|Home | About | Journals | Submit | Contact Us | Français|
Vitamin K modulates cytokines involved in bone turnover, including interleukin-6 (IL-6) and osteoprotegerin in vitro.
The objective of this study was to assess 1) associations between measures of vitamin K status [plasma phylloquinone and serum percentage of undercarboxylated osteocalcin (%ucOC)] and IL-6, osteoprotegerin, and C-reactive protein (CRP) concentrations and 2) the effect of daily 500 μg phylloquinone supplementation for 3 y on cytokine concentrations.
Concentrations of IL-6, osteoprotegerin, and CRP and bone mineral density (BMD) were measured at baseline and after 3 y of follow-up in 379 healthy men and women (60–81 y; 58.5% women) participating in a randomized trial that studied the effect of vitamin K supplementation on bone loss.
Cross-sectionally, plasma phylloquinone was inversely associated with IL-6 and CRP, whereas serum %ucOC was inversely associated with IL-6. Osteoprotegerin was associated positively with plasma phylloquinone and inversely with %ucOC. No differences were observed in the 3-y change in IL-6, osteoprotegerin, and CRP concentrations between participants who received phylloquinone supplementation and those who did not. Overall, no association was observed between the 3-y changes in circulating cytokines and BMD.
Poor vitamin K status was associated with high concentrations of cytokines involved in bone turnover, but vitamin K supplementation did not confer a decrease in cytokine concentrations. The healthy status of this cohort may explain a lack of effect of vitamin K supplementation on cytokine concentrations. This trial was registered with www.clinicaltrials.gov as NCT00183001.
A protective role for vitamin K against age-related bone loss has been proposed that is attributed to its function as a cofactor for the posttranslational γ-carboxylation of certain proteins, including osteocalcin (1). It is assumed that the γ-carboxylation reaction confers function of these proteins. When vitamin K status is inadequate, the percentage of osteocalcin that is not carboxylated (%ucOC) increases (2). The %ucOC, therefore, is considered to be a sensitive indicator of vitamin K status in bone (3, 4). Additional roles of vitamin K, independent of its function as a cofactor in the γ-carboxylation reaction, were proposed (5, 6). It is plausible that a potential role for vitamin K in skeletal health exists through the modulation of cytokines, independent of its role as a cofactor.
Recent evidence suggests inflammation may contribute to the disorder of osteoporosis (7-9). Among older women (age: 70–79 y), higher concentrations of inflammatory markers were associated with an increased risk of incident fracture (10). In vitro data suggest that vitamin K treatment modulates production of certain cytokines, including interleukin-6 (IL-6) and osteoprotegerin (11-13). The production of osteoprotegerin, a cytokine member of the tumor necrosis factor receptor family that is thought to protect against bone loss, is regulated in part by IL-6, a proinflammatory cytokine (9, 14). IL-6 is also a potent stimulator of the hepatic production of C-reactive protein (CRP), a marker of systemic inflammation (15). Cross-sectional associations between vitamin K status and several proinflammatory cytokine concentrations were observed in a community-based sample of men and women (the Framingham Offspring Cohort) (16). More specifically, plasma phylloquinone was inversely associated with overall inflammation, as well as with individual proinflammatory biomarkers, including IL-6, intercellular adhesion molecule-1, and tumor necrosis factor receptor 2. In the same cohort, plasma phylloquinone was also inversely associated with serum osteoprotegerin. Although in vitro and cross-sectional data support an association between vitamin K and certain cytokines, currently no intervention data are available that examine the possible effect of vitamin K supplementation on concentrations of those cytokines that are associated with bone turnover.
The purpose of this study was to test the hypotheses that the circulating concentrations of cytokines (IL-6, osteoprotegerin, and CRP) in older men and women are inversely associated with vitamin K status. Furthermore, circulating cytokine concentrations will decrease after 3 y of supplementation with a multivitamin that included phylloquinone (vitamin K1, the principal dietary form of vitamin K) compared with a multivitamin without phylloquinone. In addition, we examined the relation between the 3-y change in circulating cytokine concentrations and the 3-y change in bone mineral density (BMD) in the same men and women.
