Abstract

A compelling genetic association with osteoarthritis (OA) of a functional SNP (rs143383, T/C) in the 5′-UTR of the GDF5 gene was recently reported in case–control cohorts from Japan and China. GDF5 is a pro-chondrogenic growth factor. The T-allele frequency of the gene was elevated in cases, with an odds ratio (OR) of 1.79, and in vitro functional studies demonstrated that this allele mediated a moderate but significant reduction in the activity of the GDF5 promoter in several cell lines. Our initial objective was to assess whether the SNP was also associated with OA in a broad European population by genotyping the SNP in 2487 cases and 2018 age-matched controls from the UK and Spain. The T-allele was associated with OA (P = 0.03, OR = 1.10) as was carrier status for this allele (P = 0.004, OR = 1.28), demonstrating that the SNP is associated with OA in two diverse ethnic groups, Asians and Europeans. We subsequently assessed the functional effect of the SNP on GDF5 allelic expression using RNA extracted from the cartilage of OA patients who had undergone joint-replacement surgery. The associated T-allele showed up to a 27% reduction in expression relative to the C-allele (P = 0.00007), revealing that the functional effect mediated by SNP rs143383 on GDF5 expression is active in patients who have severe disease up to the point at which they require surgery. A small but persistent imbalance of GDF5 expression throughout life therefore appears to render an individual more susceptible to OA.

INTRODUCTION

Osteoarthritis (OA) is a common multifactorial disease characterized by the degeneration of the protective cartilage layer in articulating synovial joints. It is largely a disease associated with ageing, being rare in individuals below the age of 45 years with an increase in incidence and prevalence up to 80 years of age. A common treatment for patients who have severe disease of the hip or the knee is total joint replacement surgery (THR and TKR, respectively). Growth and differentiation factor 5 (also known as cartilage-derived morphogenetic protein 1) is a member of the transforming growth factor-β (TGF-β) superfamily and participates in the development, maintenance and repair of bone, cartilage and other soft tissues of the synovial joint (1–5). An association with hip and knee OA of a single nucleotide polymorphism (SNP) (rs143383, T/C) located in the 5′-UTR of the growth and differentiation factor 5 gene, GDF5, was recently reported in Japanese and in Chinese case–control cohorts (6). The major, T-allele, of the SNP was common in the Asian populations, with frequencies > 70% in controls, and was at an elevated frequency in OA cases, with odds ratios (ORs) of 1.79 in the Japanese hip cases (n = 998), 1.30 in the Japanese knee cases (n = 718) and 1.54 in the Chinese knee cases (n = 313). In vitro cell transfection studies using luciferase reporter assays revealed that the T-allele-mediated a moderate but significant reduction in the activity of the GDF5 promoter in chondrogenic and non-chondrogenic cell lines. Our initial objective was to assess the role of this SNP on OA development in Europeans by case–control association analysis. Since the SNP was associated in both Japanese and Chinese, representing a relatively broad group of ethnic Asians, we hypothesized that if the SNP were associated in Europeans then it would be associated in a broad rather than a narrow European group. We therefore chose to study North European (UK) and South European (Spanish) populations in a combined analysis of 4505 cases and controls. We subsequently assessed whether the SNP influenced allelic expression of GDF5 in vivo by extracting the RNA from the cartilage of nine OA patients who had undergone a THR or a TKR. This RNA was then used to measure GDF5 allelic expression by a single base-pair extension assay that can discriminate and quantify the mRNA synthesized from each allele.

RESULTS

Power of the association study

We determined that the combined cohort (n = 4505), the UK cohort alone (n = 2491) and the Spanish cohort alone (n = 2014) were all adequately powered, at a significance level of 5%, to detect an OR comparable to those reported in the Asian study (data not shown). There was no significant difference in genotype or allele frequencies when the UK controls were compared with the Spanish controls, with a genotype P of 0.46 and an allele P of 0.25, implying that for this SNP the two populations are equivalent and can therefore be combined.

Single nucleotide polymorphism rs143383 is associated with osteoarthritis in Europeans

In the combined analysis, the SNP demonstrated association at both the genotypic (P = 0.01) and the allelic [P = 0.03, OR = 1.10 (95% CI 1.01–1.20)] level (Table 1). This association involved an elevation of the T-allele, in line with the Asian study. The significances of these associations did not increase following stratification by sex or by site of disease (hip OA, knee OA or hand OA), implying that a particular stratum does not drive them. Carriers of the T-allele [(TT)+(TC)] were also significantly more common in cases than in controls (Table 2), with a P of 0.004 and an OR of 1.28 (95% CI 1.08–1.51). In the T-allele carrier analysis, the knee stratum, the hip stratum and the hip and knee stratum (cases who had undergone joint replacement at both a hip and a knee) were all significant (P ≤ 0.05) and all had positive ORs whose 95% CI did not include 1, whereas the hand stratum was not significant (P = 0.86) and had an OR whose 95% CI did encompass 1 (Fig. 1). This implies that the rs143383 T-allele carrier association is restricted in Europeans to larger, weight-bearing synovial joints such as those of the hip and knee. It is noteworthy that the highest frequency of T-allele carriage is in the hip and knee stratum, with a carrier frequency of 91.9% in hip and knee cases versus 84.1% in controls [Table 2; P = 0.04; OR = 2.14 (95% CI 1.03–4.46)], implying that SNP rs143383 is a particular risk factor for OA development at multiple large joints in an individual.

