Abstract
It has been demonstrated in several studies that serum calcidiol (25 OH vitamin D3) concentration is in a reversed and linear relationship with cancer risk. However, there are also studies showing no such association and some even suggest the opposite. The risk of pancreatic and oesophageal cancer seems to increase, when serum calcidiol concentration increases. A bias in these studies might be that their basic assumption is linear dependence of cancer on serum calcidiol concentration. Some studies suggest a U-shaped association between the disease and the serum calcidiol concentration. Evidence, in the literature, of the relationship between serum calcidiol concentration and disease is reviewed and an optimal level of 40-80 nmol/L (16-32 ng/ml) is suggested. Serum calcidiol seems to be a better predictor of cancer development than calcitriol (1α, 25 (OH)2 vitamin D3). A calcidiol insufficiency, as well as an insufficient solar exposure, is associated with an increased risk of several solid cancinomas. In a recent study, our group demonstrated that calcidiol is an active hormone in CYP24 (24-hydroxylase) deficient cells. In these cells, calcidiol and calcitriol act synergistically, therefore fluctuations of the serum calcidiol concentration may define the hormonal activity and cancer development. Conclusion: Serum calcidiol concentration and the risk of many common diseases and aging phenomena seem to show a U-shaped association suggesting a lower and upper limit for healthy serum calcidiol concentration. An imbalance of hormonal calcidiol rather than that of calcitriol is a risk factor in carcinomas and chronic diseases, which might be prevented by an optimal serum calcidiol concentration. Multiple daily dosing of cholecalcipherol or skin patches could best provide an optimal dosing and stable serum concentration. Alternatively, narrow-band UV-B lamps are a possible optimal solution, when given by trained personnel.
Rickets, osteomalasia and osteoporosis are the well-known outcomes of vitamin D3 insufficiency or deficiency (1). Recently our group found that bone stress fractures even without any significant osteoporosis are associated with low calcidiol (25OH vitamin D3) serum concentration among young men (2). The higher the serum calcidiol concentration the better it seems to be for bone health (3) and a trend for the recommendation of increased daily vitamin D intake is apparent. The latest recommendation by the Endocrine Society is 50-100 μg/D or that the calcidiol serum concentration should be 75-150 nmol/L (30-60 ng/ml) (4). However, it seems that there is no benefit of higher dose (6500 IU) compared to a low dose (800 IU) in the prevention of osteoporosis (5). In fact, a high vitamin D hormonal activity has a harmful effect on the bone (6, 7). In particular, a single annual dose (500,000 IU) seems to be harmful for bone health (8). An excess of calcitriol (1α,25-dihydroxy vitamin D3) action in FGF-23 (fibroblast growth factor-23) deficient mice causes osteoporosis (9) as does a lack of vitamin D action in VDRKO (vitamin D receptor knockout) mice (10, 11). Thus, the action of hormonal forms of vitamin D on bone health seems to be U-shaped, too high as well as too low action is harmful. This article reviews evidence that too little as well as too much vitamin D might be harmful for health and may increase the risk of some types of cancer.
During the past 10-15 years accumulating evidence has shown that vitamin D3 insufficiency appears to be associated with several chronic diseases. Therefore it has been hypothesized that calcidiol in sufficient serum concentrations could prevent such diseases as cancer (12), muscle weakness (13), respiratory infections (14), autoimmune diseases (15), diabetes (16), hypertension (17), cardiovascular diseases (18), multiple sclerosis (19), seasonal affective disorder (SAD) (20) and behavioral abnormalities (21), which has been supported by some double-blind placebo-controlled prevention trials (22). However, it is not known, what is the optimal cholecalcipherol dose and what are possible adverse effects, when supplementation is used for long periods of time.
Numerous in vitro and in vivo studies have shown that vitamin D potently inhibits cell proliferation in a wide range of normal cell types and carcinomas such as cancer of the mammary gland, prostate, colon, skin and brain, myeloid leukaemia cells and many others (23). Several cancer cell studies in vitro have suggested that hormonal forms of vitamin D3 (cholecalcipherol hormones) can regulate their mitotic activity and differentiation (for a review see (24)). Mechanisms in vivo may include regulation of the immune defense system and expression of growth inhibitory factors as well as a direct regulation of the cell cycle (25). Epidemiological studies suggest that nearly 20 types of cancer are inversely correlated with solar ultraviolet-B (UV-B) levels or with the availability of vitamin D (26). If the hypothesis on the role of calcidiol in cancer were valid, it could be expected, that the risk of some carcinomas would be reduced in individuals with skin cancer, since skin carcinomas are markers of high sun exposure. The high sun exposure would produce more vitamin D, which would protect against some solid carcinomas. In fact, in a large scale study of primary skin cancer including melanoma, basalioma and squamous cell carcinomas (27) a reduced risk of all solid carcinomas was found by our group, but only in the sunny countries (Spain, Singapore and Australia) and not in the Nordic countries. A clear protective effect of sun exposure was found for cancer of the stomach, colorectum, liver and gallbladder, pancreas, lung, mammary gland, prostate, bladder and kidney, which has been supported by other studies (26, 28). While the reason for the lack of a protective effect in individuals with primary skin cancer in Nordic countries is not known, a possible explanation could be the high annual fluctuation of serum calcidiol in Nordic countries (29). In addition, childhood sunburns are a risk for melanoma, but they seem to protect against prostate cancer in adulthood (30). Thus, it seems that several cancer types are associated with sun exposure and vitamin D availability.
Until now the view on the vitamin D endocrine system has been simple. There is only one hormone, calcitriol acting via a single vitamin D receptor (VDR). However, the serum concentration of calcitriol is extremely low and it has a low binding affinity to vitamin D binding protein (DBP), therefore calcitriol does not enter the cell easily, because DBP endocytosis via cubilin-megalin is necessary for its hormonal action (31). Therefore the paradigm has now been modified: calcidiol bound to DBP enters the cell and is there converted via 1α-hydroxylation to an intracrine factor, calcitriol mediating all the actions. Only some of the vitamin D responses can be explained by the circulating calcitriol. New candidate hormones make the vitamin D endocrine system more complex as discussed below.
New Cholecalcipherol Hormones
Vitamin D3 (cholecalcipherol) is biologically inactive, and therefore it should not be classified as a vitamin despite being partially obtained from food. Cholecalcipherol needs to be converted through hydroxylations to active cholecalcipherol hormones in several tissues.
It is interesting that calcidiol was first regarded as an active metabolite in 1968 (32), but the idea was immediately rejected, when the more active calcitriol was found (33). Since 1971, only calcitriol, but not calcidiol, has been thought to be the active cholecalcipherol hormone. During the past decade, 1α-hydroxylase (CYP27B1), responsible for the conversion of calcidiol to 1α-calcitriol, has been found in many vitamin D target tissues besides the kidney (34, 35). Thus an assumption that the local production of 1α-calcitriol is regulating cell growth and differentiation (intracrine regulation) was made. However, the low concentrations of locally produced calcitriol (36) and the normally low serum level make it unlikely that calcitriol could elicit the effects. The half-life of calcidiol in the circulation is about two weeks while that of calcitriol is less than four hours (37). In vitro, calcitriol can produce biological responses in serum-free medium thus in favoring the “free hormone hypothesis”. However, the concentrations of 1α-calcitriol needed for the in vitro responses are usually 100-1000-fold above the physiological levels. Moreover, the serum level of calcidiol is approximately 1000 times higher than that of calcitriol. Since only 0.04% of calcidiol and 0.4% of calcitriol are free in plasma and the rest are tightly bound to either DBP or serum albumin (38). The concentration of free calcidiol is 100 times higher than that of free calcitriol. Based on the free hormone hypothesis (39), calcidiol is accessible to the target cells is at 100-fold higher concentration than calcitriol, but the absolute concentrations of free hormone in the serum are below the concentrations known to give any biological response. On the other hand, based on the bound hormone hypothesis (31, 40, 41), calcidiol could be taken up by the target cells approximately 1000 times more effectively than calcitriol. If calcidiol were an inactive metabolite, it would thus competitively inhibit the action of calcitriol. The binding affinities of the metabolites to the VDR have been measured in different cell types with varying results, but the binding affinity of calcitriol is only approximately 50-fold higher than that of calcidiol (42) suggesting that in the physiological situation most of the VDR molecules are occupied with calcidiol and therefore calcitriol would not act in vivo.
In order to solve the above dilemma, our group studied the action of calcidiol in experiments, where the role of 1α-calcitriol was eliminated either chemically or genetically. By blocking 1α-hydroxylase activity with a specific enzyme inhibitor, calcidiol itself was found to regulate gene expression in human primary prostate stromal cells (43) andmouse primary prostate cells (44). In addition, calcidiol promoted the activity of the 24-hydroxylase gene promoter in MCF-7 human breast cancer cells and inhibited the growth of LNCaP human prostate cancer cells (43). Also in isolated the 1α-hydroxylase knockout mouse kidney and skin cells, the gene regulatory action of calcidiol was observed (44). In addition, calcidiol seems to inhibit precancerous and alveolar lesions in mouse mammary organ culture systems derived from 1α-hydroxylase knockout mice (45). All these data demonstrate clearly that calcidiol has an inherent hormonal activity regulating genes and cell proliferation.
