Association of Plasminogen Activator Inhibitor-1 (PAI-1) Gene Polymorphisms with Osteoporotic Vertebral Compression Fractures (OVCFs) in Postmenopausal Women

Osteoporosis and osteoporotic fractures are strongly associated with mortality and morbidity, both in developing and developed countries. Menopause accelerates bone loss due to estrogen deficiency and age-related linear bone loss. We investigated plasminogen activator inhibitor-1 (PAI-1) gene polymorphisms in postmenopausal women with osteoporotic vertebral compression fractures (OVCFs). In this case-control study, 355 postmenopausal women were genotyped for the presence of PAI-1 gene polymorphisms −844A > G, −675 4G > 5G, 43G > A, 9785A > G, and 11053T > G. Genetic polymorphisms of PAI-1 were analyzed by the polymerization chain reaction restriction fragment length polymorphism assay, and their association with disease status and folate and homocysteine levels was determined in 158 OVCF patients and 197 control subjects. The PAI-1 −675 5G5G (adjusted odds ratio (AOR), 3.302; p = 0.017) and 43GA + AA (AOR, 2.087; p = 0.042) genotype frequencies showed significant association with the increased prevalence of OVCFs in postmenopausal women. In addition, we performed gene–environment interaction studies and demonstrated an association between PAI-1 gene polymorphisms and OVCF prevalence. Our novel finding is the identification of several PAI-1 genetic variants that increase susceptibility to OVCF. Our findings suggest that polymorphisms in PAI-1 may contribute to OVCF, and that they can be developed as biomarkers for evaluating OVCF risk.


Introduction
Osteoporosis is a common metabolic bone disorder characterized by reduced bone mass, increased skeletal fragility, microarchitectural deterioration, and an increase in bone fractures [1]. Osteoporosis leads to decreased skeletal strength and increased fracture susceptibility. Osteoporotic fractures are a

Results
Demographic characteristics and clinical profiles of osteoporosis patients (with/without osteoporotic vertebral compression fracture, OVCF) and controls are summarized in Table 1 and  Table S1.

Combined Effects between PAI-1 Polymorphisms and Environmental Factors
To determine additional clinical significance, we performed stratified (Table S6) and gene-environment interaction (Table S7) analyses according to hypertension, diabetes mellitus, 25-hydroxyl vitamin D (25-OH vit. D), vitamin B 12 , folate, and plasma total homocysteine (tHcy). We divided subjects into two groups representing the upper and lower 15% cutoff values for plasma tHcy and folate levels (folate: 4.40 ng/mL, tHcy: 12.41 µmol/L). We performed stratified analyses comparing clinical factors and polymorphisms (Table S6). Clinical risk factors strongly correlated with PAI-1 −844 G > A, and −675 4G > 5G polymorphisms. In gene-environment interaction studies (Table S7), the PAI-1 −675 polymorphism, coupled with low plasma folate levels or high homocysteine levels, was associated with an increased odds ratio when comparing osteoporosis patients with control subjects. In addition, when comparing the association of PAI-1 +43 and plasma folate levels with osteoporosis ( Figure 1), the PAI-1 +43 GA + AA genotype combined with low plasma folate levels produced a significant result (GG vs. GA + AA: AOR, 9.247; 95% CI, 1.049-81.51). In addition, we evaluated the effects of PAI-1 genotypes on bone mineral density (BMD), plasma vitamin B 12 , 25-OH vit. D concentrations, and bone metabolism markers (osteocalcin and deoxypyridinoline) in Table S8. PAI-1 −844G > A was significantly associated with BMD. Reduced osteocalcin and deoxypyridinoline were influenced by PAI-1 −844GA + AA and 11053TG + GG genotypes, respectively ( Figure S2).

Discussion
Osteoporosis and low BMD commonly occur in postmenopausal women and lead to increased susceptibility to fractures. BMD, along with other factors, is known to contribute to skeletal health [24]. In particular, the role of genetic factors, including gene polymorphisms, takes center stage in pathogenic studies [8,25,26]. Recent studies investigated synergic effects between gene polymorphisms and OVCF risk factors [27][28][29].
PAI-1 is a downregulator of the fibrinolysis gene and the key factor for plasminogen ratios [23,30]. Rundle et al. [31] showed that PAI-1 deficiency results in an increase in not only callus size but also callus cartilage regeneration in PAI-1 knockout (KO) mice recovering from fractures. In addition, fracture callus development was promoted in PAI-1 KO mice in comparison to control mice. These results showed strengthened tissue fibrinolysis in PAI-1 deficiency, which may enhance fracture recovery by facilitating extracellular matrix regeneration [23,31]. This is consistent with our current finding that an abnormal plasminogen level adversely affects bone recovery and regeneration. Furthermore, fibrinolysis in tissue can be a crucial factor in both acute inflammation and fracture restoration [23,32].
A previous study [33] described the 10 most well-known polymorphisms of PAI-1, as follows: rs2227631 G > A, rs6092 G > A, rs2227708 C > T, rs2227662 C > T, rs2227666 G > A, rs2227667 A > G, rs2227672 G > T, rs2227683 A > G, rs2227694 G > A, and rs7242 T > G. The rs1799889 (PAI-1 −675 4G > 5G) polymorphism is widely genotyped in PAI-1 genetic studies. The five polymorphisms selected for our study affect transcription and are expected to regulate PAI-1 levels. Therefore, we hypothesized that regulation of PAI-1 activity by PAI-1 polymorphisms results in an abnormality in the bone repair system or bone loss mechanism. Furthermore, postmenopausal women have an increased risk for OVCF because of sex hormone dysregulation. We have identified an association between PAI-1 polymorphisms and OVCF prevalence. Based on statistical analyses, this result demonstrates a significant correlation between PAI-1 −844A > G, −675 4G > 5G, and 43G > A, and OVCF risk. However, another study reported the lack of any association between the rs1799889 (PAI-1 −675 4G > 5G) polymorphism and osteoporosis in Turkish postmenopausal women [34]. Interestingly, this finding conflicts with the results of our research; the discrepancy could be due to Combinations of the PAI-1 +43 genotype and plasma folate level subgroups. We divided subjects into two groups representing the lower 15% cutoff values for plasma folate levels: >4.40 ng/mL and ≤4.40 ng/mL.

Discussion
Osteoporosis and low BMD commonly occur in postmenopausal women and lead to increased susceptibility to fractures. BMD, along with other factors, is known to contribute to skeletal health [24]. In particular, the role of genetic factors, including gene polymorphisms, takes center stage in pathogenic studies [8,25,26]. Recent studies investigated synergic effects between gene polymorphisms and OVCF risk factors [27][28][29].
PAI-1 is a downregulator of the fibrinolysis gene and the key factor for plasminogen ratios [23,30]. Rundle et al. [31] showed that PAI-1 deficiency results in an increase in not only callus size but also callus cartilage regeneration in PAI-1 knockout (KO) mice recovering from fractures. In addition, fracture callus development was promoted in PAI-1 KO mice in comparison to control mice. These results showed strengthened tissue fibrinolysis in PAI-1 deficiency, which may enhance fracture recovery by facilitating extracellular matrix regeneration [23,31]. This is consistent with our current finding that an abnormal plasminogen level adversely affects bone recovery and regeneration. Furthermore, fibrinolysis in tissue can be a crucial factor in both acute inflammation and fracture restoration [23,32].
A previous study [33] described the 10 most well-known polymorphisms of PAI-1, as follows: rs2227631 G > A, rs6092 G > A, rs2227708 C > T, rs2227662 C > T, rs2227666 G > A, rs2227667 A > G, rs2227672 G > T, rs2227683 A > G, rs2227694 G > A, and rs7242 T > G. The rs1799889 (PAI-1 −675 4G > 5G) polymorphism is widely genotyped in PAI-1 genetic studies. The five polymorphisms selected for our study affect transcription and are expected to regulate PAI-1 levels. Therefore, we hypothesized that regulation of PAI-1 activity by PAI-1 polymorphisms results in an abnormality in the bone repair system or bone loss mechanism. Furthermore, postmenopausal women have an increased risk for OVCF because of sex hormone dysregulation. We have identified an association between PAI-1 polymorphisms and OVCF prevalence. Based on statistical analyses, this result demonstrates a significant correlation between PAI-1 −844A > G, −675 4G > 5G, and 43G > A, and OVCF risk. However, another study reported the lack of any association between the rs1799889 (PAI-1 −675 4G > 5G) polymorphism and osteoporosis in Turkish postmenopausal women [34]. Interestingly, this finding conflicts with the results of our research; the discrepancy could be due to differences in the ethnic group or sample size used. On the other hand, the Turkish postmenopausal women exhibited only osteoporosis [35], whereas our study participants exhibited symptoms of greater severity, such as compound fracture due to osteoporosis; this could also be the reason for the discrepancy between the results of the two studies. There are several limitations to our study. First, serum PAI-1 concentrations were not examined in our study population. Previous studies have evaluated the contribution of each PAI-1 polymorphism to serum PAI-1 expression [35][36][37]. Second, functional studies for PAI-1 SNP were not performed to elucidate OVCF-related pathogenesis. Although several studies have reported an association between PAI-1 polymorphisms and bone repair, few have evaluated the pathogenesis by which PAI-1 polymorphisms affect osteoporosis in postmenopausal women. This study cannot, therefore, propose a detailed pathogenesis by which PAI-1 polymorphism affects vertebral artery and tissue. Finally, PAI-1 expression depends on the PAI-1 single nucleotide polymorphism (SNP) genotypes in the local tissue; therefore, PAI-1 expression in vertebral tissue is more important than that in serum.

