Classical role of vitamin D in metabolism and prevalence of deficiencies in US populations

It is well established that vitamin D functions to regulate Ca homeostasis. Studies on the hormonal response to dietary Ca deprivation have identified the vitamin D metabolite 1,25-dihydroxyvitamin D (1,25(OH)₂D) and parathyroid hormone (PTH) as major hormonal regulators of Ca homeostasis. Low serum Ca is sensed at the level of the parathyroid gland through a Ca-sensing receptor⁽Reference Hebert, Brown and Harris⁶⁾. The Ca-sensing receptor relays a signal that leads to the increased production and release of PTH into the circulation. Finally, PTH is a strong stimulator of the renal enzyme 1α-hydroxylase that catalyses the conversion of 25-hydroxyvitamin D (25(OH)D) to 1,25(OH)₂D, the hormonally active form of vitamin D⁽Reference Zierold, Mings and Deluca⁷⁾. The 1,25(OH)₂D produced by the kidney and released into the serum acts on intestine, bone and kidney to regulate Ca homeostasis. Thus, a dietary Ca load results in an acute lowering of serum PTH⁽Reference Heaney⁸⁾, whereas higher vitamin D status as well as overall better Ca status leads to a lower serum PTH even in the fasting state⁽Reference Harkness and Cromer^9, Reference Krall, Sahyoun and Tannenbaum¹⁰⁾.

The primary dietary sources of vitamin D are fortified dairy products; however, the availability of vitamin D for individuals is heavily influenced by the exposure to sunlight, as vitamin D is also produced in the skin. With trends towards reduction in milk intake and less sun exposure, much of the population of the USA is considered functionally vitamin D deficient⁽Reference Weaver and Fleet¹¹⁾; however, the prevalence of vitamin D deficiency is highly dependent on the working definition of deficiency. When vitamin D deficiency is defined as ≤ 37·5 nm-25(OH)D⁽Reference Nesby-O'Dell, Scanlon and Cogswell¹²⁾, a marker for vitamin D status, there are 4·2 % Caucasians and 42 % African-Americans that can be considered vitamin D deficient. Other definitions used indicate that serum levels of at least 80 nm-25(OH)D are necessary for individuals to be considered adequate for vitamin D status. The latter greatly elevates the estimates of vitamin D inadequacy⁽Reference Holick^13, Reference Heaney and Weaver¹⁴⁾. The net effect of the latter may be dietary recommendations to promote vitamin D intakes that promote maximal bone health⁽Reference Weaver and Fleet¹¹⁾ but may not reflect vitamin D requirements for overall health. In order to establish appropriate recommendations for vitamin D it is critical to consider the plethora of effects of vitamin D as a metabolic regulator that extend far beyond the well-recognised role of this nutrient in maintaining bone. Of particular importance to the discussions presented in this review are the roles of vitamin D in modulating body composition, energy homeostasis, insulin sensitivity and the development of insulin resistance, a predisposing factor for type 2 diabetes.

Roles of vitamin D status, calcium and diet as risk factors for diabetes

In discussing the role of vitamin D in the development of insulin resistance and diabetes, it is important to also consider both the level of vitamin D supplied and the dietary source. Because vitamin D is derived, at least in part, through the same dietary source of dairy products as Ca, it is difficult in epidemiological studies to separate the impact of dietary Ca from dietary vitamin D in the regulation of body composition. Assessing the impact of vitamin D from the diet is also particularly difficult since vitamin D produced by UV exposure is a much greater influence on vitamin D status than dietary vitamin D.

Inadequate vitamin D status is proposed to play a role in insulin resistance⁽Reference Boucher^2, Reference Mathieu, Gysemans and Giulietti^4, Reference Procopio and Borretta^15–Reference Palomer, Gonzalez-Clemente and Blanco-Vaca¹⁷⁾. The primary source of dietary vitamin D is fortified dairy products. The relationship of dairy intake with risk for type 2 diabetes was explored in a prospective analysis utilising the Health Professionals Follow-up Study (n 41 254 male participants; Table 1)⁽Reference Choi, Willett and Stampfer¹⁸⁾. The results of this study show that in 12 years of follow-up, the relative risk for type 2 diabetes in men in the top quintile of dairy intake was 0·77 (95 % CI 0·62, 0·95; P for trend = 0·003) compared with those in the lowest quintile. Thus, each serving per d increase in total dairy intake was associated with a 9 % lower risk for type 2 diabetes, independent of BMI⁽Reference Choi, Willett and Stampfer¹⁸⁾. However, there are many components of dairy foods, in addition to vitamin D, that may influence the risk for diabetes. As a consequence, there is considerable confounding interpretation of experimental outcomes.

