Abnormalities in this elaborate control system underlie most cases of hypercalcemia. The many clinical and laboratory similarities between primary hyperparathyroidism and the hypercalcemia associated with various nonhematologic tumors led to the suspicion that parathyroid hormone was also responsible for so-called "malignancy-associated" hypercalcemia. With the identification, isolation, and cloning of parathyroid hormone-related protein [1-4], this theory has been disproved—at least its classic formulation as an "ectopic parathyroid hormone" syndrome. Nonetheless, it is ironic that parathyroid hormone-related protein, a protein quite distinct in structure from parathyroid hormone (save for partial homology in the 13 amino acid N-terminal region), produces hypercalcemia by masquerading as parathyroid hormone [5].

Little doubt now exists that parathyroid hormone-related protein is an important cause of malignancy-associated hypercalcemia [6-9]. What is less clear is whether other humoral factors act permissively or synergistically with parathyroid hormone-related protein in the genesis of the hypercalcemia. For example, although high circulating levels of this protein have been documented in 50% to 80% of cases of malignancy-associated hypercalcemia, there is often little correlation between the magnitude of the hypercalcemia and circulating parathyroid hormone-related protein levels. Furthermore, some patients with high levels of this protein remain normocalcemic. One wonders if already indicted factors [10]—cytokines such as transforming growth factor-, tumor necrosis factor-, tumor necrosis factor-ß (lymphotoxin) and interleukin-1 and arachidonic acid metabolites such as prostaglandin E—or as yet unindicted humoral factors are also involved.

Patients with primary hyperparathyroidism generally have either normal or elevated circulating calcitriol levels. In contrast, many patients with malignancy-associated hypercalcemia have low calcitriol levels [11]. This is usually ascribed to suppression of the parathyroid hormone-calcitriol axis by the greater degree of hypercalcemia and prevailing hypoparathyroidism. Although this is readily understandable in the case of tumors that do not secrete parathyroid hormone-related protein, tumors that do secrete this protein present a more complex situation. Although parathyroid hormone-related protein, like the hormone itself, can stimulate renal 1 -hydroxylase [12-14], most patients with parathyroid hormone-related protein-associated hypercalcemia have low calcitriol levels [11, 15-17]. The hypercalcemia and associated hypoparathyroidism apparently suppress 1 -hydroxylase activity despite high levels of circulating parathyroid hormone-related protein. It is also conceivable that solid tumors produce, along with parathyroid hormone-related protein, a substance that directly inhibits 1 -hydroxylase activity [16].

Although abnormalities in the metabolism of parathyroid hormone or parathyroid hormone-related protein account for many cases of hypercalcemia, primary disorders of vitamin D metabolism can also lead to hypercalcemia. In addition to its normal production site—the proximal tubule of the kidney—calcitriol can be produced by many other cell types, including decidual cells, keratinocytes, bone cells, spleen cells, peripheral monocytes, and activated macrophages [18-20]. Various granulomatous disorders, including sarcoidosis and tuberculosis, have been associated with hypercalcemia; in sarcoidosis, the underlying cause of the hypercalcemia—unregulated production of calcitriol by pulmonary alveolar macrophages [21, 22]—has been well delineated.

Although uncommon, hypercalcemia is also a well-recognized complication of lymphoma, including both Hodgkin disease and non-Hodgkin lymphoma. Unlike most patients with malignancy-associated hypercalcemia, in whom circulating calcitriol levels are reduced (reflecting suppression of the parathyroid hormone-calcitriol axis), many hypercalcemic patients with lymphoma have calcitriol levels that are either frankly elevated or inadequately suppressed given the degree of hypercalcemia [23].

In this issue, Seymour and colleagues [24] examine calcium homeostasis in the largest series of patients with non-Hodgkin lymphoma yet reported. They first defined a reference range for circulating calcitriol levels in a control group—hypercalcemic patients with multiple myeloma. There are no intrinsic abnormalities in the parathyroid hormone-calcitriol axis in multiple myeloma, and the hypercalcemia appropriately and reversibly suppresses the parathyroid hormone and calcitriol levels [25]. Moreover, parathyroid hormone-related protein levels are generally, although not universally, very low or undetectable [6-9]. The major cause of hypercalcemia in multiple myeloma appears to be local osteolysis mediated by tumor necrosis factor-ß [26] and perhaps by other cytokine stimulators of bone resorption as well.

