Case Control Association Study of WNT Signaling Pathway-related Genes and Osteoporosis Risk Among Chinese Postmenopausal Women

Background: The WNT signaling pathway is involved in the regulation of bone homeostasis, and the effects of WNT signaling pathway related genes (WLS, WNT16 and LRP5) on osteoporosis risk among Chinese postmenopausal women is still unknown. Hence, we performed a case-control study to assess the association of WNT signaling pathway related genes and osteoporosis risk. Methods: This study involved 1026 women (515 osteoporosis patients and 511 controls) of postmenopausal age who were randomly sampled from Xi'an 630 Hospital, Shaanxi Province, China. Eleven genetic polymorphisms in WLS (rs2566755, rs12407028, rs2566752 and rs7554551), WNT16 (rs3779381, rs3801387, rs917727 and rs7776725) and LRP5 (rs2291467, rs11228240 and rs12272917) were selected and genotyped using the Agena MassARRAY iPLEX system. The association of the genetic polymorphisms and osteoporosis risk was assessed by odds ratios and 95% condence intervals. The Multifactor Dimensionality Reduction (MDR) method was conducted to analyze SNP-SNP interaction. Results: We found that LRP5 polymorphisms (rs2291467, rs11228240 and rs12272917) were signicantly associated with a decreased risk of osteoporosis in homozygotes, both in recessive and additive models (P < 0.05). Stratication analysis showed that LRP5 polymorphisms (rs2291467, rs11228240 and rs12272917) signicantly decreased the osteoporosis risk in the subgroup of BMI ≤ 24 (P < 0.05) and that individuals carrying a heterozygote genotype of WNT16 polymorphisms (rs3779381, rs3801387, rs917727 and rs7776725) had a higher osteoporosis risk in the subgroup of BMI > 24 (P < 0.05). We observed that three blocks (block 1: rs2566755 and rs12407028, block 2: rs3779381, rs3801387, rs917727 and rs7776725, block 3: rs2291467 and rs11228240), T rs2291467 T rs11228240 and C rs2291467 C rs11228240 had a strong association with a lower risk of osteoporosis. Additionally, MDR analysis revealed that a four-locus model (rs2566752 and rs2566755 in WLS, rs7776725 in WNT16, rs12272917 in LRP5) was signicantly associated with osteoporosis risk. Conclusions: Our ndings suggested that WLS, WNT16 and LRP5 genetic polymorphisms were associated with osteoporosis risk among Chinese postmenopausal women. association between related genes of the WNT rs2566755, rs12407028, rs2566752 and rs7554551; WNT16: rs3779381, rs3801387, rs917727 and rs7776725; LRP5: rs2291467, rs11228240 and rs12272917) and osteoporosis risk among Chinese postmenopausal women. to L1 = 0.006) the of 0.010) = 0.007). PLT (P = 0.010), PDW (P = 0.017), (P = 0.011) and PCT (P = 0.045) no signicant relationships between WLS rs12407028, WNT16 rs3779381 and WNT16 rs3801387 and clinical indicators (P 0.05). the of two LRP5 haplotypes rs2291467 T rs11228240 C rs2291467 C rs11228240


