Rickets as a bone disease of young children was clearly described by Whistler in 1645 ¹ and Glisson in 1650.² The relationship of this disease with lack of exposure to sunlight was already suspected in the 19th century, since the incidence of rickets was higher in children living in large industrialized towns than in children living in rural districts (see Chapter 15).^3,4 Early in the 20th century, Huldshinsky,⁵ Chick et al.,⁶ and Hess and Weinstock⁷ demonstrated that rachitic children were cured after exposure to sunlight. In the United Kingdom, following an independent line of research in search of essential nutritional factors, Mellanby and Cantag⁸ raised dogs on a diet of oatmeal (the basic food in parts of the United Kingdom where rickets was endemic) and observed that they developed rickets, curable by cod liver oil.⁹ However, McCollum et al.¹⁰ could demonstrate that cod liver oil made vitamin A-deficient by aeration and heating was still able to cure rickets and thus contained a new essential nutrient called vitamin D.¹⁰ The two discoveries of vitamin D were unified by the demonstration of Goldblatt and Soames¹¹ that irradiation of 7-dehydrocholesterol in the skin could produce the antirachitic vitamin D. Similar observations were made by Hess and Weinstock.⁷ Windaus,¹² a German chemist, then identified the structure of vitamins D₂ and D₃ after irradiation of plant sterols (ergosterol) or 7-dehydrocholesterol,¹² which earned him the Nobel Prize in chemistry in 1928.

The elucidation of the mode of action of vitamin D can be separated in several phases: the discovery of (1) the endogenous activation of vitamin D by sequential hydroxylations at C₂₅ and C₁; (2) the molecular mechanisms following the binding of 1,25-dihydroxyvitamin D [1,25(OH)₂D] to a specific and quite ubiquitous nuclear transcription factor, vitamin D receptor (VDR), a receptor now known to recruit a large number of proteins; and (3) the regulation of the expression of a very large number of genes (between 1% and 5% of the human genome) involved in either calcium homeostasis or related to cell proliferation or differentiation.

Origin of Vitamin D: Nutrition and Photosynthesis

Vitamin D can be obtained from dietary sources of vegetal (vitamin D₂ or ergocalciferol) or animal origin (vitamin D₃ or cholecalciferol). About 50% of dietary vitamin D is absorbed by the enterocytes and transported to the blood circulation via chylomicrons. Part of this vitamin D is taken up by a variety of tissues (fat and muscle) before the chylomicron remnants and its vitamin D finally reaches the hepatocytes. The best food sources are fatty fish or its liver oils, but it is also found in small amounts in butter, cream, and egg yolk. Both human and cow’s milk are poor sources of vitamin D, providing only 15 to 40 IU/L, and equally minimal concentrations of 25(OH)D or 1,25(OH)₂D.¹³ Only an intake of pharmacologic amounts of vitamin D (6000 IU/d) can increase the vitamin D concentration of milk to a level equivalent to the daily requirements of an infant.¹⁴ Vitamin D intake is a poor predictor of serum 25(OH)D concentrations in subjects with an intake between 2 and 20 µg/d.^15,16 It is very difficult to obtain adequate vitamin D from a natural diet. However, in North America, 98% of fluid and dried milk (≥400 IU/L), as well as some margarine, butter, and certain cereals, are fortified with vitamin D₂ (irradiated ergosterol) or D₃, but the real vitamin D content is frequently quite different from the labeling standard. Skim milk and even proprietary infant formula frequently do not have the stated vitamin D content.^17,18 Vitamin D is remarkably stable and does not deteriorate when food is heated or stored for long periods. The Second National Health and Nutrition Survey (NHANES II) reported a median intake of about 3 µg/d in adults (range 0 to 49 µg),¹⁹ whereas a slightly lower median intake (2.3 µg) was recorded in older women.²⁰ In view of the low vitamin D content of a vegetarian diet (natural vitamin D intake is indeed related to intake of animal fat), vitamin D deficiency and rickets is a risk factor for strictly vegetarian children with insufficient sun exposure or vitamin D supplementation.²¹

Nature probably intended that most vitamin D would be generated by photosynthesis in the skin, with minor contribution from food sources. However, exposure to sunlight also increases the risk of dermal photodamage and several skin cancers, including melanoma. This was no real problem during human evolution, but with increasing life expectancy, the benefits of UV light for the photosynthesis of vitamin D should be compared with the lifetime risk of skin damage, especially since vitamin D supplementation can safely replace the skin synthesis. The recommended dietary allowances by the U.S. Food and Nutrition Board of the National Research Council and the 1998 updated recommendations are given in Table 3-1, and similar recommendations are still valid in Europe.^22,23 However, these recommendations were based on rudimentary knowledge of optimal vitamin D status and need to be revised upwards.

Table 3-1

Adequate Intake, Previous Recommended Dietary Allowance, Reasonable Daily Allowance, and Tolerable Upper Limit for Vitamin D

AI, Adequate intake; RDA, recommended dietary allowance; UL, upper limit.

^*Data from the Food and Nutrition Board, National Research Council, NAS: Recommended Dietary Allowances, 10th ed. Washington, DC: National Academy Press, 1989.

^†Data from the Food and Nutrition Board: Dietary Reference Intakes for Calcium, Phosphorus, Magnesium, Vitamin D and Fluoride. Washington, DC: National Academy Press, 1997. Fairly similar advice is given by the European Food Safety Authority.²³

^‡Similar upper levels for vitamin D intake were defined by the European Food Safety Authority,²³ defining 25 µg or 1000 IU/d as the upper limit for children 0 to10 years of age and 50 µg or 2000 IU/d for children older than 11 years and adults.

Hypervitaminosis can occur when pharmaceutical vitamin D is taken in excess, with a wide variety of symptoms and signs related to hypercalciuria, hypercalcemia, and metastatic calcifications (Table 3-2). The toxic dosage has not been established for all ages, but infants and children are more susceptible. Toxicity should always be monitored when daily doses markedly exceeding the present upper limit of more than 50 µg are given for a longer period. Overproduction of renal 1,25(OH)₂D by abnormal hormonal stimuli (as seen in fibroblast growth factor-23 [FGF-23] or Klotho-null mice) or absence of CYP24A1 (see later), the main catabolizing enzyme, causes the same calcemic side effects, with severe multiple-organ calcification (especially kidney, vascular wall, and heart valves) leading to premature death.²⁴

Table 3-2

Symptoms of Vitamin D Toxicity

Hypercalciuria

Kidney stones

Hypercalcemia

Hyperphosphatemia

Polyuria

Polydipsia

Decalcification of bone

Ectopic calcification of soft tissues (kidney and lung)

Nausea and vomiting

Anorexia

Constipation

Headache

Hypertension

Most vertebrates also accomplish their needs for vitamin D by photochemical synthesis in the skin; therefore, vitamin D is not a true vitamin. It is formed from 7-dehydrocholesterol (7DHC or provitamin D₃), which is present in large amounts in cell membranes of keratinocytes of the basal or spinous epidermal layers. By the action of ultraviolet B (UVB) light (290 to 315 mm), the B ring of 7DHC can be broken to form previtamin D₃. Previtamin D₃ is unstable, and in the lipid bilayer of membranes, it is rapidly isomerized to vitamin D₃ by thermal energy, followed by transport to the serum vitamin D–binding protein and uptake into the liver for further metabolization.

The production of previtamin D₃ is a nonenzymatic photochemical reaction which is not subject to regulation other than substrate (7DHC) availability and intensity of UVB irradiation. 7DHC is the last precursor in the de novo biosynthesis of cholesterol. The enzyme 7HDC-Δ7-reductase (or sterol Δ7-reductase) catalyses the production of cholesterol from 7DHC. Inactivating mutations of the 7DHC-Δ7-reductase gene ²⁵ are the hallmark of the autosomal recessive Smith-Lemli-Opitz syndrome, characterized by high tissue and serum 7DHC levels and multiple anomalies, including craniofacial dysmorphism and mental retardation due to the lack of cholesterol synthesis.²⁶ These patients may exhibit sometimes increased serum vitamin D and 25(OH)D concentrations.²⁷ Likewise, animals pretreated with a specific sterol-Δ7-reductase inhibitor also exhibit an augmented vitamin D synthesis following UVB irradiation.²⁸ With increasing human age, cutaneous stores of provitamin D decrease, together with decreased photoproduction of vitamin D.¹⁶ In cats and the feline species in general, the high cutaneous sterol-Δ7-reductase activity hampers photoproduction of vitamin D, making it a true vitamin.²⁹ Apart from substrate (7DHC) availability, the photochemical synthesis of vitamin D₃ in the skin largely depends on the amount of UVB photons that strike the basal epidermal layers. Glass, sunscreen, clothes, and skin pigment absorb UVB and blunt vitamin D₃ synthesis. Latitude, time of day, and season are factors that influence the intensity of solar radiation and the cutaneous production of vitamin D₃. Therefore, there is a risk for a shortage of vitamin D supply during winter and spring. In both the Northern and Southern hemispheres above 40 degrees latitude, vitamin D₃ synthesis of the skin decreases or disappears during winter months, owing to the low inclination of the sun and the atmospheric filtration of the shortest (but effective for vitamin D₃ synthesis) UV waves of sunlight. The importance of skin synthesis of vitamin D₃ to maintain normal vitamin D status is best reflected by the vitamin D deficiency observed in submarine personnel or inhabitants of Antarctica³⁰ during prolonged absence of sun exposure, and also by the extremely high prevalence of vitamin D deficiency in countries where exposure to sunlight is extremely low for cultural and religious reasons, as in several Arabian countries with strict adherence to Islamic rules for body covering.^31–34 Solar exposure of 2 hours per week of the face and hands is probably sufficient for maintaining normal 25(OH)D concentrations in children³⁵ and adults but should be further fine-tuned according to the climate and latitude.³⁶

Nature has built in several feedback mechanisms to minimize the risk that prolonged sun exposure would cause vitamin D intoxication. Cutaneous vitamin D and especially previtamin D are photosensitive and will be degraded to inactive sterols (lumisterol, tachysterol) before they are translocated to the circulation (Fig. 3-1). Only a maximum of 10% to 15% of the provitamin D will be converted to vitamin D. Sunlight-induced melanin synthesis, acting as a natural sunscreen, provides an additional negative feedback.

