Answers are grounded in this paper and cite their sources.
IN THIS issue of JBMR, Takeuchi et al.1 describe a patient with humoral hypercalcemia of malignancy (HHM) lasting for 7 years resulting from a neuroendocrine islet cell tumor. The case is unusual in several ways, but the primary focus of this case report is on increased bone formation in a patient with otherwise typical HHM, creating a clinical and biochemical profile more similar to classical primary hyperparathyroidism (HPT) than to classical HHM. The case stimulates one to recall the differences in the pathophysiology between the two syndromes and to try to understand the reasons that underlie these differing pathophysiologies. It is instructive to review the pathophysiological mechanisms of HPT and HHM before reviewing the case. Classical HPT (Bilezikian provides a detailed review2) is characterized by (1) increases in circulating parathyroid hormone (PTH) secreted by abnormal parathyroid tissue; (2) increased osteoclastic resorption of bone; (3) increased osteoblastic bone formation; (4) increased renal calcium reabsorption; (5) increased renal tubular production of 1,25-dihydroxyvitamin D [1,25(OH)2D] with resultant increased absorption of calcium from the gastrointestinal (GI) tract; and (6) as a result of these combined effects, hypercalcemia. Primary HPT is characterized also by hypophosphatemia. One important feature of the skeletal disease of primary HPT is that the increases in osteoblastic and osteoclastic activities are tightly coupled. Thus, despite oversecretion of PTH, net skeletal mass in patients with primary HPT commonly remains normal or near normal for decades or if it declines, the annual rates of net bone loss are relatively low in HPT. These coupled increases in bone formation and bone resorption are easily appreciated using invasive techniques such as bone biopsy and quantitative bone histomorphometry, which reveal increases in osteoclast number and resorption surface, as well as increases in osteoblast number, osteoblast surface, osteoid, and mineral apposition rate. They also can be shown using noninvasive techniques such as measurement of biochemical markers of bone formation and resorption. Typically, HHM is produced by squamous carcinomas (e.g., lung, larynx, esophagus, cervix, and vulva), renal carcinomas, breast carcinomas, ovarian carcinomas, and certain lymphomas (Stewart and Broadus provide a detailed review3). Although these tumors are the most common causes of HHM, HHM has been reported to result from virtually every kind of human tumor, including islet cell and other neuroendocrine tumors. The clinical hallmark of HHM described in the 1950s is hypercalcemia occurring in the absence of or a paucity of skeletal metastatic disease. Classical HHM shares several of the cardinal biochemical and skeletal histological features of HPT. Both syndromes cause hypercalcemia and between them account for 90% of cases of hypercalcemia in humans. Both syndromes are characterized by increased osteoclastic bone resorption. Both are caused by secretion of humoral factors: PTH in HPT and tumor-derived PTH-related protein (PTHrP) in HHM. Both share renal phosphate wasting and hypophosphatemia. Both are characterized by increased renal tubular reabsorption of filtered calcium (although there are no data as of yet in humans comparing the efficiencies of PTH vs. PTHrP in mediating this process). In contrast to these similarities, there are two key differences. First, in contrast to the increase in circulating 1,25(OH)2D in HPT, 1,25(OH)2D concentrations typically are low or undetectable in patents with HHM.4 This has been confirmed repeatedly and appears to be unique to humans; in rodent models of HHM, circulating 1,25(OH)2D is elevated. Despite this difference having been observed more than 20 years ago, the pathophysiological and cellular mechanisms responsible for this discordance between HPT and HHM remain unexplained. Some have suggested that tumors produce factors that inhibit 1,25(OH)2D production by the kidney in response to PTHrP.5 This seems an unlikely explanation for the reductions in plasma 1,25(OH)2D in most cases of HHM because bisphosphonate treatment, which lowers calcium and stimulates secretion of endogenous PTH, results in an increase in circulating 1,25(OH)2D in HHM.6 The discordance between 1,25(OH)2D concentrations in HPT and HHM and more particularly the reduction in circulating 1,25(OH)2D in HHM is particularly perplexing, because normal human volunteers infused or injected with either PTH or PTHrP display increases in circulating 1,25(OH)2D.7-9 The second hallmark that distinguishes HPT and HHM relates to