Vitamin D Safety and Requirements

Abstract

Vitamin D an ancient secosteroid is essential for mineral homeostasis, bone remodeling, immune modulation, and energy metabolism. Recently, debates have emerged about the daily vitamin D requirements for healthy and elderly adults, the safety and efficacy of long term supplementation and the role of vitamin D deficiency in several chronic disease states. Since this molecule acts as both a vitamin and a hormone, it should not be surprising that the effects of supplementation are multi-faceted and complex. Yet despite significant progress in the last decade, our understanding of vitamin D physiology and the clinical relevance of low circulating levels of this vitamin remains incomplete. The present review provides the reader with a comprehensive and up-to-date understanding of vitamin D requirements and safety. It also raises some provocative research questions.

Introduction

Cholecalciferol (vitamin D₃) and ergocalciferol (vitamin D₂), the two main forms of vitamin D are both pro-hormones and vitamins. Vitamin D₃ is formed in the skin after exposure to sunlight or consumed from natural (especially fatty fish) or supplemented foods and vitamin D₂ is obtained by irradiation of plants or supplemented foods. The non-enzymatic conversion of pro- to pre- and subsequently to vitamin D in response to solar UVB radiation leads thereafter to the production of the active hormone, through hydroxylation in the liver [forming 25-hydroxyvitamin D (25-OHD)] and then 1 alpha hydroxylation in the kidney (synthesizing 1,25 dihydroxyvitamin D). 1,25 Dihydroxyvitamin D [1,25(OH)₂D] subsequently binds to the vitamin D receptor (VDR) in the three classic target tissues (e.g. intestine, kidney and bone) to exert its effect on calcium balance. But, vitamin D has pleiotropic properties in “off-target” sites and can influence cell proliferation, muscle performance, energy metabolism and bone strength independent of its actions on calcium absorption [1,2]. It is noteworthy that endogenous vitamin D production becomes less efficient with age and the rate of decline is similar to age-associated declines in serum levels of growth hormone (GH) [3-5], insulin-like growth factor I (IGF-I) [6-8] and gonadal steroids [8,9].

Since the last half of the 20^th century, increasing longevity has led to great interest in the investigation of whether quality of life could be maintained by restoring hormone levels to their pre-aging concentrations. This has been pursued in relation to several classic hormones [3,7,10-13] and more recently to vitamin D [14,15]. For the latter, a simpler paradigm might lead to the following question: do increased levels of vitamin D improve health related quality of life and better long term outcomes? Or expressed another way, does more means better [16-18] (Fig. 1). Not surprisingly, the answer to this question is complex and confounded by multiple factors [19,20]. Indeed, there is an inherent paradox when considering low vs high for concentrations of any vitamin as recently emphasized by data from NHANES III and presented in the IOM report. For example, for all cause mortality there appears to be a “J” or “U”-shaped relationship for 25-OHD with increased mortality at low serum 25OHD levels(<20 ng/mL) and an upturn in mortality at significantly higher 25OHD levels (>50 ng/mL) (Fig 1) [21,22]. However, these data are based on observational cohorts and therefore must be considered in that light. Only long term trials will be able to sort out the benefits and potential risks of vitamin D supplementation, particularly with high doses.

Fig 1.

This figure shows some of the most relevant studies, which are used to justify beneficial and detrimental effects of high serum levels of vitamin D as well as efficiency and inefficiency of vitamin D treatment.

Physiology of Vitamin D

In contrast to other hormones and factors, vitamin D₃ generation is a complicated process, comprising different sources (endogenous and exogenous), different tissue production (skin, liver and kidney), a less active intermediary metabolite/pro-hormone (25-OHD), and a critical set of modulators (PTH, calcium, phosphorus and FGF-23), [23,24]. Taken together, these factors complicate the direct measurement of vitamin D, as well as which parameter would best reflect overall vitamin D status [25,26].

Beyond the well known effects of vitamin D deficiency on calcium homeostasis (e.g. osteomalacia and rickets), compelling evidence has recently demonstrated that severe vitamin D deficiency such as that observed in genetically engineered animals lacking or resistant to vitamin D [27,28], as well as in patients with chronic kidney disease, is associated with cardiovascular disorders [29,30]. Experimental and clinical investigations have suggested but not proven that vitamin D deficiency may also be associated with immune deficiency, diabetes mellitus, arterial hypertension, and cancer [31-35].

As noted previously, endogenous synthesis of vitamin D₃ starts with a photochemical reaction, where 7-dihydrocholesterol is converted to pre-vitamin D, which is transformed to vitamin D₃ by isomerization through a temperature-dependent reaction. Alternatively, vitamin D can be supplied exogenously as a nutrient. Limitations of both sources come from concern about the adverse effects of sun exposure on the skin and the scarcity of vitamin D in food widely consumed by the population [36]. Vitamin D circulates within the bloodstream after binding to the vitamin D binding protein (DBP) and is hydroxylated in the liver in a constitutive process, which is largely dependent on the substrate concentration. The 25-OHD formed in the liver, and bound to DBP is the circulating metabolite with the highest concentration and can be readily measured by several techniques such as competitive protein binding ( i.e. radioimmunoassay, enzyme linked immunoassay, or chemiluminescence) and chromatography (HPLC and LC-MS) [37-42]. The circulating level of 25OHD reflects overall vitamin D contact from diet and sunlight exposure. It is not a biomarker of disease. Yet the widespread utilization of the vitamin D assay to define disease risk has led to multiple issues from both a clinical and research perspective. Moreover, there remain significant differences in assay performance and worldwide standardization is lacking [22].

The most biologically active metabolite of the vitamin D synthesis pathway is the 1,25(OH)₂D hormone. This final step takes place in the kidney and is rigidly controlled by the stimulus of PTH and the inhibition by FGF-23, calcium, phosphorus and 1,25(OH)₂D itself [43]. This system is very tightly controlled such that 1,25(OH)₂ D synthesis can be suppressed during vitamin D intoxication. Additionally, during vitamin D deficiency low serum levels of 25-OHD can be associated with high, normal or low 1,25(OH)₂D serum levels. This occurs because mammalian systems have an array of mechanisms to maintain renal synthesis of 1,25(OH)₂D during negative calcium balance, thereby preserving calcium homeostasis [23,24]. For example, even a slight decline in serum calcium levels triggers a prompt PTH response, which stimulates 1,25(OH)₂D synthesis [24]. Therefore, the biochemical profiles of increased levels of 25-OHD and decreased 1,25(OH)₂D during vitamin D intoxication and diminished 25(OH)D and enhanced 1,25(OH)₂D in vitamin D deficiency are not totally unexpected. These mechanisms are highly efficient except during renal failure when 1 alpha hydroxylation is impaired despite high levels of PTH.

Relationship between 1,25(OH)₂D, PTH and FGF-23

The control of 1,25(OH)₂D synthesis was formerly considered to be solely dependent on the levels of PTH, calcium and phosphorus. However, FGF-23, an osteocyte derived hormone, has emerged as a major modulator of serum phosphate levels, by its ability to decrease tubular reabsorption and intestinal absorption of phosphate. FGF-23 is integrated into the control of 1,25(OH)₂D and PTH synthesis, inhibiting the production of both hormones [43]. PTH, 1,25(OH)₂D and FGF-23 have an intricate relationship in which increased PTH secretion leads to a positive effect on the production of 1,25(OH)₂D and a negative effect on FGF-23; 1,25(OH)₂D inhibits PTH production but stimulates FGF-23 synthesis, and FGF-23 inhibits synthesis of PTH and 1,25(OH)₂D, respectively [44].

The metabolic role of FGF-23 was first investigated in genetic models of hypophosphatemic rickets [44] and tumor-induced osteomalacia where circulating levels of FGF-23 have been reported [45]. Mineral homeostasis in VDR-null mice has allowed a more clear understanding of the regulation of serum FGF-23 [46]. For example, administration of calcitriol to thyro-parathyroidectomized rats results in a dose-dependent increase in FGF23 levels. However, this response is not observed in VDR-null mice, indicating that vitamin D works by genomic actions to directly increase FGF23 production [47]. VDR-null mice, which exhibit hypocalcemia, hypophosphatemia, and secondary hyperparathyroidism, have decreased or undetectable FGF23 concentrations compared to normal mice. Taken together, these lines of evidence indicate that 1,25(OH)₂D concentrations may directly regulate FGF23, whereas serum phosphate, and to a lesser extent, serum calcium, may indirectly modulate FGF23 generation in the skeleton. The role of FGF-23 in the pathophysiology of vitamin D deficiency and in the homeostatic adaptation after vitamin D replacement has not been fully evaluated in clinical investigations.

Solar exposure and serum vitamin D: Seasons and skin pigmentation

Endogenous vitamin D₃ and exogenous vitamin D₃ and D₂ are alternative sources supplying the physiological requirements of this secosteroid. Since a limited number of foods contain vitamin D, vitamin D might be considered to be a unique micronutrient due to its auto-renewable, solar-induced endogenous production. However, from a teleological perspective, this alternative source indicates the importance of vitamin D in human physiology and thereby the necessity for its stable production [48]. Additionally, the dependence on sunlight exposure predictably leads to a seasonal variation in the body’s supply of vitamin D. The disadvantage of living in areas with well defined seasons and expected periodical vitamin D shortage has been suggested in studies exhibiting variations in bone mineral density that can be attributed to changes in serum levels of 25-OHD [49]. Rapuri et al [49] observed that 25-OHD and BMD were lower, whereas PTH and N-telopeptide excretion were higher, in postmenopausal women during wintertime (December, January, February and March) than during summertime (June, July, August and September). This study had a cross-sectional design and comprised only 251 women. The authors did not observe differences in the levels of 1,25(OH)₂D between groups assessed in winter and summer. On the other hand, Gerdhem et. al. [50] evaluated seasonal variation in BMD in 2337 postmenopausal women in a cross-sectional study performed in Sweden and observed different results. The BMD of women assessed during July–December was 5% lower compared to values in women assessed between January–June. There was no difference between groups of women evaluated in April-September versus October-March.

Not unexpectedly equatorial areas demonstrate the least changes in serum 25OHD with seasons. However, latitude is not the exclusive variable to be considered; for instance, some studies have shown a high prevalence of vitamin D deficiency in Brazil. A multicentre study showed that the prevalence of vitamin D deficiency among osteoporotic Brazilian patients was 42.4%, but the cities where the data were collected were not mentioned [51]. The results observed in Brazil were slightly higher than those detected in Sweden (37%), but lower than those detected in Mexico (67%) and in the United Kingdom (74%) [51]. While other studies have supported these results by demonstrating a high prevalence of vitamin D deficiency in the city of São Paulo [52,53], in Riberão Preto, a city located 300 km from São Paulo, vitamin D deficiency was not observed even in patients harboring chronic diseases associated with bone loss [54,55]. These apparently paradoxical data actually reflect the complex process involved in vitamin D metabolism. Serum levels of 25-OHD are a marker of exposure and depend on an integrated response of multiple factors including geographic region, ethnicity, life style (social behavior, nutritional habits), environment, age, and medication. Accordingly, all of these components should be taken into account to recognize individuals at risk to develop vitamin D deficiency.

Animal models demonstrate even more complexity regarding seasonal changes in the vitamin D axis. C57BL6 mice were studied cross sectionally after being maintained in a stable laboratory environment on a fixed diet with constant light dark cycles and humidity [56]. Male and female C57BL6 mice had higher volumetric BMD in summer than in winter. Females showed reduced trabecular bone mass during winter, whereas males showed deleterious changes in cortical bone volume. Males, but not females, had higher IGF-I in summer than in winter, and only males showed an increase in body weight during the winter. Remarkably Trap5b was higher and 25(OH)D was lower in winter than in summer among female C57BL6 mice (Rosen and Beamer, personal communication) despite chow supplementation and controlled environmental factors. These unexpected results suggest that a variation in bone mass during the year is not only dependent on seasonal changes of vitamin D levels due to solar exposure, but other factors such as gender, dietary phytoestrogens, and genetic background. Further studies are needed to address the occurrence of seasonal bone mass variation and whether such variation is responsible, at least in part, for the slow bone loss process that is likely to start early after the peak bone mass [57]. One mechanism may be related to polymorphisms in hepatic 25 hydroxylase (Cyp2R1) that enhance the conversion of calciferol to calcidiol. Further support for that comes from Kiel and colleagues who reported in genome wide association studies that there is strong heritability for 25OHD and that polymorphic differences in the Cyp2R1 locus are associated with serum levels [58].

Skin pigmentation also results in lower serum levels of 25OHD as a direct result of less endogenous production in the skin. However, it is not clear that these changes alter normal homeostatic relationships in the PTH-vitamin D axis, particularly relative to calcium absorption and PTH secretion. Moreover heavily pigmented individuals tend to have higher bone mass and lower rates of fractures than light skinned individuals [59-62]. A similar relationship may also be found with obesity although there are many confounding factors that make this extrapolation somewhat tenuous [63-65].

The complex association between Vitamin D Deficiency and obesity

Obesity is often accompanied by low levels of 25-OHD and normal or high levels of 1,25(OH)₂D [66]. The mechanism(s) for this inverse relationship between body weight and serum 25OHD is not well defined. Vitamin D is fat soluble and it is likely that with greater fat mass more vitamin D is stored in adipocytes. But the precise mechanism whereby vitamin D storage is enhanced by greater fat mass leading to lower circulating levels of the 25OHD metabolite has not been described. On the other hand, 1,25(OH)₂D , which is much more constant in the circulation even in obese individuals, appears to be necessary for both insulin secretion and insulin action (Fig 2). As such, several studies have suggested an association of vitamin D deficiency with obesity and insulin resistance [33]. But this concept reflects only a partial view of a complex series of metabolic changes (Fig 2).

Fig 2.

Vitamin D has a complex relationship with obesity: top left) obesity is considered a condition associated with vitamin D deficiency, showing low serum levels of 25-OHD and high 1,25(OH)₂D. Previous studies suggest association between vitamin D levels and insulin resistance; top right) VDR null mice, VDR heterozygote male mice and CYP27B null mice have a lean phenotype and there are studies showing that VDR null mice have metabolic advantages when exposed to a diabetogenic diet; botton) 1,25(OH)₂D signaling through membrane and nuclear receptors inhibits lipolysis and promotes lipogenesis, there are studies showing that hypercalcemic diet induces weight loss [2,23,61-66].

VDR- vitamin D receptor, CYP27B- 25-hydroxyvitamin D-1α-hydroxylase.

To understand the possible relationship between obesity and the vitamin D axis, investigators have turned to the vitamin D receptor- knockout (VDR) and 1α-hydroxylase-knockout CYP27B mice. Both genotypes exhibited a lean phenotype [67,68] and, in addition, VDR null mice showed greater insulin sensitivity when exposed to a high-calorie diet [67] (Fig 2). More recently, it was observed that male, but not female, VDR heterozygote mice also show a lean phenotype [2]. Conversely, only female mice with VDR haploinsufficiency showed impairment in bone acquisition. In agreement with these results, in vitro data suggest that 1,25(OH)₂D can favor fat storage by signaling through genomic and membrane receptor pathways [69]. Additionally, it has been observed that a high calcium diet induces weight loss; it has been hypothesized that the mechanism behind the weight loss is the inhibition of 1,25(OH)₂D synthesis [70] (Fig. 2). 1,25(OH)₂D promotes the increase of intracellular ionized calcium, thus triggering lipogenesis and inhibiting lipolysis [71]. On the other hand, the lack of 1,25(OH)₂D may be associated with decreased lipogenesis through a diminished activity of fatty acid synthase [71].

Clinically, obesity is associated with low 25OHD levels [63-65], hyperinsulinemia, high gonadal steroids and increased peripheral leptin levels. Paradoxically obese individuals have high bone mass, lower bone turnover rates, longer periods of bone formation, increased calcium absorption, and in African-American subjects skeletal resistance to PTH despite the low to very low levels of serum 25OHD. As arterial hypertension is highly prevalent in black individuals and insulin resistance is associated with obesity, several studies have examined the role of vitamin D deficiency in hypertension and insulin resistance in both groups [72-74]. As expected, these studies suggested that vitamin D deficiency may in part be responsible for the emergence of hypertension and insulin resistance. Both conditions, heavily pigmented skin [59-61] and obesity [63-65], are classified as states of vitamin D deficiency in which low serum levels of 25-OHD exist concomitantly with increased levels of 1,25(OH)₂D.

The diagnosis of vitamin D insufficiency

The reference values of a clinical parameter can be established by collecting samples from a representative number of normal individuals and calculating the mean of the values obtained. The normal range corresponds to values situated between the mean and two standard deviations above and below the mean. Specifically, when a parameter is associated with the development of chronic complications such as glucose in diabetes mellitus [75,76], cholesterol in cardiovascular disorders [77] and bone mineral density in osteoporosis [78], more sophisticated, expensive and time-consuming studies are required to establish a threshold value for the development of these complications [75-78]. Previously, vitamin D deficiency was defined by clinical signs and symptoms (bone pain and proximal muscle weakness) and by measuring serum calcium, phosphate and alkaline phosphate levels. However that syndrome is relatively rare in North America and Europe, and has now evolved to a point where most healthy and virtually all chronically ill individuals have had serum 25OHD levels [79]. The current challenge in clinical medicine is to define a standard reference value for serum 25OHD that may be helpful for clinicians in managing patients with chronic diseases such as osteoporosis, cardiovascular disease, cancer or falls [22,80].

The measurement of 25(OH)D is challenging because circulating 25(OH)D is moderately lipophilic, binds avidly to vitamin D binding protein, is present in low (nanomolar) concentrations, and exists in two structurally similar forms, 25(OH)D₃ and 25(OH)D₂. Four decades ago, it became possible to assay serum 25-OHD through competitive protein-binding assays (CPBA) [37,38]. These methods are time-consuming, need organic extraction, sample purification by column chromatography, and involve ³H-25-OH-D₃ as a tracer. Quantitative assessment of 25-OH-D by high performance liquid chromatography was developed shortly after and the final refinement of the measurement of this metabolite in the research environment was the combination of liquid chromatography with mass spectrometry [39]. The increased demand for clinical evaluation of vitamin D deficiency stimulated the development of a rapid, simple, but valid methodology. The measurement of 25-OHD by radioimmunoassay progressed from the use of ³H-25-OH-D₃ as a tracer to a ¹²⁵I-labeled marker [40,41] which was the first test approved by the US Food and Drug Administration for the diagnosis of vitamin D disorders. The latest advance in the clinical assessment of vitamin D status was the development of automated, nonextracted chemiluminescent immunoassay [42], recently approved by the FDA. However, several reports have demonstrated large inconsistencies and variability in 25(OH)D measurements between methods and laboratories [81-83]. As a result, some groups have emphasized the need for appropriate reference materials and a standardization of 25-OHD assays [81,82].

Until recently the low normal level for serum 25-OHD was 10 ng/ml, (25 nmol/L). This threshold had the advantage of high specificity for rickets and osteomalacia but the disadvantage of being too insensitive to permit the recognition of ‘insufficient’ vitamin D. Because vitamin D plays an essential role in calcium homeostasis and bone remodeling, a growing awareness emerged about circulatory levels of 25-OHD and bone mass acquisition and maintenance. Furthermore, with the discovery that the vitamin D receptor (VDR) was expressed in many non-classical target tissues, it was hypothesized that higher serum levels of 25-OHD might be necessary for vitamin D to exert its extra-mineral effects on the neuromuscular, cardiovascular, endocrine and immune systems [84-86]. Different definitions of both vitamin D deficiency and insufficiency have been proposed that depend on the particular 25(OH)D assay used as well as the functional outcome measured. There is no clear consensus about the optimal definitions of either vitamin D deficiency or insufficiency. Recent relatively high estimates of the prevalence of vitamin D insufficiency in the general population may be attributed to the use of higher 25(OH)D thresholds to define low vitamin D status [87,88].

There is intense interest in defining the optimal vitamin D status for skeletal health. Diverse cut points for serum levels of 25(OH)D have been suggested, ranging from 16 to 48 ng/ml. This uncertainty has arisen from two main sources: differences in the functional endpoint (e.g., fractures, PTH concentration, calcium absorption) as well as in the biochemical analysis of 25OHD [89]. The feasibility of identifying the circulatory level of 25-OHD needed to prevent secondary hyperparathyroidism has caused this parameter (e.g. PTH) to be the preferred option to define the optimal level of 25-OHD. In general, a recent guideline set the level of 30 ng/ml (75 nmol/l) for 25-OHD as the desired threshold to prevent secondary hyperparathyroidism [80]. In support of this publication there are studies showing improved muscular performance and decreased fall risk in individuals exhibiting higher serum levels of 25-OHD due to supplemental vitamin D [90-92]. On the other hand, recent work suggests there may be no reliable plateau value for PTH related to serum 25OHD in part because of the many confounders for serum PTH (e.g. renal function, age, body composition, calcium intake) [93]. In that same vein, the Institute of Medicine recently reported that serum levels of 25 OHD levels of 20 ng/ml would protect 97.5% of the healthy population from skeletal disorders such as osteoporosis and osteomalacia [22].

In contraposition to the importance attributed to PTH for the diagnosis of vitamin D deficiency, 1,25(OH)₂D serum levels have been considered an alternative approach since it is the active hormone. However, physiological adaptations (i.e., increased PTH) permit the maintenance of normal or even increased levels of 1,25(OH)₂D in conditions of vitamin D deficiency. Only in extreme conditions does serum 1,25(OH)₂D fall in vitamin D deficiency, although with increased serum 25OHD, 1,25 OHD levels do tend to increase. Finally, although 1,25 dihydroxyvitamin D is the active metabolite of vitamin D and the true hormone, its concentration is 100 times less than 25 OHD and measurement of this compound is confounded by numerous factors. First, the assay is technically more challenging and subject to greater error than 25-OHD. Second, conditions such as renal insufficiency and advanced aging reduce 1, alpha hydroxylase (CYP27B) activity and is a major cause for low levels of 1,25(OH)₂D independent of vitamin D stores [94]. Currently there are few studies defining 1,25(OH)₂D) status in obese individuals with modestly low 25-OHD relative to severe vitamin D deficiency (low 25-OHD but high 1,25(OH)₂D) [66]. The lean phenotype and the high insulin sensitivity observed in VDRKO and CYP27BKO mice contrast with the clinical state of insulin resistance in obese individuals {2,33,67]. Taken together, it is clear that further studies are necessary to define the physiological role of systemic 1,25(OH)₂D and its participation on the emergence of diseases. Hence there is justification for the recent position of Endocrine Society to recommend against using the serum 1,25(OH)₂D to evaluate vitamin D status [80].

The search for the D-limit

Recently, the expert panel of IOM updated the dietary reference for vitamin D in the United States and Canada [22]. The committee proposed a modest increase in daily intake of vitamin D and recognized that current evidence does not support non-skeletal benefits for vitamin D, that most North Americans have adequate levels of this secosteroid, and that higher intakes could have adverse health consequences. While the IOM placed the recommended dietary allowance at 600 IU/day for almost all groups of individuals, it raised the tolerable upper intake level to 4000 IU/day, which means that exceeding this limit implies an increase risk of adverse effects [22].

Five years ago, Hatchkock et al., (2007) suggested a revision to the Tolerable Upper Intake Level (UL) for vitamin D [95]. The authors concluded that the UL established by the Food and Nutrition Board (FNB) for vitamin D (50 μg, or 2000 IU) was too restrictive, thus curtailing research, commercial development, and optimization of nutritional policy. In addition, they concluded that the absence of toxicity in trials conducted on healthy adults that used a vitamin D dose ≥250 μg/d (10 000 IU vitamin D₃) supported the confident selection of this value as the UL. However, the authors based their opinion on studies performed between 1982 and 2005, using different methods for 25-OHD measurement. While the results were consistent in showing that high doses of vitamin D₂ or D₃ are safe in relation to hypercalcemia, the studies taken into account were insufficient to show long-term benefits of high doses of these secosteroids on bone and beyond. Their proposal must be balanced by studies showing that even supplements as low as 400 IU/day, combined with supplemental calcium, as shown in the WHI trial, may be associated with a higher frequency of urinary lithiasis [96]. Furthermore there is a suggestion that a high doses of vitamin D administered less frequently, might not be prevent falls but rather increase the rate of both falls and fractures [97]. Currently, it is becoming evident that research efforts should be devoted to performing dose responses for vitamin D using the UL to enhance efforts to understand how higher doses of D supplementation may impact normal homeostatic processes as well as causing adverse effects in “off-target” sites.

Endogenous synthesis, from solar-mediated production, supplies 60-80% of vitamin D requirements. That process is fine-tuned to avoid excessive production of vitamin D in a hormone-independent pathway, (Fig. 3). A prolonged period of sun exposure drives conversion of previtamin D₃ to an inactive photoisomer, lumisterol or tachysterol, rather than to the synthesis of vitamin D [98]. In addition to the former step, there is another preventive mechanism of sun-induced vitamin D intoxication. Once formed, vitamin D can enter into the circulation or absorb solar ultraviolet photons and be isomerized to inactive molecules, such as suprasterol I, suprasterol II and 5,6-transvitamin D [99]. These mechanisms enable human beings to reach a significant increase in serum levels of 25-OHD, although to a non-toxic level. There is no case report of vitamin D intoxication due to excessive sun exposure, even in case of sunstroke. Conversely, oral supplementation of vitamin D₂ or D₃ bypasses skin thermoregulation and directs the secosteroid to the liver for 25-OHD synthesis (Fig. 3). Therefore, supplemental vitamin D does not mimic the physiological control of substrate availability for the production of the intermediate “pro-hormone”.

Fig 3.

A) Sun-induced synthesis of vitamin D includes an ingenious process to prevent intoxication in cases of excessive exposure. High temperature drives pre-vitamin D and vitamin D to form inactive molecules, respectively lumisterol or tachisterol and suprasterol I, suprasterol II or 5,6-transvitamin D₃; B) Oral vitamin D₃ (cholecalciferol) and Vitamin D₂ (ergocalciferol) bypass skin control in synthesis of vitamin D and deliver substrates that will be forwarded to the synthesis of 25-OHD and 1,25(OH)₂D.

It should be noted that vitamin D intoxication occurs in the presence of very high levels of 25-OHD and this leads to decreased production of 1,25(OH)2D in part because of suppressed PTH [100,101]. Therefore, one can hypothesize that the supply of vitamin D should be preferentially based on natural sun exposure. Furthermore, standardization of high vitamin D dosage to achieve non-mineral benefit can expose some subjects to the risk of adverse effects. The prolonged ingestion of excessive amounts of vitamin D can lead to hypercalcuria, hypercalcemia and eventual metastatic calcification of soft tissues. Certainly calcification of vascular tissue, and the ensuing complications thereof, has long been known to be associated with prolonged vitamin D toxicity [102-104]. This has recently emerged as a concern in individuals with renal failure treated with either vitamin D or the active metabolite calcitriol. Whether high dose vitamin D supplementation in elderly individuals with impaired renal function also accelerates cardiovascular risk is still an open question but requires more studies.

Experimental evidence provides some insight into the concerns about excess vitamin D. In genetically manipulated mice lacking FGF-23 (Fgf-23) or klotho (Kl) genes, both exhibit increased serum levels of vitamin D and hyperphosphatemia; these features are associated with premature aging [105]. FGF-23KO mice have been a useful model for understanding of the in vivo effects of FGF-23 in the regulation of phosphate homeostasis [106,107]. These mice have a unique biochemical profile (Pi > 16 mg/dl vs control= 8-10 mg/dl; Calcium > 11 mg/dl vs control= 8-10 mg/dl; PTH= undetectable vs control= 20-80 pg/ml; 1,25(OH)₂D >400 pg/ml vs control= ~ 150 pg/ml; FGF-23= undetectable) [108]. The laboratory phenotype is approximately the same in KLOTHO KO, but the latter exhibit increased serum FGF23 levels [108]. In human beings, an inactivating mutation in FGF-23 and mutation in the KLOTHO gene leads to familial tumoral calcinosis, a disorder marked by hyperphosphatemia and ectopic calcification [109,110]. In contrast to FGF-23 and KLOTHO KO mice, these patients have normal circulatory levels of PTH and 1,25(OH)₂D and do not have accelerated aging [108]. Fgf23- and Kl-deficient mice have a shortened lifespan, growth retardation, infertility, emphysema, muscle atrophy, hypoglycemia, ectopic calcification and elevated serum levels of phosphate and vitamin D [105]. Prolonged vitamin D receptor activation by increased levels of vitamin D might exacerbate premature aging-like phenotypes with or without development of abnormal calcification. Moreover, the aging phenotype of Kl- can be rescued by providing a vitamin D-deficient diet to these rodents [105]. Accordingly, silencing the 1α-hydroxylase gene in Fgf-23 mice, consequently decreasing the synthesis of 1,25(OH)₂D, circumvents the development of premature aging and of ectopic calcification in kidney, heart and lung [100]. Taken together, these results suggest that hypervitaminosis D may be a key factor in the accelerated aging phenotype of Fgf-23 and Kl- deficient mice [105,111]. In human beings, inactivating mutation in Fgf-23 and mutation in Kl gene lead to familial tumoral calcinosis a disorder marked by hyperphosphatemia and ectopic calcification. In contraposition to FGF-23 and KLOTHO KO mice, these patients have normal circulatory levels of PTH and 1,25(OH)₂D and do not have accelerated aging [106].

Few studies have been conducted to evaluate the safety of high dose vitamin D supplementation and there is no consensus about ideal levels of vitamin D intake. In the clinical context, the adverse effects of vitamin D have been mostly associated with acute manifestations of severe hypercalcemia secondary to vitamin D toxicity. The spectrum of symptoms ranges from fatigue, constipation, back pain, and forgetfulness to nausea, vomiting, weight loss and refractory status epilepticus. Hypercalcemia due to excess ingestion of vitamin D in dietary supplements [112-116] or in milk products [117,118] has been sporadically reported, almost always in individuals with impaired renal function. Aloia et al [119] conducted a six month prospective study on 138 white and African American subjects to analyze the intake of vitamin D₃ needed to achieve a targeted plasma 25OHD level (> 30 ng/ml). The dosage reached 3,915 ± 840 IU/day for blacks and 3,040 ± 1,136 IU/day for whites. No patient exceeded a calcium value of 2.65 mmol/L (or 10.6 mg/dL). Hathcock et al. [95] conducted an analysis of more than 20 publications, and concluded that there was no association between harm and intake of 10,000 IU/day vitamin D₃. These lines of evidence suggest that a high dose of vitamin D is necessary to induce toxicity related to hypercalcemia. On the other hand, no randomized clinical trial has been conducted in order to evaluate the long-term benefits and adverse effects of different vitamin D doses.

Indeed, there are also some clues indicating the potential harm effect of the indiscriminate use of vitamin D on a long-term basis. The Women’s Health Initiative Study showed no effect of daily calcium and 400 IU of vitamin D₃ on fracture risk, but as noted, there was a significant increase in the proportion of women who developed renal calculi [96]. Additionally, a recent study was undertaken to evaluate if a high-dose of cholecalciferol (500 000 IU) given orally once a year to community-dwelling older women would reduce falls and fractures [97]. The study was designed so that the vitamin D treatment would prevent decreases in 25-hydroxycholecalciferol over winter, address low adherence, and be a practical intervention easily translated to clinical practice. Unexpectedly, the study showed that, among older community-dwelling women, annual oral administration of high-dose cholecalciferol resulted in an increased risk of falls and fractures [97]. In another recent study Ensrud et al [120] assessed the relationship of 25-OHD in a cross-sectional and longitudinal prospective cohort study of more than 6000 women aged 69 years or more from the Study of Osteoporotic Fractures. The results showed that levels of 25-OHD ≤ 15 ng/ml (37.5 nmol/L) were strongly associated with risk of frailty. Moreover, those women exhibiting serum levels of 25-OHD between 15-19 ng/ml (37.5-47.5 nmol/L) at baseline were also at higher risk of death and frailty than those with levels of 20 to 29 ng/ml (50-72.5 nmol/L). Unexpectedly, women with baseline serum levels higher than 30 ng/ml (75 nmol/L) also showed a greater prevalence of frailty. These results agree with reports suggesting a U-shaped curve for serum 25-OHD and health parameters; in this case the incidence of frailty increases in elder women showing low levels (< 20 ng/ml) as well as high levels (> 30 ng/ml) of 25-OHD.

Even fewer data are available regarding the safety of vitamin D supplementation for children. In the first half of the last century, impairment of growth development was observed in a group of infants receiving 1800 to 4500 IU daily in comparison to a group receiving only 340 IU/day [121]. Another study was designed to compare the influence of a diet containing 1380 to 2170 IU or 350 to 550 IU of vitamin D during the first semester of life. No adverse effects were observed for any of the doses.

At this point, it has been well established that a small dose of vitamin D is necessary to prevent rickets in infants (400 IU) and children (600 IU) who are consuming low calcium diets. However, the additional benefit of a higher dose during this period of life is still to be determined. There are some data suggesting that excessive intake of vitamin D-fortified foods by children is associated with hypercalcemia. The number of pediatric patients with idiopathic hypercalcemia decreased from 7.2 cases per month during 1953-1955 to 3.0 cases during 1960-1961, as reported by the British Paediatric Association. The decline was coincident with the introduction of the new guidelines for fortification of food with vitamin D. It is noteworthy to observe that the estimated consumption of vitamin D decreased from 4,000 IU/ day to a range of 724-1343 IU/day between the two studies [122,123].

In summary, new insights into the integrative physiology of vitamin D has shed light on the functional plasticity of this unique micronutrient/sun-derived hormone. The ubiquitous presence of both the VDR and the machinery to synthesize 1,25(OH)₂D locally in diverse tissues and cells enables vitamin D to participate in the modulation of energy metabolism, cell proliferation, immune response, as well as the control of bone and mineral metabolism. However, limitations imposed by technical barriers, which are typical of steroid measurements, have hampered the standardization of the clinical diagnosis of vitamin D disorders. Consequently, limited advances have been made in furthering our understanding about the most appropriate regimen of vitamin D therapy for skeletal disorders. In this review, we identified gaps in knowledge about the impact of vitamin D deficiency on non-skeletal tissues and support further studies that define the relationship between serum 25OHD and chronic diseases.

Highlights.

• Vitamin D is one of the molecules whose production declines during aging.

• Vitamin D is a pleiotropic hormone, potentially influencing all human economy.

• Low as well as high vitamin D levels can have harmful effects.

• Great effort should be spent to uncover the ideal vitamin D requirement.