Introduction
Osteoporosis is a skeletal disorder characterized by low bone mineral density (BMD) and a compromised bone microarchitecture, which lead to increased risks of bone fragility and fracture [1]. Osteoporosis has become a major healthcare problem worldwide, affecting more than 200 million people [2]. There are gender differences in the occurrence of osteoporosis; postmenopausal women are more likely to have osteoporosis due to the sharp decline of estrogen production [3]. In China, the overall prevalence of osteoporosis ranges from 6.6-19.3% and increases with age [4]. It is estimated that the osteoporosis prevalence of Chinese women aged 50 years or older is more than 40% [4]. The causes of osteoporosis are multifactorial, including age, sex, hormones, physical activity, diets, genetic factors and environmental factors. Among these, genetic factors play an important role in osteoporosis development. Moreover, increasing ndings suggest that genes related to the WNT signaling pathway have effects on the onset and development of osteoporosis [5][6][7][8][9].
Wntless (WLS) is the key component required for WNT sorting and secretion and is also known as EVI or GPR177. WLS encodes a transmembrane protein that is required for the proper shuttling of WNT ligands involved in speci ed secretory pathways [10][11][12]. Overexpression of WLS leads to an interaction with WNT3A and VPS35, which suggests a nonredundant role of WLS on WNT regulation [13]. Fu et al. reported that disruption of WLS impairs the development of the body axis in mice [14]. In humans, Augustin et al. found that WLS is overexpressed in malignant astrocytoma and is related to the proliferation, survival and migration of glioma cells. Meanwhile, the expression level of WLS plays an essential role in some diseases, such as breast cancer, ovarian and gastric cancer [15][16][17][18]. Moreover, WLS has been identi ed as a novel locus for BMD and osteoporosis [19]. However, knowledge of the role of WLS in osteoporosis risk is limited to ethnicity, and the impact of the WLS pathway on the regulation of osteoporosis is unknown. Wnt family member 16 (WNT16) spans 16 kb from the initiation to termination codons and encodes two protein isoforms (WNT16A and WNT16B) [20]. WNT16A is expressed in the pancreas, whereas WNT16B is highly expressed in the kidney, placenta, brain, heart and spleen [21]. WNT16 belongs to the WNT family, which regulates bone and mineral homeostasis [22]. The disruption of WNT16 in mice leads to lower cortical bone density and strength and an increased risk of fracture [23]. In premenopausal women, WNT16 polymorphisms are associated with peak BMD [24]. Recently, several genome-wide association studies (GWAS) have revealed that WNT16 polymorphisms are associated with BMD and fracture risk across multiple populations [24][25][26][27]. These studies suggest that WNT16 is vital for bone mass regulation, which might be a promising biomarker of osteoporosis.
Low-density lipoprotein receptor-related protein 5 (LRP5) is located on chromosome 11q13. 4 and encodes a WNT receptor that regulates osteoblastic activity and bone mass [28]. LRP5 is also related to glucose and cholesterol metabolism, and thus, LRP5 is involved in the development of metabolic diseases, such as diabetes, obesity and vascular diseases [29]. Moreover, numerous studies have indicated that the LRP5 gene in uences BMD and fracture [30]. Mouse models show that LRP5 regulates bone formation by controlling osteoblast proliferation [31]. In humans, loss of LRP5 expression is linked to osteoporosis-pseudoglioma syndrome (low BMD and skeletal fragility) [28]. However, there are no studies on the association of LRP5 polymorphisms (rs2291467, rs11228240 and rs12272917) with osteoporosis among Chinese postmenopausal women.

Study population
This case-control study was conducted in 515 osteoporosis patients (women) and 511 healthy postmenopausal women who were recruited from Xi'an 630 Hospital. All cases were de ned by a bone mineral density (BMD) T score of less than − 2.5 in the lumbar spine (L1 to L4). The exclusion criteria included (1) patients with secondary osteoporosis, (2) patients treated with medications that may in uence BMD, (3) postmenopausal individuals who were not on medication, and (4) individuals with other diseases, such as immune disorders, cancer, or metabolic or endocrine diseases. The controls were generally healthy subjects aged approximately 63 years. We collected basic and clinical characteristics from their medical records. This study was performed in accordance with the World Medical Association Declaration of Helsinki and was approved by the Ethics Committee of Xi'an 630 Hospital. Written informed consent was obtained prior to the study.

Genotyping
Four WLS SNPs (rs2566755, rs12407028, rs2566752 and rs7554551), four WNT16 SNPs (rs3779381, rs3801387, rs917727 and rs7776725) and three LRP5 SNPs (rs2291467, rs11228240 and rs12272917) were selected in this study based on data from the dbSNP database (https://www.ncbi.nlm.nih.gov/snp/) and 1000 Genomes Project (http://www.internationalgenome.org/). Genomic DNA was isolated from whole blood using a DNA extraction kit (GoldMag Co. Ltd., Xi an, China), and the DNA concentration was measured by a Nanodrop 2000 (Thermo Scienti c, Waltham, Massachusetts, USA). Genotyping was carried out using the MassARRAY iPLEX system (Agena Bioscience, San Diego, CA, USA). The primers for this study were designed using the Agena MassARRAY Assay Design 3.0 Software (San Diego, California, USA, Supplemental Table 1). We managed and analyzed the data on the Agena Typer 4.0 Software (San Diego, CA, USA). Additionally, we used Haploview software (version 4.2) to perform haplotype analysis.

