Take Another CYP: Confirming a Novel Mechanism for “Idiopathic” Hypercalcemia

Three recent studies have described a novel mechanism of hypercalcemia related to the effects of vitamin D on calcium homeostasis. These studies collectively confirm that loss of function of the vitamin D 24-hydroxylase enzyme may result in significant clinical hypercalcemia. The initial studies of this disorder, by Schlingmann et al. (1) and Dauber et al. (2), identify homozygous loss-of-function mutations in CYP24A1, the gene encoding the 24-hydroxylase, in the setting of otherwise unexplained infantile hypercalcemia. The report by Schlingmann et al. (1) identified several loss-of-function mutations in eight hypercalcemic infants who were investigated at 6–8 months of age. Their data are most consistent with recessive transmission of the syndrome. Transfection experiments document impaired in vitro 24-hydroxylase activity using constructs of six mutations identified clinically; clinical evidence of impaired enzyme activity was not reported. The Dauber et al. (2) study identified a hypercalcemic infant with a homozygous loss-of-function mutation identical to one of the mutations observed in a compound heterozygote presented by the Schlingmann group (1) and provided strong in vivo evidence relating this mutation to disease. The 10-month-old proband had low circulating 24,25-dihydroxyvitamin D [24,25(OH)₂D] levels compared with age-matched normal children and markedly elevated fractional intestinal absorption of calcium (∼90% vs. ∼45% in age-matched controls). The absence of a coding region mutation in 27 other hypercalcemic infants examined by the Dauber team would suggest that loss of function of CYP24A1 is probably an uncommon cause of this syndrome. In the current issue of the JCEM, Tebben et al. (3) extend these observations, reporting a similar syndrome in adults with presumed loss-of-function mutations in the same enzyme. Several interesting features in the current report differ from the pediatric studies. First, the severity of the hypercalcemia in the infants decreased or corrected with age. Secondly, the Tebben et al. (3) study reports an overt clinical phenotype in the heterozygous state. Of interest, the possibility of a partial or mild phenotype in heterozygotes is raised by the maternal history of nephrolithiasis in the Dauber et al. (2) report, and the identification of a (deletion) abnormality of only one allele in a proband is reported by Schlingmann et al. (1). Finally, the current report describes a clinical response to therapy with the cytochrome inhibitor, ketoconazole, an agent also used in a report of 20 infants with idiopathic hypercalcemia by Nguyen et al. (4). Decreased enzymatic function of the 24-hydroxylase was identified in skin fibroblasts from one of the patients in the Nguyen et al. (4) study.

Interest in the 24-hydroxylase enzyme previously focused on its provision of an alternative pathway for the precursor vitamin D metabolite, 25-hydroxyvitamin D (25-OHD). The 24-hydroxylase provides an energy-favorable conversion of 25-OHD to 24,25(OH)₂D and, when up-regulated, allows a relative enhancement of the pathway to 24,25(OH)₂D rather than to 1,25-dihydroxyvitamin D [1,25(OH)₂D]. Many have considered that this alternative to 1-hydroxylation is primarily of importance in providing the initial step by which abundant stores of 25-OHD can be metabolized to an inactive product, thus leading to eventual catabolism and accelerated clearance of the major circulating vitamin D metabolite. Numerous conditions up-regulate 1-hydroxylase in concert with down-regulation of the 24-hydroxylase, and vice versa (see Ref. 5 for review). Thus, many (but not all) physiological conditions calling for vitamin D activation would limit 24,25(OH)₂D production, and under conditions where vitamin D functions were well met, synthesis of 24,25(OH)₂D would be increased to limit further 1,25(OH)₂D production.

The 1,25(OH)₂D metabolite is the most systemically active form of vitamin D, and the physiological activity of 24,25(OH)₂D is controversial. A body of evidence suggests that 24,25(OH)₂D has distinct physiological actions itself, including inhibition of PTH secretion (6, 7), enhancement of fracture healing (8), and various functions of growth plate cartilage (9).

1,25(OH)₂D serves as a substrate for the 24-hydroxylase as well, thereby generating 1,24,25-trihydroxyvitamin D. Thus, the enzyme is a 25-OHD/1,25(OH)₂D 24-hydroxylase, and it serves as a key enzyme in catabolism of the most active vitamin D metabolite. Indeed, an argument has been made that this is the primary function of the enzyme because the preferred substrate for the 24-hydroxylase is 1,25(OH)₂D (and not 25-OHD) (10). An informative tool in the study of the in vivo physiology of this enzyme has been the Cyp24a1 knockout (KO) mouse (11). Early survival of KO mice is limited and appears to be favorable in pups where down-regulation of 1,25(OH)₂D production occurs. The KO animals have normal 1,25(OH)₂D levels at baseline, but clearance of administered 1,25(OH)₂D is impaired, and levels increase excessively with addition of a vitamin D-supplemented diet.

A reduction in cortical bone occurs in the KO mouse, and this finding is a consequence of 1,25(OH)₂D-induced bone resorption. Thus initial studies appear to relate the observed skeletal pathology to limited clearance of 1,25(OH)₂D as opposed to the inability to generate 24,25(OH)₂D. Likewise, the mutant phenotype does not include abnormal growth plate cartilage, suggesting that in vitro effects on the growth plate may be compensated for by redundant mechanisms in vivo. The KO mouse has been shown to have impaired fracture healing, suggesting that the increased CYP24A1 expression and increased 24,25(OH)₂D production in forming callus (see Refs. 8, 12, and 13 for review) may have important physiological consequences.

The current case report (3) also highlights the use of ketoconazole in blocking hormone synthesis dependent upon enzyme complexes with a cytochrome component. Ketoconazole is a cytochrome inhibitor primarily used as an antifungal agent, but it has been used widely to treat other causes of hypercalcemia due to excess 1,25(OH)₂D production. Paradoxically, it has been used to inhibit CYP24 itself, specifically for the purpose of allowing 1,25(OH)₂D to accumulate by decreasing its degradation. This strategy has been employed to optimize antitumor effects of 1,25(OH)₂D in peripheral tissues (14). With no functional 24-hydroxylase (as in the cases at hand), the relevant clinical effect is presumably on the 1α-hydroxylase and its inhibition of 1,25(OH)₂D synthesis. Nevertheless, ketoconazole-induced inhibition of CYP in general also affects testosterone and cortisol metabolism, so that potential complications of hypogonadism and adrenal insufficiency must be watched for, along with monitoring hepatic function because of well-known toxicity to the liver. It does appear to be generally safe in infancy (4).

Of particular interest are the many important issues that the current study (3) raises with respect to vitamin D metabolism and the relationships of the reported mutations to disease:

Is This a Dominant or Recessive Disorder? Is It a Common Cause of Idiopathic Hypercalciuria?

It appears that the murine KO model, generated by gene ablation, requires homozygosity to produce disease. In this model, one normal allele coexists in the setting where the ablated gene produces no product. Thus one assumes, at least in mice, that gene dosage is not a consideration if the expressed gene encodes a normal protein. It is of interest that reports of the KO mouse only compare the homozygous mutant to heterozygotes (11, 15). Although these differences are large and convincingly show a difference between the heterozygote and mutant homozygote, no data from homozygous wild-type mice are shown. Furthermore, no in vitro evidence that the splice site mutation reported by Tebben et al. (3) generates enzymatic loss of function is provided in their report, but if so, disease in the setting of heterozygosity would suggest a dominant-negative mechanism (whereby the mutant splice variant could impair the function of the product of the normal gene). It is of interest that splice variants of the 24-hydroxylase have been previously reported and can act in a dominant-negative manner (16).

We do not know how common 24-hydroxylase loss of function is in the renal stone-forming population or in otherwise identified patients with idiopathic hypercalciuria. Could CYP24A1 variants account for a large number of such individuals, as suggested by the Tebben et al. (3) report? Or is this a relatively rare recessive disorder that would account for a small minority of hypercalcemic patients, as suggested by the Dauber et al. (2) experience? Further investigation will shed light on this issue, and a better understanding of the phenotype/genotype relationship will no doubt emerge with time.

Is This a Disease of Infantile Onset?

Another interesting difference between the current (3) and previous (1, 2) reports is the age at which disease is manifest. In the infantile forms, it appears that the severity of disease diminishes with age, although with an established diagnosis calcium and vitamin D avoidance have become a component of the chronic clinical management of the affected individuals. In the adult-onset form, disease is manifest throughout life and does not appear to diminish in severity with age. Perhaps the disorder is of somewhat lesser severity and not clinically evident at a young age, but cumulative calcium loading is such that disease is overtly manifest in later adulthood.

What Might Account for Variability of Disease and Variable Survival of Cyp24a1 KO Mice?

The absence of elevated circulating 1,25(OH)₂D levels in untreated or vitamin unsupplemented Cyp24a1 null mice suggests that there is a mechanism by which the synthesis of 1,25(OH)₂D is down-regulated, or that other redundant mechanisms for effective catabolism of this hormone are able to compensate for the absent functional 24-hydroxylase. This capacity to down-regulate 1,25(OH)₂D likely allows for survival in selected Cyp24a1 KO mice. The finding that the KO mice appear to outgrow the systemic effects on calcium metabolism with maturity also suggest that other mechanisms may come into play that allow for degradation of 1,25(OH)₂D that were not available in the newborns. Interestingly, regulation of 24-hydroxylase itself, includes down-regulation by PTH and up-regulation by 1,25(OH)₂D, fibroblast growth factor 23, or glucocorticoids (17) (see Ref. 18 for review).

What is the Cartilage Phenotype of Humans with Documented Loss-of-Function Mutations? How Well Do Fractures Repair in These Individuals?

The growth plate cartilage phenotype and fracture repair capacity need to be assessed in this human condition. Although no abnormal growth plate phenotype is reported in the KO mice (11), growth plate cartilage should be examined carefully in affected humans in light of reported effects of 24,25(OH)₂D in resting chondrocytes of the growth plate and in associated matrix vesicles (9). In contrast, fracture repair appears to be altered in the mutant mice (13), and this may be recapitulated in affected humans.

What Are the Implications of These Findings for Vitamin D Supplementation in Infants?

If functional variants in CYP24A1 are common, there may be a clinically significant influence on degradation of both 25-OHD and 1,25(OH)₂D. Such genetic variance tends to support a cautious approach to recommendations for aggressive vitamin D therapy. It seems that such problems would only be rarely encountered in infancy, but we do not yet know whether we are dealing with a proverbial tip of the iceberg.

Although these questions remain unresolved, CYP24A1 is of major physiological importance. It is highly regulable and is stimulated by 1,25(OH)₂D, fibroblast growth factor 23, and glucocorticoids (17, 18). It is critical to calcium homeostasis; loss of function results in a definitive skeletal and metabolic phenotype in mice, and hypercalcemia is a consequence of a variety of loss-of-function mutations in humans. Interestingly, the identified infantile cases have been shown to have primarily homozygous (or compound heterozygous) variants, whereas the cases identified by the Tebben et al. (3) group may manifest a clinically significant phenotype in the heterozygous state. Most of the latter cases are older, and the phenotype may represent cumulative effects over time that would not be evident in infancy.