Contents - Previous - Next

Hormonal regulation of longitudinal bone growth

1. Introduction
2. The cellular organization of the epiphyseal growth-plate
3. The effects of hormones and growth factors
4. The effect of nutrition on longitudinal bone growth
5. Summary
Discussion of papers by Price et al. and Nilsson et al.

A. Nilsson 1, C. Ohlsson 3, O.G.P. Isaksson 2, A. Lindahl 4 and J. Isgaard 2

1 Department of Orthopedics (Hand Surgery)
2 Department of Internal Medicine
3 Department of Physiology, University of Göteborg, S-413 45 Göteborg, Sweden
4 Department of Clinical Chemistry, Sahlgren's Hospital

Correspondence to: A. Nilsson.

The regulation of postnatal somatic growth is complex. Genetic, nutritional factors and hormones exert regulatory functions. Hormones that have an established role in the regulation include growth hormone (GH), thyroid hormone and sex steroids. GH promotes mainly the growth of the long bones in terms of final height, while the action of the sex steroids and thyroid hormone is less well known. Longitudinal bone growth is the result of chondrocyte proliferation and subsequent endochondral ossification in the epiphyseal growth-plates. The growth-plate is a cartilaginous template that is located between the epiphysis and the metaphysis of the long bones. GH and insulin-like growth factor-I (IGF-I) have different target cells in the epiphyseal growth-plate. GH stimulates the slowly dividing prechondrocytes in the germinative cell layer while IGF-I promotes the clonal expansion in the proliferative cell layer of a GH primed cell. Thyroid hormone blocks the clonal expansion and stimulates chondrocyte maturation. IGF-I mRNA is primarily regulated by GH, and IGF-I is produced in several tissues such as the liver, muscle, fat and epiphyseal growth plates. However, IGF-I mRNA is also increased during compensatory growth of heart and kidneys and by estrogen in the Fallopian tube in the rat. Nutrition, i.e. energy from fat and carbohydrates and proteins, also influences the final height, but the cellular mechanism of action is not known. The aim of this article is to review hormonal action on longitudinal bone growth.

1. Introduction

Longitudinal bone growth is the result of chondrocyte proliferation and subsequent endochondral ossification in the epiphyseal growth-plates. The growth-plate is a cartilaginous template that is located between the epiphysis and the metaphysis of the long bones. The plate is formed after birth, when the epiphysis is developed as a bone nucleus. During subsequent growth the growth-plate and the articular cartilage are separated by accumulated bone. The process of longitudinal bone growth is complex and regulated by several factors, such as nutritional, neuronal and hormonal mechanisms which are all necessary for optimal bone growth. In the present review, recent findings regarding regulation of growth of the long bones will be discussed.

2. The cellular organization of the epiphyseal growth-plate

The chondrocytes in the growth-plate are strictly organized in a fascicular pattern according to their stage of maturation. The germinal cell layer is considered to contain the progenitor cells which divide less frequently (Kember, 1960). The progenitor cells are surrounded by matrix which consists of randomly orientated collagen type II fibrils and proteoglycans. The type II collagen acts as a barrier towards the bony epiphysis by inhibiting calcification (Robertson, 1990). In the proliferative cell layer, most of the cell replication takes place, (Kember, 1971) and the chondrocytes are organized in cell columns parallel to the longitudinal bone axis. The ratio between cells and matrix is higher than in the germinative layer, and the type II collagen fibrils are oriented longitudinally (Robertson, 1990). In the zone of maturation, the chondrocytes stop dividing and the cytoplasmatic cell volume increases. Production of collagen type X dominates (Leboy et al., 1989; Oshima et al., 1989). Alkaline phosphatase activity is found in the plasma membranes of the hypertrophic chondrocytes as well as in the matrix vesicles between the cells. Mineral deposition occurs, initially in the matrix vesicles, then in the matrix, and the zone of mineralization is formed with calcium phosphate which subsequently is replaced with crystalline hydroxyapatite. Endothelial cells are recruited from the metaphyseal vessels, proliferate and fill the space between the solid matrix scaffold (Robertson, 1990). Osteoblasts are recruited from mesenchymal stem cells and produce a new matrix named osteoid, rich in proteoglycans and collagen type I. The surface of the formed bone is covered with a thin layer of nonmineralized bone and a layer of lining cells to protect the new bone from being engulfed by osteoclasts.

In the rat, one cell division in the germinative cell layer gives rise to approximately 30 hypertrophic chondrocytes, resulting in an accumulated bone growth of 0.9 mm (Kember 1969; 1978). Thus, 40 cell divisions in the germinative cell layer are required for the complete growth of the rat tibia.

3. The effects of hormones and growth factors

3.1. Growth hormone

Several hormones are needed for normal postnatal somatic bone growth but GH is the only hormone that stimulates longitudinal bone growth dose-dependently (Cheek & Hill, 1974). Besides its growth promoting effects, GH has important regulatory functions in protein, carbohydrate and lipid metabolism (Kostyo & Nutting, 1974). In adult male rats, GH is secreted in a pulsatile pattern, with large peaks at 3-4 h intervals. In females the pattern is more irregular, with a higher frequency and a lower amplitude (Edén, 1978; Jansson et al., 1984). A pulsatile GH secretion is more effective in promoting body growth in the rat than a continuous secretion (Jansson et al., 1982a,b; Clark et al., 1985). In humans, GH concentration is high in the newborn and decreases after birth. Plasma levels of GH then increase during the prepubertal period, and sex differences in the pattern of GH become apparent. Although not extensively studied, results suggest that GH secretion in humans decreases with age (Jansson et al., 1985). The physiological significance of GH in adults has only recently been studied, due to a limited supply of GH. Adults with hypopituitarism have so far received replacement therapy with thyroid hormones, adrenal and sex steroids, but not GH. Deficient adults are overweight (Rosén et al., 1993a), have reduced bone mineral content (Rosén et al., 1993b) and show lower rates of employment and marriage (Björk et al., 1989., Dean et al., 1985). Replacement therapy with GH decreases body fat mass, increases the extracellular fluid volume, total body nitrogen, muscle volume and physical strength as well as vigor, ambition, sense of well-being and quality of life (Bengtson et al., 1993).

The GH-receptor was cloned in 1987 (Leung et al., 1987) and is a member of a large family including receptors of prolactin, erythropoietin, interleukins-2, 4, 6 and granulocyte-macrophage colony-stimulating factor (Patthy, 1990). The human GH-receptor gene is located on chromosome 5 and contains at least 10 exons spanning over more than 85 kb (Barton et al., 1989; Godowski et al., 1989). Noncoding sequences are found in exon 1 and a signal peptide in exon 2. Exons 3-7 encode the extracellular domain, exon 8 the transmembrane and exons 9-10 the intracellular domains of the receptor (Godowski et al., 1989). It is of interest that the prolactin receptor gene does not share any sequence homology with the exon 3 of the GH-receptor, suggesting that exon 3 may be crucial for the specificity between somatogenic and lactogenic receptors.

In the rat most tissues, including the liver, adipose tissue, cartilage growth-plate, muscle, heart and skin express GH-receptor mRNA (Carlsson et al., 1991). Recently it has been shown that GH can increase GH-receptor mRNA levels both in vivo and in vitro (Vikman et al., 1991a; Nilsson et al., 1990a).

The cloned GH-receptor is of importance for GH signal transduction. This was recently shown by transfection experiments with the cloned GH-receptor, where GH stimulation results in both protein and insulin synthesis (Billestrup et al., 1990; Emtner et al., 1990).

Mutations in the GH-receptor gene may result in GH resistance. GH insensitivity syndrome, also known as Laron dwarfism, is an inherited disease which is characterized by growth failure, low serum IGF-I and increased GH secretion (Laron et al., 1971). Recently it was shown that one family with Laron dwarfism had a substitution of phenylalanine for serine at position 96 of the extracellular domain at a highly conserved region of the GH-receptor (Amselem et al., 1989; Godowski et al., 1989).

In the germinative cell layer of newborn rabbit epiphyseal growth-plates, no GH-receptor immunoreactivity was found. However, in 20- and 50-day-old rabbits GH-receptor staining was present (Barnard et al., 1988). This suggests that the GH-receptor is developmentally regulated. In vitro the messenger for the GH-receptor, GH-binding and GH effects have been demonstrated in rat epiphyseal chondrocytes. Maximal GH-receptor mRNA levels and binding are obtained if the epiphyseal chondrocytes are seeded at low density (Nilsson et al., 1989; 1990a). This finding is supported by results where GH stimulated thymidine uptake in confluent cells that had been pre-cultured for several days. No stimulatory effect was seen in rapidly proliferating preconfluent or in postconfluent cells under these conditions (Ohlsson et al., 1992a).

