Genetics of osteoporosis in Chinese


: Sumai Xiao, Department of Medicine, The University of Hong Kong, Queen Mary Hospital, 102 Pokfulam Road, Hong Kong. Email:


Osteoporosis is a common skeletal disease characterized by low bone mineral density (BMD) and deterioration in bone microarchitecture, resulting in increased bone fragility and susceptibility to fractures. As a complex disease, it is determined by both genetic and environmental factors, as well as their interactive effects. Studies have suggested that different genetic determinants may be involved in different ethnic groups. In this paper, we reviewed the genetic studies of osteoporosis in A Chinese population, focusing on the genes affecting BMD, a surrogate phenotype of osteoporosis.


Osteoporosis is a systematic skeletal disease characterized by low bone mineral density (BMD) and deterioration in bone microarchitecture, resulting in increased bone fragility and susceptibility to fractures. Like other common chronic diseases, it is predicted to increase rapidly worldwide as a result of the ageing population. Asia is expected to carry the major burden of this disease. It is projected that more than 50% of the world's hip fractureS will occur in Asia by 2050, mainly in China.1 The fifth census of China reported that the percentage of the elderly population above the age of 60 was 10.5% in 2000, and this is expected to be around 30%, that is about 300 million people, in 2050. In 2000, 6.97% of the Chinese population (about 88 million) were found to suffer from osteoporosis, with the direct cost for managing this disease being no less than 15 billion RMB.2


Osteoporosis is a complex disease determined by both genetic and environmental factors as well as their interactive effects. Its strong genetic control is supported not only by the clinical observation of a strong family history, but also by the high heritability estimated from family and twin studies. BMD is the most widely used surrogate phenotype to dissect the complex genetic structure of osteoporosis, and one of the important indices for the clinical diagnosis of osteoporosis. BMD is highly concordant between monozygotic than dizygotic Caucasian twins with heritability raging from 0.5 to 0.9.3–5 A strong familial aggregation of BMD is similarly confirmed in Chinese.6,7 The risk of osteoporosis among siblings is 2.6-fold in Chinese.6 Using the complex segregation model in 1260 individuals from 401 Chinese families, a high heritability estimate was observed at the spine (0.807 ± 0.099) and hip (0.897 ± 0.101) with a major additive effect on hip but not spine BMD.8 However, this finding differed from the inheritance model reported in Caucasians. Recently, Ng et al. (2006) used a much larger sample of 3320 subjects from 1019 southern Chinese families, and reported a sex-specific heritability with the heritability estimate of 0.63–0.71 for females and 0.74–0.79 for males.9 These results supported a strong genetic effect on BMD, which may be site and sex specific. Furthermore, the difference in the genetic model between Chinese and Caucasians may provide an evidence for genetic heterogeneity among various ethnic groups.


More than 200 candidate genes are reported to be associated with osteoporosis. The Sp1 polymorphism of collagen type 1 α1 (COLIA1) gene appears to be consistently related to BMD and its functionality has been suggested.10–12 However, studies performed in Asian populations revealed that the Sp1 polymorphism is almost nonexistent and hence its application to the risk assessment of osteoporosis in Asians is limited.13,14 It seems that the cohort of genes predisposing to the risk of osteoporosis varies among different ethnic groups, depending on the frequency of occurrence of the variant allele. Linkage studies and candidate gene association analysis have been performed in Chinese to search for genes underlying the aetiology of osteoporosis.

