Pituitary Regulation of Human Growth Hormone Binding Sites in Rat Liver Membranes A. C. Herington, We have studied the binding human growth hormone (hGH)

L. S. Phillips, and W. H. Daughaday of ‘151to crude

100,000 x g membmne preparations from rat liver, and have studied factors which might regulate the capacity and affinity of hGH binding sites. Membrane prepamtions from livers of pregnant mts bound between 8% and 13% of the “‘IhGH initially added, and 70%~80% of that bound was displaced by 1 pg of unlabeled hGH. Human prolactin (hPrl) displaced 12’l-hGH in a manner parallel to hGH itself but with about one-third the potency. Ovine, porcine, and rat Prl, and mt and bovine GH were much less effective. Scatchard analysis of specific hGH binding by a variety of different mt liver membmne prepamtions revealed a single order of binding site in each case with a binding afTinity of 0.93-1.62 x 10s M-‘. Membmnes from pregnant mts had twice the binding capacity of membmnes from nonpregnant female mts, and about six times the capacity of sites present in preparations from normal adult male rats and hypophysectomized (Hx) male or female rats. Female or male rats with extremely

high circulating OH and Prl levels, due to the presence of tmnsplantable GH/Prl secreting pituitary tumors showed a significantly greater binding capacity than did the pregnant mts. Estmdiol (Es) treatment (25 pg/day for 10-12 days) of normal male rats led to an increase in specific hGH binding. Treatment of hypophysectomized male rats with bovine OH (100 or 500 pg/day) f Es (25pglgday) for S-10 days stimulated both body weight gain and the incorporation of sulfate by cartilage from the treated mts, but no significant increase was observed in the characteristics of ‘251-hGH binding. These results indicate that high levels of Es, GH, and/or Prl play an important role in the regulation of hGH binding sites in mt liver membmnes. The restoration of binding sites in liver from hypophysectomized mts, however, apparently additional factors which are as identified. The role of the hGH sites in the physiologic actions of remains to be determined.

requires yet unbinding GH also

T

HE INITIAL STEP in growth hormone (GH) action may be an interaction of the hormone with specific cell membrane receptors, although this has not been unequivocally established. Binding of GH by liver membranes has been recently reported from several laboratories. Sonenberg et al.lJ observed that bovine GH caused specific perturbation of rat liver membrane conformation as measured by fluorescent and circular dichroism techniques. Tsushima and Friesed described specific binding of “‘I-human GH (hGH) to crude 100,000 x g membrane preparations from the livers of pregnant rabbits and Kelly et a1.4reported hGH binding to similar membrane preparations from the

From the Metabolism Division, Department of Internal Medicine, Washington University School of Medicine, St. Louis, MO. Received for publication July 24. 1975. Supported by Research Grant AM01526 and Training Grant AM05027 from the National Instttute of Arthritis. Metabolism. and Digestive Diseases, National Institutes of Health, Bethesda. Md. Reprint requests should be addressed to Dr. William H. Daughaday, Internal Medicine, Metabolism Division, Washington, University School of Medicine, 660 South Euclid, St. Louis, MO. 63110. c 1976 by Grune & Stratton, Inc. Metabolism, Vol. 25, No. 3 (March), 1976

341

342

HERINGTON,

PHILLIPS,

AND

DAUGHADAY

livers of rabbits, rats, sheep, and pigeons. The binding of hGH by pregnant rabbit liver membranes has provided a successful radioreceptor assay for hGH in serum.‘p6 These latter studies prompted us to investigate the binding of hGH to 100,000 g particulate preparations of rat liver membranes, and to study some of the factors Which might regulate the capacity and affinity of the hGH binding sites in these j3i@arations. MATERIALS