FIGURE 3-1 Origin of vitamin D: photosynthesis and metabolism.

Metabolism of Vitamin D

Vitamin D is biologically inert and requires two successive hydroxylations in the liver (on C₂₅) and kidney (on the α position of C₁), using cytochrome P450 enzymes 37,38 to form its hormonally active metabolite, 1α,25-dihydroxyvitamin D (see Fig. 3-1).

25-Hydroxylation

25(OH)D was the first metabolite identified after the availability of radiolabeled vitamin D₃.39,40 Although the liver is probably the main tissue responsible for 25-hydroxylation of vitamin D, extrahepatic 25-hydroxylation has been observed in vitro in a large number of tissues. In vivo observations after hepatectomy in rats also revealed that the conversion rate of [³H]-vitamin D was still about 10% when compared with intact rats.⁴⁰ The hepatic 25-hydroxylation step is probably performed by more than one enzyme, localized either in the inner mitochondrial membrane (CYP27A1 or sterol 27-hydroxylase) or in the microsomes (including CYP2D11, CYP2D25, CYP3A4, and especially CYP2R1).^37,41,42 CYP27A1 is a multifunctional enzyme with broad substrate specificity and is mainly involved in the 26- or 27-hydroxylation of cholesterol and bile-acid precursors.³⁷ The rather mild (if any) disturbance of vitamin D metabolism in animals or humans lacking this enzyme⁴³ indicates that the 25-hydroxylation of vitamin D does not rely exclusively on the activity of CYP27A1. It is therefore likely that a microsomal enzyme is the more physiologic enzyme, as initially suspected on the basis of hepatic enzyme activity being much higher and with lower K_m in the microsomal fraction when compared with mitochondrial 25-hydroxylase activity.⁴⁴ The microsomal enzyme activity is up-regulated by vitamin D deficiency or by prior exposure to phenobarbital. The most important 25-hydroxylase is probably CYP2R1, since a homozygous mutation was identified in a patient with classical rickets with low 25(OH)D levels.⁴²

1α-Hydroxylation

25(OH)D is biologically inactive and requires further hydroxylation in the kidney 45,46 to the active hormone, 1,25(OH)₂D, by 25-hydroxyvitamin D-1α-hydroxylase (CYP27B1). The production of 1,25(OH)₂D is regulated primarily at this final step by several factors (vide infra). The rat, mouse, and human 1α-hydroxylase have been cloned by several groups^47–51 and mapped on human chromosome 12q13.3 in close vicinity to the VDR gene. The proximal renal tubule is the principle site of 1α-hydroxylation, but high levels of 1α-hydroxylase mRNA have also been found in human keratinocytes,⁴⁸ and its gene expression is also observed in mouse macrophages⁵² and about ten other tissues.²⁴ 1α-Hydroxylase activity is under tight control by 1,25(OH)₂D (negative but probably indirect feedback); parathyroid hormone (PTH), calcitonin, and insulin-like growth factor 1 (all positive feedback); and phosphate, calcium, and especially FGF-23 (negative regulation).^38,53,54 The promoters of the mouse and human 1α-hydroxylase genes have been characterized with a profound responsiveness to PTH and a negative regulation by 1,25(OH)₂D^55,56 by complex chromatin and DNA modifications.⁵⁷

Pseudovitamin D–deficiency rickets (PDDR), also known as vitamin D-dependency rickets type I, is an autosomal recessive disease characterized by failure to thrive, muscle weakness, skeletal deformities, hypocalcemia, secondary hyperparathyroidism, normal to high serum levels of 25(OH)D, and low serum 1,25(OH)₂D concentrations, all caused by impaired activity of the renal 1α-hydroxylase.⁵⁸ These patients recover with supplementation of physiologic doses of 1,25(OH)₂D. The human CYP27B1 maps to the previously identified PDDR locus and mutations found in this gene in patients with PDDR provide the molecular genetic basis for the disease.^48,59

24-Hydroxylation: Catabolism or Specific Function?

An alternative hydroxylation of 25(OH)D occurs on carbon 24 by the multifunctional enzyme, 24-hydroxylase (CYP24A), mapped on human chromosome 20q13.⁶⁰ This enzyme not only initiates the catabolic cascade of 25(OH)D and 1,25(OH)₂D^61,62 by 24-hydroxylation but catalyzes also the dehydrogenation of the 24-OH group and performs 23-hydroxylation, resulting in 24-oxo-1,23,25(OH)₃D.⁶² This C₂₄ oxidation pathway finally leads to calcitroic acid, which is the major end product of 1,25(OH)₂D (Fig. 3-2). In vivo evidence for this catabolic role of 24-hydroxylase was provided by the generation of mice deficient in the 24-hydroxylase gene, resulting in pathology consistent with systemic excess of 1,25(OH)₂D.⁶³ The expression of the 24-hydroxylase gene has been detected in virtually all nucleated cells. The induction of CYP24A belongs to the most sensitive biomarkers for responsiveness and is explained by the presence of several vitamin D responsive elements in its promoter.^64,65 As a consequence, CYP24A mRNA levels appear to fall under the detection limits in VDR knockout mice.^50,66 No human mutation in the 24-hydroxylase gene have yet been identified, but local overexpression of the enzyme may be involved in cancer.⁶⁷

FIGURE 3-2 Catabolism of 1,25(OH)₂D.

Other Metabolic Pathways For Vitamin D and Its Metabolites

Apart from the multifunctional 24-hydroxylation pathway, C₂₃ and C₂₆ hydroxylation of 1,25(OH)₂D is also possible in the absence of prior 24-hydroxylation. The 23-hydroxylation probably only becomes important in the case of vitamin D excess; its major locus is the kidney. In contrast, 26-hydroxylation is mainly performed outside the kidney. Both activities are necessary for the formation of 25(OH)D- or 1,25(OH)₂D-23,26-lactones (see Fig. 3-1). The A-ring metabolism involves the oxidation of C₁₉ and the recently discovered 3-epimerisation. The latter, irreversible, reaction occurs only in a limited number of cells (e.g., keratinocytes, bone, and parathyroid cells) and is performed by hydroxysteroid dehydrogenases.⁶⁸

The enzymes involved in the metabolic degradation of 1,25(OH)₂D do not recognize all vitamin D analogs in the same way. Indeed, analogs with either 20-epi or 20-methyl configuration or 16-ene structure show an impaired 23-hydroxylation. These or other analogs are then preferentially hydroxylated on C₂₆ or on new terminal carbons of the side chain. Such alternative metabolism can certainly explain part of the specific selectivity profile of a number of analogs (vide infra).

Vitamin D is mainly excreted in the bile after esterification in the liver, but some of its more polar metabolites (e.g., calcitroic acid) are excreted via the urine. The enterohepatic recirculation of vitamin D esters is probably devoid of biological relevance.

Vitamin D Transport

Nutritional vitamin D is absorbed by the gut and then transported via the lymphatic system by chylomicrons ⁶⁹ and stored in several tissues (e.g., fat and muscle). Skin-produced vitamin D probably binds directly to an α-globulin known as vitamin D binding protein (DBP) and is then transported to the liver, where it is hydroxylated and thereafter released as 25(OH)D.

Human DBP,⁷⁰ detected immunologically in 1959 as a group-specific component, or Gc-globulin,⁷¹ is a 43-amino-acid glycoprotein synthesized by the liver. Long before DBP’s functions had been characterized, its polymorphicity was already used in population genetics, parentage testing, and forensic medicine.⁷² Worldwide, over 120 Gc alleles have been detected,⁷³ making the DBP locus one of the most polymorphic known. The Gc1F, Gc1S and Gc2 are the three most common alleles. Since in the many thousands of sera tested, none had been found of DBP deficiency, such a mutation was for a long time considered to be lethal; but this was contradicted by the generation of viable and fertile homozygous, DBP-deficient mice (DBP-null animals).⁷⁴ The existence of similarity among the genes and protein structure of DBP, albumin, and α-fetoprotein is long recognized.⁷⁵ The crystal structure, with or without actin, is now available and identified a surface cleft to bind 25(OH)D.⁷⁶

Role of Vitamin D Binding Protein For Vitamin D Homeostasis

DBP, the major plasma carrier of vitamin D₃, all its metabolites, and the vitamin D₃ analogs, has one vitamin D sterol-specific binding site.⁷⁵ The relative binding affinity is 25(OH)D-23,26-lactone > 25(OH)D = 24,25(OH)₂D = 25,26(OH)₂D (K_a = 5.10⁸ mol/L at 4° C for human DBP) >> 1,25(OH)₂D (4.10⁷ mol/L) >> vitamin D >> previtamin D.⁷⁷ The affinity for D₂ metabolites is slightly lower than for D₃ metabolites in mammals, but especially in birds. Since probably only non-DBP-bound vitamin D metabolites can readily cross the plasma membrane, and since the VDR has a much higher affinity for 1,25(OH)₂D than for 25(OH)D (100-fold difference), while the opposite is true for DBP, it is clear that 1,25(OH)₂D has substantially higher cellular uptake than 25(OH)D. This is also confirmed by the distribution space of (radiolabeled) metabolites: 25(OH)D has a distribution space similar to that of DBP and the plasma volume, whereas the distribution space of 1,25(OH)₂D is closer to that of intracellular water. The half-life of 25(OH)D and 1,25(OH)₂D in the human circulation is about 2 to 3 weeks and 4 to 6 hours, respectively.^78,79 DBP’s function in the vitamin D endocrine system is assumed to reflect the “free hormone” hypothesis, which states that the unbound (free) rather than the protein-bound fraction of the active vitamin D hormone is responsible for the biological activity. The plasma concentration of DBP is increased by estrogens in most mammalian species and birds. In women, the DBP concentration therefore doubles at the end of pregnancy.⁸⁰ Recent studies with megalin knockout mice indicate that megalin, a lipoprotein-like receptor present at the surface of the proximal tubular cells in the kidney, is responsible for the reabsorption of DBP and of DBP complexed with vitamin D sterols. This megalin reabsorption mechanism may control the availability of the 25(OH)D/DBP complex for the 25(OH)D-1α-hydroxylation enzyme and explain the severe bone disease of megalin-deficient mice.⁸¹

Other Functions of Vitamin D Binding Protein

DBP binds globular actin with a high affinity (K_a = 2 × 10⁹ mmol/L).82,83 Actin is the most abundant intracellular protein. The cell motility, shape, and size depend on the ability of globular actin to polymerize into filaments (F-actin). Upon cell injury or cell necrosis, actin is released into extracellular space. However, when actin is released from cells, it may rapidly form filaments with detrimental effects for the microcirculation. Two plasma proteins, DBP and gelsolin, bind actin avidly, thereby acting as “actin-scavenger” system.^84,85

DBP-null (KO) mice, however, develop normally. They are nevertheless more sensitive to vitamin D deficiency and less sensitive to vitamin D excess, probably by an enhanced urinary loss of vitamin D metabolites.⁷⁴ The DBP and megalin KO mice, however, suggest that the main function of DBP is indeed to transport all vitamin D metabolites and preserve them from rapid clearance or urinary loss.

Action and Mode of Action

General Characteristics of The Vitamin D Receptor

Protein

1,25(OH)₂D, the hormonally active form of vitamin D, exerts its effects mainly by activating the nuclear VDR, a member of the nuclear-receptor superfamily of ligand-activated transcription factors. Based on structure and function similarities between members of this family, different functional domains can be distinguished in these nuclear-receptor proteins. The short A/B domain at the N-terminus of VDR lacks the usual ligand-independent activation function (AF1). Two highly conserved zinc finger DNA binding motifs constitute the DNA-binding C domain, which also harbors the nuclear localization signal. The D domain or hinge region may regulate the receptor’s flexibility between DNA-binding and ligand-binding domains and may be crucial to allowing the heterodimer complex of the ligand-binding domains to interact with two differently oriented response elements (direct repeat or palindrome orientation with variable number of spacer nucleotides). The large multifunctional E region contains the ligand-binding domain, as well as a dimerization surface and a ligand-dependent activation function (AF2) at the extreme C-terminus, represented by helix 12.⁸⁶

Gene

The human VDR gene, consisting of 14 exons, spans more than 60 kb on chromosome 12.87,88 The major VDR transcript is a 4.8 kb mRNA species, but multiple promoters and alternative splicing give rise to a multitude of less abundant transcripts that mostly vary in their 5′ untranslated region but encode the same 427-amino-acid protein.⁸⁸ However, two of these mRNAs are translated into VDR proteins that contain an additional 23 or 50 amino acids at the N-terminus.⁸⁸

Genomic Actions

Binding of 1,25(OH)₂D to VDR generates conformational changes of VDR followed by heterodimerization with unliganded RXR and binding to vitamin D response elements (VDREs) in the promoter region of vitamin D target genes, with subsequent release of corepressors and recruitment of coactivators and general transcription factors for the assembly of an active transcriptional complex.⁸⁹ A putative crucial event in this respect is the mousetrap-like intramolecular folding of helix 12, closing off the ligand-binding pocket and exposing the AF2 domain for interaction with coactivators.⁹⁰ Corepressors bind and silence unliganded steroid receptors by recruitment of histone deacetylases, maintaining chromatin in a transcriptional repressive state.⁹¹ Coactivators are a group of proteins that allow gene transcription in several waves of activities. First, coactivators of the CBP/p300 family and of the p160 protein family, including the steroid receptor coactivators (SRCs), are recruited.^92–94 These proteins possess intrinsic histone acetyltransferase (HAT) activity and by acetylating histone tails open up the chromatin structure, creating a chromatin environment permissive for gene transcription.⁹⁵ In a second wave, the vitamin D receptor interacting protein (DRIP) multimeric complex is recruited, followed by recruitment of basal transcription factors, as well as RNA polymerase II. Finally, target gene transcription is induced.⁹⁶ Gene expression can also be mediated by ATP-dependent chromatin remodeling complexes such as SWI/SNF-type and ISWI-type complexes and the multiprotein complex WINAC.^97–99 Distinct regulation of transcriptional coregulators may provide species-specific, tissue-specific, or developmental stage–specific regulation of nuclear receptor function.¹⁰⁰ Furthermore, the expression or the recruitment of these regulatory proteins is regulated by several intracellular signaling pathways¹⁰¹ and by steroids themselves,¹⁰⁰ with receptor agonists or antagonists inducing preferential recruitment of coactivators or corepressors, respectively.¹⁰¹

A hexanucleotide direct repeat spaced by three nucleotides (DR3) is the cognate vitamin D response element (VDRE) to which RXR and VDR bind the 5′ and 3′ half-site, respectively, although alternative options appear to be possible, both with respect to dimer formation (VDR/VDR, VDR/RAR) and target-gene VDRE structure (DR4, DR6; IP9).¹⁰²

Nongenomic Actions

Aside from the VDR transcriptional or genomic effects, several research groups have described rapid effects by 1,25(OH)₂D that are independent of transcription and would be mediated by a membrane receptor for 1,25(OH)₂D or by the localization of the nuclear VDR near the membrane.¹⁰³ These so-called nongenomic effects include the opening of calcium or chloride channels and the activation of second messenger signaling pathways (phosphoinositide turnover, activation of protein kinase C, and the Ras/Raf/ERK/MAPK pathway). A wide variety of rapid and transient modifications in the second messenger signaling system have also been observed for other steroid hormones.¹⁰⁴ At the tissue or cellular level, however, nongenomic activity of vitamin D and its analogs or metabolites have only been described for intestinal calcium absorption (transcaltachia) or cellular differentiation of leukemia cells.^105–107 This pathway seems to prefer 6-s-cis to the 6-s-trans configuration of vitamin D.¹⁰⁵ Moreover, the agonist/antagonist specificity differs for that of the genomic pathway.¹⁰⁸

Classic Target Tissues

The action of 1,25(OH)₂D on bone, intestine, kidney, and parathyroid glands and its role in mineral metabolism is the result of a complex interplay between calcium and phosphate 1,25(OH)₂D, PTH, and phosphatonins. PTH induces calcium mobilization from bone and stimulates 1,25(OH)₂D production, but its secretion is inhibited by the action of 1,25(OH)₂D on the parathyroid glands (negative feedback). In a second negative feedback loop, 1,25(OH)₂D limits its own availability by inhibition of 1α-hydroxylase and stimulation of 24-hydroxylase, inducing 1,25(OH)₂D catabolism. In the last few years, considerable progress has been made in the understanding of phosphate homeostasis.¹⁰⁹ The phosphaturic hormone phosphatonin, or FGF-23, is produced by osteocytes and osteoblasts and inhibits the activity of the NPT2 protein. The NPT2 gene encodes a renal sodium/phosphate cotransporter responsible for reabsorption of phosphate and represents a newly identified target gene for 1,25(OH)₂D.¹¹⁰ Phosphatonin can be indirectly inactivated by a protease encoded by the PEX gene, which was identified as the gene that is defective in X-linked hypophosphatemic rickets. FGF-23 secretion is stimulated by 1,25(OH)₂D and impairs renal 1α-hydroxylase, creating an additional feedback system so that the production of 1,25(OH)₂D is tightly feedback regulated (Fig. 3-3).

FIGURE 3-3 Feedback regulation of renal synthesis of 1,25(OH)₂D. A, Regulation of renal 1,25(OH)₂D synthesis by parathyroid hormone and calcium, with multiple feedback control mechanisms. CaSR, Calcium-sensing receptor. B, Regulation of renal 1,25(OH)₂D synthesis by FGF23 and phosphate, with several feedback control mechanisms.

Effects on Intestine

The absorption capacity of calcium along the gastrointestinal tract of the rat is dependent on the segment and follows the order ileum > jejunum > duodenum. The efficiency of the small intestine to absorb dietary calcium is increased by 1,25(OH)₂D,¹¹¹ and the abundance of the vitamin D receptor is highest in the duodenum, followed by jejunum and ileum. Although the exact mechanism by which 1,25(OH)₂D alters the flux of calcium across the intestinal absorptive cell is not known, 1,25(OH)₂D increases the production and activity of several proteins in the small intestine, including TRPV6 and V5, calbindin-D9K, alkaline phosphatase, and low-affinity Ca-ATPase (PMCA). The entry of Ca²⁺ from the intestinal lumen across the brush border membrane into the enterocyte is mainly regulated by the epithelial channels TRPV6 and V5.¹¹² The intracellular calcium transfer is considered to be dependent mainly on calbindin-D9K.¹¹³ The transfer of Ca²⁺ from the cytoplasm to the extracellular space requires energy input because of an uphill concentration gradient and an unfavorable electrochemical gradient. Both the plasma membrane calcium pump and a sodium-calcium exchanger play important roles in this process. The stimulatory effect of 1,25(OH)₂D on the ATP-dependent uptake of Ca²⁺ at the basolateral membrane involves an increase in PMCA gene expression.¹¹⁴ The essential role of the intestine for calcium and phosphate homeostasis was clearly demonstrated by the phenotype of VDR KO mice. Such VDR-null mice are phenotypically normal at birth, but after weaning, they develop hypocalcemia, secondary hyperparathyroidism, and hypophosphatemia despite very high levels of 1,25(OH)₂D. They become growth retarded and develop severe rickets.^{66,112,115,116} Mice deficient in 1α-hydroxylase display a similar phenotype.^117,118 This bone and calcium phenotype can be largely corrected by a high dietary calcium intake (especially in combination with high lactose intake) in both knockout models or 1,25(OH)₂D treatment of 1α hydroxylase–null mice.^{112,119–125} These findings confirm previous observations in humans.^59,126,127 The data strongly suggest that the intestine is the primary target for 1,25(OH)₂D’s action on calcium/bone homeostasis. This is largely confirmed by genetic mouse models of selective rescue or deletion of VDR in the intestine of transgenic mice.¹²⁸ The primary molecular targets, however, merit further exploration; ablation of CaBP-9k or TRPV6 or even their combined deficiency have shown no major effects on basal intestinal calcium absorption or serum calcium levels when calcium intake is normal.¹²⁹ Paracellular intestinal calcium transport may also be part of the picture of vitamin D’s action in that the expression of claudin 2 and claudin 12, both known to form paracellular calcium channels, are induced by 1,25(OH)₂D and decreased in the intestine of VDR-null mice.¹³⁰

Effects on Kidney

The kidney is important both for the metabolism of 1,25(OH)₂D and the reabsorption of calcium and phosphate, processes regulated by 1,25(OH)₂D. The kidney and more specifically the proximal tubule is the central tissue for 1α-hydroxylation of 25(OH)D. Chronic renal failure reduces 1α-hydroxylase activity, which ultimately results in renal osteodystrophy or uremic bone disease. 1,25(OH)₂D also increased the distal tubular reabsorption of calcium; as in the intestine, TRP channels (now TRPV5), calbindin-D9K and 28K, and the plasma membrane calcium ATPase are involved. Whereas in the intestine, active calcium absorption in the duodenum takes place before the less-regulated diffusion process in the ileum, reabsorption of filtered calcium follows a more logical sequence of massive calcium-sodium reabsorption in the proximal convoluted tubuli, followed by specific, actively regulated calcium reabsorption in the distal parts of the nephron. The crucial role of 1,25(OH)₂D-regulated renal calcium reabsorption was demonstrated by persistent hypercalciuria and reduction in bone mass in TRPV5-deficient mice.¹³¹ The kidney is also the major component in phosphate homeostasis, as both PTH and FGF-23, in complex interplay with 1,25(OH)₂D, are able to reduce renal phosphate reabsorption (see Fig. 3-3).

