Vitamin D Therapy and the Era of Precision Medicine

In this issue of the Journal of Clinical Endocrinology and Metabolism, Enlund-Cerullo et al. shed some welcome light (pun intended!) on the influence of polymorphic variations in the GC gene encoding the vitamin D binding protein (DBP) on vitamin D status and the response to vitamin D supplementation in infants (1). The authors genotyped 913 healthy term infants for 4 different single nucleotide polymorphisms (SNPs) of the GC gene, and measured plasma levels of 25-hydroxyvitamin D (25(OH)D), the principal circulating form of vitamin D and the conventional biomarker for vitamin D status, in cord blood (basal state) and after daily supplementation with 10 or 30 μg of vitamin D3 from 2 weeks to 24 months of age (intervention). The authors found that minor allele homozygosity of all studied GC SNPs, their combined haplotypes and rs4588/rs7041 diplotype 2/2 were associated with lower concentrations of 25(OH)D at all time points in one or both intervention groups (analysis of covariance P < .043), with the exception of rs7041 which did not affect 25(OH)D at birth. Based on these studies, the authors conclude that in infants GC genotype affects 25(OH)D concentration and efficiency of high-dose vitamin D3 supplementation, providing further evidence that genetic background influences vitamin D status.

The present study adds to the growing evidence that nature, as well as nurture, plays an important role in vitamin D physiology. Normal vitamin D homeostasis depends upon an intricate interplay between environmental, photochemical, and biological processes. The parent forms of vitamin D, cholecalciferol (vitamin D3) and ergocalciferol (vitamin D2), may be obtained through the diet or from food supplements. Nevertheless, the chief source of vitamin D for most people is cutaneous production of cholecalciferol, which is generated by photolysis of provitamin D3 (7-dehydrocholesterol) into previtamin D3, with subsequent thermal isomerization to cholecalcifierol. This process requires exposure of the skin by UVB radiation (290–315 nm) usually from sunlight, and is diminished by increased melanin skin pigmentation and use of sunscreens or extensive body coverage (2, 3). Vitamin D is then sequentially hydroxylated in the liver and kidney by CYP2R1 (4) and CYP27B1(5), respectively, to generate 1,25-dihydroxyvitamin D (1,25(OH)₂D, also known as calcitriol), the ligand for the vitamin D receptor and the metabolite that is responsible for most biological actions of vitamin D. Understanding the biology of vitamin D has led to improved ability to quantify plasma levels of vitamin D metabolites (6). On the other hand, considerable controversy persists regarding the normal level of plasma 25(OH)D. Vitamin D deficiency has historically been defined as circulating levels of 25(OH)D lower than 20 ng/mL, a reference adopted by the Institute of Medicine to meet the needs of at least 97.5% of the normal population (7). By contrast, other groups, including the Endocrine Society (8, 9), have proposed vitamin D deficiency as a 25(OH)D level that is less than 20 ng/mL with vitamin D insufficiency as 25(OH)D of 21 ng/mL to 29 ng/mL, and indicated that a 25(OH)D level of at least 30 ng/mL was required for adults to ensure bone health. These recommendations for higher levels of 25(OH)D have been influenced by a variety of human observational and association studies that have suggested a link between vitamin D deficiency and nonskeletal benefits. However, the results of randomized vitamin D intervention trials have generally been disappointing (10), thus reducing enthusiasm for additional supplementation for the general population.

Although environmental and lifestyle factors account for most of the variation in plasma levels of 25(OH)D, genetic variability also is important. Twin studies show a strong (23–80%) genetic contribution to circulating levels of 25(OH)D, but genome-wide association studies (GWAS) suggest a lesser effect of only 7 to 9%. These GWAS have identified SNPs in or near several genes involved in vitamin D homeostasis that are associated with plasma levels of 25(OH)D (11–15). Not surprisingly, these genes include CYP2R1, the principal vitamin D 25 hydroxylase, and CYP24A1, which encodes the vitamin D 24 hydroxylase that inactivates 25(OH)D. And perhaps of greatest relevance to the present work, there is also very strong association between circulating concentrations of 25(OH)D and the GC gene, which encodes the DBP. Although these SNPs have been associated with vitamin D levels, it is unknown until now if they may also be associated with the response to vitamin D supplementation. Among these genes, SNPs near GC have the greatest effect on serum vitamin 25(OH)D and therefore represent the highest possibility for targeted intervention.

GC is a highly polymorphic gene, and more than 120 GC variants have been described, although combinations of 2 common SNPs, which produce 3 major DBP isoforms (Gc1f, Gc1s, and Gc2), are present in most people and appear to segregate based on race and skin pigmentation (16). Over 90% of circulating vitamin D metabolites are tightly bound to DBP, which has a far greater apparent affinity for vitamin D metabolites than albumin, an ancestrally related protein. Therefore, it should not be surprising that conditions that result in changes in the plasma concentration of DBP (eg, pregnancy or chronic liver disease (17)) are associated with corresponding changes in circulating levels of vitamin D metabolites (17), even if levels of free 25(OH)D remain normal (18, 19). Although the authors of this study did not measure circulating levels of DBP, it has been proposed that differences in binding affinity for vitamin D metabolites may affect the plasma half-life of 25(OH)D (17, 20, 21). By contrast, different isoforms of DBP appear to have similar plasma concentrations when quantified by liquid chromatography-tandem mass spectrometry or a polyclonal enzyme-linked immunosorbent assay (22, 23).

Notwithstanding the biochemical observations reported here, it seems obvious to wonder what the biological or clinical significance might be. The free hormone hypothesis states that only unbound or free molecules are active. While free 25(OH)D may be internalized into many cells such as monocytes (20, 24), entry of 25(OH)D into proximal renal tubule cells (and parathyroid cells) is achieved by interaction of the 25(OH)D–DBP complex with the multifunctional endocytic receptor proteins megalin and cubulin (25). Once internalized, 25(OH)D3 is thought to dissociate from DBP for delivery to the renal mitochondria where it can be metabolized to 1,25(OH)2D3. The essentiality of the endocytic process for vitamin D action has been confirmed in vivo, as mice lacking megalin (26), as well as humans and dogs with loss of function mutations in cubilin (27, 28), develop signs of vitamin D deficiency secondary to impaired renal cellular uptake of 25(OH)D3–DBP and the inability to generate 1,25(OH)2D3. By contrast, complete absence of DBP in knockout mice (29) and a woman with homozygous deletion of GC genes (30) is associated with a more modest phenotype in which bone structure is relatively intact and serum levels of calcium and PTH are normal despite nearly undetectable plasma levels of 25(OH)D3 and low levels of 1,25(OH)2D. These observations might argue that the most important role that DBP serves is to buffer serum concentrations of 25(OH)D from rapid changes due to variation in vitamin D supply by prolonging the half-life of 25(OH)D in the serum.

One aspect of this work that is both a strength and weakness is the relative homogeneity of the intervention population; this decreases background genetic effects, but may also mean that this result may not be applicable to diverse populations. A second weakness is the possibility that the effects observed here only apply to high-dose supplementation of vitamin D-sufficient individuals and not to supplementation of vitamin D-deficient individuals. These issues aside, perhaps the most important aspect of this work is the refined focus on the role of GC, and other genes, on vitamin D homeostasis. In this context, genotypes for GC as well as for CYP2R1, CYP24A1, and even CYP3A4, which also can inactivate vitamin D metabolites (31, 32), may provide at least a partial explanation for the wide variation in baseline 25(OH)D levels in the normal population, as well as for the reduced response of some patients to conventional vitamin D supplementation. Notwithstanding the ongoing controversy regarding normal 25(OH)D concentrations and the wide therapeutic window for vitamin D supplementation, we submit that for at least some patients with unexplained vitamin D deficiency, an approach to vitamin D therapy that considers the patient’s specific genetic background will be advantageous. Moreover, given the growing enthusiasm for a precision medicine approach to recommendations involving diet, exercise, stress management, and dietary supplements (33), and the concept of a personal vitamin D response index (34), it is conceivable that clinicians will one day be guided by a patient’s genetic background as they calculate an optimal daily intake for vitamin D. We look forward to the next era of vitamin D biology!

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability: All data generated or analyzed during this study are included in this published article or in the data repositories listed in References.