AND METHODS

Normal male, female, and pregnant rats were of the Sprague-Dawley strain, obtained from Eldridge Laboratory Animals, St. Louis. The male hypophysectomized (Hx) rats (Sprague-Dawley) were obtained from Hormone Assay Laboratory, Chicago. Animals bearing the MStT/W15, MStT/WS, MStT/F45, and MtT/F4 tumors were obtained from the Mason Research Institute, Worcester, Mass. The MStT/WlS and MStT/WS tumors were transplanted into the female Wistar-Furth rats obtained from Microbiological Associates, Walkersville, Md. The MStT/F45 and MtT/F4 tumors were transplanted into female Fischer rats obtained from A. R. Schmidt Co., Madison, Wise. All purified hormone preparations used in these studies were gifts provided by the National Pituitary Agency and the National Institute of Arthritis, Metabolism and Digestive Diseases. Lacto eroxidase (B grade) was obtained from Calbiochem, estradiol from Sigma Chemical Co. p . and ‘* I sodium lodlde from Amersham. Iodination (‘*‘I) of hGH (HS 1544C) to specific activities of approximately 100 &i/fig was carried out with lactoperoxidase by a modificatior? of the method of Thorell and Johansson.’ All iodinated hormones were stored frozen in small aliquots and purified prior to use by gel filtration on Sephadex G-100. Crude 15,000-100,000 g membranes from rat liver were prepared according to the method described for pregnant rabbit liver.3 All preparatory steps were performed at 4 C. The livers were removed from the animals, rinsed in 0.3 M sucrose, weighed wet, minced with scissors and either stored frozen or used immediately for membrane isolation. In a limited number of control studies, specific binding of ‘2SI-hGH was indistinguishable in membranes prepared either from fresh liver or from liver stored at - 17 C for up to 7 wk (e.g., in the case of membranes from livers of MtT/F4 rats, specific binding after 0, 5, 12, 48, or 52 days at - 17 C was 20.2x, 18.9%. 19.3x, 22.9x, and 22.4x, respectively). The liver mince was homogenized in 2-3 volumes of 0.3 M sucrose with seven strokes of a glass Dounce homogenizer and then centrifuged for 20 min at 1500 x g. The supernatant was then centrifuged serially at 15,000 x g (20 min) and 100,000 x g (90 min). The 100,000 x g pellet was resuspended in 0.025 M Tris-HCI, pH 7.6, 10 mM in Car&, to a volume of 1 ml per 1 g original wet weight of liver. Each membrane preparation was stored at - 17 C in 1 or 2 ml aliquots and the protein concentration determined.’ Hormone binding was measured in triplicate in a reaction mixture consisting of 200 pl 0.025 M Tris-HCI containing IO mM CaC12 and 0.1% bovine serum albumin (BSA), 100 p1 unlabeled hormone (as standard), 100 ~1 membrane preparation (200 pg protein), and 100 al ‘*‘Ihormone (15,000-25,000 cpm). Total binding (Bo) was measured in the absence of unlabeled hormone and nonspecific binding was taken as the ‘*‘I remaining bound in the presence of I pg unlabeled hormone. Following incubation at room temperature, bound and free hormone were separated by centrifuging a 300 ~1 aliquot of the reaction mixture in a Beckman microfuge (15,000 rpm: 4 min). After aspiration of the supernatant, the microtube tip containing the pellet was cut off and counted. Binding displacement curves (B/b versus hormone concentration) were constructed for each different membrane preparation. The binding site capacity (fmoles/ mg membrane protein) and the hormone binding affinity were obtained for each preparation by conventional Scatchard analysis.’ The stability of the binding sites on the membranes prepared from pregnant rats during incubation at room temperature was examined by allowing the membranes to stand for periods up to 6 hr before addition of ‘*‘I-hGH. The binding of the hormone was then determined after an additional 3 hr. In one experiment no binding activity was lost and in a second experiment the number of counts bound fell by 20%.

343

PITUITARYREGULATION

we have studied the ability of the liver membranes from pregnant rats to degrade the “‘I-hGH by preincubating the labeled hormone with membranes for 2 hr. After removing the membranes by centrifugation the ability of the supernatant “‘1-hGH to bind to fresh membranes was compared to the binding of the labeled hormone which had not been preincubated. There was a 41% decrease in the counts bound. This appears to be greater than the 11.4% degradation of label reported by Posner et al.” In view of the fact that the amount of ‘251-hGH bound to membranes continues to rise slowly with prolonged incubation rather than fall, we suspect that hGH bound to membranes may be protected against degradation. The major errors introduced in the Scatchard calculations by degradation of label to this extent would be in the calculation of binding affinity rather than in the capacity of the binding sites. Although bound growth hormone may be protected against degradation it is still capable of being displaced. In one experiment “‘1-hGH was allowed to react with pregnant rat liver membranes for 2 hr. One microgram of hGH was then added to half the tubes and the incubation was continued for an additional 3 hr. It was noted that 69% of the counts bound at 2 hr were displaced by the addition of excess hGH. Measurements of growth hormone and prolactin in rat serum were made by radioimmunoassay “X using material provided by the National Institute of Arthritis, Metabolism, and Digestive Diseases. Estrone and estradiol were measured by radioimmunoassay by Miss Joan Bauman of the Reproductive Biology Research Foundation using the method of Mikhail et al.,13 and using antibody supplied by Dr. Burton Caldwell. Uptake of 35S-sulfate by cartilage was measured by methods previously described from this laboratory.‘4 RESULTS

The characteristics of hGH binding were initially studied with liver membranes prepared from pregnant female rats (18-20 days pregnant). Total binding of ‘251-hGH to these membranes reached near maximal levels by 150- 180 min (half-maximal binding occurred at 60-90 min) (Fig. 1). There was a gradual increase in total binding between 3 and 7 hr, but for convenience, a reaction time of 3 hr was routinely used for all subsequent binding studies. Nonspecific binding was relatively constant after 1 hr of incubation. Figure 2 shows a typical displacement curve for “‘1-hGH binding at increasing concentrations of unlabeled hormone. In this case, 16.8% of the i251-hGH initially added to the reaction mixture was bound in the absence of unlabeled hormone, and 80% of that bound was displaced by 1 pg of unlabeled hGH. In multiple studies with three different pregnant rat liver preparations, between 8% and 18% of 1251-hGH was bound, of which 70x-90% was displaced by un12 I

Fig. 1. Time course of binding at room kmpamture of ‘*‘I-hOH to pregnant mt liver mombmnes. Both total binding (0) and nonspecific binding (0) a* expressed as a percentage of the total cpm originally added.