Table 1

Epidemiological studies investigating relationship of vitamin D and diabetes or insulin resistance endpoints

25(OH)D, 25-hydroxyvitamin D; OGTT, oral glucose tolerance test; HOMA-IR, homeostasis model assessment of insulin resistance; PTH, parathyroid hormone.

Body composition, vitamin D and diabetes

Obesity is a major risk factor for the development of type 2 diabetes. A number of studies have been conducted to study the effects of dietary Ca on body fat mass⁽Reference Parikh and Yanovski^19–Reference Zemel, Shi and Greer²²⁾. Many studies, both epidemiological and intervention studies, support an inverse relationship between higher Ca intakes, enhanced by dairy product intake, and body fat mass⁽Reference Parikh and Yanovski¹⁹⁾. However, other intervention studies⁽Reference Barr^23–Reference Gunther, Legowski and Lyle²⁵⁾ do not support that Ca or dairy products will be effective in every situation but instead alternatively suggest that another dietary factor, probably vitamin D, may be responsible for inducing changes that improve body composition. Evidence that supports a link between vitamin D status and an increase in energy expended from a meal⁽Reference Teegarden, White and Lyle²⁶⁾ provides an explanation of the role of vitamin D to reduce adiposity. In the study of 250 overweight and obese adults of different ethnicities, serum 25(OH)D was shown to be inversely related to weight (r − 0·21), BMI (r − 0·18) and waist circumference (r − 0·14), but not fat mass⁽Reference McGill, Stewart and Lithander²⁷⁾. These results suggest that the relationship of vitamin D status was specific to waist circumference, an independent risk factor for disease, and not to fat mass.

Muscle mass is an important determinant of overall body insulin sensitivity and vitamin D status clearly has effects on muscle and physical activity. Changes in gait, difficulties in rising from a chair, inability to ascend stairs and diffuse muscle pain are classic symptoms of vitamin D deficiency⁽Reference Pfeifer, Begerow and Minne²⁸⁾. Results of epidemiological studies support a need for diet and lifestyle factors conducive to higher vitamin D status, assessed by 25(OH)D and greater muscle function⁽Reference Pfeifer, Begerow and Minne^28–Reference Mowe, Haug and Bohmer³⁴⁾. Several intervention studies also support the conclusion that improved vitamin D status improves muscle function⁽Reference Bischoff-Ferrari, Dawson-Hughes and Willett^35–Reference Sato, Iwamoto and Kanoko³⁷⁾. In a meta-analysis which included five double-blind randomised, controlled trials in elderly populations (mean age 60 years; n 1237), vitamin D supplementation reduced the corrected OR of falling by 22 % compared with patients receiving Ca or placebo, independent of Ca supplementation⁽Reference Bischoff-Ferrari, Dawson-Hughes and Willett³⁵⁾. A randomised controlled trial of patients (ninety-six women) with post-stroke hemiplegia receiving vitamin D supplementation of 1000 IU (25 μg)/d for 2 years reduced injurious falls by 59 %⁽Reference Sato, Iwamoto and Kanoko³⁷⁾.

The efficacy of vitamin D to promote muscle growth is supported by laboratory experiments. Rodents receiving diets containing high levels of vitamin D for 12 weeks had 8 % greater muscle mass compared with animals receiving suboptimal vitamin D levels⁽Reference Siddiqui, Chang and Zou³⁸⁾. In support of an important physiological role for 1,25(OH)₂D on muscle, vitamin D receptor (VDR) mice experience myopathy characterised by smaller muscle fibres⁽Reference Endo, Inoue and Mitsui³⁹⁾.

The purported impact of vitamin D on insulin sensitivity may be, at least in part, through an increase in muscle mass, which will improve overall body insulin sensitivity. Therefore, the role of vitamin D in potentially reducing adiposity, improving insulin sensitivity indirectly through improving muscle mass, and the reduction in vitamin D status with increased adiposity (Fig. 1) are factors that need to be carefully considered when reviewing the epidemiological literature on the relationship of vitamin D with diabetes or insulin sensitivity.

