Main

Patients with vitamin D deficiency and vitamin D-resistant states are associated with extremely short stature, which is usually cured by treatment with vitamin D except if there is severe resistance to 1,25-D3 (1,2). In the growth plate, chondrocytes undergo cellular changes including proliferation, formation of a proteoglycan matrix, terminal differentiation to hypertrophic cells, and calcification until cessation of linear bone growth after puberty (1–4). The role of 1,25-D3 in growth plate cartilage is to regulate the differentiation and proliferation of chondrocyte (3–7). To act on its target cells, 1,25-D3 binds to a specific VDR (2,8). Human VDR cDNA has two potential initiation sites (8). An A(T/C)G polymorphism exists in the first initiation codon in exon 2 of the VDR gene (9). Individuals with the ATG polymorphism have two translation start sites and may initiate VDR translation from the first ATG site and encode a 427-amino acid VDR protein, whereas those with the ACG, VDR translation would initiate at the second site and encode a 424-amino acid VDR protein (8,9). This polymorphism and consequent difference in the VDR mRNA and protein may potentially affect the final adult height achieved.

One study reported a marked association of the VDR gene intron 8 BsmI polymorphism with bone mineral density (10,11); however, the meta-analysis of subsequent reports revealed that the largest differences between BsmI genotypes and bone mineral density were seen in the mixed menopausal status groups and the effect of BsmI polymorphism on bone mineral density may be minor (12,13). Furthermore, the genetic influence on bone mineral density and also height is better assessed in the young population with the same social and ethnic background (13,14).

The purpose of this study was to determine whether the exon 2 polymorphism in the VDR gene is related to final height. To pursue this study, 90 young healthy Japanese women aged 18-20 y, 159 healthy subjects aged 13 y, and 24 children with constitutional short stature less than -1.5 SD score in height, mostly with parents of normal height, were enrolled in this study.

METHODS

Subjects

Ninety healthy women aged 18-20 y (height SD score, -2.2 ∼ 2.4; body mass index, 18.4-27.3), who were third-year high school or college students with regular menstruation and regular physical activity, were recruited through the school administrative offices, and 159 healthy subjects (72 male and 87 female) aged 13 y were studied. We also examined 24 children with constitutional short stature, mostly with parents of normal height, aged 6-10 y, less than -1.5 SD score in height (-3.0-1.5 SD) who have normal thyroid function, IGF-I level, no shift in the height growth curve, and without perinatal events and any diseases affecting growth. Information was obtained from detailed questionnaires in 90 healthy women aged 18-20 y.

Height was measured with a fixed stadiometer. Blood was collected for DNA isolation. All studies were approved by the Chiba University Hospital Ethics Committee, and informed consent was obtained.

DNA Extraction and PCR Primers

DNA was prepared from peripheral blood with the Easy-DNA extraction kit (Invitrogen, San Diego, CA). The strategy for PCR amplification of the VDR gene is shown in Figure 1. PCR primers were prepared according to the published human VDR gene sequence (8,9). The primers for exon 2 polymorphism analysis from genomic DNA were 2a, 5′-AGCTGGCCCTGGCACTGACTCTGCTCT-3′ and 2b, 5′-ATGGAAACACCTTGCTTCTTCTCCCTC-3′; and the amplified PCR product was 265 bp in length.

Figure 1
figure 1

PCR strategy for RFLP analysis of the vitamin D receptor gene. The VDR gene consists of nine exons and intervening introns. The locations of primers 2a and 2b for exon 2 polymorphism analysis from genomic DNA are shown. The amplified PCR products are 265 bp.

Full size image

Exon 2 Polymorphism

PCR-SSCP. Fifty nanograms of genomic DNA were amplified in 20 µL of a mixture containing 10 mmol each of dATP, dCTP, dTTP, and dGTP, 0.5 mmol of [32P]dATP (3000 dpm/µL), 20 mmol of Tris-HCl, pH 7.5, 100 mmol of KCl, 1 mmol of DTT, 0.1 mmol of EDTA, 0.5% Tween 20 (vol/vol), 0.5% Nonidet P-40, 50% glycerol (vol/vol), 10 pmol of primers 2a and 2b, and 1 U of Expand Taq™ DNA polymerase (Boehringer Mannheim, Indianapolis, IN). After initial denaturation for 2 min at 94°C, samples were subjected to 25 cycles of amplification, consisting of denaturation for 40 s at 95°C, annealing for 30 s at 60°C, and extension for 1 min at 72°C. The PCR fragment 265 bases in length was analyzed with SSCP. Two microliters of PCR products were denatured by heating with formamide and electrophoresed on a 6% polyacrylamide gel (acrylamide:N,N′-methylenbisacrylamide, 49:1) containing 5% glycerol, in 0.5× Tris borate-EDTA buffer with 950 V, for 8 h. After electrophoresis, the gels were dried and autoradiographed.