$

;/$?-y--:y?~ 2

I

2

3 Time

(hours)

4

5

6

7

HERINGTON,

5 10 20 HORMONE

50 100

500 1000

ADDED(ng/tube)

PHILLIPS,

AND DAUGHADAY

Fig. 2. The offeet of unlabeled pnpamtionr of humon (hGH), bovine (bGH), ond mt growth hormone (rGH) on the binding of lz51-hGH by pngnont mt liver membmnes. Binding (B) is expnssed or the percentage of the binding in the absence of unlabeled hGH (Bo). Bo in this experiment wos

16.8%.

labeled hGH. Figure 2 also shows that rat GH and bovine GH were less effective than hGH itself in competitively inhibiting the binding of “‘1-hGH. In contrast to the substantial specific binding shown by hGH, labeled preparations of rat GH did not exhibit significant binding to pregnant rat liver membranes (data not shown). The specificity of binding of ‘251-hGH was also tested with several prolactin preparations (Fig. 3). Human prolactin (hPr1) competitively inhibited hGH binding in a manner parallel to unlabeled hGH but with about one-third the potency. Ovine, rat, and porcine prolactins were much less effective in displacing ‘*‘I-hGH and generally showed nonparallel displacement. Since crude 100,000 g membrane preparations from pregnant rat liver did bind ‘251-hGH in a specific manner, a single membrane preparation was used

1

IOO-

EO-

Fig. 3. The effect of unlobeled prepamtions of human growth hormone (hGH) ond human (hPrI), ovine (oPrl), porcine (pPr& and rot proloctin (rPrI) on the binding of I-hot4 by pmgnont mt liver membmnes. Binding (B) is expressed OS described for Fig. 2. Bo was 13.1%.

$ -

20 .

5 IO 20

50 100

HORMONE ADDED(ng/tube)

500 lOO(

345

PITUIVARYREGULATION

‘251-hGH BlNDlNG

Figure 4. Displacement curves and katchard plots of ‘s’l-hGH binding to liver mombmnos pmpared from a pregnant rat (a), a normal female mt (o), and a normal male mt (A). For the displacemont curve, binding (6) is expressed as described for fig. 2. For the kotchard plots the honnone bound by each membmno prepamtion is expressed as fmole/mg membmno protein. go was 16.8% for membranes from the pregnant mt, 15.1% for those from the fomale rat, ond 5.6% for those from the male rot.

--iizzq

TO RAT LIVER MEMBRANES o,20/

hGH (ng/tute)

HORMONE

BOUND(fmolr/mo)

as a standard preparation for comparison of the binding of ‘251-hGH by a variety of different rat liver membrane preparations. As shown in a representative experiment, Scatchard analysis of the data of Fig. 4 revealed a single order of a high affinity, low capacity binding sites for the pregnant and nonpregnant female preparations. The intercept on the abscissa of the membranes from pregnant rats showed a greater binding capacity (4 15 fmoles per mg membrane protein) than did the membranes from nonpregnant female rats (340 fmoles/mg), although each showed comparable binding affinity (slope) (K = 1.04 x 10’ M-’ and 1.OO x lo9 M-‘, respectively). The binding site capacity of the liver membranes from male rats, on the other hand, was much lower than for the pregnant rat liver membranes, but the binding by the male preparation was so scattered that calculation of the binding capacity lacked precision. Figure 5 shows the results of another experiment in which membranes from female rats bearing the transplantable GH/Prl secreting pituitary tumor MStT/ W 15 and membranes from hypophysectomized (Hx) rats were compared to the standard pregnant rat membrane preparation. Rats bearing the MStT/WlS tumor have extremely high circulating GH and Prl levelsI and in this case the membranes prepared from one such animal exhibited three times greater initial ‘%hGH

LGH (r&tubs)

BINDING

TO RAT LIVER MEMBRANES

HORMONE

BC4_%D (fmddng)

Pig. 5. Displocemont curves and Scotchard plots of ‘=I-hGH binding to liver membmna prepamd from o pregnant rot (o), o hypophysectomired (Hx) mt (X), ond a hmale ml bearing the tmnsplantoble pituitary tumor MWT/Wl5 (0). All binding pammeters are expressed as doscribod for pigs. 2 ond 4. go for liver membranes from the pregnant mt in this experiment wos 7.9%; it wos 19.1% in membrana from the tumor-bearing animal, and 2.4% in those from the hypophysectomized mt.