Fig. 1

Potential mechanisms underlying the putative relationship between vitamin D and insulin sensitivity or diabetes. 25(OH)D, 25-hydroxyvitamin D; 1,25(OH)₂D, 1,25-dihydroxyvitamin D; PTH, parathyroid hormone.

Epidemiological evidence linking vitamin D status to diabetes

Several epidemiological analyses of large datasets support that dietary vitamin D is associated with abnormal glucose homeostasis (Table 1). Cross-sectional results from the Women's Health Study show that in women (n 10 066; aged>45 years) dietary vitamin D is inversely associated with the prevalence of the metabolic syndrome, but this relationship is not independent of total Ca intake⁽Reference Liu, Song and Ford⁴⁰⁾. Women (n 83 779) participating in the in the Nurses' Health Study with no history of diabetes were followed for approximately 20 years⁽Reference Pittas, Dawson-Hughes and Li⁴¹⁾. The results suggest that although vitamin D intake had a minor influence on the risk of developing diabetes, a combined daily intake of>1200 mg Ca and>800 IU (20 μg) vitamin D was associated with a 33 % lower risk of type 2 diabetes with a relative risk of 0·67 (95 % CI 0·49, 0·90) compared with an intake of < 600 mg and 400 IU (10 μg) Ca and vitamin D, respectively. These results suggest that dietary Ca may enhance the impact of dietary vitamin D⁽Reference Pittas, Dawson-Hughes and Li⁴¹⁾. A primary shortcoming of linking vitamin D intake to health outcomes is a potential confounding of vitamin D source and adequacy of supply through the diet and sunlight exposure. For the study described above the dietary vitamin D may not adequately predict vitamin D status, but may instead be an indicator of a healthier diet that includes dairy products or may suggest greater physical activity which may involve increased exposure to the sun.

Serum 25(OH)D concentration, a preferred indicator of vitamin D status, has been correlated with improved glucose homeostasis and increased insulin sensitivity (Table 1). In cross-sectional analyses, serum 25(OH)D (using quartiles) of participants (n 6228) in the Third National Health and Nutrition Examination Survey (NHANES) was negatively related to high fasting glucose levels in non-Hispanic whites (OR 0·25) and Mexican blacks (OR 0·17) after adjusting for sex, age, BMI, leisure activity and time of year⁽Reference Scragg, Sowers and Bell⁴²⁾. In addition, serum 25(OH)D concentration was inversely associated with fasting and 2 h glucose, fasting insulin and the homeostasis model assessment of insulin resistance in Mexican Americans, with a trend towards significance for non-Hispanic whites (P = 0·06)⁽Reference Scragg, Sowers and Bell⁴²⁾. However, no relationship between 25(OH)D concentration and measures of glucose homeostasis was observed in non-Hispanic blacks⁽Reference Scragg, Sowers and Bell⁴²⁾. These contradictory results suggest a potentially race-specific inverse relationship of vitamin D status with serum glucose and possibly insulin resistance.

Analysis of the third NHANES database shows that serum 25(OH)D levels were lower in participants with diabetes (OR 1·98 lowest compared with highest quartile)⁽Reference Martins, Wolf and Pan⁴³⁾. In a secondary analysis of the results from the Rancho Bernardo Study (1070 men and women aged 44–96 years), PTH, but not serum 25(OH)D, was inversely associated with hyperglycaemia in men, but not women⁽Reference Reis, von Mühlen and Kritz-Silverstein⁴⁴⁾. Finally, the results of a study which included over 8000 adult (aged ≥ 20 years) men and women showed serum 25(OH)D concentration, in quintiles, was negatively related to fasting hyperglycaemia⁽Reference Ford, Ajani and McGuire⁴⁵⁾. In a cross-sectional analysis including 142 men (aged 70–88 years), controlling for age, BMI, physical activity, month of sampling, cigarette smoking and alcohol, serum 25(OH)D concentration was inversely correlated with glucose levels at 1 h and area under the curve of an oral glucose tolerance test (r − 0·23 (P < 0·01) and r − 0·26 (P < 0·01), respectively)⁽Reference Baynes, Boucher and Feskens⁴⁶⁾. In postmenopausal women (n 753), serum 25(OH)D level, but not PTH or 1,25(OH)₂D, was inversely related to fasting serum glucose, when controlled for BMI, weight and age⁽Reference Need, O'Loughlin and Horowitz⁴⁷⁾. Thus, there is substantial epidemiological data to support an inverse relationship between vitamin D status and abnormal glucose homeostasis.

Several case–control studies also support a relationship between vitamin D status and risk of diabetes or glucose intolerance (Table 1). For example, the prevalence of hypovitaminosis D (25(OH)D ≤ 37·5 nmol/l) was higher in diabetic patients (24 v. 16 %; P < 0·001) than in controls (390 subjects per group)⁽Reference Targher, Bertolini and Padovani⁴⁸⁾. In addition, serum 25(OH)D concentration was lower in newly diagnosed diabetics and those with impaired glucose tolerance compared with controls (total n 5677, aged 40–64 years)⁽Reference Scragg, Holdaway and Singh⁴⁹⁾. However, epidemiological studies cannot demonstrate cause and effect.

Several smaller epidemiological studies explored the relationship of vitamin D status with other indices of glucose homeostasis (Table 1). Serum 25(OH)D levels correlated positively with insulin sensitivity index and negatively with first- and second-phase insulin response in normal healthy glucose-tolerant subjects in a cross-sectional study⁽Reference Chiu, Chu and Go⁵⁰⁾, suggesting a better insulin sensitivity with higher vitamin D status in healthy young individuals independent of changes in weight. In addition, in another study 95 % of at-risk (defined by glucose levels) and 80 % of low-risk subjects were vitamin D deficient (serum 25(OH)D < 11 ng/ml) and 30 min blood glucose during a glucose tolerance test, serum insulin and C-peptide levels were correlated with serum 25(OH)D concentrations in at-risk subjects (n 44) and even in the ‘not-at-risk’ subjects (n 15)⁽Reference Boucher, Mannan and Noonan⁵¹⁾. These results support that lower vitamin D status may be a significant risk factor for glucose intolerance.

Increased serum PTH may contribute to the risk for diabetes due to its relationship with obesity. Increased levels of fasting PTH have been hypothesised to influence increased levels of body fat mass⁽Reference McCarty and Thomas⁵²⁾. In a meta-analysis of studies investigating the relationship of primary hyperparathyroidism with indices of body weight, in thirteen studies, patients with primary hyperparathyroidism were 3·34 (95 % CI 1·97, 4·71) kg heavier (P < 0·00001) than age-matched controls or had an increased BMI of 1·13 (95 % CI − 0·29, 2·55) kg/m² (P = 0·12) compared with controls⁽Reference Bolland, Grey and Gamble⁵³⁾. Studies in young adults with PTH in normal limits show that serum PTH levels are higher in the obese than in the non-obese⁽Reference Bell, Epstein and Greene^54, Reference Atkinson, Dahms and Bray⁵⁵⁾. Thus, serum PTH may be involved in regulating adiposity, a major risk factor for diabetes.

Vitamin D status, diabetes and controlled interventions

Several intervention studies support that vitamin D supplementation may affect glucose homeostasis or insulin resistance (Table 2). Non-diabetic adults (aged ≥ 65 years; n 314) received Ca (500 mg/d) and vitamin D (700 IU; 17·5 μg) for 3 years in a double-blind, randomised, controlled trial which was designed for bone-related outcomes⁽Reference Pittas, Harris and Stark⁵⁶⁾. In a post hoc analysis, there was no effect of the intervention in participants when fasting glucose concentrations were within normal limits at the initiation of the study ( < 5·6 mmol/l; n 222). However, non-diabetic subjects with fasting glucose above normal (5·6–6·9 mmol/l; n 92) displayed a reduced change in fasting plasma glucose at 3 years compared with those on placebo (0·02 mmol/l (4 mg/l) v. 0·34 mmol/l (61 mg/l), respectively; P = 0·042) and a lower increase in homeostasis model assessment of insulin resistance (0·05 v. 0·91; P = 0·031)⁽Reference Pittas, Harris and Stark⁵⁶⁾. Therefore supplementation with Ca and vitamin D may attenuate the development of insulin resistance that occurs over time and rescue the progression of abnormal glucose metabolism toward insulin resistance and type 2 diabetes.

Table 2

Intervention studies investigating the relationship of vitamin D and insulin resistance or diabetes

HOMA-IR, homeostasis model assessment of insulin resistance; GTT, glucose tolerance test; 1,25(OH)₂D, 1,25-dihydroxyvitamin D; OGTT, oral glucose tolerance test.

* 1 IU = 0·025 μg vitamin D.

The largest intervention study to examine the influence of dietary vitamin D and Ca on health risks is the Women's Health Initiative, and diabetes risk was examined in secondary analysis from this trial⁽Reference de Boer, Tinker and Connelly⁵⁷⁾. In this study, postmenopausal women (n 33 951) without diabetes at baseline were assigned to received either 1000 mg Ca/d plus 400 IU (10 μg) vitamin D or placebo. The women were followed for a median of 7 years and there was no change in the risk for developing diabetes with the treatment. Unfortunately, these study results need to be interpreted with caution, as the level of vitamin D supplementation may have been too low to achieve a significant change in vitamin D status. Thus, recent studies demonstrate that a higher level of intake may be required for optimal health⁽Reference Weaver and Fleet¹¹⁾.

Shorter-term intervention studies also demonstrate a positive impact of vitamin D or its active metabolites on insulin resistance. Treatment of type 2 diabetic females with oral vitamin D for 1 month led to an anticipated increase in 25(OH)D but also led to a reduced first-phase insulin secretion and caused a substantial reduction in insulin resistance of 21·4 %⁽Reference Borissova, Tankova and Kirilov⁵⁸⁾. Treatment of uraemic patients with vitamin D and Ca⁽Reference Allegra, Luisetto and Mengozzi⁵⁹⁾ or treatment with intravenous 1,25(OH)₂D⁽Reference Mak⁶⁰⁾ also improved insulin response. In contrast, in a double-blinded, placebo-controlled, cross-over trial in diabetics, acute administration of 1,25(OH)₂D did not affect fasting or stimulated glucose, insulin or C-peptide⁽Reference Orwoll, Riddle and Prince⁶¹⁾. On the other hand, a single dose of vitamin D (300 000 IU (7500 μg) intramuscular injection) in type 2 diabetics resulted in improvements in the oral glucose tolerance and glucose-stimulated increase in serum insulin levels in 4 weeks compared with placebo controls⁽Reference Raghuramulu, Raghunath and Chandra⁶²⁾. However, a single oral dose up to 450 000 IU (11 250 μg) vitamin D did not alter these parameters, even up to 12 weeks following the dose⁽Reference Ayesha, Raghunath and Sesikeran⁶³⁾, suggesting that higher doses of vitamin D may be required orally to achieve improvements in insulin sensitivity compared with intramuscular injections. The results of these intervention studies provide support that vitamin D specifically can improve insulin sensitivity.

Genotypic links to vitamin D action and insulin resistance

Support for a relationship of vitamin D in mediating improved insulin resistance is found in assessment of responsiveness of gene transcripts whose protein products regulate vitamin D metabolism. There are two known polymorphisms in exon 11 of the vitamin D-binding protein (DBP) gene that result in amino acid variants: at codons 416 GAT - - >GAG (Asp- - >Glu) and 420 ACG- - >AAG (Thr- - >Lys) which are the genetic basis for the three common electrophoretic variants of DBP (Gc1F, Gc1S and Gc2). These variants of DBP, the serum carrier of vitamin D metabolites, are associated with higher risk for type 2 diabetes or prediabetic phenotypes in several populations⁽Reference Szathmary^64–Reference Hirai, Suzuki, Hinokio and Hirai⁶⁸⁾, but not others⁽Reference Ye, Dubois-Laforgue and Bellanne-Chantelot^69, Reference Klupa, Malecki and Hanna⁷⁰⁾. While the precise association of variant subtypes with predisposition to insulin resistance and type 2 diabetes is not completely understood, the relationship between variant expression and predisposition to insulin resistance provides further evidence of the importance of vitamin D and maintenance of glucose homeostasis.

An important mediator of vitamin D action is the VDR which functions as a transcription factor when bound to 1,25(OH)₂D. There is evidence that VDR genotype may affect insulin resistance, both in regards to insulin secretion (the Apa1 VDR polymorphism) and insulin resistance (the BsmI VDR polymorphism)⁽Reference Oh and Barrett-Connor⁷¹⁾. The association of the Fokl, Apal, BsmI and TaqI polymorphisms of the VDR gene with type 2 diabetes was explored in a case–control design (308 type 2 diabetic patients and 240 control cases). In this study, there was no association of the four VDR polymorphisms examined with type 2 diabetes⁽Reference Malecki, Frey and Moczulski⁷²⁾. In another study, the influence of BsmI VDR genotype in young males with low (n 752) and high (n 787) physical activity was investigated. Those with the BsmI VDR BB genotype had significantly higher levels of fasting glucose (n 137; 5·61 (sd 0·49) mmol/l) than gene carriers with the genotype Bb (n 370; 5·44 (sd 0·44) mmol/l) or bb (n 245; 5·38 (sd 0·44) mmol/l). Of the BB gene carriers, 47 % had fasting glucose levels>5·55 mmol/l compared with 36 % of Bb gene carriers and 34 % of bb gene carriers (P = 0·02)⁽Reference Ortlepp, Metrikat and Albrecht⁷³⁾. In another study, the distribution of alleles and genotypes of the four single nucleotide polymorphisms in intron 8 (BsmI, Tru9I, ApaI) and exon 9 (TaqI) of the VDR gene was similar in type 2 diabetics (n 309) and controls (n 143)⁽Reference Ye, Reis and Dubois-Laforgue⁷⁴⁾. Thus, the evidence supporting an association of VDR genotypes with risk for diabetes is conflicting.

Another critical protein in the regulation of vitamin D action is the vitamin D 1α-hydroxylase gene, which converts 25(OH)D to the active vitamin D metabolite 1,25(OH)₂D. The prevalence of two variants in the 1α-hydroxylase was determined in type 2 diabetic patients (n 291) and controls (n 231). There were no differences in the distribution of genotypes, haplotypes and haplotype combinations between the groups. However, the T-C/T-T heterozygous haplotype combination was more prevalent in the subgroup of obese type 2 diabetics (BMI ≥ 30 kg/m²) than in the controls (41·5 v. 28·6 %; P = 0·01), suggesting an association with the risk factor for diabetes, obesity⁽Reference Malecki, Klupa and Wolkow⁷⁵⁾.

Tissue, cellular and molecular actions of vitamin D to alter glucose homeostasis

Several mechanisms have been proposed to explain the impact of vitamin D on insulin sensitivity and glucose homeostasis (Fig. 1). Because dietary vitamin D and elevated serum 25(OH)D are well known as regulators of PTH and 1,25(OH)₂D, these are likely candidates to mediate systemic changes in glucose metabolism. The regulation of serum Ca via PTH and 1,25(OH)₂D following changes in dietary Ca has been proposed to mediate the effects of vitamin D, at least in part, on insulin resistance. Vitamin D and PTH have also been associated with a variety of other actions beyond their classical functions, including cell growth, differentiation and apoptosis. Both hormones have been shown to increase levels of intracellular Ca and other rapid signalling pathways in a variety of tissues including adipocytes and muscle cells⁽Reference Zemel, Shi and Greer^22, Reference Smogorzewski, Zayed and Zhang^76, Reference Ni, Smogorzewski and Massry⁷⁷⁾. In addition to its rapid actions in cells, 1,25(OH)₂D also mediates genomic regulation through the VDR, a member of the steroid hormone receptor family⁽Reference Christakos, Dhawan and Liu⁷⁸⁾. Thus, both hormones have the capability of regulating a variety of processes far beyond their classical actions in mediating Ca homeostasis.

It is well established that PTH regulates the activity of the renal 1α-hydroxylase to convert 25(OH)D to 1,25(OH)₂D; however, extra-renal 1α-hydroxylase enzymes have also been identified in a variety of tissues which may lead to the local production of 1,25(OH)₂D under conditions of high vitamin D status⁽Reference Zehnder, Bland and Williams⁷⁹⁾. The evidence suggests that these enzymes are not regulated by PTH⁽Reference Flanagan, Wang and Tangpricha⁸⁰⁾. When vitamin D status is improved (as evidenced by higher 25(OH)D levels), PTH levels are reduced. Total serum concentrations of 1,25(OH)₂D are subsequently reduced by the lower PTH levels. However, the increased 25(OH)D may lead to an increase in 1,25(OH)₂D locally at the tissues even with decreased fasting PTH. Thus, while the dogma exists that PTH and 1,25(OH)₂D levels are coordinated, this remains a controversial issue in normal physiology. Many tissues have been shown to express the 1α-hydroxylase⁽Reference Zehnder, Bland and Williams⁷⁹⁾, including muscle and adipocytes⁽Reference Li, Byrne and Chang⁸¹⁾.

There is also evidence to support an effect of 1,25(OH)₂D at multiple levels of insulin release and action. Insulin release is low in vitamin D-deficient rats⁽Reference Ishida, Suzuki and Someya^82, Reference Kadowaki and Norman⁸³⁾ and is enhanced by treatment with 1,25(OH)₂D⁽Reference Bourlon, Faure-Dussert and Billaudel⁸⁴⁾, potentially via synthesis of proteins and increased conversion of pro-insulin to insulin⁽Reference Bourlon, Billaudel and Faure-Dussert⁸⁵⁾. Results of studies suggest that the increase in insulin release mediated by 1,25(OH)₂D may involve increases in intracellular Ca through the phosphoinositide/protein kinase C pathway and facilitating Ca entry by Ca channels⁽Reference Billaudel, Bourlon and Sutter⁸⁶⁾ and that activation of the cyclic AMP-mediated pathway restores insulin release in vitamin D deficiency⁽Reference Bourlon, Faure-Dussert and Billaudel⁸⁷⁾. Further, the 1α-hydroxylase is expressed in pancreatic islet cells⁽Reference Bland, Markovic and Hills⁸⁸⁾. Consistent with the ability of vitamin D to enhance insulin secretion, following a single intramuscular injection of 100 000 IU (2500 μg) vitamin D, insulin and C-peptide concentrations at 30 min of an oral glucose tolerance test increased 8–12 weeks later in patients with elevated glucose concentrations, but not in normal subjects⁽Reference Boucher, Mannan and Noonan⁵¹⁾. Therefore, vitamin D is important for insulin secretion.

The active metabolite of vitamin D, 1,25(OH)₂D, also may be directly involved in altering insulin action. These actions may be either direct at the level of the adipocyte to alter insulin action or at the level of the pancreatic B cell to alter insulin release. Supporting data indicate that when adipocytes are incubated with 1,25(OH)₂D there is an observed decrease in insulin-stimulated glucose uptake⁽Reference Huang, Ishizuka and Miura⁸⁹⁾. Likewise there is a vitamin D response element sequence in the insulin receptor gene promoter⁽Reference Maestro, Davila and Carranza⁹⁰⁾ and cellular tests indicate increased transcription and protein expression of the insulin receptor induced by 1,25(OH)₂D⁽Reference Maestro, Campion and Davila⁹¹⁾. Therefore an adequate vitamin D status or adequate Ca intake will act to promote higher levels of 25(OH)D and reduced levels of PTH, and reduce 1,25(OH)₂D levels. An increase in 25(OH)D, the substrate for 1α-hydroxylase, could drive a local increase in 1,25(OH)₂D to promote an increased insulin receptor expression. These changes may occur despite a suppression of PTH levels with a net effect to insulin responsiveness in adipocytes with global consequences to induce insulin sensitivity. Diets containing whey protein that are also high in Ca and vitamin D act to increase insulin receptor mRNA expression in rodents. Therefore the data suggest that 1,25(OH)₂D may affect insulin resistance by enhancing insulin; however, the effect of improved vitamin D status alone on insulin receptor expression has not been explored.

PTH levels are also regulated by vitamin D status. Serum levels of 25(OH)D, a marker for vitamin D status, are inversely correlated with fasting levels of serum PTH⁽Reference Harkness and Cromer^9, Reference Krall, Sahyoun and Tannenbaum^10, Reference Pepe, Romagnoli and Nofroni⁹²⁾. Studies also support a negative relationship between serum 25(OH)D levels and PTH in healthy young women⁽Reference Teegarden, Legowski and Gunther⁹³⁾. It is intriguing that this study also demonstrates that 1,25(OH)₂D is positively related to 25(OH)D, and 1,25(OH)₂D shows no relationship with fasting serum PTH levels, contrary to the model that only PTH controls the rate of conversion of 25(OH)D to 1,25(OH)₂D to regulate serum concentration. Overall, Ca homeostasis mediated by higher vitamin D status limits levels of fasting serum PTH. Therefore vitamin D may improve insulin action is by its ability to reduce PTH levels⁽Reference Krall, Sahyoun and Tannenbaum¹⁰⁾. There is evidence that increased blood PTH is associated with insulin resistance or glucose intolerance⁽Reference Kumar, Olukoga and Gordon^94, Reference Malecki, Klupa and Wanic⁹⁵⁾. It is well established that there is an increased prevalence of type 2 diabetes mellitus (8 %) and glucose intolerance (40 %) in patients with primary hyperparathyroidism⁽Reference Taylor and Khaleeli⁹⁶⁾. In addition, fasting PTH levels were shown to be inversely correlated with insulin sensitivity index in fifty-two normotensive, healthy subjects, even after adjustment for potentially confounding factors⁽Reference Chiu, Chuang and Lee⁹⁷⁾, supporting the hypothesis that higher fasting serum PTH may mediate a reduction in glucose tolerance. These results support a role of physiological levels of PTH and 1,25(OH)₂D hormones in insulin action.

