Pharmacological manipulation of height: qualitative review of study populations and designs.

OBJECTIVE: Various substances that can have an important effect on height are increasingly available. However, research into pharmacological manipulation of height in children has been criticized. There are concerns about diagnostic criteria; about the medical, ethical, and economic ramifications of modulating growth in children with no endocrinological abnormalities; and about biased results due to weak study designs. The authors reviewed articles published since Jan. 1, 1995, to characterize recent research into this area. METHODS: 70 peer-reviewed articles published in 18 journals in 1995 describing effects of hormonal interventions to affect height were reviewed. Study population, intervention, main purpose (safety, physiology, or therapeutic effect), and methodology were examined. The search was expanded after 1995 to list randomized controlled trials (RCTs) investigating pharmacological manipulation in children and its effect on ultimate height in adults. RESULTS: The inexpensive and brief androgen therapy for pubertal delay has been examined in RCTs, but expensive, long-term treatments to alter final adult height in children have rarely been subjected to RCTs. Some outcome reports pooled subjects with different causes of short stature. Documentation of growth hormone deficiency is problematic. CONCLUSIONS: There is a lack of RCTs in which target populations and growth outcomes are explicitly defined. Further research into overcoming barriers to relevant RCT studies is needed.

Growth hormone therapy for non-growth hormone-deficient children with short stature.

A summary of the majority of the available uncontrolled studies of 322 children with idiopathic short stature treated with growth hormone showed that the final height attainment over predicted adult height was only +2.85 cm (+0.49 SD score). Furthermore, a summary of seven studies reported that the spontaneous outcome in children with untreated idiopathic short stature was more than +1 SD score in final height compared to height at presentation; patients with delayed puberty spontaneously gained more than +2 SD score as adults. Recent reevaluations have concluded that short stature is not associated with clinically significant psychologic morbidity, and the psychologic outcome in response to growth hormone treatment of the short normal child showed no discernable difference in psychologic benefit, despite a difference in height gained. A recent editorial has strongly advised against the expanded use of growth hormone in the normal short child.

How do we best measure growth hormone action?

Growth hormone (GH) is a powerful anabolic hormone with a broad spectrum of action that has been assessed with three general parameters: auxological to assess the growth response; biochemical to measure anabolic effects; and body composition. In childhood, linear growth response is assessed with height, short-term changes in height velocity (HV), and attainment of final adult height, which may not be concordant. In both children and adults, the biochemical indices utilized to predict and/or monitor response to GH therapy have included: (1) nonspecific indices: glucose, insulin, urea, protein synthesis, lipid metabolism, and lipoproteins; (2) more specific indices of the GH-IGF axis: GH binding protein, IGF-I, IGFBP-3, and acid-labile subunit; or (3) indices of bone and mineral metabolism: calcium, phosphate, bone alkaline phosphatase, osteocalcin, propeptides of procollagen type I and III, and bone mineral content. For body composition, body mass index, total body % fat, total body or extracellular water, and bone mineral density have been addressed most frequently. Modest changes with wide variability have been observed with most measurements. GH dose is a very significant positive factor for all parameters. Few of the currently available tests can reliably predict and/or monitor response to GH therapy. Of these, serum IGF-I appears to offer the best integrated indicator of the action of GH throughout all age groups.

Does growth-hormone supplementation affect adult height in Turner's syndrome?

BACKGROUND: By comparison with historical controls, the effect of treatment with growth hormone on adult height in Turner's syndrome was initially reported as uniformly and strongly positive. Because randomised controlled trials are not near completion, we report our experiences in an open study. METHODS: We examined adult height, projected and attained, in 31 patients (17 treated with subcutaneous recombinant human growth hormone, up to 15 mg a week, outside of a controlled trial and 14 untreated contemporaries). FINDINGS: Contingency table analysis of attained versus projected height showed significantly higher values in treated patients although only 4 of 17 had final heights of 5 cm or more over projection. Patients' and treatment variables (height, bone-age delay, oestrogen replacement) that interfere with adult height projection confounded the analysis of adult height data. INTERPRETATION: Girls with Turner's syndrome should be counselled cautiously about the expectation of a strongly positive effect of treatment on adult height. Completion of the randomised controlled trials to adult height is needed to establish the effect of growth-hormone supplementation on adult height in Turner's syndrome and the psychological effect of treatment.

