Introduction

Osteonecrosis of the femoral head (ONFH) is a devastating disease that frequently leads to the progressive collapse of the femoral head, followed by degenerative arthritis of the hip joint. Non-traumatic ONFH has been associated with various factors such as corticosteroid usage, alcoholism, infections, marrow infiltrating diseases, coagulation defects, and some autoimmune diseases (Mont et al., 1998; Jones, 1999; Glueck et al., 2003; Boss et al., 2004; Childs, 2005). Among several confounding pathogenic mechanisms regarding ONFH, a vascular hypothesis appears to be the most persuasive, presuming that a decrease in the local blood flow in the femoral head plays a pivotal role in the pathogenesis of ONFH (de Camargo et al., 1984; Kerachian et al., 2006).

Particularly, some autoimmune disorders, including systemic lupus erythematosus (SLE), polymyalgia rheumatica (PMR), rheumatoid arthritis (RA), and inflammatory bowel disease (IBD), are related to the development of osteonecrosis (Migliaresi et al., 1994; Abu-Shakra et al., 2003; Boss et al., 2004; Tektonidou and Moutsopoulos, 2004; Lane, 2006). These diseases cause enough vasculature defects to cause a loss of microvascular flow to intraosseous tissue (Tektonidou and Moutsopoulos, 2004; Childs, 2005). Several studies reported that immunologic factors such as interleukins and TNFs might influence the development of osteonecrosis (Mont et al., 1998; Boss et al., 2004; Tektonidou and Moutsopoulos, 2004; Weitzmann and Pacifici, 2005; Sato and Takayanagi, 2006).

IL23 is a proinflammatory cytokine and consists of a p19 subunit and a p40 subunit of IL12. IL23R pairs with IL12R and both are necessary for IL23 signal pathway. IL23 regulates the activity of an immune response and promotes inflammation through the engagement of IL23R. These cytokines are primarily expressed in immune system cells such as a T cell, macrophage, and dendritic cell (Oppmann et al., 2000; Parham et al., 2002). Recently, it has been reported that IL23 deficient mice were resistant to collagen-induced arthritis (Murphy et al., 2003; Sanchez et al., 2007) and IL23 also inhibited osteoclastogenesis by RANKL (Receptor Activator for Nuclear Factor κB Ligand) through T cell activation (Kamiya et al., 2007).

Recently, many studies have reported that IL23R was related to inflammatory disorders. Several variations of the IL23R gene have been reported to influence the risk of a developing skin disorder called psoriasis (Capon et al., 2007; Cargill et al., 2007). Also, polymorphisms of the IL23R gene were associated with susceptibility to IBD and Crohn's disease (CD) in the genome wide association studies previously published (Cardon, 2006; Duerr et al., 2006; Raelson et al., 2007). In order to determine whether polymorphisms of the IL23R gene are associated with the susceptibility of ONFH, the genotype and allele frequencies between ONFH patients and controls were compared in this study.

Materials and Methods

Subjects

The study population was comprised of 443 unrelated patients (366 men, 77 women; mean age: 49.7 ± 13.3) with ONFH and 273 (206 men, 67 women; mean age: 52.1 ± 10.6) unrelated control subjects consecutively enrolled at the Kyungpook National University Hospital (Daegu, Korea) from 2002 to 2006. Patients were diagnosed and subgrouped by criteria which were described previously (Kim et al., 2007). Briefly, patients were subgrouped, according to etiological factors, into osteonecrosis, idiopathic (181 cases), steroid-induced (56 cases), and alcohol-induced (206 cases) groups. The control subjects were defined by the following criteria: those having no hip pain and others whose anteroposterior and frog leg lateral pelvic radiographs did not show any lesions, or with a sclerotic margin or subchondral collapse consistent with ONFH. We excluded all persons related to patients from our control group. This study was approved by the Institutional Review Board of Kyungpook National University Hospital, and all subjects gave written informed consent.

Selection of SNP and genotyping

Genomic DNA was isolated from peripheral blood of each individual by using a FlexiGene DNA Kit (QIAGEN, Valencia, CA). A total of 10 polymorphic sites regarding IL23R were selected by considering their location, allele frequencies and relevance to diseases based on public databases (dbSNP; http://www.ncbi.nlm.nih.gov/SNP/, HAPMAP; http://www.hapmap.org/index.html.en). The genotype identification was performed by using a Taqman probe (Applied Biosystems, Foster city, CA), according to the manufacturer's instructions. Primer Express (Applied Biosystems) was used in order to design both the PCR primers and the MGB TaqMan probes. One allelic probe was labeled with the FAM dye and the other was labeled with fluorescent VIC dye. Detailed procedures regarding the PCR reaction for Taqman assay have been described previously (Kim et al., 2006). Fluorescence data files from each plate were collected and analyzed by using automated allele-calling software (SDS 2.2, Applied Biosystems).

Statistical analyses

Chi-square tests were used in order to determine whether individual variants were in intra-locus equilibrium in the samples (Hardy-Weinberg equilibrium). Logistic regression analyses were used to calculate the odds ratios (OR), 95% confidence intervals (CI) and corresponding P values of each SNP and haplotypes controlling for age and sex as covariates with three alternative models (codominant, dominant and recessive). The linkage disequilibrium (LD) between loci was measured by using the absolute value of Lewontin's D' (|D'|) (Hedrick, 1987). Haplotype structures and their frequencies were estimated from genotyped data within the LD block by using the Haploview 3.32 (http://www.broad.mit.edu/mpg/haploview/), which estimates haplotype by an accelerated EM algorithm similar to the partition ligation method (Qin et al., 2002). All analyses were two-tailed and a P-value of < 0.05 was considered to be statistically significant. Statistical analyses were performed by using SAS 9.1 (SAS Institute Inc., Cary, NC).

Results

In order to investigate the association of IL23R gene polymorphisms with respect to ONFH, ten polymorphic sites in the IL23R gene were genotyped regarding 443 ONFH patients and 273 control subjects. Among the 10 SNPs genotyped, three SNP sites did not fulfill our criteria in a call rate (CR) > 0.90, minor allele frequency (MAF) > 0.05 and Hardy-Weinberg equilibrium (HWE) > 0.05. Seven SNP sites were in the HWE (P > 0.05) (Table 1). The P values of each polymorphism were obtained using logistic regression analyses by comparison between ONFH patients and the controls. It was found that rs4655686, rs1569922 and rs7539625 polymorphisms of the IL23R gene were significantly associated with the risk of ONFH in an alternative analysis model (P values; 0.0198-0.0447, OR; 1.30-1.49) (Table 2). These results suggested that the minor allele of rs4655686, rs1569922 and rs7539625 (A, T and G, respectively) contributes to an increase in the risk of ONFH.

Table 1 Frequencies of IL23R gene polymorphisms in ONFH patients and normal controls. aFrequencies of rare alleles. bP values of deviation from Hardy-Weinberg equilibrium among ONFH patients and control
Full size table
Table 2 Association analyses of IL23R gene polymorphisms with the risk of ONFH. Genotype distributions, odds ratio (OR; 95% confidence interval) and corresponding P values for logistic analyses of three alternative models and alleles are shown. Logistic regressions were applied in codominant (homozygote for major allele vs heterozygote vs homozygote for minor allele), dominant (homozygote for major allele vs heterozygote + homozygote for minor allele), recessive (homozygote for major allele + heterozygote vs homozygote for minor allele), and allele (major vs minor) models
Full size table

Further analysis based on pathological etiology (alcohol-, steroid- or idiopathic) showed that the genotypes of rs4655686, rs1569922, rs7539625 and rs7518660 of IL23R gene were specifically associated with the risk of ONFH in the idiopathic ONFH subgroup in all alternative models (P values; 0.0001-0.0150, OR; 1.45-2.17) (Table 3). These results suggested that the IL23R polymorphisms were possibly important risk factors in the idiopathic ONFH. However, there was no significant association with regard to alcohol- and steroid-induced subgroups (data not shown).

Table 3 Association of IL23R gene polymorphisms between idiopathic ONFH patients and controls
Full size table

Since the LD pattern has been hypothesized to be highly structured as conserved blocks of sequence separated by hotspots of recombination, the final function of a conserved haplotype may be the result of interaction among polymorphisms within the block. LD coefficients (|D'|) between all SNP pairs were calculated, and complete LD (i.e. D' = 1) were found between the rs10789229 and rs1008193 (Figure 1A). Additionally, we calculated the haplotype frequencies with two loci (rs10789229, rs1008193) within the LD block, and tested for a haplotype association between the controls and ONFH patients. Two major haplotypes among three haplotypes explained more than 98% of the variation (Figure 1B). As shown in Table 4, haplotypes were not associated with susceptibility to ONFH. In addition, haplotype analysis in subgroup patients also showed no association with the risk of ONFH (data not shown).

Figure 1
figure 1

Linkage disequilibrium coefficients and haplotypes of IL23R (A) Linkage disequilibrium coefficients (|D'|) and LD block among IL23R polymorphisms. (B) Haplotypes of IL23R

Full size image
Table 4 Association of IL23R gene haplotypes with ONFH patients and normal controls. Haplotypes: rs10789229 - rs1008193
Full size table

Discussion

ONFH is a devastating disease that frequently leads to a progressive collapse of the femoral head, followed by degenerative arthritis of the hip joint. The precise pathophysiology of ONFH is not known, but it has been suggested that a common pathogenesis of ONFH involves an interruption of the circulation of blood to the anterior-superior-lateral region of the femoral head (Atsumi and Kuroki, 1992). A previous study reported that idiopathic ONFH coincided with an increased release of a VEGF and proinflammatory factors (Boss et al., 2004). These factors include the recruitment of endothe lial progenitor cells, macrophages, osteoclasts, fibroblasts, and osteoblasts, which, pervading throughout the necrotic areas, initiate the reparative processes. The bone effect resulting from systemic inflammation may lead to a variety of bone diseases including RA, osteoporosis and osteonecrosis (Migliaresi et al., 1994; Abu-Shakra et al., 2003; Lane, 2006). According to a previous report (Tektonidou and Moutsopoulos, 2004), immunologic factors may be related with pathogenesis of osteonecrosis.

Cytokines are recognized as being involved in the bone metabolism process. IL3, IL6, and IL11 may act to increase the proliferation as well as the differentiation of osteoclast precursors (Goldring and Goldring, 1996; Mont et al., 1998). A previous study had suggested that IL23, an IL6/IL12 cytokine member, was related to collagen-induced arthritis in mouse models and also inhibited osteoclast formation (Murphy et al., 2003; Kamiya et al., 2007). IL4 and IFN-γ, induced by IL23, have been shown to block osteoclastogenesis by means of inhibition of the transcription factors NFATc1 and c-Fos, and induction of TNF receptor-associated factor 6 (TRAF6) (Mundy, 1996; Takayanagi et al., 2002; Moreno et al., 2003; Kamel Mohamed et al., 2005; Gao et al., 2007). Therefore, these cytokines may represent a major target for the prevention of inflammation related bone diseases and promote bone metabolism. A genome wide association study has shown that the IL23R gene was associated with autoimmune diseases, including IBD, CD and psoriasis (Cardon, 2006; Duerr et al., 2006; Raelson et al., 2007). The association between IL23R and these diseases has been replicated in many other studies with different populations (Cargill et al., 2007; Cummings et al., 2007; Dubinsky et al., 2007; Glas et al., 2007; Rioux et al., 2007; Roberts et al., 2007). The recent findings of the association studies suggested that the IL-23/IL-23R cytokine receptor system is also involved in the pathogenesis of several diseases.

In this study, we have tried to determine the contribution of IL23R gene polymorphisms in regards to ONFH. We found that polymorphisms of the IL23R gene were significantly associated with ONFH in the Korean population. In particular, the genotypes of rs4655686, rs1569922, rs7539625 and rs7518660 of the IL23R gene were specifically associated with the risk of ONFH in the idiopathic subgroup (Table 3). These results suggested that inflammatory factors might be one of the important risk factors in regards to idiopathic ONFH. However, there were no significant associations shown in both alcohol- and steroid-induced subgroups. No association with ONFH in these subgroups could be attributable to different etiological characteristics of subgroup patients. Corticosteroid therapy and alcoholism, which can change fat metabolism, and accumulating cell stress, which causes occlusion of the blood flow, were major risk factors with respect to osteonecrosis (Gold and Cangemi, 1979; Asano et al., 2003). In addition, alcohol and steroids promote the generation of reactive oxygen species (ROS), resulting in oxidative stress. Therefore, it was generally thought that the effect of alcohol abuse or excessive steroid administration was stronger than genetic factors in regards to alcohol- and steroid-induced ONFH patients. On the other hand, genetic factors may be more strongly related to ONFH among idiopathic patients due to no obvious reasons having been identified.

Our study had several concerns. Firstly, since our hospital is a tertiary institution covering the Daegu-Kyungpook region, more severely ill patients are usually referred to our hospital, thus possibly resulting in selection bias. Secondly, patients in the early stage of osteonecrosis might have been included in the control group since only patients were identified by a MRI and we did not check the MRI results of the control group. The numbers of controls were less than numbers of cases since the age and sex-matched normal controls were difficult to obtain. Thirdly, the incidence of ONFH in Korea is relatively higher and sex-biased (male:female = 4:1) in comparison to other populations. The majority of osteonecrosis in Korea is either idiopathic or related to excessive alcohol use. Most male patients represent alcohol-induced cases, whereas female patients represent idiopathic or steroid-induced cases. These facts suggested that the etiology of ONFH may be quite complex and possibly attributable to gender and ethnic differences.

ONFH is one of the most common diseases of the hip in Korea, responsible for more than half of the underlying causes of total hip arthroplasty. Therefore, early diagnosis of the disease using several noted genetic markers will provide beneficial information regarding individual susceptibility to ONFH and may also help high-risk individuals to take precautions against further aggravating symptoms. In conclusion, the present study suggests, for the first time, to our knowledge, that the polymorphisms of the IL23R gene are likely to be associated with the risk of ONFH, at least in the Korean population. Further replication studies in a large cohort would be needed to confirm the suggested association.