Free-living men and postmenopausal women (n = 452; ± SD: 68 ± 6 y of age) enrolled in a 3-y, double-blind, randomized controlled trial designed to assess the effect of vitamin K supplementation on age-related bone loss (17). Exclusion criteria included history of osteoporosis or known coronary disease or use of osteoporosis medication, anticoagulant medication, or, for women, hormone replacement therapy. Study participants were randomly assigned to either treatment (n = 229) or nontreatment (n = 223). A total of 401 (164 men and 237 women) completed the 3-y intervention, 204 in the treatment group and 197 in the nontreatment group. The treatment group received 500 μg of phylloquinone, a dose that was deemed nutritionally optimal, safe, and attainable within the diet, as part of a daily effervescent multivitamin formulation (1 tablet). The nontreatment group received the multivitamin formulation without phylloquinone (1 tablet). All study participants also received a second daily effervescent tablet that provided 600 mg of elemental calcium and 10 μg (400 IU) of vitamin D as cholecalciferol. The composition of these supplements was described elsewhere (17). Adherence of supplement use, defined as the percentage of tablets consumed over the 3 y, was 89.1% for the group supplemented with the multivitamin with phylloquinone and 88.5% for the group who received the multivitamin without phylloquinone. Measures of circulating cytokine concentrations were available for 379 participants (189 in the treatment group and 190 in the nontreatment group) (Figure 1). All participants signed a written informed consent, and this study was approved by the Institutional Review Board at Tufts University-New England Medical Center.
All blood samples were drawn between 0700 and 1000 after a 12-h fast. Dedicated aliquots of EDTA plasma were stored −80 °C and protected from light until the time of analysis. For all cytokine measures both baseline and year 3 samples for each study participant were run together, to minimize interassay variation between visits.
Plasma IL-6 was assayed in duplicate by enzyme-linked immunoassay with the use of high-sensitivity IL-6 kits commercially available from R&D Systems (Minneapolis, MN) (total CV: 8.7%). Plasma CRP was assayed in singlet, with the use of high-sensitivity Immulite CRP kits commercially available from Diagnostic Products Corporation (Los Angeles, CA) with the use of the COBAS MIRA (Roche Instruments, Belleville, NJ) (total CV: 3.7%). Plasma osteoprotegerin was assayed in duplicate by enzyme-linked immunoassay with the use of high-sensitivity osteoprotegerin kits commercially available from Immunodiagnostics (Biomedica, Vienna, Austria) (total CV: 13.7%). Although the assessment of receptor activator of nuclear factor-κB ligand, the receptor for osteoprotegerin, would have been valuable to better understand the potential influence of vitamin K on cytokines involved in bone turnover, because of limited resources we were only able to measure osteoprotegerin, a cytokine for which in vitro data suggest modulation by vitamin K. The 3-y change in cytokine concentrations were calculated by subtracting the baseline from the year 3 measure.
Plasma concentrations of phylloquinone (done as singlet determinations) were determined by reversed-phase HPLC with the use of postcolumn, solid-phase chemical reduction of phylloquinone to hydroquinone, followed by fluorometric detection (18). Serum total osteocalcin and ucOC were measured by radioimmunoassay, with the use of the method of Gundberg (19). The antibody recognizes both carboxylated osteocalcin and ucOC. Carboxylated osteocalcin was separated from ucOC by adsorption on hydroxyapatite. Total osteocalcin was determined in the serum before adsorption and ucOC was measured in the adsorbed serum. The total CVs for the 3 control serums with an average total osteocalcin result of 6.4, 14.7, and 23.8 μg/L were 8.8%, 8.9%, and 7.6%, respectively.
BMD of the spine, femoral neck, and whole body was measured at baseline and after 3 y of follow-up by dual-energy X-ray absorptiometry (model GE Lunar Prodigy; GE, Madison, WI). Software version 5.0 was used for acquisition and analysis. Three-year change in BMD was calculated by subtracting baseline from the year 3 measure.
Information was collected every 6 mo about changes in medical history, medication use, and smoking status. Dietary intakes, including phylloquinone, were assessed yearly with the use of the Willett Food-Frequency Questionnaire (FFQ). This FFQ queries frequency of intake of reference portions of 126 foods during the previous year (20). Daily intakes of individual foods and specific nutrients were calculated by Harvard University. This FFQ has been validated for the assessment of vitamin K intake (21). Participation in regular physical activity was also assessed with the use of a previously validated survey (22). Measurements of weight and height were taken, and body mass index (BMI; in kg/m2) was calculated.
Because the variability in response was seen to increase with increasing concentration, a (natural) logarithmic transformation was applied to IL-6 and CRP before formal analysis. Independent samples t test (for continuous data) or chi-square test for differences in proportions (for categorical data) were used to examine differences between treatment and nontreatment groups at baseline.
