| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Research Centre for Endocrinology and Metabolism (C.O., B-Å.B., O.I., M.S.), Sahlgrenska University Hospital, S-41345 Göteborg, Sweden; Diabetes Branch (C.O.), The National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland; Department of Connective Tissue Biology (T.T.A.), Institute of Anatomy, University of Aarhus, Aarhus, Denmark; and Eli Lilly, Netherlands (M.S.), Nieuwegein, Netherlands
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Effects of GH...III. Effect of GH...IV. Effects of GH...V. Effects of GH...VI. Summary and ConclusionsReferences |
|---|
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Effects of GH...III. Effect of GH...IV. Effects of GH...V. Effects of GH...VI. Summary and ConclusionsReferences |
|---|
Due to limitations in the supply of GH, a limited number of animal and clinical studies were performed until the mid-1980s when recombinant human GH became available. The initial use of recombinant human GH was restricted to treatment of growth-retarded GH-deficient (GHD) children. However, it is now well established that GH also exerts important effects in adults, and GH treatment of GHD adults is now approved in several countries. Recent studies, in both animals and humans, have demonstrated that GH exerts potent effects on bone remodeling.
In this article we will discuss the role of GH in the process of bone growth until peak bone mass is achieved and present evidence that an increased endogenous production of GH or treatment with GH might increase bone mass in adults. Recent studies of the cellular mechanism of action for GH in the regulation of bone growth are given in Section II. It is proposed that GH stimulates longitudinal bone growth directly by stimulating prechondrocytes in the growth plate followed by a clonal expansion caused both by the GH-induced local production of insulin-like growth factor I (IGF-I) and by a GH-induced increase in circulating levels of IGF-I. However, the main purpose of this article is to present recent data indicating that GH is important in the regulation of bone remodeling. Finally we will present a hypothetical model for the mechanism of action of GH in the regulation of bone remodeling and bone mass.
|
II. Effects of GH on Longitudinal Bone Growth |
|---|
|
TopAbstractI. IntroductionII. Effects of GH...III. Effect of GH...IV. Effects of GH...V. Effects of GH...VI. Summary and ConclusionsReferences |
|---|
Several hormones are important for normal postnatal longitudinal bone growth, but it is generally accepted that GH is the most important hormone in this respect. Furthermore, it has been demonstrated that GH stimulates growth of cartilage and other tissues by increasing the number of cells rather than by increasing cell size (11, 12, 13, 14). A widely discussed question during the last two decades has been whether GH acts on tissues directly, or whether the effect is mediated by a liver-derived growth factor, initially called sulfation factor, but later renamed somatomedin, and subsequently shown to be identical to IGF-I. According to the original somatomedin hypothesis, GH stimulates skeletal growth by stimulating liver production of somatomedin which, in turn, stimulates longitudinal bone growth in an endocrine manner (15, 16, 17).
In the early 1980s the somatomedin hypothesis was challenged by a study demonstrating that injection of GH directly into the rat tibia growth plate stimulated longitudinal bone growth at the site of injection (18). This initial observation has subsequently been confirmed and extended, and it is now well documented that GH stimulates growth of many different tissues directly (19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34).
By studying the effects of GH and IGF-I in 3T3 preadipocytes, Green and co-workers (35, 36) made the observation that GH and IGF-I act on cells at different stages of maturation. Thus, GH was found to stimulate young preadipocytes, whereas IGF-I stimulated cells at a later stage of development. The hypothesis by Green and co-workers (35), that GH acts on progenitor cells and that IGF-I stimulates the subsequent clonal expansion, was named the "dual effector theory." The finding that GH stimulates longitudinal bone growth directly (18) and increases the local production of IGF-I by stimulating transcription of the IGF-I gene (37) led to the proposal that the dual effector theory of GH action is valid for the regulation of longitudinal bone growth as well (10). Subsequent in vitro studies using cultured epiphyseal chondrocytes in suspension revealed that GH and IGF-I stimulate cells at different stages of maturation. Thus, GH stimulates the colony formation of young prechondrocytes, whereas IGF-I stimulates cells at a later stage of maturation, giving support to the hypothesis that cell maturation is indeed an important factor determining responsiveness of epipyseal chondrocytes to GH and IGF-I (38, 39, 40, 41, 42). The hypothesis that GH preferentially acts on chondrocyte progenitor cells in vivo is directly supported by another study from our laboratory. By labeling slowly cycling prechondrocytes with radioactive thymidine and studying the subsequent labeling pattern in sagittal sections of the tibia growth plate by means of autoradiography, the observation was made that local injection of GH increased the number of labeled cells in the prechondrocyte layer of the growth plate. In contrast, IGF-I did not stimulate incorporation of radioactive thymidine in cells in the same layer (43). Using a histomorphometric technique, it was found that GH as well as IGF-I has the capacity to stimulate prechondrocytes, as both GH and IGF-I reduced the cell cycle time of prechondrocytes. However, it was found in the same study that growth plate prechondrocytes from GH-treated animals had a 50% shorter cell cycle time compared with IGF-I-treated animals (44). These data indicate that at least some of the growth-promoting effect of GH is exerted via direct stimulation of prechondrocytes.
B. Results supporting an important physiological role of IGF-I for
bone growth
Studies performed during the last 20 yr involving systemic
administration of IGF-I to GH-deficient animals and man suggest that
both IGF-I and GH have the capacity to stimulate longitudinal bone
growth in vivo (45, 46, 47, 48, 49, 50, 51, 52, 53, 54). Elimination of IGF-I and the IGF-I
receptor by homologous gene recombination have demonstrated that the
IGF-I-signaling pathway is very important for tissue development and
growth. Thus, mice with IGF-I deficiency show severe retardation of
statural growth that first becomes apparent at day 12 in embryonic
life, and subsequent postnatal growth is severely retarded (55, 56, 57, 58).
