, U.K. 0303-7207/89/$03.50
these hormones and their down-regulation following binding (e.g. Lesniak and Roth, 1976; Lesniak et al., 1977; Hughes et al., 1983; Asakawa et al., 1986). There have been few reports on biological actions following hormone-receptor interaction in this cell line, however. In particular, the binding of human growth hormone (hGH) to IM-9 cells has not been associated with any biological action, other than modulation of receptor number. During the past few years it has been shown that several cell types incubation in vitro are able to synthesize and secrete IGF-I, and that in some cases GH can enhance production of this peptide. The effect has been demonstrated in several established fibroblast cell lines, including some of human origin (Atkison et al., 1980; Clemmons et al., 1981) and also various primary cell cultures
6 1989 Elsevier Scientific Publishers Ireland, Ltd.
168
including hepatocytes (Spencer, 1979; Atkison et al., 1984; Scott et al., 1985). In view of the occurrence of GH receptors in IM-9 cells it was of interest to examine whether these cells can produce IGF-I and if so to assess the effect of GH on the process in relation to binding of the hormone. Materials and methods Materials Culture media, sera and plasticware for cell culture were obtained from Gibco, Paisley, U.K. Sep-paks for extractions of IGF-I were from Millipore/Waters, Harrow, Middlesex, U.K. Iodine125 was from DuPont (U.K.), Stevenage, Herts, U.K. or Amersham International, Amersham, Bucks, U.K. hGH was obtained from Wellcome Laboratories, Becker&am, Kent, U.K. Ovine GH and prolactin and anti-IGF-I serum were gifts from the National Institute of Arthritis. Digestive and Kidney Diseases, Bethesda, MD, U.S.A. IGF-I (Thr-59 human IGF-I) was obtained from Amersham International, Amersham, Bucks, U.K. Cell culture IM-9 cells were obtained from Flow Laboratories, Irvine, Ayrshire, U.K. The cells were cultured in RPM1 medium 1640 containing 2.0 g/l NaHCO,, 10% (v/v) fetal calf serum, 30 mg/l penicillin and 50 mg/l streptomycin, at pH 7.4. The cells were maintained at 37” C in an atmosphere of 95% air : 5% COz. For receptor assays and for incubations to measure IGF-I release the cells were transferred to serum-free medium (RPM1 1640 containing 2.0 g/l NaHCO,, 25 mM Hepes, 0.1% bovine serum albumin (BSA), pH 7.4) after first washing 3 times in sterile phosphate-buffered saline (PBS; 8 g/l NaCl, 0.2 g/l KCI, 1.15 g/l Na,HPO,, 0.2 g/l KH,PO,, 0.2 g/l glucose, pH 7.4) to remove any trace of serum. Cell numbers were determined using a haemocytometer, and viability was estimated by ability to exclude trypan blue. For studies on IGF-I production cells were diluted to about 1 x lo5 cells/ml in serum-free medium and 1 ml of this suspension was pipetted into each 75 X 12 mm culture tube (Sterilin). The appropriate concentration of hormone dissolved in PBS was added and the cells were incubated at
37 o C under 95% air : 5% CO, for the required period. The tubes were then centrifuged (2000 x g for 30 min) and the supematant was removed for assay of IGF-I. Labelling of hormones IGF-I and hGH used in the receptor assays and immunoassays were ‘251-labelled by the iodogen method (Fraker and Speck, 1978). The iodinated hormones were partially purified on a Sephadex G-50 column (12-20 cm x 1 cm) eluted with PBS containing 0.1% BSA and 0.6 mM merthiolate (pH 7.4) (hGH) or Tris-HCl buffer (25 mM Tris/HCl, 0.1% BSA, 0.6 mM merthiolate, pH 7.6) (IGF-I). Aliquots of labelled hormone were stored at 4’ C until required. Binding of ‘251-labelled hGH to IM-9 cells IM-9 cells were washed 3 times in sterile PBS and then resuspended in serum-free medium (RPMI; 25 mM Hepes, 0.1% BSA, pH 7.4) at a concentration of about 1.5 x lo6 cells/ml. 1.0 ml of this cell suspension was then pipetted into each of the plastic culture tubes (75 mm X 12 mm, Sterilin) used in the assay. 100 ~1 of unlabelled hormone (of the appropriate concentration, dissolved in PBS) was added, followed by 100 ~1 of ‘251-labelled hGH (about lo5 cpm). After mixing of the contents, tubes were incubated at 30 o C for 2 h in an atmosphere of 95% air: 5% CO,. The cells were then precipitated by centrifugation (2000 X g, 30 mm). Supernatants were removed and the precipitates were subjected to gamma counting. IGF-I extraction 0.25 ml of sample were acidified with 1 ml of 0.5 M HCl and then incubated at 20-23” C for 30 min. Adsorption on Sep-pak Cl8 cartridges was used to separate the IGF-I from protein (including binding protein) in the sample (Wallis et al., 1987). Each cartridge was prepared by washing successively with isopropanol (5 ml), methanol (5 ml) and 4% (v/v) acetic acid (15 ml). The sample was then applied (flow rate approximately 0.4 ml/mm) and the cartridge was washed with 4% (v/v) acetic acid (20 ml) to remove unadsorbed material. IGF-I was then eluted with 4 ml of methanol. This elute was dried at 37’ C in a stream of air, dissolved in
169
IGF-I radioimmunoassay buffer (0.5 ml), and incubated at 37°C for 10 min to ensure complete solution. The extraction method used gave a recovery for IGF-I of 70-908, and reproducibility between assays was generally good. For a series of ten consecutive assays the interassay coefficient of variation for a control sample subjected to extraction and assay was 18%. Radioimmunoassay
of IGF-I
The IGF-I content of medium samples, before or after extraction, was carried out by radioimmunoassay (Wallis et al., 1987) using antiserum to human IGF-I obtained from NIADDK, 1251labelled IGF-I, and as standard a partially purified IGF-I preparation (N4G50 II) obtained from Dr. A. Holder at the National Institute for Child Health, London, U.K. This standard has a potency of about 30 units/mg and has been calibrated by radioimmunoassay with reference to the Amgen somatomedin C as containing 1.31 pg/mg (Wallis et al., 1987). Results are expressed as ng equivalents IGF-I/ml, based on this potency for the standard. The antibody was reported to show less than 5% crossreactivity with IGF-II, and to crossreact minimally with insulin (notes received from NIADDK). Treatment