All the products of 24-hydroxylation (CYP24) were earlier presumed to be inactive degradation products. However, accumulating evidence suggests that at least 24,25 dihydroxy cholecalcipherol (24,25-dihydroxy vitamin D3) might be biologically active in chondrocytes (46). This action seems to be mediated by the membrane receptor (mVDR), because the effect can be detected in nuclear VDR−/− cells. 24,25 dihydroxy cholecalcipherol seems to act also through the nuclear VDR stimulating differentiation of human osteoblasts as well as bone mineralization (47). Furthermore, 24,25 dihydroxy cholecalcipherol seems to be necessary for bone fracture healing, since the callus formation is impaired in CYP24−/− mice (48). It is possible that new active calcipherol hormones will be found in the future. It has been proposed that 20-hydroxy vitamin D3 (49) and 1α,25 (OH)2-3-epi vitamin D3 (50) are biologically active. Before accepting them as new calcipherol hormones, it is necessary to demonstrate that they are physiologically present in sufficient concentrations to induce biological response. Calcitriol, calcidiol and 24,25 dihydroxy cholecalcipherol fulfill the basic criteria (sufficient physiological concentration and specific responses) to be hormones (Figure 1). Using DNA microarray, the gene inductions by calcitriol, calcidiol and 24,25 dihydroxy cholecalcipherol were measured by our group at concentrations slightly above the physiological levels in CYP27B1−/− cells (Lou et al., to be published). Only a few genes are induced by all the metabolites, but the majority of the genes are unique to each.
Synergistic Action of Calcitriol and Calcidiol
A strong argument against the “hormonal calcidiol” concept is the phenotype of 1α-hydroxylase knockout mice (CYP27B1−/−), which is quite similar to that of VDR−/− mice (51). The CYP27B1−/− mice have an elevated serum calcidiol concentration, but it cannot prevent the phenotype. This discrepancy could be explained by our recent finding of a mutual synergistic action of calcidiol and 1α-calcitriol on CYP24 expression(44) (Figure 2). The combined effect of 250 nM calcidiol and 0.1 nM calcitriol was significantly higher than the effect of either alone. In another experiment with mouse CYP27B1−/− primary keratinocytes, the synergistic effect was even more pronounced: 100 nM calcidiol or 0.05 nM calcitriol alone caused 0.9-fold and 1.4-fold expression of CYP24, respectively, whereas their combined effect was 475-fold induction of the gene (Figure 2). Although in vitro experiments cannot be directly compared with the situation in vivo, it is important to note that in this combination experiment the concentrations used were closer to the physiological levels than those normally used in vitro (10-1000 fold higher). It is interesting that the response to the higher combined doses was lower suggesting a kind of resistance. Furthermore, changing the concentration ratio of calcidiol and calcitriol the synergistic effect varied significantly. The mechanism of synergism is not known, but a possible explanation could be homodimerization of two VDRs occupied by two different ligands. Calcidiol could bind to the AP-site and 1α-calcitriol to the GP-site causing different conformational changes allowing apoVDR homodimer stabilization (52-58). Another possibility is that VDRs occupied with calcidiol or calcitriol heterodimerize with RXR (retinoic acid × receptor), but they recruit different transcription factors and act in concert at the chromatin level. Because the two ligands cause different conformational changes of the VDR (44), it is possible that different transcription factors are bound. A third possibility is that the ligands bind different receptors such as glucocorticoid or thyroid receptors, because calcidiol and calcitriol can bind firmly to them (59).
The finding of synergism suggests that, in fact, both calcidiol and calcitriol are important hormones and they act in concert. Therefore, calcidiol with its variable serum concentrations has always been clinically and epidemiologically more important than 1α-calcitriol, which shows less fluctuation of its serum concentration. The synergism between the calcipherol hormones seems to be bell-shaped and may explain the U-shaped responses to serum calcidiol concentrations as described below.
U-shaped Health Responses to Serum Calcidiol
A basic assumption in almost all the epidemiological studies is that the association between the serum calcidiol concentration and disease risk is linear. Therefore the reference is either the lowest or highest quartile or quintile. If the association were non-linear, the analysis would lead to a non-significant result. The problem is evident in the studies of Garland's group where moderate (not high) elevations of the serum calcidiol concentration reduced the risk of colon cancer (60), but later higher serum concentrations were found to protect from colon cancer (61). The highest quartile would give variable results, because high serum calcidiol values are rare especially during wintertime. Also very low values are not common in a normal population. This means that serum calcidiol concentrations of the quartiles are very close. Physiologically, this kind of comparison does not make sense. When our group used a physiologically more meaningful quintile division (each with equal concentration limits of 20 nmol/L) a U-shaped association was found between the prostate cancer risk and the serum calcidiol concentration in the Scandinavian countries (62). Our group found that the smallest risk of prostate cancer was found at the serum calcidiol 40-60 nmol/L (16-24 ng/ml). Both lower and higher calcidiol concentrations were associated with a significantly increased risk of the prostate cancer. Later Faubel-Badger et al. (63) argued that neither reversed linear nor significant U-shaped association could be found for prostate cancer in Finland. However, comparison of their results with ours (Figure 3) shows a remarkable resemblance between the two, although the U-shaped association by Faubel-Badger et al. was statistically not significant.
Until now several studies have demonstrated a U-shaped association between the disease risk and the serum calcidiol concentration or an increased risk at higher concentrations: Prostate cancer (62, 63); Mammary cancer (64, 65); Colon cancer (60); Oesophageal cancer (66); Pancreatic cancer (67); Cancer mortality (68, 69); Cell proliferation (70); All cause mortality (69, 71); Cardiovascular mortality (69, 72); Cardiovascular complications in chronic kidney disease (73); Vascular calcification (74); Small-for-gestational-age births (75); Falls and fractures in elderly women (8); Osteoporosis (5, 6, 9, 10, 51, 76); Schizophrenia (77); Common single nucleotide polymorphism (78); Deafness (79-82); Aging (9, 11, 51).
Most of the above mentioned diseases or biological responses show a U-shaped association with calcidiol concentration, but some diseases show an increased risk only at higher calcidiol serum concentration (mammary, oesophageal and pancreatic cancers). It is not known whether the variation in serum calcidiol concentration in these studies is caused by oral vitamin D substitution, sun exposure or endogenous metabolism. In their review article, Muenstedt and El-Safadi (65) come to the conclusion that arbitrary high doses of vitamin D could be harmful in breast cancer even though the results from clinical trials suggested that a high dose (1100 IU/day) of vitamin D is beneficial (83). However, a smaller dose (400 IU/day) also affected the breast cancer rate (84). Higher serum calcidiol concentrations were associated with significantly increased risk of oesophageal squamous cell carcinoma in men, but not in women (66). Similarly, Finnish male smokers with higher prediagnostic vitamin D status had an increased pancreatic cancer risk compared with those with lower status (67). The most convincing U-shaped associations with serum calcidiol have been shown for all cause, cardiovascular and cancer mortality rates (69, 71). Some of the diseases showing U-shaped association with serum calcidiol (all cause and vascular mortality, prostate cancer, vascular calcification, deafness and osteoporosis) are aging-related diseases. Since aging phenomena appear to show a U-shaped association with serum calcidiol (51), it is understandable that aging-related diseases show a U-shaped association. It is evident that besides vitamin D insufficiency, non-calcemic hypervitaminosis has significant harmful effects on health.
The U-shaped response is not unique to cholecalcipherol hormones, but is a common feature of many, if not all known hormones such as vitamin A, thyroid hormone and steroids. The mechanism of the U-shaped responses to calcidiol concentration is not known. A possible explanation is that a too low concentration is not sufficient to regulate beneficial genes, whereas a high calcipherol hormone concentration activates harmful genes. Another possible explanation is that hypervitaminosis D could lead to resistance. Haussler (85) pointed out that the up-regulation of CYP24 by calcitriol led to the consequent degradation of calcitriol (self-destruction). Since calcitriol is highly calcemic and has a high affinity to the VDR, its concentration needs to be tightly controlled. Our group demonstrated that calcidiol is also able to induce CYP24 and, in turn, to cause degradation of calcitriol (44). High serum calcidiol leads to an suppression of the synthesis of the biologically more active intracrine metabolite, calcitriol (86) and to weaker activity on the target cell. Although the concentration of calcidiol is high, the low calcitriol concentration leads to a weak synergistic action between these two metabolites and vitamin D resistance (44). The best indicator of the resistance is the serum 24,25 dihydroxy cholecalcipherol which does not increase significantly, when the serum calcidiol concentration is below or close to 50 nmol/L by moderate vitamin D substitution (87), but it does increase when calcidiol is above 75 nmol/L (30 ng/ml) using a high weekly dose (88). In order not to activate resistance responses, moderate daily doses of vitamin D should be used.
Could too much Vitamin D Be Harmful?
Because vitamin D insufficiency seems to be associated with a number of adverse health effects, many scientists and health professionals are advocating increasing vitamin D status to above 75 nmol/L (30 ng/ml) of calcidiol, a concentration that has been suggested to maintain normal parathyroid hormone levels (89). This could be accomplished either throughincreased exposure to ultraviolet radiation or by taking high doses of vitamin D supplements (>2000 IU/D=50 μg/D) (89). The new clinical practice guideline of the Endocrine Society released calling for calcidiol serum concentrations of 75-150 nmol/liter (4) would mean that almost all people in the United States and in countries located at temperate latitudes need oral substitution, at least during the winter.
The main health risks from excessive vitamin D are associated with its role in the regulation of plasma calcium concentrations (ectopic calcification of blood vessels, kidney stones etc.). Excessive vitamin D can cause hypercalcemia by increasing intestinal calcium absorption or by increasing mobilization of bone calcium. While hypercalcemia is uncommon with intakes less than 10,000 IU/day (90), knowledge of non-calcemic adverse effects that could be associated with the maintenance of high vitamin D status is currently increasing.
The following three reasons explain why the new recommendation by the Endocrine Society is too high.
Oral vitamin D3 is rapidly absorbed from the intestine. Its conversion to an active hormone, calcidiol (44), in the liver is not rate limited, thus leading to a peak serum concentration within 2-3 h. The high calcidiol concentration leads to vitamin D resistance via 24-hydroxylation of calcitriol. The higher the baseline calcidiol serum concentration is the smaller the increase after the supplementation (91) and the higher the dose of vitamin D the smaller is the increase in the serum calcidiol per μg or IU of the supplement (3). Furthermore, with increasing vitamin D doses the 24,25 dihydroxy vitamin D3 concentration in the serum also increases. All the findings suggest that a defense system against too high vitamin D supplements exists. The dose or the method of administration should be such that the inactivation mechanisms are not activated.
Several significant health responses, all cause death rate, cancer and cardiovascular mortality, prostate cancer (perhaps some other carcinomas), osteoporosis, bone fractures in elderly women and all ageing phenomena show U-shaped dependency on serum calcidiol concentration (8, 51, 62, 64, 65, 68, 92). Based on these findings, the optimal serum concentration appears to be 40-80 nmol/L (6-32 ng/ml), i.e., half of that recommended. This means that the daily dose should be below 20 μg, most likely 10 μg, except it could be higher for elderly people (=20 μg).
Besides the serum concentration for optimal health, oral administration of vitamin D might not be the best choice. Skin is the main source of vitamin D in humans and food is the secondary source when no fortification is used. In contrast to the fluctuation after oral vitamin D, UV-B exposure of the skin yields an even calcidiol serum concentration within 2 weeks after the exposure. Of course, oral vitamin D could be divided into multiple small daily doses, but that is inconvenient. Vitamin D containing topical patches would be more physiological and easy to use. Alternatively, use of UV-B lamps could be a physiological solution. They could replace the hazardous UV-A lamps of the solarium and be used under the supervision of trained personnel.
The Institute of Medicine (IOM) came to moderate conclusions namely for practically all persons serum 25OHD levels of at least 50 nmol/L (20 ng/ml) are sufficient and serum concentrations above75 nmol/L (30 ng/ml) are not consistently associated with increased benefit (93, 94). The present review very much agrees with their conclusions, but one problem remains. The IOM is recommending 10 μg/day for the children, thus a child weighing 5 kg would get a 10-fold dose compared to a 50 kg adult. It seems that the pediatric vitamin D dosing needs careful re-evaluation, because the possible increased risk of some chronic diseases would appear several decades later.
In conclusion, vitamin D is a typical hormone like vitamin A, steroids, thyroid hormone etc in that it is harmful in both too low and too high concentrations. From the endocrine point of view, the most successful fortification, substitution or hormone replacement therapy is an appropriate dosage and route of administration, which does not activate the feedback mechanisms, vitamin D resistance and health risks.
- Received September 28, 2011.
- Revision received November 30, 2011.
- Accepted November 30, 2011.
- Copyright© 2012 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved