Proc. Natl. Acad. Sci. USAVol. 83, pp. 6854-6857, September 1986Cell Biology

Growth hormone-releasing factor stimulates proliferation ofsomatotrophs in vitro

(DNA synthesis/cyclic AMP/somatostatin/growth hormone/pituitary)

NILS BILLESTRUP*, LARRY W. SWANSONt, AND WYLIE VALE*t*The Clayton Foundation Laboratories for Peptide Biology, and tThe Howard Hughes Medical Institute, The Salk Institute, 10010 North Torrey Pines Road,La Jolla, CA 92037

Communicated by Renato Dulbecco, June 11, 1986

ABSTRACT The mitogenic effect of the hypothalamicpeptides growth hormone-releasing factor (GRF) and somato-statin on cultured growth hormone (GH)-producing cells(somatotrophs) was studied. Using autoradiographic detectionof [3H]thymidine uptake and immunocytochemical identifica-tion of GH-producing cells, we show that 5 nM GRF causes a20-fold increase in the percentage of somatotrophs labeled with[3H]thymidine. The total number of somatotrophs in GRF-treated cultures was increased by 60%. Somatostatin had nomeasurable effect on the labeling index by itself, but it partlyinhibited the GRF-induced increase in both the labeling indexand the total number of cells. Forskolin caused an increase inboth the percentage of somatotrophs with a [3H]thymidine-labeled nucleus and the somatotroph number similar to thatcaused by GRF. GH secretion as well as cellular GH content inthe GRF- or forskolin-treated cells increased with culture timeover the entire period, whereas secretion and content of GHgradually decreased in control or somatostatin-treated culturesduring the entire culture period. These data suggest that GRFand somatostatin regulate the mitotic activity of GH-producingcells and that the effect of GRF is possibly mediated by cyclicAMP.

The biosynthesis and secretion of growth hormone (GH) inthe anterior pituitary gland is under complex hormonalregulation. The hypothalamic peptide growth hormone-releas-ing factor (GRF) stimulates both release (1-3) and synthesis(4, 5) of GH, as well as transcription of the GH gene (6, 7),whereas somatostatin depresses GH secretion (8). Patientswith GRF-producing tumors (9-12) and transgenic miceexpressing large amounts of GRF (13) have chronicallyincreased plasma GRF with resulting high levels of plasmaGH, pituitary enlargement, increased somatic growth, andacromegaly. This suggests but does not establish that GRFstimulates the growth ofpituitary somatotrophs, the cells thatsecrete GH. Not only might the effects ofGRF in these in vivostudies be indirect or require the interaction of othermitogenic substances, but also other products of the GRF-producing cells might be involved. We therefore chose toexplore the effect of long-term GRF administration on thegrowth and function of cultured rat somatotrophs. The resultsshow that GRF stimulates the mitotic activity in somato-trophs as measured by cell counts in pituitary cultures treatedwith [3H]thymidine and immunohistochemically stained forGH. The specificity of this response was indicated by theobservation that somatostatin attenuates the growth-pro-moting effect of GRF, whereas prepro-GRF-(75-104) had nomeasurable influence.

MATERIALS AND METHODSCell Culture. Rat anterior pituitary cells from male

Sprague-Dawley rats (Holtzmann, Madison, WI) were pre-pared as described (14) and seeded in two-well slides, coatedwith an extracellular matrix from bovine pituitary endothelialcells. Cells were cultured in 1.5 ml of minimal essentialmedium containing D-valine instead of L-valine (GIBCO) andsupplemented with 5 nM dexamethasone, 30 pM triiodothy-ronine, 0.8 ,tM insulin, and 60 nM transferrin. After 2 days,new medium containing 2% fetal calf serum was added andsynthetic rat GRF (rGRF), prepro-GRF-(75-104), somato-statin-14 (generously provided by J. Rivier, The Salk Insti-tute), forskolin, or dilution buffer (1 mM acetic acid) wasadded to start treatment. Cells were incubated in a humidifiedatmosphere containing 7.5% CO2 at 37°C, and the culturemedium was changed every 4 or 5 days.Immunocytochemistry and Autoradiography. After 18 days