of results
All experimental and control treatments were performed in triplicate, at least. Results are expressed as mean values k SEM. The statistical significance of differences between groups of observations was tested using Student’s t-test. Results Earlier work (Lesniak et al., 1974,1977; Lesniak and Roth, 1976) has shown that the IM-9 cell line can bind ‘251-labelled hGH, and we have confirmed this observation. Intact IM-9 cells bound approximately 14.5% of the ‘251-labelled hormone added, of which about 8% was specifically bound (displaced by excess, 10 pg, unlabelled hGH). Unlabelled hGH competed for binding sites in a dose-dependent manner (Fig. 1). Ovine GH also displaced ‘251-labelled hGH, with a potency approximately 10% that of hGH. Rat GH and bovine insulin did not displace ‘251-labelled hGH. Rat
Fig. 1. Binding of 1251-labelled hGH to IM-9 cells. IM-9 cells (approx. 1.5 x lo6 cells/tube) were incubated with ‘251-labelled hGH (approx. 10s cpm) in the absence or presence of unlabelled human (0) ovine (0) or rat (W) GH at the concentration shown. Binding to cells was measured as described in the text, and is expressed as a percentage of the total cpm present. Values are means of four replicates; vertical lines show + SEM.
and bovine prolactin did so with very low potency (< 2% that of hGH). When IM-9 cells were incubated in serum-free medium in the presence of hGH, immunoreactive IGF-I was released into the medium. The effect of hGH was slow; a significant effect (P < 0.05) was seen only after 24 h (Fig. 2). However, substantial effects required 48-120 h of treatment. Cells incubated without added hGH showed a small, but significant, increase in medium content of IGF-I during the course of the incubation. Medium at time zero contained a low IGF-I content, possibly due to carry-over from the serum-containing medium in which the cells were grown, as assays performed on non-conditioned serum-free medium failed to detect significant quantities of IGF-I. The effect of hGH on IGF-I production by IM-9 cells was dose-related. Fig. 3 shows the effect of incubating cells with various concentrations of hGH on the release of IGF-I from IM-9 cells after a 5-day incubation. The response to increasing concentrations of hGH was bimodal with maxima at 15 and 1000 ng/ml hGH. The size of the response obtained at 1000 ng/ml was rather variable, being as great as that seen at 15 ng/ml in some experiments, lower in others. The bimodal dose-response relationship was seen both when
170
i4
48
72 lncllbatii
ee
120
time(h)
Fig. 2. Actions of hGH on IGF-I production by IM-9 cells. IM-9 cells (lo5 cells/ml) in serum-free medium were incubated in the absence (m) or presence of 10 (0) or 1000 (0) ng/ml hGH. IGF-I concentration in the medium was determined at the times shown by radioimmunoassay after extraction on Sep-pales. Values shown are means of four replicates; vertical lines show f SEM. Significant stimulation of IGF-I production by hGH (comparing treated and untreated for any given time) is indicated: * P < 0.05; * * P c 0.01.
IGF-I was measured directly in medium and when it was extracted first using Sep-paks (Fig. 3) suggesting that it is not a consequence of interac-
tkmentmtii
of hGHhg/ml)
Fig. 3. Actions of hGH on IGF-I production by IM-9 cells: dose-response curve. IM-9 cells (lo5 cells/ml) were incubated for 5 days in serum-free medium in the presence of the hGH concentrations shown. IGF-I concentrations in the media were then determined by radioimmunoassay either directly (shaded bars) or after extraction on Sep-paks (open bars). Values are means of four replicates; vertical lines show f SEM. hGH gave a significant stimulation of IGF-I secretion at all concentrations tested (P < 0.05 for unextracted values at 500 ng/ml and lo4 ng/ml hGH; P < 0.01 for all other observations).
Concentration d GH(ng(m11
Fig. 4. Actions of human, ovine and rat GH on IGF-I production by IM-9 cells. IM-9 cells (lo5 cells/ml) in serum-free medium were incubated for 5 days in the presence of the concentrations shown of human (a), ovine (0) or rat (m) GH. IGF-I concentration in the media was then determined by radioimmunoassay, without extraction on Sep-paks. Values are means of eight replicates; vertical lines show f SEM. * Values significantly different (P i 0.01) from control (no GH).
tion between IGF-I and other components in the medium, such as IGF binding protein. No correction has been made to allow for small losses during the Sep-pak extraction procedure, so the corresponding values in Fig. 3 (open bars) may slightly underestimate IGF-I concentration after extraction. Incubation of conditioned medium from IM-9 cells with ‘251-labelled IGF-I followed by gel filtration suggested that only traces of binding protein were present. The specificity of the effect of hGH on IGF-I release was studied (Fig. 4). In accordance with the results of binding studies, human GH was most effective in stimulating IGF-I release. Ovine GH gave a smaller response, and a less marked bimodal dose-response curve. Rat GH had little effect on IGF-I release from the cells, though it did give a significant stimulation (P < 0.001) at the highest concentration tested. These results suggest that the response seen at lower concentrations (- 10 ng/ml) of hormone is more specific for hGH than the response brought about by much higher hormone concentrations.
171
mained at over incubation.
80% throughout
the
period
of
Discussion
I
O
1
1
.