Use of Growth Hormone in children with short stature and normal Growth Hormone secretion: a growing problem.

The ethical, economic, psychologic, social, and growth attainment outcome issues related to the use of GH therapy in normal children with short stature are discussed. Although some short children accelerate their growth velocity with GH treatment, the limited available data do not suggest a significant benefit in final height attainment. An international survey of 99 normal short children treated with GH for at least 3 years reported a net improvement in mean height gain of <1 cm/year. Only in one-third was the result considered very good or good; 40% stated that there was no benefit. Thus, it seems unlikely that GH will dramatically increase final height in short children. On this basis, the use of GH for the treatment of the normal child with short stature outside of carefully controlled clinical trials cannot be recommended at present.

Expression of transforming growth factor-alpha and epidermal growth factor receptor in human fetal kidneys.

Beginning at the fifth week of fetal life, successive generations of individual nephrons are induced by contact between metanephric mesenchyme and ureteric bud. Following phenotypic transformation, cells of each primitive renal vesicle undergo a phase of rapid cell division. In order to identify genes which might regulate nephron development in man, we screened adult and fetal kidney RNA for expression of a panel of growth-related genes. Among the genes which were expressed at higher levels in fetal kidney was the epidermal growth factor (EGF) receptor. There is controversy as to the most likely physiologic EGF receptor ligand in fetal kidney; we were able to identify a transcript for transforming growth factor-alpha (TGF-alpha) but not EGF on Northern blots of fetal kidney RNA. Since the abundance of TGF-alpha mRNA is low, we confirmed its presence by polymerase chain reaction amplification. Using specific radioimmunoassays, we also provide direct evidence for TGF-alpha but not EGF peptide in extracts of fetal kidney and mid-gestational amniotic fluid. We suggest that TGF-alpha/EGF receptor interactions may serve an important function in development of human fetal kidney.

The differential regulation by glucagon and growth hormone of insulin-like growth factor (IGF)-I and IGF binding proteins in cultured rat hepatocytes.

The liver is a major site of production of insulin-like growth factor-I (IGF-I) and IGF binding proteins (IGF-BPs). GH decisively influences IGF-I production. To study the role of GH and glucagon in the regulation of IGF-I and IGF-BP production, we examined IGF-I and IGF-BPs secreted by primary rat hepatocytes cultured in a serum-free medium. Glucagon (1 x 10(-8) M) stimulated IGF-I secretion and IGF-BP secretion. Bovine GH (bGH, 300 ng/ml) stimulated IGF-I secretion but suppressed IGF-BP secretion. Combining bGH and glucagon significantly augmented IGF-I secretion above the level seen with each individual agent. The inhibitory effect of bGH on IGF-BP secretion was reversed by glucagon. The major species of IGF-BPs secreted by hepatocytes were found, on Western ligand blotting, to be 24K and 30-34K. All species of secreted IGF-BPs appeared to be comparably affected by glucagon, bGH, and their combination. Northern analysis of IGF-I mRNA revealed three transcripts of 0.7-1.1 kilobases (kb), 1.8 kb, and 7.0 kb. Glucagon stimulated IGF-I mRNA levels 1.8- to 2.0-fold, whereas bGH stimulated IGF-I mRNA levels 2.0- to 2.5-fold. When hepatocytes were incubated with glucagon and bGH for 6 h, IGF-I mRNA levels were augmented 10-fold. Glucagon, in the presence of 50 ng/ml bGH, had a dose-dependent effect on IGF-I mRNA accumulation from a 6-fold level of stimulation at 50 ng/ml of glucagon to a 9-fold level of stimulation at 1000 ng/ml glucagon to a 9-fold level of stimulation at 1000 ng/ml glucagon. This study has demonstrated that glucagon, as well as GH, has significant effects on the production of both IGF-I and IGF-BPs. Of particular interest was the marked augmentation of hepatic IGF-I messenger RNA levels and the reversal of the low levels of IGF-BP production seen on adding glucagon to bGH.

Metabolic effects of growth factors and polycyclic aromatic hydrocarbons on cultured human placental cells of early and late gestation.

The metabolic effects of epidermal growth factor (EGF), insulin, insulin-like growth factor-I (IGF-I), and IGF-II were determined on human placental cells in monolayer culture obtained from early gestation (less than 20 weeks) and late gestation (38-42 weeks). Parameters studied were uptake of aminoisobutyric acid (AIB), uptake of 3-O-methylglucose and [3H]thymidine incorporation into cell protein. Since benzo[alpha]pyrene (BP) inhibits EGF binding and autophosphorylation in cultured human placental cells, particularly in early gestation, we also studied the effect of benzo[alpha]pyrene and other polycyclic aromatic hydrocarbons (PAHs) on EGF-mediated AIB uptake. The metabolic effects of EGF, insulin, and the IGFs in cultured human placental cells varied with gestational age and the growth factor studied. All three classes of growth factors stimulated AIB uptake in both early and late gestation at concentrations from 10-100 micrograms/L, well within a physiological range. However, insulin stimulation of AIB uptake was maximal at a high concentration (200 micrograms/L) in both early and late gestation cells, suggesting an action via type 1 IGF receptors rather than via insulin receptors. EGF stimulated 3-O-methylglucose uptake only in term placental cells. No significant stimulation of [3H]thymidine incorporation by any of the growth factors tested was seen with either early or late gestation cells. The effect of PAHs on AIB uptake by cultured placental cells was variable. BP alone stimulated AIB uptake by both very early and late gestation cells and enhanced EGF-stimulated AIB uptake. alpha-naphthoflavone alone inhibited AIB uptake at all gestational ages and inhibited EGF-stimulated AIB uptake. beta-Naphthoflavone and 3-methylcholanthrene minimally inhibited AIB uptake by early gestation cells and did not modify EGF-stimulated uptake at any gestational period. Our prior results demonstrated that BP more significantly inhibited EGF than IGF or insulin receptor binding as well as autophosphorylation in early gestation placenta, and that BP was the most potent of the PAHs tested. Thus, the direct effect of the PAHs on placental EGF receptors and amino acid transport may indicate altered function of EGF in the regulation of placental growth in smoking mothers that is developmentally regulated.

Effects of mannose-6-phosphate on receptor-mediated endocytosis of insulin-like growth factor-II.

The insulin-like growth factor-II (IGF-II) and lysosomal enzymes containing a mannose-6-phosphate (M6P) recognition site bind to different sites of the same receptor molecule. We have observed that M6P increases the receptor-mediated uptake of IGF-II into IM-9 cells. We now confirm this phenomenon in a different line, the H-35 rat hepatoma cells, and present additional characterization of the underlying mechanisms. When incubated in the presence of radiolabeled IGF-II, H-35 cells accumulated, in a time-dependent fashion, radioactivity that was resistant to removal by trypsin digestion at 15 C, indicating that it was endocytosed. In the presence of 3 mM M6P, endocytosed counts were approximately 2-fold higher after 5 min of incubation, an enhancement that peaked at 10 min, then declined, but was still evident after 40 min (1.5-fold). The rate of release of cell-associated IGF-II, degraded or intact, as measured in a chase experiment, was not affected by M6P. These observations indicate that M6P increased accumulation of IGF-II by accelerating its rate of endocytosis rather than by interfering with IGF-II degradation or with the recycling of intact hormone-receptor complexes to the cell surface. Electrophoresis after affinity cross-linking of labeled cells demonstrated that the enhancement in radioactivity could be located at a molecular size of approximately 250 kDa, corresponding to IGF-II-receptor complexes. Preincubation with M6P did not significantly alter the specific binding of IGF-II to the cell surface of H-35 cells, as measured by a binding assay at 4 C. Finally, pretreatment with cycloheximide for up to 8 h, to remove all newly synthesized lysosomal enzymes bound to the M6P/IGF-II receptor, did not affect IGF-II endocytosis beyond what could be accounted for by some loss of receptor, suggesting that the observed effect of M6P is due to the binding of M6P itself to the receptor and not to displacement of lysosomal enzymes. We conclude that simultaneous occupancy of the M6P/IGF-II receptor by ligands on both binding sites enhances its rate of endocytosis.

A serum-free system for culturing human placental trophoblasts.

We have compared hormone production by early gestation and term human placental trophoblasts cultured in Ham's F10 medium containing 10% fetal bovine serum with that by cells cultured in serum-free HB102 medium. Mean daily production of progesterone on Days 3 to 7 was approximately 25% less by both early gestation and term cells cultured in HB102 as compared to Ham's F10, but production was maintained at a stable level for at least 7 d longer than the cells in Ham's. Estradiol production from 10(-6) M dehydroepiandrosterone by both early gestation and term cells was comparable in both media. Human placental lactogen production on Days 3 to 7 was 40% less by cells cultured in HB102. Human chorionic gonadotropin (hCG) output by early gestation cells was also 50% less in HB102 but term cells in HB102 produced twice as much hCG as those in Ham's F10. 3B-Hydroxysteroid dehydrogenase (3BHSD) activity in early gestation and term cells and 11B-hydroxysteroid dehydrogenase (11BHSD) activity of early gestation cultures was comparable in the two media. 11BHSD activity was decreased in the term cultures, and this decrease was more marked in Ham's than in HB102. Sulfatase and aromatase activities in the early gestation cultures were comparable in both media; sulfatase activity was comparable and aromatase activity only 20% less in the term cells cultured in HB102. These results indicate that serum-free HB102 supports differentiated function of human trophoblast cells and is useful for studies of placental activity for as long as 14 d in culture.

Evaluation of placental morphology and growth factor receptors in women receiving antiepileptic drugs: a pilot study.

A direct teratogenic effect of antiepileptic drugs (AEDs) on the fetus has been postulated. However, there may also be a primary effect of AEDs on placenta, with secondary consequences to the fetus. Thirteen carefully medically controlled epileptic women were followed prenatally and perinatally. Placentas from all were investigated morphologically and in two with electronmicroscopy. In nine patients, the placentas were studied with respect to insulin and insulin-like growth factor (IGF-I and IGF-II) receptor activity. In this pilot study of a group of carefully medically controlled epileptic women with low levels of AEDs, the results did not differ from normal controls. The possibility of adverse effects of higher concentrations of AEDs cannot be excluded, however.

Benzo(a)pyrene inhibits epidermal growth factor binding and receptor autophosphorylation in human placental cell cultures.

Studies investigated the effects of benzo(a)pyrene (BP) treatment on epidermal growth factor (EGF) receptor binding and kinase activity in human placental cell cultures. Specific binding of 125I-EGF to cells from early gestation placentae was significantly decreased by 37 and 60% following exposure to 1 and 10 microM BP, respectively, for 24 hr. In contrast, cells cultured from term placentae showed no inhibitory effect of either concentration of BP. Specific binding of 125I-labeled insulin and insulin-like growth factors-I and -II to early gestation cells was decreased only 15-18% at 10 microM BP, which indicates that loss of membrane receptors appears to be selective for EGF. Scatchard analysis of early gestation cells revealed that BP was associated with a dose-dependent loss in the number of high affinity EGF binding sites. Evidence from cross-linking and autophosphorylation experiments confirmed that the Mr 170,000 binding protein was decreased in a dose-dependent manner following BP treatment. In comparison, term placental cells exhibit a 26% loss of EGF receptor autophosphorylation without alteration in binding following exposure to 10 microM BP. Thus, early gestation cells exhibit a BP-related down-regulation of EGF receptors, whereas term placental cells show receptor desensitization. No adverse effect of BP treatment was observed on the incorporation of [35S] methionine into proteins secreted by early gestation cells. Further experiments compared the effects of BP with the related poly-cyclic compounds beta-naphthoflavone, alpha-naphthoflavone, and 3-methylcholanthrene. In early gestation cells, EGF binding and receptor autophosphorylation were measurably decreased at 10 microM concentrations of these polycyclic compounds, but to a lesser extent than observed with BP. In term placental cells, however, EGF binding was unchanged or increased, whereas receptor autophosphorylation was decreased 10-26%. Thus, exposure of term placental cells to these polycyclic compounds leads to a dissociation between EGF binding and receptor protein kinase activity. Finally, aryl hydrocarbon hydroxylase activity was induced 20- to 200-fold in early placental cells exposed to BP, beta-naphthoflavone, and 3-methylcholanthrene. In summary, the direct effects of BP and related compounds observed on placental EGF receptors may indicate altered function of EGF in the regulation of cell growth and differentiation in the human placenta.

Determination of insulin and insulin-like growth factors in the ovarian circulation.

Recent in vitro studies implicate the ovary as an extra-hepatic source of insulin-like growth factors (IGFs) with production regulated by gonadotropins and local steroids. Because previous studies have failed to show any significant variations in IGF levels in peripheral blood during the menstrual cycle, we measured the concentrations of IGF-I, IGF-II, and insulin in ovarian and peripheral venous blood samples obtained simultaneously from nine women undergoing abdominal hysterectomy to obtain more detailed data on the ovarian contribution. A significant decreased ovarian gradient was found for IGF-II but not for IGF-I or insulin. Although there was no significant ovarian vein insulin gradient, insulin levels were higher in follicular than in luteal phase ovarian samples. These data suggest that IGF-II may be locally regulated by the ovary. Both insulin and IGFs may regulate ovarian function in vivo.

A prospective study of the development of growth hormone deficiency in children given cranial irradiation, and its relation to statural growth.

Although GH deficiency (GHD) is the most frequent hormonal abnormality that occurs after cranial radiation, the natural course of this complication and its relationship to growth in children are not known. Therefore, we undertook a 2-yr prospective study of 16 children, aged 1.7-15 yr at the time of treatment, who received cranial [31-42 Gy (1 Gy = 100 rads)] and spinal radiation for medulloblastoma or ependymoma (group I). Their growth was compared to that of 11 children given similar doses of cranial radiation only (group II). The mean plasma GH response to arginine-insulin test (AITT) was 9.1 +/- 1.5 (+/- SE) micrograms/L in group I and 8.5 +/- 1.8 micrograms/L in group II (P = NS). After 2 yr, 16 of the 27 children had a peak plasma GH value below 8 micrograms/L after AITT, and 10 children had a peak response less than 5 micrograms/L. In addition, in group I, AITT and sleep-related GH secretion were compared; at the 2 yr follow-up only 3 of 13 children had discrepant results. At the 2 yr follow-up children treated by cranial and spinal radiation had a mean height of -1.46 +/- 0.40 SD below the normal mean. In contrast, the children given only cranial radiation had a mean height of -0.15 +/- 0.18 SD; P less than 0.02. Therefore, most of the growth retardation appeared to be due to lack of spinal growth. GHD is thus an early complication of cranial radiation in these children, and no significant growth retardation can be attributed to GHD during the first 2 yr. These data contribute to the organization of follow-up in irradiated children in order to decide when human GH treatment is necessary.

Mannose 6-phosphate increases the affinity of its cation-independent receptor for insulin-like growth factor II by displacing inhibitory endogenous ligands.

The insulin-like growth factor II (IGF-II) and glycoprotein lysosomal enzymes containing mannose 6-phosphate (M6P) bind with high affinity to two separate sites on the same receptor molecule (Morgan et al. Nature 329:301). The addition of free M6P significantly increases the affinity of some preparations of the M6P/IGF-II receptor (M6P/IGF-II-R) for IGF-II. We conducted this study to test the hypothesis that this effect is the result of displacement of M6P-related ligands that inhibit IGF-II binding. First we found that although M6P caused a 66% increase in the binding of IGF-II to microsomes prepared from IM9 cells, it had no effect, under identical conditions, on binding to receptor on the surface of intact cells. Secondly, extensive washing of rat liver microsomes in the presence of 3 mM M6P, followed by removal of the M6P by further washing, abolished the effect by raising binding to levels seen in the presence of M6P. M6P, then, had no additional effect. Finally, when IGF-II-affinity purified receptor was repurified by ultracentrifugation on a sucrose gradient, binding to the pure receptor peak was not affected by M6P. We conclude that there is no intrinsic positive cooperativity between free M6P and the IGF-II-binding site of the M6P/IGF-II-R. The reported M6P-induced increase in IGF-II binding appears to be due to the displacement of contaminating inhibitory endogenous ligands.

Transient growth deceleration in normal short children. A potential source of bias in growth studies.

We have examined the records of 21 short children evaluated for growth hormone deficiency and found not to be deficient. Their growth velocity was evaluated for at least 6 months, both before and after testing. Without any specific therapeutic intervention, growth velocity was significantly higher after testing, as compared with before. We attribute this apparent "therapeutic" effect of testing to a selection bias, due to the fact that, in normal clinical practice, children are selected for testing immediately following a period of slow growth, and that decelerations of growth are very often transient. Studies of growth-stimulating treatments using children as their own controls should, for this reason, be interpreted with caution.

Insulin-like growth factor-binding proteins in hypophysectomized rat liver: characterization and subcellular localization.

We have characterized binding proteins for insulin-like growth factors (IGFs) in hepatic subcellular fractions and in the washed supernatants of these fractions in normal and hypophysectomized (hypox) rats. In the course of assessing IGF-II-binding sites on rat liver microsomes, we observed that [125I] IGF-II binding to the microsomal membranes of hypox rats was much lower than that in normal rats. Paradoxically, binding increased in hypox animals at low concentrations (0.5-5 ng/ml) of unlabeled IGF-II. After resuspension and centrifugation (washing) of the microsomes, no difference was found in [125I]IGF-II binding to hypox vs. normal microsomes. However, the binding of [125I]IGF-II to the washing supernatant (SN) from hypox rat microsomes was greater than binding to that from normal animals. Binding to SN was inhibited by unlabeled IGF-II in a dose-dependent manner. Scatchard analyses indicated that the affinity constant for binding by hypox SN was higher than that of normal SN and the microsomal fractions of both hypox and normal rats. After further subfractionation of the liver, no binding activity was found in SN from plasmalemma, whereas it was about 20% of input counts per min of [125I]IGF-II in SN from combined Golgi-endosome fractions of hypox rat liver. We next compared IGF-binding moieties in microsomal SN with those in plasma using cross-linking of [125I]IGF-II followed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. In normal rat plasma, we observed the presence of 42K, 39K, 31K, and 27K binding complexes. In hypox rat plasma only a 42-39K doublet was found. In the SN of normal rat microsomes, the predominant complex migrated at 39K and was distinguishable only after acidification. In the SN of hypox rat microsomes, the 42K complex was predominant, with a minor 34K complex. These studies have identified IGF-binding moieties in hepatic tissues, particularly in hepatic vesicular elements, which interfere in the binding of IGF-II to membrane receptors. Their presence in these receptor-rich elements may influence IGF binding to intracellular receptors and, hence, the biological activity of the peptide.