Linkage study

More than 10 whole genome linkage scans of osteoporosis in humans have been reported in both Caucasian and Chinese populations. A few linkage regions have been identified and replicated.15 The first small linkage study with 218 Chinese subjects from 96 families showed modest linkage to chromosomes 2p21 and 13q34 for forearm BMD.16 Recently, we conducted two region-wide linkage studies in 306 Chinese families with 1459 subjects and replicated previous linkage findings on chromosome 1q and 11q with BMD.17,18 Chromosome 11q13–2417 was linked with hip BMD with LOD (log of odds) score ranging of 1.2–1.6 and chromosome 1q2118 with spine BMD with LOD score of 2.36. Subgroup analysis showed the sex-specific quantitative trait loci (QTL) in chromosome 11q with 11q21 (LOD = 1.62) for women and chromosome 11q13 (LOD = 1.57) for men.17 Recently, Hsu et al. (2007)19 reported a large-scale genome-wide linkage study in 3093 Chinese sibships selected on the traits of extreme BMD at different skeletal sites, and found significant QTLs on chromosome 7p21.2 for femoral neck BMD (LOD = 3.68) and chromosome 2q24.3 for total hip BMD (LOD = 3.65). It also revealed a sex-specific linkage for lumbar spine BMD on chromosome 13q21.1 (LOD = 3.62) in women. In order to search for common QTL among different but highly correlated skeletal sites, the authors performed a multivariate linkage analysis by combining BMDs at whole body, total hip, femoral neck and lumbar spine, and revealed an additional significant QTL at chromosome 5q21.2 (LOD = 4.56). Interestingly, none of these significant QTLs found in this study overlapped with major QTLs reported by previous studies in non-Asian populations. This ethnic difference may imply that there is a Chinese-specific background of genetic determination of osteoporosis. Of course, the difference in the findings could also be caused by the different study designs and within ethnic group heterogeneity.

Candidate gene association study

About 100 association studies on 34 osteoporosis candidate genes in Chinese populations have been published (Table 1). These genes include receptors for calciotropic hormones, cytokines and growth factors that regulate bone turnover, bone matrix, and miscellaneous genes that have been implicated in the regulation of bone metabolism. Here we summarize and discuss the association results of four genes with BMD in the Chinese population, that is VDR, ESR1, ESR2 and LRP5.

Table 1.  The information of 34 osteoporosis candidate genes investigated in Chinese
GeneChromosomal locationGeneChromosomal location
  1. Notes: osteoprotegrin, OPG; interleukin-6, IL6; transforming growth factors β1, TGF-β1; leptin receptor, LEPR; tumour necrosis factor receptor 2, TNFR2; interleukin-1 receptor antagonist, IL1RN; collagen type 1 α2, COL1A2; COL1A1; alpha 2 Hs-glycoprotein, AHSG; osteocalcin, BGP; oestrogen receptor 1, ESR1; vitamin D receptor, VDR; oestrogen receptor 2, ESR2; parathyroid hormone, PTH; glucocorticoid receptor, GCCR; calcitonin receptor, CTR; aromatase, CYP19A1; calcium sensing receptor, CASR; androgen receptor, AR; cytochrome P450 1B1, CYP1B1; PTH-related peptide receptor 1, PTHR1; cytochrome P450c 17alpha, CYP17A1; cytochrome P450 1A1, CYP1A1; low density lipoprotein receptor-related protein 5, LRP5; 12/15-lipoxygenase, ALOX15; myostatin, MSTN; 5,10-methylenetetrahydrofolate reductase, MTHFR; PPAR gamma coactivator-1, PPARGC1; glutaminyl-peptide cyclotransferase, QPCT; apolipoprotein E, Apo E; adiponectin, ADIPOQ; calmodulin 2, CALM2; estradiol 17-beta-dehydrogenase-1, HSD17B1 and proopiomelanocortin, POMC.

Cytokines/growth factors related genesBone matrix related genes
Calciotropic hormones receptor genesMiscellaneous genes


VDR gene is the first reported candidate gene with BMD regulation and most attention has focused on BsmI, ApaI and TaqI polymorphisms situated at the 3′ flank of VDR gene. These three polymorphisms are located in the same block with strong linkage disequilibrium (LD) between each other in White and Asian, but not Black populations.20 The original report of VDR gene effects on BMD variation was published over 10 years ago. The follow-up replication studies were numerous, but results were conflicting. In Chinese, the frequencies of minor alleles are different from those in Caucasians but are similar to Koreans.21 Studies investigating these three polymorphisms with BMD in Chinese were similarly conflicting. FokI polymorphism is independent from any LD block of VDR gene and was suggested to create an alternative translational start site, resulting in the production of two isoforms of VDR protein that differ in length by three amino acids.22 None of the three studies focused on FokI polymorphism in Chinese showed significant individual effects on BMD variation, but the haplotype analysis displayed that FfTt, FfAa and FFaa predicted a higher BMD in postmenopausal women.23–25 Haplotype analysis may be more statistically powerful due to increased information by combining multiple polymorphisms. One of the major reasons for the inconsistency may be caused by the relatively modest effect of VDR gene on BMD variation, as the effect size of the BsmI polymorphism on BMD variation is only 0.1 SD (standard deviation) in postmenopausal women.26,27 A sample with more than 3000 random subjects is required to detect such modest effect of 0.1 SD at P = 0.01 with 80% power. However, most studies were conducted with a much smaller sample size.


Estrogen acts through the estrogen receptor to inhibit bone loss and ESR1 gene is a strong candidate for osteoporosis. Two of the frequently studied polymorphisms (PvuII and XbaI) are located in the intron 1 of ESR1 gene. Like VDR gene, the results were inconsistent about the individual contribution of ESR1 polymorphism to osteoporosis. A meta-analysis of ESR1 polymorphisms (PvuII and XbaI) and BMD in 4297 Chinese women28 showed that PvuII polymorphism is marginally associated with BMD at the femoral neck but not the spine in Chinese postmenopausal women. PP homozygosity predicts lower BMD than Pp/pp genotype. This finding is not observed in pre/perimenopausal Chinese women. These results differ from those of another meta-analysis of 18,917 Caucasian women whereby no significant association was detected between ESR1 polymorphisms and BMD variation.29 Reasons for the discrepancy among these meta-analysis studies are unclear. However, Chinese populations have a quite different lifestyle and also a varied distribution of gene polymorphisms from Caucasians. The genetic effects of ESR1 gene may be modulated by age, lifestyle, and other genes. Furthermore, since a meta-analysis is a summation of data, and if a very large but poorly done study is part of the meta-analysis, it may adversely impact the results. It should be pointed out that to date ESR1 gene has not been comprehensively screened by any study. The ongoing debate about its involvement in BMD variation may be simply due to the omission of a real causative marker.


As the second identified estrogen receptor in humans, ESR2 gene was mapped to human chromosome 14q22–24. The DNA binding domain of ESR2 possessed a similar structure and considerable homology with ESR1, consequently binding estrogen responsive elements with similar specificity and affinity.30 So far, the six association studies of ESR2 polymorphism with BMD in Chinese all showed positive results.23,31–35 We first reported the association of CA repeat polymorphism (D14S1026) located in intron 5 of ESR2 gene with hip BMD in Chinese premenopausal women.33 We subsequently also studied the relation of six tagSNPs with BMD in 752 case-control samples.35 The results suggested that the T-1213C polymorphism, the most significant SNP which is situated in the promoter region and is in LD with D14S1026, explained the association with BMD variation. The variant C allele predicted lower BMD and a 2–3-fold increased risk of osteoporosis in both Chinese men and women. The mechanism by which this variant affects ESR2 signalling has not been investigated.


LRP5 gene is a recently identified novel candidate gene for osteoporosis and is a key regulator of osteoblast proliferation and bone formation. The importance of this gene is originally substantiated by two rare monogenic bone disorders. Inactivation mutations of LRP5 are the cause of the recessive disorder osteoporosis pseudoglioma syndrome, whereas activating mutations give rise to autosomal dominant inheritance of high bone mass.36,37 The findings then led several investigators to evaluate the role of LRP5 as a candidate gene for BMD regulation in the normal population. The major association findings of BMD were related to 2 exon polymorphisms of LRP5 gene, rs4988321 (V667M) and rs3736228 (A1330V). Quite similar LD patterns of LRP5 gene were observed between Chinese and Caucasians, and the above two polymorphisms all belong to the LD block nearing-3′-UTR.38 In Chinese, the tested polymorphisms were different among studies, while all of the studies found the significant association result of BMD with polymorphisms located in the LD block nearing-3′-UTR of LRP5 gene. Zhang et al. (2005) reported that subjects with the Q89R (rs41494349) QQ genotype or the N740N (rs2306862) TT genotype had a significantly higher BMD at the femoral neck in 647 postmenopausal women.39 We observed the associations of BMD with these two polymorphisms in 674 subjects from 177 families as well.23 Besides, we also found an association between SNP rs682429 and BMD, but this finding was not replicated by Xiong et al.38,40 They studied 29 SNPs of LRP5 gene in 733 random Chinese and found the significant association of SNP rs491347 in intron 7 with spine BMD in both Chinese and Whites.38 Based on the HapMap and published data, the four SNPs with positive results located in the LD block nearing-3′-UTR of LRP5 gene (rs491347, rs4988321, rs2306862 and rs3736228) were in high LD with each other (D′ > 0.9) in Chinese.