*
~
234567 Incubation time(days]
Fig. 5. Growth of IM-9 cells in serum-containing and serum-free medium. IM-9 cells were incubated in medium containing 10% foetal calf serum (closed symbols) or without serum (open symbols) in the absence (0, 0) or presence of 10 ng/ml (& 0) or 1000 ng/ml (A, A) hGH. Cell numbers were determined at the times shown, using a haemocytometer. Values shown are means of four determinations; vertical bars show SEM.
The studies on IGF-I production were carried out in serum-free medium, and the ability of IM-9 cells to survive and grow in such medium was investigated (Fig. 5). In the presence of foetal calf serum, cells grew rapidly during the first 2 days of a 7-day incubation; growth then slowed, and there was no change in cell number between days 3 and 7. After transferral of cells to serum-free medium, cell proliferation stopped, but after 2-3 days cell numbers began to increase slowly. The presence of hGH had no significant effect upon growth either in the normal growth medium or in serum-free medium. The viability of the cells (measured by trypan blue exclusion and by ability to recommence growth when transferred to fresh serum-containing medium) fell rapidly when they were placed in serum-free medium, dropping to 50-55% and then remaining constant (results not shown). In serum-containing medium viability re-
Previous studies on IM-9 cells have shown that they bind ‘251-labelled hGH (Lesniak et al., 1974, 1977; Lesniak and Roth, 1976; Hughes et al., 1983; Asakawa et al., 1986), but have provided no evidence for a biological response to such binding apart from modulation of receptor levels (Lesniak and Roth, 1976). We have confirmed previous results showing that IM-9 cells can bind 1251labelled hGH specifically. Binding was species specific in that ‘251-labelled hGH could be displaced by unlabelled hGH, but not by rat GH, and only with low potency by ovine GH. These results agree with those presented previously (Lesniak et al., 1974) although the apparent potency of ovine GH was rather greater than that found by others; they also accord with the observed biological specificity of the actions of GH in man. We have also shown that hGH, and to a lesser extent ovine GH, can stimulate somatomedin C/IGF-I production by IM-9 cells. The effect was dose-dependent, but showed an unusual, bimodal dose-response relationship (Figs. 3 and 4), with maximal activity at lo-20 ng/ml and 1000 ng/ml, and decreased activity at intermediate concentrations. The effect was also of rather slow onset (Fig. 2), requiring exposure of cells to the hormone for more than 5 h for significant stimulation, and exposure for at least 48 h before IGF-I output from the cells was doubled. This slow onset accords with the idea that stimulation of production of IGF-I rather than of secretion of preformed growth factor is involved, though direct studies of effects on synthesis of IGF-I are needed to confirm this idea. The slow onset applied to stimulation by both low (10 ng/ml) and high (1000 ng/ml) GH concentrations. The explanation for the bimodal dose-response relationship observed here is not clear. It is possible that it results from the presence of more than one active (or inhibitory) component in the hGH preparation used, though the preparation was a highly purified one. Another possibility, that the bimodal dose-response curve might be due to
172
secretion of IGF binding protein, was explored by assaying the apparent growth factor concentration both before and after extraction using Sep-pak columns. Such extraction did not affect the pattern of response observed, suggesting that production of binding protein does not contribute to its bimodal nature. It is possible that if a very large quantity of binding protein was produced the Sep-pak procedure might not extract IGF-I effectively, though it would be surprising in that case if the dose-response pattern was completely unaltered, as was actually observed. Incubation of 125I-labelled IGF-I with conditioned medium provided no evidence for the presence of large amounts of binding protein. Other cellular systems showing enhanced IGF-I production in response to GH, such as liver cells (Spencer, 1979) and cultured fibroblasts (Atkison et al., 1980; Clemmons et al., 1981) do not appear to show a bimodal response of this kind. It is possible that it is associated with the prolonged exposure to GH necessary for demonstration of a good response, or to the transfer to serum-free media in which cells are unable to proliferate. Down-regulation/ internalization of GH receptors is known to follow hormone binding to IM-9 cells (Lesniak and Roth, 1976; Rosenfeld and Hintz, 1980b) and it is possible that such down-regulation caused by a relatively high concentration of GH could lead to lowered responsiveness to the hormone. It is notable, however, that extensive receptor down-regulation can be produced by a concentration of hGH (about 15 ng/ml) that gives a maximal response for IGF-I production. The specificity of the response to the hormones tested was similar to the specificity of their ability to bind to receptors on IM-9 cells. Thus, the response to ovine GH was much less than that to the human hormone, and the response to rat GH was very small. It is perhaps surprising that such differences related mainly to the absolute response rather than the concentrations at which responses were observed, but the bimodal nature of the dose-response curve makes this feature difficult to assess. Since calf serum contains IGF-I which would crossreact with and mask IGF-I produced by IM-9 cells, the studies carried out on these cells were performed in serum-free medium. The experiment