346

HERINGTON,

PHILLIPS,

AND

DAUGHADAY

specific binding of “‘1-hGH than did the standard pregnant preparation. Inhibition of binding by unlabeled hGH occurred in a manner comparable to that for the standard pregnant membrane preparation. Scatchard analysis again demonstrated a single binding site for each of the pregnant, tumor, and Hx rat membranes. The binding site capacity of the membranes prepared from the tumor rat was 635 fmole/mg, which was significantly higher than for the membranes prepared from the pregnant rats (172 fmole/mg). The binding affinities were comparable (K = 0.97 x 10’ M-‘) in each case as indicated by the parallel slopes. Liver membranes prepared from the Hx male rats had negligible specific It5I-hGH binding (0.83%). A large number of membrane preparations have been studied and the composite results for the single order of binding site generally observed are given in Table 1. The standard pregnant female preparation (assayed 18 separate times) had a binding capacity of 319 f 27 fmoles/mg (mean + SE) and a binding affinity of 1.16 f 0.07 x 10’ M-’ (mean =t SE). This preparation was used as an interassay control and the values obtained demonstrate the high reproducibility of multiple measurements of the binding parameters, even though different batches of 1251-hGH were used. Including the mean data for the standard pregnant membrane preparation, five different pregnant preparations gave a mean capacity of 327 f 53 fmoles/mg and a mean affinity of 1.37 + 0.30 x 10’ M-l. The mean binding site capacity for five different membrane preparations from normal, nonpregnant female rats was only 50% (p < 0.02) of that for the pregnant rat preparations and the binding affinities were almost identical in Table 1. Binding Characteristicsof “‘I-hGH

by Different Rat liver

Membmne

Preparations

Specific Binding Preparation Standard Total

(n)

pregnant

pregnant

females

Normal

females

Normal

males

Hx Males

female

Binding Affinit

Preparation)

(M--l

(18)

100

(5)

x lo-

J

Binding Site Capacity

)

(fmols/mg

protein)

1.16

& 0.07

319

f

27

iz A.5

1.37

& 0.30

327

f

53

60.7

f

13.8

1.26

f

0.24

162

f

A7*

8.3

f

1.5

1 .A8 iz 0.56

22.5

f

10***

18.2

f

1.9

1.21

52.5

f

20**

108.9

(5)

(3)t

(3)t

Hx Females

(% Pregnant

13.2

(2)

zt 0.30

62.5

1.10

Tumor MStT/W15 MtT/FA

Females Females

145.2

f

48.4

1.03

f

0.30

A97 i

152*

170.8

+

20.9

1.43

f

0.17

A62 f

67*

MStT/WS

Females

(A)

168.7

zk 69.2

0.95

f

0.07

500

f

125*

MStT/FA5

Females

(3)

189.4

f

1.62

& 0.44

665

f

155**

MStT/FA5

Males

All values unlabeled

are

hGH

control obtained

by

compared tSince

for

rat

different

conventional

to that specific

male

as

determined

pregnant

to account

rat

for

binding

was

preparations

**p < 0.01. ***p < 0.001.

mean

& SEM

(for

liver

membrane

batches

the so and

of

low, three

and

Specific

For rot

liver

meaningful of five

The

which binding

statistical

of

as

per

analysis

a

was

run

affinity

analysis

membrane

A52

binding

expressed

preparation,

‘2sl-hGH.

analysis. pregnant

1.07

n > 2).

in each experiment

Scatchard

obtained

lp < 0.05.

12.7

111.2

(2)

expressed was

of the standard

normal

(5)

(6)

of the data

cent

the assay

binding

(Students

could

in the

of

in each

and

preparations

Hx male rat preparations.

‘251-hGH

“t”

site test)

absence

specific as

an

of

binding internal

capacity

were

oil

dato

were

in only

three

of six

(n = 5). be made

347

PITUITARY REGULATION

Table 2. Serum levels of Growth Hormone, Prdoctin, ond htrogon in hmole

Rots Booring Tmnspiontoblr Pituitary Tumon

Growth Tumor

Type

Hormone

fig/ml

Prolactin p9/ml

Estronc ng/lOO

ml

Estrodiol ng/lOOml

7.5 f 3.9

9.3 f 3.0

(6)’ 61.3 f 25.7

(6) 03.5 f 46.6

(3) 3.2

(3) 5.2

MStT/‘WS

(4) 8.2 zt 1.3

(4) 1.9 f 0.5

(2) 3.1

(2) 2.5

MT/F4

(5) 3.9 f 1.5

(5) 1.4 zk 0.8

(2) 3.6 f 1.5

(2) 4.1 + 3.1

(3) 18.9 i 5.1

(3) 0.34 f 0.15

(3) 2.9

(3) 5.7

(6)

(6)

(2)

(2)

Normal

female

0.014 f

MStT/W 15

MStT/F45

All results given as mean f *Figures in parenthew

1.6

0.102 f 60

SE. When only two sera were examined only the mean is given.

indicate number of sera examined.

each case. Although six preparations from normal male rats were studied, only three preparations showed any significant specific binding. These preparations showed a marked reduction in binding capacity compared to the pregnant rat membrane preparations (p < 0.001). No meaningful calculation of binding site capacity was possible with the three preparations showing negligible specific binding. Hypophysectomy of either male or female rats was also accompanied by a marked lowering of binding capacity. Of five membrane preparations from Hx male rats studied, three exhibited specific binding of “‘I-hGH with binding affinity comparable to that of the pregnant preparations. Although these three preparations appeared to have a greater mean binding capacity than did membranes from normal male rats of comparable age, the difference was not significant. (The apparently greater binding affinity for the three normal male preparations probably results in a low calculated value for normal male rat liver binding sites). The total lack of specific binding by two of the membrane preparations from male Hx rats prevented a calculation of binding characteristics in these cases. Membranes prepared from two postpubertal female rats 1 wk after hypophysectomy had a threefold reduction in the binding capacity compared to membrane preparations from normal females, but retained comparable binding affinity. Membrane preparations from female rats each bearing one of four different GH/Prl secreting transplantable pituitary tumors (MStT/WlS, MStT/F45, MtT/F4, MST/WS) were also studied. (The characteristic serum levels of GH, Prl, and estrogens for these rats are shown in Table 2.) All preparations from tumor bearing rats showed a significant increase (approximately threefold) in the capacity of binding sites compared to normal female rat preparations, but showed a similar binding affinity. The increase in capacity resulting from the implanted pituitary tumors was even more apparent in membranes prepared from two male rats bearing the MStT/F45 tumor. With binding affinities of 0.93 x lo9 M-’ and 1.21 x lo9 M-‘, respectively, the binding capacities (697 and 217 fmole/mg) were markedly increased compared to membranes from normal male rats and significantly greater than those from normal female rats.

348

HERINGTON,

lablo 3.

PHILLIPS,

AND

DAUGHADAY

The induction by Estmdiol of hGH Binding Sites in Normal Male Rat liver Specific Binding (% Pregnant

Binding Alhit

Preparation

Preparation)

(M-1

Pregnantfemale

100

Normal El

)

(fmole/mg

1.54

protein)

182.5

9.0

2.50

10.0

treated

male

51.1

2.89

47.5

Es treated

male

47.0

2.94

40.0

Normal liver

male

Binding Site Capacity J

x lo-

male

rats

membranes.

were

treated

All binding

with

estradiol

characteristics

(E2)

(25

are expressed

pg/day)

for

as described

10

days

for Table

prior

to

preparation

of

1.

In an attempt to investigate the cause of the progressive increase in binding site capacity from male to female to pregnant female rat membrane preparations, both normal male rats and Hx male rats were treated with estradiol (25 rg/day) for 5 or 10 days. In the experiment shown in Table 3, estradiol (E,) treatment of normal male rats resulted in a fourfold increase in binding capacity compared to the control male membrane preparation, with little change in the binding affinity (which in this case was much higher than observed previously for the standard pregnant preparation). Estradiol treatment of Hx male rats (Table 4) had no effect on specific binding, binding affinity or the binding capacity. In an attempt to increase the specific binding of hGH in Hx male rats, these rats were treated with high doses of bovine GH (bGH) (100 or 500 pg/day) with or without Ez (25 pg/day) for 5 or 10 days. Table 4 shows that bGH treatment, with or without Ez, failed to increase specific binding. Scatchard analysis showed that treatment with bGH alone or bGH with Ez also resulted in no significant changes in binding affinity or binding capacity. In another experiment (not shown) the concurrent administration of thyroid hormone (T4) with Table 4. The Effectof Estmdiol, Growth Hormone, and Prolactin on hGH Binding Sites in Hypophysectomixed Mole Rat liver Specific Binding Preparation Standard

(n)

pregnant

Hx Male

(3)t

Hx + E2

(2)

female

(% Pmgno”~

Binding Affini

Preparation)

(K

Binding Site Capacity r

x lo-

)

(fmole/mg

protein)

100%

(18)

18.2

f

1.9

1.21

f

0.30

52.5

f

0.80

11.2

20

72.5

Hx + bGH

(4)t

11.6

&

1.9

0.07

f

0.14

83.1

f

24.9

Hx + bGH

+ E2 (3)t

10.9

f

3.6

0.69

f

0.02

90.0

f

30.0

Hx + oPrl

(1)

l

0.63

112.5

Hx + oPrl

+ E2 (1)

l

0.49

107.5

Hx male

rats

were

treated

for 5 or 10 days,

or ovine

liver

All values

membranes.

expressed

as

cient data

described

points,

revealed

with

Prl

(oPrl)

ore

for

bovine (500

GH

expressed

Table

1.

(LGH)

pg/day) OS the

Stotisticol

no significant

f

(100 El

mean

500

Pg/day)

pg/day)

f

analysis

differences

or

(25

SEM

for

(n

>

(Student’s

between

these

f

5

estradiol

days,

2).

All

t test) groups

(E2)

prior

to

binding

of

those

and

the

(25

pg/day)

preparation

of

characteristics groups control

ore

having Hx

suffi-

male

rat

preporations. t.Since Hx mole treated ‘In of Hx

specific rot

binding

preparations,

was four

so low, of

meaningful

six

bGH

treated

analysis Hx

of the data rat

could

preparations,

be made and

three

in only of

four

three

of five

bGH

+

E2

Hx rot preparations. these +

oPrl

experiments was

2.2%

comparison

to

of added

label

pregnant ond

for

rat liver

membranes

Hx + oPrl

+

ore

not

E2 this was only

performed. 1.4%

Specific

of added

binding

lobel.

