View Report Details

The Interaction of Vitamins K & D in Bone & Vascular Health: New Functional Markers of Vitamin K Status

Project Code: N05036

31/05/2004

The Heamophilia Reference Centre, St Thomas' Hospital
Shearer, M

Background

The broad strategy behind this project was to make new, timely and cost-effective assays on samples collected during a previous project sponsored by The Food Standards Agency (N05001). Project N05001 had been a 2-year intervention study designed to find out whether taking regular, extra dietary amounts of vitamin K, either alone, or in combination with vitamin D plus calcium, could reduce the rate of bone mineral loss in healthy, elderly women. The study population comprised healthy elderly women aged over 60 years who had been randomised into four groups to receive (1) placebo (2) 200 μg vitamin K (phylloquinone, K₁) (3) 10 μg vitamin D plus 1g calcium (4) combined vitamins K and D plus calcium. The rationale was previous work that had suggested that vitamin K might play an important role in bone health and that bone reserves are commonly insufficient to fully support the post-translational γ-carboxylation of bone matrix proteins such as osteocalcin. Osteocalcin (OC) is a well-established marker of bone formation, and its three γ-carboxyglutamic acid (Gla) residues are essential for binding to hydroxyapatite, both in vivo and in vitro. Levels of under γ-carboxylated osteocalcin (GluOC) are a measure of vitamin K status. An integral part of the study was to study the putative interaction of vitamins K and D on bone health. An additional stimulus was previous evidence that high levels of GluOC were independently predictive of a higher risk of hip fracture risk and a lower BMD.

Objectives and Methods

The objectives of this project were to clarify and extend our knowledge of biochemical markers that reflected vitamin K status and/or bone metabolism. The new measurements were carried out on the serum and urine samples that had collected at 6-monthly intervals from the 209 subjects who had completed the intervention trial. In the first phase, we collaborated with Dr Caren Gundberg from Yale University to make new analyses of total osteocalcin using a well-validated assay that had been used in previous USA population studies. In addition, we have used well-validated radioimmunoassay (RIA) and hydroxyapatite (HAP)-binding methodologies to make new measurements of total OC, and GluOC (expressed as a percent of total OC, %GluOC). In the second phase we measured the urinary excretion of free Gla, an amino acid that is derived from the catabolism of all Gla proteins in the body and is a measure of overall vitamin K status. In the third phase, in collaboration with Dr Cees Vermeer from Maastricht University, we measured matrix Gla protein (MGP), and assessed its relationship to other biochemical markers. Recent work has shown MGP functions as an inhibitor of calcification in tissues such as bone and arteries.

Results

Using the HAP-binding assay, the mean concentration of %GluOC of 24% was comparable to that found in women of the same age range in the USA Framingham Offspring Study. This value was about half that originally determined in N05001 from separate monoclonal antibody (MAb)-based assays of GluOC and GlaOC fractions. Assay comparisons showed that values measured by the HAP-binding assay explained <40% of the variance of the MAb-based values. Further interpretation suggested that the much higher values of GluOC by MAb-based assay might reflect the overestimation of the large N-terminal-mid fragment of OC, which may be artefactual or circulate naturally. Although the absolute values of %GluOC were assay dependent, both methodologies showed similar responses to vitamin K supplementation and significant negative correlations with serum vitamin K. The mean fractional decreases from baseline in the vitamin K and vitamin K/D groups were 43 and 50% by the HAP-binding assay and 48 and 54% by the MAb assay respectively. Multiple regression analysis showed that the new assessment of total OC by RIA was a significant contributor to the variance of BMD at all bone sites. Circulating %GluOC, by either methodology, was not predictive of BMD variance at any site and neither total OC nor %GluOC made any significant contribution in models examining factors that influenced changes in BMD.

A new in-house assay using capillary electrophoresis was developed to measure urinary free Gla in mid-morning spot urine samples. The values (corrected for creatinine) were very widely distributed with no significant differences between the 4 groups at any time point. However, when the supplemented versus non-supplemented groups were combined, a trend was seen (borderline significance) to an increased median urinary Gla/creatinine ratio (~20%) after vitamin K supplementation. The validity of the results was supported by the finding of a significant negative correlation between the changes in urinary Gla over 24 months with the corresponding changes in %GluOC.

Measurements of total MGP with a newly developed assay are the first in any British population. The results were within the normal reference range established in the Netherlands but showed a significant increase over 2 years. No associations were found between MGP and bone markers, BMD or measures of vitamin K status.

Conclusions

This is the first comparative study of the effect of vitamin K supplementation on %GluOC values as assessed by either direct MAb assays of GluOC and GlaOC or by HAP-binding assay. The results confirm that, while there are considerable differences in the assay specificities, both are suitable for assessing vitamin K status and response to supplementation. Major caveats are that meaningful comparisons between different populations can only be made using the same assay, and by expressing the data as a percent of the total OC. With the HAP-binding assays it is also important to take account of the dependence of the binding curves on the basal level of total OC. The HAP-binding assay has the advantage over the MAb-based assay that a separate RIA measure of total OC, which is also a validated marker of bone formation, is an integral part of the assay. This was confirmed by the finding that total OC by RIA was a significant contributor to BMD variance, previously only found for the GlaOC fraction. The lack of association of the sum of GluOC + GlaOC with BMD shows that this is not equivalent to total OC by RIA and that the former measure lacks any credential as a marker of bone formation. We could not confirm observations by others that either GluOC or %GluOC is an independent determinant of BMD in elderly women, which might be expected given the evidence that high levels of GluOC are associated with increased fracture risk. Possible explanations for the lack of such a relationship include the greater level of fitness of our cohort or that high levels of GluOC reflect deleterious changes in bone matrix biochemistry that are independent of BMD changes. One possibility, also considered for collagen, is that changes in post-translational modifications might represent an early marker of reduced bone quality, rather than BMD.

As far as we are aware this is the first time that an inverse relationship between changes in %GluOC and urinary Gla has been recorded for any supplementation study and is notable given the large intra-and inter-variability of urinary Gla in spot urines. It shows that vitamin K supplementation was reflected in increased γ-carboxylation at the level of protein synthesis (%GluOC) and protein catabolism (urinary Gla).

BMD measurements in N05001 had suggested a possible synergistic effect of vitamins D and K in preventing bone loss, a finding supported by other recent intervention studies. Although the study design was well suited to the purpose, we found no evidence that vitamin D can significantly improve the γ-carboxylation status of OC, either alone or in concert with vitamin K. Thus, we were unable to confirm that vitamin D might independently improve the γ-carboxylation of OC as suggested by a previous study in institutionalised French elderly women. Again, this might reflect a difference in population characteristics. Further work is necessary to find out whether the putative synergism between vitamins D and K operates by other mechanisms.

Some of the files on this site may be in a format that your computer can't read. However, you can download Readers and Viewers for the following document types below: