Molecular and Cellular Endocrinology, 30 (1983) 63-7 1 Elsevier Scientific Publishers Ireland, Ltd.

63

MODULATION OF HEPATIC INSULIN RECEPTORS BY A HUMAN GROWTH HORMONE FRAGMENT (hGH 6-13)

Frank NG, Thrift HENDERSON, Janet BROMLEY and Joseph BORNSTEIN

Department of Biochemistty, Monash Universit_v, Clayton. Victoria 3168 (Australia)

Received 10 June 1982; revision received 13 December 1982: accepted 13 December 1982

A synthetic amino-terminal fragment of human growth hormone (hGH) containing the sequence H,N-Leu-Ser-Arg-Leu-Phe-Asp-Asn-Ala-COOH (hGH 6-13) was shown

to increase [ ‘251]iodoinsulin binding to rat hepatic receptors in vivo. Analysis of the binding

data indicated an increase in the capacity of the insulin receptors or the number of available

receptors. In vitro experiments with isolated hepatocytes and hepatic plasma membranes

revealed no direct interaction between the hGH 6-13 and the hepatic insulin receptors.

When the isolated hepatocytes were preincubated with the synthetic hGH fragment (10

pg/ml) at 37T for 30 min prior to tests for insulin binding, the binding of [‘251]iodoinsu-

lin to the hepatic receptors was significantly enhanced. The variance between the in viva

and in vitro findings was considered in terms of the mechanism of hormonal actions

mediated through a secondary cellular mediator.

Kqwords: hCH fragment; hepatocytes; plasma membranes; insulin binding.

The mechanism by which growth hormone produces its diverse effects on cells is complex and poorly understood. Accumulated evidence sug- gests that growth hormone as isolated from pituitaries may, in fact, be a

prohormone (Lewis et al., 1977; Hales, 1978; Wade et al., 1982), and various forms or regions of the molecule could be responsible for specific

metabolic activities (Kostyo and Wilhelmi, 1976; Paladini et al., 1979). Hypoglycaemic activity of growth hormone has long been documented (Milman and Russell, 1950), and was shown to be due to the insulin

synergetic effect of the molecule (Mahler and Szabo, 1969). The direct effects of growth hormone on insulin action on rat adipocytes (Maloff et al., 1980; Schoenle et al., 1981) and hepatocytes (Fix and Moore, 1981) have been reported. Our previous studies suggest that the entire molecule of growth hormone is not required to elicit the biological activity, and the amino-terminal region of the hormone constitutes the active centre of the molecule for the insulin-potentiating action (Bornstein et al., 1973; Ng et

0303-7207/83/0000-0000/$03.00 0 1983 El sevier Scientific Publishers Ireland. Ltd.

al., 1974; Ng and Larner, 1976; Newman et al., 1978). Synthetic frag- ments of human growth hormone (hGH) containing the active sequence

H,N-Arg-Leu-Phe-Asp-Asn-Ala-COOH (hGH 8-13) have recently been found to induce hypoglycaemia in normal and streptozotocin-di-

abetic rats as the result of increased insulin receptor function and glycogen synthase activity in vivo (Ng and Bornstein, 1979; Ng et al.,

1980; Pullin et al., 1981). This study was undertaken to investigate the

mechanism of action of an active hGH fragment, H,N-Leu-Ser-

Arg-Leu-Phe-Asp-Asn-Ala-COOH (hGH 613), on rat hepatic in-

sulin receptors.

MATERIALS AND METHODS

Animuls Male albino Wistar rats of 100 + 20 g, housed under conditions of

constant temperature (21°C) with a normal 12 h light cycle and fed ad libitum, were used for the present work. Prior to experimentation, all animals were fasted for 16 h and then anaesthetized with an i.p. injection

of pentobarbital (50 mg/kg body weight). For the studies of the in viva

effect of hGH 6-13, the anaesthetized animals were given i.v. a prede- termined dose of the hGH fragment or saline (as control) 30-45 min

before the commencement of hepatocyte or hepatic plasma membrane preparations.

Chemicals

Synthetic fragment of human growth hormone, hGH 6-13, was pre- pared by solid phase synthesis and purified by filtration-chromatography as previously described (Pullin et al., 1981). Purified crystalline beef insulin was supplied by Commonwealth Serum Laboratories (Melbourne, Australia). Sodium [ ‘251jiodide was obtained from the Radiochemical Centre (Amersham, Great Britain), bovine serum albumin (Fraction V) from Sigma Chemical Co. (St. Louis, MO, U.S.A.), collagenase (type II) from Worthington Chemicals (Freehold, NJ, U.S.A.). All other reagents used were of analytical grade.

Preparations of isolated hepatocytes and hepatic plasma membranes The liver of the anaesthetized rats was exposed; the portal vein and

right atrium were cannulated. The liver was then perfused for 25-30 min with calcium-free Krebs-Ringer bicarbonate buffer, pH 7.5, containing glucose (1 mg/ml), bovine serum albumin (10 mg/ml) and collagenase (0.5 mg/ml) at a flow rate of 10 ml/min. Hepatocytes were isolated by

Modulutron of insulin receptors by hGH 65

the method of Berry and Friend (1969) as modified by Olefsky et al.

(1975) and resuspended in the same buffer without collagenase at a concentration of lo6 cells/ml for binding studies. Purified hepatic plasma membranes were prepared according to the method of Ray (1970). Membrane protein determinations were performed according to the method of Lowry et al. (1951) as modified by Peterson (1977).

Binding of “‘I-labelled insulin

Insulin was iodinated with chloroamine-T according to the method of

Cuatrecasas (1971). When prepared in this fashion, the [‘251]iodoinsulin

was at least 95% precipitable by trichloroacetic acid. Studies of

[ ‘251]iodoinsulin binding to hepatic plasma membranes in the presence or

absence of synthetic hGH 6- 13 were conducted at 30°C in 50 mM

Tris-HCl buffer, pH 7.5, containing 0.5% bovine serum albumin as

previously described (Ng et al., 1980). Binding studies with hepatocytes were performed similarly, except that the isolated cells were also prein-

cubated for the period of time as indicated and with or without the synthetic hGH 6-13 (10 pg/ml) in Krebs-Ringer phosphate buffer, pH

7.6, containing 0.5% bovine serum albumin at 15°C or 37°C in a shaking bath. The treated cells were washed, resuspended, and assayed for insulin binding using the same procedure as for hepatic plasma membranes.

In these studies, non-specific binding is defined as the residual [ ‘251]iodoinsulin in the hepatic plasma membrane, or hepatocyte pellet, in the presence of a large excess (20 mM) of unlabelled insulin. For all

experiments total and non-specific binding was determined, and non- specific binding was subtracted from total binding to yield specific binding.

RESULTS

Both hepatocytes and hepatic plasma membranes from rats treated with hGH 6-13 bound significantly more [ ‘251]iodoinsulin than those from untreated control animals in 6 separate paired experiments (Fig. 1). The significant difference was observed at insulin concentrations ranging from 0.15 to 50 nM. All data were corrected for non-specific binding. The non-specific binding with hepatocytes from control and treated rats was 2.81 f 0.43% and 2.75 + 0.52% of the [ ‘251]iodoinsulin, respectively, whereas the non-specific binding with hepatic plasma membranes from control and treated rats was 1.63 + 0.45% and 1.77 + 0.38%, respectively. The data from the same series of experiments were also analysed by the Scatchard method as described by Kahn et al. (1974). If one assumes two

66 F. Ng et al.

0 1 10 102 lOSO I 10 102 103

TOTAL INSULIN (nM)

Fig. I. [‘251]10doinsulin binding to isolated hepatocytes and hepatic plasma membranes

from control (0) and hGH 6-13 treated (e) rats. Hepatocytes (control,~O.52~0.10~ 106;

treated, 0.58 + 0.13 X 10’ cells/ml) and hepatic plasma membranes (control, 205 i 27;

treated, 2OOi_ 32 pg/ml) were incubated with 0.15 pmoles of [ i2sI]iodoinsulin in the

absence or presence of unlabelled insulin over a range of insulin concentrations from 0.15

to 1000 nM. Data are corrected for non-specific binding and expressed as the percentage of

insulin bound per IO6 cells or per mg membrane protein. Each point represents the

mean i SD of 6 separate paired experiments. Significant differences (P < 0.05) between the

test and the corresponding control values of hepatocytes and ptasma membranes were

observed at insulin concentrations from 0.15 to 50 nM.

classes of insulin receptors, the capacity or the number of the high-affin- ity as well as the low-affinity receptors was increased (Fig. 2).