Figure 1.

Odds ratios (OR) and 95% CI (bar) for carriers of the T-allele of GDF5 SNP rs143383. TKR, knee cases (UK and Spanish data combined), THR, hip cases (UK and Spanish data combined), K + H, cases who have undergone a total knee and a total hip replacement (UK data only), HOA, cases with hand OA (Spanish data only). Also shown are the ORs for the UK cohort alone and for the Spanish cohort alone, for males and for females from the UK and Spanish data combined, and finally for the combined unstratified study.

Figure 1.

Odds ratios (OR) and 95% CI (bar) for carriers of the T-allele of GDF5 SNP rs143383. TKR, knee cases (UK and Spanish data combined), THR, hip cases (UK and Spanish data combined), K + H, cases who have undergone a total knee and a total hip replacement (UK data only), HOA, cases with hand OA (Spanish data only). Also shown are the ORs for the UK cohort alone and for the Spanish cohort alone, for males and for females from the UK and Spanish data combined, and finally for the combined unstratified study.

Table 1.

Genotype and allele association analysis of SNP rs143383 in cases with osteoarthritis and controls

Group  Genotype P-value Allele P-value 
  (CC) (CT) (TT) (C) (T) 
All cases Count 319 1194 974 0.01 1832 3142 0.03 
(n = 2487) Frequency (%) 12.8 48.0 39.2  36.8 63.2  
All controls Count 320 935 763  1575 2461  
(n = 2018) Frequency (%) 15.9 46.3 37.8  39.0 61.0  
Female cases Count 207 785 631 0.29 1199 2047 0.20 
(n = 1623) Frequency (%) 12.8 48.4 38.9  36.9 63.1  
Female controls Count 154 493 388  801 1269  
(n = 1035) Frequency (%) 14.9 47.6 37.5  38.7 61.3  
Male cases Count 111 409 342 0.05 631 1093 0.08 
(n = 862) Frequency (%) 12.9 47.4 39.7  36.6 63.4  
Male controls Count 166 442 375  774 1192  
(n = 983) Frequency (%) 16.9 45.0 38.1  39.4 60.6  
All knees Count 76 304 243 0.08 456 790 0.12 
(n = 623) Frequency (%) 12.2 48.8 39.0  36.6 63.4  
Female knees Count 52 193 157 0.62 297 507 0.38 
(n = 402) Frequency (%) 12.9 48.0 39.1  36.9 63.1  
Male knees Count 24 111 86 0.07 159 283 0.19 
(n = 221) Frequency (%) 10.9 50.2 38.9  36.0 64.0  
All hips Count 198 728 599 0.06 1124 1926 0.06 
(n = 1525) Frequency (%) 13.0 47.7 39.3  36.9 63.1  
Female hips Count 118 461 369 0.29 697 1199 0.21 
(n = 948) Frequency (%) 12.4 48.6 38.9  36.8 63.2  
Male hips Count 80 267 230 0.28 427 727 0.19 
(n = 577) Frequency (%) 13.9 46.3 39.9  37.0 63.0  
All hips and kneesa Count 57 34 0.04 73 125 0.54 
(n = 99) Frequency (%) 8.1 57.6 34.3  36.9 63.1  
Female hips and knees Count 37 21 0.09 45 79 0.59 
(n = 62) Frequency (%) 6.5 59.7 33.9  36.3 63.7  
Male hips and knees Count 20 13 0.47 28 46 0.79 
(n = 37) Frequency (%) 10.8 54.1 35.1  37.8 62.2  
All handsb Count 37 105 98 0.65 179 301 0.46 
(n = 240) Frequency (%) 15.4 43.7 40.8  37.3 62.7  
Female handsb Count 33 94 84 0.72 160 262 0.76 
(n = 211) Frequency (%) 15.6 44.5 39.8  37.9 62.1  
Male handsb Count 11 13 0.52 17 37 0.24 
(n = 27) Frequency (%) 11.1 40.7 48.1  31.5 68.5  
Group  Genotype P-value Allele P-value 
  (CC) (CT) (TT) (C) (T) 
All cases Count 319 1194 974 0.01 1832 3142 0.03 
(n = 2487) Frequency (%) 12.8 48.0 39.2  36.8 63.2  
All controls Count 320 935 763  1575 2461  
(n = 2018) Frequency (%) 15.9 46.3 37.8  39.0 61.0  
Female cases Count 207 785 631 0.29 1199 2047 0.20 
(n = 1623) Frequency (%) 12.8 48.4 38.9  36.9 63.1  
Female controls Count 154 493 388  801 1269  
(n = 1035) Frequency (%) 14.9 47.6 37.5  38.7 61.3  
Male cases Count 111 409 342 0.05 631 1093 0.08 
(n = 862) Frequency (%) 12.9 47.4 39.7  36.6 63.4  
Male controls Count 166 442 375  774 1192  
(n = 983) Frequency (%) 16.9 45.0 38.1  39.4 60.6  
All knees Count 76 304 243 0.08 456 790 0.12 
(n = 623) Frequency (%) 12.2 48.8 39.0  36.6 63.4  
Female knees Count 52 193 157 0.62 297 507 0.38 
(n = 402) Frequency (%) 12.9 48.0 39.1  36.9 63.1  
Male knees Count 24 111 86 0.07 159 283 0.19 
(n = 221) Frequency (%) 10.9 50.2 38.9  36.0 64.0  
All hips Count 198 728 599 0.06 1124 1926 0.06 
(n = 1525) Frequency (%) 13.0 47.7 39.3  36.9 63.1  
Female hips Count 118 461 369 0.29 697 1199 0.21 
(n = 948) Frequency (%) 12.4 48.6 38.9  36.8 63.2  
Male hips Count 80 267 230 0.28 427 727 0.19 
(n = 577) Frequency (%) 13.9 46.3 39.9  37.0 63.0  
All hips and kneesa Count 57 34 0.04 73 125 0.54 
(n = 99) Frequency (%) 8.1 57.6 34.3  36.9 63.1  
Female hips and knees Count 37 21 0.09 45 79 0.59 
(n = 62) Frequency (%) 6.5 59.7 33.9  36.3 63.7  
Male hips and knees Count 20 13 0.47 28 46 0.79 
(n = 37) Frequency (%) 10.8 54.1 35.1  37.8 62.2  
All handsb Count 37 105 98 0.65 179 301 0.46 
(n = 240) Frequency (%) 15.4 43.7 40.8  37.3 62.7  
Female handsb Count 33 94 84 0.72 160 262 0.76 
(n = 211) Frequency (%) 15.6 44.5 39.8  37.9 62.1  
Male handsb Count 11 13 0.52 17 37 0.24 
(n = 27) Frequency (%) 11.1 40.7 48.1  31.5 68.5  

aCases who had undergone both a hip and a knee joint replacement.

bTwo hand osteoarthritis cases had unrecorded sex status.

Table 2.

Association analysis of carriers of the T-allele of SNP rs143383 in cases with osteoarthritis and controls

Group  Carriers of the T-allele (TT)+(TC) Non-carriers (CC) P OR (95% CI) 
All cases Count 2168 319 0.004 1.28 
(n = 2487) Frequency (%) 87.2 12.8  (1.08–1.51) 
All controls Count 1698 320   
(n = 2018) Frequency (%) 84.1 15.9   
Female cases Count 1416 207 0.12 1.20 
(n = 1623) Frequency (%) 87.2 12.8  (0.95–1.50) 
Female controls Count 881 154   
(n = 1035) Frequency (%) 85.1 14.9   
Male cases Count 751 111 0.02 1.37 
(n = 862) Frequency (%) 87.1 12.9  (1.06–1.78) 
Male controls Count 817 166   
(n = 983) Frequency (%) 83.1 16.9   
All knees Count 547 76 0.03 1.36 
(n = 623) Frequency (%) 87.8 12.2  (1.04–1.77) 
Female knees Count 350 52 0.35 1.18 
(n = 402) Frequency (%) 87.1 12.9  (0.84–1.65) 
Male knees Count 197 24 0.03 1.67 
(n = 221) Frequency (%) 89.1 10.9  (1.06–2.63) 
All hips Count 1327 198 0.02 1.26 
(n = 1525) Frequency (%) 87.0 13.0  (1.04–1.53) 
Female hips Count 830 118 0.12 1.23 
(n = 948) Frequency (%) 87.6 12.4  (0.95–1.59) 
Male hips Count 497 80 0.11 1.26 
(n = 577) Frequency (%) 86.1 13.9  (0.95–1.69) 
All hips and kneesa Count 91 0.04 2.14 
(n = 99) Frequency (%) 91.9 8.1  (1.03–4.46) 
Female hips and knees Count 58 0.07 2.53 
(n = 62) Frequency (%) 93.5 6.5  (0.91–7.08) 
Male hips and knees Count 33 0.33 1.68 
(n = 37) Frequency (%) 89.2 10.8  (0.59–4.80) 
All handsb Count 203 37 0.86 1.03 
(n = 240) Frequency (%) 84.6 15.4  (0.71–1.48) 
Female handsb Count 178 33 0.78 0.94 
(n = 211) Frequency (%) 84.4 15.6  (0.63–1.42) 
Male handsb Count 24 0.43 1.63 
(n = 27) Frequency (%) 88.9 11.1  (0.48–5.46) 
Group  Carriers of the T-allele (TT)+(TC) Non-carriers (CC) P OR (95% CI) 
All cases Count 2168 319 0.004 1.28 
(n = 2487) Frequency (%) 87.2 12.8  (1.08–1.51) 
All controls Count 1698 320   
(n = 2018) Frequency (%) 84.1 15.9   
Female cases Count 1416 207 0.12 1.20 
(n = 1623) Frequency (%) 87.2 12.8  (0.95–1.50) 
Female controls Count 881 154   
(n = 1035) Frequency (%) 85.1 14.9   
Male cases Count 751 111 0.02 1.37 
(n = 862) Frequency (%) 87.1 12.9  (1.06–1.78) 
Male controls Count 817 166   
(n = 983) Frequency (%) 83.1 16.9   
All knees Count 547 76 0.03 1.36 
(n = 623) Frequency (%) 87.8 12.2  (1.04–1.77) 
Female knees Count 350 52 0.35 1.18 
(n = 402) Frequency (%) 87.1 12.9  (0.84–1.65) 
Male knees Count 197 24 0.03 1.67 
(n = 221) Frequency (%) 89.1 10.9  (1.06–2.63) 
All hips Count 1327 198 0.02 1.26 
(n = 1525) Frequency (%) 87.0 13.0  (1.04–1.53) 
Female hips Count 830 118 0.12 1.23 
(n = 948) Frequency (%) 87.6 12.4  (0.95–1.59) 
Male hips Count 497 80 0.11 1.26 
(n = 577) Frequency (%) 86.1 13.9  (0.95–1.69) 
All hips and kneesa Count 91 0.04 2.14 
(n = 99) Frequency (%) 91.9 8.1  (1.03–4.46) 
Female hips and knees Count 58 0.07 2.53 
(n = 62) Frequency (%) 93.5 6.5  (0.91–7.08) 
Male hips and knees Count 33 0.33 1.68 
(n = 37) Frequency (%) 89.2 10.8  (0.59–4.80) 
All handsb Count 203 37 0.86 1.03 
(n = 240) Frequency (%) 84.6 15.4  (0.71–1.48) 
Female handsb Count 178 33 0.78 0.94 
(n = 211) Frequency (%) 84.4 15.6  (0.63–1.42) 
Male handsb Count 24 0.43 1.63 
(n = 27) Frequency (%) 88.9 11.1  (0.48–5.46) 

Data is presented un-stratified and following stratification by sex and by joint.

aCases who had undergone both a hip and a knee joint replacement.

bTwo hand osteoarthritis cases had unrecorded sex status.

When we studied the UK case–control cohort alone, we did not observe a significant difference in genotype or allele frequencies for the un-stratified analysis between cases and controls, although the UK sample did approach significance for the genotype comparison, with a P of 0.07, and was significant for the stratum of cases who had undergone both a hip and a knee joint replacement, with a genotype P of 0.04 (Supplementary Material, Table S1). Carriers of the T-allele were also more common in the UK cases than in the UK controls [P = 0.02; OR = 1.32 (95% CI 1.04–1.68); Supplementary Material, Table S3). As mentioned above, in the combined analysis, the hip and knee cases showed the highest frequency of T-allele carriage. All of these cases were from the UK cohort. When these hip and knee cases were compared with only UK controls, the difference in T-allele carriage remained significant [P = 0.05; OR = 2.06 (95% CI 0.98–4.35)] implying that this association is not an artefact of examining two combined populations (UK and Spain). For the Spanish case–control cohort alone, there were no significant differences between the cases and controls (Supplementary Material, Table S2), although there was a non-significant increase in the frequency of carriers of the T-allele in the cases (Supplementary Material, Table S3).

Significantly less GDF5 transcript is produced by the T-allele of SNP rs143383 in the cartilage of osteoarthritis patients

We next set out to assess whether the 5′-UTR SNP correlated with differences in allelic expression using RNA extracted from the articular cartilage of nine OA patients who were heterozygous for the SNP and who had undergone either a THR or a TKR. These nine individuals had been collected subsequent to our association analysis. All nine patients demonstrated a reduction in the expression of the T-allele relative to the C-allele, with allelic ratios below 1.0, and for six of the nine patients, the reductions were significant (P < 0.005, two-tailed Mann–Whitney exact test, Fig. 2A). The greatest allelic difference was for patient 7, a male who had undergone a TKR at the age of 66 years. His GDF5 T-allele showed a 27% reduction in expression relative to his GDF5 C-allele (P = 0.00007). When the data for all nine patients was studied together, the T-allele demonstrated an average 12% reduction in relative expression (P = 0.006, two-tailed t-test, Fig. 2B).

Figure 2.

Allelic expression analysis at GDF5 SNP rs143383 using RNA extracted from articular cartilage. (A) The expression analysis was performed on the cartilage of nine UK patients who had severe end-stage OA, seven of whom had undergone a TKR (K) and two of whom had undergone a THR (H). For each patient, 19 (patients 1, 4, 5 and 7) or 20 (patients 2, 3, 6, 8 and 9) individual cDNA PCR amplifications and SBE extension reactions were performed. Forty-five individual PCR and SBE reactions were performed on genomic DNA (five reactions per patient). The cDNA allelic ratios from each patient were then compared with the 45 genomic allelic ratios by a two-tailed Mann–Whitney exact test. Data shown are the mean + SD. *P < 0.005; **P < 0.0005; ***P < 0.0001. (B) The mean allelic ratios for the nine patient cDNAs were compared with the mean allelic ratios for the nine patient genomic DNAs using a two-tailed t-test. Data shown are the mean + SD.

Figure 2.

Allelic expression analysis at GDF5 SNP rs143383 using RNA extracted from articular cartilage. (A) The expression analysis was performed on the cartilage of nine UK patients who had severe end-stage OA, seven of whom had undergone a TKR (K) and two of whom had undergone a THR (H). For each patient, 19 (patients 1, 4, 5 and 7) or 20 (patients 2, 3, 6, 8 and 9) individual cDNA PCR amplifications and SBE extension reactions were performed. Forty-five individual PCR and SBE reactions were performed on genomic DNA (five reactions per patient). The cDNA allelic ratios from each patient were then compared with the 45 genomic allelic ratios by a two-tailed Mann–Whitney exact test. Data shown are the mean + SD. *P < 0.005; **P < 0.0005; ***P < 0.0001. (B) The mean allelic ratios for the nine patient cDNAs were compared with the mean allelic ratios for the nine patient genomic DNAs using a two-tailed t-test. Data shown are the mean + SD.

DISCUSSION

Using a large cohort of 2487 OA cases and 2018 controls, we have replicated in Europeans the association of the GDF5 5′-UTR SNP rs143383 that is reported as associated to OA in Asians (6). We hypothesized that since rs143383 showed association in a relatively broad group of ethnic Asians, then any association in Europeans would also be ethnically broad, that is it would be unlikely to be associated in Asians and to only a specific ethnic subgroup of Europeans. We therefore studied Europeans from the North and South of the continent. There were no significant differences in genotype or allele frequencies between the UK controls and the Spanish controls, implying that for this SNP the two populations are equivalent and can therefore be combined. In our combined study we detected association at the genotype level, at the allele level and when carriers of the T-allele were combined [(TT)+(TC)]; this latter result implying that the allele has a dominant effect on OA susceptibility. Association was detected in the UK cohort alone but not in the Spanish cohort alone, although trends towards association were evident in the Spanish cohort. The UK cohort is larger than the Spanish cohort (2491 individuals versus 2014 individuals) and does therefore have more power to detect an association.

The unstratified ORs detected in our study are not as large as those reported in the Asian study. We must conclude therefore that the GDF5 SNP rs143383 is not as major a susceptibility locus in Europeans as it is in Asians. This may reflect differences in the genetic background or in non-genetic, environmental influences between Europeans and Asians. Other loci that are risk factors for OA development in Asians have been studied in Europeans and their effects have not been replicated to the same degree as in the primary report (7–14). An example of this is the asporin gene ASPN, which harbours an aspartic-acid repeat polymorphism that has a highly significant effect upon OA susceptibility in Asians, but which shows only a moderate association in Europeans (9–11,14–16).

Growth and differentiation factor 5 is a member of the TGF-β superfamily of signalling molecules. Post-development, this superfamily facilitates synovial joint homeostasis in a pro-anabolic manner (2). Disruption to the anabolic–catabolic balance is often cited as a likely mechanism that influences OA development (17). Our in vivo allelic expression data demonstrates that the GDF5 gene is also expressed in the cartilage of elderly adults and that the functional difference mediated by SNP rs143383 on GDF5 expression that was intimated by the in vitro reporter experimemt carried out in the Asian study (6) is active in patients who have severe end-stage disease up to the point at which they require surgery. It could be therefore that a small but persistent imbalance of GDF5 expression throughout life renders an individual more susceptible to OA. This is potentially a very important observation as it suggests that a therapy designed to counteract the functional susceptibility of GDF5 may be beneficial even when administered to mature individuals.

A number of other genes encoding proteins mediating cell signalling and signal transduction have been implicated in OA susceptibility (18). GDF5 joins this growing group. Manipulation of these genes and their encoded proteins, even in adult tissue, may be a fertile area for the development of novel OA therapeutics.

MATERIALS AND METHODS

UK case–control cohort

Cases were ascertained through the Nuffield Orthopaedic Centre in Oxford (n = 1669; 1006 females and 663 males). They had undergone total joint replacement of a hip (THR, n = 1221; 765 females and 456 males), of a knee (TKR, n = 349; 179 females and 170 males) or of a hip and a knee (n = 99; 62 females and 37 males) for primary OA. The cases were ascertained using the criteria of signs and symptoms of OA sufficiently severe to require joint replacement surgery. The radiological stage of the disease was a Kellgren and Lawrence grade of 2 or more in all cases with over 90% being grade 3 or 4. Inflammatory arthritis (rheumatoid, polyarthritic or autoimmune disease) was excluded, as was post-traumatic or post-septic arthritis. No cases suggestive of a skeletal dysplasia or developmental dysplasia were included. The average age of the cases at replacement surgery was 65 years with an age range of 56–85 years. The control group comprised 822 individuals (368 females and 454 males) with no signs or symptoms of arthritis or joint disease (pain, swelling, tenderness or restriction of movement). The average age of the controls at recruitment was 69 years with an age range of 55–89 years. Because of ethical and financial constraints, the hip and knee joints of the controls were not subjected to radiographic analysis. All cases and all controls were UK citizens of European ethnicity. Ethical approval for the study was obtained from appropriate ethics committees.

Spanish case–control cohort

The selection criteria and case characteristics have been described previously (7). In brief, OA cases were recruited among patients undergoing total hip replacement (THR, n = 304; 183 females and 121 males) or total knee replacement (TKR, n = 274; 223 females and 51 males) or among patients attending the Rheumatology Unit because of hand OA [HOA, n = 240; 211 females and 27 males (two hand OA cases had unrecorded sex status)] complaints and fulfilling ACR classification criteria (19). Controls (n = 1196; 667 females and 529 males) were older than 55 years. This study was conducted following the Helsinki Declaration, was approved by the Ethical Committee for Clinical Research of Galicia and all participants gave their written informed consent. All subjects were of Spanish ancestry and resided in the reference area of the Hospital Clinico Universitario de Santiago, Spain.

Genotyping SNP rs143383 in the UK case–control cohort

The SNP alters a BsiEI restriction enzyme site and was genotyped using a PCR-restriction enzyme analysis. The forward primer had the sequence 5′-AGCACACAGGCAGCATTACG-3′ and the reverse primer the sequence 5′-CCAGTCCCATAGTGGAAATG-3′, creating a 197 bp PCR product. Following digestion with BsiEI (New England Biolabs, Hitchin, UK) a T-allele remains uncut, while a C-allele generates two fragments of 106 bp and 91 bp. Digestion products were electrophoresed through 3% agarose and scored following ethidium bromide staining.

Genotyping SNP rs143383 in the spanish case–control cohort

A TaqMan (Applied Biosystems, Foster City, CA, USA) assay was used, with primers and fluorescence-labelled probes designed and synthesized by Applied Biosystems (sequences are available from the authors). TaqMan reactions were performed in a total volume of 10 µl containing 24 ng of genomic DNA following the Applied Biosystems protocol. A Chromo4 real-time PCR system (MJ Research, Waltham, MA, USA) was used to run the assays. Several samples with different genotypes were sequenced with the Big Dye Ready Reaction Kit (Applied Biosystems) to confirm the accuracy of genotypes.

Power calculations and statistical analysis

The minimum detectable OR under the log-additive model with power greater than or equal to 80% and significance level of 5% was calculated using Quanto version 1.1 (http://hydra.usc.edu/gxe) (20,21) and using the Power and Sample Size software (22). The T-allele frequency was set to 62% (the frequency for all 822 of the UK controls) or to 60.2% (the frequency for all 1196 of the Spanish controls) and the population risk of OA was set to 5%.

Genotype and allele distributions in cases and controls were compared using standard χ2 analysis-of-contingency tables. ORs were calculated with 95% CIs. For stratification analysis, female cases were compared with female controls, and male cases were compared with male controls.

