Background: Clinical assessment of vitamin D status often relies on measuring total circulating 25-hydroxyvitamin D3 (25OHD3), but much of each vitamin D metabolite is bound to plasma vitamin D–binding protein (DBP), such that the percentage of free vitamin is very low. We hypothesized that measurement of free rather than total 1,25-dihydroxyvitamin D3 [1,25(OH)2D3] and 25OHD3 may provide better assessment of vitamin D status. We therefore aimed to assess vitamin D status in men with idiopathic osteoporosis, in whom possible secondary causes of osteoporosis had been excluded, and to determine the extent of change in biologically active “free” vitamin D caused by variation in plasma DBP concentrations.
Methods: We measured 1,25(OH)2D3 and 25OHD3 in plasma samples from 56 men with idiopathic osteoporosis [mean (SD) age, 59.6 (13.6) years; range, 21–86 years] and 114 male controls [62.4 (10.4) years; range, 44–82 years].
Results: Mean total plasma 25OHD3 in the 56 men with osteoporosis and the 114 controls was 44.7 (21) and 43.3 (17) nmol/L, respectively; total plasma 1,25(OH)2D3 measured in randomly selected men with osteoporosis (n = 50) and controls (n = 50) was 90 (37) and 103 (39) pmol/L, respectively. Mean plasma DBP was significantly higher (P <0.001) in men with osteoporosis [224 (62) mg/L; n = 56] than in the controls [143 (34) mg/L; n = 114], but calculated free plasma 25OHD3 and 1,25(OH)2D3 were significantly lower in the osteoporotic men than in controls [6.1 (3.1) vs 9.1 (4.4) pmol/L (P <0.00001) and 77 (37) vs 142 (58) fmol/L (P <0.00001), respectively].
Conclusions: Measurement of total vitamin D metabolites alone, although providing a crude assessment of vitamin D status, may not give an accurate indication of the free (biologically active) form of the vitamin. The ratio of total 25OHD3 and 1,25(OH)2D3 to plasma DBP, rather than total circulating vitamin D metabolites, may provide a more useful index of biological activity. Further studies are required to substantiate this hypothesis.
The importance of 1,25-dihydroxyvitamin D3 [1,25(OH)2D3] 1 as a critical calciotropic hormone is well established, as it plays a pivotal role in plasma calcium homeostasis and skeletal mineral balance. It is now recognized that vitamin D has other functions and may play a role in immunity and cell differentiation (1)(2)(3). Recent animal and human studies have suggested that vitamin D deficiency may be an etiologic factor in the pathogenesis of autoimmune diseases such as autoimmune thyroid disorder (4), multiple sclerosis(5), juvenile idiopathic arthritis (6), and type 1 diabetes mellitus(7). In view of the growing awareness of the roles of vitamin D, accurate assessment of vitamin D status has become increasingly important. However, assessment of vitamin D status is relatively unreliable because it depends on measuring total circulating 25-hydroxyvitamin D3 (25OHD3) concentrations rather than determining the concentration of the free, biologically active component.
25OHD3 and 1,25(OH)2D3 are hydrophobic molecules with low plasma solubility and are therefore transported in the circulation bound to plasma proteins. A very large proportion of circulating 25OHD3 and 1,25(OH)2D3 is bound to vitamin D–binding protein (DBP) and albumin, and only 0.02%–0.05% of 25OHD3 and 0.2%–0.6% of total 1,25(OH)2D3 remain free or unbound (8). DBP has greater affinity for 25OHD3 than for 1,25(OH)2D3; the dissociation constants for the 2 metabolites differ by 10-fold (8). In mammals, DBP exhibits the same relative affinities for vitamin D2 and D3 metabolites. DBP (also known as Gc globulin, or Gc) is a 52- to 58-kDa glycoprotein that is synthesized in the liver. DBP circulates in plasma at concentrations 20-fold higher than the total amount of vitamin D metabolites (9). 25OHD3 and 24,25-dihydroxyvitamin D circulate in blood at concentrations ∼1000-fold higher than those of 1,25(OH)2D3; however, because the molar concentration of the circulating DBP in relation to all of the vitamin D metabolites is 20-fold higher, many DBP binding sites remain unoccupied. These unoccupied DBP sites bind most of the circulating 1,25(OH)2D3, and consequently, the percentage concentration of free hormone is very low. The physiologic consequence of the large molar excess of circulating DBP, which differs from other steroid hormone carrier proteins, is unclear. DBP has a single sterol binding site, and only 5% of the total DBP is usually occupied with vitamin D compounds. Therefore, under physiologic conditions, nearly all circulating vitamin D compounds are protein bound, which has a great influence on vitamin D pharmacokinetics. Protein-bound vitamin D metabolites will have limited access to target cells (10) and are likely to be biologically inactive. An implication of this observation is that the concentrations of the free rather than total forms of 1,25(OH)2D3 and 25OHD3 are likely to provide better assessment of functional vitamin D status. Because malabsorption of calcium has previously been reported in men with osteoporosis (11), we compared calculated free and total 25OHD3 and 1,25(OH)2D3 in men with idiopathic osteoporosis and male controls.
Materials and Methods
The study population comprised 56 men with idiopathic osteoporosis [mean (SD) age, 59.6 (13.6) years; range, 21–86 years] attending the Bone Clinic, Freeman Hospital, Newcastle Upon Tyne (United Kingdom) and 114 male controls [62.4 (10.4) years; range, 44–82 years] from Derby (United Kingdom). All participants in the study were Caucasian and were from 2 industrial cities in Northern England. The osteoporotic men all had a bone mineral density (BMD) t-score below −2.5 at either the femoral neck or lumbar spine. Of the osteoporotic men, 45 had a history of fractures (33 symptomatic vertebral, 6 hip, 2 forearm, and 4 other fractures) and 11 had no previous fractures. The fractures had all occurred at least 6 months before inclusion of the men in the study, and all blood samples were taken before the initiation of therapy. Underlying secondary causes of osteoporosis were excluded by medical history, physical examination, and laboratory investigation, which led to the exclusion of 66 men with osteoporosis. The laboratory investigation included full blood counts, erythrocyte sedimentation rates, biochemical profiles, thyroid function tests, and measurement of serum testosterone, sex hormone–binding globulin (SHBG), and gonadotropins (12). Previous work showed that 54% of men attending the Bone Clinic in Newcastle with vertebral fractures have an underlying secondary cause of osteoporosis (13).
The 114 male controls were a subset of the 198 men without low-trauma fractures who had previously taken part in a case–control study of distal forearm fractures (14) and subsequently agreed to have blood tests. Men were recruited from a preexisting database in Derby of 692 male controls and had all attended clinic for bone densitometry determinations in 1998. They had been recruited from the city of Derby and its environs via notices in general practitioners’ offices and in local newspapers. The study was approved by the local ethics committee, and informed written consent was obtained from all participants.
biochemical and bmd measurements
Possible differences attributable to the diurnal variation in vitamin D status were minimized by ensuring that all venesections were performed in the morning. Differences attributable to seasonal variation in vitamin D status were minimized by undertaking venesection throughout the year in both groups, so that the number of samples collected through the seasons were balanced. The following analyses were performed on serum, on an Olympus AU640: creatinine (inter- and intraassay CVs, 0.7%–1.0% and 2.2%–2.3%) and alkaline phosphatase (0.6%–3.5% and 2.1%–4.6%). All laboratory tests were subject to validation by use of National External Quality Assurance Schemes. Glomerular filtration rate was estimated from serum creatinine, body weight, and age by use of the Cockcroft–Gault formula.
All bone density measurements were performed by dual x-ray absorptiometry performed on a QDR 2000 (Hologic). In vivo imprecision for measurement with this system is 1.0% at the lumbar spine (L1–L4) and 1.5% for the femoral neck. BMD results were obtained as the areal density in g/cm2, but are also given as t- and z-scores. The t-score is the number of SD above or below the mean for healthy young men, whereas the z-score is the number of SD above or below the age-related normal mean (calculated by use of the manufacturer’s standard reference database).
ria for measurement of dbp
The total plasma DBP was determined by use of monospecific polyclonal goat anti-human DBP antibodies (Diasorin) in a competitive RIA (15). Briefly, microtiter plates were coated with the anti-human Gc-globulin antibody in triplicate and incubated overnight at 4 °C. The microtiter plates were then washed 6 times with phosphate buffer containing Tween. 125I-labeled DBP was added in fixed amounts (100 μL containing 80 000 cpm) and allowed to bind for 48 h. After incubation, the microtiter wells were again washed 6 times with phosphate buffer containing Tween and slapped dry; the radioactivity in the wells was then measured with a gamma counter (Packard Cobra). The intraassay CV, determined using 24 replicates of samples, was between 4.3% and 5.4%, and the interassay CV was 5.3%–7.8%.
assay for 25ohd3
The assay used was a liquid-phase RIA for the quantitative determination of 25OHD3 in human plasma (Immunodiagnostics Ltd.). Briefly, plasma 25OHD3 was extracted according to the manufacturer’s instructions, and portions of the supernatant were incubated with 125I-labeled 25OHD3 tracer and sheep anti-25OHD3 polyclonal antibody for 90 min. Separation of the antibody-bound tracer from free analyte was achieved by a short incubation with Sac-Cel® (solid phase anti-sheep IgG cellulose). The intra- and interassay CVs of the method are 5.3%–6.1% and 7.3%–8.2%, respectively.
assay for 1,25(oh)2d3
The assay for plasma 1,25(OH)2D3 was performed with commercially available reagents (Gamma-B reagent set; Immunodiagnostic Ltd.). Immunoextraction was carried out with an anti-1,25(OH)2D3 monoclonal antibody attached to a solid phase. Primary antibody was added to tubes containing calibrators and to the tubes containing the extracts, which were then incubated overnight at 4 °C. The following day, tracer was added and incubated at 25 °C for 2 h, after which Sac Cel solid-phase second antibody was added. After 30 min, wash solution was added to dilute the assay mixture, and the tubes were subjected to centrifugation. The liquid was decanted, the tubes were allowed to drain, and the 125I activity remaining in the pellets was counted. The intra- and interassay CVs of the method were 9.1%–11% and 9.6%–14%, respectively.
calculated free vitamin d metabolites
The free plasma 25OHD3 and 1,25(OH)2D3 concentrations were calculated by use of the following equations: where [T] and [F] are the total and free vitamin concentrations, respectively, and KaALB and KaDBP are the association constants for 25OHD3 and 1,25(OH)2D3 with albumin (ALB) and DBP. Most plasma 25OHD3 is bound to circulating DBP (85%–90%); however, plasma DBP concentrations show considerable individual variability. The remainder of plasma 25OHD3 is bound mainly to plasma albumin, which in our study can be ignored because this smaller fraction is almost constant. The plasma albumin concentrations in the male controls and men with osteoporosis were quite similar and within the reference interval. The affinity constant of albumin (KaALB) for 25OHD3 is much lower (6 × 105 L/mol) than the affinity constant of DBP (KaDBP) for 25OHD3 (0.6 × 109 L/mol) (8)(16)(17). 1,25(OH)2D3 was assumed to be bound primarily to DBP with the affinity constants of DBP and albumin for 1,25(OH)2D3 being 2.8 × 108 L/mol (18) and 4 × 105 L/mol (19), respectively. Previous studies have identified that a small number of phenotypes, such as 1S1S and 21S, account for ∼70% of all DBP phenotypes in Caucasian men in the United Kingdom (17); therefore, for calculating the free vitamin D metabolites, we used KaDBP values for the most common DBP phenotypes and 25OHD3 and 1,25(OH)2D3, respectively (8)(16)(17)(18).
Results are presented as the mean (SD). Data were analyzed by appropriate statistical packages (Microsoft Excel XL). We used the Mann–Whitney test to determine significant differences in BMD, anthropometric measurements, DBP, and vitamin D metabolites between osteoporotic men and male controls.
The ages and results of anthropometric and BMD measurements in men with osteoporosis and controls are summarized in Table 1⇓ . Although there was no statistically significant difference in body mass index (BMI), age, or height between the 2 groups, weight and BMD measurements were significantly lower in the osteoporotic men than in the controls. Plasma 1,25(OH)2D3 concentrations were determined in a randomly selected subgroup (n = 50) of male controls. The age, height, weight, BMI, BMD, and DBP and total vitamin D concentrations in these individuals were similar to those in the total control group (Table 2⇓ ).
The mean (SD) plasma DBP concentration was significantly higher (P <0.001) in men with osteoporosis [224 (62) mg/L; n = 56] than in the control group [143 (34) mg/L; n = 114; Fig. 1⇓ ]. Total plasma concentrations of 25OHD3 did not differ significantly between men with osteoporosis [44.7 (21) nmol/L; n = 56] and controls [43.3 (17) nmol/L; n = 114; Fig. 2A⇓ ]. There also was no significant difference between plasma concentrations of 1,25(OH)2D3 measured in randomly selected men with osteoporosis [90 (37) pmol/L; n = 50] and male controls [103 (39) pmol/L; n = 50; Fig. 2B⇓ ]. the circulating concentrations of calculated “free” 25OHD3 and 1,25(OH)2D3 in the osteoporotic men and controls are shown in Fig. 3⇓ . Mean calculated free plasma 25OHD3 concentrations were ∼33% lower (P <0.00001) in the osteoporotic men [6.1 (3.1) pmol/L] than in the controls [9.1 (4.4) pmol/L]. Calculated free plasma 1,25(OH)2D3 was also significantly lower (P <0.00001) in osteoporotic men [77 (37) fmol/L] than in the control group [142 (58) fmol/L]. However, the magnitude of difference in 1,25(OH)2D3 between osteoporotic men and controls was greater (42%) than that for plasma 25OHD3 concentrations (33%). Interestingly, although there was no correlation between total 25OHD3 and total 1,25(OH)2D3 (r <0.058), total plasma 25OHD3 and free 25OHD3 (r = 0.0.60), or free 25OHD3 and total 1,25(OH)2D3 (r = 0.138), calculated free plasma 25OHD3 and 1,25(OH)2D3 showed a significant positive correlation (r = 0.614; P <0.01).
All serum follicle-stimulating hormone, luteinizing hormone, SHBG, and testosterone values for the controls and men with osteoporosis were within laboratory reference intervals. The mean (SD) total serum alkaline phosphatase activity in men with osteoporosis was 81 (25) U/L (adult reference interval, 30–130 U/L).
The creatinine clearance [expressed as mL · min−1 · (1.73 m2)−1 and calculated by use of the Cockcroft–Gault formula] in the control group [72 (20); range, 36–129] and in men with osteoporosis [74 (21); range, 34–114] did not differ significantly. Thus, the mean estimated glomerular filtration rates for both groups were in the range thought to be consistent with mild renal impairment. However, none of the control group or patients with osteoporosis had an estimated glomerular filtration rate <30 mL · min−1 · (1.73 m2)−1, which is regarded as consistent with severe renal impairment.
The mean total 25ΟΗD3 concentrations for both osteoporotic men and controls in our study were below the concentration that is generally regarded as sufficient (50 nmol/L). Although we found no significant differences in plasma total 25ΟΗD3 and 1,25(OH)2D3 between osteoporotic men and controls, the calculated free plasma 25ΟΗD3 and 1,25(OH)2D3 concentrations were markedly lower in men with osteoporosis. Furthermore, there was no correlation between total 25ΟΗD3 and calculated free 25ΟΗD3, total 1,25(OH)2D3 and calculated free 1,25(OH)2D3, or total 25ΟΗD3 and total 1,25(OH)2D3 (r ≤0.138). There was nevertheless a significant positive correlation between free plasma 25OHD3 and free 1,25(OH)2D3 (r = 0.614; P <0.01). These findings suggest that total circulating concentrations of 25ΟΗD3 or 1,25(OH)2D3, although providing information about gross excess and deficiency of the hormone, do not provide information about the physiologically relevant portion of the hormone. The consequences of excessive DBP production and plasma concentrations have never been addressed.
The existence of vitamin D deficiency despite the presence of total circulating concentrations of 25ΟΗD3 or 1,25(OH)2D3 within reference values needs to be established by further investigations, including estimations of parathyroid hormone concentrations. The total mean serum alkaline phosphatase concentration in men with osteoporosis was within the reference interval; therefore, the gross sign of vitamin D deficiency, leading to secondary hyperparathyroidism and consequent increases in alkaline phosphatase, was not evident. Calculated creatinine clearance showed that mild renal impairment of almost similar extent and frequency affected both groups, but none of the men had any evidence of severe renal impairment.
Several studies have demonstrated that the free forms of vitamin D metabolites are more accessible to target cells and therefore produce a higher biological response in vitro and in vivo (20)(21)(22). For example, DBP-null mice fed a vitamin D–replete diet showed no biochemical or histologic evidence of secondary hyperparathyroidism, despite low total plasma concentrations of 25ΟΗD3 and 1,25(OH)2D3 (23). This observation supports the “free hormone hypothesis” of simple diffusion of the unbound sterol through the plasma membrane into the target cell. Furthermore, the extent of inhibition of lymphocyte proliferation and stimulation of 24-hydroxylase in keratinocytes and monocytes by 1,25(OH)2D3 correlates with the free rather than the total concentration of the metabolite (24).