Separate linear models were used to determine cross-sectional associations between baseline measures of vitamin K status and circulating cytokine concentrations, with the cytokine concentration as the outcome and measure of vitamin K status as the main exposure. No significant interactions were observed between sex and vitamin K status for any of the cytokines measured. For each association, 3 multivariable models were constructed. 1) To identify general associations between vitamin K and inflammation, we initially adjusted for sex and age only. 2) Because statin and antiinflammatory medications influence measures of inflammation, we subsequently adjusted for age, sex, and use of antiinflammatory and statin medications. 3) BMI, physical activity, and smoking may also influence inflammation in healthy persons, so we included these covariates as well in fully adjusted models. Models that tested the association between cytokines and plasma phylloquinone were also adjusted for triacylglycerols. We excluded participants with CRP ≥ 10 mg/L (n = 45) from CRP analyses, because concentrations ≥10 mg/L are indicative of acute infection (23).
Primary analyses were limited to those participants for whom measures of inflammation and BMD were available, regardless of adherence to treatment. Selected secondary analyses were restricted to the study participants who completed the study and took the supplements throughout the study period. To determine the effect of vitamin K supplementation on the change in circulating cytokines we used multiple linear regression with the year 3 cytokine concentration measures as outcome variables and treatment as the main exposure, controlling for baseline cytokine concentrations, as well as age, sex, BMI, triacylglycerols, antiinflammatory and statin medication use, physical activity, and smoking. To verify that there was no differential response to vitamin K supplementation because of differences in inflammation before randomization, we also checked for interactions between baseline cytokine concentrations and treatment by entering a product term (baseline cytokine concentration × treatment) for each marker (IL-6, osteoprotegerin, CRP). Significant interactions were observed between baseline osteoprotegerin and treatment and between baseline IL-6 and treatment as a result of 2 extreme values. These measurements were set aside so that the fitted models would describe the bulk of the data (24). Consequently, no significant interactions were observed between baseline cytokine concentrations and treatment.
To determine the association between the 3-y change in BMD and the change in cytokine concentration, we examined partial correlations between changes in circulating cytokines and BMD, adjusted for treatment, age, sex, baseline concentrations of inflammation markers, baseline measures of BMD, BMI, antiinflammatory and statin medication use, physical activity, and smoking status. All analyses were performed with the use of SAS version 9.1 (SAS Institute, Cary, NC) and were considered to be statistically significant at P < 0.05 (2-sided).
Cytokine concentrations, BMD, and covariate data for study participants in the treatment and nontreatment groups at baseline and year 3 are presented in Table 1. At baseline, the mean (±SD) age of this sample (58.5% women) was 68 ± 6 y (range: 60–81 y), and 70% were characterized as overweight or obese according to BMI. Overall, use of statin medication significantly increased to 36%, and antiinflammatory medication use (steroidal or non-steroidal) significantly decreased to 13.4% by year 3 (both P < 0.01). Plasma phylloquinone increased [P < 0.01, repeated-measures analysis of variance (ANOVA)] and %ucOC decreased (P < 0.01, repeated measures ANOVA) after 3 y in the group that received phylloquinone supplementation, but they did not change in the group that received the multivitamin without phylloquinone. Reported intakes of phylloquinone were higher than current recommendations at baseline in both groups (183 ± 108 μg/d in the vitamin K treatment group and 172 ± 111 μg/d in the non–vitamin K treatment group), and dietary intakes remained unchanged in both groups during the first 12 mo of the study (data not shown).
At baseline, plasma phylloquinone was significantly and inversely associated with CRP and IL-6 in age- and sex-adjusted analyses (P = 0.04 and 0.02, respectively, 2-sided) (Table 2). These associations remained significant when further adjusted for antiinflammatory and statin medication use (P < 0.05), but they were attenuated by inclusion of BMI, physical activity, and smoking as covariates. We chose to include all 3 statistical models to show the influence of the lifestyle factors, including BMI, on these associations. In contrast, plasma phylloquinone was positively associated with circulating osteoprotegerin in all models (P = 0.02). The %ucOC was inversely associated with IL-6 and osteoprotegerin in all models (P < 0.01) (Table 2). Total osteocalcin, a measure of bone formation, the synthesis of which is not influenced by vitamin K, was inversely associated with CRP and IL-6, such that an increase of 10 ng/mL in circulating osteocalcin was associated with a 0.73-fold decrease in CRP and a 0.84-fold decrease in IL-6 in fully adjusted models (all P < 0.01). Serum osteoprotegerin was not associated with total osteocalcin in any model (all P > 0.05). The mean concentrations of CRP and IL-6 at the end of the 3-y study did not significantly differ from baseline concentrations in either group or when data from the 2 groups were combined (Table 1). A statistically significant increase overall was observed in the mean concentration of osteoprotegerin in both groups (P < 0.01, repeated-measures ANOVA). However, the mean difference in change in osteoprotegerin between the treatment groups was not statistically significant (Figure 2). When analyses were limited to include participants who were taking the multivitamin supplement for 3 y, results were unchanged.
Because there was no effect of vitamin K supplementation on circulating cytokine concentrations or on age-related bone loss (17) and no significant interactions between change in concentration of any of the circulating cytokines and treatment with respect to change in BMD at any site, the treatment and nontreatment groups were combined to examine associations between change in BMD and change in cytokines (Table 3). Increases in BMD at the lumbar spine ( ± SD: 0.02 ± 0.05 g/cm2) and decreases in total body BMD (–0.01 ± 0.03 g/cm2) were observed between baseline and year 3 in both groups (P < 0.05). No consistent associations were observed between change in circulating cytokine concentrations and change in BMD at the end of the 3-y follow-up, with the exception of a significant positive association between change in CRP and change in total body BMD (P = 0.01)
Low plasma phylloquinone was associated with high circulating concentrations of proinflammatory markers (CRP and IL-6) and low concentrations of osteoprotegerin in older men and women. Our cross-sectional findings are generally consistent with in vitro (11-13) and observational (16) data, and collectively they suggest a potentially protective role for vitamin K in inflammation.
Osteoprotegerin is an antiresorptive cytokine that is believed to protect against bone resorption. High %ucOC was associated with low osteoprotegerin, which is consistent with high plasma phylloquinone being associated with low osteoprotegerin in the same cohort. In vitro, vitamin K treatment increased osteoprotegerin production in bone marrow cells (25), and, in one small study of patients being treated with glucocorticoid medication for kidney disease, vitamin K treatment prevented a glucocorticoid-induced reduction in serum osteoprotegerin (26). In contrast, an inverse association between plasma phylloquinone and serum osteoprotegerin was previously reported in a community-based cohort (16). Circulating osteoprotegerin increases with age (27, 28), and our cohort was older (mean age: 68 y) and within a narrower age range (60–81 y) compared with the community-based cohort, in which the mean age was 59 y (range: 35-89 y) (16). This age-related increase would also explain the significant 3-y increases in circulating osteoprotegerin we observed in both the treatment and nontreatment groups.
When vitamin K status is low, osteocalcin is not fully carboxylated, so a high %ucOC is indicative of poor vitamin K status. Therefore, the association between high %ucOC (ie, poor vitamin K status) and low circulating IL-6 does not support our overall hypothesis that improved vitamin K status is associated with less inflammation. Circulating IL-6 and CRP were significantly inversely associated with total osteocalcin in our cohort, as was reported elsewhere (29). Measurement of total osteocalcin captures both carboxylated and uncarboxylated forms of osteocalcin, so it is not a measure of vitamin K status. Instead, it is a measure of bone formation, independent of the posttranslational role of vitamin K as an enzyme cofactor. Therefore, we cannot dismiss the possibility that the total osteocalcin was responsible for an inverse association between %ucOC and IL-6. In contrast, total osteocalcin was not significantly associated with circulating osteoprotegerin, whereas low osteoprotegerin concentrations were consistently associated with low phylloquinone and high %ucOC in our study.
Phylloquinone supplementation did not change cytokine concentrations in this cohort of older men and women. The null findings of our intervention trial, which are not in agreement with the cross-sectional associations, are not readily explained by the in vitro studies (12, 25). Changes in circulating cytokine concentrations were small in our cohort in comparison to other nutrient intervention studies that have reported improvements in inflammatory cytokines (30) and serum osteoprotegerin (26) among persons with heart and kidney diseases respectively. This attenuated response may reflect a selection bias because participants in our study were generally free of chronic disease because of the exclusion criteria used for acceptability into the study. This is consistent with the results of other supplementation studies in which the participants were generally healthy and inflammatory cytokines remained unchanged (31, 32). That neither group in our study showed a detrimental change in circulating cytokine concentrations after 3 y may also contribute to the lack of measurable effect of phylloquinone supplementation on these cytokines. In contrast, this does not explain why we observed a consistent, cross-sectional association between vitamin K status and concentrations of circulating cytokines. One possible explanation is that biochemical measures indicative of a poor vitamin K status are consistently associated with a less healthy diet and lifestyle compared with those of persons with high serum concentrations of phylloquinone and low %ucOC (33, 34). The lack of cytokine response to supplementation of 500 μg phylloquinone/d suggests that in the cross-sectional data the biochemical measures of vitamin K are surrogate measures of other dietary or lifestyle factors that may influence circulating cytokine measures in this cohort. In addition, although a supplemental dosage of 500 μg/d exceeds the current adequate intake of 90–120 μg/d (35), those studies that reported a beneficial effect of phylloquinone supplementation on health outcomes supplemented with a dosage of 1000 μg/d (36, 37). The dosage chosen for the current study may have been insufficient to confer any beneficial effect on cytokine concentrations. Furthermore, dietary intakes of phylloquinone were above the current recommended adequate intakes at baseline and did not change significantly throughout the study. It is plausible that the high intakes attenuated any beneficial effect of vitamin K on our measures of inflammation.
