The Role of Parathyroid Hormone in the Pathogenesis of Glucocorticoid-Induced Osteoporosis: A Re-Examination of the Evidence

One of the most important of the secondary causes of osteoporosis is chronic exposure to glucocorticoids, which are used for an extraordinarily large number of disorders. The adverse effects of hypercortisolism on bone metabolism were recognized more than half a century ago (1). Today, glucocorticoid exposure in the context of medicinal use has become far more common than excess endogenous exposure (Cushing’s syndrome). Glucocorticoid-induced osteoporosis (GIO) is the third most common cause of osteoporosis, trailing only postmenopausal and age-related osteoporosis (2). As many as 50% of individuals on chronic glucocorticoid therapy will suffer an osteoporotic fracture (3). Recently, a large-scale retrospective cohort study by Van Staa et al. (4) in England clearly demonstrated that fracture risk is increased across virtually the entire dosage range of oral glucocorticoids. A large number of subjects with a history of glucocorticoid exposure (n = 244,235) were matched to the same number of control patients who had no history of glucocorticoid exposure. The average age of the subjects was 57 yr; respiratory diseases were the common indication for therapy, being prescribed in 40% of the patients (4). Referent to nonglucocorticoid users, subjects with a history of glucocorticoid therapy had significantly greater risk for fractures at the spine (rr = 2.6), the hip (rr = 1.6), and at any nonvertebral site (rr = 1.3). The magnitude of the fracture risk was directly related to dosage, with subjects receiving as little as 2.5 mg of prednisolone at significantly greater risk than control subjects (4). Bone loss from glucocorticoid use was also found to occur rapidly, within the first 3 months of treatment. A similarly precipitous loss of bone mass has also been observed prospectively when glucocorticoids are used in the setting of organ transplantation (5, 6), and in other clinical situations (7–10). Even inhaled steroids have been implicated as a cause of bone loss (11, 12). The cardinal feature of GIO on skeletal dynamics is a reduction in bone formation. Bone formation is inhibited, in part, through a decrease in osteoblast life span and function. Histomorphometric studies demonstrate a marked reduction in indices of bone formation, such as reduced mineral apposition rate and prolonged mineralization lag time. The amount of bone that is replaced in each remodeling cycle can be reduced by as much as 30% (13–17). Biochemical markers of bone formation, osteocalcin and bone-specific alkaline phosphatase, are suppressed. In addition to this primary suppressive effect on bone formation, glucocorticoids also induce an early phase of accelerated bone resorption (18). Osteoclast number and activity increase, along with an increase in the fraction of eroded bone surface (14, 15). Biochemical markers of bone resorption, urinary N-telopeptide and pyridinoline cross-link excretion, rise during early glucocorticoid exposure (17, 19, 20). This early phase of glucocorticoid use, therefore, can be associated with rapid bone loss due to both reduced bone formation and accelerated bone resorption. With continued use of glucocorticoids, the rapid rate of osteoclast-mediated bone resorption slows (14), but suppression of bone formation continues as the dominant skeletal dynamic. Thus, bone loss is progressive because bone resorption chronically exceeds bone formation. Although bone loss due to glucocorticoid use tends to be diffuse, the axial skeleton is targeted preferentially. The cancellous bones of the vertebral spine are typically affected, whereas cortical bone sites of the appendicular skeleton (i.e. forearm) are affected to a lesser extent (21, 22). Spontaneous fractures of the vertebrae or ribs are common complications of GIO (23). The pathogenesis of GIO is multifactorial (Fig. 1; Ref 24). Reduction of gonadal hormones is an important mechanism through the inhibitory effects of glucocorticoids on the pituitary gonadotropins (19). Glucocorticoids blunt the secretion of LH in response to GnRH in men and women (25, 26), inhibit the action of FSH, and reduce gonadal sex steroid production (27–31). In one study, asthmatic men treated with prednisone at an average daily dose of 12 mg had significantly lower free and total testosterone levels and higher LH and FSH levels than age-matched control subjects (29). Similarly, estrogen levels decline with glucocorticoid administration (30, 31). The classical pathophysiology of GIO. The shaded area is the focus of attention in this paper. However, the adverse effects of glucocorticoids on bone are not mediated exclusively by sex-steroid deficiency (29, 32, 33). Other proposed mechanisms for GIO include the effects of glucocorticoids on the expression of locally produced growth factors and related proteins (18). Prolonged exposure to high levels of glucocorticoids results in reduced production of IGF-I, a trophic factor for bone, as well as alterations in IGF-binding proteins in osteoblasts (34, 35). Hepatocyte growth factor, a polypeptide with a mitogenic effect on osteoblasts (36, 37), is also decreased with glucocorticoid administration (38, 39). Loss of muscle strength and reduced physical activity also probably contribute to the bone loss that occurs with glucocorticoids (40). Another proposed mechanism for the adverse skeletal effects of glucocorticoids is a direct effect on calcium metabolism. At daily doses of prednisone, 10 mg, intestinal calcium absorption is reduced (41, 42). Alterations in vitamin D metabolism or an independent noncompetitive effect to counter the actions of vitamin D (43) could account, in part, for reduced intestinal calcium absorption. When prednisone doses are increased to 20 mg/d or higher, a direct effect to increase renal calcium excretion adds to negative calcium balance (44, 45). Physiologically, if the serum calcium is reduced as a result of reduced calcium absorption and increased urinary calcium excretion, one would expect to find evidence for a secondary increase in PTH secretion. The state of secondary hyperparathyroidism would then be expected to lead to patterns of bone loss typified by excessive PTH secretion, namely cortical bone loss. Hahn et al. (46) compared 17 glucocorticoid-treated patients with normal subjects and found increases in PTH, along with decreases in 47Ca absorption and forearm bone mass. Similar findings were noted in other small studies when patients on glucocorticoid therapy were found to have increases in PTH (47, 48). In a larger study by Suzuki et al. (45), 44 glucocorticoid-treated patients were also found to have elevated PTH, nephrogenous cAMP, and fasting urinary calcium. Recently, elevated PTH levels were observed in infant piglets administered dexamethasone for 15 d (49). Thus, a classic model of GIO on bone invariably includes a compensatory increase in PTH due to these proposed effects of glucocorticoids on gastrointestinal and urinary calcium metabolism (19). Some studies have suggested that concomitant vitamin D deficiency or resistance could contribute to the reductions in calcium absorption or conceivably also lead more directly to PTH secretion (50). In some trials, when calcium or vitamin D was replaced, the rise in PTH levels was attenuated (46, 48, 51), but this has not been consistently found in other studies (52, 53). A possible mechanism by which glucocorticoids increase PTH levels is through a direct effect on the glandular secretion of PTH (54). This might occur via increased PTH gene transcription (55) and enhanced efficiency of postreceptor signaling (56, 57). Another mechanism by which glucocorticoids might affect the parathyroid-bone axis is by increasing the sensitivity of bone cells to PTH. Glucocorticoids have been shown to increase the expression (58) and availability (56) of PTH receptors on osteoblasts. Increased numbers of PTH receptors could be associated with enhanced sensitivity to PTH. Alternatively, enhanced sensitivity to PTH could be due to changes in the affinity of the receptor for PTH. In mouse cell cultures, the addition of a defined concentration of glucocorticoids has a synergistic effect on PTH-mediated bone resorption (59, 60). These models and the supporting evidence for them might be expected to have specific consequences. For example, PTH levels might be consistently elevated when glucocorticoids are used. Second, measurements of bone mass might reflect the known skeletal effects of excessive PTH on bone to erode cortical bone and to help maintain cancellous bone. Third, histomorphometric analysis of bone biopsies from individuals on glucocorticoid therapy might be expected to show evidence for the actions of PTH on bone to increase indices of bone remodeling. Evidence available in support of or against these expectations is reviewed here. As early as 1980, Seeman et al. (61) observed that excess glucocorticoid levels, whether endogenous or exogenous in origin, were not associated with alterations in PTH levels. Similarly, Slovik et al. (53) found normal PTH levels in asthmatic patients on short-term and long-term glucocorticoid treatment. In addition, 