PITUITARY

349

REGULATION

bGH and Ez also failed to elicit a response in hGH binding characteristics to the crude 100,000 g liver membrane preparations or to purified liver plasma membranes prepared according to the method described by Neville.16 Although the administered bGH failed to induce a rise in the binding site capacity, it did produce a significant rise in body weight gain and also a marked stimulation of the in vitro incorporation of 35S-sulfate by costal cartilage taken from the treated rats (saline injected controls 17.5 + 0.6 pg SO.,/100 mg dry weight; bGH treated 46.2 + 4.4 rg S04/100 mg dry weight). The administration of E2 alone did not affect the incorporation of sulfate by cartilage. The incorporation of sulfate by Hx rat cartilage is used as a sensitive measure of the growth effects of GH (possibly mediated by somatomedin” and indicate in these experiments that the test rats were responsive to the administered bGH. Table 4 also shows that treatment of a male Hx rat with ovine Prl (500 pg/ day) and another male rat with ovine prolactin and 25 pg/day of Ez appeared to increase the binding capacity over Hx controls. DISCUSSION It is now widely accepted that the initial step in peptide hormone action is an interaction with specific binding sites (receptors) on the surface membranes of target tissue cells. The observations of high affinity low capacity GH binding by liver membranes prepared from a number of species’-6 led us to quantitate the GH binding capacity of liver membranes prepared from rats in a variety of physiological states. Initial characterization of hGH binding was carried out using liver membranes prepared from pregnant rats. There was a rapid and very significant specific binding of hGH (8%-l 8% of ‘25I-hGH initially added), which was competitively displaced to varying degrees by bovine and rat GH and by ovine, human, rat, and porcine Prl preparations. Scatchard analysis of the data for hGH binding to pregnant rat liver revealed a single order of binding site with an affinity constant of 1.37 f 0.3 x lo9 M-‘. The binding site capacity was 327 =t 53 fmole/mg protein. The present analysis does not permit us to calculate the degree of occupancy of these sites by endogenous growth hormone. From studies of several different rat liver membrane preparations, hGH binding appeared to be greatly influenced by the endocrine status of the animal from which the membranes were prepared. For almost all preparations showing measurable specific binding, the affinity constant was quite comparable to that for the pregnant rat preparations. The binding site capacity, however, differed widely. Male or female rats bearing the transplantable GH/Prl secreting pituitary tumors MStT/W15, MStT/F45, SMtT/WS, MtT/F4 showed approximately double the capacity of nonpregnant females. The binding capacity or normal male rats or of Hx male or female rats was very low (normal male l/12 that of pregnant female) and in some cases was unmeasurable. Estradiol treatment of normal male rats, but not of Hx male rats, increased binding capacity with little effect on binding affinity. However, treatment of Hx male rats with high doses of bovine GH or ovine GH or ovine Prl, in both cases with and

without I&, may have induced a modest the observations are too few for statistical

increase in binding confirmation.

site capacity,

but

350

HERINGTON,

PHILLIPS,

AND DAUGHADAY

Since this work was completed, Posner, Kelly, and Friesen” have reported that hypophysectomy of female rats lowered lz51-hGH binding by liver membranes but renal implantation of pituitaries into female hypophysectomized rats partially restored binding activity. Similarly, hypophysectomy blocked the ability of estrogens to increase ‘*‘I-hGH binding by liver membranes but renal implantation of pituitaries completely restored estrogen responsiveness. The authors conclude that the implanted pituitaries exerted their influence by secretion of prolactin which in synergism with estrogens induced the hGH binding site which they interpret as mainly a prolactin binding site. Because binding of ‘*‘I-hGH was greatest in those animals whose concentration of endogenous rGH or rPr1 was highest, our results could not be attributed to occupation of available binding sites by these hormones. In fact, the total number of binding sites may be higher than observed. The results obtained in the present study with the pregnant, normal female and normal male rat membrane preparations are in agreement with those previously reported by Kelly and co-workers. 4 They found a similar sex difference in the binding capacity for hGH and also for oPr1. Since both hGH and oPr1 had very similar binding characteristics for any one type of rat liver membrane (the same affinity and number of sites), they suggested that they were dealing with a common “lactogenic” receptor site. This hypothesis was strengthened by the observations that in rat liver only lactogenic hormones competed with either hGH or oPr1 for binding and that nonlactogenic hormones (bovine GH) exhibited little specific binding of their own and did not effectively compete with hGH or oPr1 for binding. The specificity experiments carried out in the present study suggest that the situation is more complex. Bovine GH and rat GH (both nonlactogenic hormones) did show competitive inhibition of hGH binding although with approximately 10% of the potency of hGH itself. Human Prl was an effective competitor of hGH binding but ovine, rat, and procine Prl exhibited much lower competition. Some of the inhibition of hGH binding exhibited by these prolactins may have been due to small amounts of contaminating GH. It should be noted that the maximum displacement occurring with oPr1 and bGH was not as great as that with hGH. Although the present studies are inconclusive they suggest that hGH binding by rat liver membranes occurs on separate binding sites with prolactin and growth hormone specificities. A similar proposal has been made by Ranke et a1.19It is paradoxical that these membranes exhibit marked sensitivity to both the human Prl and GH molecules, but have a much lower apparent sensitivity to the rat hormones. In an attempt to explain the differences observed between hGH binding by male, female, and pregnant female rat liver membranes, the role of estrogens in the induction of these binding sites was studied. Estrogen levels are higher in female than male rats and show a further increase during pregnancy.M The observed increase in binding capacity with little change in binding affinity, following E2 treatment of normal male rats, indicates a possible role for estrogens in the regulation of hGH receptors. Our results confirm the earlier report of Posner et al.*’ who also demonstrated increased binding by liver membranes prepared from male rats following Ez treatment. In addition, they reported that treatment of normal females with E2 raised the capacity of binding sites

