Clinical Usefulness of Bioavailable Vitamin D and Impact of GC Genotyping on the Determination of Bioavailable Vitamin D in a Korean Population

Background Bioavailable 25-hydroxy vitamin D (25(OH)D) has been suggested for the accurate determination of vitamin D status. The purpose of this study was to determine the utility of bioavailable 25(OH)D in assessing vitamin D status when vitamin D-binding protein (VDBP) was significantly altered by pregnancy and liver cirrhosis (LC). The role of genotyping of GC, a gene encoding VDBP, in the determination of bioavailable 25(OH)D concentration in a Korean population was also evaluated. Methods This prospective study enrolled a total of 136 subjects (53 healthy controls, 45 patients with LC, and 38 pregnant women) from 2017 to 2018. The concentrations of total 25(OH)D and VDBP were measured, and bioavailable 25(OH)D concentrations were calculated. GC genotyping was performed to determine rs4588 and rs7041 polymorphisms. Clinical and laboratory data were compared among the three groups of subjects. Results Median VDBP and total 25(OH)D concentrations were 165.2 μg/ml and 18.5 ng/ml in healthy controls, 76.9 μg/ml and 10.5 ng/ml in patients with LC, and 368.9 μg/ml and 17.7 ng/ml in pregnant women, respectively. Compared with controls, patients diagnosed with LC had significantly lower VDBP and total 25(OH)D concentrations (all P < 0.001) while pregnant women had significantly higher VDBP concentrations (P < 0.001). Although total 25(OH)D concentrations in pregnant women were similar to those in controls (P = 0.394), their bioavailable 25(OH)D concentrations were significantly lower (1.2 vs. 3.0 ng/ml; P < 0.001). Among all the three groups combined, the genotype-specific bioavailable 25(OH)D and the genotype-independent bioavailable 25(OH)D concentrations did not differ significantly (P = 0.299). Conclusions Our study has demonstrated that bioavailable 25(OH)D concentration reflects vitamin D status more accurately than the total 25(OH)D concentration, especially in pregnant women. In addition, GC genotyping did not significantly affect bioavailable 25(OH)D concentration. Therefore, if VDBP concentration is significantly altered, the measurement of bioavailable 25(OH)D concentration might facilitate the accurate determination of vitamin D status. However, GC genotyping might be unnecessary.


Introduction
Humans synthesize vitamin D in skin following exposure to sunlight. Vitamin D can also be obtained from diet. Vitamin D does not exhibit biological activity until a two-step hydroxylation occurs. Following hydroxylation in the liver to 25-hydroxy vitamin D (25(OH)D), the vitamin D metabolite is transported to kidneys where it is converted to 1α, 25-dihydroxyvitamin D (1α, 25(OH) 2 D) [1,2]. Active forms of vitamin D exhibit varying function in tissues and organs throughout the body. Most (85%-90%) of the circulating vitamin D is tightly bound to vitamin D-binding protein (VDBP) and only a smaller amount (10%-15%) is loosely bound to albumin. Less than 1% of circulating vitamin D exists in free unbound form [3][4][5]. The fraction that is not bound to VDBP (free and albumin-bound) is considered bioavailable 25(OH)D [3].
Currently, vitamin D status is assessed via measurement of the total 25(OH)D concentrations. The commonly used criteria for vitamin D status are as follows: vitamin D deficiency (<20 ng/ml) and vitamin D insufficiency (20-30 ng/ml) [2,6,7]. However, a few studies have reported a stronger correlation between serum calcium, parathyroid hormone [3], bone mineral density [8], and vascular outcomes [9] with bioavailable 25(OH)D concentration compared with total 25(OH)D concentration, suggesting the clinical significance of bioavailable 25(OH)D concentration. However, other studies have failed to detect stronger associations between bioavailable 25(OH)D concentration and clinical outcomes [10]. Thus, the clinical utility of bioavailable 25(OH)D in the assessment of vitamin D status remains unclear.
The bioavailable 25(OH)D concentration is affected by serum VDBP concentration and GC genotype. The VDBP concentration is altered depending on various conditions. For example, VDBP is increased by up to 50% under elevated estrogen levels (e.g., pregnancy), whereas it is decreased in certain disease states (e.g., severe hepatic disease) [11][12][13][14]. The GC gene encoding VDBP exhibits more than 120 polymorphisms. Two single nucleotide polymorphisms (SNPs), rs7041 and rs4588, generate three major polymorphic isoforms of VDBP: Gc1f, Gc1s, and Gc2 [15,16]. Because the affinity of VDBP for 25(OH)D depends on the polymorphic isoform, the GC genotype may also play a significant role in determining bioavailable 25(OH)D [13,14,17]. However, GC genotyping entails DNA extraction and PCR that are expensive, labor-intensive, and time-consuming procedures. To date, no published studies have evaluated the necessity or the impact of GC genotyping on bioavailable 25(OH)D in a Korean population.
Thus, the objectives of this study were (1) to determine whether calculation of bioavailable vitamin D might facilitate the assessment of vitamin D status in individuals with altered VDBP concentrations such as pregnant women and patients with liver cirrhosis (LC) and (2) [18] were collected from electronic medical records. At the time of study enrollment, blood samples were collected and serum and leukocytes were separated and stored at -80°C. The study protocol was approved by the Institutional Review Board (IRB) of Gyeongsang National University Hospital (approval number: 2017-01-005). Written informed consent was obtained from all participants.
2.5. Statistical Analysis. We compared clinical and laboratory data and GC genotypes between groups using the chi-square test for categorical variables and the Kruskal-Wallis test for continuous variables, with the Tukey test based on ranks as post hoc analysis. Patient characteristics (Child-Pugh class or trimester) were compared using the Mann-Whitney test or two-sample t-test. GC genotype and allele frequencies were compared between groups using the chi-square test or Fisher's exact test. Genotype-independent and genotypespecific bioavailable 25(OH)D concentrations were compared using the Wilcoxon signed-rank test. Percent difference between genotype-independent and genotype-specific bioavailable 25(OH)D was calculated as [genotype-specific bioavailable 25(OH)D − genotype-independent bioavailable 25(OH)D]/genotype-independent bioavailable 25(OH)D × 100. The relationship between continuous variables was evaluated by Spearman's correlation. Values were expressed as median and interquartile range (IQR). P < 0 05 was considered significant. All statistical analyses were performed using IBM SPSS Statistics version 21 (IBM Corp., Armonk, NY, USA).

