This hospital-based case–control study was conducted in the Department of Orthopaedic Surgery, King George’s Medical University (KGMU), Lucknow, Uttar Pradesh, India. A total of 450 consecutive patients diagnosed with knee OA between March 2010 and May 2015 were evaluated for inclusion.

The radiographic evaluation was performed on anteroposterior standing and lateral X-Ray views of the knee by a single investigator using the Kellgren–Lawrence (KL) score ranging between 0 and 4. OA were classified into mild (K-L grade 2), moderate (K-L grade 3) severe (K-L grade 4) [22]. Only patients with radiographic OA, defined as KL score of ≥2, were included in the study. Persons suffering from other knee joint ailments like rheumatoid arthritis, gouty arthritis, septic arthritis, post traumatic or dysplasias were excluded.

A total of 458 age and sex-matched healthy volunteers were recruited from the same hospital during the same period. Selection criteria for the controls included no history of OA. Age, sex, weight, height, and Body Mass Index (BMI) were recorded for all study participants. Written informed consent was provided by all participants, and the study protocol was approved by the Ethics Committee of the KGMU, Lucknow, Uttar Pradesh, India.

Association of genotypes (rs12901499 and rs921126) of SMAD3 and BMP5 gene with OA knee was determined using standard chi-square test. Frequencies for the rs12901499 and rs921126 polymorphisms in the controls agreed with that expected according to the Hardy–Weinberg principle. Associations between the SMAD3 or BMP5 variants and OA risk were estimated by calculating the odds ratios (ORs) and 95% confidence intervals (CIs) using both univariate and multivariate logistic regression analyses with adjustments for age, sex, and BMI. Two sided tests were used for statistical analyses, and a p-value < 0.05 was considered to be statistically significant for comparison of continuous data of general characteristic of cases and controls, unpaired student t-test was performed. The statistical power of the study was calculated using formula “n=(r + 1/r) p (1-ɸ) (Zβ +Zα/2)2/ (p1 – p2)2” with proportion of Allele (A) exposed in controls (37%) and considering to detect Odd ratio of 2.0 and r =1. the calculated sample size was 150 in each case and controls. But, for increasing power of study we recruited 450 cases and 458 controls.

All statistical analysis was performed with the SPSS software package (version 16.0 for windows; SPSS Chicago, IL).

DNA isolation and genotype analysis: Venous blood samples were obtained in ethylene diamine tetra-acetic acid anticoagulant (20g/l) for genomic DNA extraction. Samples were stored at -80 ºC until analysis. Isolated DNA was subsequently used for genotyping.

Detection of rs921126 and rs12901499 polymorphism of BMP5 and SMAD3 gene: Reaction was performed with 50–200 ng of genomic DNA, 0.2 μmol of each primer, 200 μmol of dNTP (Fermentas, USA), 1.2 mmol of MgCl₂, 1 unit of Taq polymerase (Fermentas, USA) and sterile MilliQ water in a total volume of 20 μl. Amplification was performed on GeneAmp PCR (Polymerase Chain Reaction) system 9700 (Applied Bio System) using the following specific PCR conditions for each polymorphism.

rs12901499 polymorphism: Initial denaturation at 94°C for 5 minutes followed by 35 cycles of denaturation at 94°C for 45 seconds, annealing at 58.6°C for 45 seconds and extension at 72°C for one minute. It was further followed by final extension at 72°C for 10 minutes. PCR reaction resulted in a 226 bp product. PCR products (10 μl) were digested with 10 U of MboII restriction enzyme (MBI Fermentas, Germany) to identify the presence of polymorphic sites in SMAD3 gene. Digestion of 226bp PCR product of SMAD3 gene with MboII restriction enzyme into two fragments of 176 bp and 50 bp indicated the presence of GG genotype and the presence of fragments of three sizes (226 bp, 176 bp and 50 bp) was indicative of AG genotype while the undigested 226 bp PCR fragment was indicative of AA genotype.

rs921126 polymorphism: Initial denaturation at 94°C for five minutes followed by 35 cycles of denaturation at 94°C for 45 seconds, annealing at 58°C or 45 seconds and extension at 72°C for one minute. It was further followed by final extension at 72°C for 10 minutes. PCR reaction resulted in a 794 bp product. PCR products (10 μl) were digested with 10 U of Dra1 restriction enzyme (MBI Fermentas, Germany) to identify the presence of polymorphic sites in BMP5 gene. Digestion of 794 bp PCR product of BMP5 gene with Dra1 restriction enzyme into two fragments of 424 bp and 360 bp indicated the presence of AA genotype and the presence of fragments of three sizes (794 bp, 424 bp and 360 bp) was indicative of the AG genotype while the undigested 794 bp PCR fragment was indicative of GG genotype.

A total of 958 subjects were evaluated. Of these, 450 had radiographic knee OA. In addition, 458 age- and sex- matched healthy controls were recruited. There were no significant differences in demographic characteristics between the two groups [Table/Fig-1].

The observed genotype frequencies for the rs12901499 in SMAD3 and rs921126 in BMP5 polymorphisms in the controls agreed with that expected according to the Hardy–Weinberg principle [23] (data not shown). The results of repeat genotyping of randomly selected samples were 100% concordant.

The genotype and allele distributions of the rs12901499 and rs921126 polymorphisms in patients with knee OA and healthy controls are shown in [Table/Fig-2]. The genotype distribution for rs12901499 andrs921126 was significantly different between the two groups (p<0.03 and p<0.001). After adjustment for age, sex, and BMI, a significantly increased risk of knee OA was associated with the genotype GG and GA of rs12901499 and rs921126 compared with the AA genotype. In addition, those bearing at least one G allele (GG + GA) had a significantly increased risk of knee OA compared with those without the G allele (AA) in rs921126. However, in rs12901499 were not significantly associated with the risk of knee OA.

When stratified by age, both young (<55 years) and old (>55 years) patients showed significant differences in genotype frequencies compared with controls (p<0.001 and p<0.0001) [Table/Fig-3].

Stratification analysis was performed to evaluate the potential association of genetic variants of SMAD3 rs12901499 and BMP5 rs921126 with knee OA risk in subgroups based on demographic characteristics. When stratified by age, both young (≤55 years) and old (>55 years) patients showed significant differences in genotype frequencies compared with controls (p<0.0001 and <0.04) and similarly in male and females (p<0.0001 and <0.03) [Table/Fig-3]. Likewise, in polymorphism rs9382564 both young (≤55 years) and old (>55 years) patients showed significant differences in genotype frequencies compared with controls (p<0.01 and <0.0001) and similarly in male and females (p<0.0001 and <0.0008) [Table/Fig-4].

When the association between G allele carriers and the risk of OA was evaluated using logistic regression analysis, GG homozygotes carried a 1.3 and 1.5 fold risks in the young and old patients group respectively. Likewise 1.3and 1.5-fold increased risk of OA compared with AA homozygotes in male and female group [Table/Fig-3].

Similarly, in the [Table/Fig-4], GG homozygotes carried a 1.91 and 3.4 fold increased risk of OA compared with AA homozygotes in the both age group and 2.93 and 2.25 fold increased in both gender groups. In addition, GA heterozygote carried a 1.6- and 1.7-fold increased risk of knee OA in young and old patients, respectively, compared with AA homozygotes. Likewise 1.59 and 1.76 fold increased risk of OA in male and female group [Table/Fig-4].

In this case control study, SNP rs921126 and rs12901499 were found associated with Knee OA. The result showed that SNP rs921126 polymorphic both (GG and GA) genotype of BMP5 and variant G may contribute to the risk of knee OA and that this risk increased with age (>55) and in both gender.

To the best of our knowledge, this is the first report linking BMP SNP rs921126 with knee OA in North Indian population and first report with age, gender stratified association in any population. Further on this age, gender stratification, the association between risk of knee osteoarthritis and SNP rs921126, GG genotype compared with AA genotype was stronger in both genders and in those >55-year-old.

Many evidences are documented between OA pathogenesis and TGF-b signaling pathways [24,25]. BMP5 is involved in the maintenance of synovial joint development and tissue homeostasis [21]. The SNPs in BMP5 have been well documented. It was reported that SNP 921126 polymorphism in BMP5 gene was associated with female hip OA [16]. In our case-control study we found that SNP 921126 was associated with Knee OA. Our results support the hypothesis that the SNP 921126 of BMP5 gene might be a risk of OA. In the present study, the individuals carrying BMP5 SNP 921126, the GG and GA or GG + GA genotypes had a higher risk of knee OA than those carrying the AA genotype. The G allele of BMP5 SNP 921126 might be associated with the development of knee OA. This result is in consonant with a study by Wilkins JM et al., on female hip osteoarthritis [16].

Similarly, our study shows that SMAD3 rs12901499 genetic variants, GG and GA genotype, were significantly associated with knee OA and variant G may contribute to the risk of knee OA. This result is similar to a study by Valdes AM et al., on knee and hip OA and in the Northeastern Chinese population between knee OA and SMAD3 polymorphisms (rs12901499A/G and rs6494629T/C) [13], However Su SL et al., found no such associations [14]. Wu Q et al., suggested that Loss of SMAD3 appears to enhance bone morphogenetic protein signaling in the articular chondrocytes, leading to hypertrophy and OA-like changes [26]. In addition Cherlet T et al., reported that SMAD3 levels are lower in women than in men, which is consistent with other data showing that estrogens inhibit SMAD3 transcriptional activity [27]. In addition, Valdes AM et al., found a significant association of SMAD3 intronic SNP with OA in both the genders [13]. Remarkable similarity in effect sizes between male and female subjects strengthened this finding of their study. This is consistent to our study in which gender stratified analysis found that there was an association between risk of knee osteoarthritis and SMAD3 SNP rs12901499, but none of the genotypes were strongly associated. However, the association of this polymorphism with age was also found and this finding is being reported for the first time.

Our study has certain limitations like a moderate sample size of 450 cases and 458 controls; therefore the results should be confirmed with larger sample size. Secondly only two SMAD3 and BMP5 SNPs were investigated; SNPs at other loci may also be associated with susceptibility to OA in TGF-b signaling pathway.

The present study was the first to show that the genotype distribution of the BMP5 rs921126 was significantly associated with the risk of knee OA. Further, it validates the significant association of the SMAD3 rs12901499 polymorphisms in our population. Following validation the effect of levels or functional roles of the rs921126 and rs12901499 polymorphism may help us in determining the aetiology of knee OA and thereby in formulating further research in its prevention and management.