IGF-I receptor "knock-out" mice are affected more profoundly and
die of respiratory failure early postnatally due to poor development of
respiratory muscles (57). Furthermore, a patient with a deletion of the
IGF-I gene demonstrated intrauterine growth retardation and postnatal
growth failure (59). These experimental studies and the clinical
observation clearly demonstrate that a normal expression of IGF-I, as
well as its receptor, plays a critical role for normal growth and
tissue development. However, these experiments are unable to answer the
question of whether locally produced (autocrine/paracrine acting) IGF-I
is more important for normal tissue growth and development than
circulating (endocrine acting) IGF-I.
Several studies have shown that systemic administration of recombinant IGF-I stimulates longitudinal bone growth as well as body weight gain in hypophysectomized rats, giving support to the theory that IGF-I has endocrine actions on statural growth. Interestingly, administration of IGF-I particularly promoted the growth of nonskeletal tissues. Thus, the effect of IGF-I on kidney, spleen, and thymus growth was larger in magnitude compared with other tissues (46, 47, 60). The quantitative difference in tissue response to IGF-I has also been found in transgenic mice overexpressing IGF-I, suggesting that IGF-I has particularly important functions in nonskeletal tissues (61, 62). These data demonstrate that systemic delivery of IGF-I has the capacity to increase growth in animals.
C. Evaluation of the somatomedin theory vs. the dual effector
theory
The fact that both GH and IGF-I stimulate tissue growth
makes an analysis of the relative importance of the peptides for this
effect, in terms of spatial and temporal patterns, quite complex. It
seems justified to critically analyze available experimental and
clinical data and find out how available data fit into the two
different theories, the somatomedin theory and the dual effector
theory. The effect of systemic administration of GH and IGF-I to
hypophysectomized rats has shown that GH and IGF-I have independent and
differential functions (45, 46, 63). When the two compounds are given
together, they exert additive or synergistic effects (45, 60, 64).
Also, administration of GH to animals treated with maximal doses of
IGF-I stimulates growth further (63). Furthermore, the differences
between GH and IGF-I are quite obvious in transgenic animals
overexpressing either GH or IGF-I. Thus, GH-transgenic animals grow to
approximately twice the size of their normal littermates (62, 65). In
contrast, mice generated from a cross of mice overexpressing IGF-I and
mice lacking GH-expressing cells demonstrate an increase in
longitudinal bone growth and body weight when compared with their
GH-deficient controls. However, IGF-I transgenic mice do not grow more
than their nontransgenic siblings (61, 66), demonstrating that
overexpression of GH, but not IGF-I, causes supranormal growth. Local
administration of GH, but not IGF-I, stimulates the local production of
IGF-I by stimulating the transcription of the IGF-I gene (22, 37),
giving direct experimental support to the notion that there is an
interplay between GH and IGF-I. Administration of antibodies to IGF-I
abolishes the stimulatory effect of locally administered GH (31),
supporting the theory that the locally produced IGF-I has an important
functional role in the expression of the effect of GH at the site of
the local tissue level (10, 67).
Treatment of GH insensitivity syndrome (GHIS) patients with recombinant IGF-I has shown that IGF-I is quite effective in stimulating statural growth for 1–2 yr (49, 50, 51, 53, 54, 68, 69, 70, 71, 72), supporting the somatomedin theory. However, available clinical data suggest that the effect of IGF-I subsequently becomes less effective, perhaps due to a decreased rate of stimulation of prechondrocytes, a lack of GH- induced IGF-binding protein 3 (IGFBP-3) and/or a suboptimal IGF-I administration. However, from these clinical studies it is difficult to make a general conclusion whether IGF-I stimulates tissue growth by endocrine or autocrine/paracrine mechanisms under physiological circumstances in the intact organism. The question whether autocrine/paracrine or endocrine IGF-I is the more important factor for the stimulation of tissue growth will probably not be solved until tissue growth can be studied in transgenic animals with tissue-specific gene deletions of IGF-I or the IGF-I receptor.
Taken together, available data suggest that GH stimulates longitudinal
bone growth directly by stimulating prechondrocytes in the growth plate
followed by a clonal expansion caused both by the GH-induced local
production of IGF-I, and by a GH-induced increase in circulating levels
of IGF-I. GH is the major determinant for the stimulation of progenitor
cells, although it is possible that IGF-I might stimulate progenitor
cells to some extent (Fig. 1
).
|
|
III. Effect of GH in Vitro |
|---|
|
TopAbstractI. IntroductionII. Effects of GH...III. Effect of GH...IV. Effects of GH...V. Effects of GH...VI. Summary and ConclusionsReferences |
|---|
B. Effects of GH on osteoblasts
1. GH directly stimulates osteoblasts. The effect of GH has
been studied in a number of osteoblastic cell lines and primary
isolated cells of various origin, including human, chicken, rat, and
mouse primary cells, and the SaOS-2 human and UMR 106.01 rat
osteosarcoma cell line. GH induces proliferation of primary isolated
rat (21, 73), mouse (74), chicken (32), human (24, 75, 76, 77, 78), and rat
osteosarcoma cells (19, 79), as well as cells from a rat
osteoblast-like cell line (80) and human osteosarcoma cells (75, 81)
(Fig. 2). The effective concentrations of GH are in the physiological
range (half-maximal stimulation at 10–50 ng/ml), suggesting that GH
exerts direct actions on osteoblasts. Not only does GH stimulate the
proliferation of osteoblasts, but, in some studies, it also stimulates
differentiated functions of these cells. Thus, typical phenotypic
functions of osteoblasts such as AP, osteocalcin, and type I collagen
are stimulated by GH (4, 21, 24, 74, 80, 82). For osteoblasts it is
difficult to find a good model system for the identification of the
actual target cell of GH action. However, bone marrow- derived
precursors of human bone cells are responsive to GH (76, 77),
suggesting, in analogy to the actions of GH in early progenitor cells
in adipose tissue and cartilage, that GH interacts with progenitor
cells.
|
Whereas circulating levels of IGF-I are GH dependent, GH may not be the chief determinant of local IGF-I production in bone. Thus, in vitro regulatory effects of estrogen, PTH, and cortisol, as well as a variety of local growth factors, on IGF I production have been demonstrated (86, 88, 94, 95, 96, 97, 98, 99, 100, 101, 102). The regulation of local IGF-I by growth factors and hormones is of potential clinical importance, but in this article only GH-modulated effects in bone are discussed. A stimulation of osteoblastic IGF production by GH has been demonstrated by some authors (80, 82) but not by others (24, 83). To examine the importance of IGF-I as a mediator of GH action, endogenous IGF-I was sequestered by an antiserum to IGF in bone cell cultures. As a result, the proliferative action of GH was abolished (21), indicating that local IGF-I is important for GH-induced cell proliferation. In another study, by Scheven et al. (75), it was demonstrated that GH induced osteosarcoma growth but not growth of human osteoblast-like cells when the cells were cultured in the presence of IGF-I antibodies. In summary, GH induces IGF-I expression in rodent osteoblasts, while the induction of IGF-I by GH in human osteoblasts is uncertain.