Using this reference range, Seymour and colleagues classified hypercalcemic patients with lymphoma into two groups. Twelve of 22 patients (55%) had elevated circulating calcitriol levels, and the rest had levels within the established hypercalcemic reference range. That abnormalities in vitamin D metabolism might be even more pervasive than these data indicate is suggested by the observation that 18% of newly diagnosed, normocalcemic patients with lymphoma had serum calcitriol levels greater than the normocalcemic reference range. Moreover, 71% of the normocalcemic patients for whom measurements were available had an elevated fasting urinary calcium excretion, which perhaps indicates even more subtle abnormalities in the parathyroid hormone-calcitriol axis.

That calcitriol is a seminal cause of hypercalcemia in at least certain patients with lymphoma is perhaps best shown by the patients' response to cytotoxic chemotherapy. Seymour and colleagues studied four patients sequentially before and after the initiation of chemotherapy. During chemotherapy in each case, the plasma calcium concentration returned to normal and the circulating calcitriol level was significantly reduced. In three patients, calcitriol levels decreased to the normocalcemic reference range. The remaining patient, who also initially responded, subsequently developed progressive lymphoma and recurrent hypercalcemia, at which time calcitriol levels were again markedly elevated.

Under normal circumstances, physiologic concentrations of calcitriol inhibit 1 -hydroxylase activity [19, 27]. Feedback inhibition of 1 -hydroxylase is not product inhibition but rather is mediated by a genomic effect that has yet to be fully characterized. Extrarenal 1 -hydroxylases appear to be more sensitive to feedback inhibition than the renal 1 -hydroxylase [19], which could explain why the kidney is normally the major source of circulating calcitriol. In addition to 1 -hydroxylation, all tissues with vitamin D receptors can also convert 25-hydroxycholecalciferol to 24R,25-dihydroxycholecalciferol. The latter has yet to be assigned a unique biological function, and, despite ongoing debate, most experts believe that it is nothing more than a component of the vitamin D catabolic pathway [28, 29]. Thus, calcitriol not only impedes its own synthesis (by feedback inhibition of 1 -hydroxylase) but also enhances the entry of its immediate precursor into a competing catabolic pathway (by inducing the 24R-hydroxylase).

What is the meaning of the elevated circulating calcitriol levels in lymphoma-associated hypercalcemia? Calcitriol is strongly bound to a specific vitamin D-binding protein, with less than 1% circulating as the free, biologically active hormone [30]. Seymour and colleagues report only total calcitriol levels and therefore do not definitively exclude alterations in vitamin D-binding protein turnover or activity as an explanation of their findings. However, because elevated vitamin D-binding proteins have only been described in pregnancy and in patients taking estrogens, it is likely that total calcitriol levels accurately mirror free calcitriol levels in lymphoma.

Calcitriol has a half-life of only about 10 hours [31], so elevated circulating levels could reflect increased synthesis, decreased clearance, or both. Metabolic clearance studies will be necessary to definitively exclude alterations in calcitriol catabolism in patients with lymphoma. However, given the usual tight control of 1 -hydroxylase activity, especially the prominent role of feedback inhibition by calcitriol itself, it is unlikely that impaired clearance alone would persistently elevate circulating calcitriol levels. Therefore, it seems reasonable to assume, as do Seymour and colleagues, that the elevated levels in patients with lymphoma reflect increased calcitriol synthesis.

Is the calcitriol renal or extrarenal in origin? Renal 1 -hydroxylase activity is enhanced by parathyroid hormone, parathyroid hormone-related protein, and hypophosphatemia and is inhibited by hypercalcemia, hyperphosphatemia, and calcitriol. Interestingly, parathyroid hormone levels in this study were not as low as those usually reported in patients with malignancy-associated hypercalcemia [11]. Seymour and associates ascribe this to the nature of the particular assay they used and also suggest that their patients may have had less severe hypercalcemia than those patients described previously. Whatever the case, despite markedly different calcitriol levels, the hypercalcemic and normocalcemic patients had essentially identical parathyroid hormone levels. Moreover, parathyroid hormone levels were no different from those found in the control patients with myeloma. Therefore, hyperparathyroidism cannot be considered the cause of the excessive calcitriol production.

Nor, it appears, is parathyroid hormone-related protein to blame. Elevated serum levels of this protein were found in only 2 of the 22 patients (9%), only 1 of whom had an increased calcitriol level. Hypophosphatemia was likewise excluded: Serum phosphate levels were within the normal range and were not related to circulating calcitriol levels. Thus, it seems unlikely that the kidneys are the source of the calcitriol. However, the possibility that certain lymphomas acquire the ability to synthesize a currently unidentified trophic stimulator of renal 1 -hydroxylase activity should be tested.

The regulation of extrarenal 1 -hydroxylases is much less well characterized. However, like the renal enzyme, most appear to be subject to feedback inhibition by calcitriol [18, 19]. A notable exception that has potential relevance to this study is the 1 -hydroxylase of activated macrophages. The experience with the pulmonary alveolar macrophage in patients with sarcoidosis [21, 22]—the paradigm of unregulated calcitriol synthesis—makes the macrophage a logical candidate for the source of calcitriol in patients with lymphoma as well.

Although lymphoma tissue has been shown to convert 25-hydroxycholecalciferol into calcitriol in vitro [32], the precise cell type involved remains to be established. In the case of non-Hodgkin lymphoma, Seymour and colleagues speculate that the calcitriol is produced by tumor-infiltrating macrophages, which are prominent components of intermediate and high-grade neoplasms—those tumor types most commonly associated with hypercalcemia. However, lymphoid cells themselves should also be considered. Indeed, T-lymphocytes that have been transformed by the human T-cell lymphotropic virus type 1 (HTLV-1) have been shown to synthesize calcitriol in vitro [33, 34]. One wonders whether patients with lymphoma who are destined to develop hypercalcemia have a particular clone of cells—a clone that can synthesize calcitriol but, like the activated macrophage, lacks the capacity for feedback inhibition of 1 -hydroxylase.

Calcitriol cannot be the only mediator of hypercalcemia in patients with lymphoma. If it were, how could the hypercalcemia in those patients with appropriately suppressed calcitriol levels (45% of Seymour and colleagues' patients, for example) be explained? Additional circulating or local osteolytic factors must be involved. Parathyroid hormone-related protein, for example, has recently been implicated as a cause of hypercalcemia in non-Hodgkin lymphoma [35, 36], and two patients in Seymour and colleagues' study also had elevated parathyroid hormone-related protein levels. The hypercalcemia commonly associated with HTLV-1-related adult T-cell leukemia-lymphoma [34, 37-40] was initially ascribed to calcitriol [33, 41], but parathyroid hormone-related protein has also been implicated as a cause of the hypercalcemia. Because of the heterogeneity of the disorders collectively described as "lymphoma," it would certainly be a mistake to assume a common cause for hypercalcemia in all such patients.

It seems unlikely that calcitriol is the sole cause of the hypercalcemia even in patients with elevated circulating levels of the hormone. Calcitriol and parathyroid hormone-related protein, for example, may act together to produce hypercalcemia in some patients [24, 42]. The potential roles of cytokines such as interferon-, which is a potent stimulator of macrophage calcitriol synthesis [21, 43], and interleukin-1, which increases the expression of monocyte-macrophage interferon- receptors [44], and other modulators of hematolymphopoietic and bone cell function should also be explored. It will be important to define whether such factors act permissively or synergistically with calcitriol in the genesis of lymphoma-associated hypercalcemia.

Although calcitriol has now been established as an important—and perhaps the dominant—cause of hypercalcemia in both Hodgkin disease and non-Hodgkin lymphoma, only when the cellular biology of vitamin D-metabolizing tissues in general, and hematolymphopoeitic cells in particular, are much better defined will it be possible to fully understand this intriguing disorder of calcium homeostasis. Only then will we be able to translate the etiologic insights provided by Seymour and colleagues into mechanism-specific therapy of this life-threatening complication of uncontrolled lymphoid cell proliferation.