of treatment, the cells were incubated in fresh mediumcontaining 10 ,uCi (1 Ci = 37 GBq) of [3H]thymidine (6.7Ci/mmol; New England Nuclear) for 3 hr and fixed in 4%(wt/vol) paraformaldehyde for 15 min. The cells were thenwashed in potassium phosphate-buffered saline (PBS) andincubated overnight at 4°C in PBS containing 0.3% TritonX-100 and 2% normal goat serum, followed by a 48-hrincubation with a 1:8000 dilution of monkey antiserum to ratGH in PBS containing 2% normal goat serum. Cells were thenwashed and incubated with a 1:250 dilution of biotinylatedgoat anti-human IgG antiserum (Vector Laboratories,Burlingame, CA) for 45 min, rinsed, and incubated in 1:100dilution of avidin-biotinylated peroxidase complex (VectorLaboratories) for 30 min and finally developed for 5 min ondiaminobenzidine tetrahydrochloride (0.5 mg/ml) solutioncontaining 0.01% H202. No GH staining could be observedwhen the GH antiserum had been preincubated with anexcess of highly purified rat GH. The cell culture slides werethen dehydrated through rinses in ascending concentrationsof ethanol, defatted in 1:1 (vol/vol) chloroform/ethanol, andair dried. Slides were dipped in Kodak NTB2 photographicemulsion at 42°C, air dried at 24°C in a humidified atmo-sphere, and left at 4°C. After 10-14 days of exposure, theslides were developed in Kodak D-19 for 3 min, washed inwater and fixed in Kodak rapid fixer for 6 min and finallyrinsed in water for 45 min. Cells were then dehydratedthrough ascending concentrations of ethanol, counter-stained with eosin, and coverslipped in D.P.X. (BDH). Thecells were viewed with a Leitz Dialux 22 microscope with ax20 objective by phase-contrast microscopy. Somatotrophscould clearly be identified as cells having an intensered/brown cytoplasmic staining, whereas cells containing no

Abbreviations: GH, growth hormone; GRF, GH-releasing factor;rGRF, rat GRF.tTo whom correspondence should be addressed.

6854

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

May

24,

202

0

Proc. Natl. Acad. Sci. USA 83 (1986) 6855

GH had only a pale pink eosin counterstaining. Somatotrophsand somatotrophs with a [3H]thymidine-labeled nucleus werecounted by systematically scanning the slides. Experimentswere performed four times, each in triplicate. More than 2000somatotrophs were counted per well. Thus, over 24,000somatotrophs were counted for each experimental variable.GRF and GH Assays. GRF was measured by radio-

immunoassay as described (14). GH in the culture medium orin Nonidet P-40 extracts of cells was measured by radio-immunoassay (kit kindly provided by the National Institute ofArthritis, Diabetes, and Digestive and Kidney Diseases) asdescribed (3).

RESULTS AND DISCUSSIONCultures of rat anterior pituitary cells grown in completemedium (14) containing fetal calf serum are heavily contam-inated with rapidly proliferating fibroblasts and were thus notsuitable for the current study. A 48-hr incubation of the cellsin serum-free medium, followed by culture in medium withD-valine substituted for L-valine and containing 2% fetal calfserum, resulted in cultures virtually without fibroblasts.Under these culture conditions it was, however, necessary togrow the cells on an extracellular matrix (15, 16) to obtain areproducible high plating efficiency with pituitary cells thatassume an epithelial morphology.Rat anterior pituitary cells were cultured in the absence or

presence of 5 nM rGRF. After 18 days of treatment, the cellswere labeled with [3H]thymidine for 3 hr, fixed, stained forGH, and processed for autoradiography. Fig. 1 shows theappearance of pituitary cells cultured in the presence of 5 nM

_.

rGRF. Six somatotrophs can be identified clearly, one ofwhich (arrow) displays a [3H]thymidine-labeled nucleus.rGRF caused a 20-fold increase in the percentage ofsomatotrophs labeled with [3H]thymidine (Fig. 2A); an av-erage of 0.97% were labeled in the rGRF-treated group ascompared to an average of 0.05% in control cells. The totalnumber of somatotrophs in the rGRF-treated cultures wasalso significantly higher compared to controls (Fig. 2B),confirming that the increased uptake of [3H]thymidine actu-ally represents increased mitotic activity of the somato-trophs. We also examined the possible proliferative effect ofa pro-GRF peptide in view of the proposed role of parts of thecorticotropin precursor in the regulation of adrenal corticalgrowth (17). No effect ofprepro-GRF-(75-104) on the uptakeof [3H]thymidine was observed, nor was the number ofsomatotrophs in these cultures different from controls (datanot shown). Somatostatin at 5 nM had no measurable effectby itself, but it did attenuate the rGRF-induced increase inboth the percentage of somatotrophs labeled with[3H]thymidine and the total number of somatotrophs. Prepro-GRF-(75-104) did not influence the effects of rGRF on[3H]thymidine uptake or somatotroph growth.The labeling index in both control and GRF-treated cells

after 9 days was indistinguishable from that at 18 days. Thenumber of somatotrophs in control cultures was the same atday 9 and day 18, indicating that the data cannot be explainedby a survival effect of GRF. At day 9 the number ofsomatotrophs in the GRF-treated cultures was 31% abovecontrol vs. 61% at day 18, suggesting a gradual increase insomatotroph number with culture time, caused by themitogenic action of GRF.