Results
3.1. Subject Characteristics. Median age of subjects was 50 years (IQR, 40-60 years) for healthy controls, 52 years (IQR, 46-62 years) for patients with LC, and 33 years (IQR, 32-36 years) for pregnant women. The proportion of women was 35.8% in the control group and 17.8% in the LC group. Albumin concentration was significantly lower in patients with LC (2.9 g/dl) than that in healthy controls (4.5 g/dl) and pregnant women (3.6 g/dl; both P < 0 001). Of the 45 patients with LC, 4 patients were Child-Pugh class A, 26 patients were class B, and 15 patients were class C. Of the 38 pregnant women, 2 were in the first trimester, 4 were in the second trimester, and 32 were in the third trimester (Table 1).

Discussion
In our study, the highest levels of serum VDBP concentrations were recorded in pregnant women and the lowest values were found in patients with LC as expected. The total 25(OH)D and genotype-independent bioavailable 25(OH)D concentrations were lower in patients with LC compared with those in healthy controls (P < 0 001 for both). This finding is consistent with previous studies reporting a high prevalence of total 25(OH)D deficiency in patients with chronic liver disease [14,[22][23][24][25]. The result could be attributed to hepatic dysfunction (e.g., liver cirrhosis) and lifestyle of chronic patients. In other words, impaired 25-hydroxylation of vitamin D and hepatic VDBP synthesis,    [14,25] have found no significant difference in total or free 25(OH)D concentrations between pregnant women and healthy controls. However, Kim et al. [29] have reported that bioavailable 25(OH)D concentrations were significantly lower in pregnant women (1.9 vs. 2.6 ng/ml; P = 0 003), although the total 25(OH)D concentrations were comparable between pregnant women and healthy controls (18.3 vs. 18.3 ng/ml; P = 0 808), similar to our results. Most previous studies assessing bioavailable vitamin D in pregnant women have enrolled Caucasian and African-American participants. The vitamin D status of ethnic minorities has been less studied [30], and few studies have enrolled Korean participants. Thus, differences in ethnicity may account for these discrepant results, which may be attributed to the effect of genetic polymorphisms in vitamin D metabolism and transport [30]. Our results imply that bioavailable 25(OH)D might reflect vitamin D status more accurately than the total 25(OH)D, especially in pregnant women. Thus, measurement of the bioavailable 25(OH)D levels may be essential to accurately assess vitamin D status, at least in pregnant Korean women. Ethnicity and pregnancy-specific vitamin D status need to be further evaluated.
The use of GC genotype-specific VDBP binding affinity values to calculate the bioavailable 25(OH)D was expected to reveal the differences between genotype-independent and genotype-specific values. Arnaud and Constans [15] have reported different VDBP affinities for total 25(OH)D according to the GC genotype. The affinity of Gc1f was fourfold higher than that of Gc2 and twofold higher than that of Gc1s. However, in our study, the genotype-independent and genotype-specific bioavailable 25(OH)D concentrations did not differ significantly. In patients with LC, genotype-independent bioavailable 25(OH)D concentrations appeared to be lower than genotype-specific 25(OH)D concentrations (2.4 vs. 3.0 ng/ml), although the difference was not significant (P = 0 077). Our results indicate that the GC genotype has a limited effect on the calculated value of bioavailable 25(OH)D. In addition, we found that the correlation between total and genotype-independent bioavailable 25(OH)D concentrations was higher than the correlation between total and genotype-specific bioavailable 25(OH)D concentrations.
The minimal impact of GC genotype on the calculated bioavailable 25(OH)D concentration might be explained by the distribution of GC alleles in subjects enrolled in this study. To calculate genotype-independent bioavailable 25(OH)D, we used the binding affinity constant of VDBP (K VDBP ), which represents the average affinity constant of three isoforms. In our study, the frequency of GC alleles in all subjects was 41.9% for Gc1f, 35.3% for Gc2, and 22.8% for Gc1s, which showed relatively even distribution, which may explain the lack of significant difference between genotype-independent bioavailable 25(OH)D calculated using the average K VDBP and genotype-specific bioavailable 25(OH)D calculated using genotype-specific K VDBP . In other words, the even distribution of GC allele in the present study might arithmetically compensate the effect of genotype-specific K VDBP on the calculation of bioavailable 25(OH)D. In fact, a previous study [31] of 360 Korean subjects has reported that the allele frequencies of Gc1f, Gc1s, and Gc2 were 44%, 25%, and 31%, respectively, which were similar to our results. Thus, although our study included a relatively small number of subjects, our results suggesting a limited effect of the GC genotype on bioavailable 25(OH)D value may still be meaningful. However, further studies enrolling larger numbers of Korean subjects are needed to corroborate these results.
In our study, patients with severe hepatic dysfunction (Child-Pugh class C) had lower VDBP concentrations than patients with Child-Pugh class A and B (45.3 vs. 91.6 μg/ml; P < 0 001). The total 25(OH)D appeared to be lower in patients with Child-Pugh class C, but not significant (8.6 vs. 11.7 ng/ml; P = 0 057). Bioavailable 25(OH)D concentrations did not differ between groups (all P > 0 05), indicating similar total and bioavailable 25(OH)D concentrations despite differences in hepatic dysfunction. Arteh et al. [23] have shown that severe total 25(OH)D deficiency is more common in patients diagnosed with LC than in patients with     noncirrhotic liver disease. Lai et al. [24] have reported that in patients with LC, those with low albumin concentrations also carry lower concentrations of VDBP, total 25(OH)D, and free 25(OH)D than patients with normal albumin concentrations. The analysis of VDBP, total 25(OH)D, and bioavailable 25(OH)D concentrations in pregnant women showed no significant differences according to the trimester. However, only two women in the first trimester were enrolled in our study. Therefore, we only compared the two groups (first and second trimesters vs. third trimester). Ma et al. demonstrated that VDBP, 1α-hydroxylase, 24-hydroxylase, and vitamin D receptor were expressed on the placenta [32], and Cleal et al. showed that maternal 25(OH)D and VDBP concentrations may mediate the regulation of amino acid transfer to the fetus [33], suggesting that dysregulation of VDBP as well as vitamin D could be a risk factor in the pregnancy outcome (i.e., preeclampsia, preterm birth, and gestational diabetes) [34]. Further studies are needed to analyze the VDBP, total 25(OH)D, and bioavailable 25(OH)D concentrations in pregnant women according to the trimester and pregnancy outcome.
Our study has a few limitations. First, we measured VDBP concentrations using monoclonal ELISA. Nielson et al. [35] have reported that VDBP concentrations measured with monoclonal ELISA are strongly correlated with GC genotypes. Therefore, this assay underestimated VDBP concentrations in African-Americans because of the high frequency of Gc1f. A study evaluating VDBP concentrations in Korean subjects by polyclonal ELISA may be necessary. Second, we have not used liquid chromatography-tandem mass spectrometry (LC-MS/MS) to measure the total 25(OH)D concentrations. However, in healthy controls, the measured total 25(OH)D concentration (18.5 ng/ml) was similar to that reported by a previous study (18.3 ng/ml) in Korea using LC-MS/MS [29]. Third, the small sample size limited our ability to determine the effect of GC genotype on VDBP concentrations and compare subgroups according to the Child-Pugh class in patients diagnosed with LC and trimester in pregnant women.
Fourth, we have not investigated vitamin D supplementation or sun exposure. We have not considered seasonal variation either. Fifth, we failed to exclude subjects with other conditions (e.g., malnutrition, infection, and nephrotic syndrome) that could affect VDBP concentrations other than LC and pregnancy.

Conclusions
In this study, we demonstrated that bioavailable 25(OH)D levels reflect vitamin D status more accurately than the total 25(OH)D concentrations, especially in pregnant women. In addition, we have shown that the GC genotype did not significantly affect bioavailable 25(OH)D concentration in Koreans. Therefore, if the VDBP concentration is significantly altered, calculation of bioavailable 25(OH)D levels might facilitate the accurate determination of vitamin D status obviating the need for GC genotyping.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest relevant to this study.