Statistical analysis
We performed statistical analysis using SPSS version 21.0 software (SPSS, Chicago, IL, USA). The χ2 test or t test was applied to examine the distribution difference of characteristics. Deviation from Hardy-Weinberg equilibrium (HWE) of each SNP in the controls was tested using Fisher's exact test. The effects of genetic polymorphisms on osteoporosis risk, adjusted with or without age, were assessed by odds ratios (OR) and 95% con dential intervals (95% CI) using logistic regression. We also conducted haplotype analysis of the association of genetic polymorphisms with osteoporosis risk and calculated linkage disequilibrium (LD) using Haploview software (version 4.2) and PLINK software. Moreover, we used multifactor dimensionality reduction (MDR, version 3.0.2) to evaluate the effects of SNP-SNP interactions on osteoporosis risk. All tests were twosided, and the signi cance threshold set at P < 0.05.

Characteristics of the study population
The characteristics of all the participants are shown in Table 1, including age, body mass index (BMI), uric acid (UA), urea, fasting glucose, white blood cell (WBC), red blood cell (RBC), platelets (PLT), platelet distribution width (PDW), mean platelet volume (MPV), plateletocrit (PCT), L1, L2, L3 and L4. The mean ages of the cases and controls were 63.72 ± 5.58 and 62.87 ± 4.68 years, respectively. There were signi cant differences in age, UA, urea, fasting glucose, WBC, RBC, PLT, PDW, MPV and PCT between the two groups (P < 0.05). As predicted by HaploReg, the functions of the genetic polymorphisms were related to promoter histone marks, enhancer histone marks, motif changes, NHGRI/EBI GWAS hits, GRASP QTL hits, selected eQTL hits, SiPhy cons, DNAse and proteins bound. In Table 2  We then performed strati cation analysis of the genetic polymorphisms with osteoporosis risk (Table 3). We found that rs2291467, rs11228240 and rs12272917 signi cantly decreased osteoporosis risk in the subgroup of age > 60 years (rs2291467: homozygote, OR For individuals with higher BMI (BMI > 24), WNT16 polymorphisms (rs3779381, rs3801387, rs9117727 and rs7776725) were signi cantly associated with a higher risk of osteoporosis in multiple models (P < 0.05). In addition, LRP5 polymorphisms (rs2291467, rs11228240 and rs12272917) had a strong relationship with an increased risk of osteoporosis among individuals with lower BMI (BMI ≤ 24). We further conducted haplotype analysis of the genetic polymorphisms with osteoporosis risk (Table 4, Fig. 1-3). We observed three blocks: block 1 (rs2566755 and rs12407028), block 2 (rs3779381, rs3801387, rs917727 and rs7776725) and block 3 (rs2291467 and rs11228240). As shown in Table 4, the haplotype T rs2291467 T rs11228240 (with adjustment:  Bold values mean statistical signi cance.

Association of genetic polymorphisms and clinical indicators
In Table 5 and Supplemental Table 3

