Work

Vitamin D and Health: Perspectives From Mice and Man

Year: 2008
Type: review
Abstract: The discovery of the dual origin of vitamin D either by synthesis in the skin from 7-dehydrocholesterol under the influence of UVB sunlight or from dietary origin quickly enabled the eradication of endemic rickets. It then took about one century to unravel the complex metabolism of vitamin D into a large number of metabolites and a key hormone, 1,25(OH)2D, that acts as a ligand for a nuclear transcription factor, VDR, which like all other nuclear receptors, regulates gene transcription and cell function. The vitamin D endocrine system undoubtedly plays a key role in calcium and bone homeostasis. The widespread presence of VDR and the key activating (1α-hydroxylase, CYP27B1) and inactivating (24-hydroxylase, CYP24A1) enzymes in most mammalian cells and the data that 1,25(OH)2D regulates ∼3% of the mouse and human genome suggest that vitamin D metabolites may also function as a hormone or cytokine for many other cellular functions. The development of mice lacking either VDR or 1α-hydroxylase provided the tools for a more careful examination of the global physiologic role of vitamin D. The phenotype of these mice can then also be used to explore whether similar consequences of vitamin D deficiency or excess can be observed in humans. As for other ligands of nuclear receptors, the spectrum of activities of the vitamin D hormone not only covers a primary target system (calcium and bone homeostasis) but also modifies nearly all cells or tissues including skin and muscle, the cardiovascular and immune systems, glucose homeostasis, and cell proliferation in general. Again, similarly to other nuclear receptor ligands (e.g., thyroid hormones), the active vitamin D hormone produced in the kidney plays a hormonal role, whereas the local, extrarenal expression of the 1α-hydroxylase enzyme enables a paracrine role for the vitamin D endocrine system. The relative role of the hormonal versus the paracrine function is, however, still largely unsolved. This report compares the essential findings in VDR-null mice with the mostly observational data generated in men to better define the possible complex role of vitamin D for global health. VDR- or 1α-hydroxylase-deficient mice develop normally in utero but show progressive evidence of rickets after weaning, with growth retardation, growth plate hypertrophy, impaired mineralization of primary and secondary ossification centers and decreased mineralized bone mass, hypocalcemia and hypophosphatemia, secondary hyperparathyroidism and decreased fibroblast growth factor 23 (FGF23) serum concentrations, and ultimately increased mortality.1-4 This exactly reflects the phenotype of children with inactivating mutations of VDR or CYP27B15-7 or with severe vitamin D deficiency. A very high calcium intake (equivalent to >5 g of calcium for humans) with appropriate calcium/phosphate ratio can correct the systemic and bone calcium homeostasis of mice and man with VDR or CYP27B1 mutations. Selective rescue or inactivation of VDR in the intestine confirms that the presence of a functional VDR in the intestine is both essential and sufficient for calcium and bone homeostasis.8 The effects of more modest vitamin D deficiency on bone are better known in humans than in rodents and include decreased intestinal calcium absorption, secondary hyperparathyroidism, decreased BMD, and ultimately increased fracture risk. Cross-sectional studies have shown secondary hyperparathyroidism in most but not all vitamin D-insufficient subjects with variable threshold 25(OH)D levels (from 12 to 40 ng/ml). Short-term intervention studies indicate that PTH levels fall after vitamin D supplementation when baseline 25(OH)D levels were <20 ng/ml.9 Decreased BMD values are associated with low 25(OH)D concentrations in several cross-sectional studies, but again, the absolute 25(OH)D threshold differ in different studies or populations from as low as 1210, 11 to >30 ng/ml.12 The largest analysis was based on 13,432 individuals of NHANES III, including individuals with different ethnic racial backgrounds.12 Compared with the lowest quintile of 25(OH)D, the highest quintile had higher mean BMD by >4% in whites (test for trend; p < 0.0001), by 2–4% in Mexican Americans (p ≤ 0.01), and by 1–2% in blacks. In the regression plots, higher serum 25(OH)D levels were associated with higher BMD throughout the reference range of 9–37 ng/ml in all subgroups. Fracture incidence is also increased in patients with the lowest 25(OH)D levels (<19 ng/ml).13 The true benefit of vitamin D supplementation on fracture incidence is ideally studied in randomized controlled trials (RCTs). Several recent meta-analyses have come to slightly different conclusions,14-18 but nevertheless indicate that daily vitamin D supplementation (preferably 700–800 IU of vitamin D3) and adequate calcium intake (>1000 mg) reduce the risk for hip or nonvertebral fractures of older white U.S. and European subjects by slightly >20%, whereas calcium supplementation alone is largely ineffective.19 Unfortunately, most RCTs either did not measure serum 25(OH)D or used methods with poor accuracy so that the optimal 25(OH)D level for bone health is still disputed. There is, however, consensus that 25(OH)D should be >20 ng/ml in all postmenopausal women and elderly men20 or, according to some experts,21 a mean population value of ≥30 ng/ml. VDR-deficient but not 1α-hydroxylase– or vitamin D-deficient mice develop total alopecia after first catagen because of a skin stem cell defect. This is thus a VDR and not vitamin D phenotype. As for the calcium and bone phenotype, there is a perfect concordance with the clinical presentation of children with VDR mutations, and this skin phenotype represents the first proof of noncalcemic actions of the vitamin D-VDR system. Mice lacking VDR develop a skeletal muscle phenotype with smaller and variable muscle fibers and persistence of immature muscle gene expression during adult life. These abnormalities persist after correction of systemic calcium metabolism by a rescue diet.22 In humans, four lines of evidence support a role of vitamin D in muscle health. First, proximal muscle weakness is a prominent feature of the clinical syndrome of vitamin D deficiency. Second, VDR is expressed in human muscle tissue, and VDR activation may promote de novo protein synthesis in muscle. Third, several observational studies suggest a positive association between 25(OH)D and muscle strength or lower extremity function in older persons.23, 24 Finally, in several double-blind randomized controlled trials, vitamin D supplementation increased muscle strength and balance and reduced the risk of falling in community-dwelling individuals, as well as in institutionalized individuals.23, 24 Fall reduction with vitamin D seems to depend on dose based on data from one multiple-dose trial and a meta-analyses of high-quality trials, with effects seen at higher intake of at least 700 IU but not below.23 A dose-response relationship between vitamin D status and muscle health was examined in NHANES III, including 4100 ambulatory adults ≥60 years of age. Muscle function was poorest in subjects with the lowest 25(OH)D (<20 ng/ml) levels. Similar results were found in a Dutch cohort of older individuals.24 The cross-sectional data obtained in the NHANES study, however, also indicated that muscle function continued to increase throughout the reference range of 25(OH)D (9–37 ng/ml). VDR-null mice, even on a high calcium rescue diet, develop high renin hypertension and cardiac hypertrophy that can be prevented by treatment with an angiotensin blocker.25 This high renal renin production is also observed in 1α-hydroxylase-deficient mice, whereas in vitro or in vivo exposure to 1,25(OH)2D decreases renin production probably by direct regulation of the gene expression through a vitamin D responsive element (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 among both normotensive and hypertensive individuals. The results of these observational studies are supported by two RCTs, in which vitamin D treatment reduced blood pressure in hypertensive subjects and elderly community-dwelling women. Two large prospective cohort studies among men and women reported that the serum level of 25(OH)D associated with the lowest incidence of hypertension is 30 ng/ml or higher, with a significant inverse trend between baseline 25(OH)D and incidence of hypertension over a 4-yr follow-up.26 Consistent with this, another large epidemiological study showed a significant relation between serum 25(OH)D concentrations and systolic, diastolic, and pulse pressure among the total adult population of NHANES III; the lowest pressure levels were observed >36 ng/ml.27 VDR-null mice display increased thrombogenicity when exposed to inflammatory stimuli.28 Moreover 1,25(OH)2D has beneficial effects on most cells of the vascular wall and enhances fibrinolysis. In humans, a low vitamin D status is associated with several cardiovascular risk factors,29 including the metabolic syndrome. The largest prospective intervention trial (WHI) with calcium (1 g/d) and vitamin D (a low dose of 400 IU/d), however, showed no increased or decreased coronary or cerebrovascular risk after 7 yr of follow-up.30 Several large-scale observational studies showed that, among patients on chronic renal replacement therapy, survival and especially cardiovascular events were lower when treated with 1,25(OH)2D analogs compared with either untreated or calcitriol-treated groups.31 These data are in line with a beneficial effect of the vitamin endocrine system on cardiovascular targets but certainly need confirmation by a proper RCT. Vitamin D toxicity caused by endogenous 1,25(OH)2D or by excess of exogenous vitamin D can lead to vascular and valvular calcifications and thereby cause multiple organ failure and early mortality. This has been amply shown in animals with genetic deletion of the 24-hydroxylase catabolizing enzyme and in animals with overexpression of 1α-hydroxylase activity caused by genetic deletion of FGF23 or klotho. In all these mouse strains, additional deletion of either 1α-hydroxylase or VDR (double knockout mice) avoided these ectopic calcifications. It thus seems that both deficiency and excess vitamin D might have deleterious effects on the cardiovascular system. All cells of the immune system (antigen-presenting cells such as dendritic cells, macrophages, and T and B cells) express VDR either constitutively or after appropriate immune stimulation. Dendritic cells and monocytic cells also express 1α-hydroxylase after immune stimulation.32 Moreover, all immune cells including mast cells are sensitive to 1,25(OH)2D action. 1,25(OH)2D is a potent inhibitor of dendritic cell maturation and generates a coordinated action on T-cell gene expression of key cytokines (interleukin [IL]1, IL2, IL12, IL17, IFNγ) and genes needed for antigen presentation to T cells (MHC class II and co-signaling proteins). These wide set of immune modulating actions of 1,25(OH)2D therefore taper down the activation of the acquired immune system and should have beneficial effects on the occurrence or evolution of autoimmune diseases. Indeed 1,25(OH)2D and especially some more potent and selective analogs are able to significantly reduce spontaneous or experimental autoimmune diseases in rodents such as type 1 diabetes in NOD mice or multiple low-dose streptozotocin-induced diabetes, experimental allergic encephalitis, nephritis, or inflammatory bowel disease.32-34 However, no such intervention studies have yet been performed in human autoimmune diseases. Based on these observations, VDR-null or vitamin D-deficient mice should be predisposed to autoimmune diseases, and indeed, VDR-null mice are more prone to develop inflammatory bowel disease.35 Moreover, genetically predisposed NOD mice with transient vitamin D deficiency early in life have a much higher incidence of type 1 diabetes than vitamin D-replete mice.36, 37 Consistent with these data from NOD mice, several epidemiological studies in humans report that vitamin D intake in early life may reduce their later risk of type 1 diabetes. Risk reduction varied between 26% with cod liver oil to 78% with 2000 IU/d.32 Because 25(OH)D serum levels have not been measured in the cohorts of children, a threshold cannot be addressed for optimal reduction of type 1 diabetes. These observations point to the potential importance of preventing vitamin D deficiency in the perinatal period and early childhood. 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 >7 million U.S. military personnel showed that, among whites, low 25(OH)D levels, especially before 20 yr of age, was a strong risk factor for later occurrence of multiple sclerosis [OR for a 20 ng/ml increase of 25(OH)D of 0.59, confidence limits 0.36–0.97].38 Therefore, compared with the lowest quintile of 25(OH)D levels (<25 ng/ml), the OR for multiple sclerosis in each of the subsequent quintiles varied between 0.74 and 0.38 (p = 0.02 for trend across quintiles). However, as for type 1 diabetes, no controlled randomized intervention studies have yet proven 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 be high priority. In contrast to the tapering down effects of 1,25(OH)2D on the acquired immune system, positive effects of the vitamin D endocrine system on the innate immune system have been well described. Exposure of monocytes/macrophages to bacterial infections upregulates toll-like receptor, VDR, and 1α-hydroxylase expression, leading to the induction of natural defensins (especially cathelicidin), which increase killing of intracellular bacteria by their microbicidal action.39 This is consistent with decreased monocytic differentiation and antibacterial activity of monocytes from VDR- or vitamin D-deficient rodents. The human equivalent would be a greater risk for bacterial or other infections in vitamin D-resistant or vitamin D-deficient children or adults. Higher infection rates in ricketic children is clinically well recognized but could be multifactorial. Low serum levels of 25(OH)D are associated with a higher risk of active tuberculosis according to a recent meta-analysis.40 Several clinical trials are under way according to the NIH clinical trial register to evaluate the potential benefit of vitamin D supplementation on the speed or cure rate of tuberculosis. Pancreatic β cells express VDR, 1α-hydroxylase, and calcium-binding proteins and respond to 1,25(OH)2D in vitro with increased insulin secretion.41 Vitamin D-deficient rodents and rabbits secrete less insulin and display impaired glucose tolerance. VDR-null mice either have a normal glucose tolerance or are mildly glucose intolerant, probably related to a different genetic background. Small-scale studies of vitamin D-deficient adults reported a modest improvement of their glucose handling after vitamin D repletion. Large epidemiological studies support a relation between serum 25(OH)D concentrations and insulin secretion and/or sensitivity. In NHANES III, including 6228 adults ≥20 yr of age, there was a dose-dependent inverse association between serum 25(OH)D concentration and the prevalence of diabetes. For 25(OH)D levels of 32 ng/ml or higher compared with 18 ng/ml or lower, the OR was 0.25 (95% CI, 0.11–0.60) among whites and 0.17 (95% CI, 0.08–0.37) among Mexican Americans, without an association among blacks.42 Furthermore, in the same cohort, serum 25(OH)D concentrations were inversely associated with insulin resistance (HOMA-IR), with best levels observed in the top quartile of 25(OH)D of 32 ng/ml or higher. In elderly Dutch individuals with a mean age of 75 yr, the risk of type 2 diabetes increased with lower serum 25-hydroxyvitamin D in a crude analysis, whereas adjustment for waist circumference, physical activity, and alcohol intake made the association disappear.43 Because obese subjects are known to have much lower 25(OH)D levels than lean subjects because of multifactorial mechanisms, it is unclear whether 25(OH)D deficiency independently increases the risk of metabolic syndrome and type 2 diabetes. Small-scale intervention studies with vitamin D supplementation or 1,25(OH)2D have generated inconsistent results. In a larger 3-yr randomized controlled trial among 445 older individuals ≥65 yr of age treated with either 700 IU vitamin D3 plus 500 mg calcium or placebo, no overall effect on glucose tolerance was observed.44 However, among participants with impaired fasting glucose at baseline, those who took combined calcium-vitamin D supplements had a lower rise in fasting blood glucose at 3 yr compared with those on placebo (p = 0.04) and a lower increase in HOMA insulin resistance (p = 0.03). Because this study combined vitamin D with calcium, the effect may have been caused by either nutrients. Thus, more data are needed to confirm the independent effect of vitamin D on glucose and insulin metabolism. In the absence of data from RCTs, an alternative explanation could be that higher 25(OH)D serum concentrations are a surrogate of outdoor exercise. Slightly supraphysiologic concentrations of 1,25(OH)2D nearly universally retard the cell cycle progression of normal and most malignant cells and even frequently enhance their differentiation.45 This prominent G0/G1 cell cycle arrest is the result of a coordinated action of 1,25(OH)2D on at least 100 genes, with major effects on the activity of E2F transcription factors, phosphorylation of retinoblastoma family proteins, cyclins and cyclin-dependent kinase inhibitors, and actions on epidermal growth factor (EGF), IGF signaling, and prostaglandin synthesis and action. In some cells, it also enhances the apoptosis cascade. This activity is maintained in most cancer cells, but several escape mechanisms (loss of VDR, VDR resistance, or enhanced catabolism of vitamin D metabolites) are found in the most malignant cells. VDR-null mice should therefore be more prone to tumor formation. Spontaneous cancer incidence is apparently not increased, but VDR KO mice have mammary gland hyperplasia and show increased proliferation response to exogenously administered estrogen and progesterone46 and also develop hyperproliferation of their intestinal cells. Moreover, they are more prone than wildtype littermates to develop chemocarcinogen or oncogen-induced tumors.45 This obviously raises the question of whether less severe vitamin D deficiency in humans is associated with increased risk of cancer. This was initially suggested by a higher incidence of a wide variety of cancers in populations living in areas with lower potential exposure to natural UVB sunlight, but the relation between latitude or potential exposure to sunlight and actually measured 25(OH)D levels is inconsistent. When the relationships between colorectal cancer and dietary or supplementary vitamin D have been studied in cohort studies or in case-control studies, the majority of studies suggested inverse associations of vitamin D intake with colon or rectal cancer or both.33, 45, 47 The evidence for an association between estimations of vitamin D intake is much weaker for other cancers. Serum 25(OH)D levels are therefore a more reliable marker for evaluation of the association of vitamin D deficiency and cancer because they reflect both intake and sun exposure. Several observational studies have found a higher risk with low 25(OH)D for especially colorectal cancer but also for breast and prostate cancer. According to a pooled analysis, women with serum 25(OH)D of ∼48 ng/ml (median of the top quintile) had a 50% lower risk of breast cancer than those with serum <13 ng/ml in the lowest quintile.48 Meta-analysis of studies addressing the association between 25(OH)D levels and colorectal cancers showed that individuals with serum 25(OH)D levels ≥32 ng/ml had a 50% lower incidence of colorectal cancer than those with relatively low levels (≤12 ng/ml).49 The inverse association between colorectal cancer mortality and serum 25(OH)D levels was confirmed by the results of the Third National Health and Nutrition Examination Survey.50 The association between plasma 25(OH)D and prostate cancer are more mixed, because most studies show a negative relation, but a substantial number of studies show no relation, and a few indicate even a positive relation between (higher) 25(OH)D levels and prostate cancer.49, 51 Finally, lending further support to the vitamin D-cancer hypothesis, an increment of 10 ng/ml in predicted 25(OH)D levels (not actual measured values) was associated with a 17% reduction in total cancer incidence, a 29% reduction in total cancer mortality, and a 45% reduction in digestive system cancer mortality in a comprehensive prospective analysis from a large male U.S. cohort (Health Professionals Follow-Up Study).52 The largest intervention study (WHI) did not show an effect on colorectal cancer incidence after 7 yr of treatment with vitamin D (400 IU/d) and calcium (1 g/d),53 although the same study suggested a significant inverse trend between lower baseline serum 25(OH)D levels and an increased risk of colo-rectal cancers. The low dose of vitamin D, serious compliance problems in the WHI trial, and the duration of treatment may all have influenced this negative outcome. In a much smaller 4-yr study in postmenopausal women, higher doses of calcium (1.4–1.5 g) and vitamin D (1100 IU), which raised serum 25(OH)D to levels >32 ng/ml,54 significantly reduced the overall cancer risk in postmenopausal women. Similar to epidemiological studies, the authors documented a 35% decrease in cancer risk per 10 ng/liter increase in 25(OH)D levels, The small number of cancer deaths and a major confounding factor of calcium intake limits the interpretation of the study. Larger prospective RCTs are in principle necessary to evaluate the possible causal link between vitamin D status and the major human cancers. This was also the conclusion of two NIH supported meetings on the nutritional role of vitamin D.55, 56 However, because cancers have long latencies, evidence from RCTs may be difficult to achieve. Thus, it may well be that the best arguments will come from multiple prospective epidemiological studies combined with strong mechanistic evidence. The careful analysis of the phenotype of VDR-null and other transgenic mice has generated a wealth of new information on the spectrum of activities of the vitamin D endocrine system. The concordance between the bone, growth plate, and skin phenotype of VDR- and vitamin D-deficient mice and humans raises the question of whether the many other noncalcemic effects of severe vitamin D resistance (on immune system, cardiovascular system, muscle and cell proliferation) are also operational in humans with less severe vitamin D deficiency. The work of epidemiologists suggests that suboptimal vitamin D status does indeed exacerbate the risks of major human diseases as predicted from the animal models. Specific RCTs are now desirable to confirm the causal relationships and to define the optimal vitamin D status for bone and global health. At present, it seems that vitamin D requirements for prevention of rickets [25(OH)D > 10 ng/ml] are clearly lower than for the maintenance of bone quality in older age (>20 ng/ml), whereas the threshold for other endpoints might well be even much higher. A large part of the human population therefore seems to have a compromised vitamin D status with potentially major health implications. The vitamin D field is thus a very good example of how data generated in cells and mice have stimulated clinical research and how epidemiological data can guide research should be to further for vitamin D
Language: English
Field: Medicine
Open Access status: bronze
OpenAlex ID: W2069981790
FWCI: 3.014
Cites: 55
Cited by: 264
Related to: 10