22 women with rheumatoid arthritis were found to have the same PTH levels whether or not they were receiving low dose prednisone (6.6 mg/d; Ref. 62). When women were administered even higher amounts of prednisone chronically (for an average of 13 yr), PTH levels were the same as in age-matched controls (63). Hattersley et al. (64) also found no difference in PTH levels in patients treated with chronic glucocorticoids for obstructive airway disease when compared with controls. The expectation that PTH levels should have been elevated in these studies, given increases in urinary calcium excretion and decreases in intestinal calcium absorption is not confirmed by the data. It is important to note, however, that the subjects in these trials had underlying medical illnesses, which could conceivably have modified both the secretion of and skeletal response to PTH. Other data indicating that PTH levels are not changed by glucocorticoid administration come from studies in normal, healthy subjects. In three separate investigations examining the effects of a short course (5–14 d) of prednisone (doses ranging from 15–40 mg daily) in healthy adults, no significant change in PTH levels was found (52, 64, 65). In a longer study of 9 healthy men who were administered 50 mg prednisolone daily for as long as 6 months, there was no change in PTH levels (Fig. 2; Ref. 66). These studies in normal subjects confirm those in specific medical illnesses and help to minimize any potential confounding effects of underlying illness on circulating PTH levels. However, as already noted, it remains possible that skeletal responsiveness to PTH is differentially affected by glucocorticoids without any change in PTH concentrations. Thus, even a normal PTH level might elicit significant physiological effects, such as an elevation in the level of cAMP (45, 48). Glucocorticoids do not increase PTH levels. Nine healthy men receiving 50 mg of prednisolone daily for infertility (due to the presence of antisperm antibodies) for 3.7 ± 0.6 months. PTH did not increase during corticosteroid treatment. The area above the dotted line represents the normal area for PTH. [Reprinted with permission from G. Pearce et al.: J Clin Endocrinol Metab 83:801–806, 1998 (66 ) ©The Endocrine Society.] Most recently, Manelli et al. (67) have studied PTH secretory dynamics in glucocorticoid-treated men, as compared with normal age- and sex-matched controls. Six men (ages, 31–64 yr) treated chronically with glucocorticoids (daily dose > 7.5 mg of prednisone or equivalent for more than 6 months) and control subjects underwent peripheral blood sampling every 3 min for 6 h. Basal PTH secretory rate was reduced in the glucocorticoid-treated group (4.3 vs. 8.8 pg/ml·min; P = 0.017) despite an increase in amplitude of fractional pulsatile PTH secretion (42 vs. 18 pg/ml·min; P = 0.006) as compared with controls. Chronic glucocorticoid treatment appeared to induce a redistribution of the spontaneous PTH secretory profile by reducing the amount tonically released while increasing the amount released by pulsatile secretion (67). Although the overall dominant effect was a reduction in PTH secretion, this study could not determine whether tonic or pulsatile secretion was the driving force. If such a distinction could have been made, the principal action of glucocorticoids on PTH, positive or negative, could have been elucidated. Thus, although the implication of the studies in which PTH levels are measured favor an inhibition by glucocorticoids, more direct evidence requires an examination of the skeleton per se. Densitometric findings further support the idea that PTH is not involved in the pathophysiological mechanisms of bone loss in GIO. The typical pattern of bone loss in GIO is a preferential reduction of lumbar spine and trochanteric bone mineral density (BMD), significantly greater than any reduction at the distal radius (Fig. 3A) (22, 68–71). Reid et al. (69), however, reported that the mean decrement in BMD in their steroid-treated patients was approximately 20% in the distal forearm, lumbar spine, and proximal femur as assessed by single-photon absorptiometry or dual-energy x-ray absorptiometry (DXA). The application of quantitative computed tomography (QCT) helped to distinguish cancellous elements in the lumbar spine from its cortical elements. Reid et al. (69) showed, using QCT, that the mean decrement in lumbar spine BMD was much greater, approximately 40%. These observations lend further support to the idea that glucocorticoids reduce cancellous bone density preferentially. Some of this cancellous reduction, however, could be due to expanded marrow space. A, Bone mineral content in prednisone-treated patients with rheumatoid arthritis, as compared with age- and sex-matched patients with rheumatoid arthritis never treated with glucocorticoids. [Adapted from R. F. Laan et al.: Calcif Tissue Int 52:5–9, 1993 (68 ).] B, Bone densitometry in primary hyperparathyroidism. Data are shown in comparison with age- and sex-matched normal subjects. Divergence from expected values is different at each site (P = 0.0001). [Adapted from S. J. Silverberg et al.: J Bone Miner Res 4:283–291, ).] hyperparathyroidism is an model for bone loss. The preferential action of PTH to be at cortical sites distal one third radius is typically in primary hyperparathyroidism In the lumbar spine, a site of cancellous bone, is well (Fig. Ref. Even in postmenopausal women with primary cancellous bone of the lumbar spine is well This pattern is in those with primary hyperparathyroidism. With more disease and prolonged PTH however, this pattern can be and skeletal loss at sites can of BMD in primary hyperparathyroidism from those observed in GIO the patterns of primary hyperparathyroidism and GIO be directly compared because each results from different mechanisms of skeletal loss. In primary bone is increased at both formation and resorption whereas in bone resorption is from bone formation which is suppressed. One can for preferential loss of cancellous bone in GIO due to the higher rate of cancellous bone compared with cortical bone. If PTH is bone dynamics in bone formation would not be expected because PTH increases bone formation It is that of cancellous bone in primary hyperparathyroidism is in to the pattern of bone loss in GIO in which cancellous bone loss is If compensatory hyperparathyroidism is an important in one might expect to evidence that it the actions of glucocorticoids on cancellous bone much in the that PTH the effects of estrogen deficiency on cancellous bone loss. although there is some evidence of cortical bone loss with GIO as there is evidence for some cancellous bone loss in primary it is the different patterns that are most of the and fracture risk of GIO and primary hyperparathyroidism findings in primary hyperparathyroidism. of the and fracture risk of GIO and primary hyperparathyroidism findings in primary hyperparathyroidism. Bone density fracture In of glucocorticoid the vertebral fracture is the reduction in BMD at this In of PTH vertebral fractures are most Although data on fracture is it would that the forearm fracture is much more to occur in primary hyperparathyroidism than the vertebral fracture However, it is important to that fracture risk is due to many in addition to bone PTH, and glucocorticoids. The of an fracture to specific such as PTH or glucocorticoids might be and direct histomorphometric analysis of bone biopsies also a for PTH in the bone loss associated with glucocorticoid In there is a decrease in of in and in cancellous bone (Fig. 15 and In primary the is namely cortical with of cancellous bone (Fig. effects of glucocorticoids on bone. of vertebral cancellous bone in a mouse receiving and a mouse receiving prednisone the marked reduction in vertebral cancellous bone area in the mouse vertebrae receiving [Reprinted with permission from R. S. et al.: J Clin ).] of biopsies of a with primary hyperparathyroidism and an age- and sex-matched control the of in the with primary hyperparathyroidism as well as the of cancellous bone and [Reprinted with permission from et al.: J Clin Endocrinol Metab ) The Endocrine Society.] is well in primary hyperparathyroidism whereas in GIO is typically with a reduction in In primary the expected decline in number and the increase in do not and their are more than would be expected through the hyperparathyroidism to the normal age-related associated with loss. In number is reduced and is findings different from the analysis in primary hyperparathyroidism. histomorphometric indices further PTH and bone loss. The early phase of glucocorticoid use is associated with increased activity in this is to the accelerated bone resorption of primary hyperparathyroidism. Chronic glucocorticoid use decreases the bone formation rate and the life span of the The mineralization rate and apposition rate with a decrease in bone is normal or reduced (14, 15). PTH, in increases bone formation by the number and activity of osteoblasts rate and apposition rate along with a of the A of these is in However, it is important to that the histomorphometric effects of PTH and glucocorticoids when could be different from their of histomorphometric in