MDR analysis
We used MDR analysis to assess the effect of SNP-SNP interactions among the eleven selected SNPs in WLS, WNT16 and LRP5 (Table 6, Supplemental Table 4 -6). In total, we found four models, and a four-locus model including rs2566752 and rs2566755 in WLS, rs7776725 in WNT16 and rs12272917 in LRP5 was the best model (CV consistency = 9/10, accuracy = 0.605, P < 0.001). We then showed SNP-SNP interactions with genotype combinations (Fig. 4-6) and revealed that the impact of the SNP genotypes of three genes on osteoporosis risk are dependent on one another.
Osteoporosis is a disease in which bone density and strength decrease, leading to skeletal dysfunction [32]. The WNT signaling pathway is one of the major signaling pathways involved in maintaining bone homeostasis and is also a key signaling pathway of other biological processes, such as embryogenesis, organogenesis and tumorigenesis [33]. It is reported that reduced WNT signaling results in increased osteoclast activity and decreased osteoblast activity [5]. Several loci clustered within the WNT signaling pathway are strongly associated with BMD, including WLS, WNT16 and LRP5. WLS is an integral component of the WNT ligand secretion pathway, which is required for the development of various organs [34]. Genetic variation in the WLS gene has been studied in several GWAS; rs2566752 is the maximally associated variant for lumbar spine and femoral neck BMD [35,36]. However, the association of WLS polymorphisms and the risk of BMD or osteoporosis is inconsistent [37], which may be attributed to a limited sample size and ethnicity differences. In our study, we did not observe a strong relationship between WLS polymorphisms (rs2566755, rs12407028, rs2566752 and rs7554551) and osteoporosis susceptibility among Chinese postmenopausal women.
Nevertheless, we found a signi cant association between some clinical indicators and different genotypes of WLS polymorphisms, suggesting that genotypes of WLS polymorphisms have effects on the level of clinical factors.
WNT16 is a noncanonical WNT ligand that regulates bone resorption and thus inhibits osteoporosis [38]. It is reported that genetic polymorphisms of WNT16 are associated with BMD at different bone sites. Although the relationship of rs3801387 and BMD is controversial [39], we found that rs3801387 signi cantly increased the osteoporosis risk of Chinese postmenopausal women with higher BMI (> 24), suggesting the potential importance of this variant in osteoporosis. Rs917727 is another WNT16 polymorphism; a GWAS showed that rs917727 in uences skull BMD [25]. Consistent with previous studies, we observed a strong linkage of rs917727 and osteoporosis risk in strati cation analysis. In addition, we assessed the relationship of two other polymorphisms (rs3779381 and rs7776725) with osteoporosis. We rst found that individuals carrying the heterozygote genotype of rs3779381 had higher osteoporosis risk. In Caucasians, rs7776725 is associated with hip, spine, wrist and whole-body BMD [26], whereas rs7776725 is associated with osteoporosis in Chinese women with higher BMI. These studies revealed that WNT16 polymorphisms (rs3779381, rs3801387, rs917727 and rs7776725) might affect the risk of osteoporosis. Further studies are needed to con rm the role of WNT16 polymorphisms on the pathology of osteoporosis.
LRP5 encodes a transmembrane protein in the WNT signaling pathway and regulates bone mass in humans and mice [40]. Several studies have shown that LRP5 polymorphisms are a determining factor of bone mass [41]. In this study, we rst found that rs2291467, rs11228240 and rs12272917 of LRP5 were signi cantly associated with a decreased risk of osteoporosis and that the relationships were BMI-dependent. We also observed the association of two LRP5 haplotypes (T rs2291467 T rs11228240 and C rs2291467 C rs11228240 ) with osteoporosis in Chinese postmenopausal women. These results provided further insight into the cause of osteoporosis; however, the complete mechanism of osteoporosis requires further exploration. Moreover, MDR analysis elucidated the association between gene-gene interactions and osteoporosis risk in the Chinese population.
Additional studies to determine the exact mechanism of gene-gene interactions on osteoporosis treatment are necessary.
Some limitations of this study are noteworthy. First, we did not have information about lifestyle habits, physical activity or other clinical indicators, and therefore, we could not explore the associations with these variables or adjust them. Second, we were not able to genotype all the polymorphisms in WLS, WNT16 and LRP5. Third, all the participants in this study were enrolled from the same hospital, and therefore, selection biases are not neglected. Furthermore, functional experiments should be conducted in a larger sample to con rm the relationship of WNT signaling pathway related genes and osteoporosis risk.

Conclusions
In summary, we found a signi cant association of WNT signaling pathway related genes (WLS, WNT16 and LRP5) and osteoporosis risk among Chinese postmenopausal women. More individuals from different areas and countries in further studies are needed to verify our ndings.

Abbreviations List
Multifactor   Linkage disequilibrium (LD) plots containing three LRP5polymorphisms. Block 1 includes rs2291467 and rs11228240. The numbers inside the diamonds indicate the D for pairwise analyses.

Figure 4
Multifactor dimensionality reduction (MDR) analysis of the WLS rs2566755-LRP5 rs12272917 interaction. In each box, the left bar represents cases and the right bar represents controls. The light gray boxes indicate a low risk of osteoporosis, and the dark gray boxes indicate a high risk of osteoporosis; empty boxes indicate no data.

Figure 5
Multifactor dimensionality reduction (MDR) analysis of the WLS rs2566752-WLS rs2566755-WNT16 rs3801387 interaction. In each box, the left bar represents cases and the right bar represents controls. The light gray boxes indicate a low risk of osteoporosis, and the dark gray boxes indicate a high risk of osteoporosis; empty boxes indicate no data.

Figure 6
Multifactor dimensionality reduction (MDR) analysis of the WLS rs2566752-WLS rs2566755-WNT16 rs7776725-LRP5 rs12272917 interaction. In each box, the left bar represents cases and the right bar represents controls. The light gray boxes indicate a low risk of osteoporosis, and the dark gray boxes indicate a high risk of osteoporosis; empty boxes indicate no data.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementalTables.docx