Effects on Bone

1,25(OH)₂D has dual effects on bone: it can stimulate osteoclastogenesis and bone resorption as well as modify osteoblast function and bone mineralization. The overall effects of vitamin D metabolites on bone are thus extremely complex. From observations in man and animals, it is clear that vitamin D deficiency or resistance impairs bone-matrix mineralization, whereas osteoblast activity and matrix synthesis are even stimulated. Excess 1,25(OH)₂D can clearly enhance osteoclastogenesis and bone resorption (Fig. 3-4). Because bone mineralization and bone structure can be largely normalized in vitamin D- or 1,25(OH)₂D-deficient or resistant mice by sufficient supply of minerals via active or passive intestinal calcium absorption, it seems that direct effects of vitamin D metabolites on chondrocytes and bone cells are redundant if calcium and phosphate supply are guaranteed. However, most of the genes and proteins typically expressed in osteoblasts and osteoclasts are vitamin D regulated, so it is likely that 1,25(OH)₂D can fine-tune bone mineral homeostasis. Moreover, pharmacologic use of vitamin D metabolites or analogs might positively influence bone balance, as shown by human and animal experiments^132,133 and transgenic mice overexpressing osteoblast VDR.¹³⁴

FIGURE 3-4 Effect of 1,25(OH)₂D on bone cells and osteoclastogenesis. In osteoblast/stromal cells, 1,25(OH)₂D induces ODF expression, down-regulates OPG, and stimulates M-CSF production. It also stimulates production of IL-6 and IL-11, which represent distinct signals in osteoclastogenesis. On osteoclast precursors, 1,25(OH)₂D induces the expression of RANK (or ODF receptor) and several osteoclast differentiation markers, such as the vitronectin receptor αvβ3 and carbonic anhydrase-II. CA-II, Carbonic anhydrase-II; c-fms, M-CSF receptor; M-CSF, macrophage colony-stimulating factor; ODF, osteoclast differentiation factor; OPG, osteoprotegerin; V-ATP-ase, vacuolar adenosine triphosphatase.

Effects on Growth Plate

The absence of VDR or 1α-hydroxylase creates no detectable phenotype in overall growth or growth plate of prenatal animals, but the longitudinal growth of long bones is impaired after weaning. X-ray analysis reveals advanced rickets, including widening of the epiphyseal growth plate, with an increased width and marked disorganization of the growth plate on histology, including impaired mineralization of hypertrophic chondrocytes.66,115–118 This increased growth-plate width in VDR or 1α hydroxylase–null mice cannot be explained by their (normal) chondrocyte proliferation and differentiation, including collagen X and osteopontin expression. The expansion of the growth plate can be largely explained by decreased apoptosis of hypertrophic chondrocytes.¹³⁵ Based on analysis of several genetic models with abnormal phosphate homeostasis, serum phosphate levels are probably crucial for hypertrophic chondrocyte apoptosis in vivo. This is confirmed in vitro: apoptosis of hypertrophic chondrocytes is regulated by phosphate levels via the activation of the caspase-9-mediated mitochondrial pathway.¹³⁶

In accordance with these findings, chondrocyte-specific inactivation of the VDR did not cause a growth-plate phenotype and certainly not rickets.¹³⁷ Critical analysis of these mice, however, revealed that VDR action in chondrocytes regulates bone development and phosphate homeostasis by inducing expression of paracrine factors such as vascular endothelial growth factor and receptor activator of nuclear factor κB (NFκB) ligand expression, leading to impaired vascular invasion and decreased osteoclast number in the metaphysic, as well as increased bone mass of long bones of juvenile chondrocyte-specific VDR-null mice. In addition, FGF-23 expression in osteoblasts was decreased, probably linked to the increased gene expression profile of NPT2 and 1α-hydroxylase in the kidney and resulting in increased serum levels of phosphate and 1,25(OH)₂D.¹³⁷

Noncalcemic or Nonclassic Actions of Vitamin D Endocrine System

The virtual ubiquitous expression of the VDR in all nucleated cells, the presence of a functional 1α-hydroxylase in at least 10 different tissues apart from the kidney, and the very large number of genes that are under direct or indirect control of 1,25(OH)₂D all point toward a more universal role for the vitamin D endocrine system than just regulation of calcium/phosphate/bone metabolism. This is not totally unexpected; most other ligands for nuclear receptors also have a very wide spectrum of activities such as androgens, estrogens, glucocorticoids, and retinoids.²⁴ Based on controlled observations in cells, tissues, and transgenic mice and on observational studies in humans, it seems that the functioning of nearly all major tissues or systems of the organism is modulated by vitamin D.

Skin

The combined presence of vitamin D production, 25-hydroxylase, 1α-hydroxylase, and VDR expression in the epidermis suggests the existence of a unique vitamin D intracrine system in which UVB-irradiated keratinocytes may supply their own needs for 1,25(OH)₂D. A role for vitamin D in epidermal homeostasis can also be expected from the prominent effects of vitamin D compounds on keratinocyte growth and differentiation.¹³⁸ The epidermal keratinocyte represents the major cell type in the epidermis and most likely the major cutaneous target cell for vitamin D, but many other cell types present in the epidermis are also vitamin D targets.

Based on studies of 1α hydroxylase–deficient mice, the repair of the essential barrier function of the skin is impaired in the absence of vitamin D action.¹³⁹ The major skin phenotype of both VDR-null mice and children with VDR mutations is, however, the development of total alopecia. Hair development at birth is normal, but hair loss starts after the first catagen and ultimately leads to alopecia totalis associated with large dermal cysts. The absence of alopecia in vitamin D–deficient WT mice or in mice with CYP27B1 mutations clearly suggests that the absence of receptor and not its ligand is the cause of the skin phenotype. Mutations in the NR corepressor, hairless, or keratinocyte-specific loss of interaction of Lefl with β-catenin (part of the Wnt signaling pathway) produce a strikingly similar alopecia. There is therefore little doubt that ligand-independent effects of the VDR are required for normal keratinocyte stem cell function.

As expected from animal studies, no clear skin disorders are linked to vitamin D deficiency or insufficiency in humans. The very same photons that can generate the photoconversion of 7-dehydrocholesterol into previtamin D are also able to cause DNA damage and, ultimately, photoaging and increase the risk for skin cancer, so exposure to the UVB or sunlight needed to produce vitamin D always involves a small but cumulative risk of skin damage. This risk is especially relevant for humans with a fair skin type (phototypes 1 and 2).¹⁴⁰ Although 1,25(OH)₂D is able to generate a strong photoprotective effect against UVB-mediated events in cultured keratinocytes,^141,142 the overall effect is negative.

Cell Proliferation and Cancer

Exposure to 1,25(OH)₂D of virtually all normal cells and even most malignant cells results in an accumulation in the G0/G1 phase of the cell cycle.143–145 This inhibition of cell proliferation involves a large number of mechanisms and genes, and the exact sequence of events between VDR-mediated transactivation of genes and the actual G0/G1 arrest is probably cell-type specific. A general downstream effect is the regulation of the E2F family of transcription factors, which act as master switch for a very large number of genes involved in cell-cycle progression. These EF factors are under the control of the retinoblastoma protein members (especially pocket proteins p107 and p130), and their phosphorylation state is regulated by cyclins and cyclin-dependent kinases (p18, p19, p21, or p27), many of which are regulated by 1,25(OH)₂D.^143,146 However, 1,25(OH)₂D may also inhibit cell growth by interfering with signaling pathways initiated by TGF-β, epidermal growth factor (IGF), prostaglandins,¹⁴⁷ and Wnt ligands,¹⁴⁸ as well as by intervening in other mitogenic signaling pathways (e.g., ERK/MAPK pathway and c-myc) (Fig. 3-5).^149–156 Moreover, 1,25(OH)₂D can regulate apoptosis and angiogenesis, mechanisms well known to be important for cancer cell expansion. In view of these well-established in vitro effects, one might expect a greater sensitivity for carcinogenesis in VDR-null mice. Epidermal, mammary, and intestinal cells of such animals do indeed show signs of hyperproliferation. Moreover, when exposed to chemocarcinogens or oncogens, VDR-null mice develop more mammary cancer–type lesions, skin tumors, and lymphomas.^157,158

FIGURE 3-5 Effect of 1,25(OH)₂D on cell-cycle progression. 1,25(OH)₂D treatment leads to a cell cycle phase–specific effect characterized by an accumulation of cells in G₁ through modulation of different signaling pathways. (−), Inhibitory effect; (+), stimulatory effect; EGF, epidermal growth factor; IGF-1, insulin-like growth factor 1; PGE2, prostaglandin E2; pRb, retinoblastoma tumor-suppressor gene; TGF-β, transforming growth factor β.

In humans, absolute VDR or CYP27B1 deficiencies are rare, but vitamin D deficiency is highly frequent. This obviously raises the question whether such vitamin D deficiency is associated with increased risk of cancer in humans. Such a hypothesis was originally reinforced by observations of higher cancer prevalence in areas of the United States and Japan with lower UVB exposure. Serum concentrations of 25(OH)D are, of course, a much better indication of the real vitamin D status, and a vast literature links lower levels of 25(OH)D with higher prevalence of the major cancers, especially colon and breast cancer, with more mixed results for prostate cancer. The inverse association between colorectal cancer or breast cancer and serum 25(OH)D levels was confirmed by the results of the Third National Health and Nutrition Examination Survey (NHANES III),¹⁵⁹ although 25(OH)D levels were not related to overall cancer mortality. In most of the vast number of cross-sectional or observational studies, the higher cancer risk was found in subjects with serum 25(OH)D levels below 20 ng/mL, but most studies also revealed a significant trend across the different 25(OH)D subgroups and risk of cancer. Meta-analysis of studies addressing the association between 25(OH)D levels suggest that women with serum 25(OH)D of approximately 48 ng/mL (median of the top quintile) had a 50% lower risk of breast cancer than those with serum less than 13 ng/mL in the lowest quintile,¹⁶⁰ and that individuals with serum 25(OH)D levels greater than 32 ng/mL had a 50% lower incidence of colorectal cancer than those with relatively low levels (≤12 ng/mL).¹⁶¹ However, a number of studies also link higher vitamin D nutritional status with a higher prevalence or more aggressive type of cancer.^24,162

The final question is, of course, whether serum 25(OH)D is a predictor or has a causative relation with the overall cancer risk. Intervention studies should be able to provide the answer. In the Women’s Health Initiative (WHI) study, a significant inverse relationship was found between baseline levels of serum 25(OH)D and subsequent colorectal cancer incidence, but postmenopausal women receiving calcium (1 g) plus vitamin D (400 IU) did not develop less colon cancer than control patients.¹⁶³ In a much smaller 4-yr study in postmenopausal women, higher doses of calcium (1.4 to 1.5 g) and vitamin D (1100 IU), which raised serum 25(OH)D to mean levels above 80 nmol/L, did significantly reduce overall cancer risk.¹⁶⁴ The small number of cancer deaths and a major confounding factor of calcium intake, however, limit the value of this study to hypothesis, and much larger prospective studies with substantial vitamin D supplementation are essential.

Immune Function and Vitamin D

All immune cells (antigen-presenting cells, T and B cells, natural killer [NK] cells, and even mast cells) express at certain stages of their differentiation a functional VDR. Antigen-presenting cells (dendritic cells and equivalent resident cells, as well as monocytes/macrophages) can synthesize 1,25(OH)₂D using the same enzyme as in the kidney but controlled by immune stimuli instead of calciotropic hormones.52,165–167 Finally, 1,25(OH)₂D regulates a wide range of genes that play crucial roles in the immune system. The overall effects are different for the innate immune system (largely mediated by monocytes/macrophages) than for the acquired immune system. The innate immune system, upon exposure to bacterial agents, is first stimulated to produce 1,25(OH)₂D (Fig. 3-6A) and thereafter activated by this paracrine 1,25(OH)₂D to become a more active macrophage, including the local production of a number of defensins (including cathelicidin). Defects in immune functions indispensable for antimicrobial activity have been observed in vitamin D–deficient mice.^168,169 The overall effects suggest that 1,25(OH)₂D enhances the natural defense against bacterial infection. In the human situation, low serum levels of 25(OH)D were repeatedly associated with increased susceptibility to and more rapid disease progression of tuberculosis.^170–174 This hypothesis is confirmed by the lower induction of cyclic adenosine monophosphate (cAMP) by monocytes incubated with 25(OH)D-deficient serum from sunlight-deprived African Americans.¹⁷⁵ Moreover, oral administration of vitamin D markedly improved tuberculosis outcome in small-scale studies.^176,177 Prospective clinical trials are ongoing to evaluate the effect of vitamin D supplementation on the evolution of various infections, including tuberculosis, to confirm the cause/effect relationship between vitamin D status and the native immune defense system.

FIGURE 3-6 Immune effects of the vitamin D endocrine system. In the innate immune system, 1,25-(OH)₂D strengthens the antimicrobial function of monocytes/macrophages, for example, through enhanced expression of the cathelicidin antimicrobial peptide, eventually leading to better clearance of pathogenic microorganisms. In the acquired immune system, the immunomodulatory effects of 1,25(OH)₂D on players of the adaptive immune system can lead to the protection of target tissues in autoimmune diseases and transplantation. 1,25(OH)₂D inhibits the surface expression of MHC II–complexed antigen and of costimulatory molecules, as well as the production of the cytokine IL-12 in antigen-presenting cells (such as dendritic cells), thereby shifting the polarization of T cells from an (auto-)aggressive effector (Te) towards a protective or regulatory (Tr) phenotype. 1,25(OH)₂D also directly exerts its immunomodulatory effects at the level of T cells.

The acquired immune system reacts in opposite ways to the native immune system (see Fig. 3-6B). Indeed, 1,25(OH)₂D inhibits dendritic cell maturation and generates a coordinated action on T cell gene expression of key cytokines (IL-1, IL-2, IL-12, IL-17, INF-γ) and genes needed for antigen presentation to T cells (MHC class II and cosignaling proteins). The global effect of these immune-modulating actions is thus a down-regulation of the acquired immune system. This should have beneficial effects on the occurrence or evolution of autoimmune diseases. Vitamin D–deficient mice more easily develop a more severe type of such autoimmune diseases. For example, genetically predisposed NOD mice exposed to transient vitamin D deficiency early in life have a much higher incidence of type 1 diabetes than vitamin D–replete mice,^169,178 and 1α hydroxylase–deficient mice are more prone to several types of inflammatory bowel disease.¹⁷⁹ Moreover, 1,25(OH)₂D and more potent and selective analogs can significantly reduce spontaneous or experimental autoimmune diseases in rodents, such as type 1 diabetes in NOD mice, experimental allergic encephalitis, nephritis, or inflammatory bowel disease in rodents.^16,24,180

Studies in human autoimmune diseases are, however, more complicated. VDR polymorphism is not clearly related to the risk for type 1 diabetes, according to a large meta-analysis.¹⁸¹ However, polymorphism of the 1α-hydroxylase gene is associated with the risk of type 1 diabetes in cross-sectional as well as in family studies.¹⁸² Consistent with data from NOD mice, several epidemiologic studies in humans report that vitamin D intake in early life may reduce later risk of type 1 diabetes. Risk reduction varied between 26% with cod liver oil to 78% with 2000 IU/d, with an overall effect of 30% reduction in five published reports.¹⁸⁰ Since 25(OH)D serum levels have not been measured in the cohorts of children, a threshold cannot be defined for optimal reduction of type 1 diabetes.

A low vitamin D status has also been repeatedly associated with a higher risk for multiple sclerosis. A large prospective, nested, case-control study among more than 7 million U.S. military personnel revealed that low 25(OH)D level was a strong risk factor for later occurrence of multiple sclerosis (odds ratio of about 2 for serum 25[OH]D <20 ng/mL, with possibly even greater “protection” by higher levels).¹⁸³ However, as for type 1 diabetes, no controlled, randomized intervention studies have yet proved a causal relation between vitamin D (deficiency or insufficiency) and later occurrence of autoimmune diseases, but randomized trials (especially in genetically at-risk groups) should receive high priority. All these observations, however, suggest that preventing vitamin D deficiency in the perinatal period, early childhood, or adolescence may have long-lasting effects on autoimmune diseases. Moreover, there are several observations in mouse models that could explain such effects, such as the increased apoptosis of (autoreactive?) thymocytes or lymphocytes, the generation of regulatory T cells, and NKT cells after exposure to 1,25(OH)₂D.^180,184

Cardiovascular System

VDR-null mice, as well as 1α hydroxylase–null mice, develop high-renin hypertension and cardiac hypertrophy that can be prevented by treatment with an angiotensin blocker.¹⁸⁵ In vitro or in vivo exposure to 1,25(OH)₂D decreases renin production, probably by direct regulation of the gene expression via a VDRE in the promoter of renin. Observational studies in humans found an inverse association between 1,25(OH)₂D levels and blood pressure or plasma renin levels in normotensive or hypertensive individuals. Prospective cohort studies reported that incident hypertension over a 4-year follow-up is lowest when the serum 25(OH)D is 30 ng/mL or higher.¹⁸⁶ Also, in the NHANES III population, a significant negative relation between serum 25(OH)D concentrations and systolic, diastolic, and pulse pressure among the total adult population was observed.¹⁸⁷ The results of these observational studies are supported by two randomized controlled trials in which vitamin D treatment reduced blood pressure in hypertensive subjects and elderly community-dwelling women.

VDR-null mice display increased thrombogenicity and decreased fibrinolysis when exposed to inflammatory stimuli,¹⁸⁸ whereas 1,25(OH)₂D has beneficial effects on most cells of the vascular wall. In humans, a low vitamin D status is associated with a number of cardiovascular risk factors,¹⁸⁹ including the metabolic syndrome. The largest prospective intervention trial (WHI) with calcium (1g/d) and vitamin D (400 IU/d), however, revealed no increased nor decreased coronary or cerebrovascular risk after 7 years of follow-up.¹⁹⁰ Several large-scale observational studies demonstrated that survival and especially cardiovascular events were lower in patients on chronic renal replacement therapy treated with 1,25(OH)₂D analogs, compared with either untreated or 1,25(OH)₂D-treated groups.¹⁹¹ However, a large meta-analysis casts doubt on this conclusion.¹⁹² In addition, vitamin D excess can have deleterious effects on all structures of the vascular wall, with ectopic calcification and organ failure of kidney, cardiac valves, myocardium, and most other soft tissues. These data suggest a beneficial effect of the vitamin endocrine system (within specific optimal limits) on cardiovascular targets but certainly need confirmation by a proper prospective, large-scale randomized trial.

Muscle and Muscle Function

VDR is expressed in myoblasts and is also present in low concentrations in mature striated muscle cells. VDR-null mice, even on a high calcium diet, show maturation problems of their muscle fibers, with smaller muscle fibers and expression of embryonic markers even after weaning.¹⁹³ Genes that are typically expressed early in life (e.g., myf-5) are under negative control by 1,25(OH)₂D. Evaluation of muscle performance in VDR-resistant or vitamin D–deficient mice is difficult to interpret, because hypocalcemia may have major effects on calcium fluxes in muscle cells. Patients with chronic renal failure and vitamin D deficiency (thus combined deficiency of 25[OH]D and 1,25[OH]₂D) can develop severe myopathy and inability to walk that can be promptly restored by appropriate vitamin D and/or analog treatment. Sarcopenia (progressive loss of muscle mass and strength) is highly prevalent in the elderly and frequently associated with vitamin D deficiency. Vitamin D supplementation can modestly and inconsistently improve muscle function and improve body sway. Meta-analysis of several prospective intervention studies revealed that vitamin D supplementation in vitamin D–deficient elderly subjects can modestly reduce the risk of falls;¹⁹⁴ this may explain, together with beneficial effects on bone, a reduced fracture risk.¹⁹⁵

Glucose and Energy Metabolism

Several tissues which are important for energy and glucose metabolism, such as endocrine β cells, muscle, and fat cells, are also targets for vitamin D, so the obvious question is whether vitamin regulates or modulates overall metabolism apart from the effects of vitamin D on the immune system and autoimmune diabetes. Vitamin D deficiency in experimental animals (rodents and rabbits) impairs glucose tolerance.180,196,197 VDR-null mice, however, did not have a consistently abnormal glucose tolerance; one strain differs from another strain.^24,180 This discrepancy between effects of ligand and receptor deficiency is not unique (see alopecia and immune effect) nor specific for vitamin D. Similar effects are known for thyroid hormone/thyroid receptor function. 1,25(OH)₂D has modest stimulatory effects on insulin production and secretion, probably mediated by the well-known effects of calcium on β-cell functions.

Most observational studies in humans link vitamin D insufficiency with nearly all aspects of the metabolic syndrome—including obesity, insulin resistance, fasting blood glucose or type 2 diabetes, hypertension, and hyperlipidemia.24,180 This was confirmed in the NHANES study¹⁹⁸ and a Scandinavian cohort study,¹⁹⁹ as well as in a prospective British cohort study.²⁰⁰ Baseline 25(OH)D levels were inversely associated with a 10-year risk of fasting hyperglycemia and insulin resistance, with the greatest risk in subjects with 25(OH)D levels below 20 ng/mL and the lowest risk in subjects with above 30 ng/mL.²⁰⁰ Short-term interventional studies to correct severe vitamin D deficiency in a few subjects indicate improved glycemic control, but larger studies using 400 IU/d, such as in the WHI trial, could not demonstrate an effect on glucose levels.²⁰¹ A much smaller study was also negative, but post hoc analysis revealed a modest effect of a higher dose of vitamin D (700 IU/d) on subjects with fasting hyperglycemia at randomization.²⁰² In view of the high prevalence of vitamin D insufficiency and of metabolic syndrome and their association, it is highly desirable to demonstrate a causal link and then take appropriate actions.

Mortality

Because vitamin D status is associated with so many major diseases, it is worthwhile to explore whether vitamin D deficiency is associated with increased mortality. Indeed, in a prospective 8-year cohort study, all-cause and cardiovascular mortality was twofold higher when 25(OH)D levels were in the lower 2 quartiles (<17 ng/mL), compared with the upper quartile.²⁰³ Similar results were obtained from the NHANES III study: mortality was about 1.5-fold higher when 25(OH)D levels were well below 20 ng/mL, and the lowest mortality was observed in subjects with 25(OH)D levels between 20 and 50 ng/mL.²⁰⁴ Also, in patients on chronic hemodialysis, all-cause mortality was greater (significant 1.6-fold increase) when 25(OH)D levels fell below 10 ng/mL.²⁰⁵ In addition, a meta-analysis of all studies using vitamin D supplementation for the prevention of fractures revealed a modest but significant (−8%) reduction in overall mortality in the vitamin D supplemented groups.²⁰⁶

Diagnostic and Therapeutic Aspects of Vitamin D

Assays For Vitamin D and Metabolites: Methodology and Applications

Vitamin D and about 30 of its metabolites are found in plasma. Measurements of their concentrations may be essential for clinical or research purposes.207,208 Most techniques require a lipid extraction to free these compounds from their binding proteins (especially DBP). In view of the high molar extinction of vitamin D, UV absorptiometry can be used for measurement of vitamin D₂, vitamin D₃, or 25(OH)D after high-performance liquid chromatography (HPLC). However, competitive-binding assays or radioimmunoassay (RIA) are preferred for measurements of 25(OH)D, 1,25(OH)₂D, or 24,25(OH)₂D.²⁰⁹ These assays remain difficult, as demonstrated by remarkably poor intralaboratory and especially interlaboratory quality-control studies.^210,211 Non-chromatographic assays using DBP overestimate the true 25(OH)D concentration by 10% to 20%, but nonspecific interferences (DBP, lipids?) in some assays can result in up to 100% higher values. The quality control of routine 25(OH)D assays revealed extreme problems with accuracy,²¹⁰ and since the definition of vitamin D deficiency or insufficiency is defined in absolute concentration (see later), there is urgent need for improved quality assurance.²¹⁰ The use of tandem mass spectrometry after liquid chromatography and serum extraction is now considered the gold standard, but this assay is not yet routinely available for purposes other than research.²¹²

The measurement of serum concentrations of vitamin D₂ or D₃ is of little clinical value. Indeed, because of their short half-life in plasma, it reflects only recent exposure to UV light or nutritional intake. Serum 25(OH)D concentration is, however, an excellent reflection of the vitamin D status because of the rapid conversion of vitamin D into 25(OH)D and its long plasma half-life.208,209 Its plasma concentration varies widely in normal subjects because of large variations in endogenous and exogenous supply of vitamin D (Table 3-3).

Table 3-3

Plasma Concentration of 25(OH)D

Normal Fluctuation According to:

Dietary intake (+)*

Sun (UV light) exposure (+) influenced by seasonal life style and cultural habits

Age (−)

Skin pigmentation (−)

Latitude (−)

Sunscreen use (−)

Increased 25(OH)D Concentration

Exposure to pharmaceutical vitamin D^†

Excess exposure to nutritional vitamin D

Excess exposure to UV light

Decreased 25(OH)D Concentration^‡

Combined deficiency of access/exposure to nutritional vitamin D and UV light

Major risk groups include:

Infants, especially when born in late winter

Women and children of immigrants with pigmented skin living in temperate climates

Elderly population with limited mobility

Subset of population with low exposure to sunshine because of socioeconomic, religious, or cultural reasons

Decreased intestinal absorption of vitamin D associated with fat malabsorption (e.g., associated with biliary cirrhosis)

Short bowel syndrome

Exocrine pancreas insufficiency

Gluten enteropathy

Increased loss or catabolism of vitamin D

Nephrotic syndrome

Chronic liver P450 activation by drugs (e.g., barbiturates or antiepileptic drugs)

Low calcium intake or absorption

^*Positive or negative effects are indicated by + or −, respectively.

^†Vitamin D toxicity with hypercalciuria, hypercalcemia, nephrocalcinosis, kidney stones, metastatic calcification, etc., are only observed if 25(OH)D concentrations exceed 100 ng/mL. Without access to pharmaceutical vitamin D, it is therefore unlikely to acquire clinical vitamin D toxicity.

^‡For definition of vitamin D insufficiency or deficiency, see chapter-recommended daily intake and Table 3-5.

Plasma 25(OH)D concentration thus behaves as a true vitamin whose concentration depends on nutritional supply or synthesis in the skin after exposure to UVB light. Low exposure to UV light and low vitamin D intake is quite common in infants and elderly subjects if food sources are not supplemented with vitamin D. Plasma 25(OH)D concentrations are indeed low at birth (about half the maternal concentration because of the 2:1 ratio of maternal to fetal DBP concentration), and the natural vitamin D content of milk is low. Sun exposure was therefore evolution’s solution to prevent rickets. However, in view of the relation between exposure to UVB light (especially in young children) and subsequent risk for skin malignancies, it is probably wise to advocate systematic vitamin D supplementation of all infants and young children. Whereas widespread vitamin D deficiency in infants was recognized and prevented in the beginning of the 20th century, a similar endemic deficiency in the elderly was only recognized and addressed at the end of the same century (see Treatment, later). Intestinal malabsorption of fat-soluble vitamins interrupts the absorption of exogenous (probably also endogenous) hepatobiliary excretion of vitamin D and therefore requires either substitution with large amounts of vitamin D or more physiologic doses (10 to 20 µg/d) of the more soluble 25(OH)D. A low calcium intake can markedly (twofold) increase the catabolism of 25(OH)D and will thus facilitate substrate deficiency if the “nutritional” supply is marginal.²¹³

The metabolism of 25(OH)D into 1,25(OH)₂D and 24,25(OH)₂D is tightly controlled by hormones, ions, and humoral factors. The plasma concentration of 1,25(OH)₂D is therefore regulated as a true hormone (Table 3-4), and measurements of its concentration can be useful for clinical exploration of unusual cases of rickets, osteopenia, and hypo- or hypercalcemia.

Table 3-4

Plasma Concentration of 1,25(OH)₂D

DBP, Vitamin D-binding protein.

Monocyte activation can result in extrarenal synthesis of 1,25(OH)₂D, such as in sarcoidosis, tuberculosis, foreign body inflammation, lymphopenia, and some fungal infections.

^*In many cases of rickets or osteomalacia, 1,25(OH)₂D concentrations are still measurable or even nearly normal. This may be due to recent (and insufficient) access to vitamin D after long-term vitamin D deficiency. Nevertheless, such concentration is too low in comparison with the degree of secondary hyperparathyroidism. In any way, 25(OH)D is a better marker for vitamin D deficiency than 1,25(OH)₂D. A similar situation is observed in hypothyroidism when precursor hormone T₄ is a better marker for clinical hypothyroidism than the real hormone, T₃.

^†Vitamin D excess only increases serum 1,25(OH)₂D when renal function remains normal and/or PTH secretion is elevated. Frequently, 1,25(OH)₂D levels are low or normal in vitamin D toxicity.

The serum concentration of 24,25(OH)₂D and 25,26(OH)₂D usually reflects the concentration of 25(OH)D and therefore does not contribute additional valuable clinical information. The 25(OH)D- and 1,25(OH)₂D-lactone concentrations are only increased in case of important substrate excess, but their measurement is not (yet) introduced in clinical practice.