Although a role for inflammation in bone resorption was suggested (7, 8), we did not observe associations between plasma IL-6 or CRP and BMD either cross-sectionally at baseline, before randomization (data not shown), or during the 3 y of follow-up. It was suggested that the predictive effect of serum IL-6 on bone loss in women is attenuated after the first decade of menopause (38). Because the women in our study were, on average, >10 y postmenopausal, the possible association between circulating IL-6 and BMD may have been attenuated.
Associations between serum osteoprotegerin concentrations and BMD are inconsistent in men (39, 40) and women (40, 41), which may be related to the potential modulation of osteoprotegerin by sex hormones (40). Because the clinical significance of the measure of circulating osteoprotegerin remains to be clarified (42), the meaning of the association between vitamin K status and circulating osteoprotegerin with respect to bone health is also uncertain.
All participants received daily supplementation with 600 mg elemental calcium and 400 IU vitamin D, which has been shown to reduce bone loss in older men and women over 3 y (43) The attenuation in bone loss because of the calcium and vitamin D supplementation, in addition to the nonstatistically significant changes in circulating IL-6 and CRP from baseline to year 3, may partially explain our findings of no association between circulating proinflammatory cytokine concentrations and the 3-y change in BMD. Furthermore, the influence of cytokines on bone is local, so circulating measures may not be sensitive enough to reflect the cytokine activity in the microenvironment of skeletal tissue (38, 44).
In summary, poor vitamin K status was associated with high concentrations of the cytokines, IL-6 and CRP, in a cross-sectional analysis of older men and women. The use of %ucOC as a measure of vitamin K status was limited by the potential confounding effect of cytokine concentrations on total osteocalcin. However, supplementation with phylloquinone (vitamin K1) at a dosage of 500 μg/d for 3 y did not confer a decrease in inflammatory cytokine concentrations in our cohort. Similarly, although cross-sectional results suggest a positive association between vitamin K status and plasma osteoprotegerin, phylloquinone supplementation did not change plasma osteoprotegerin in this cohort. Because markers of vitamin K status are reflective of an overall healthy diet, the cross-sectional observations may reflect a general influence of healthy dietary patterns on serum IL-6, osteoprotegerin, and CRP. Alternatively, in this generally healthy cohort of older men and women, the concentrations of IL-6 and CRP did not change during the 3-y study, so any putative effect of vitamin K supplementation would have been attenuated. The health status of this cohort may also partially explain the lack of association between circulating cytokines and BMD changes over 3 y. However, the role of vitamin K supplementation in modulating measures of circulating cytokines among persons at greater risk of inflammation, such as the elderly (45) or those with cardiovascular, metabolic, or rheumatic diseases (30, 46, 47), merits investigation.
None of the authors had a personal or financial conflict of interest.
2Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors, and do not necessarily reflect the view of the US Department of Agriculture.
3Supported by the US Department of Agriculture, Agricultural Research Service (cooperative agreement 58-1950-7-707), National Institutes of Health (AG14759, HL69272, and T32 HL69772-01A1), and American Heart Association (0515605T).
The author’s responsibilities were as follows—MKS: designed the study, performed the statistical analyses, and drafted the manuscript; SLB, GED, and BD-H: contributed to the design of the analyses, the interpretation of the data, and writing of the manuscript; CMG: contributed to the interpretation of the data and writing of the manuscript; JWP: contributed to the laboratory analyses and writing of the manuscript; CJO and JMO: contributed to design of the analyses and interpretation of the data. All authors reviewed the final manuscript.