osteoblastic activity. Osteoblastic activity is increased in a fashion coupled to the increased osteoclastic activity in patients with HPT, as described previously. This contrasts to patents with HHM who have even greater increases in osteoclastic activity than those with HPT but who have a dramatic decrease in osteoblastic activity, evidenced by markedly reduced numbers of osteoblasts and little or no osteoid.10 This dramatic “uncoupling” of bone formation and bone resorption is apparent on bone biopsy specimens histologically both in humans10 as well as in rodent models of HHM.11, 12 It was one of the original observations that suggested that the humoral factor responsible for HHM was something different than PTH.10 This uncoupling also is strikingly apparent when markers of bone turnover are measured; bone resorption markers such as deoxypyridinoline and N-telopeptide are increased in both HPT and HHM, but markers of bone formation such as osteocalcin and bone-specific alkaline phosphatase are increased in HPT but decreased in HHM.13 The uncoupling is so dramatic that one would predict rapid losses of skeletal mineral in patients with HHM, and this in fact occurs; skeletal biopsy specimens from patients with HHM of only 3 months duration show an approximate 50% reduction in trabecular bone volume.10 Indeed, it is precisely the dramatic net skeletal calcium losses that are primarily responsible for hypercalcemia in HHM. Again, like the discordance between HPT and HHM regarding 1,25(OH)2D, despite our awareness of this osteoclast-osteoblast uncoupling for 20 years, the pathophysiology and cell biology underlying this phenomenon remain completely mysterious. Against this background, the case described by Takeuchi et al. is unusual. Like a typical patient with HHM, the patient described here had metastatic malignant tumor; hypercalcemia; hypophosphatemia; renal phosphorus wasting (a low tubular maximum for phosphorus [TmP]/glomerular filtration rate [GFR]); an elevated circulating PTHrP and undetectable PTH; no evidence of bone metastases by bone scintigraphy or autopsy; and marked increases in three bone resorption markers, pyridinolines, deoxypyridinoline, and N-telopeptide. However, unlike a typical patient with HHM, but like a patient with HPT, his circulating 1,25(OH)2D was at the upper limits of normal. Moreover and of central importance in the case, the patient's serum osteocalcin was elevated consistently, suggesting that bone formation was increased. Supporting the concept that bone formation was increased, the bone-specific alkaline phosphatase was mildly elevated also (instead of suppressed as is typical in HHM) and the bone at autopsy showed evidence of osteoblastic bone formation (although it would have been more convincing to have confirmed an increase in bone formation using quantitative bone histomorphometry). In further support of an increase in bone formation is the observation that despite 7 years of markedly increased bone resorption, the patient's bone mineral density, although low, was not extremely low. Other features of this case are unusual. First, the patient survived 7 years before succumbing to metastatic tumor. This is distinctly unusual for patients with HHM in whom 50% survival is typically 3 months from the time of diagnosis of hypercalcemia.14 This is, of course, a reflection of the type of tumor; this prolonged survival is typical of patients with islet cell tumors. Second, the patient had a malignant neuroendocrine tumor. Although these tumors are well known to produce PTHrP, they rarely cause HHM, accounting for <1% of patients with HHM. Third, the patient had no particular response to treatment with bisphosphonates, which in most cases are very effective in lowering serum calcium in patients with HHM. We are told that the patient did not even have transient reductions in serum calcium after bisphosphonate treatment. Because there is no reason to think the bisphosphonates should not have at least transiently reduced bone resorption in this patient, the lack of response to bisphosphonates suggests that the primary pathophysiological cause of hypercalcemia in this patient may have increased renal tubular reabsorption of calcium. PTH-mediated increased renal tubular calcium reabsorption is typical of HPT, and PTHrP-mediated renal tubular calcium reabsorption has been suggested to play an important role in some patients with HHM as well.15-17 So the question arises, what is different in this patient that might account for the hybrid HPT-HHM pathophysiology? Takeuchi et al. suggest that this case provides further evidence that the suppressed, uncoupled bone formation typical of HHM