The bioactivity of IGFs in bone tissue is modulated by several IGFBPs, mainly IGFBP-3, -4, and -5 (103). Therefore, some of the GH effect may be mediated via a regulation of the local production of IGFBPs in osteoblasts. It is well known that GH treatment increases serum levels of IGFBP-3 (104, 105, 106, 107, 108, 109), and the complex of IGF-I and IGFBP-3 is more effective in stimulating cortical thickness in ovariectomized (OVX) rats than IGF-I alone (110). IGFBP-3 is produced by osteoblasts. GH increases IGFBP-3 production in rat cells (73, 96, 111, 112), while no effect of GH is seen on IGFBP-3 expression in human cells (24, 83, 113). IGFBP-4 was originally isolated from bone as the inhibitory IGFBP (114), while IGFBP-5 is regarded as a stimulatory IGFBP for osteoblastic proliferation (115, 116, 117). In rat, as well as in human osteoblasts, it was found that GH decreases IGFBP-4, as determined by ligand blotting (113, 118), while no effect of GH was seen on IGFBP-4 protease activity in human osteoblast-like cells (119). In primary rat osteoblasts, IGFBP-5 mRNA levels were increased 2-fold after GH treatment (96). Interestingly, a recent clinical study has demonstrated that GH treatment increases serum levels of IGFBP-5 in GHD children (109). Whether this effect of GH is a direct effect on osteoblastic IGFBP-5 production remains to be shown. In conclusion, there are some indications of a GH-induced regulation of IGFBPs that potentially might have a regulatory role in bone metabolism.
Many functions of GH can be exerted without prior synthesis of IGFs. Thus, the expression of the protooncogenes c-fos, c-jun, jun B, and c-myc are expressed in the presence of protein synthesis inhibitors (120). Recently, using human osteoblast-like cells, Melhus and Ljunghall (121) demonstrated that different sets of genes were induced by IGF in some cases and GH in others, indicating that these factors have separate actions. In conclusion, it appears that some of the effects of GH on osteoblasts are mediated by IGFs, but others are not.
3. Regulation of GH receptor (GHR) expression. Barnard and
colleagues (19) were the first to show specific high-affinity GHRs on
osteoblast-like cells (UMR 106.06 cells). The presence of these
receptors has been confirmed in primary isolated cultured human (78)
and mouse (122) osteoblasts. Serum decreases the number of GHRs (79).
In a search for the factors in serum responsible for this reduction in
GHR expression, it was found that IGF-I and -II decrease the number of
GHR in a dose- and time-dependent manner (123). This decrease was
accompanied by a decrease in the levels of mRNA encoding the GHR.
Similarly, the action of GH on osteoblastic proliferation was decreased
after preincubation of the cells with IGFs (124). Conversely, it was
found that IGFBPs up-regulated GHR number and activity, possibly
through inhibition of IGF activity (123, 124) (Fig. 3). These findings suggest a local
feedback of the GH/IGF axis at the tissue level. Thus, hypothetically,
IGF decreases GHR number and activity, fine-tuned by the presence of
IGFBPs (123, 125). Leung et al. have suggested that the
negative feedback of the GH/IGF-I axis in skeletal tissue might involve
three different mechanisms: a) liver-derived IGF-I inhibits pituitary
GH secretion, b) bone-derived IGF-I inhibits pituitary GH secretion,
and c) bone-derived IGF-I inhibits local action of GH by reducing GHR
availability (125) (Fig. 4).
|
|
In conclusion, a regulation of GHR expression in osteoblasts may be important for 1) a local autocrine feedback loop in the GH/IGF-I axis and 2) for sex steroids, glucocorticoids, and other factors modulating the effect of GH in osteoblasts.
4. GH signal transduction in osteoblasts. The GHR is a member of the cytokine/hemopoietic growth factor receptor family (129). GH signaling via its receptor has now been shown to be mediated through cascades of protein phosphorylation resulting in activation of nuclear proteins and transcription factors. The GHR itself is not a tyrosine kinase. Instead, after binding of GH to its receptor, an association with a protein, JAK2, occurs. JAK2 is then phosphorylated and in turn phosphorylates the GHR (130, 131, 132). A number of signaling pathways may transduce the signal from this complex to the nucleus. The first cascade is via STAT proteins (133, 134, 135), which upon phosphorylation are translocated to the nucleus and bind to DNA. Also, the Ras-Raf signaling pathway plays a role in the GH-induced signaling (136, 137). IRS-1 and -2 are also proteins functioning as signal transducers for the GHR after they have been phosphorylated on tyrosine residues (138, 139, 140).
In mouse osteoblasts, it has been shown that GH induces the nuclear protooncogenes c-fos, c-myc, c-jun, and Jun-B (120, 141). The formation of diacylglycerol was induced, and the signaling was found to be dependent on a form of protein kinase C. In these cells, the involvement of the phorbol esther-sensitive transregulating transcription factor AP1 in GH-induced gene transcription was demonstrated for the first time (141).
In rat osteosarcoma cells, another signaling protein, annexin 1, was detected recently as being tyrosine phosphorylated upon stimulation of the cells with GH (142). The tyrosine phosphorylation of this protein also occurs after stimulation of cells with epidermal growth factor, pp60v-scr, angiotensin II, and insulin (143). Although it appears as if this is a general mechanism of signal transduction in cells, the exact function of this protein is unknown, as is its place in the already known signaling cascades (143).