Study Population
This was a case-control study, in which the case group consisted of 158 osteoporotic postmenopausal women (mean age ± SD, 69.50 ± 8.24 years; age range, 56-81 years), defined as having a T-score of −2.5 or lower and/or at least one non-traumatic fracture of the spine. Women were recruited from the Neurosurgery and Orthopedic Surgery Departments at CHA Bundang Medical Center, CHA University of South Korea between March 2005 and December 2008. The control group consisted of 197 postmenopausal women (mean age ± SD, 66.57 ± 8.05 years; age range, 50-85 years) without disease and not taking any medications known to influence bone mass or bone turnover. BMD at the lumbar spine was measured using a dual-energy X-ray (Hologic Discovery W, Waltham, MA, USA). An OVCF was defined as a ≥15% reduction in the anterior, posterior, or central height of the vertebra [38]. All examinations were performed by a trained neurosurgeon and orthopedic surgeon according to previously described methods [39]. All subjects provided their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of the Institutional Review Board of CHA Bundang Medical Center on 20 January 2014 (reference No. BD2015-043). We provided each participant with a full explanation of participation in this study and obtained confirmed written consent from all study participants. Bone mineral density (BMD) of the lumbar spine (L2-L4), femur neck (FN), Ward's triangle, and trochanter (Tro) were measured using dual-energy X-ray absorptiometry (Norland Medical Systems, White Plains, NY, USA). Osteoporosis was defined according to the 1994 classification of the World Health Organization (WHO) [40].

Genotyping
We performed DNA extraction from peripheral blood with an anticoagulant tube using the G-DEX™ II kit (Intron Biotechnology, Seongnam, Korea). The PAI-1 polymorphisms were amplified by polymerase chain reaction and digested with a restriction enzyme to identify them via restriction fragment length polymorphism (RFLP ; Table S9) [41]. We selected 30% of the PCR product at random and checked the RFLP results for concordance with DNA sequencing results. The concordance of the quality control samples was 100%.

Bone Measurements
The diagnosis of osteoporosis was based on WHO criteria (dual energy X-ray absorptiometry (DXA) hip or lumbar T-score <−2.5 standard deviations), and an OVCF was diagnosed when a progressive or newly generated compression fracture was identified after low-energy trauma. The BMD T-score was measured in g/cm 2 using Hologic densitometers (Hologic Inc., Waltham, MA, USA) [39]. The control group had a BMD T-score greater than −1.0 and had no spine or hip fractures based on the results of simple radiography. The BMD T-score analysis in this study was conducted with reference to the WHO report [40].

Measurement of Vitamin B 12 , Plasma Total Homocysteine (tHcy), and Folate Levels
We collected plasma samples to measure vitamin B 12 , tHcy, and folate levels within 48 h of osteoporosis detection. Twelve hours after the patient's previous meal, we collected whole blood in a tube containing anticoagulant. We centrifuged the tube for 15 min at 1000× g to separate the plasma and stored it at −80 • C. Plasma vitamin B 12 concentrations were measured using the BioRad Quantaphase II radioassay (Hercules, CA, USA). We measured plasma tHcy concentration using a fluorescent polarizing immunoassay with the IMx system (Abbott Laboratories, Chicago, IL, USA) and the plasma folate concentration with a radio assay kit (ACS 180; Bayer, Tarrytown, NY, USA).

Statistical Analysis
The genotype and allele combination frequency differences between case and control subjects were analyses by logistic regression and Fisher's exact test. The adjusted odds ratio (AORs) and confidence intervals (CIs) were used as a measure of the association between genotype frequencies and osteoporosis. All of the alleles were in Hardy-Weinberg equilibrium (p > 0.05). The genotypes with osteoporosis and OVCF occurrence was calculated with AORs and 95% CIs from logistic regression adjusted for age, hypertension, and diabetes mellitus. The linkage disequilibrium (LD) with blocking calculation with HaploView software and haplotype analysis for the models chosen by multifactor dimensionality reduction (MDR) methods were checked with the HAPSTAT program version 3.0 (www.bios.unc.edu/~lin/hapstat). The correlations of genotypes with BMD, as well as plasma folate, vitamin B 12 , 25-hydroxyl vitamin D, osteocalcin and deoxypyridinoline levels, were evaluated by Kruskal-Wallis test and Mann-Whitney test. Statistical analyses were performed by GraphPad

Conclusions
In summary, we found an association between PAI-1 polymorphisms and osteoporosis risk in postmenopausal women. Furthermore, PAI-1 polymorphisms were shown to strongly association with OVCF occurrence. Interestingly, our results suggest differences between non-OVCF and OVCF patients in this association with PAI-1 polymorphisms. In particular, the PAI-1 −844G > A, −675 4G > 5G, +43G > A polymorphisms were associated with increased susceptibility for OVCF rather than non-OVCF and osteoporosis. Therefore, we expect a potentially higher fracture risk in postmenopausal women who are osteoporosis patients with these PAI-1 polymorphisms. Our findings present the first report on the association between PAI-1 polymorphisms and OVCF. In conclusion, this study suggests that PAI-1 polymorphisms (−844G > A, −675 4G > 5G, +43G > A) may contribute to OVCF and could be used as genetic biomarkers for OVCF risk.
Supplementary Materials: Supplementary materials can be found at www.mdpi.com/1422-0067/17/12/2062/s1. and analyzed the data. Woo Sik Lee, In Bo Han, Dong Eun Shin and Soo Hong Han discussed and interpreted the data and results. Jung Oh Kim wrote the first draft of the manuscript. All authors contributed to and have approved the final manuscript.

Conflicts of Interest:
All authors declare no conflicts of interest.