The hypothesis that DBP may buffer free vitamin D metabolites and therefore guard against vitamin D intoxication (24) has been challenged by the observation that DBP-null mice are less susceptible to hypercalcemia and its toxic effects than are normal mice (23). A likely explanation for this finding is that DBP and DBP-bound vitamin D metabolites filtered through the glomerulus are reabsorbed by the endocytic receptor megalin into cells in the proximal tubules (25). Megalin-null mice have high urinary excretion of 25OHD3 and DBP, severe vitamin D deficiency, and bone disease (24), but the contribution of megalin-mediated endocytosis in the extrarenal cellular uptake of 1,25(OH)2D3 from circulating 1,25(OH)2D3–DBP is unclear.
Albumin and lipoproteins also serve as important carriers for vitamin D metabolites but have lower affinities than DBP. In the 2 groups of men in this study, albumin and lipoprotein concentrations were similar, with very little variation among individuals or between the 2 groups. Albumin binds a relatively small proportion (<10%) of total blood 25ΟΗD3 and 1,25(OH)2D3, whereas DBP binds ∼90% of 25ΟΗD3 and 1,25(OH)2D3 (16). Furthermore, circulating plasma DBP concentrations in osteoporotic men were significantly different from those in male controls, and we observed considerable variation between individuals. Thus, the plasma DBP concentration, rather than albumin concentration, is likely to have a more substantial effect on the biologically activity of 25OHD3 and 1,25(OH)2D3. The ratio of total vitamin D metabolites to plasma DBP may provide a useful index of the biological activity of the vitamin. In view of the relatively low KaALB values for both 25OHD3 and 1,25(OH)2D3, it follows that: Because KaDBP [DBP] is much greater than 1 and KaDBP is constant, [F] ≈ [T]/[DBP]; this suggests that the ratio of total vitamin D metabolites to plasma DBP may provide a useful index of the biological activity of the vitamin. Therefore, the conclusions of the study would be unaltered if, instead of the calculated free vitamin D metabolites, only the ratio of the respective metabolite to plasma DBP had been used.
The current assays for measuring free 25OHD3 and 1,25(OH)2D3 are technically demanding and time-consuming and not amenable to routine laboratory use; this is primarily attributable to the extreme hydrophobicity of 25OHD3 and 1,25(OH)2D3 (26)(27). However, in view of the increasing evidence for the importance of vitamin D deficiency as an etiologic factor in a variety of diseases (1)(2)(4)(5)(6)(7), there is a strong case for establishing a suitable method that can provide information about free (i.e., biologically active) 25OHD3 and 1,25(OH)2D3. The above considerations suggest that the measurement of total blood 25OHD3 and 1,25(OH)2D3 provides a gross index of the vitamin status but that more subtle and borderline cases of apparent vitamin deficiency may be missed. A more suitable method for the routine clinical assessment of these metabolites is therefore required.
We found higher DBP concentrations in men with osteoporosis than the male controls, which led to lower calculated free vitamin D metabolite concentrations. There are 2 possible explanations for this observation: an alteration in liver function or altered gene expression. Indeed, in a previous case–control study, we showed significantly higher concentrations of SHBG in men with symptomatic vertebral fractures compared with age-matched male controls (28). Another explanation for the increased DBP concentrations in men with osteoporosis may be altered gene expression, e.g., attributable to a DBP gene polymorphism (15).
This work was supported by the National Osteoporosis Society and Newcastle Hospital Special Trustees (H.K.D. and R.M.F.) and an EC grant for FP6 project ‘Osteogene’ (H.K.D.).
a,b P values (Mann-Whitney test):
1 P <0.005;
2 P <0.0001.
1 Comparison of the 2 groups revealed no significant differences.
↵1 Nonstandard abbreviations: 1,25(OH)2D3, 1,25-dihydroxyvitamin D3;25OHD3, 25-hydroxyvitamin D3; DBP, vitamin D–binding protein; BMD, bone mineral density; SHBG, sex hormone–binding globulin; and BMI, body mass index.
- © 2006 The American Association for Clinical Chemistry