C. Effects of GH on osteoclasts
GH increases the number of osteoclasts in the metaphysial bone of
the proximal tibia of hypophysectomized rats (25). However, the
mechanism for this effect is less clear. GHR mRNA has been detected in
mouse marrow cultures (144) and in mouse hemopoietic blast cells (145).
In a recent study by Nishiyama et al. (145), using mouse
stromal cells and hemopoietic blast cells, it was found that GH
stimulates osteoclastic bone resorption through both direct and
indirect actions on osteoclast differentiation and indirect activation
of mature osteoclasts. Factors that may mediate the indirect
GH-regulated osteoclast formation include IGF-I and IL-6, both of which
are involved in osteoclast formation and have been shown to be
regulated by GH (33, 81, 145, 146, 147, 148, 149). It has earlier been demonstrated
that IGF-I supports activation and formation of osteoclasts in cultures
of unfractionated mouse bone cells (146, 149) and that osteoblasts
mediate IGF-I-stimulated formation of osteoclasts in mouse marrow
cultures and activation of isolated rat osteoclasts (150). Furthermore,
human osteoclasts express functional IGF-I receptors (151). In another
study by Ransjö et al. (144), using mouse marrow
cultures, GH caused an inhibition of osteoclast formation by an
IGF-I-independent mechanism. In summary, available data suggest that GH
regulates osteoclast formation but both stimulatory and inhibitory
mechanisms have been presented, probably due to differences in culture
conditions. Future studies are necessary to improve the understanding
of the physiology of GH-induced effect on osteoclast.
|
IV. Effects of GH on Bone Metabolism in Animals |
|---|
|
TopAbstractI. IntroductionII. Effects of GH...III. Effect of GH...IV. Effects of GH...V. Effects of GH...VI. Summary and ConclusionsReferences |
|---|
A. Effects of GH deficiency and GH replacement on bone parameters
Hypophysectomy (HX) of rats with subsequent replacement with
T4 and glucocorticoid is followed by a rapid and pronounced
decrease in the amount of metaphyseal and vertebral body cancellous
bone. Bone volume, trabecular number, and trabecular thickness are
decreased, and bone formation is minimal (159, 160, 161, 162). The cancellous
bone resorption is also enhanced. Using tetracycline labeling, dynamic
histomorphometry demonstrates that bone resorption is enhanced after HX
(161). The result is in accordance with the previously observed decline
in bone mass but perhaps unexpected because static histomorphometric
investigations have clearly shown that HX decreases the number of
osteoclasts and bone surface area covered by osteoclasts (25, 162).
Biochemical markers for bone formation are also decreased after HX.
Thus, circulating osteocalcin declines, and the mRNA levels of
osteocalcin and 
1(I)-procollagen in the bone are decreased (163, 164).
When GH is given to HX rats, increases in both bone formation and the
number of osteoclasts are seen (25, 161). Correspondingly, an increase
in serum osteocalcin and bone mRNA levels of osteocalcin and
1(I)-procollagen is observed (163, 164, 165). Furthermore, GH, but not
IGF-II, increases incorporation of radioactive thymidine and proline in
femur and tibia of HX rats (165). In bone from HX rats a decrease in
mRNA levels of IGF-I is found, and the levels are restored after GH
replacement (163). This observation strongly suggests that GH has a
direct effect on bone cells. However, the bone content of IGF-I protein
was not influenced by HX. In summary, HX of rats results in a decreased
bone formation with a concomitant decrease in bone mass.
GH replacement therapy restores bone formation and bone mass. Conflicting results have been presented regarding the specific effect of GH on bone resorption after HX, probably due to the fact that these animals are also lacking gonadotropins and are sex steroid deficient. Thus, the dwarf rat (dw/dw) with a normal pituitary function, except for GH deficiency, is probably more appropriate for studying the specific effect of GH deficiency. This animal model was recently used in bone mass and metabolic experiments (166, 167, 168, 169, 170). Cancellous bone volume, bone mineral density (BMD), and serum AP are decreased in the dwarf rats, compared with normal rats fed ad libitum and food-restricted animals, although the food restriction caused growth retardation that was similar to that in dwarf rats (169). Dwarf rats treated with GH showed no difference in bone volume when compared with normal animals, while BMD was decreased and serum AP increased in these animals (169). In cortical bone from these dwarf rats, GH treatment caused increased periosteal bone formation and collagen deposition and a slight decrease in BMD (170). Taken together, these studies support the earlier observations in HX rats, i.e., that GH increases bone formation and bone mass in GHD animals.
B. Effects of GH treatment on bone parameters of animals with
normal GH secretion
Normal rats have been used widely for studying the influence of GH
on intact bone, and experiments have been performed in young, adult,
and old rats. However, in almost all of these experiments GH
administration has induced linear bone growth because the growth plates
do not close until the rats are very old (171). Therefore, the data
have to be evaluated in relation to both growth/modeling and remodeling
(172, 173). The response pattern in rats should be compared with the
situation in primates (monkeys, humans), which will be discussed later
on in this section. In primates the growth plates are closed after
sexual maturation and confounding factors, due to stimulation of bone
growth and bone modeling, are of less importance.
1. Effects of GH in rodents. GH administration increases
cortical bone mass in normal rats. Tetracycline labeling of the
mineralization front demonstrates that GH induces subperiosteal bone
formation without influencing the endosteal bone surface (174, 175, 176)
(Fig. 5). The new bone is organized in a
manner similar to that of adjacent bone that was formed before the
start of GH injection, i.e., in concentric lamellae and with
the same direction of the collagen fibers. After withdrawal of GH
administration, the subperiosteal bone formation ceases quickly in
areas with minimal bone formation before the start of GH treatment. A
remaining effect of GH, however, was found in areas where active bone
formation occurred before the start of treatment. The new bone formed
during GH administration is preserved after discontinuation of the
treatment (176). Corresponding to the increased bone mass, there is
also an increase in mechanical strength of the whole bone, and the
mechanical quality of the bone itself is almost the same in GH-injected
animals as in controls (175, 176, 177). The GH-induced subperiosteal bone
formation also shows regional differences. At the outer surface around
the lumbar vertebrae, new bone deposition is seen whereas no effect of
GH is observed at the surface of the vertebrae toward the vertebral
canal (178). GH also causes formation of cavities inside the cortical
shell of the vertebral body in contrast to diaphyseal cortical bone in
rats (178, 179), suggesting that GH exerts site-specific effects on
bone.