results from the secretion by typical HHM-associated tumors of agents, cytokines, factors, etc. that suppress bone formation that otherwise would have been increased by PTHrP, and that such bone formation-suppressing factors were not produced by this unusual neuroendocrine tumor, thus permitting the increased osteoblastic bone formation. This is one possible explanation, but it remains unproved in this case or any other case. There are other possible explanations as well. Human PTHrP is translated as three large precursors of 139, 141, and 173 amino acids, together with a signal and propeptide. These PTHrP precursors undergo extensive proteolytic post-translational processing to result in a family of PTHrP peptides such as PTHrP(1-36), PTHrP(38-94), a larger form that is approximately PTHrP(1-86), a carboxy-terminal form, and others as well that have been less well defined. In addition, PTHrP has been shown glycosylated by some but not all tissues. Moreover, there is evidence for cell-type specificity in the profile of PTHrP secretory forms.18, 19 For example, islet tumor cell lines related to that which the Takeuchi patient had have been shown to post-translationally process PTHrP differently than other cell lines.18, 19 Thus, another explanation for the differing pathophysiology in this case as compared with “garden variety” HHM is that this unusual neuroendocrine tumor secreted different forms of PTHrP than those typically secreted by renal and squamous carcinomas and that the family of PTHrP peptides produced by this tumor resulted in activation of bone formation. This possibility, like the first, remains unproved. I favor yet another explanation. PTH and amino-terminal PTHrP peptides bind to the same receptor, the PTH1 receptor. They bind with the same affinity; they activate downstream signaling pathways (protein kinase C [PKC], protein kinase A [PKA], cyclic adenosine monophosphate [cAMP], and mitogen-activated protein kinase [MAPK]) with equal efficacy. Thus, it is difficult to see how two such interchangeable peptides could have opposite effects on bone formation. Interestingly, PTH secretion by the parathyroids has been shown repeatedly to occur in pulses or bursts.20, 21 This is not surprising for a neuroendocrine peptide. In contrast, PTHrP has been shown to be produced by squamous carcinomas in a constitutive or continuous fashion.22 This is not surprising either, because squamous carcinomas do not contain neuroendocrine regulatory secretory machinery such as dense core secretory granules. Thus, one could hypothesize that both PTH and PTHrP when secreted continuously will lead to suppression of bone formation, perhaps through the down-regulation and desensitization of PTH receptors in bone and kidney cells, which has been so well documented.23 On the other hand, when either peptide is secreted intermittently, it might increase bone formation. There is ample precedent for this in the GnRH-LH-FSH system, in which continuous administration of GnRH desensitizes and down-regulates the pituitary GnRH receptor, resulting in hypogonadotropism. This desensitization/down-regulation is used clinically to suppress gonadotrophin secretion, for example, in prostate cancer. In contrast to continuous administration of GnRH, pulsatile administration of GnRH has the opposite effect: it up-regulates LH and FSH responses. Accordingly, pulsatile therapy with GnRH is used clinically to induce FSH/LH secretion. Thus, by analogy, perhaps intermittent secretion of PTH and PTHrP would have one effect, and continuous secretion of either peptide would have the opposite effect. More specifically, continuous, constitutive secretion of PTHrP by a squamous carcinoma or renal carcinoma may have one effect, whereas intermittent or pulsatile secretion of the same peptide(s) by a neuroendocrine tumor may have directionally opposite effects. This could be the explanation for the unusual findings in the Takeucki case. Is there any support for this hypothesis? There is very strong support for the fact that continuous exposure of the skeleton to PTH in high doses results in a net catabolic or resorptive response, whereas intermittent (i.e., once daily) exposure to the same doses of PTH leads to dramatic increases in bone formation and bone density. This has been documented using both bone formation markers as well as using bone biopsy and histomorphometry (Neer et al. and Dempster et al. provide recent reviews24, 25). Indeed, recently, we have come full circle with PTH in this regard; although HPT (with continuous secretion of very high levels of PTH) was described 80 years ago as the quintessential skeletal “catabolic” disease, we now