Perspective: PTH/PTHrP Activity and the Programming of Skeletal Development In Utero

Authors


  • The authors have no conflict of interest.

INTRODUCTION

EPIDEMIOLOGICAL STUDIES SUGGEST that men and women who were undernourished during intrauterine life, and therefore had low birthweight or poor growth in infancy, are at increased risk of developing several chronic diseases, including coronary artery disease, hypertension, non-insulin-dependent diabetes, and hypercholesterolemia.(1) A possible explanation for these associations is the phenomenon known as programming; this term describes persisting changes in structure and function caused by environmental stimuli acting at critical periods in early development.

There is increasing evidence to suggest that fetal programming also contributes to the risk of developing osteoporosis. For example, cohort studies from the United Kingdom, North America, and Australia have shown that poor growth during fetal life and infancy are associated with decreased bone mass in adulthood.(2–6) Recently, studies of a unique Finnish cohort have demonstrated that a reduced trajectory of linear growth during intrauterine and early postnatal life is directly linked with an increased risk of hip fracture six to seven decades later.(7) This relationship between adult bone mass and growth in early life seems to be independent of known environmental influences on bone loss during later life such as calcium, nutrition, and physical activity. Although shared genetic factors are likely to contribute to this association, a causal relationship between nutritional deficiency in fetal life and subsequent bone development may also be involved. Consistent with this suggestion, protein restriction in maternal rats has been reported to reduce bone area and bone mineral content (BMC) in the offspring, which also show altered growth plate morphology and a delay in the formation of osteogenic precursors within the bone marrow compartment.(8, 9)

As for the mechanisms responsible for programming, an adaptive response to fetal nutritional deficiency is thought to be involved, characterized by a reduction in the rate of cell proliferation secondary to altered concentrations of growth factors and hormones.(1) Consistent with this view, modification of the set-point of the growth hormone (GH)-insulin-like growth factor (IGF)1 axis has been suggested to play a role in the programming of bone development by early life factors.(10, 11) This axis is known to regulate cell proliferation and growth in a wide range of organs including bone. Whether fetal nutritional deficiency also affects tissue-specific processes involved in controlling the development of particular organs is currently unclear. To the extent that organ-specific mechanisms are also involved in mediating effects of programming on the skeleton, parathyroid hormone (PTH) and PTH-related peptide (PTHrP), which play a central role in calcium homeostasis in fetal as well as adult life, represent likely candidates.

PTH is released from the parathyroid glands in response to hypocalcemia as detected by the calcium sensing receptor (CaSR) and acts on several target organs including bone to increase extracellular fluid (ECF) calcium concentrations. In serving this endocrine function, PTH uses bone tissue as the calcium reservoir, which contains approximately 1–2 kg of calcium in adult humans compared with 1-2 g present in the ECF. In utero, calcium homeostasis is regulated not only by PTH, but also by PTHrP, which is present at relatively high concentrations in the fetal circulation. These hormones act to ensure that sufficient calcium is available to meet the needs of the developing fetal skeleton, which typically accumulates 21 g of calcium by term, of which 80% is gained in the third trimester.(12) PTH and PTHrP together exert a diverse range of cellular effects, which are now known to include major influences on skeletal structure and development in utero, making this system a prime candidate for a role in skeletal programming.

In this perspective article, we propose that PTH/PTHrP activity represents an important mechanism involved in the programming of bone development by early life factors, based on evidence that (1) PTH/PTHrP activity plays a major role in fetal calcium homeostasis; (2) PTH/PTHrP activity is an important regulator of skeletal development in utero; and (3) fetal calcium homeostasis is related to bone development in childhood. Although it is recognized that these findings only amount to relatively indirect evidence, taken together, they provide sufficient justification for future studies aimed at investigating the role of PTH/PTHrP activity in the programming of bone development in more detail.

PTH/PTHrP ACTIVITY AND FETAL CALCIUM HOMEOSTASIS

Several lines of evidence indicate that PTH/PTHrP activity plays a central role in calcium homeostasis in utero. Observations that parathyroidectomy induces fetal hypocalcemia in a range of mammals implies that PTH is required for maintaining serum calcium levels in the fetus, in which the parathyroid glands are the principle source of circulating PTH but not PTHrP.(12) PTH immunoreactivity is detected in human fetal parathyroid glands from as early as 10 weeks of gestation,(13) and PTH levels are approximately 50% that of maternal levels in human fetuses at mid-gestation.(14) By the third trimester, when the majority of skeletal mineralization occurs, fetal PTH secretion seems to play an active role in fetal calcium homeostasis, in light of a case report of increased fetal PTH secretion after maternal hypocalcemia caused by removal of a maternal parathyroid adenoma in the third trimester.(15)

Therefore, regulation of calcium exchange by PTH is likely to contribute to maintenance of serum calcium levels in the human fetus as in the adult, at least in the later stages of pregnancy. However, fetal PTH levels fall significantly by the end of gestation,(12) which is thought to be secondary to placental active transport of calcium. The latter results in relatively high levels of fetal serum calcium concentration being achieved, which is generally maintained significantly above that of the mother by a process which is largely independent of PTH (see paragraph below). Hence, while regulation of calcium exchange by PTH at organs such as the skeleton and kidneys contributes to maintenance of serum calcium levels in the fetus, independently regulated calcium exchange at the placenta seems to be the predominant determinant of fetal calcium concentration by the end of pregnancy.

Although PTH has been reported to exert weak independent effects on placental calcium transport, several studies suggest that the latter is largely under the control of PTHrP.(12) For example, in studies of thyroparathyroidectomized fetal sheep, full length and mid-region PTHrP fragments were found to stimulate placental calcium transport in perfused placentas, whereas PTH was without effect.(16, 17) Because PTHrP is synthesized by the placenta, stimulation of placental calcium transport by this protein largely represents paracrine activity, as also described for regulation of longitudinal bone growth by PTHrP (see next section). However, PTHrP can also be detected in the fetal circulation at significant concentrations, and so an endocrine action may also be involved. As well as the trophoblast cells of the placenta, circulating fetal PTHrP is thought to arise from production at numerous sites including the fetal parathyroid glands, skeletal growth plate, liver, amnion, chorion, and umbilical cord.(12)

The suggestion that, whereas PTH is required for maintaining fetal serum calcium levels, PTHrP is the major regulator of placental calcium transport is supported by a recent study of the relative roles of PTH and PTHrP in fetal calcium regulation using Hoxa3−/− knockout mice that lack parathyroid glands.(18) Serum PTH levels were found to be undetectable in these animals, whereas PTHrP levels were considerably elevated as a consequence of increased placental and hepatic synthesis, suggesting that circulating fetal PTHrP largely originates from the placenta and liver rather than the parathyroids. Associated with these changes, fetal mice developed significant hypocalcemia, but placental calcium transport was maintained.

Fetal calcium levels and placental calcium transport seem to be preserved in the face of significant calcium stress in a range of mammals. For example, no adverse effect on fetal-placental calcium transfer was observed in sheep with maternal hypocalcemia secondary to parathyroidectomy or dietary calcium restriction.(12) Presumably, compensatory increases in fetal PTH and/or PTHrP secretion are responsible for preserving fetal calcium exchange under these conditions (Fig.1). Maternal calcium deficiency is likely to stimulate fetal PTH production, based on observations of fetal parathyroid gland enlargement in response to maternal hypocalcemia in the rat and intrauterine hyperparathyroidism in children of mothers with hypoparathyroidism.(12) Increased placental PTHrP production may also contribute to the maintenance of fetal calcium transport and serum calcium levels in response to maternal calcium stress. Consistent with this suggestion, the CaSR has recently been identified within the human placenta,(19) although evidence that this regulates placental PTHrP production in response to ECF levels of calcium in a similar manner to PTH synthesis in the parathyroid glands is lacking.

Figure FIG. 1..

PTH signaling pathways and fetal calcium homeostasis. The major sites involved in fetal calcium homeostasis are shown. The fetus is thought to respond to a fall in maternal ECF calcium concentration by increasing PTH and PTHrP production, thereby preserving placental calcium transport and maintaining fetal ECF calcium levels above those in the mother. Other sites involved in fetal calcium homeostasis include organs contributing to amniotic fluid production such as the fetal kidneys.