|
Cancellous bone mass of the vertebral body does not seem to be affected by GH administration in normal old rats as no differences in bone volume and bone surface/bone volume have been found (178). Apart from increasing cortical bone mass, GH also increases bone turnover. GH administration increases serum osteocalcin and increases formation of bone collagen in both cancellous and cortical bone as determined by in vivo labeling with radioactive proline (181, 182, 183). Bone resorption is also augmented, as shown by measuring excretion of pyridinolines and the specific marker [3H]tetracycline in rats labeled with [3H]tetracycline before GH treatment (184). Because bone matrix is a major reservoir for IGF-I, Yeh et al. measured the content of bone matrix IGF-I after 9 weeks of GH treatment. However, these investigators found no increase in bone matrix content of IGF-I in the GH-treated animals (181).
Intermittent PTH injection to rats increases bone mass primarily by inducing endosteal and cancellous bone deposition, whereas the subperiosteal bone deposition is modest (185, 186). When GH and PTH are given simultaneously, a substantial increase in bone mass of vertebral bodies is seen because the GH-induced subperiosteal bone deposition takes place together with the PTH-induced endosteal and cancellous bone deposition (187), suggesting that different treatment protocols using combinations of PTH and GH might be clinically useful.
In summary, these results from GH-treated rats with a normal GH secretion clearly demonstrate that GH increases cortical bone mass by inducing subperiosteal bone formation while no large effect on cancellous bone mass is seen.
2. Effect of GH in transgenic mice. The creation of the first giant GH-transgenic mouse in 1982 attracted considerable attention from scientists as well as the popular press (65, 188). The extent of GH expression and tissue distribution of GH in the transgenic mice depend on which promoter is attached to the GH gene. In most bone metabolic studies, the metallothionein promoter (MT) fused to the GH gene has been used, resulting in very high serum levels of GH (188, 189, 190, 191, 192, 193, 194). However, two new GH-transgenic lines with a tissue-specific expression resulting in high local concentrations of GH without affecting serum concentrations of GH have recently been described: 1) Baker et al. (158) used the osteocalcin promoter, resulting in GH expression in osteoblasts; 2) Saban et al. (156) used GH driven by ß-globin regulatory-elements, resulting in an erythroid expression with an "adult" expression in the bone marrow.
The femora of MT-GH-transgenic mice with very high serum concentrations of GH demonstrate an increased bone growth, an increased BMC, no change in BMD (BMC/vol), and an increased mechanical strength (188, 194). The increase in mechanical strength was due to an increased cortical width and not due to an improved quality of the bone. Rather, one of the parameters measuring the quality of the cortical bone, the E-module, was decreased in GH transgenic mice (194). It should be emphasized that these mice have been exposed to supraphysiological serum levels of GH (more than 10 times increased) from late prenatal life (194). Interestingly, disproportionate skeletal gigantism has been found in adult MT-GH-transgenic mice, suggesting that supraphysiological GH levels exert differential effects on different parts of the skeleton (189). These studies in GH-transgenic mice with increased serum concentrations of GH give support for the fact that GH increases cortical bone formation, resulting in an increased mechanical strength of the bone. However, the net result on bone mass in old MT-GH-transgenic mice is also highly dependent on bone growth and bone modeling.
The erythroid-specific GH-transgenic mice had increased cortical bone thickness, and the authors suggested that the local effect of GH from erythroid cells in the bone marrow is a major contributor to the increased bone deposition in these GH-transgenic mice (156). However, a slight increase in serum levels of GH was seen, indicating that some of the effect of GH may have been systemic. In the osteoblast promoter-driven GH-transgenic mice the femora demonstrated an increased growth, increased cortical width, and an increased mechanical strength (157, 158). Similar to the situation in the MT-GH-transgenic mice the quality of the bone, as measured with the E module, was decreased (157, 194). As the serum levels of GH were not increased in these GH-transgenic mice, it was concluded that the stimulatory effect on bone formation was caused by local effect of GH.
GH-transgenic mice have also been used as a model to study the functional interaction between male and female sex steroids with increased expression of GH. It was found that preserved gonadal function was a prerequisite for the increase in bone mass caused by overexpression of GH (192, 193).
3. Effects of GH in primates. As discussed above, GH exerts
potent effects in rodents, resulting in an increased bone formation.
However, it is possible that some of these effects are due to bone
growth and bone modeling, as rodents close the epiphyseal plate late in
life. The monkey is a primate and the bone metabolism in these animals
is more similar to that of humans. The effect of GH has been studied in
hypogonadal female monkeys. The monkeys were made hypogonadal by
treatment with a GnRH agonist for 10 months, resulting in a 12%
decrease in BMD (BMC/area). GH supplementation (100 µg/kg/day)
reduced the decline of BMD in GnRH agonist-treated monkeys (195). A
recent study, using old female monkeys, demonstrated that GH (100
µg/kg/day), but not IGF-I (120 µg/kg/day), given for 7 weeks
increased bone formation as measured with mineral apposition and bone
formation rates (196) (Fig. 6). No
additional effect was seen when IGF-I was given together with GH. The
effect was seen both in the tibia and in the femur whereas no
significant effect was seen in the vertebrae. The difference between
the long bones (predominantly cortical bone) and the vertebrae
(predominantly cancellous bone) in terms of GH responsiveness is
similar to what has been described earlier in old rats. These
experimental studies of primates are promising for future human
clinical studies. However, further long-term studies with GH treatment
of primates are needed to elucidate whether the increased bone
formation results in an increased bone mass and mechanical strength.
|
These studies, using OVX rats, indicate that GH alone or in combination with another hormone may be useful in the treatment of postmenopausal osteoporosis. However, further studies need to be performed in old OVX rats and primates with closed growth plates.