Likewise, PTH may mediate insulin resistance by reducing glucose uptake by liver, muscle and adipose cells. PTH treatment (16 h) decreased insulin-stimulated glucose transport⁽Reference Thomas, Rogers and Sleeman⁹⁸⁾ in an osteoblast-like cell type. GLUT1 mRNA was reduced in osteogenic sarcoma cells following the 16 h PTH treatment⁽Reference Thomas, Rogers and Sleeman⁹⁸⁾, suggesting a mechanism for the PTH-mediated reduction in glucose transport. In addition, PTH decreased insulin-stimulated glucose uptake in rat adipocytes⁽Reference Reusch, Begum and Sussman⁹⁹⁾. These studies suggest that PTH may elicit insulin resistance by reducing the number of glucose transporters (both GLUT1 and GLUT4) available in the membrane to promote glucose uptake. On the other hand, PTH has been shown to suppress insulin release⁽Reference Perna, Fadda and Zhou¹⁰⁰⁾ and to promote insulin resistance in adipocytes⁽Reference Chang, Donkin and Teegarden¹⁰¹⁾. Therefore, cumulatively, the results of these studies suggest that PTH may negatively affect insulin sensitivity through altering body composition and inhibiting insulin signalling.

Vitamin D, inflammation and emerging roles of vitamin D in reducing insulin resistance

The pathogenesis of diabetes is complex, and one factor proposed to mediate an increase in insulin resistance is inflammation, such as occurs in obesity. Several studies support a role for vitamin D and 1,25(OH)₂D, as an anti-inflammatory agent. For example, 1,25(OH)₂D inhibits the release of the pro-inflammatory cytokine TNFα and regulates the activity of NFκB⁽Reference Zhang, Yuan and Sun^102–Reference Cohen-Lahav, Douvdevani and Chaimovitz¹⁰⁶⁾, which functions as a mediator of TNFα pro-inflammatory actions, at multiple levels. In addition, 1,25(OH)₂D down-regulates the increased levels of inflammatory markers (TNFα, IL-6, IL-1, IL-8, cyclo-oxygenase-2, intercellular adhesion molecule-1 and B7-1) in monocytes from type 2 diabetic patients compared with monocytes from healthy controls⁽Reference Giulietti, van Etten and Overbergh¹⁰⁷⁾. In other models, 1,25(OH)₂D inhibits the synthesis and actions of pro-inflammatory PG by inhibiting cyclo-oxygenase-2 expression, increasing the expression of the enzyme which inactivated PG (15-PG dehydrogenase) and decreasing PG receptors. 1,25(OH)₂D influences several pathways known to regulate inflammatory responses, including increasing mitogen-activated protein kinase phosphatase 5 which down-regulates p38 mitogen-activated protein kinase activity⁽Reference Krishnan, Moreno and Nonn¹⁰⁸⁾. Therefore, vitamin D may also function to reduce the risk of diabetes by acting to reduce inflammatory responses.

New roles of vitamin D to regulate insulin resistance are emerging. One such area is the role of vitamin D in the non-enzymic glycation of proteins. The process occurs more rapidly in diabetic patients than in normal individuals and measures of HbA_1c in blood provide a clinical indicator of integrated blood glucose homeostasis during the previous 2–3 months. Glycated albumin similarly indicates glucose metabolism during the previous 21 d period. In addition to the role of glycation products in vascular complications, there is evidence that links impaired insulin signalling in skeletal muscle cells to glycated albumin⁽Reference Miele, Riboulet and Maitan¹⁰⁹⁾. Recent studies support a role for vitamin D as a vascular protective agent against the effects of advanced glycation endproducts, which are proposed to mediate the devastating consequences of diabetes on cardiac complications⁽Reference Talmor, Golan and Benchetrit¹¹⁰⁾.