A growth hormone (GH) analog can antagonize the ability of native GH to promote differentiation of 3T3-F442A preadipocytes and stimulate insulin-like and lipolytic activities in primary rat adipocytes

The effect of amino acid substitutions introduced to the third alpha-helix in bovine GH (bGH) was investigated. A GH analog (bGH-M8), in which three amino acids were substituted to form an idealized amphiphilic alpha-helix, possessed the same specific binding affinity as wild-type bGH to cell membranes prepared from 3T3-F442A cells or rat adipocytes. However, bGH-M8 failed to stimulate preadipocyte differentiation, as measured by the level of glycerol-3-phosphate dehydrogenase activity. An equimolar concentration of bGH-M8 was inhibitory for this adipogenic effect caused by bGH at a concentration of 30 pM. bGH-M8 also failed to induce an insulin-like response and reduced lipolytic potency in rat primary adipocytes. A 10-fold excess of bGH-M8 abolished the effect of wild-type bGH in the insulin-like and lipolytic assays. Thus, bGH-M8 inhibited these actions of wild-type bGH and, therefore, appears to be a competitive antagonist. These results suggest that a major biologically active domain resides in the third alpha-helix of bGH, which is independent of amino acids important in the initial interaction of GH with its receptor.

Up-regulation of vinculin expression in 3T3 preadipose cells by growth hormone

In an effort to define biochemical events relevant to the adipogenic action of GH, the effect of GH on expression of the cytoskeletal element vinculin was studied in 3T3-F442A preadipose cells. Results from Western blotting indicated that in serum-free medium 2 nM met-hGH induced an approximately 100% increase in vinculin expression relative to that in cells maintained in serum-free medium alone. GH-elicited alterations in vinculin expression were dose dependent. GH treatment elevated levels of tubulin to a lesser extent, whereas actin expression was unaffected by GH. Immunoprecipitation experiments revealed that GH treatment promoted a 200% increase in vinculin synthesis on day 4 relative to that in control cells. GH had no effect on phosphorylation of vinculin in 3T3-F442A cells. Based on Northern blotting, we noted that GH induced approximately a 200% increase in levels of vinculin mRNA on day 4. GH responsiveness as well as levels of vinculin were similar in 3T3-GI-16 (a highly adipogenic subclone of the 3T3-F442A cell) and 3T3-F442A cells. The GH-dependent increase in vinculin protein expression was not observed in nonadipogenic 3T3-C2 cells, suggesting that this effect of GH was related to the program of differentiation. Interestingly, levels of vinculin in nontreated 3T3-C2 cells were approximately 10-fold lower than levels in 3T3-F442A cells. GH-mediated alterations in vinculin expression in 3T3-F442A cells were abolished by treatment with fetal bovine serum, a potent mitogen. Our data indicate that increased expression of vinculin is a component of the GH-induced portion of the adipose differentiation program.

Growth hormone and adipose differentiation: growth hormone-induced antimitogenic state in 3T3-F442A preadipose cells

An additional activity for pituitary growth hormone is described--i.e., the in vitro induction of an antimitogenic state in murine 3T3-F442A preadipocyte fibroblasts. We previously developed a serum-free, hormonally defined medium permissive for the adipose differentiation of 3T3-F442A cells. When 3T3-F442A fibroblasts were maintained in serum-free medium without insulin but with growth hormone (2 nM), typical adipose differentiation did not occur. However, we found that growth hormone induced a state of cellular refractoriness to the mitogenic stimulus of fetal bovine serum as assayed by de novo DNA synthesis. The mitogen refractory condition (i.e., the antimitogenic state) was time-dependent (half maximal at approximately 2.5 days) and growth hormone concentration-dependent (half maximal and maximal at approximately 0.05 and 2.0 nM, respectively). The antimitogenic state was specifically induced by growth hormone and was not mediated by insulin-like growth factor I or prolactin. The growth hormone-induced antimitogenic state was completely reversible. The antimitogenic state was not induced by growth hormone in 3T3-C2 cells, a sister clone of 3T3 cells that exhibits essentially no adipose conversion. The kinetics for growth hormone-dependent commitment to adipose differentiation and induction of the antimitogenic state were similar. We suggest a relationship of growth hormone-induced antimitogenic state and the growth hormone-induced adipose differentiation of 3T3-F442A cells.

Growth hormone-dependent events in the adipose differentiation of 3T3-F442A fibroblasts: modulation of macromolecular synthesis

GH is necessary but not sufficient to induce adipose differentiation of 3T3-F442A fibroblasts in serum-free medium. Human (h) GH (2 nM) treatment of 3T3-F442A cells in serum-free medium caused a time-dependent (maximal at 48-72 h) and dose-dependent (EC50, approximately 0.2 nM) decrease (40-60%) in de nova protein synthesis. Insulin-like growth factor-I (IGF-I; 17 nM), PRL (2 nM), and glucagon (20 nM) did not decrease de novo protein synthesis, whereas bovine GH was equipotent with hGH. The half-lives of 35S-labeled proteins of 3T3-F442A cells were 21 and 57 h for cells maintained in serum-free medium for 3 days without or with hGH (2 nM), respectively. The total protein content of cells maintained in hGH (2 nM) for 1-4 days was unaffected compared to cells in serum-free medium alone. IGF-I (17 nM) treatment of cells for 4 days doubled the protein content of cells compared to control values in serum-free medium. hGH (2 nM) pretreatment of cells for 1-4 days had no effect on total RNA synthesis. hGH (2 nM) but not IGF-I (17 nM) treatment (3 days) resulted in a 7-fold decrease in cytoplasmic 18S rRNA content (as measured by DNA-RNA hybridization) of cells compared to that of control cells maintained in serum-free medium. When 3T3-F442A cells were transferred to serum-free medium there was a progressive decrease in DNA synthesis. The presence of hGH enhanced the rate at which DNA synthesis decreased for 3T3-F442A cells. IGF-I (17 nM) increased DNA synthesis by 6- and 8-fold after 2 and 3 days of IGF-I exposure. 3T3-F442A cells maintained in serum-free medium for 3 days responded to the addition of platelet-derived growth factor (2 U/ml) and insulin (1.6 microM) with a 56-fold increase in DNA synthesis, assayed 24 h later. 3T3-F442A cells treated with hGH (2 nM) for 3 days before platelet-derived growth factor and insulin addition exhibited a diminished DNA synthetic response, demonstrating that GH-exposed cells were partially refractory to mitogenic stimulation. GH had no effect on any aspect of macromolecular synthesis in 3T3-C2 cells, which have a low frequency of adipogenesis. Based upon these results a cell cycle model for the role of GH in the adipose differentiation of 3T3-F442A cells was proposed.

Growth hormone-induced alteration of morphology and tubulin expression in 3T3 preadipose cells

Effects of growth hormone on morphology and cytoskeletal protein expression were examined in 3T3-F442A preadipocytes in serum-free medium. Between 2 and 5 days of culture 2 nM methionyl human growth hormone converted 3T3-F442A cells from a flat fibroblastic morphology to a rounded form with numerous membrane convolutions. Growth hormone treated cultures manifested a 30-40% reduction in cell volume. Growth hormone induced changes in morphology and volume preceded and were independent of lipogenesis. In cells treated with growth hormone, expression of alpha and beta-tubulin as determined by Western blotting was found to increase approximately 50% within 72 h as compared to untreated cells. After 7 days, tubulin levels in growth hormone treated cells were approximately 40% of control levels. This indicated that morphological changes and alteration of tubulin expression were signatures of growth hormone action on 3T3-F442A cells.

Expression of growth hormone-independent adipogenesis by a 3T3 cell variant

We have examined the regulation of adipogenesis of a 3T3-F442A cell variant. The variant, designated 3T3-GH-independent clone 16 (GI-16), was isolated after serum-induced adipogenic commitment. 3T3-GI-16 fibroblasts displayed a lower serum requirement for adipogenesis than the 3T3-F442A parent cell. Insulin-stimulated adipogenesis of 3T3-GI-16 cells in serum-free medium (SFM) was extensive in the absence of GH, as judged by oil red O staining or glycerol-3-phosphate-dehydrogenase activity, a property not associated with the 3T3-F442A cell. In SFM devoid of GH the concentration of insulin required to promote half-maximal adipogenesis of 3T3-GI-16 fibroblasts was 5 nM. The expression of GH-independent adipogenesis by 3T3-GI-16 cells was not due to exposure to adipogenic stimuli during routine passage, as insulin-stimulated differentiation was not a function of the inoculation density in nonadipogenic cat serum. We noted that nine proteins resolved by polyacrylamide gel electrophoresis behaved in a differentiation-dependent manner during adipogenesis of 3T3-GI-16 and 3T3-F442A fibroblasts in SFM. The concentrations of all nine proteins were regulated in a GH-independent manner during insulin-stimulated adipogenesis of 3T3-GI-16 fibroblasts. In contrast, the presence of insulin alone markedly altered the expression of only two of the proteins during differentiation of 3T3-F442A cells. The observed changes in the expression of five presently uncharacterized differentiation-dependent proteins were most likely due to employment of SFM. Our results suggest that expression of GH-independent insulin-induced adipogenesis of 3T3-GI-16 fibroblasts reflects a prior commitment by GH during our selection protocol. These results are discussed in the context of a model in which adipogenesis in vivo is postulated to proceed through the sequential action of GH and insulin on target cells.

Role of insulin in growth hormone-stimulated 3T3 cell adipogenesis

The role of insulin during GH-stimulated adipogenesis of 3T3-F442A fibroblasts was investigated. Adipogenesis in defined medium (DM), as quantified by the level of glycerol-3-phosphate dehydrogenase activity, revealed that there existed a strict requirement for both insulin and GH during adipogenesis. The concentration of insulin required to elicit half-maximal adipogenesis was approximately 20 nM. Insulin-like growth factor I was less effective than insulin in promoting adipogenesis, indicating that insulin action during differentiation was most likely mediated through the insulin receptor. Cellular viability was not compromised by the absence of insulin, as judged by colony-forming efficiency or trypan blue exclusion. Deletion of insulin from DM supplemented with 1 nM recombinant human GH reduced glycerol-3-phosphate dehydrogenase activity to uninduced levels. Removal of other individual DM constituents did not have this effect. The growth factors fibroblast growth factor, platelet-derived growth factor, and bombesin did not substitute for insulin during GH-stimulated adipogenesis. The characteristic increase in cell number observed during serum-based differentiation, reflecting clonal expansion of young adipocytes, did not occur in DM supplemented with insulin, and insulin-like growth factor I were necessary for this event. These results suggest that insulin functions in concert with GH as a coinducer of the differentiating signals.

Steroid metabolizing enzymes associated with the microvillar membrane of human placenta

17 beta-Hydroxysteroid oxidoreductase, as well as estrone sulfate and dehydroepiandrosterone sulfate sulfatases, were found in the plasma membrane of microvilli of the fetal syncytiotrophoblast. Because of their location, these enzymes may influence feto-maternal transfer of steroids circulating as sulfates, the utilization of sulfated estrogen precursors and the proportion of estrone and estradiol delivered towards fetal and maternal circulations. Microvillar vesicles isolated from human term placentas were disrupted in hypotonic medium to obtain a membrane preparation. A fraction of the estradiol 17 beta-oxidoreductase (E2DH) activity in the vesicle remained associated to the membrane after disruption and treatment with 2 M NaCl. The membrane-associated activity was resistant to inhibition with trypsin and did not react with a polyclonal antibody which neutralized cytosolic E2DH activity. The membrane-associated enzyme was solubilized with a cholate-glycerol buffer solution and purified on Sephadex G-100. The estimated molecular weight of the solubilized enzyme (137 kDa) appears to correspond to a tetramer since it was found to be about twice the size of the cytosolic enzyme. Both enzymes focused in polyacrylamide gels at pH 5.2. The Km relative to E2 of the membrane-associated E2DH (1.3 microM) differs from those of mitochondrial (0.43 microM), microsomal (0.69 microM) and cytosolic (11 microM) fractions. The cytosolic and the microvillar membrane associated 17 beta-hydroxysteroid oxidoreductases also differ in their specificity for C18 and C19 steroid substrates and in their pH dependence patterns. Sulfatases acting on estrone sulfate and dehydroepiandrosterone sulfate in microvillar membranes were insensitive to trypsin and as resistant to washes with 2 M NaCl as alkaline phosphatase. This data indicated that steroid sulfatases are also microvillar membrane associated enzymes of potential physiologic importance in the hydrolysis of estrogen precursors.

Relative distribution of estrone, estradiol and estriol between fetal and maternal perfusates during perfusions of human term placentas with labelled C19 precursors

Published results from in vivo experiments carried out in Rhesus monkeys and from in vitro perfusions of human term placentas have indicated that placental estradiol (E2) is preferentially released towards the mother whereas estrone (E1) is about evenly distributed between fetal and maternal circulation. In order to examine the distribution of estriol, relative to that of E1 and E2, we have now prepared [3H]16-hydroxyandrostenedione by incubation of [6,7-3H]androstenedione with Streptomyces roseochromogenus and perfused placental cotyledons with mixtures of these two labeled precursors. Measurement of the concentrations of tritiated E1, E2 and E3 in the maternal and fetal perfusates, flowing at approx 10 and 5 ml/min, respectively, indicated that the distribution of E3 is different from that of E2 and resembles the distribution of E1. The simple perfusion system being used shows differences in the distribution of various estrogens between fetal and maternal perfusates which may reflect the in vivo situation and offers the opportunity for experimental examination of various explanations for these differences, e.g. existence of specific carrier systems in the syncytial membranes, specific binding of the estrogens to secreted placental proteins, and actions of placental and decidual 17 beta-hydroxysteroid dehydrogenases.