Many of the cellular effects of PTH and PTHrP are mediated by the G-protein-coupled cell membrane receptor, PTH/PTHrP receptor (PPR), which is activated by the N-terminal region of these two hormones. However, mid-region PTHrP fragments fail to bind to this despite their activity in stimulating placental calcium transport, suggesting that PTHrP stimulates placental calcium transport through a distinct and as yet uncloned receptor. Consistent with this possibility, PPR−/− mice have been found to be hypocalcemic but to have elevated levels of PTHrP and placental calcium transport.(20) Although a second PTH receptor has subsequently been cloned, PTH2, this is activated by PTH but not PTHrP, including mid-region fragments, the functional significance of which is currently unclear.(21) In addition, the C-terminal portion of PTHrP has been suggested to exert distinct actions on bone cell activity through a further receptor.(22)

PTH/PTHrP ACTIVITY AND BONE DEVELOPMENT IN UTERO

PTH/PTHrP activity is known to play an important role in bone growth and development in utero. For example, PPR is an important regulator of growth in early life as a consequence of its actions on growth plate chondrocytes; PPR−/− mice show premature differentiation of hypertrophic chondrocytes and a marked reduction in growth plate width, suggesting that PPR acts to delay chondrocyte differentiation.(23) Trabecular bone is formed by endochondral bone formation, which is linked to longitudinal bone growth, and the abnormality in growth plate function reported in PPR−/− mice is associated with impaired formation of primary and secondary spongiosa.(24) In contrast, mice that overexpress PPR show evidence of delayed chondrocyte differentiation and enhanced trabecular bone formation.(25)

As for the physiological ligands responsible for regulating these actions of PPR, PTHrP−/− mice show a similar growth plate morphology to that observed in PPR−/− animals,(23) suggesting that PTHrP acts to inhibit chondrocyte differentiation through PPR. PTHrP is synthesized by growth plate chondrocytes at the end of long bones, where it acts in a paracrine manner to stimulate chondrocyte proliferation and delay chondrocyte maturation; a negative feedback loop also exists, whereby PTHrP suppresses secretion of Indian hedgehog, which is the major stimulus for the local secretion of PTHrP within the growth plate.(26) Because the growth plate is a relatively avascular tissue, it is likely that this action of PTHrP reflects a paracrine mechanism involving locally produced PTHrP as opposed to systemic PTH or PTHrP. Consistent with this suggestion, animals with deficiency of PTH, which signals through PPR in a similar manner to PTHrP, have a relatively normal growth plate morphology as assessed by analysis of PTH−/− mice.(27) On the other hand, trabecular bone was found to be deficient in PTH−/− mice as well as PPR−/− animals but unaffected in PTHrP−/− mice.(27) These observations suggest that PTH is the major ligand for the stimulation of trabecular bone formation by PPR during skeletal development and may represent an important endocrine mechanism by which fetal skeletal development is regulated by mechanisms involved in calcium homeostasis.

As well as promoting endochondral and trabecular bone formation during fetal skeletal development, PPR has also been reported to inhibit cortical bone development. For example, cortical bone formation seems to be increased in PPR−/− mice(24) and reduced in animals with transgenic overexpression of PPR in osteoblasts.(25) In addition, increased cortical thickness was observed in PTH−/− and PTHrP−/− mice,(27) suggesting that PTH and PTHrP both act as physiological ligands as far as effects of PPR on cortical bone are concerned. A possible explanation for this reciprocal influence of PPR on cortical and trabecular bone is provided by results of studies of Gsα−/− chimaeras, in which Gsα-deficient osteoblasts were found to populate the growth plate and cortical bone, but were not identified within trabecular bone.(28) Because Gsα is a major downstream mediator of the PPR receptor, these findings suggest that PPR influences the entry of osteoblasts into the trabecular compartment and/or their subsequent proliferation or survival at this site and therefore helps to regulate the balance between formation of the cortical and trabecular bone compartments (Fig. 2).

Figure FIG. 2..

PTH/PTHrP signaling pathways and fetal bone development. PTHrP is produced locally by the growth plate (GP), where it acts to regulate chondrocyte differentiation. In addition, PTH and PTHrP affect skeletal development by increasing trabecular bone formation (top box) and suppressing cortical bone formation (bottom box). Because trabecular bone provides a more readily available source of calcium than cortical bone, this action is likely to reflect an adaptive response to calcium stress.

Taken together, these studies show that PTH/PTHrP activity plays an important role in regulating skeletal development in utero, acting to increase trabecular bone mass with a reciprocal reduction in size of the cortical bone envelope. In light of its higher surface to volume ratio compared with cortical bone, trabecular bone is better adapted to provide a calcium reservoir to exchange with the ECF compartment. Therefore, this action of PTH/PTHrP activity may represent an attempt to adapt the skeleton for postnatal life associated with significant calcium stress. Hence, as well as controlling ECF calcium levels, PTH/PTHrP activity may act to regulate skeletal structure to ensure that this is adapted to serve the needs of calcium mineral homeostasis. This conclusion is consistent with the known effects of PTH in increasing trabecular bone mass in the adult.(29, 30)

In light of its role in regulating growth plate function in utero, PTHrP activity may also affect skeletal size in utero by influencing the rate of longitudinal bone growth. However, previous studies show that deficiency or overactivity of PTHrP both result in significant inhibition of longitudinal growth. For example, PTHrP−/− mice have reduced longitudinal growth,(27) whereas activating mutations of PTHrP have been reported to underlie the short-limbed dwarfism seen in Jansen's metaphyseal chondrodysplasia.(31)

The studies described above have examined the effects of major perturbations in PTH/PTHrP activity on skeletal development, and it is currently unknown whether lesser changes, such as those that occur in response to maternal calcium stress, are associated with equivalent effects. However, genetic studies suggest that smaller changes in the function of these genes associated with single nucleotide polymorphisms (SNPs) also result in detectable effects on bone development. For example, SNPs located within the promoter of the PPR gene, which have been found to influence its rate of transcription, have been reported to affect height in young adults.(32)

CALCIUM HOMEOSTASIS AND PROGRAMMING OF BONE DEVELOPMENT

The experimental data discussed above show that PTH/PTHrP activity influences trabecular and cortical bone development in early life and raise the possibility that altered fetal secretion of PTH and PTHrP in response to maternal calcium stress regulates skeletal development in utero. Because trabecular bone is remodeled throughout life, any tendency for trabecular bone mass to be elevated in babies born to mothers under calcium stress may be relatively transitory. Conversely, any reduction in the size of the cortical bone envelope at birth might be expected to affect the trajectory of bone growth in subsequent childhood, and therefore represents a plausible mechanism by which early life exposures influence bone development.

Several studies support the possibility that maternal and fetal nutritional deficiency adversely affect early skeletal development. For example, in a cohort of normal term deliveries, neonatal bone mass as measured by DXA was related to lifestyle characteristics of the mother in pregnancy and a range of anthropometric factors.(33) After adjustment for sex and gestational age, neonatal BMC was found to be strongly, positively associated with birthweight, birth length, and placental weight. Other positive determinants of neonatal BMC included maternal and paternal birthweight and maternal triceps skinfold thickness at 28 weeks (a measure of maternal fat stores). In contrast, maternal smoking and excessive levels of maternal physical activity were negatively associated with neonatal BMC at both the spine and whole body.

These findings suggest that factors related to maternal and fetal nutrition are associated with skeletal development in utero. Whether these observations reflect an effect of nutritional deficiency in early life on the balance between trabecular and cortical bone formation is unclear, because DXA measurements do not accurately distinguish between these two compartments. However, the associations described above between neonatal BMC and maternal nutrition and physical activity persisted after adjustment for overall skeletal size based on linear regression between BMC and skeletal area, which may provide a more accurate reflection of volumetric bone density compared with areal bone mineral density (BMD). Because trabecular bone contributes to volumetric bone density, these findings are consistent with the possibility that maternal nutrition and physical activity exert a positive effect on fetal trabecular bone development. On the other hand, significant associations with cigarette smoking and parental birthweight were no longer observed when BMC was adjusted for overall skeletal size in this way, suggesting that these variables primarily affect size of the skeletal envelope, to which trabecular bone makes no contribution.

There is limited evidence to suggest that maternal calcium stress contributes to this apparent relationship between early life nutrition and bone development in utero. For example, calcium supplementation was reported to increase neonatal BMD in undernourished Indian children(34) and to increase total body BMC in neonates born to U.S. women with low habitual calcium intake.(35) In a preliminary analysis, we examined whether calcium intake in pregnancy influences bone development in subsequent childhood in the Avon Longitudinal Study of Parents and Children (ALSPAC), which is a birth cohort initially comprising approximately 14,000 children.(36) Interestingly, we found that maternal calcium intake, as assessed by a food frequency questionnaire in the third trimester, was positively associated with whole body BMC of the child as measured at 9 years of age by DXA in 6500 individuals, after adjustment for total energy intake, maternal education, paternal social class, and housing (Tobias JH, Steef G, Ennett P, and Ness A, unpublished observations, 2003).

Because fetal calcium homeostasis is also likely to be influenced by maternal vitamin D status, observations that seasonal variations in maternal vitamin D status are related to neonatal BMC provided further support for a link between fetal calcium homeostasis and bone development.(37) Maternal vitamin D status also seems to exert a persistent effect on bone development in childhood, because maternal serum levels of 25(OH) vitamin D in late pregnancy were found to be significantly related to whole body BMC of the child as measured at 9 years of age in 211 individuals followed from birth.(38) In this study, maternal use of vitamin supplements and sunshine exposure were the major determinants of maternal vitamin D status.

Therefore, factors that are likely to influence calcium homeostasis in utero seem to be associated with subsequent bone development in childhood. Although it is unclear whether this relationship can be attributed to fetal PTH/PTHrP activity, recent evidence supports the suggestion of a direct link between maternal nutrition, fetal calcium homeostasis, and skeletal development. In the study of 211 children followed from birth described above, significant positive associations were also observed between umbilical vein calcium concentration (corrected for albumin) and whole body BMC at the age of 9 years.(39) Interestingly, although maternal smoking, nutrition, and physical activity were also found to be related to childhood BMC, these associations were dependent on umbilical venous calcium concentration, consistent with the hypothesis that maternal nutrition influences skeletal development of the child through an effect on calcium homeostasis. On the other hand, the relationship between maternal serum 25(OH) vitamin D and whole body BMC at 9 years of age was in part independent of umbilical venous calcium levels, suggesting that maternal vitamin D status also influences bone development of the child through mechanisms that do not involve calcium homeostasis.

Large birth cohorts such as ALSPAC, in which umbilical cord blood samples are also available, provide an important opportunity to explore the relationship between fetal calcium homeostasis and skeletal development in childhood, to investigate whether this reflects associated changes in PTH/PTHrP activity in utero, and to identify environmental exposures in the mother that underlie this. Based on the limited data that are presently available, these maternal exposures are likely to include factors directly related to calcium homeostasis such as calcium intake and determinants of vitamin D status like sunlight exposure and use of vitamin D supplements. In addition, other influences on maternal nutrition are likely to play a role, such as smoking and physical activity. An important justification for these studies is that, having identified such factors, it should be possible to develop relatively simple and cost-effective population-based interventions aimed at optimizing skeletal development in utero, thereby reducing the risk of osteoporotic fracture in later life.

SUMMARY AND CONCLUSIONS

There is increasing evidence that nutritional deficiency in utero adversely affects bone development and the risk of developing osteoporosis in later life. Although the mechanisms involved are unknown, circumstantial evidence points to an important role of PTH/PTHrP activity. It is recognized that PTH and PTHrP are critically involved in regulating fetal calcium homeostasis, actions that are mediated at least in part by PPR. As well as playing a central role in the maintenance of calcium homeostasis in the fetus, studies in transgenic mice show that PTH, PTHrP, and PPR exert similar effects on skeletal development in utero, acting to increase the size of the trabecular envelope and decrease that of the cortical envelopes.

Taken together, these observations raise the possibility that stimulation of PTH/PTHrP activity in the fetus in response to calcium deficiency acts to increase the size of the trabecular envelope but to reduce that of the cortical envelope. Although any increase in trabecular bone at birth is likely to be relatively transient, a decrease in size of the cortical envelope may have a persistent effect on the trajectory of bone growth in subsequent childhood. Consistent with this proposal, preliminary findings from birth cohort studies suggest that maternal calcium intake and cord blood calcium levels are positively related to bone mass of the offspring as assessed later in childhood. Further studies are justified to determine whether alterations in fetal PTH/PTHrP activity caused by calcium stress lead to a reduction in size of the cortical envelope at birth that persists into childhood and later adult life and to identify modifiable maternal factors that are responsible for these changes.

Acknowledgements

We thank Wellcome Trust for organizing the meeting (“Why do peoples' bones break? The structural, mechanical and cellular basis of the gain and loss of bone strength during growth and ageing,” Dormy House, Worcestershire, UK, June 28 to July 1, 2003), where discussions were held that formed the basis of this publication, and H Kronenberg for kindly reviewing a draft of the manuscript.

Ancillary