Sequencing to determine Exon 2 polymorphism. To examine the relationship between PCR-SSCP pattern and polymorphism, in several samples, the reaction products were purified with a QIA quick-gel extraction kit (Qiagen GmbH, Hilden, Germany). Nucleotide sequencing was performed by a Sequenase™ version 2 sequencing kit (U. S. Biochemical, Cleveland, OH) with the modification of adding 0.1% Nonidet P-40.

PCR-RFLP. The exon 2 polymorphism creates a restriction site of FokI endonuclease (FokI, GGATG), and the PCR products were digested with FokI (New England BioLabs, Beverly, MA).

Statistical Analysis

Data were evaluated by analysis of variance with Fisher's protected least significant difference post hoc test, two-tailed Fisher exact test, and linear regression analysis using StatView J 4.02 software (Abacus Concepts, Berkeley, CA) and presented as mean ± SD, unless otherwise noted.

RESULTS

Exon 2 Polymorphism

The PCR-direct sequencing and PCR-SSCP of PCR-amplified DNA are shown in Figure 2, A and B. The PCR-direct sequencing revealed that three SSCP patterns of lower single band, two bands, and upper single band correspond to genotype CC (ACG/ACG), CT (ACG/ATG), and TT (ATG/ATG), respectively (Fig. 2, A and B). The cleavage of the PCR products (265 bp) by FokI generates two fragments of 69 and 196 bp in size in the case of the ATG allele (Fig. 2C). All the DNA samples of female subjects aged 18-20 y were examined and confirmed by both PCR-SSCP and restriction enzyme analysis. The other DNA samples were examined by PCR-RFLP. The allele frequencies for the exon 2 polymorphism were ACG, 0.62, and ATG, 0.38, in 498 alleles examined from healthy subjects. The distribution of the VDR genotypes for the exon 2 was CC, 0.37; CT, 0.51; and TT, 0.12.

Figure 2
figure 2

PCR direct sequence (A) and SSCP (B) of the PCR-amplified products. The reverse strand was sequenced, and the sequence was flipped to show the sense strand sequence (A). PCR direct sequencing of the PCR-amplified products revealed that three SSCP patterns of the lower single band, the two bands, and the upper single band correspond to ACG homozygote designated as CC, ACG/ATG heterozygote CT, and ATG homozygote TT. They were also examined by PCR-RFLP using the restriction enzyme FokI (GGATG) (C). The results obtained by the PCR-RFLP confirmed those by the PCR-SSCP.

Full size image

Exon 2 Polymorphism and Height

Ninety healthy female subjects aged 18-20 y. The mean heights of subjects with the genotype CC, CT, and TT were 155.9 ± 4.4 cm (mean ± SD), 160.3 ± 4.3 cm, and 157.6 ± 3.3 cm, respectively (CC versus CT, p < 0.0001: CT versus TT, p = 0.0302) (Table 1, Fig. 3). Thus, heterozygotes are significantly taller than the two types of homozygotes (3.8 cm, which corresponds to 0.76 SD, p < 0.0001, the 95% confidence limit of the difference is 2.0-5.6 cm). But there was no relationship between height and other BsmI polymorphism of the vitamin D receptor gene (data not shown). There was neither significant difference in the age for onset of menarche between three genotypes (Table 1) nor correlation between height and menarche.

Table 1 Characteristics of 90 female subjects aged 18-20 y
Full size table
Figure 3
figure 3

Exon 2 polymorphism and adult height in 90 female subjects. The height of individuals with CT genotype was 160.3 ± 4.1 cm (mean ± SD), whereas the mean heights of those with CC and TT were 4.5 cm (p < 0.0001) and 2.8 cm (p = 0.0302) shorter than those with type 2, respectively. Note that the distribution of height is different between genotypes.