4D

FIG. 1. Photomicrograph of GH-immunostained rat anterior pituitary cells, showing six somatotrophs, one of which (arrow) displays a

[3H]thymidine-labeled nucleus. The cells were cultured on slides in the presence of 5 nM rGRF for 18 days. (x 1300.)

Cell Biology: Billestrup et al.

A

i,

N4It

:t.-4.

Dow

nloa

ded

by g

uest

on

May

24,

202

0

6856 Cell Biology: Billestrup et al.

0C

E

I.0

-o

C,

m2

.o

0

E

co

0

a,

Control rGRF SS-14 rGRF Forskolin+SS-14

20cJ0

41-0

0

30

0

20)0

E0C/)

Control rGRF SS-14 rGRF Forskolin+SS-14

FIG. 2. Effect of 5 nM rGRF, 5 nM somatostatin-14 (SS-14), or 10 AM forskolin on rat anterior pituitary somatotroph proliferation. The doseof GRF was chosen to resemble the GRF concentration in plasma of patients with GRF-producing tumors (9) and transgenic mice (10). Somedegradation of GRF occurred during the culture period, since only 60-70%o of the added GRF was immunoreactive after 5 days of incubation.(A) Percent somatotrophs with a 3H-thymidine-labeled nucleus. (B) Total number of somatotrophs per well as percent of control. The mean ±SD for four experiments each done in triplicate is shown.

The function of the somatotrophs was evaluated by mea-suring GH secretion as well as cellular GH content during theculture period. A decrease in GH secretory rates over first 5days in culture was observed in all cultures. Both GHsecretory rates and GH content of somatotrophs culturedwithout rGRF gradually decreased during the entire cultureperiod (Fig. 3). In contrast, GH secretion and content in therGRF-treated cells increased by 50% from day 9 to day 18.This corresponds well to the 60% increase in the number ofsomatotrophs observed at day 18. Thus, the increase in GHproduction in long-term GRF-treated cultures could in part beexplained by an increase in the number of somatotrophs.

Cyclic AMP has been proposed as a mediator for the GRFstimulation of GH secretion and gene transcription (7). Thediterpene compound forskolin has been shown to stimulateGH secretion and transcription of the GH gene by increasingintracellular cyclic AMP concentration (7). Treatment of ratpituitary cells with 10 ,uM forskolin for 18 days resulted in a25-fold increase in the percentage of somatotrophs with a[3H]thymidine-labeled nucleus. The total number of somat-otrophs in each well was also increased to an extent similar

to that caused by rGRF itself (Fig 2). Secretion of GH andcellular GH content in forskolin-treated cells also increasedwith time in culture (Fig. 3).These data suggest that GRF may stimulate the mitotic

activity of rat somatotrophs and that GRF can maintainelevated GH production by somatotrophs in long-term cul-ture. Because only GH-positive cells were analyzed in thisstudy, these results show thatDNA synthesis is stimulated byGRF in fully differentiated somatotrophs. However, it cannotbe ruled out that GRF also stimulates differentiation of ahypothetical stem cell to a somatotroph phenotype. Further-more, it has been demonstrated that somatostatin can partlyinhibit the growth-promoting effect of rGRF on so-matotrophs, perhaps by virtue of the ability of somatostatinto reduce the GRF-stimulated rise in cyclic AMP (18).Because forskolin stimulates somatotroph growth, theseresults also suggest that the GRF-induced growth ofsomatotrophs is mediated at least in part by cyclic AMP andthat the level of cyclic AMP is an important factor indetermining the mitotic activity of somatotrophs. CyclicAMP is also able to stimulate proliferation of Schwann cells

10

Days of treatment15 20

1250 -

C

= 1000-

03.

CL 750c

08 500-

2I5

250-

BT

T

Days of treatment: 9 18Control

9 18 9 18 9 18rGRF SS-14 rGRF

+SS-14

FIG. 3. Time course of action of 5 nM rGRF, 5 nM somatostatin-14 (SS-14), or 10 ,uM forskolin on GH release from (A) and GH contentin (B) rat anterior pituitary cells. Immunoreactive GH in the culture media or in Nonidet P-40 extracts of cells was measured byradioimmunoassay (4). (A) GH secretion per day per well is plotted as a function of culture time. The mean ± SD for four experiments is shown.(B) GH content per well in cells treated for 9 or 18 days. The mean ± SD for four experiments is shown.