In contrast, the hGH 6- 13 from 0.2 up to 40 pg/mI exerted no direct effect on insulin binding of either hepatocytes or hepatic plasma mem-

branes (data not shown). Fig. 3 unequivocally shows no significant

difference in the percentages of [‘251]iodoinsulin bound in the absence and the presence of the hGH fragment at a concentration of 20 pg/mI. Furthermore, the mean values of the control experiments revealed in both Fig. 1 and 2 are similar and confirm that the in vivo and in vitro studies are comparable.

Hepatocytes of untreated rats preincubated with 10 pg/ml of hGH 6-13 at 37°C for 30 min prior to receptor assays significantly increased insulin binding capacity (Fig. 4), similar to the hepatocytes of hGH 6- 13 treated animals. The in vitro effect of hGH 6- 13 on the isolated cells was observed at as low as 2 pg/ml and the maximal effect was achieved at 10 pg/ml as demonstrated in Fig. 4. This effect of the synthetic hGH fragment on hepatocytes depends also on the length of the preincubation period, as illustrated in Table 1. At the optimal concentration (10 pg/ml), hGH 6- 13 enhanced insulin binding of the isolated cells after a 15 min

Modulation of insulin receptors by hGH 67

HEPATOCYTES 0.3

0.2

w w cc 0.1 tI_

0 8 4 I I I 1 ,

MEMBRANES

01234567

BOUND (pMOLE)

Fig. 2. Scatchard analysis of data in Fig. 1, plotting bound/free insulin vs. insulin bound. Each curve represents the mean of 6 separated paired experiments with control (0) and

hGH 6- 13 treated (0) rats.

Q z 3

0 25 m

2 20

? 3 15 w

_z 10

Lo 5 ”

g_ 0

HEPATOCYTES MEMBRANES

TOTAL INSULIN (nM)

Fig. 3. [ ‘251]IodoinsuIin binding to isolated hepatocytes and hepatic plasma membranes in

the absence (0) and the presence (0) of 20 pg/ml of hGH 6- 13. The data are corrected for

non-specific binding and normalized to lo6 cells and 1 mg protein for hepatocytes and

plasma membranes, respectively. Each point is the mean rt SD of 3 separate experiments,

each carried out with two different hepatocytes or plasma membrane preparations.

0123456

8 (p MOLE i 10’ CELLS)

F. Ng et al.

Ot, 0 1 10 IO2 ff?

TOTAL INSULIN (nM)

Fig. 4. [ ‘2SI]Iodoinsulin binding to rat hepatocytes. Prior to the binding assays isolated

hepatocytes were preincubated in the absence (0) and the presence (0) of hGH 6-13 as

described in Materials and Methods. The data represent the meankSD of triplicate

determinations of 4 separate experiments. The curves are significantly different (P < 0.05)

at insulin concentrations from 0.15 to 50 nM. Inset: Scatchard analysis of the same binding

data.

Table 1

Effect of time and temperature of preincubation with hGH 6-13 on insulin binding of

isolated hepatocytes

Temperature

(“C)

Time

(min)

% [‘251]iodoinsulin specificahy bound

Buffer control hGH 6- 13 added

37 0 22.1*:1.2

15 21.0*0.9

30 20.8+ 1.2

60 19.7* 1.0

15 0 20.5 & 0.8

15 21.6i 1.2

30 19.8+ 1.1

60 20.0 + 0.9

21611.0

23.2 f 1.4

27.0* 1.3 *

26.4+ 1.3 *

20.25 I.1

19.s* 1.3 22.6 k 1.0 *

23.5 f I.4 *

Isolated hepatoeytes were initialty preincubated with hGH 6- 13 (10 tg/mi) for the length

of time and at the temperature indicated, then washed with and resuspended in assay buffer

at 30°C. Binding assays were performed in the presence of 0.15 nM [“‘I]iodoinsulin.