GIO and primary hyperparathyroidism findings in primary hyperparathyroidism. of histomorphometric in GIO and primary hyperparathyroidism findings in primary hyperparathyroidism. Glucocorticoids and PTH have effects on the life span of bone with an elevation in bone formation glucocorticoids reduce osteoblast number both and of osteoblasts and A mouse model of glucocorticoid bone loss increased osteoblast in the vertebrae and increased of in the cortical bone These changes were confirmed in patients receiving long-term glucocorticoid therapy when the of the femur was at and the bone was directly PTH, in has an effect on osteoblasts. PTH in with normal bone mass or due to increased their bone formation rate without the of osteoblasts. PTH increased the life span of osteoblasts by PTH and glucocorticoids also have different effects at the when the receptor of is increase in the in cells to enhanced activity and bone resorption Glucocorticoids the production of a receptor that and its at the The and on the cell line is This mechanism could for the early phase of bone resorption with glucocorticoids. Recently, administration of was shown to bone loss in a model of When PTH is administered the effects on and are to those observed with glucocorticoids, with an increase in and a decrease in However, when PTH is administered in an the effects from those observed with glucocorticoids and PTH. The alterations in and are only or do not occur at an osteoblast effect in bone. Most recently, evidence has that a of can bone formation in The effects of glucocorticoids and PTH on the more The area to on the effect of PTH on cancellous bone in clinical trials of This observed both in postmenopausal osteoporosis and men with osteoporosis, and for this in further an important effect of PTH in mechanisms of bone loss. The potential of PTH as an for osteoporosis treatment was first noted ago trials have shown that PTH administration results in increases in spine The potential for skeletal actions of PTH is in the common of PTH primary hyperparathyroidism. As noted in the of primary the cancellous skeleton of the lumbar spine is well whereas the cortical skeleton is reduced The enhanced cancellous bone and have the idea that PTH could be a therapy for the clinical trials of PTH ranging from in have been in The principal common to studies in both men and women is a marked increase in spine BMD with PTH This increase in BMD is greater than the increase observed of increases of and by increases of 40% or more are The difference these the of cancellous bone by in to which both cortical and cancellous elements in the lumbar Bone density at the hip site but not as density remains the same or with treatment. bone mineral The to by et al. daily administration of PTH in women with postmenopausal In addition to increases in BMD total total bone there was a significant risk reduction of vertebral fractures and risk reduction of nonvertebral PTH was shown by et al. to be in GIO. In a postmenopausal women on therapy and steroids were to PTH or no PTH for were not In the PTH treatment vertebral bone density increased by and by (Fig. hip bone density increased by in the PTH group by whereas forearm density did not significantly change indices of bone formation in the first 3 months, whereas resorption markers at 6 months with a that bone formation. months PTH but continued estrogen use, bone density increased further changes in the PTH group were by in the vertebral spine and and in the total and by The in did not demonstrate any significant changes in spine or hip bone had a decrease of forearm bone density the Bone markers to within 6 months PTH. in lumbar spine BMD in postmenopausal women on therapy and steroids given PTH or estrogen for 12 months, as measured by and 12 and months, both estrogen P at 12 months and months. [Reprinted with permission from et al.: J Bone Miner Res ).] These data that administration of PTH, in with is in increasing bone density postmenopausal women being treated with glucocorticoids. It is not however, whether administration of PTH would lead to the same response as when PTH is used to the of et al. This would help to the that PTH could contribute to and be in GIO. The evidence reviewed in this that the many mechanisms for secondary increases in PTH would no longer to have Although not without the balance of evidence from bone bone and PTH from the that of PTH is involved in GIO. In an that can only be as the and for are the potential of PTH in GIO than its with and Silverberg of of and This was by of Bone mineral dual-energy x-ray quantitative computed receptor of