5. Effects of GH in animals treated with glucocorticoids. In rats, rabbits, and dogs, glucocorticoid treatment has been shown to decrease bone formation and bone mass (211, 212, 213, 214). The effects, however, vary with species and in the rat model low doses of glucocorticoids increase bone mass and mechanical strength of bone whereas higher doses decrease bone formation, bone mass, and bone strength (215, 216, 217). In mice, simultaneous administration of GH and glucocorticoids prevents the catabolic effect of glucocorticoids whereas this does not seem to be the case in rats. However, the number of experiments is still very limited, which is why the interpretation should be cautious. Using mice, Altman et al. (218) showed that glucocorticoids caused a decline in linear bone growth, trabecular bone volume, cortical bone width, mineral bone content, and bone alkaline- and acid-phosphatase activity. The observed declines in different bone parameters were inhibited when glucocorticoid and GH were given simultaneously. The results correspond well with histological data obtained in a trial in children when GH and glucocorticoid were given either separately or simultaneously (219). In rats a short dose-response study showed that GH is able to prevent glucocorticoid-induced growth inhibition (220). In long-term experiments, however, GH does not seem to counteract the glucocorticoid-induced decline in linear growth, bone formation, and bone mass, although GH alone increases these parameters (179, 221).
C. Effects of GH on fracture healing
When movements between the ends of a fractured bone are possible,
bone healing is initiated by formation of a thick periosteal callus of
woven bone with a central area of cartilage. Through endochondral
ossification the cartilage is subsequently replaced by woven bone.
Later in the healing phase, a marked modeling takes place and hereby
the callus volume declines and the density is enhanced (222, 223). As
GH stimulates both periosteal bone formation and linear growth where
bone formation takes place by endochondral ossification, it has been
natural to examine the effect of GH treatment on healing bones.
In rats, GH administration increases callus formation and mechanical strength of healing fractures (224, 225, 226, 227, 228, 229). The enhanced rate of healing continues after withdrawal of GH (230, 231). A considerable delay in mechanical strength development of healing fractures is seen in old rats and GH treatment partly prevents this delay (232, 233). Augmented callus formation is found in the rat bone defect model, when GH is administered both systemically and locally (155). GH has not previously been applied locally either to intact bone or healing fractures, and the data imply that GH exerts a direct, non-liver-mediated effect on bone tissue (155). In studies in which rats were used, only a few papers show no effect of GH on healing bone defects and fractures (234, 235), and GH has not been able to stimulate formation of new bone in titanium bone conduction chambers (236).
In rabbits, GH has not been able to increase callus formation or mechanical strength in healing fractures and bone defects (234, 237, 238, 239). However, when subperiosteal bone formation was induced by applying a cerclage band around the femur, GH was able to enhance bone formation in rabbits (240).
In dogs, GH administration augments callus formation in bone defects, and in human trials GH treatment stimulates healing of fractures and pseudoarthroses, when evaluated by radiographs and clinical examination (241, 242, 243, 244).
In summary, GH treatment in rats obviously increases callus formation and the mechanical strength of healing bones, whereas the response in the rabbit model seems to be much weaker. At present, it is not possible to evaluate whether GH treatment has any role in human fracture healing because only a few clinical trials and experiments in higher animals have been performed.
|
V. Effects of GH on Bone Metabolism in Humans |
|---|
|
TopAbstractI. IntroductionII. Effects of GH...III. Effect of GH...IV. Effects of GH...V. Effects of GH...VI. Summary and ConclusionsReferences |
|---|
2. GH deficiency (GHD). There is no conclusive data on the effects of GHD on bone remodeling in adults. Serum levels of osteocalcin, reflecting osteoblast activity and bone formation, have been found to be decreased (263, 264, 265, 266, 267), increased (268), or unchanged (269). Most studies have shown that there is no difference in resorption markers between controls and adult GHD patients (264, 268, 270).
a. Bone mass in adult patients with childhood-onset GHD.
In
children with GH deficiency, a relative osteopenia is found before the
start of exogenous GH treatment, an effect that might be due to a delay
in skeletal maturation (271, 272, 273). Several studies have shown low bone
mass in adults with childhood-onset GHD (268, 273, 274, 275, 276, 277, 278, 279, 280). In a
cross-sectional study of 30 young adult males, with childhood-onset
GHD, Kaufman et al. (273) found decreased bone mineral
content in the lumbar spine and forearm compared with age- and height-
matched controls. The BMC in the lumbar spine was shown to be between 9
and 19%, and in the forearm 20 and 30% lower compared with controls,
using dual- and single-photon absorptiometry, respectively. A similar
decrease in BMC was observed in patients with multiple pituitary
deficiencies and isolated GHD. These observations were recently
confirmed by de Boer et al. (275) who performed a similar
cross-sectional study in 70 adult men with childhood-onset GHD. This
investigations found that the BMD area (BMC/bone area) in GHD patients
was significantly reduced at the lumbar spine as well as the
nondominant hip. In fact, in 33% of the patients the lumbar spine BMD
area was at least 2 SD lower than normal. They also
observed a positive relationship between body height and BMD area.
Patients and controls differed in body height, which partly explained
the difference in BMD area. However, also after correction for bone
size, the difference in BMD area between patients and controls still
remained. Similar results were obtained in patients with multiple
pituitary deficiencies and isolated GHD.
The similar results observed both by Kaufman et al. and de Boer et al. in patients with multiple pituitary deficiencies and isolated GHD suggest that lack of GH is the most important factor behind the observed low bone mass in childhood onset GHD (273, 275). A reported reduction in vertebral trabecular bone density assessed by CT technique in 10 males with childhood-onset isolated GHD further supports this conclusion (274). There is no evidence suggesting that bone loss is enhanced after cessation of GH treatment in young adults (273). Thus, it is conceivable that insufficient acquisition of bone mass during childhood and thus reduced peak bone mass explain the reduced BMC and BMD observed in these patients. The cause of the reduced bone mass is probably suboptimal GH therapy in these patients. Patients included in the cited studies were mainly treated with GH when the supply of GH was limited, and the doses used were lower and cessation of treatment occurred earlier than current pediatric practice. In patients with hypopituitarism of childhood onset, the induction and timing of puberty are also important in reaching the optimal peak bone mass. Boys with constitutionally retarded puberty will achieve a lower peak bone mass than boys with puberty of normal onset (281). At present there are no studies showing that GH replacement during childhood results in a normalization of BMD when peak bone mass is reached, suggesting that GH also is important for the additional increase of bone mass that occurs after completion of linear growth. It has been suggested that GH treatment should be continued until the attainment of peak bone mass, irrespective of the height achieved (282).
b. Bone mass in adult patients with adult onset GHD.