Non-specific binding was determined in the same manner except that 20 PM unlabelled

insulin was also present. Data are corrected for non-specific binding, expressed as %

[‘~‘I]i~oinsulin bound per lo6 cells, and represent the mean i SD of triplicate determina- tions of 3 separate experiments. * Values significantly different from the corresponding control (P < 0.05).

preincubation and had no further increase after 30 min. The results in

Table 1 indicate also that the effect of the hGH fragment appeared to be temperature-dependent.

DISCUSSION

Early in vivo studies have shown that the primary effect of the amino-terminal region of human growth hormone (hGH) on carbohydrate

metabolism is to alter insulin action and glucose utilization at the peripheral levels (Bornstein et al., 1973; Ng et al., 1974, 1980). The present in vivo experiments have revealed that the short-term administra- tion of synthetic hGH 6-13 to normal rats increased insulin binding to hepatic receptors. The Scatchard analysis of the insulin binding data indicates that the hGH 6-13 induced in vivo an increase in the capacity

of the insulin receptor or the number of available receptors on hepato-

cytes similar to its effects on adipocytes as previously observed (Sims et

al., 1980). In contrast to the in vivo findings, the results of the in vitro studies

have shown the absence of the direct biological effect of the synthetic hGH 6 13 on insulin binding to both hepatic plasma membranes and isolated hepatocytes under our binding assay conditions. These in vitro

findings, which have been confirmed by the lack of observable effect of

an active analogous fragment (hGH 4-15) on Triton X-100 solubilized insulin receptors (Asznowicz and Ng, unpublished observation), indicate

that the amino-terminal fragment of hGH may not directly modulate the insulin receptors. There are several possibilities for the variance between the in vivo and the in vitro findings. The most obvious one is that the hGH fragment may have to be activated in vivo in some way before exerting its biological actions. Considering the possible conformational changes or prosthetic group modifications of the peptide molecule in response to varying physical and chemical conditions, it is difficult to rule out this occurrence in vivo, although technically there is no simple way to prove it at present. The present study shows that the effect of the hGH fragment on insulin receptor of hepatocytes during preincubation

was time- and temperature-dependent, indicating an active process. Therefore, the other possibility is that there may be a chemical mediator generated from the intact cells by the hGH fragment, in which case the mediator but not the hGH fragment acts on the insulin receptors. The concept of hormone second messenger has long been established. Cur- rently, evidence has demonstrated the involvement of a chemical media- tor in insulin action and it appears to be a soluble peptide-like compo-

70 F. Ng et 01.

nent generated from plasma membrane by proteolytic reaction (Larner et al., 1979, 198 1; Seals and Jarett, 1980; Seals and Czech, 1980). The investigation of the possible existence of a cellular mediator for the

action of the hGH fragment is in progress in our laboratory. It is evident from the in vitro studies with the isolated hepatocytes

after preincubation with the synthetic hGH 6- 13 that the synthetic hGH fragment enhanced insulin binding to the treated cells. The data are

compatible with either one or both of the above-mentioned possibilities. Furthermore, these two possibilities are not necessarily exclusive of each

other. If the latter is finally proven true, this increase of insulin binding

to hepatocytes after in vitro preincubation with the synthetic hGH 6- 13

could mean that the freshly isolated intact cells still retained the molecu-

lar mechanism to generate the cellular mediator in response to the hGH fragment and subsequently increased their insulin receptor function. Although in vitro data should always be interpreted in physiological contexts with caution, the isolated cell system may provide a useful means of clarifying the molecular action of the growth hormone fragment and of identifying the cellular mediator responsible for its action on insulin receptor.

In conclusion, the present findings indicate that the synthetic hGH 6- 13 induced hypoglycaemia (Ng et al., 1974) by, at least in part,

modulating the plasma insulin receptors with an indirect but so far unknown mechanism. As hGH 6- 13 is a part-sequence of the hGH

molecule, the question that still remains to be resolved is whether the

fragment is the active centre for the hypoglycaemic action of the entire hGH molecule or whether it is in a separate hormone entity from

post-pituitary modification of growth hormone which, as generally de- fined, is a prohormone as previously suggested (Lewis et al., 1977; Sims

et al., 1980; Wade et al., 1982). However, our recent studies reveal that the intact molecule of hGH has no effect on the binding of insulin to hepatic receptor, and suggest that post-pituitary molecular modification of the pituitary hGH for biological actions may occur in vivo (Ng and Bornstein, 1982).

ACKNOWLEDGEMENTS

This work was supported by a Research Project Grant of the National Health and Medical Research Council of Australia, and a Special Re- search Fund Grant of Monash University, Victoria, Australia.

Modulatron of insulin receptors by hCH

REFERENCES

Adamson, U. and Efendic, S. (1979) J. Clin. Endocrinol. Metab. 49, 459-464.

Berry, M.N. and Friend, D.S. (1969) J. Cell Biol. 43, 506-520.

Bornstein, J., Armstrong, J.McD., Taft, H.P., Ng, F.M. and Gould, M.K. (1973) Postgrad.

Med. J. 49, 219-241.

Cuatrecasas. P. (1971) Proc. Natl. Acad. Sci. (U.S.A.) 68, 1264-1268.

Fix, J.A. and Moore, W.V. (1981) Endocrinology 108, 239-246.

Hales, C.N. (1978) FEBS Lett. 94, lo- 16.

Kahn, C.R., Freychet, P. and Roth, J. (1974) J. Biol. Chem. 249, 2249-2257.

Kostyo, J.L. and Wilhelmi, A.E. (1976) Metabolism 25, 105-124.

Larner, J., Galasko, G., Cheng, K., De Paoli-Roach, A.A., Huang, L., Daggy, P. and

Kellogg, J. (1979) Science 206, 1408- 1410.

Larner, J., Cheng, K., Schwartz, C., Dubler, R., Creacy, S., Kikuchi, K., Tamura, S.,

Galasko, G., Pullin, C. and Katz, M. (1981) Mol. Cell. Biochem. 40, 155- 161.

Lewis, U.J., Singh, R.N.P., Vanderlaan, W.P. and Tutwiler, G.F. (1977) Endocrinology 101,

1587-1603.

Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 193,

265-275.

Mahler, R.J. and Szabo, 0. (1969) Diabetes 18, 550-555.

Maloff, B.L., Levine, J.H. and Lockwood, D.H. (1980) Endocrinology 107, 538-544.

Milman, A.E. and Russell, J.A. (1950) Endocrinology 47, 114- 128.

Newman, J.D., Armstrong, J.McD. and Bornstein, J. (1978) Biochim. Biophys. Acta 544,

234-244.

Ng, F.M. and Bornstein, J. (1979) Diabetes 28, 1126-l 130.

Ng, F.M. and Bornstein, J. (1982) Diabetologia 23, in press.

Ng, F.M. and Lamer, J. (1976) Diabetes 25, 4133419.

Ng. F.M., Bornstein, J., Welker, C., Zimmett, P.Z. and Taft, P. (1974) Diabetes 23.

943-949.

Ng, F.M., Bornstein, J.. Pullin, CO., Bromley. J.O. and Macaulay, S.L. (1980) Diabetes 29,

782-787.

Olefsky, J.M., Johnson, J., Liu, F., Edwards, P. and Baur, S. (1975) Diabetes 24, 801-810.

Paladini, A.C., Pena, C. and Retegui, L.A. (1979) Trends Biochem. Sci. 4. 256-260.

Peterson, J.L. (1977) Anal. B&hem. 85, 346-356.

Pullin, C.O., Harcourt, J.A., Ng, F.M. and Bornstein, J. (1981) Int. J. Peptide Protein Res.

18, 318-323.

Ray, T.K. (1970) Biochim. Biophys. Acta 196, 1-9.

Schoenle, E., Zapf, J. and Froesch, E.R. (1981) Endocrinology 109, 561-566.

Seals, J.R. and Czech, M.P. (1980) J. Biol. Chem. 255, 6529-6531.

Seals, J.R. and Jarett, L. (1980) Proc. Natl. Acad. Sci. (U.S.A.) 77, 77-81.

Sims, P., Ng, F.M. and Bornstein, J. (1980) Biochem. Int. 1, 120-125.

Wade, J.D., Ng, F.M., Bornstein, J., Pullin, CO. and Pearce, J.S. (1982) Acta Endocrinol.

(Kbh.) 101, 10-14.