Materials and methods

Patients and controls

A total of 550 unrelated patients with hip OA (232 from Hospital Clínico Universitario, Santiago de Compostela and 318 from Hospital Universitario A Coruña) were included in the study. Patients meeting the inclusion criteria included 340 women and 210 men aged >50 years (mean (SD) 69.10 (7.43) years; range 51–95) and diagnosed with hip OA according to the ACR criteria. The 505 asymptomatic donors who met the inclusion criteria for normal subjects (235 from Hospital Clínico Universitario in Santiago de Compostela and 270 from Hospital Universitario A Coruña) included 261 women and 244 men aged >50 years (mean (SD) age 68.78 (9.43) years; range 50–94) who lacked the ACR criteria for a diagnosis of hip OA. The cohort from Santiago has been described previously7 and included patients with severe primary OA who had undergone total hip replacement and controls with no OA-related complaints or hand deformities as assessed from the patient history using a specific questionnaire and clinical examination. The cohort from A Coruña included patients with hip OA with different radiographic Kellgren-Lawrence grades, and controls free of hip OA as assessed from the patient history, clinical examination and radiographic studies. A set of clinical variables such as sex, age and smoking status was also collected for most of the subjects.

In this study an additional experimental group was included which consisted of patients who developed knee OA, hip OA or knee plus hip OA (n = 977) and controls with no knee and hip complaints (n = 179; table 1). A cohort of 252 patients with rheumatoid arthritis (RA) was also analysed to compare the frequency distribution of mtDNA haplogroups in order to validate the specificity of the results for OA.

Genotyping of mtDNA haplogroups

We assessed the most common European mtDNA haplogroups and both J1c (m.14798T→C) and J2 (m.15257G→A) genotypes, as previously described.6 The polymorphic sites analysed in this study have been previously reported8 (see table 1 in online supplement).

Statistical analyses

Statistical analyses were performed using SPSS software release 15 (Chicago, Illinois, USA). Haplogroup analysis was performed following a previously described approach.6 A model of multivariate logistic regression analysis was used to test the influence of age, sex, smoking and haplogroup J.

Results

Distribution of mtDNA haplogroups

The frequency of haplogroup J was significantly lower in patients with hip OA than in the controls when each haplogroup was compared with all other pooled haplogroups (OR = 0.661; 95% CI 0.440 to 0.993; p = 0.045; table 2). The frequency of sub-haplogroup J1c was significantly lower in patients with hip OA than in controls when this sub-haplogroup was compared with the other pooled haplogroups (OR = 0.492; 95% CI 0.281 to 0.862; p = 0.012; table 2). However, analyses of the single nucleotide polymorphisms (SNPs) that characterise the European mtDNA haplogroups did not show statistical significance (see table 2 in online supplement).

When stratified by sex, only the frequency for sub-haplogroup J1c bordered on statistical significance in men with OA (OR = 0.444; 95% CI 0.191 to 1.030; p = 0.053; table 3). None of the SNPs showed statistical significance.

Multivariate analysis

For the hip study, after adjusting for sex, age, haplogroup J and smoking, only haplogroup J (OR = 0.519; 95% CI 0.311 to 0.946; p = 0.048) and smoking (OR = 0.543; 95% CI 0.311 to 0.946; p = 0.031) were significantly independently associated with hip OA (table 4). These data support the conclusion that patients with hip OA are represented less frequently in haplogroup J and smoke less than controls with healthy hips.

For the knee and hip study, the results showed that, after adjusting for sex, age, haplogroup J and smoking, haplogroup J (OR = 0.406; 95% CI 0.214 to 0.772; p = 0.006) was independently associated with all the forms of OA represented (table 4). These data indicate that patients with OA are less frequently represented in haplogroup J than controls with healthy knees and hips, and that individuals carrying haplogroup J are at a lower risk for developing OA than non-J haplogroup carriers of the same age (p<0.001) (see fig 1 in online supplement).

Validation of disease specificity

We compared mtDNA haplogroup frequencies in patients with RA with controls without hip RA. No relationship was found between RA and the mtDNA haplogroup J or other haplogroups or SNPs (see table 3 in online supplement).

Discussion

A previous study reported by our group revealed that mtDNA haplogroups have a role in the prevalence and severity of knee OA.6 In the present study we attempted to confirm the results obtained in the study of knee OA in patients with hip OA. The incidence of smoking in patients with hip OA was also examined.

The results obtained here suggest that mtDNA haplogroup J offers protection from the occurrence of hip OA. However, the apparent protective effect related to the 4216C allele and the overexpression of SNP 14766C in patients with knee OA found in our previous study were not detected. These data suggest that the protective effect could be due either to an overexpression of some unidentified “OA-protective” mutation(s) linked to this haplogroup, as in the case of the m.150C→T mutation associated with centenarians,9 or the under-representation of some unidentified “OA-risk” mutation(s) on haplogroup J carriers, as in the case of the m.3243A→G mutation in the French population.10 However, we cannot rule out the possibility that the set of polymorphisms characteristic of haplogroup J, including the SNP m.14798T→C, are responsible for this protective effect. This latter supposition would be more in accordance with the hypothesis proposed elsewhere that climate influenced the selection of European mtDNA haplogroups towards a strong selective advantage for mtDNA mutations that partially uncoupled mitochondrial oxidative phosphorylation (OXPHOS).11 These uncoupling mutations would decrease mitochondrial ROS production by increasing the oxidation of the electron transport chain, thus reducing oxidative damage and apoptosis.12

Although ROS are involved in several intracellular signalling pathways as secondary messengers under normal and pathological conditions in chondrocytes,13 it has also been reported that an important biochemical factor involved in the development of OA is mitochondrial ROS production1 and apoptosis.4 Our group recently found that the inhibition of complexes III and V of the MRC induces cyclooxygenase (COX-2) expression and PGE₂ production in chondrocytes through the generation of ROS, among other factors.3

After adjusting for sex, age and haplogroup J, both haplogroup J and smoking were independently associated with hip OA only in the hip study group, indicating that individuals who carry mtDNA haplogroup J and smoke are less prone to develop hip OA, regardless of sex and age. Several retrospective studies have suggested a protective role for tobacco use against OA, and some of them indicated a high level of nicotine stimulation for the anabolic activity of chondrocytes.14 However, further studies are required to determine if nicotine has a chondroprotective role.

In this study we compared mtDNA haplogroup frequencies of patients with RA with controls without hip RA. No statistically significant differences were found between the two groups. This suggests that, although both pathologies cause joint damage, the protective effect of haplogroup J affects only an age-related and oxidative stress disease such as OA, which is in agreement with the hypothesis proposed here and in our previous work.

Because of the controversial associations found within the mtDNA haplogroup J,5 some authors propose the “haplogroup J paradox” in which a haplogroup that may increase the expression of mtDNA pathogenic mutations can also be beneficial in certain environmental contexts.10 The results from this study and our previous study on knee OA strengthen the hypothesis that European mtDNA haplogroup J, which is clearly biochemically different from those of other population-specific mtDNA lineages,15 protects against the development of OA.