1750 -

= 1500 -

0.3a, 1250 -

-o 1000-0caC 750-0._o0o 500-0)

I(D 250-

A

I 1-IrGRF+SS-14

Control

SS-14

0 59 18

Forskolin

Proc. Natl. Acad Sci. USA 83 (1986)

Dow

nloa

ded

by g

uest

on

May

24,

202

0

Proc. Natl. Acad. Sci. USA 83 (1986) 6857

in vitro (19), whereas it exhibits an inhibitory effect onfibroblast growth (20).A stimulatory effect of GRF on somatotroph proliferation

could explain the pituitary hyperplasia observed in patientswith GRF-producing tumors (9-12, 21). In addition, it hasbeen observed in a number of patients that the presence ofneuronal tumors containing GRF immunoreactivity is asso-ciated with pituitary somatotroph adenomas and acromegaly(9). The ability of somatostatin to reduce the GRF-stimulatedproliferation of somatotrophs supports a possible therapeuticuse of somatostatin or more selective analogs in the man-agement of GH-secreting adenomas. The role played byGRF, acting either alone or in combination with otherfactors, in inducing somatotroph transformation andadenoma formation has yet to be established.

We thank Gayle Yamamoto, Diane Jolley, Anne Corrigan, andDonna Simmons for technical assistance and Bethany Connor andSusan McCall for manuscript preparation. This research was sup-ported in part by National Institutes of Health Grant AM-26741 andThe Keck Foundation. The research was conducted in part by theClayton Foundation for Research, California Division. W.V. is asenior Clayton Foundation investigator. N.B. is the recipient offinancial support from the Danish Medical Research Council(12-5157) and the Carlsberg Foundation.

1. Rivier, J., Spiess, J., Thorner, M. & Vale, W. (1982) Nature(London) 300, 276-278.

2. Guillemin, R., Brazeau, P., Bohlen, P., Esch, F., Ling, N. &Wehrenberg, W. B. (1982) Science 218, 585-587.

3. Spiess, J., Rivier, J. & Vale, W. (1983) Nature (London) 303,532-535.

4. Bilezikjian, L. M. & Vale, W. (1984) Endocrinology 115,2032-2034.

5. Fukata, J., Diamond, D. S. & Martin, J. B. (1985) Endocrinol-ogy 117, 457-467.

6. Barinaga, M., Yamamoto, G., Rivier, C., Vale, W., Evans, R.& Rosenfeld, M. G. (1985) Nature (London) 306, 84-85.

7. Barinaga, M., Bilezikjian, L. M., Vale, W. W., Rosenfeld,M. G. & Evans, R. M. (1985) Nature (London) 314, 279-281.

8. Brazeau, P., Vale, W., Burgus, R., Ling, N., Batcher, M.,Rivier, J. & Guillemin, R. (1973) Science 179, 77-79.

9. Asa, S. L., Scheithauer, B. W., Bilbao, J. M., Horvath, E.,Ryan, N., Kovacs, K., Randall, R. V., Laws, E. R., Singer,W., Linfoot, J. A., Thorner, M. 0. & Vale, W. (1984) J. Clin.Endocrinol. Metab. 58, 796-803.

10. Zafar, M. S., Mellinger, R. C., Fine, G., Szabo, M. & Froh-man, L. A. (1979) J. Clin. Endocrinol. Metab. 48, 66-71.

11. Frohman, L. A., Szabo, M., Berelowitz, M. & Stachura,M. E. (1980) J. Clin. Invest. 65, 43-54.

12. Leverston, S. A., McKeel, D. W., Buckley, P. J.,Deschryver, K., Greider, M. H., Jaffe, B. M. & Daughaday,W. H. (1981) J. Clin. Endocrinol. Metab. 53, 682-689.

13. Hammer, R. E., Brinster, R. L., Rosenfeld, M. G., Evans,R. M. & Mayo, K. E. (1985) Nature (London) 315, 413-416.

14. Vale, W., Vaughan, J., Yamamoto, G., Bruhn, T., Douglas,C., Dalton, D., Rivier, C. & Rivier, J. (1983) MethodsEnzymol. 103, 565-577.

15. Vlodavsky, I., Lui, G. M. & Gospodarowicz, D. (1980) Cell19, 607-616.

16. Spira, O., Vlodavsky, I., Vlmansky, R., Atzmon, R., Fuks, Z.,Gordon, A. & Gross, J. (1983) Acta Endocrinol. (Copenhagen)104, 279-286.

17. Lowry, P. J., Silas, L., McLean, C., Linton, E. A. &Estivariz, F. E. (1983) Nature (London) 306, 70-73.

18. Bilezikjian, L. M. & Vale, W. W. (1983) Endocrinology 113,1726-1731.

19. Raff, M. C., Hornby-Smith, A. & Brockes, J. P. (1978) Nature(London) 273, 672-673.

20. Pastan, I. H., Johnson, G. S. & Anderson, W. B. (1975) Annu.Rev. Biochem. 44, 491-522.

21. Thorner, M. O., Perryman, R. L., Cronin, M. J., Rogol,A. D., Draznin, M., Johanson, A., Grindeland, R., Vale, W.,Horvath, E. & Kovacs, K. (1982) J. Clin. Invest. 70, 965-977.

Cell Biology: Billestrup et al.

Dow

nloa

ded

by g

uest

on

May

24,

202

0