A circulating GH-binding protein (GHbp) has been identified in several species, including human, rabbit, rat and mouse. The GHbp is identical to the extracellular part of the GH-receptor (Leung et al., 1987). In rodents, a specific mRNA has been cloned, encoding the extracellular domain of the receptor (Baumbach et al., 1989; Smith et al., 1989). However, in humans and rabbits such a specific mRNA has not been found, and it has been suggested that the GHbp is produced by proteolytic cleavage of the extracellular domain of the GH-receptor (Trivedi & Daughaday, 1988). The physiological function of the GHbp is unclear. In vivo, it has been shown to prolong the half-life of GH (Baumann et al., 1987).

3.2. Insulin-like growth factor (IGF)

In 1957 Salomon & Daughaday demonstrated that GH stimulated sulphate incorporation into cartilage indirectly through a serum factor, initially termed sulphation factor, later designated somatomedin and shown to be identical to insulin-like growth factor I and II. This finding formed the basis of the somatomedin hypothesis, stating that GH stimulates the production of IGF-I in the liver, which subsequently stimulates longitudinal bone growth.

There are two different forms of IGF peptides: IGF-I (70aa) and IGF-II (67aa) with an amino acid sequence homology of 70% (Daughaday & Rotwein, 1989). In rats, IGF-II is the major IGF in fetal tissues, while IGF-I expression dominates in adult tissues (Norstedt et al., 1988). The highest IGF-I expression is found in the liver; however most tissues express IGF-I mRNA. GH is probably the most important regulator of the IGF-I mRNA levels (Mathews et al., 1986; Doglio et al., 1987; Norstedt & Möller, 1987; Isgaard et al., 1988; Vikman et al., 1991b).

IGF-I transcripts (Nilsson et al., 1990b), IGF-I immunofluorescence (Nilsson et al., 1986) and IGF-I binding (Tripper et al., 1986; Makower et al., 1989) are predominantly found in the proliferative cell layer of growth-plates. Moreover, IGF-I binding to bovine (Tripper et al., 1983), rabbit (Poster Vinay et al., 1983) and rat (Makower et al., 1989) growth-plate chondrocytes in monolayer culture has been shown.

At least six binding proteins (bp) have been identified which are thought to mediate IGF-I and II action (Shimasaki & Nicholas, 1991; Baxter, 1992). Possible functions are: (i) constituting a storage pool for IGFs in the circulation; (ii) prolonging the half-life of the IGFs; and (iii) binding to the extracellular matrix and thereby trapping IGFs from the circulation and targeting the growth factor to its receptor. The IGF-bp3 is the most abundant bp and the main transporter in the circulation. The IGF-bp3 is directly GH dependent, while IGF-bp1 and IGF-bp2 are inversely correlated with GH status. In GH insufficiency IGF-bp3 is decreased and IGF-bp1 and IGF-bp2 are increased (for review see Guyda, 1991). In cord serum from fetuses with intrauterine growth retardation, IGF-bp3 was markedly decreased, while IGF-bp2 and IGF-bp1 were increased (Crystal & Giudice, 1991). Nothing is known about the hormonal regulation of the IGF-bps in the growth-plate; three separate bps have been detected in medium from cultured fetal ovine growth-plates, but they have not been further characterized (Hill & Han, 1991).

3.3. Thyroid hormones

Thyroid hormones are important for optimal bone growth both in humans (Rochiccioli, 1985) and rats (Thorngren & Hansson, 1973). Three different possible mechanisms for stimulation of longitudinal bone growth have been demonstrated. In vivo, thyroid hormones have been shown to stimulate GH secretion (Hervas et al., 1975; Coiro et al., 1979). Moreover, thyroid hormones increased IGF-I mRNA in hypophysectomized rats (Wolf et al., 1989) and stimulated IGF-I production in perfused rat livers (Ikeda et al., 1989), suggesting that thyroid hormones increase the circulating levels of IGF-I. A direct effect of thyroid hormones on the growth-plate has also been shown. In hypophysectomized rats, thyroid hormones stimulate longitudinal bone growth (Thorngren & Hansson, 1973) and are required for the formation of hypertrophic cells in normal rats (Ray et al., 1954; Lewinson et al., 1989).