Result inconsistency among studies is the major problem of gene identification of osteoporosis, which may result from the insufficient power of individual studies. Several reviews have recently addressed this issue and provided useful guidelines for subsequent genetic epidemiology studies.41,42 To increase the power of the study, a larger sample size and consideration of multiple variants and haplotypes should be included. So far most identified QTL showed small individual effects (< 5%) in the general population. A large population size in terms of several thousand subjects is generally necessary to gain enough power to detect weak association signals and generate reliable results. However, very few reported studies could reach this level of significance, and this could only be achieved by multicentre collaborations. Additionally, association studies presumed that the tested polymorphism is in LD with the causative variants. The structure of the human genome is suggested to be in a series of high LD regions separated by short segments of low LD. A small fraction of variants can explain most information of the high LD region, whereas low LD regions can only be characterized sufficiently by high dense markers. Furthermore, the complete resolution of common haplotypes highly depends on marker density. However, the majority of studies in Chinese tested only one or a few variants. The complete information of genetic variation and LD structure within a candidate gene will help to thoroughly examine the gene and identify the real disease variants.

With the progress of genetic structure dissection of osteoporosis, it has been recognized that interactions of gene-gene/gene-environment are indispensable modulators to the determination of disease variation, for example, interactions of ESR1 gene with VDR and IL6 gene on BMD variation were observed in Chinese women.43–46 These observations are consistent with the important function of estrogen, vitamin D and IL6 in bone metabolism. IL6 gene is tightly related to hip BMD variation in postmenopausal women, especially those without estrogen replacement therapy, and those with insufficient calcium intake.47 One of the actions of estrogen is suggested to act via inhibition of IL6 production.48 Vitamin D is involved in the regulating of balance between androgens and estrogens in peripheral tissues, which in turn modulates the availability of estrogens for their receptors.49 Interestingly, ESR1-VDR and ESR1–IL6 interactions were not found in Chinese men, which may suggest a sex-dependent phenomenon. Sex as an environmental factor might be particularly important in studying diseases with skewed sex ratios such as osteoporosis. Recent studies suggested that a sex-specific genetic contribution to BMD and gene–sex interactions may exist.9,50,51 Many other environmental factors may also have interactions with genes of osteoporosis through epigenetic mechanisms.52,53 These findings might imply that the above referred result inconsistency may also possibly be due in part to the omission of assessing interaction effects of gene-gene/gene-environment from the statistical models.

Major advances have been made in understanding the role of genetic factors of osteoporosis in Chinese and several candidate genes have been identified. However, to fully understand the multigenic nature of the disease, more studies with efficient methods are needed to unravel the other genes involved. The recently developed genome-wide association (GWA) study is a powerful new tool for unlocking the genetic basis of complex diseases systematically. Compared with the linkage approach, GWA is able to identify the actual disease genes efficiently by localizing small regions of chromosome containing a single to a few genes directly. This allows the identification of disease genes of small effect size. In contrast to candidate gene studies, GWA permits a systematical scan of genome in an unbiased fashion and thus has the potential to identify totally novel susceptibility genes.

Apart from targeting the gene level, gene expression studies are also needed to understand the pathogenesis of osteoporosis. The correlation between candidate genes and disease phenotype may not be identified through studying the DNA level under certain circumstances. RNA microarray and proteomics techniques can monitor gene expression for tens of thousands genes in parallel with relatively high sensitivity. The large volume of information obtained from the three combined approaches would enable the identification of multiple interacting disease genes and their respective pathways, allowing a comprehensive understanding of the aetiology of osteoporosis. These new strategies when applied in different ethnic groups will boost the discovery of osteoporosis candidate genes and provide totally new and ethnic-specific information on the aetiology of osteoporosis.