An
increased prevalence of osteoporosis has been found in several recent
studies of patients with adult-onset GHD (283, 284, 285, 286, 287, 288, 289). In a population of
122 hypopituitary patients, Wüster et al. (283)
observed that 57% of the patients had low bone mass of lumbar spine as
assessed with dual photon absorptiometry, and 73% of the patients had
low bone mass of the proximal forearm as assessed with single photon
absorptiometry. Johansson et al. (284) studied 17 adult GHD
men and found that total, but not spinal, BMD, measured with DEXA, was
lower in the patients compared with controls. In a study by Rosén
et al. (286) of 95 (55 males and 40 women) patients with
adult-onset GHD with a mean age of 54 yr, BMC was assessed in the third
lumbar vertebra with dual-photon absorptiometry. The control population
comprised 214 women aged 35–80 yr and 199 men between 16 and 79 yr of
age. BMC was found to be lower in all males and in females with
untreated as well as treated gonadal deficiency. BMC was lower in
patients below 55 yr of age and normal in patients above 55 yr of age.
Holmes et al. (287) measured vertebral trabecular BMD with
quantitative computed tomography (QCT), total BMD and BMD in lumbar
spine and hip with DEXA, and BMC in the forearm with single photon
absorptiometry (SPA) in adult patients with GHD. There was a highly
significant reduction in QCT and in DEXA of the lumbar spine and in SPA
of the forearm in these patients. Similarly, as has been shown in
patients with childhood-onset GHD, there was no difference in Z-scores
between those patients with isolated GH deficiency and those with GH
and gonadotropin deficiency. In a subgroup analysis of patients with an
estimated age above 30 yr at the onset of the disease, a reduction was
still present in QCT and in DEXA of the lumbar spine, and in SPA of the
forearm. Interestingly, older adults had less reduction in bone mass
than younger, confirming the observation by Rosén et
al. (286, 287). In a cross-sectional study comprising 64
hypo-pituitary patients, Beshyah et al. (288)
demonstrated a significant reduction in lumbar spine BMD and BMC in
both male and female patients compared with controls. The area and the
width of the vertebra were similar in patients and controls. In
contrast, Degerblad et al. (289) observed normal total,
spine, and hip BMD in males with adult onset GHD and Kaji et
al. (290) found a normal BMD in the spine and midradius of
patients with adult onset GHD. However, Degerblad et al.
(289) found low total, spine, and hip BMD in women with adult onset
GHD. A recent study in elderly patients (over 60 yr old) with
adult-onset GHD demonstrated normal BMD in the hip and the lumbar spine
(291). In summary, most studies demonstrate that patients 55 yr of age
or less with adult-onset GHD have decreased bone mass.
c. Fracture rate in GHD patients.
Few studies have
investigated whether or not GHD patients have an increased fracture
rate. The reasons for this are probably that a huge number of GHD
patients are required for a meaningful study and/or that additional
pituitary hormone deficits may confound the results. However, an
increased risk of osteoporotic vertebral fractures has been suggested
in hypopituitary patients (283). The consequences of low BMD in GHD
adults have only recently been delineated by Rosén et
al. (292) who found a higher fracture rate in patients with
adult-onset GHD compared with that in healthy controls. The fracture
rate was studied in 107 patients with adult-onset GHD, and a subsample
of the Göteborg WHO MONICA Study was used as a reference
population. The total fracture frequency was 2- to 3 times higher in
the patients compared with the controls. Confounding factors such as
longstanding untreated hypogonadism might have contributed to the low
bone mass in some subjects, since most of the studied patients also had
other pituitary deficiencies. On the other hand, Holmes et
al. (287) found a similar reduction in bone mass in patients with
isolated GHD and in those with multiple pituitary deficiency. Since
peak bone mass may not be reached until the third or fourth decade of
life (293), failure of accretion of bone mass may also be partly
responsible for the reduced BMD in adult-onset GHD. Again, since
patients who acquired their GHD after the age of 30 also have reduced
bone mass, it is likely that GH per se is important for the
maintenance of the adult bone mass (287).
3. Treatment of GHD patients with GH. GH treatment of GHD
adults has consistently been shown to have marked effects on markers
for bone formation [serum osteocalcin, serum levels of C-terminal
propeptide of PICP, and AP] and bone resorption [urinary
hydroxyproline, collagen cross-links, and serum concentrations of
collagen type I telopeptide (CITP)] and serum IGF-I levels (263, 266, 267, 268, 269, 270, 279, 289, 294, 295, 296, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306, 307, 308, 309). There is a dose-dependent increase of
bone markers during GH treatment (266, 309), and the increase in
resorption- and formation markers was maximal after 3 and 6 months,
respectively (300, 304). The similarity in time courses of the bone
markers supports the concept of a temporal coupling between bone
resorption and bone formation, with resorption preceding formation
during bone remodeling (300). After 2 yr of GH treatment these markers
were still elevated, suggesting that the increased rate of bone
remodeling was sustained (304). An increased bone turnover and an
increased cortical thickness, as studied by histomorphometric indices,
was found after GH treatment in a study of GHD men (310). Similarly,
increased bone turnover has been observed in patients with
long-standing acromegaly (251), indicating that bone turnover can be
elevated for many years as a result of high plasma levels of GH.