PITUITARY

REGULATION

351

to that seen during pregnancy, but administration of progesterone or human chorionic somatomammotropin was without effect. Although it seems likely that estrogens do play a role in the development of hGH binding sites, it is clear that the pituitary also plays a definite role. In addition to a possible primary effect on liver, E2 may have a secondary or indirect role resulting from the known stimulus by E2 of the secretion of both GH and Pr1.22,23Removal of the pituitary gland (Hx) resulted in a marked reduction in the capacity of the hGH binding sites which could not be increased by treatment with E2. One probable role of the pituitary in regulating hGH binding sites is the maintenance of circulating Prl and/or GH levels. The observed differences in binding sites of pregnant, female, and male rat liver membrane preparations parallels the circulating levels of GH and Prl which tend to be higher in pregnant females than in nonpregnant females, which in turn often have higher levels than normal males. In the case of the pregnant rats, induction of hGH binding sites might be influenced by a placental GH-like peptide. In addition, the results obtained in this study with the rats bearing the transplantable GH/Prl secreting pituitary tumors also suggest that circulating GH/ Prl levels may have an important influence on the induction of binding sites. Rats bearing these tumors have extremely high circulating GH and Prl” (Table 2) and, as shown in Table 1, have a significantly greater binding capacity than either control female or male rats. The circulating serum estrogen levels in these animals were within normal limits, ruling out the possibility of a major direct estrogen effect. The demonstration that high circulating levels of GH/Prl are associated with an increase in the capacity of hGH binding sites and that hypophysectomy is associated with a decrease in the capacity of the binding sites, is in marked contrast to recent observations by Lesniak et a1.24of the binding of hGH to cultured lymphocytes. Cultured human leukemic lymphocytes have been shown to possess binding sites with high specificity for hGH,25*26and an affinity constant of 1.3 x lo9 M-‘,26 very similar to that obtained in this study for the liver membrane receptors. Addition of either low or moderately elevated levels of hGH to the culture medium for 24 hr prior to the binding studies resulted in a loss of up to 85% of hGH receptors. Subsequent removal of hGH from the medium was followed by a restoration of receptors which could be inhibited by cycloheximide. 24Similar results were obtained in studies of the effect of insulin on lymphocyte insulin receptors. 27The reason for the contrasting effect of GH concentration on GH binding sites in the two systems is unexplained. Our results do not permit us to state that the biologic effect of GH is proportional to the amount bound by liver membranes. Although endogenous Prl and/or GH appear to play a role in the regulation of hGH binding, treatment of Hx rats with high doses of either bovine GH or ovine Prl produce at the most, only a modest change in the binding characteristics of liver membrane preparations. In the case of GH, this occurred despite a marked stimulation of body weight and the incorporation of sulfate by costal cartilage removed from the treated rats. The concurrent administration of E2 with the GH or Prl also had little effect on hGH binding. The failure of these attempts may have been the result of inadequate duration or dose of bGH or