illustrated in Fig. 5 showed that cell proliferation is stopped by transfer to such medium. A substantial proportion of the cells die soon after this transfer, but the remainder remain viable for at least 7 days and presumably represent the cells that respond to GH by producing IGF-I. Clearly, the serum-free medium used for the experiments is not ideal, and the transfer could be partly responsible for the very delayed response seen to GH. A serum-free medium that supports sustained growth would facilitate the use of IM-9 cells to study the actions of GH. It should be noted, however, that addition of hGH to culture medium, with or without serum, had no effect on the proliferation of the cells. A number of cell lines have now been shown to produce IGF-like material, though only in a few is this production stimulated by GH. Atkison et al. (1984) have shown that primary cultures of rat hepatocytes and cell lines derived from human epidermoid carcinoma, pancreatic carcinoma, Simian virus 40 transformed human skin fibroblasts and normal skin fibroblasts all release IGF-like factors into their growth media; only for the rat hepatocytes was it shown that GH enhanced IGF-I production. Insulin-like and IGFlike factors have also been found in media conditioned by other types of transformed cell, including lines derived from human myeloid leukaemia (HL-60) (Yamanouchi et al., 1985) human osteosarcoma (Blatt et al., 1984) and rat liver (BRL3A) (Rechler et al., 1981); in none of these cases was GH reported to stimulate growth factor production. GH has, however, been reported to stimulate IGF-I production by some human fibroblast cell lines (Atkison et al., 1980; Clemmons et al., 1981). A previous study detected no IGF-I production by IM-9 cells (Atkison et al., 1984); details of this work were not given, but it is clear that the delayed responsiveness of the cells and the bimodal dose-response curve could make it difficult to detect the action of hGH on these cells. Whether the effect seen with IM-9 cells reflects equivalent actions in normal lymphocytes is not clear, nor is the physiological significance. It does provide, however, a potentially useful model for studies on the mechanism of action of GH, using a cell line that is convenient to work with and on which much previous work on GH receptors has been
173
reported. It should also be noted that IM-9 cells themselves carry somatomedin receptors (Rosenfeld and Hintz, 1980; Duronio et al., 1986) so autocrine effects cannot be excluded. Acknowledgements We thank the Medical Research Council for financial support, Drs. S. Raiti, A. Parlow and the National Institute of Diabetes, Digestive and Ridney Diseases for gifts of hormones and antisera, Mrs. Margaret Daniels for skilled production of the figures and Mrs. Eileen Willis for expert preparation of the manuscript. References Asakawa, K., Hedo, J.A., McElduff, A., Rouiller, D.G., Waters, M.J. and Gorden, P. (1986) B&hem. J. 238, 379-386. Atkison, P.R., Weidman, E.R., Bhaumick, B. and Bala, R.M. (1980) Endocrinology 106, 2006-2012. Ad&on, P.R., Hayden, L.J., Bala, R.M. and Hollenberg, M.D. (1984) Can. J. B&hem. Cell. Biol. 62, 1343-1350. Blatt, J., White, C., Dienes, S., Friedman, H. and Foley, T.P. (1984) Biochem. Biophys. Res. Commun. 123, 373-376. Clemmons, D.R., Underwood, L.E. and Van Wyk, J.J. (1981) J. Clin. Invest. 67, 10-19. Duronio, V., Jacobs, S. and Cuatrecasas, P. (1986) J. Biol. Chem. 261, 970-975. Fahey, J.L., Buell, D.N. and Sox, H.C. (1972) Ann. N.Y. Acad. Sci. 190, 221-234.
Fraker, P.J. and Speck, J.C. (1978) B&hem. Biophys. Res. Commun. 80, 849-857. Gavin, J.R., Roth, J., Neville, D.M., De Meyts, P. and Buell, D.N. (1974) Proc. Natl. Acad. Sci. U.S.A. 71, 84-88. Hughes, J.P., Simpson, J.S.A. and Friesen, H.G. (1983) Endocrinology 112, 1980-1985. Jacobs, S., Cook, S., Svoboda, M.E. and Van Wyk, J.J. (1986) Endocrinology 118, 223-226. Lest&k, M.A. and Roth, J. (1976) J. Biol. Chem. 251, 3720-3729. Lesniak, M.A., Gorden, P., Roth, J. and Gavin, J.R. (1974) J. Biol. Chem. 249, 1661-1667. Lesniak, M.A., Gorden, P. and Roth, J. (1977) J. Clin. Endocrinol. Metab. 44, 838-849. Misra, P., Hintz, R.L. and Rosenfeld, R.G. (1986) J. Clin. Endocrinol. Metab. 63, 1400-1405. Rechler, M.M., Nissley, S.P., King, G.L., Moses, A.C., Van Obberghen-Schilling, E.E., Romanus, J.A., Knight, A.B., Short, P.A. and White, R.M. (1981) J. Supramol. Structure Cell. B&hem. 15, 253-286. Rosenfeld, R.G. and Hintz, R.L. (1980a) Endocrinology 107, 1841-1848. Rosenfeld, R.G. and Hintz, R.L. (1980b) J. Clin. Endocrinol. Metab. 51, 368-375. Scott, C.D., Martin, J.L. and Baxter, R.C. (1985) Endocrinology 116, 1102-1107. Spencer, E.M. (1979) FEBS Lett. 99, 157-161. Wallis, M., Daniels, M., Ray, K.P., Cottingham, J.D. and Aston, R. (1987) B&hem. Biophys. Res. Commun. 149, 187-193. Yamanouchi, T., Tsushima, T., Kasuga, M. and Takaku, F. (1985) Biochem. Biophys. Res. Commun. 129, 293-299.
Molecular and Cellular Endocrinology, 63 (1989) 167-173
Elsevier Scientific Publishers Ireland, Ltd. MCE 02049
Human growth hormone stimulates somatomedin C/insulin-like growth factor I production by the human lymphoid cell line, WI-9 J. Michael Palmer and Michael Wallis Biochemistry Laboratory, School of Biological Sciences, University of Sussex, Falmer, Brighton BNI 9QG, U.K.
(Received 13 January 1989; accepted 16 January 1989)
Key words: Growth hormone; Somatomedin C/insulin-like
growth factor; Insulin-like growth factor I; IM-9 cell; Growth hormone
receptor
The human lymphoid cell line, IM-9, is known to possess receptors for human growth hormone (hGH), but the only biological response that has been shown to follow binding of this hormone to the cells is receptor down-regulation. We have studied the actions of hGH on production of insulin-like growth factor I (IGF-I) by IM-9 cells. In order to demonstrate effects cells had to be transferred to a serum-free medium in which cell multip~~tion almost ceased, and cell viability fell to 50-60%. hGH stimulated IGF-I production by up to 400%. The effect was dose-related, but the dose-response curve was bimodal, with peaks of activity at approximately 15 ng/ml and 1000 ng/ml hGH. The effect of hGH was of slow onset, becoming significant only after about 24 h, and approaching a maximum after 2-5 days of treatment. hGH had a much greater stimulatory effect than non-primate growth hormones. The physiological significance of the effect observed is not yet clear, but it is apparent that the IM-9 line is a potentially useful model for study of the actions of growth hormone.