Allelic expression analysis

Using a protocol described previously (23), nucleic acid was extracted from the macroscopically normal articular cartilage of UK OA patients who had undergone hip or knee replacement surgery (THR and TKR, respectively). The cartilage genomic DNA was used to genotype the patients for SNP rs143383 using the restriction enzyme assay described above. Two THR (one female and one male) and seven TKR patients (five females and two males) were identified as heterozygous for SNP rs143383 (these nine patients were collected subsequent to the case–control cohort). The cartilage RNA from these nine patients was then taken forward for an allelic expression analysis using a single base extension (SBE) assay that we have described in detail previously (23). At least 800 ng of RNA was used for the cDNA synthesis using random hexamers and the SuperScript™ kit (Invitrogen, Paisley, UK). Two reverse transcription (RT) reactions were performed for each patient: with reverse transcriptase (+RT) and without reverse transcriptase (−RT). From each + RT reaction, ≥ 19 individual PCR amplifications were carried out using forward primer 5′-CTTCAAGCCCTCAGTCAGTTG-3′ and reverse primer 5′-CGGGTGTGTGTTTGTATCCAG-3′, both located in the 5′-UTR of GDF5. The (−RT) controls did not yield detectable PCR products. The primer 5′-CTCGTTCTTGAAAGGAGAAAGCC-3′, which is located immediately adjacent to SNP rs143383, was used for the SBE assay. To ascertain the peak pattern for an assumed 1:1 ratio between alleles, we performed five individual PCR and SBE reactions on the cartilage genomic DNA of each of the nine patients to give a total of 45 individual genomic DNA measurements. The same PCR primers and SBE primer were used for the cDNA and genomic template. This use of the same analytic conditions for the cDNA and genomic DNA measurements enabled us to use the average of the 45 genomic DNA allelic ratio measurements (representing the assumed 1:1 ratio between alleles) to correct the allelic ratios obtained from the cDNA measurements and thus to account for differences in fluorescent yield and terminator dye incorporation specific to the assay. This correction allowed us to obtain exact values of the relative allelic expression of each cDNA measurement. To determine if there was a significant difference in allelic expression for each patient, we compared the cDNA allelic ratios for that patient to the pooled genomic allelic ratios using a two-tailed Mann–Whitney exact test. To determine if there was an overall difference in expression between the T-allele and the C-allele, we compared the mean allelic ratios for the nine patient cDNAs to the mean allelic ratios for the nine patient genomic DNAs using a two-tailed t-test.

SUPPLEMENTARY MATERIAL

Supplementary material is available at HMG Online.

ACKNOWLEDGEMENTS

We thank Andrew Carr, Kim Clipsham, Bridget Watkins and Nick Athanasou, who helped organize the collection of samples from UK individuals evaluated in this report, and Yolanda Lopez-Golan and Fina Meijide for their help in recruiting Spanish study participants. Marta Picado and Cristina Fernandez provided outstanding technical assistance. This study was supported by Research into Ageing (J.L.), the Arthritis Research Campaign (J.L.), by grant PI02/0713 from the Instituto de Salud Carlos III, Spain (A.G.), with funds from FEDER of the European Union (A.G.) and by the Fundacion MMA of Madrid, Spain (A.G.). J.R.-L. is the recipient of a scholarship of the Spanish Ministry of Education. J.M.W. is the recipient of a Marshall Scholarship. M.P.-S. received a bursary from the Fundacion Española de Reumatologia.

Conflict of Interest statement. None declared.

REFERENCES

1
Luyten
F.P.
Cartilage-derived morphogenetic protein-1
Int. J. Biochem. Cell Biol.
 , 
1997
, vol. 
29
 (pg. 
1241
-
1244
)
2
Edwards
C.J.
Francis-West
P.H.
Bone morphogenetic proteins in the development and healing of synovial joints
Semin. Arthritis Rheum.
 , 
2001
, vol. 
31
 (pg. 
33
-
42
)
3
Mikic
B.
Multiple effects of GDF-5 deficiency on skeletal tissues: implications for therapeutic bioengineering
Ann. Biomed. Eng.
 , 
2004
, vol. 
32
 (pg. 
466
-
476
)
4
Francis-West
P.H.
Abdelfattah
A.
Chen
P.
Allen
C.
Parish
J.
Ladher
R.
Allen
S.
MacPherson
S.
Luyten
F.P.
Archer
C.W.
Mechanisms of GDF-5 action during skeletal development
Development
 , 
1999
, vol. 
126
 (pg. 
1305
-
1315
)
5
Hatakeyama
Y.
Tuan
R.S.
Shum
L.
Distinct functions of BMP4 and GDF5 in the regulation of chondrogenesis
J. Cell Biochem.
 , 
2004
, vol. 
91
 (pg. 
1204
-
1217
)
6
Miyamoto
Y.
Mabuchi
A.
Shi
D.
Kubo
T.
Takatori
Y.
Saito
S.
Fujioka
M.
Sudo
A.
Uchida
A.
Yamamoto
S.
, et al.  . 
A functional polymorphism in the 5′-UTR of GDF5 is associated with susceptibility to osteoarthritis
Nat. Genet.
 , 
2007
, vol. 
39
 (pg. 
529
-
533
)
7
Rodriguez-Lopez
J.
Pombo-Suarez
M.
Liz
M.
Gomez-Reino
J.J.
Gonzalez
A.
Lack of association of a variable number of aspartic acid residues in the asporin gene with osteoarthritis susceptibility: case–control studies in Spanish Caucasians
Arthritis Res. Ther.
 , 
2006
, vol. 
8
 pg. 
R55
 