352

HERINGTON,

PHILLIPS,

AND

DAUGHADAY

oPr1. It is also possible that other pituitary factors could be required. It seems unlikely that hepatic membrane receptors would respond only to endogenous rGH or rPrl and not to the heterologous hormones administered which are fully potent in the rat. ACKNOWLEDGMENT The authors wish to acknowledge the assistance of Drs. L. S. Jacobs and L. Jarett in the preparation of this paper. We are indebted to Mrs. Ida Mariz for assistance in the studies of tumor bearing rats. REFERENCES 1. Rubin MS, Swislocki NI, Sonenberg M: Modification by bovine growth hormone of liver plasma membrane enzymes, phospholipids and circular dichroism. Arch Biochem Biophys 157:243-251, 1973 2. Postel-Vinay MC, Sonenberg M, Swislocki NI: Effect of bovine growth hormone on rat liver plasma membranes as studied by circular dichroism and fluorescence using the extrinsic probe 7,12-dimethylbenzanthracene. Biochim Biophys Acta 332:1X-165, 1974 3. Tsushima T, Friesen HG: Radioreceptor assay for growth hormone. J Clin Endocrinol Metab 37:334-337, 1973 4. Kelly PA, Posner BI, Tsushima T, Friesen HG: Studies of insulin, growth hormone and prolactin binding: Ontogenesis, effects of sex and pregnancy. Endocrinology 95532-539, 1974 5. Tsushima T, Shiu RPC, Kelly PA, Friesen HG: Radioreceptor assay for human growth hormone and lactogens: Structure-function studies and clinical applications, in Raiti S (ed): Advances in Human Growth Hormone Research. DHEW Publication No. 74-612, 1974, p 372 6. Herington AC, Jacobs LS, Daughaday WH: Radioreceptor and radioimmunoassay quantitation of human growth hormone in acromegalic serum: Overestimation by immunoassay and systematic differences between antisera. J Clin Endocrinol Metab 39:257-262, 1974 7. Thorell JI, Johansson BA: Enzymatic iodination of polypeptides with “‘1 to high specific activity. Biocbim Biophys Acta 251: 363-369, 1971 8. Lowry OD, Rosebrough NJ, Farr AL, Randall RJ: Protein measurement with the Folin phenol reagent, J Biol Chem 193:265-277, 1951 9. Scatchard G: The attractions of proteins for small molecules and ions. Ann NY Acad Sci 51:660-672, 1949

10. Posner BI, Kelly PA, Shiu RPC, Friesen HG: Studies of insulin, growth hormone and prolactin binding: Tissue distribution, species variation and characterization. Endocrinology 95521-531, 1974 11. Birge CA, Peake GT, Mariz IK, Daughaday WH: Effects of cortisol and diethylstilbestrol on growth hormone release by rat pituitary in vitro. Proc Sot Exp Biol Med 126:342-345, 1968 12. Niswender CD, Chen CL, Midgley AR, Meites J, Ellis S: Radioimmunoassay for rat prolactin. Proc Sot Exp Biol Med 130~793-797, 1969 13. Mikhail G, Wu CH, Ferin M, Vande Wiele RL: Radioimmunoassay of plasma estrone and estradiol. Steroids 15:333-352, 1970 14. Daughaday WH: Somatomedin (sulfation factor), in Berson SA, Yalow RS (eds): Peptide Hormones. Amsterdam, North-Holland Publishing Co., 1973, p 292 15. Peake GT, Mariz IK, Daughaday WH: Radioimmunoassay of growth hormone in rats bearing somatotropin producing tumors. Endocrinology 83:714-720, 1968 16. Neville DM, Jr: Isolation of an organ specific protein antigen from cell surface membrane of rat liver. Biochim Biophys Acta 154: 540-552, 1968 17. Phillips LS, Herington AC, Daughaday WH: Hormone effects on somatomedin action and somatomedin generation, in Raiti S (ed): Advances in Human Growth Hormone Research. DHEW Publication No. 74-612, 1974, P 50 18. Posner BI, Kelly PA, Friesen HG: Prolactin. Science 188:57-59, 1975 19. Ranke MB, Stanley CA, Parks JS, Baker L, Greene C, Bongiovanni AM: Sex differences in human growth hormone binding to rat hepatocytes. Clin Res 23:328A, 1975 20. Skinner JD: Puberty in the male rabbit. J Reprod Fertil 14151-154, 1967 21. Posner BI, Kelly PA, Friesen HG: In-

PITUITARY

REGULATION

duction of a polypeptide hormone receptor in rat liver. Proc Nat1 Acad Sci USA, 21:1029, 1973 22. Dickerman E, Dickerman S, Meites J: Influence of age, sex and estrous cycle on pituitary and plasma GH levels in rats, in Pecile A, Muller EE (eds): Growth and Growth Hormone. Amsterdam, Excerpta Medica, 1972, p 252 23. Meites J, Lu KH, Wuttke W, Welsch CW, Nagasawa H, Quadri SK: Recent studies on functions and control of prolactin secretion in rats. Ret Prog Hormone Res 28:471-526, 1972 24. Lesniak MA, Bianco AR, Roth J, Gavin JR HI: Human growth hormone radioreceptor

353

assay using cultured human lymphocytes. Res 22:343A, 1974

Clin

25. Lesniak MA, Roth J, Gorden P, Gavin JR III: Human growth hormone, radioceptor assay using cultured human lymphocytes. Nature, New Biol241:20-22, 1973 26. Lesniak MA, Gorden P, Roth J, Gavin JR III: Binding of I25I-human growth hormone to specific receptors in human cultured lymphocytes. J Biol Chem 249:1661-1667, 1974 27. Gavin JR III, Roth J, Neville DM, DeMeyts P, Buell DN: Insulin dependent regulation of insulin receptor concentrations: A direct demonstration in cell culture. Proc Nat1 Acad Sci (USA) 71:84-88, 1974