Introduction The IM-9 cell line is one of several human lymphoid cell lines established in culture (Fahey et al., 1972). IM-9 cells bind “251-labelled insulin (Gavin et al., 1974) growth hormone (GH) (Lesniak et al., 1974, 1977; Lesniak and Roth, 1976), somatomedin C/insulin-like growth factor I (IGF-I) (Rosenfeld and Hintz, 1980a; Jacobs et al., 1986) and insulin-like growth factor II (IGF-II) (Misra et al., 1986). The cell line has been widely used for studies on the nature of the receptors for
Address for correspondence: M. Wallis, Biochemistry Laboratory, School of Biological Sciences, University of Sussex, Falmer, Brighton BNl 9QG, U.K. 0303-7207/89/$03.50
these hormones and their down-regulation following binding (e.g. Lesniak and Roth, 1976; Lesniak et al., 1977; Hughes et al., 1983; Asakawa et al., 1986). There have been few reports on biological actions following hormone-receptor interaction in this cell line, however. In particular, the binding of human growth hormone (hGH) to IM-9 cells has not been associated with any biological action, other than modulation of receptor number. During the past few years it has been shown that several cell types incubation in vitro are able to synthesize and secrete IGF-I, and that in some cases GH can enhance production of this peptide. The effect has been demonstrated in several established fibroblast cell lines, including some of human origin (Atkison et al., 1980; Clemmons et al., 1981) and also various primary cell cultures
6 1989 Elsevier Scientific Publishers Ireland, Ltd.
168
including hepatocytes (Spencer, 1979; Atkison et al., 1984; Scott et al., 1985). In view of the occurrence of GH receptors in IM-9 cells it was of interest to examine whether these cells can produce IGF-I and if so to assess the effect of GH on the process in relation to binding of the hormone. Materials and methods Materials Culture media, sera and plasticware for cell culture were obtained from Gibco, Paisley, U.K. Sep-paks for extractions of IGF-I were from Millipore/Waters, Harrow, Middlesex, U.K. Iodine125 was from DuPont (U.K.), Stevenage, Herts, U.K. or Amersham International, Amersham, Bucks, U.K. hGH was obtained from Wellcome Laboratories, Becker&am, Kent, U.K. Ovine GH and prolactin and anti-IGF-I serum were gifts from the National Institute of Arthritis. Digestive and Kidney Diseases, Bethesda, MD, U.S.A. IGF-I (Thr-59 human IGF-I) was obtained from Amersham International, Amersham, Bucks, U.K. Cell culture IM-9 cells were obtained from Flow Laboratories, Irvine, Ayrshire, U.K. The cells were cultured in RPM1 medium 1640 containing 2.0 g/l NaHCO,, 10% (v/v) fetal calf serum, 30 mg/l penicillin and 50 mg/l streptomycin, at pH 7.4. The cells were maintained at 37” C in an atmosphere of 95% air : 5% COz. For receptor assays and for incubations to measure IGF-I release the cells were transferred to serum-free medium (RPM1 1640 containing 2.0 g/l NaHCO,, 25 mM Hepes, 0.1% bovine serum albumin (BSA), pH 7.4) after first washing 3 times in sterile phosphate-buffered saline (PBS; 8 g/l NaCl, 0.2 g/l KCI, 1.15 g/l Na,HPO,, 0.2 g/l KH,PO,, 0.2 g/l glucose, pH 7.4) to remove any trace of serum. Cell numbers were determined using a haemocytometer, and viability was estimated by ability to exclude trypan blue. For studies on IGF-I production cells were diluted to about 1 x lo5 cells/ml in serum-free medium and 1 ml of this suspension was pipetted into each 75 X 12 mm culture tube (Sterilin). The appropriate concentration of hormone dissolved in PBS was added and the cells were incubated at
37 o C under 95% air : 5% CO, for the required period. The tubes were then centrifuged (2000 x g for 30 min) and the supematant was removed for assay of IGF-I. Labelling of hormones IGF-I and hGH used in the receptor assays and immunoassays were ‘251-labelled by the iodogen method (Fraker and Speck, 1978). The iodinated hormones were partially purified on a Sephadex G-50 column (12-20 cm x 1 cm) eluted with PBS containing 0.1% BSA and 0.6 mM merthiolate (pH 7.4) (hGH) or Tris-HCl buffer (25 mM Tris/HCl, 0.1% BSA, 0.6 mM merthiolate, pH 7.6) (IGF-I). Aliquots of labelled hormone were stored at 4’ C until required. Binding of ‘251-labelled hGH to IM-9 cells IM-9 cells were washed 3 times in sterile PBS and then resuspended in serum-free medium (RPMI; 25 mM Hepes, 0.1% BSA, pH 7.4) at a concentration of about 1.5 x lo6 cells/ml. 1.0 ml of this cell suspension was then pipetted into each of the plastic culture tubes (75 mm X 12 mm, Sterilin) used in the assay. 100 ~1 of unlabelled hormone (of the appropriate concentration, dissolved in PBS) was added, followed by 100 ~1 of ‘251-labelled hGH (about lo5 cpm). After mixing of the contents, tubes were incubated at 30 o C for 2 h in an atmosphere of 95% air: 5% CO,. The cells were then precipitated by centrifugation (2000 X g, 30 mm). Supernatants were removed and the precipitates were subjected to gamma counting. IGF-I extraction 0.25 ml of sample were acidified with 1 ml of 0.5 M HCl and then incubated at 20-23” C for 30 min. Adsorption on Sep-pak Cl8 cartridges was used to separate the IGF-I from protein (including binding protein) in the sample (Wallis et al., 1987). Each cartridge was prepared by washing successively with isopropanol (5 ml), methanol (5 ml) and 4% (v/v) acetic acid (15 ml). The sample was then applied (flow rate approximately 0.4 ml/mm) and the cartridge was washed with 4% (v/v) acetic acid (20 ml) to remove unadsorbed material. IGF-I was then eluted with 4 ml of methanol. This elute was dried at 37’ C in a stream of air, dissolved in