8
Mustafa
Z.
Dowling
B.
Chapman
K.
Sinsheimer
J.S.
Carr
A.
Loughlin
J.
Investigating the aspartic acid (D) repeat of asporin as a risk factor for osteoarthritis in a UK Caucasian population
Arthritis Rheum.
 , 
2005
, vol. 
52
 (pg. 
3502
-
3506
)
9
Kaliakatsos
M.
Tzetis
M.
Kanavakis
E.
Fytili
P.
Chouliaras
G.
Karachalios
T.
Malizos
K.
Tsezou
A.
Asporin and knee osteoarthritis in patients of Greek origin
Osteoarthritis Cartilage
 , 
2006
, vol. 
14
 (pg. 
609
-
611
)
10
Jiang
Q.
Shi
D.
Yi
L.
Ikegawa
S.
Wang
Y.
Nakamura
T.
Qiao
D.
Liu
C.
Dai
J.
Replication of the association of the aspartic acid repeat polymorphism in the asporin gene with knee-osteoarthritis susceptibility in Han Chinese
J. Hum. Genet.
 , 
2006
, vol. 
51
 (pg. 
1068
-
1072
)
11
Ikegawa
S.
Kawamura
S.
Takahashi
A.
Nakamura
T.
Kamatani
N.
Replication of association of the D-repeat polymorphism in asporin with osteoarthritis
Arthritis Res. Ther.
 , 
2006
, vol. 
8
 pg. 
403
 
12
Mototani
H.
Mabuchi
A.
Saito
S.
Fujioka
M.
Iida
A.
Takatori
Y.
Kotani
A.
Kubo
T.
Nakamura
K.
Sekine
A.
, et al.  . 
A functional single nucleotide polymorphism in the core promoter region of CALM1 is associated with hip osteoarthritis in Japanese
Hum. Mol. Genet.
 , 
2005
, vol. 
14
 (pg. 
1009
-
1017
)
13
Loughlin
J.
Sinsheimer
J.S.
Carr
A.
Chapman
K.
The CALM1 core promoter polymorphism is not associated with hip osteoarthritis in a United Kingdom Caucasian population
Osteoarthritis Cartilage
 , 
2006
, vol. 
14
 (pg. 
295
-
298
)
14
Valdes
A.M.
Loughlin
J.
Van Oene
M.
Chapman
K.
Surdulescu
G.L.
Doherty
M.
Spector
T.D.
Sex and ethnic differences in the association of ASPN, CALM1, COL2A1, COMP, and FRZB with genetic susceptibility to osteoarthritis of the knee
Arthritis Rheum.
 , 
2007
, vol. 
56
 (pg. 
137
-
146
)
15
Kizawa
H.
Kou
I.
Iida
A.
Sudo
A.
Miyamoto
Y.
Fukuda
A.
Mabuchi
A.
Kotani
A.
Kawakami
A.
Yamamoto
S.
, et al.  . 
An aspartic acid repeat polymorphism in asporin inhibits chondrogenesis and increases susceptibility to osteoarthritis
Nat. Genet.
 , 
2005
, vol. 
37
 (pg. 
138
-
144
)
16
Nakamura
T.
Shi
D.
Tzetis
M.
Rodriguez-Lopez
J.
Miyamoto
Y.
Tsezou
A.
Gonzalez
A.
Jiang
Q.
Kamatani
N.
Loughlin
J.
, et al.  . 
Meta-analysis of association between the ASPN D-repeat and osteoarthritis
Hum. Mol. Genet.
 , 
2007
, vol. 
16
 (pg. 
1676
-
1681
)
17
Goldring
M.B.
The role of the chondrocyte in osteoarthritis
Arthritis Rheum.
 , 
2000
, vol. 
43
 (pg. 
1916
-
1926
)
18
Loughlin
J.
Polymorphism in signal transduction is a major route through which osteoarthritis susceptibility is acting
Curr. Opin. Rheumatol.
 , 
2005
, vol. 
17
 (pg. 
629
-
633
)
19
Altman
R.
Alarcon
G.
Appelrouth
D.
Bloch
D.
Borenstein
D.
Brandt
K.
Brown
C.
Cooke
T.D.
Daniel
W.
Feldman
D.
, et al.  . 
The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hand
Arthritis Rheum.
 , 
1990
, vol. 
33
 (pg. 
1601
-
1610
)
20
Gauderman
W.J.
Sample size requirements for matched case–control studies of gene environment interaction
Stat. Med.
 , 
2002
, vol. 
21
 (pg. 
35
-
50
)
21
Gauderman
W.J.
Candidate gene association studies for a quantitative trait, using parent offspring trios
Genet. Epidemiol.
 , 
2003
, vol. 
25
 (pg. 
327
-
338
)
22
Dupont
W.D.
Plummer
W.D.
Jr
Power and sample size calculations. A review and computer program
Control Clin. Trials
 , 
1990
, vol. 
11
 (pg. 
116
-
128
)
23
Wilkins
J.M.
Southam
L.
Price
A.J.
Mustafa
Z.
Carr
A.
Loughlin
J.
Extreme context specificity in differential allelic expression
Hum. Mol. Genet.
 , 
2007
, vol. 
16
 (pg. 
537
-
546
)

Author notes

The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors.

Supplementary data