Furthermore, indices for bone formation remain elevated for several
weeks after short-term treatment with GH, suggesting that the half-life
for processes reflecting bone resorption/formation is quite long (302, 311). Administration of GH to healthy volunteers for 7 days produced no
discernible effect on serum calcium concentration, but urinary calcium
excretion increased. Serum PTH concentrations increased, as did the
concentrations of phosphate and 1,25-dihydroxyvitamin D (312). In
contrast, in other double blind, placebo-controlled trials (6 months
duration) of adult GHD patients, no effect on 1,25-dihydroxyvitamin D
concentrations (313) and no effect (313) or a decrease (296) in intact
PTH concentrations was found. These changes were accompanied by a
concomitant increase in total serum calcium concentrations (296, 313).
A similar increase in serum calcium concentrations has been observed by
others (297, 299, 300). Still, after 2 yr of GH treatment, serum
calcium concentration was elevated compared with baseline (304). GH may
enhance 1-hydroxylase activity (314), thus increasing the
concentration (315) or availability (316) of vitamin D3,
which is conceivably the mechanism behind the sustained increase in
serum calcium concentration. An alternative hypothesis is enhanced
mobilization of skeletal calcium due to increased bone turnover.
Trials involving adults with childhood onset GHD have yielded conflicting results regarding the effect of GH on bone mass. Several short-term placebo-controlled (306, 307, 308) and short-term open trials (268, 278, 294, 299) have failed to show any increase in BMD or BMC during GH treatment. In fact, some of these studies (299, 306, 307) reported a slight decrease of BMD or BMC after 3–6 months of treatment. In contrast, O‘Halloran (274) reported an increase in vertebral BMD assessed with QCT after 6 months of GH treatment but no changes in proximal or distal forearm BMC. After more prolonged treatment periods (12–30 months), several studies have disclosed more encouraging results (274, 305, 307, 317). Degerblad et al. (317) showed an increase in distal and proximal forearm BMD in six patients by 12 and 3.8%, respectively, after 24 months of treatment. Similarly, Vandeweghe et al. (307) reported a significant and progressive increase in BMC above pretreatment values, reaching 7.8% for total BMC at the lumbar spine and 9.9% for total BMC at the forearm, after 30 months of GH administration.
Short-term trials of 6–18 months in adults with adult onset GHD (263, 289, 297, 298, 300) failed to show any increase in BMC or BMD. In analogy with the findings observed in adults, with childhood onset GHD, several studies have shown a decrease in BMD and or BMC after 6–12 months of treatment. Holmes et al. (263) observed a decrease in BMD after 6 months of treatment at several skeletal sites. After 12 months of treatment, however, there was only a significant reduction in lumbar spine BMD. Similarly, Degerblad et al. (289) showed a decrease in total body and lumbar spine BMD after 6 months of GH treatment, but after 12 months of GH treatment there were no differences compared with baseline values. Furthermore, Hansen et al. (300) showed that in a placebo-controlled trial of 12 months, a decline occurred in forearm BMC and BMD by 4.2 and 3.5%, respectively. In contrast, in the longest placebo-controlled trial reported so far (18 months), Baum et al. (295) reported a significant increase in BMD in lumbar spine and femoral neck of 5.1 and 2.4%, respectively, using a daily dose of GH of only 4 µg/kg. Surprisingly, BMD increased at sites mostly composed of trabecular bone but not at sites composed of cortical bone. Since bone absorptiometry only detects the mineralized component of the bone, the reduction in BMD observed after short periods of GH treatment is best explained by the increased remodeling activity, with an increased remodeling space and an increased proportion of new unmineralized bone. Interestingly, the addition of a bisphosphonate to GH therapy in GHD adults reduced the GH-induced bone turnover and prevented the initial decrease in bone mineral content seen during GH treatment alone (318). Therefore, bisphosphonates might perhaps be an important adjunct to GH replacement therapy in adults with GHD and severe osteopenia during the early phase of GH treatment. However, if bone resorption is a prerequisite for bone formation, it is possible that an initial GH-induced bone resorption is crucial for the following GH-promoted bone formation.
Johannsson et al. (304) recently demonstrated that 2 yr of
GH treatment in 24 men and 20 women with adult-onset GHD induced a
sustained increase in overall bone remodeling activity and a net gain
in BMD in several weight-bearing skeletal locations (Fig. 7). A significant increase in BMD first
became apparent after 18 months, which might explain why previous
trials of shorter duration were unable to demonstrate an increase in
BMD. The study also demonstrates the importance of an adequate duration
of treatment to include a sufficient number of remodeling cycles and
sufficient time for mineralization to occur before a net gain in BMD
can be detected with bone absorptiometry. After 2 yr of GH treatment,
the total body BMC increased but not the total body BMD. Furthermore,
the increment in BMC was slightly more marked than the increment in BMD
at the different skeletal loci, suggesting that there was an increase
in the bone area. A similar increase in bone area after GH treatment of
rats has been described in detail in Section IV.
Interestingly, patients with a z-score of less than -1 SD
demonstrated the most pronounced increase in BMD (Fig. 7), reducing the
calculated number of patients with the greatest fracture risk by
40–50%, dependent on skeletal loci. However, this calculation is
based on the assumption that the changes in quality of bone in GHD
adults are similar to what was earlier described in postmenopausal
women. Furthermore, it should be emphasized that half of the patients
in the study by Johannsson et al. (304) were given a
supraphysiological GH dosage, which resulted in abnormally elevated
IGF-I levels (304). This study suggests that the remodeling balance
during GH treatment in GHD adults is positive, particularly in those
with a low pretreatment BMD. This is supported by a study demonstrating
a continuous increment in forearm cortical BMC 13 months after the
discontinuation of GH treatment (319).
|
B. Effects of the GH/IGF-I axis on bone metabolism and bone mass in
patients with normal GH secretion
1. Osteoporosis. The causes of osteoporosis are complex and
multifactorial. Bone mass decreases with aging, but the mechanisms
behind this decrease are unclear. Aging is associated with a decrease
in GH secretion (320, 321) and serum IGF-I concentration (322). The
GH/IGF-I axis is also influenced by lifestyle factors. For example,
smoking decreases IGF-I while physical activity increases GH secretion
(322). It has been suggested that the GH/IGF-I axis is one of the major
determinants of adult bone mass (323, 324). Thus, a positive
relationship between BMD and serum concentrations of IGF-I and IGFBP-3
w
