Genetic association studies in osteoporosis often bring discrepant results. Our study found no direct association between fracture prevalence and specific genotypes of the analyzed VDR polymorphisms. Our results are consistent with previous studies, i.e. meta-analyses by Uitterlinden [12
] and Fang [13
], where no relationship between BsmI, Apa
I polymorphisms with fracture risk was found. The results previously published by our team in a similar group of postmenopausal women are also in line with the current findings [10
Association analysis of single alleles made it possible to increase the sensitivity of the tests and revealed that the single alleles b, a, T of BsmI, Apa
polymorphisms, respectively, were overrepresented in patients with non-vertebral fractures compared to B, A, t alleles. The same was not shown for vertebral fractures. The B allele was therefore—contrary to some previous data [15
]—not shown to be protective for bone. Confirming the correlation between higher prevalence of the b allele and higher fracture risk, we made similar observations to those of Uitterlinden et al. [16
Some clinical studies proved a correlation between VDR polymorphisms and incidence of low-energy fractures. A recent large meta-analysis by Ji et al., including over 6,600 subjects, revealed a modest but significant association between hip fracture prevalence and lower frequency of bb genotype [17
]. Feskanich et al., in a group of women over 75 year of age, proved the BB genotype to be associated with a more than twofold increased risk of hip fracture compared with the bb genotype [18
]. Their findings are also consistent with the results obtained by Garnero et al. [19
Discordant results may be due to different populations studied/due to racial differences in BDM and fracture risk. Susceptibility to osteoporosis and fractures has been shown to be substantially lower in black subjects when compared to those of Caucasian and Asian origin [20
]. Studies by Morrison [9
] and Nguyen [22
] were carried out among ethnically diverse Australian population, additionally exposed to higher doses of UVB radiation, with potentially higher amounts of active vitamin D. In our study, the studied females were Caucasian.
It is more than likely that the VDR gene expression is affected by environmental factors. Several authors suggested that calcium homeostasis may play a role in this process. Stathopoulou et al. [23
], showed that, under lower calcium intake (<680 mg/d), the presence of the B allele of Bsm
I polymorphism and of the t allele of Taq
I polymorphism increased the risk of osteoporosis by 118 and 132 %, respectively. In the group of the higher (>680 mg/d) calcium intake, the influence of the VDR alleles on BMD was insignificant. Thus, the authors concluded that adequate calcium intake “masked” the VDR genetic influence on the bone. On the other hand, Gennari et al. [24
], observed that intestinal calcium absorption (self-assessment questionnaire) in healthy postmenopausal Italian women was significantly lower in BB and tt genotypes than in bb and TT genotypes, respectively, and in AABBtt genotype than in either aabbTT or AaBbTt genotypes.
Recently, as we gain new insights into vitamin D actions, interest in its non-genomic signaling pathways has increased, particularly regarding the trans-membrane calcium transport [25
]. There is growing evidence that local vitamin D hydroxylation differs and depends on calcium availability. An experimental study by Anderson et al., [26
] showed that with high dietary calcium, the activity of 1-α hydroxylase (CYP27B1) in the kidney is reduced, whereas in the bone it increases, promoting calcium incorporation into the bone. A negative correlation between serum vitamin D and mRNA CYP27B1 levels in the kidney, and positive—in the bone, was also noted. In other words, under the influence of high serum vitamin D, its hydroxylation in the bone is also more efficient, which results in adequate calcium availability in the bone. These findings further support the hypothesis that environmental factors, especially dietary calcium, may modulate the genotype-phenotype relationship, and therefore have an impact on the obtained results.
Secondly, it should be remembered the BsmI, Apa
I, and TaqI
polymorphisms do not have an effect on the final protein product, as they are found in the non-coding region of the VDR gene [27
]. This fact highlights the importance of understanding the mechanisms by which these polymorphisms affect the VDR action.
Thirdly, the activities of other polymorphisms within the large VDR gene should not be underestimated. According to the NCBI database, 180 single nucleotide polymorphisms (SNPs) and seven haplotypes [27
] have been found so far. It is therefore likely that the identification of specific polymorphisms may fail to provide definitive knowledge about the risk of fracture, and serve only as its approximation. In a recently published work, Ramagopalan and co-workers were able to find 2776 DNA-binding sites in the VDR gene by chromatin immunoprecipitation sequencing (Chip-seq) [28
], indirectly proving the pleiotropic effects of vitamin D. The binding sites were located in the promoter region of the gene, which implies that even the smallest mutations might affect the VDR gene function.
We did not find any association of the studied VDR gene polymorphisms with BMD. Morrison initially reported lower BMD in BB and tt homozygotes of BsmI i Taq
I polymorphisms [10
]. Also, meta-analysis by Thakkinstian and co-workers [29
], revealed weak but statistically significant association between B allele and lower BMD in the lumbar spine. Both mentioned polymorphisms are found in the same haplotype block of the VDR gene and are linked to each other. A paper by Ralston et al. [30
], who also did not show such relationship, is more in line with our work.
In addition to that complex scenario, there are non-BMD-related risk factors for fracture, e.g. falls. Again, they are associated with vitamin D. Proper vitamin D status has been shown to improve muscle mass and strength and to have a beneficial effect on motor coordination. Numerous (ca. 400) risk factors for falls (i.e. old age, poor vision, impaired hearing, imbalance, neurological disorders, medications, architectural barriers at home, etc.) have been identified [31
], but to consider of all them in a comprehensive assessment of risk fracture seems difficult, if not impossible. It is therefore assumed that the non-BMD fracture risk factors might have also been the reason of divergent results of the abovementioned studies.
Our results support the concept that population variants of the given genotype might be diverse and multifactorial. At the same time, we showed that there is no universal insight into the genotype-phenotype relation. Thus, each study that offers a possible explanation of this complex issue would be beneficial to clinicians.
The prevalence of the BB genotype in the studied population was 16.4 %, which is similar to the prevalence observed in other samples of Caucasians populations [32
], proving that our sample was representative. Since we investigated women with a specific disease—that is with a specific phenotype—the results seem to be no less valuable than the results of population studies. Another advantage is the large sample size and ethnic homogeneity.
In the multifactorial etiology of osteoporosis, the role of genetic factors is undeniable, but complex. We managed to show that the presence of the single alleles a, b and T of ApaI, BsmI and TaqI polymorphisms of the VDR gene may be a predictor of low-energy fractures. However, lack of association between the VDR gene polymorphisms and BMD suggests different ways in which the former contributes to fracture risk. The question whether and how VDR genotypes influence bone microarchitecture requires further investigation.