3.4. Sex steroids

In humans, androgens and estrogens are crucial for increased longitudinal bone growth during the pubertal growth spurt, and fusion of the epiphyseal plates. This is illustrated by precocious puberty, with an early pubertal growth spurt and short adult stature due to fused growth-plates (Bourguinon, 1988). The mechanism of action suggests that the sex steroids exert both an indirect and a direct effect on longitudinal bone growth. Sex steroids influence GH secretion in humans (Frantz & Raben, 1965; Illig & Prader, 1970) and rats (Jansson et al., 1985). Laron dwarfs that do not respond to GH because of a deficient GH-receptor, show a growth spurt at sexual maturation (Laron et al., 1980). In Turner's syndrome, GH therapy together with oxandrolone is more potent in stimulation of longitudinal bone growth than GH alone (Rudman et al., 1980; Rosenfeld et al., 1992). Receptors of both sex steroids have been shown on both cultured cartilage (Rosner et al., 1982; Carrascosa et al., 1990) and osteoblast-like cells (Eriksen et al., 1988). In vitro, the sex steroids stimulated proteoglycan synthesis in human epiphyseal chondrocytes and IGF-I production in human osteoblasts (Ernst et al., 1989).

3.5. Vitamin-D

The bioactive form of vitamin-D (vit-D) is 1,25-(OH)2D3 and has been shown to be important for the maturational process of epiphyseal chondrocytes. Vit-D receptors have been demonstrated in the epiphyseal growth-plate (Klaus et al., 1991), and vit-D stimulates both proliferation (Klaus et al., 1991) and maturation (Gerstenfeld et al., 1990; Schwarz et al., 1992) in cell culture. In humans, vit-D deficiency (rickets) is characterized by an accumulation of proliferating but poorly mineralized growth-plate cartilage (Klaus et al., 1991).

3.6. The dual effector theory

By studying the growth, differentiation and proliferation of cultured preadipose cells, Green and co-workers made important discoveries regarding the mechanism of action of GH (Morikawa et al., 1982; Green et al., 1985). Using 3T3-cells it was found that GH had a specific stimulatory effect on preadipocyte conversion to adipocytes. According to the dual effector theory of GH action, GH stimulates the differentiation of the progenitor cells and, as a result of differentiation, these cells become responsive to growth factors like IGF-I (Green et al., 1985). This theory of GH action has also been applied to tissues such as the epiphyseal growth-plate (Isaksson et al., 1987).

3. 7. GH versus IGF-I

Superphysiological longitudinal bone growth has only been shown in GH transgenic mice or excessive GH production in humans before the growth-plates are fused (gigantism). So far no IGF-I transgenic mice or human IGF-I disturbance has been shown to result in highly increased longitudinal bone growth.

GH receptors have been found in the germinative cell layer where the IGF-I protein or transcripts are not present (Barnard et al., 1988; Nilsson et al., 1986; Nilsson et al., 1990b). In vitro, it has been shown that the number of large sized colonies of epiphyseal chondrocytes in suspension culture were increased in the presence of GH and not in the presence of IGF-I (Lindahl et al., 1986; 1987a). Moreover, cells isolated from animals primed with GH before cell isolation, resulted in larger sized colonies than in saline primed animals, in subsequent culture in the presence of IGF-I (Lindahl et al., 1987b). Regarding thymidine incorporation, the stimulatory effect of GH was not apparent until 3 days of GH exposure, and GH but not IGF-I had a sustained effect on thymidine incorporation (Ohlsson et al., 1992a). Recently, it was shown that IGF-I bound to dense differentiated foci expressing type II collagen of epiphyseal chondrocytes, while GH bound to the less dense noncollagen producing cells (Bentham et al., 1993). Taken together, these results suggest that the target cells and thus the mechanism of action of the two peptides are different.

Until recently, the slowly cycling germinative cells of the tibial growth-plate have not been visualized, nor has it been demonstrated whether GH in vivo induces multiplication of these cells, as suggested in the dual effector theory of GH action. To determine the localization of the slowly dividing cells, tibiae were labeled by a continuous thymidine infusion in vivo between days 18 and 32 after birth, with a final 14-day period without isotope infusion, to dilute labeled cells with a high turnover rate. To study the effect of GH and IGF-I on the slowly cycling cells, GH or IGF-I was injected locally to hypophysectomized rats through an inserted cannula to one proximal tibia, and the other tibia served as a control under continuous systemic thymidine infusion. Then, systemic GH infusion was given for 14 days to dilute labeled cells. The animals were killed and the tibiae were prepared for autoradiographic examination. Results showed that tibiae of rats exposed to systemic GH infusion were only labeled in the germinative cell layer. Most label-retaining cells were found in the five most proximal cell positions, and GH but not IGF-I significantly increased the number of labeled cells in this layer (Ohlsson et al., 1992b).

Recent findings suggest that not only GH but also other hormones such as estrogen and mechanical load can induce IGF-I mRNA expression locally. For example, volume overload produced by a fistula between the aorta and the inferior vena cave in male rats resulted in significantly increased IGF-I mRNA levels in the right but not the left ventricle of the heart (Isgaard et al., 1992). Moreover, in female rats both IGF-I protein and mRNA were detected in the Fallopian tubes. The IGF-I mRNA followed the ovarian cycle, and an injection of estrogen increased the IGF-I mRNA levels dose-dependently (Carlsson et al., 1993). These findings show that IGF-I is produced in several peripheral tissues besides the liver and regulated by other factors besides GH.

3.8. IGF-I versus thyroid hormones

Alkaline phosphatase activity (ALP) is expressed during the late processes of differentiation, like hypertrophy, degeneration and calcification in the growth-plate. Recently it has been shown that neither GH nor IGF-I regulate these steps. However, triiodothyronine (T3) increased ALP activity and reduced DNA synthesis as measured with thymidine incorporation. Interestingly, it was shown that increasing concentrations of T3 dose-dependently reduced the thymidine incorporation stimulated by IGF-I. No interaction was found between T3 and GH (Ohlsson et al., 1992c). This may be interpreted that T3 blocks the epiphyseal cell clonal expansion stimulated by growth factors like IGF-I, and T3 stimulates chondrocyte cell maturation. An example showing that thyroid hormones in vivo may limit the clonal expansion induced by IGF-I is perhaps a child with hyperthyroidism. Such children grow more rapidly and end up shorter than expected if not treated (Prof. Otto Westphal, personal communication).

3.9. GH versus sex steroids

GH-deficient or-insensitive children are characterized by severe growth failure. The long bones are relatively short compared to the spine length and head size. A short stature as a result of short long bones is also found in precocious puberty, where the elevated levels of sex steroids (androgens in boys and estrogen in girls) result in increased growth velocity for the age. The children are taller than their peers during childhood, but enter puberty earlier and thus stop growing, ending up shorter than expected, due to an early fusion of the growth-plates (Westphal, 1991). Moreover, children with hypogonadism are tall with relatively long limbs compared to the spine and head size. In this syndrome the growth-plates of the long bones are open for GH action over a longer period of time, due to lack of sex hormones. These findings suggest that GH is important for the final height by its effect on the growth of the long bones. The relative importance of the spine and the long bones in relation to final height remains to be elucidated.

4. The effect of nutrition on longitudinal bone growth

Nutrition is important for final height, and nutritional disorders affect hormonal secretion. This may be exemplified by the retarded bone growth and the late onset of puberty of a child suffering from intolerance of gluten. Another example is the dramatic upward secular trend of adult height in Japan over the past century, where the mean adult height in males has increased almost 10 cm from 160.9 cm in 1900 to 170.5 in 1989 (Matsuo, 1991). The mechanism behind variations of longitudinal bone growth at a cellular level is not yet understood.

In the case of starvation, GH levels are affected. In the rat, GH secretion is suppressed (Tannenbaum et al., 1978), whereas in man starvation is associated with high GH levels (Ho et al., 1988). In children with anorexia nervosa, GH secretion and cortisol levels are increased while thyroid hormones, insulin and s-IGF-I levels are low. Moreover, androgen levels are higher than expected, due to a reduced amount of adipose tissue. An early and severe development of anorexia nervosa results in reduced final height that is never regained (Prof. Otto Westphal, personal communication).

5. Summary

GH and IGF-I have different target cells in the epiphyseal growth-plate. GH stimulates the slowly dividing prechondrocytes in the germinative cell layer while IGF-I promotes the clonal expansion of a GH-primed cell in the proliferative cell layer. Thyroid hormone blocks the IGF-I induced clonal expansion, probably in the late proliferative layer, and stimulates chondrocyte maturation.

Acknowledgements - We would like to express our gratitude to Prof. Otto Westphal for valuable discussion. This work was supported by grants from the Swedish Medical Research Council (14x-04250 and K90-13F-09248-01A); KabiPharmacia, Stockholm; The Göteborg Medical Society and the Faculty of Medicine, University of Göteborg.

Contents - Previous - Next