Full size image

One hundred fifty-nine healthy subjects aged 13 y of both sexes. To clarify from when the relationship between this polymorphism and height was found, we studied in younger population aged 13 y. The mean height SD score of subjects with the genotype CT was again higher than homozygotes (CC/CT, p = 0.0363; CT/TT, p = 0.0208; and CT/CC,TT, height SDS difference, 0.308, p = 0.0087) in 159 subjects (72 male and 87 female) (Table 2, Fig. 4). But the difference is relatively small.

Table 2 Characteristics of 159 subjects aged 13 y in both sexes
Full size table
Figure 4
figure 4

Exon 2 polymorphism and height SD score in 159 subjects aged 13 y. The height SD score of individuals with CT genotype was larger than homozygotes (CC/CT, p = 0.0363; and CT/TT, p = 0.0208).

Full size image

Twenty-four children with constitutional short stature. The frequency of the exon 2 polymorphism in 24 children with constitutional short stature was different from that of other populations, therefore CT genotype was less frequent compared with normal controls (CC, 0.62; CT, 0.21; TT, 0.17) (χ2 p value, 0.0171).

DISCUSSION

We found that the frequencies of ACG and ATG alleles are 0.62 and 0.38, respectively. The figures are similar to those in a report on Caucasians as well as in a recent report on Japanese, indicating that the frequency and the distribution of VDR exon 2 polymorphism is conserved between two different ethnic groups but not in African American subjects (15–17).

Among the three genotypes, the CT heterozygotes were the tallest in young female Japanese who had already attained their peak height. The association was also noted at age 13 y in spite the biologic variability in height at this age. A group with constitutional short stature had less frequent CT genotype. When we consider the function of vitamin D in the growth plate cartilage, intestine (an increase in calcium and phosphorus absorption), and kidney, it is possible to draw the conclusion that VDR gene polymorphism affects the difference in height in the healthy young population (1,2). The maturing growth plate chondrocytes have receptors for 1,25-D3 (3,4). The function of 1,25-D3 is reported to be inhibition of terminal differentiation of chondrocytes (4). Several reports indicate that the proliferation of chondrocytes is regulated by both 1,25-D3 and PTH (3,5–7). The supply of blood calcium and phosphorus is dependent on the actions of vitamin D in the intestine and the kidney (2). Consequently, deficiency in the actions of vitamin D gives rise to an extremely short stature in growing subjects and osteomalacia in adults (1,2). Very recently, Raivio et al. (18) reported that the boys with the variant LHβ gene including heterozygotes had shorter height (18).

The height growth is under strong genetic control. The most striking phenomenon is that the child's increase in height follows an exact channel as if growth is being channeled to reach a genetically predetermined target size, and this is termed "developmental canalization" (19,20). A child's height after age 2 correlates with its adult height and parental height (r = 0.5-0.75) (19,21,22). One of the applications of the results in clinical medicine is more accurate prediction of adult height based on the mean parental stature. A mother and father with a combination of both CC and TT homozygotes will have taller children (CT heterozygotes) than the predicted adult height based on the mean parental height. Further study will be required to test this hypothesis and if proved correct will be useful to modify the present's predictions based on parental height.

The functional significance of the differently sized VDRs, either 424 or 427 amino acids forms, is unclear. One study suggested that, despite similar affinity for 1,25-D3 of the VDR variants, the longer variants were 50% less efficient in vitamin D-mediated trans-activation of a reporter gene linked to the 24-hydroxylase promoter (16). Another study revealed that they were functionally the same (23). Our preliminary results indicated that the expression of the ATG allele mRNA in the peripheral mononuclear cells of the CT heterozygotes was twice as much as that of the ACG allele (data not shown). It is possible to consider that there is a difference in the level of VDR gene expression between ATG and ACG alleles. Thus, the tallest height in the heterozygotes and the possible difference in promoter activity and mRNA expression between two alleles may indicate that there could be functional significance in the presence of both ACG and ATG alleles, and each may have different properties, which need to be clarified in the future study.

In conclusion, we have shown that the VDR gene exon 2 polymorphism which encodes variant VDR proteins is one of the determinants of height. The CT heterozygotes attained a significantly greater height than either CC or TT homozygotes. This indicates a functional significance to the presence of both ACG and ATG alleles which requires further study.