169
IGF-I radioimmunoassay buffer (0.5 ml), and incubated at 37°C for 10 min to ensure complete solution. The extraction method used gave a recovery for IGF-I of 70-908, and reproducibility between assays was generally good. For a series of ten consecutive assays the interassay coefficient of variation for a control sample subjected to extraction and assay was 18%. Radioimmunoassay
of IGF-I
The IGF-I content of medium samples, before or after extraction, was carried out by radioimmunoassay (Wallis et al., 1987) using antiserum to human IGF-I obtained from NIADDK, 1251labelled IGF-I, and as standard a partially purified IGF-I preparation (N4G50 II) obtained from Dr. A. Holder at the National Institute for Child Health, London, U.K. This standard has a potency of about 30 units/mg and has been calibrated by radioimmunoassay with reference to the Amgen somatomedin C as containing 1.31 pg/mg (Wallis et al., 1987). Results are expressed as ng equivalents IGF-I/ml, based on this potency for the standard. The antibody was reported to show less than 5% crossreactivity with IGF-II, and to crossreact minimally with insulin (notes received from NIADDK). Treatment
of results
All experimental and control treatments were performed in triplicate, at least. Results are expressed as mean values k SEM. The statistical significance of differences between groups of observations was tested using Student’s t-test. Results Earlier work (Lesniak et al., 1974,1977; Lesniak and Roth, 1976) has shown that the IM-9 cell line can bind ‘251-labelled hGH, and we have confirmed this observation. Intact IM-9 cells bound approximately 14.5% of the ‘251-labelled hormone added, of which about 8% was specifically bound (displaced by excess, 10 pg, unlabelled hGH). Unlabelled hGH competed for binding sites in a dose-dependent manner (Fig. 1). Ovine GH also displaced ‘251-labelled hGH, with a potency approximately 10% that of hGH. Rat GH and bovine insulin did not displace ‘251-labelled hGH. Rat
Fig. 1. Binding of 1251-labelled hGH to IM-9 cells. IM-9 cells (approx. 1.5 x lo6 cells/tube) were incubated with ‘251-labelled hGH (approx. 10s cpm) in the absence or presence of unlabelled human (0) ovine (0) or rat (W) GH at the concentration shown. Binding to cells was measured as described in the text, and is expressed as a percentage of the total cpm present. Values are means of four replicates; vertical lines show + SEM.
and bovine prolactin did so with very low potency (< 2% that of hGH). When IM-9 cells were incubated in serum-free medium in the presence of hGH, immunoreactive IGF-I was released into the medium. The effect of hGH was slow; a significant effect (P < 0.05) was seen only after 24 h (Fig. 2). However, substantial effects required 48-120 h of treatment. Cells incubated without added hGH showed a small, but significant, increase in medium content of IGF-I during the course of the incubation. Medium at time zero contained a low IGF-I content, possibly due to carry-over from the serum-containing medium in which the cells were grown, as assays performed on non-conditioned serum-free medium failed to detect significant quantities of IGF-I. The effect of hGH on IGF-I production by IM-9 cells was dose-related. Fig. 3 shows the effect of incubating cells with various concentrations of hGH on the release of IGF-I from IM-9 cells after a 5-day incubation. The response to increasing concentrations of hGH was bimodal with maxima at 15 and 1000 ng/ml hGH. The size of the response obtained at 1000 ng/ml was rather variable, being as great as that seen at 15 ng/ml in some experiments, lower in others. The bimodal dose-response relationship was seen both when
170
i4
48
72 lncllbatii
ee
120
time(h)
Fig. 2. Actions of hGH on IGF-I production by IM-9 cells. IM-9 cells (lo5 cells/ml) in serum-free medium were incubated in the absence (m) or presence of 10 (0) or 1000 (0) ng/ml hGH. IGF-I concentration in the medium was determined at the times shown by radioimmunoassay after extraction on Sep-pales. Values shown are means of four replicates; vertical lines show f SEM. Significant stimulation of IGF-I production by hGH (comparing treated and untreated for any given time) is indicated: * P < 0.05; * * P c 0.01.
IGF-I was measured directly in medium and when it was extracted first using Sep-paks (Fig. 3) suggesting that it is not a consequence of interac-
tkmentmtii
of hGHhg/ml)
Fig. 3. Actions of hGH on IGF-I production by IM-9 cells: dose-response curve. IM-9 cells (lo5 cells/ml) were incubated for 5 days in serum-free medium in the presence of the hGH concentrations shown. IGF-I concentrations in the media were then determined by radioimmunoassay either directly (shaded bars) or after extraction on Sep-paks (open bars). Values are means of four replicates; vertical lines show f SEM. hGH gave a significant stimulation of IGF-I secretion at all concentrations tested (P < 0.05 for unextracted values at 500 ng/ml and lo4 ng/ml hGH; P < 0.01 for all other observations).
Concentration d GH(ng(m11
Fig. 4. Actions of human, ovine and rat GH on IGF-I production by IM-9 cells. IM-9 cells (lo5 cells/ml) in serum-free medium were incubated for 5 days in the presence of the concentrations shown of human (a), ovine (0) or rat (m) GH. IGF-I concentration in the media was then determined by radioimmunoassay, without extraction on Sep-paks. Values are means of eight replicates; vertical lines show f SEM. * Values significantly different (P i 0.01) from control (no GH).
tion between IGF-I and other components in the medium, such as IGF binding protein. No correction has been made to allow for small losses during the Sep-pak extraction procedure, so the corresponding values in Fig. 3 (open bars) may slightly underestimate IGF-I concentration after extraction. Incubation of conditioned medium from IM-9 cells with ‘251-labelled IGF-I followed by gel filtration suggested that only traces of binding protein were present. The specificity of the effect of hGH on IGF-I release was studied (Fig. 4). In accordance with the results of binding studies, human GH was most effective in stimulating IGF-I release. Ovine GH gave a smaller response, and a less marked bimodal dose-response curve. Rat GH had little effect on IGF-I release from the cells, though it did give a significant stimulation (P < 0.001) at the highest concentration tested. These results suggest that the response seen at lower concentrations (- 10 ng/ml) of hormone is more specific for hGH than the response brought about by much higher hormone concentrations.
171
mained at over incubation.
80% throughout
the
period
of
Discussion
I
O
1
1
.
*
~
234567 Incubation time(days]
Fig. 5. Growth of IM-9 cells in serum-containing and serum-free medium. IM-9 cells were incubated in medium containing 10% foetal calf serum (closed symbols) or without serum (open symbols) in the absence (0, 0) or presence of 10 ng/ml (& 0) or 1000 ng/ml (A, A) hGH. Cell numbers were determined at the times shown, using a haemocytometer. Values shown are means of four determinations; vertical bars show SEM.
The studies on IGF-I production were carried out in serum-free medium, and the ability of IM-9 cells to survive and grow in such medium was investigated (Fig. 5). In the presence of foetal calf serum, cells grew rapidly during the first 2 days of a 7-day incubation; growth then slowed, and there was no change in cell number between days 3 and 7. After transferral of cells to serum-free medium, cell proliferation stopped, but after 2-3 days cell numbers began to increase slowly. The presence of hGH had no significant effect upon growth either in the normal growth medium or in serum-free medium. The viability of the cells (measured by trypan blue exclusion and by ability to recommence growth when transferred to fresh serum-containing medium) fell rapidly when they were placed in serum-free medium, dropping to 50-55% and then remaining constant (results not shown). In serum-containing medium viability re-
Previous studies on IM-9 cells have shown that they bind ‘251-labelled hGH (Lesniak et al., 1974, 1977; Lesniak and Roth, 1976; Hughes et al., 1983; Asakawa et al., 1986), but have provided no evidence for a biological response to such binding apart from modulation of receptor levels (Lesniak and Roth, 1976). We have confirmed previous results showing that IM-9 cells can bind 1251labelled hGH specifically. Binding was species specific in that ‘251-labelled hGH could be displaced by unlabelled hGH, but not by rat GH, and only with low potency by ovine GH. These results agree with those presented previously (Lesniak et al., 1974) although the apparent potency of ovine GH was rather greater than that found by others; they also accord with the observed biological specificity of the actions of GH in man. We have also shown that hGH, and to a lesser extent ovine GH, can stimulate somatomedin C/IGF-I production by IM-9 cells. The effect was dose-dependent, but showed an unusual, bimodal dose-response relationship (Figs. 3 and 4), with maximal activity at lo-20 ng/ml and 1000 ng/ml, and decreased activity at intermediate concentrations. The effect was also of rather slow onset (Fig. 2), requiring exposure of cells to the hormone for more than 5 h for significant stimulation, and exposure for at least 48 h before IGF-I output from the cells was doubled. This slow onset accords with the idea that stimulation of production of IGF-I rather than of secretion of preformed growth factor is involved, though direct studies of effects on synthesis of IGF-I are needed to confirm this idea. The slow onset applied to stimulation by both low (10 ng/ml) and high (1000 ng/ml) GH concentrations. The explanation for the bimodal dose-response relationship observed here is not clear. It is possible that it results from the presence of more than one active (or inhibitory) component in the hGH preparation used, though the preparation was a highly purified one. Another possibility, that the bimodal dose-response curve might be due to
172
secretion of IGF binding protein, was explored by assaying the apparent growth factor concentration both before and after extraction using Sep-pak columns. Such extraction did not affect the pattern of response observed, suggesting that production of binding protein does not contribute to its bimodal nature. It is possible that if a very large quantity of binding protein was produced the Sep-pak procedure might not extract IGF-I effectively, though it would be surprising in that case if the dose-response pattern was completely unaltered, as was actually observed. Incubation of 125I-labelled IGF-I with conditioned medium provided no evidence for the presence of large amounts of binding protein. Other cellular systems showing enhanced IGF-I production in response to GH, such as liver cells (Spencer, 1979) and cultured fibroblasts (Atkison et al., 1980; Clemmons et al., 1981) do not appear to show a bimodal response of this kind. It is possible that it is associated with the prolonged exposure to GH necessary for demonstration of a good response, or to the transfer to serum-free media in which cells are unable to proliferate. Down-regulation/ internalization of GH receptors is known to follow hormone binding to IM-9 cells (Lesniak and Roth, 1976; Rosenfeld and Hintz, 1980b) and it is possible that such down-regulation caused by a relatively high concentration of GH could lead to lowered responsiveness to the hormone. It is notable, however, that extensive receptor down-regulation can be produced by a concentration of hGH (about 15 ng/ml) that gives a maximal response for IGF-I production. The specificity of the response to the hormones tested was similar to the specificity of their ability to bind to receptors on IM-9 cells. Thus, the response to ovine GH was much less than that to the human hormone, and the response to rat GH was very small. It is perhaps surprising that such differences related mainly to the absolute response rather than the concentrations at which responses were observed, but the bimodal nature of the dose-response curve makes this feature difficult to assess. Since calf serum contains IGF-I which would crossreact with and mask IGF-I produced by IM-9 cells, the studies carried out on these cells were performed in serum-free medium. The experiment
illustrated in Fig. 5 showed that cell proliferation is stopped by transfer to such medium. A substantial proportion of the cells die soon after this transfer, but the remainder remain viable for at least 7 days and presumably represent the cells that respond to GH by producing IGF-I. Clearly, the serum-free medium used for the experiments is not ideal, and the transfer could be partly responsible for the very delayed response seen to GH. A serum-free medium that supports sustained growth would facilitate the use of IM-9 cells to study the actions of GH. It should be noted, however, that addition of hGH to culture medium, with or without serum, had no effect on the proliferation of the cells. A number of cell lines have now been shown to produce IGF-like material, though only in a few is this production stimulated by GH. Atkison et al. (1984) have shown that primary cultures of rat hepatocytes and cell lines derived from human epidermoid carcinoma, pancreatic carcinoma, Simian virus 40 transformed human skin fibroblasts and normal skin fibroblasts all release IGF-like factors into their growth media; only for the rat hepatocytes was it shown that GH enhanced IGF-I production. Insulin-like and IGFlike factors have also been found in media conditioned by other types of transformed cell, including lines derived from human myeloid leukaemia (HL-60) (Yamanouchi et al., 1985) human osteosarcoma (Blatt et al., 1984) and rat liver (BRL3A) (Rechler et al., 1981); in none of these cases was GH reported to stimulate growth factor production. GH has, however, been reported to stimulate IGF-I production by some human fibroblast cell lines (Atkison et al., 1980; Clemmons et al., 1981). A previous study detected no IGF-I production by IM-9 cells (Atkison et al., 1984); details of this work were not given, but it is clear that the delayed responsiveness of the cells and the bimodal dose-response curve could make it difficult to detect the action of hGH on these cells. Whether the effect seen with IM-9 cells reflects equivalent actions in normal lymphocytes is not clear, nor is the physiological significance. It does provide, however, a potentially useful model for studies on the mechanism of action of GH, using a cell line that is convenient to work with and on which much previous work on GH receptors has been
173
reported. It should also be noted that IM-9 cells themselves carry somatomedin receptors (Rosenfeld and Hintz, 1980; Duronio et al., 1986) so autocrine effects cannot be excluded. Acknowledgements We thank the Medical Research Council for financial support, Drs. S. Raiti, A. Parlow and the National Institute of Diabetes, Digestive and Ridney Diseases for gifts of hormones and antisera, Mrs. Margaret Daniels for skilled production of the figures and Mrs. Eileen Willis for expert preparation of the manuscript. References Asakawa, K., Hedo, J.A., McElduff, A., Rouiller, D.G., Waters, M.J. and Gorden, P. (1986) B&hem. J. 238, 379-386. Atkison, P.R., Weidman, E.R., Bhaumick, B. and Bala, R.M. (1980) Endocrinology 106, 2006-2012. Ad&on, P.R., Hayden, L.J., Bala, R.M. and Hollenberg, M.D. (1984) Can. J. B&hem. Cell. Biol. 62, 1343-1350. Blatt, J., White, C., Dienes, S., Friedman, H. and Foley, T.P. (1984) Biochem. Biophys. Res. Commun. 123, 373-376. Clemmons, D.R., Underwood, L.E. and Van Wyk, J.J. (1981) J. Clin. Invest. 67, 10-19. Duronio, V., Jacobs, S. and Cuatrecasas, P. (1986) J. Biol. Chem. 261, 970-975. Fahey, J.L., Buell, D.N. and Sox, H.C. (1972) Ann. N.Y. Acad. Sci. 190, 221-234.
Fraker, P.J. and Speck, J.C. (1978) B&hem. Biophys. Res. Commun. 80, 849-857. Gavin, J.R., Roth, J., Neville, D.M., De Meyts, P. and Buell, D.N. (1974) Proc. Natl. Acad. Sci. U.S.A. 71, 84-88. Hughes, J.P., Simpson, J.S.A. and Friesen, H.G. (1983) Endocrinology 112, 1980-1985. Jacobs, S., Cook, S., Svoboda, M.E. and Van Wyk, J.J. (1986) Endocrinology 118, 223-226. Lest&k, M.A. and Roth, J. (1976) J. Biol. Chem. 251, 3720-3729. Lesniak, M.A., Gorden, P., Roth, J. and Gavin, J.R. (1974) J. Biol. Chem. 249, 1661-1667. Lesniak, M.A., Gorden, P. and Roth, J. (1977) J. Clin. Endocrinol. Metab. 44, 838-849. Misra, P., Hintz, R.L. and Rosenfeld, R.G. (1986) J. Clin. Endocrinol. Metab. 63, 1400-1405. Rechler, M.M., Nissley, S.P., King, G.L., Moses, A.C., Van Obberghen-Schilling, E.E., Romanus, J.A., Knight, A.B., Short, P.A. and White, R.M. (1981) J. Supramol. Structure Cell. B&hem. 15, 253-286. Rosenfeld, R.G. and Hintz, R.L. (1980a) Endocrinology 107, 1841-1848. Rosenfeld, R.G. and Hintz, R.L. (1980b) J. Clin. Endocrinol. Metab. 51, 368-375. Scott, C.D., Martin, J.L. and Baxter, R.C. (1985) Endocrinology 116, 1102-1107. Spencer, E.M. (1979) FEBS Lett. 99, 157-161. Wallis, M., Daniels, M., Ray, K.P., Cottingham, J.D. and Aston, R. (1987) B&hem. Biophys. Res. Commun. 149, 187-193. Yamanouchi, T., Tsushima, T., Kasuga, M. and Takaku, F. (1985) Biochem. Biophys. Res. Commun. 129, 293-299.