All vitamin D metabolites are tightly bound to DBP. Since the hepatic 25-hydroxylase activity is not feedback regulated, the free (or total) 25(OH)D concentration is largely fluctuating according to substrate supply. In contrast, renal 25(OH)D-1α-hydroxylase is tightly controlled, and since the access to VDR in target tissues is dependent on the circulating free concentrations, free and not total 1,25(OH)₂D concentration is important.80,214 The circulating DBP concentration is fairly stable, except when stimulated by estrogens (or pregnancy) or decreased by reduced synthesis (liver cirrhosis) or increased urinary loss (nephrotic syndrome). The major arguments for the importance of free rather than total 1,25(OH)₂D are (1) in vitro experiments (biological activity of 1,25[OH]₂D on cultured cells)^215,216 and (2) in vivo observations such as increased steady-state concentration of 1,25(OH)₂D without signs of increased action during chronic estrogen use or in animals immunized against 1,25(OH)₂D-hapten-protein complex.²¹⁷

Clinical Aspects of Vitamin D

In contrast to some rare inborn diseases related to vitamin D production, metabolism, or action (see Chapter 11), vitamin D deficiency and insufficiency are extremely frequent worldwide, and the full scope of their prevalence has only recently been appreciated.^16,24,32 Vitamin D excess is also a serious disease but seems to occur rarely without the context of excess intake of vitamin D supplementation and is thus usually iatrogenic.

Before the discovery of the dual origin of vitamin D and the introduction of vitamin D supplementation of infants and children, rickets was highly prevalent among the poor in many European cities but also in children of wealthy families (see Chapter 15). The optimal dosage was determined over time, largely on a trial-and-error basis, since this happened well before even the concept of randomized clinical trials was conceived. The vitamin D content of 1 teaspoon of cod liver oil (later found to contain the equivalent of 400 IU of vitamin D₃) was found to be efficient and sufficient to prevent endemic rickets. Subsequently, a dose of 200 to 400 IU/d was considered protective. The recommended treatment dose for a child with established rickets was usually larger, partly because of the need for a loading dose. Most recommendations by various official nutritional boards around the world also suggested, up to recently, a daily intake of 5 or 10 µg of vitamin D for infants and children. Indeed, the standing committee on the scientific evaluation of dietary reference intakes (by the U.S. National Academy of Sciences Panel for vitamin D in partnership with the Institute of Medicine) carefully evaluated vitamin D requirements and recommendations in 1997 and suggested 5 µg as the daily supplement (see Table 3-1),²¹⁸ with similar recommendation by the European Food and Safety Authority.²³ Only one real randomized trial with regard to preventive action of vitamin D was ever published and found no cases of new rickets in young Turkish children treated with 400 IU/d for 18 months.^219,220 The recommended daily supplement for children, long set at 5 µg or 200 IU/d, was increased in 2008 to 10 µg or 400 IU of vitamin D₃/d by the American Academy of Pediatrics.²²¹ No well-designed prospective or randomized trials have ever used serum levels of 25(OH)D to define the minimal threshold for preventing or curing rickets. In clinical case studies, serum 25(OH)D levels found in simple vitamin D–deficiency rickets are well below 10 ng/mL and usually even below 5 ng/mL. Serum 25(OH)D levels are frequently low in newborn sera (cord sera); the mean level in newborns is usually only slightly above 50% of the 25(OH)D levels in the mother’s serum.⁸⁰ Despite the cheap and effective strategies to prevent rickets, many countries or regions around the world are still facing endemic rickets affecting even a few percent of the infants or children,²²² especially in many Islamic countries, rural China, in children of immigrants in Western Europe, or in children born to mothers living in areas of high frequency of vitamin D insufficiency in general.²²³ The disease prevalence or severity may be aggravated by simultaneous poor dietary calcium intake, such as in many African countries.²²⁴ Poor vitamin D status in perinatal life or during childhood may also predispose to lower bone mass much later in life.²²⁵

Vitamin D is also known to be a major factor in maintaining bone integrity later in life, and most studies have dealt with the elderly or postmenopausal women. The idea for this role came from repeated and well-documented cross-sectional studies linking increasing PTH serum levels with increasing age and decreasing 25(OH)D levels. To define an optimal 25(OH)D level for optimal bone health, numerous studies have looked at surrogate endpoints (Table 3-5), but fortunately several intervention studies are now also available, including several meta-analyses of these data. Indeed, several surrogate endpoints have been evaluated to define a minimal or optimal threshold for 25(OH)D. Serum concentrations of 1,25(OH)₂D are not related to the substrate concentration if serum 25(OH)D exceeds 20 ng/mL in adults, but a positive relation with serum 25(OH)D has been observed when subjects with low 25(OH)D were included. Intervention studies reveal that 1,25(OH)₂D rapidly increases, even transiently above the normal level, when vitamin D supplementation is given to vitamin D–insufficient patients (25[OH]D levels < 12 to 20 ng/mL).²²⁶ Thus, it seems that the plasma level of 1,25(OH)₂D is no longer substrate dependent once 25(OH)D levels exceed 20 ng/mL. PTH has been very extensively used as a surrogate marker for defining optimal vitamin D status. Undoubtedly PTH increases in groups of patients with low 25(OH)D levels; however, PTH concentrations increase only to levels above the normal range when 25(OH)D is very low, whereas in about a third of patients with low 25(OH)D levels, PTH remains normal. The threshold of 25(OH)D below which serum PTH starts to increase varies between 12 and 40 ng/mL in a large number of cross-sectional studies.³² Such differences may partly be due to differences in accuracy of the 25(OH)D assay and differences in nutritional calcium intake and kidney function. Intervention studies are therefore more reliable. Serum PTH decreases after vitamin D supplementation when baseline levels are below 20 ng/mL.^32,227 Active intestinal calcium absorption is the primary target for vitamin D action and would thus represent an ideal surrogate endpoint for defining 25(OH)D levels. However, measuring intestinal calcium absorption is difficult because only dual-isotope techniques allow accurate estimations. Other methods, such as plasma concentrations of a calcium isotope after oral intake, provide a more crude estimation, whereas change in total serum calcium concentrations after a large oral calcium loading is a very poor estimation of active calcium absorption. Cross-sectional data suggested a minimal 25(OH)D level of 32 ng/mL for optimal calcium absorption but used a poor procedure to measure calcium absorption.^228,229 Other large cross-sectional studies could not identify a true 25(OH)D threshold but only a relation with serum 1,25(OH)₂D, either in adults²³⁰ or adolescents.²³¹ Again, intervention studies with vitamin D supplementation revealed no²³² or only a minimal²³³ increase in active intestinal calcium absorption in subjects with baseline 25(OH)D levels above 20 ng/mL. Cross-sectional data on 25(OH)D and bone mineral density (BMD) values could not demonstrate a strong correlation; this may be due to a long lag time between vitamin D intake and bone turnover or bone mass. Lower BMD levels, however, were observed in subjects with the lowest 25(OH)D levels (<12 ng/mL).^32,234 Fracture prevalence in cross-sectional or prospective studies according to 25(OH)D levels revealed that a higher fracture risk is associated with single point measurements of 25(OH)D levels below 20 ng/mL.²³⁵ Finally, intervention studies with vitamin D and/or calcium supplements are most relevant to define the optimal vitamin D status. A very large number of studies have addressed this question, but only a limited number can be classified as well-designed randomized trials. Several meta-analyses came to slightly different conclusions. Vitamin D alone given to postmenopausal or elderly subjects cannot reliably reduce hip-fracture incidence,^219,236 and oral calcium supplementation alone also has no clear benefit on hip-fracture risk, with one study even revealing an increased risk.^194,237 Combined vitamin D and calcium supplementation, however, can reduce hip-fracture risk by about 20%, with a similar reduction on other nonvertebral fractures.^236,238 In these studies, a vitamin D dosage of 800 IU/d was more efficient than 400 IU/d.¹⁹⁵

Table 3-5

Strategies and Clinical Studies to Define Optimal Vitamin D Status for Bone Health

Hard Endpoints

Placebo-controlled intervention studies

Vitamin D/25(OH)D and fractures

Prospective/cross-sectional studies

25(OH)D and fractures

Surrogate Endpoints

Prospective/cross-sectional studies

25(OH)D and BMC/BMD

25(OH)D and bone turnover markers

Prospective/cross-sectional studies

25(OH)D and calcium absorption

Cross-sectional/intervention studies

25(OH)D and PTH

Cross-sectional/intervention studies

25(OH)D and 1,25(OH)₂D

BMC, Bone mineral content; BMD, bone mineral density; PTH, parathyroid hormone.

It thus seems that only combined high calcium intake (>1 g/d) and vitamin D supplementation (≥800 IU/d) has shown to be efficient for fracture reduction in target populations of elderly subjects, with the greatest effect in institutionalized patients. The corresponding 25(OH)D level is more disputed, mainly owing to lack of accuracy of 25(OH)D assays in older studies and the confusion about optimal (minimal) 25(OH)D in individual subjects and mean population levels. Because most intervention studies conclude that serum 25(OH)D levels increase by 1 ng/mL for each additional 100 IU of vitamin D supplement per day,²³⁹ serum 25(OH)D levels (calculated from baseline population level and expected increase of 8 ng/mL for a 800 IU/d supplement) reached during the vitamin D intervention studies with positive fracture effects can be estimated as over 20 ng/mL in most adults, with mean levels closer to 30 ng/mL.²⁴⁰ The amount of vitamin D needed to obtain a minimal 25(OH)D level above 20 ng/mL and a population mean closer to 30 ng/mL, of course, depends on the baseline 25(OH)D level and accesses to UVB and dietary vitamin D. Intervention studies, however, revealed that 400 IU/d is not sufficient to raise 25(OH)D above 20 ng/mL in more than 95% of the target population of postmenopausal or elderly Caucasians, whereas this can be better achieved by 800 IU/d.

Whereas a large body of literature and meta-analyses are available to define the relationship between vitamin D supplementation, 25(OH)D concentration, and future fracture risk, such data are far less available for defining optimal 25(OH)D levels and noncalcemic endpoints. For muscle function and falls, vitamin D supplements of 700 to 1100 IU/d were found to be modestly efficient, and serum 25(OH)D levels obtained by such interventions would be very similar to minimal 25(OH)D levels that seem to be effective for fracture reduction. For immune and cardiovascular effects, as well as for potential effects on cancer risks, no, very limited, or controversial intervention data are available. Observational data reveal that the greatest risks are observed in subjects with baseline 25(OH)D levels below 20 ng/mL. However, most cross-sectional studies also reveal that higher 25(OH)D levels (>30 to 40 ng/mL) are associated with the greatest risk reduction for cancer, autoimmune diseases, and metabolic endpoints. It is unlikely that these higher 25(OH)D levels increases serum 1,25(OH)₂D above the levels when substrate 25(OH)D exceeds 20 ng/mL, so the working hypothesis is that such higher 25(OH)D levels would be needed to allow local paracrine production of 1,25(OH)₂D. Although this is plausible, direct proof for this concept is still unresolved.

What are the options for vitamin D supplementation based on the available information? For infants and children, a daily intake of 400 IU/d should be assured from early life till adolescence, since this can prevent vitamin D–deficiency rickets. Real-life implementation is far from ideal in at-risk groups in Western countries and in many regions of the world where exposure to sunlight is minimal for geographic or sociocultural reasons. For elderly subjects and probably also for all adults, the minimal 25(OH)D level associated with the lowest risk for fractures, falls (based on intervention studies), and a number of major diseases (based on epidemiologic surveys) is above 20 ng/mL. This can be achieved by increasing vitamin D intake by 800 IU/d or equivalent per week or month.²⁴¹ In populations with better 25(OH)D baseline levels, 400 IU/d can be sufficient, and in some countries or regions of the world with mean 25(OH)D levels of 30 ng/mL, no vitamin D supplements may be needed. Alternative sources of vitamin D such as fatty fish are not a practical solution for many millions of mildly vitamin D–deficient subjects. Higher exposure to UVB light could certainly improve the vitamin D status but cannot be recommended to subjects with a fair-skin phototype (phototypes 1, 2) because of lifetime risk of photodamage or skin cancer. Of course, for a number of patients with specific diseases (see Table 3-3) a higher dose of vitamin D or 25(OH)D or 1,25(OH)₂D is needed because of poor intestinal absorption, increased catabolism, or impaired metabolism.

There are a number of arguments that 25(OH)D levels over 30 to 40 ng/mL may provide additional benefits for bone, muscle, and noncalcemic endpoints. To reach such levels in over 97% of the population, daily supplements of at least 2000 IU/d (and usually more) would be needed, with or without substantially greater exposure to UVB.²²⁹ Since severe vitamin D toxicity is usually observed only when 25(OH)D levels exceed 100 ng/mL, such levels of 30 to 40 ng/mL are probably safe. However, randomized, large-scale, long-term studies with supplements of 2000 IU/d or more do not exist; such doses were only evaluated in a few hundred subjects for a maximum of 6 months.²⁴² As a reminder of possible toxicity, it is worth remembering that a mild but significant increase in kidney stones was observed in the WHI trial when calcium supplements were combined with at least 400 IU/d of vitamin D for 7 years.¹⁶³ A causal relationship between vitamin D status and more major diseases has still to be proven, so it seems wise to defer recommendations for a generalized vitamin D intake above the present upper limits. However, in view of the solid hypotheses generated by observational studies, appropriate randomized controlled trials with multiple end points deserve a great priority.

Worldwide Vitamin D Status

Serum 25(OH)D levels, as the best marker for vitamin D status, vary widely in different populations around the world, and the frequency of vitamin D deficiency or insufficiency (Table 3-6) varies accordingly. In an extensive meta-analysis of cross-sectional studies on serum 25(OH)D levels in healthy subjects around the world (394 studies), average serum 25(OH)D levels were 21 ng/mL. Caucasians had slightly higher levels than non-Caucasians,²⁴³ but this may be biased, since 25(OH)D levels are lower in subjects with darker skin when living in moderate climate zones. Older subjects (>75 yrs), as well as young children (<15 yr), had lower 25(OH)D levels. Latitude had only a minimal effect, demonstrating that apart from potential exposure to UVB light, many other factors—skin pigmentation, lifestyle, and nutritional factors—define vitamin D status. In more homogenous populations, the expected North-South gradient has been confirmed (e.g., in France),²⁴⁴ whereas for all European populations, the North-South gradient was reversed, probably because of high fish intake in Scandinavian countries and differences in sun-seeking behavior.^34,245 In North America, NHANES data confirm that 25(OH)D levels remain relatively stable over time, with mean levels of 30 ng/mL¹⁸⁷ and thus substantially higher than in Europe, probably related to the widespread use of vitamin D–enriched food. Non-Hispanic U.S. blacks had mean levels of 20 ng/mL. In some countries, however, mean levels can be quite low.^246,247 Lower levels are usually observed in obese subjects or those with special risk factors (see Table 3-3). Low levels are also frequent in pregnant women and their infants, and this poses an additional risk in view of the potential late consequences of perinatal vitamin D deficiency.²⁴⁸

Table 3-6

Vitamin D Nutritional Status as Described by Circulating Levels of 25(OH)D

Optimal vitamin D status defined by serum 25(OH)D levels.

^*Different opinions exist for defining the minimal threshold for optimal 25(OH)D levels; 30 ng/mL or 50 nmol/L is suggested by others.²⁷²

It is therefore obvious that mild vitamin D deficiency (see Table 3-6), even when defined conservatively as 25(OH)D levels under 20 ng/mL, is very widespread worldwide and can affect about one third to half of the world population. When insufficiency is defined by 25(OH)D levels below 30 ng/mL, then half of the healthy U.S. population (based on NHANES data) and about two thirds of the European population (and even more in most Muslim countries) would be vitamin D deficient.^16,249 It is therefore imperative that the health consequences of vitamin D status should be better evaluated and that the most severe forms of vitamin D deficiency or insufficiency (see Table 3-2) should be corrected by appropriate strategies.

Therapeutic Potential of 1,25(OH)₂D Analogs

The combined presence of 25(OH)D-1α-hydroxylase ⁴⁸ and VDR in several tissues introduced the concept of a paracrine role for 1,25(OH)₂D.²⁵⁰ These newly discovered functions of 1,25(OH)₂D create possible new therapeutic applications for immune modulation (e.g., for the treatment of autoimmune diseases or prevention of graft rejection), inhibition of cell proliferation (e.g., psoriasis), and induction of cell differentiation (cancer). To achieve growth inhibition or cell differentiation, supraphysiologic doses of 1,25(OH)₂D are needed, causing calcemic side effects. Therefore, new analogs of 1,25(OH)₂D have been developed to dissociate the antiproliferative and prodifferentiating effects from the calcemic and bone-metabolism effects.²⁵¹

The secosteroid 1,25(OH)₂D, with its open B ring and side chain of 8 carbon atoms, is a very flexible molecule. Different modifications have already been introduced in the A, B, C, and D rings and in the side chain by addition or transposition of hydroxyl groups, introducing unsaturation, replacing a carbon atom with a hetero atom, inverting the stereochemistry, and/or shortening or lengthening the side chain.²⁵¹ During the last decade, over 500 analogs were synthesized and their biological potency reported in the non-patent literature.^251,252 Moreover, several thousands of analogs have been synthesized by pharmaceutical and academic research groups and reported briefly in patent literature. Some of these analogs demonstrate a clear dissociation between antiproliferative and calcemic effects. In the meantime, nonsteroidal analogs were synthesized with a totally new structure lacking the full CD region of 1,25(OH)₂D.^253,254

No single or simple mechanism can explain the exact mechanism of superagonistic and selective activity profile (calcemic versus noncalcemic effects) of the new vitamin D analogs. Most of the biological effects of 1,25(OH)₂D are believed to be mediated via binding to the VDR, but surprisingly the binding affinity of the analogs to VDR does not always correlate with their potency. Some analogs extend the VDR half-life and induce different conformational changes to the VDR-ligand complex as assessed by limited proteolytic digestion and site-directed mutagenesis.255–257 However, most analogs bind to the ligand-binding pocket of the VDR without major modifications of the surface of the ligand-binding domain of the receptor.^252,254,258 Nevertheless, the superagonists are able to enhance gene transcription at the level of coactivator recruitment or activity. The selective action of analogs, however, requires different mechanisms, such as different metabolism in different target cells or cell- or gene-specific regulation, depending on the VDR (e.g., homo-heterodimer configuration induced by specific analogs, presence or selective interaction with coactivators or repressor protein, etc).^251,259,260 Some analogs are selective agonists or antagonists for nongenomic rapid actions while being devoid of significant genomic activity.¹⁰⁵ Since for other steroid hormones (e.g., estrogens, androgens, glucocorticoids), it is now well established that analog-specific gene regulation can be generated by chemical modification of the parent ligand molecule, it is likely that among the many powerful selective vitamin D analogs, at least some will be found to be clinically useful for noncalcemic indications.

Skin

The epidermis is a unique tissue in that it can fully produce and activate the full vitamin D synthesis and metabolism pathways and is very sensitive to 1,25(OH)₂D. Calcitriol at pharmacologic concentrations potently induces growth arrest and differentiation of the epidermal keratinocyte.¹³⁸ The profound effects of calcitriol on keratinocyte proliferation and differentiation have led to the application of vitamin D analogs for skin diseases with disturbed keratinocyte proliferation and differentiation, primarily psoriasis.²⁶⁴ Topical vitamin D analogs that display decreased calcemic activity (calcipotriol and tacalcitol) are now widely used for mild to moderate forms of psoriasis. Monotherapy with these vitamin D compounds achieves an equal effectiveness as topical medium-potency glucocorticoids, without a risk for skin atrophy. Mild irritation is the only frequently observed side effect.²⁶⁴ During treatment with vitamin D analogs, the abnormal epidermal homeostasis is fully restored: keratinocyte proliferation is inhibited; the perturbed psoriatic differentiation profile is normalized, with a decrease of the premature expression of involucrin and type I transglutaminase and enhancement of filaggrin expression;²⁶⁵ and the aberrant expression of cell adhesion molecules (integrins, ICAM1) also returns to normal.²⁶⁶ The concomitant decrease of the inflammatory infiltrate is, however, incomplete,^265,266 which may account for the residual redness of the lesions after completion of therapy. This can, however, be further improved by combination with topical usage of corticosteroids.

Immunology

The detection of VDR in almost all cells of the immune system, especially antigen-presenting cells (macrophages and dendritic cells) and activated T lymphocytes, led to the investigation of a potential for 1,25(OH)₂D as an immunomodulator.267–270 Vitamin D supplementation is now actively explored in several clinical trials for the stimulation of the natural immune defense and as adjuvant treatment of tuberculosis or other infectious diseases. Similar studies have started for the prevention of major autoimmune diseases. However, vitamin D analogs, though effective in animal models of autoimmunity, have not reached the stage of clinical trials in humans.