appear to be on the verge of treating humans with osteoporosis with intermittently administered PTH. PTH is now regarded as the ultimate “anabolic” agent. The reasons for this diametrically opposed anabolic-catabolic physiology have everything to do with the osteoblast or osteoblast precursors being exposed to PTH continuously (where it is catabolic) or intermittently (where it is anabolic). What about PTHrP? Can it to be either catabolic or anabolic depending on secretory profile? The answer to this question is clearly “yes” as well. As noted previously, HHM is the classic skeletal catabolic disease, with uncoupled increases in resorption and suppression of formation.10-12 Of course, in HHM, bone cells are exposed to concentrations of constitutively secreted PTHrP that are continuously elevated, because typical squamous and renal carcinomas can only secrete PTHrP constitutively or continuously.22 In addition and in contrast to these catabolic skeletal effects of continuous PTHrP, there is very strong evidence that PTHrP can have potent anabolic effects on the skeleton when administered intermittently (once per day). First, daily injections of amino-terminal forms of PTHrP have been known for years to increase skeletal mass in rats,26, 27 and this has been underscored in more detail recently.28 This increase in skeletal mass is accompanied by improvements in skeletal biomechanical properties as well.28 Second, in postmenopausal women treated with PTHrP(1-36) administered as a daily subcutaneous injection for 2 weeks, the bone formation marker osteocalcin increased, indicating an osteoblastic response to intermittently administered PTHrP in humans.29 The Takeuchi case may represent a one-third example of the ability of PTHrP to stimulate bone formation in humans when “administered” chronically, with the caveat that it may need to be secreted intermittently in pulses by rare tumors such as islet tumors for this effect to occur. I emphasize that although the anabolic effects of intermittently administered PTHrP are well described, this proposed pathophysiology in the current case remains hypothetical rather than proven. Of course, these considerations raise the interesting pathophysiological and cell biological question: “What regulates osteoclast-osteoblast coupling?” A corollary question is “What are the key target cells for PTH and PTHrP in bone required for the catabolic and anabolic effects of PTH and PTHrP?” The answer to the first question is that we do not know at present in any kind of complete detail. Coupling may be caused by osteoclast-activating factors derived from PTH/PTHrP-stimulated osteoblasts or osteoblast precursors. It may be caused by osteoblast activation by osteoclast resorption-mediated release of osteoblast-stimulating factors such as transforming growth factor (TGF) β or insulin-like growth factor (IGF) 1. It may be caused by coordinated up-regulation of osteoclast precursors and osteoblast precursors via the receptor activator of NF-κB (RANK), RANK ligand (RANKL), osteoprotegerin (OPG) system or it may be caused by other coordinating mechanisms. Given this lack of complete understanding of the intimate cellular and molecular details of normal coupling, it is of course impossible to describe what is wrong when uncoupling occurs in HHM. So what have we learned from the Takeuchi case? It has been 80 years since malignancy-associated hypercalcemia was described. It has been 45 years since HHM was first described. It has been 15 years since PTHrP was identified, and 12 years since PTHrP immunoassays became available for clinical research. We now know almost everything about HHM in pathophysiological terms, and we can reproduce the cardinal features of the syndrome in laboratory animals and humans. The Takeuchi case reminds us that we still have a few things to learn about HHM. Specifically, “Why is the regulation of 1,25(OH)2D different in patients with HHM and HPT?” and “Why is normal osteoblast-osteoclast coupling dysregulated in HHM?” or more fundamentally, “What regulates osteoblast-osteoclast coupling, and why is it deranged in HHM?” Given the rate of accumulation of new information about HHM, about the anabolic effects of PTH and PTHrP, and about osteoblast-osteoclast coupling over the past 10 years, there is reason to be optimistic that the answers to these questions will soon become clear. This work was supported by NIDDK 51081. Dr. Stewart serves as a consultant for Eli Lilly regarding PTH. He also has equity interests in Osteotrophin, Inc., for PTHrP.
No takes yet. Share an insight, caveat, or question.
Andrew F. Stewart (2002) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: