Objective To analyse the influence of mitochondrial DNA (mtDNA) haplogroups on serum levels of molecular biomarkers in patients with osteoarthritis (OA).
Methods Serum levels of molecular biomarkers of cartilage metabolism (collagen type II markers: C-terminal neoepitope generated by the collagenase-mediated cleavage of collagen type II triple helix (C2C), collagen type II (Coll2-1, and its nitrated form, Coll2-1NO2), procollagen type II (CPII)), synovial metabolism (hyaluronic acid (HA)) and cartilage and synovial turnover (cartilage glycoprotein 39 (YKL-40)) were analysed in 73 patients with OA and 77 healthy controls using ELISAs. All participants had been previously genotyped for the mtDNA haplogroups J, U and H. Non-parametric and multivariate analysis were performed to test the effects of the clinical variables, including gender, age, smoking status, diagnosis, mtDNA haplogroups and radiological Kellgren and Lawrence (K/L) grade on the serum levels of the molecular markers.
Results Non-parametric analysis found increased serum levels of HA in patients with OA, while the values for C2C and the C2C/CPII ratio were significantly higher in the healthy controls. A multiple regression analysis showed a relationship between the mtDNA haplogroups and serum levels of the typical collagen type II markers. Carriers of the mtDNA haplogroup H had higher levels while carriers of the mtDNA haplogroup J showed lower levels. Statistically significant interactions between mtDNA haplogroups and diagnosis and between mtDNA haplogroups and radiological K/L grade in the serum levels of molecular markers were also found.
Conclusion A new role for mtDNA haplogroups emerges from this work. The results suggest that the mtDNA haplogroups interact significantly with the serum levels of OA-related molecular markers, suggesting the possibility of their use as a complementary assay with these molecular markers.
Statistics from Altmetric.com
Osteoarthritis (OA) is the most common joint disease related to ageing and is characterised by degeneration of articular cartilage leading to joint destruction and severe impairment of mobility.1 Globally, OA is a major cause of disability in the ageing population. The best established methods for assessing the progression of cartilage loss are the measurement of joint space width (JSW) using plain radiography and arthroscopic evaluation.2 However, OA has little efficient therapeutics, probably as a consequence of its late diagnosis, which results from physical examination and radiographic studies being relatively poor means for early detection; joint damage detectable on radiographs is considered irreversible. Because changes in JSW are relatively small compared to the precision error of radiographic measurement, at least 1 or 2 years of monitoring are usually required for the accurate assessment of the progression of joint damage or its inhibition by treatment.3
MRI is a promising non-invasive tool for the evaluation of cartilage; however, access to this technique is limited and expensive. Like standard radiography, MRI does not provide information on metabolic changes occurring in OA cartilage. The reproducibility of MRI assessment is not yet fully validated and the technique is currently being optimised for the monitoring of patients with OA.4
To overcome all these drawbacks, molecular markers have been developed in an effort to detect changes in OA with more reliability and sensitivity, preferably in an earlier stage of the disease.4 5 These markers are molecules released into biological fluids during the process of tissue biosynthesis and turnover. During normal metabolism, joint tissue macromolecules or fragments thereof are released into the synovial fluid, and from there to blood and urine, where they can be detected by immunoassays. Several molecular markers of bone, cartilage and synovium have been described as useful for early identification of patients with OA, patients at high risk for progression, monitoring disease progression and assessing therapeutic responses in OA, because of their greater sensitivity compared to radiographs.4 6,–,8 Among them, those related with collagen type II, one of the major constituents of cartilage, have been suggested as the ideal marker for monitoring cartilage degradation.9
Several data have led to the suggestion that mitochondria mediate the pathogenesis of OA.10 A significant decrease in complex II and III activity in OA chondrocytes compared with normal chondrocytes has been demonstrated, and the mitochondrial mass was also shown to be increased in OA chondrocytes11; additionally, the apoptotic mitochondrial pathway has been implicated to be one of the major cellular pathways of apoptosis in OA chondrocytes.12 In addition, the inhibition of complexes III and V of the mitochondrial respiratory chain (MRC) causes an increased inflammatory response, which could be particularly relevant to the production of prostaglandin E2 (PGE2) and reactive oxygen species (ROS).13 The mitochondrial-free radical production has been shown to compromise chondrocyte function,14 15 causing mitochondrial DNA (mtDNA) damage and poor mtDNA repair capacity.16 Our group recently found that haplogroup J is associated with a lower incidence and severity of knee17 and hip OA18 in the Spanish population. This may be explained by the decreased free radical production caused by reduced coupling efficiency of the oxidative phosphorylation system (OXPHOS).19
Based on these findings, we performed a retrospective study in a population from northern Spain. The purpose was to attempt to demonstrate a correlation between the haplogroups and the serum levels of several OA-related molecular biomarkers. To carry out this study, we analysed the serum levels of 6 molecular markers for cartilage and synovium in 73 patients with OA and 77 healthy controls carrying the mtDNA haplogroups J, U or H.
Materials and methods
Patients and healthy controls
The population analysed in this study has been previously described in detail,17 18 and all participants attended the Hospital Universitario A Coruña. The patients meeting the inclusion criteria for this study had been diagnosed as having OA according to American College of Rheumatology (ACR) criteria20 and knee and hip radiographs from 148 participants were classified according to Kellgren and Lawrence (K/L) scoring from grade I to grade IV21 (see supplementary materials and methods).
mtDNA haplogroup genotyping
Samples from all the participants included in the present study were haplogrouped by a previously described assay17 (see supplementary materials and methods).
Fasting blood samples were collected from all patients in plain tubes containing separation gel. They were allowed to stand for 20 min and then centrifuged for 10 min at 800g. The serum was aliquoted and stored at −80ºC until assayed.
The molecular markers measured were a denaturation epitope of the triple helical domain of the collagen type II (Coll2-1) and its nitrated form (Coll2-1NO2),22 a C-terminal neoepitope generated by the collagenase-mediated cleavage of collagen type II triple helix (C2C), the procollagen type II C-terminal propeptide (CPII), hyaluronic acid (HA) and cartilage glycoprotein 39 (YKL-40) (see supplementary materials and methods).
Statistical analyses were performed using SPSS software, V.17 (SPSS, Chicago, Illinois, USA). Due to the multiple tests performed, p values between 0.01 and 0.05 should be interpreted with caution, therefore we considered p≤0.01 to be significant. Non-parametric and multiple regression analysis (MRA) were performed (supplementary materials and methods).
Non-parametric analysis of serum levels of the molecular markers
Results of the analysis showed that levels of C2C and the C2C/CPII ratio are significantly elevated in the serum of healthy controls, as well as a trend towards increased serum levels of Coll2-1; while HA appeared significantly elevated in the serum of patients with OA (table 1). Interestingly, the haplogroups have a significant influence on the serum levels of all of the collagen type II markers regardless of diagnosis, except for the Coll2-1NO2/Coll2-1 ratio (Coll2-1 (p<0.001); Coll2-1NO2 (p=0.002); C2C (p<0.001); CPII (p<0.001) and the C2C/CPII ratio (p<0.001)), with the haplogroup H carriers having the highest values while those with haplogroup J showed lower values (tables 2 and 3). Similarly, radiographic severity, based on the K/L score, had a significant influence on the serum levels of some of the collagen type II markers (C2C (p=0.003) and the C2C/CPII ratio (p<0.001)), which were higher in radiographic group A (K/L grades 0 and I). Serum levels of HA were significantly (p<0.001) higher in groups B (K/L grades II and III) and C (K/L grade IV) (table 4).
Multiple regression analysis
The results of the MRA for each of the biomarkers analysed in this study are described below.
The analysis performed on the overall experimental population showed an influence of the haplogroups in the serum levels of Coll2-1 (p<0.001), being higher in participants carrying haplogroup H when compared to non-H carriers (p=0.002) (figure 1A). Interestingly, a significant interaction between diagnosis and haplogroups was also found (p=0.004) with serum levels of Coll2-1. Patients with OA carrying haplogroup H had higher levels of Coll2-1 than patients with OA carrying haplogroups J and U (p<0.001) (figure 2A, table 3), whereas no significant interaction was observed in the healthy controls. The analysis of the radiological groups showed significant interactions between the haplogroups and radiological groups on the serum levels of Coll2-1 (p=0.017). Carriers of haplogroup H within radiological group B showed higher serum levels of Coll2-1 than carriers of haplogroup J (p=0.014) and U (p=0.004) (figure 3A).
The analyses of the results for Coll2-1NO2 were similar to those of Coll2-1. Following adjustment for gender, age, diagnosis, smoking status and haplogroups, the serum levels of Coll2-1NO2 were significantly influenced by the haplogroups (p=0.002). As with Coll2-1, this finding was also detected when comparing, after Bonferroni post hoc analysis, haplogroup H to haplogroup U carriers (p=0.014) and to haplogroup J carriers (p=0.013) (figure 1B). Interestingly, a significant relationship between diagnosis and haplogroups was also found (p=0.001). Patients with OA carrying mtDNA haplogroup H showed higher levels of Coll2-1NO2 than any other group (p<0.01 for all the possible combinations) (figure 2B and table 3). As described earlier, the analysis of radiological group B showed that carriers of haplogroup H had higher serum levels of Coll2-1NO2 than carriers of haplogroup J and U (p=0.001) (figure 3B).
The results of the Coll2-1NO2/Coll2-1 ratio showed a non-significant trend towards an influence of the haplogroups in this ratio (p=0.028). Decreased values in people carrying haplogroup J when compared with people carrying haplogroup H (p=0.042) were detected. Interactions between haplogroups and diagnosis bordered the statistical significance (p=0.020). Subsequent Bonferroni post hoc analysis showed that patients with OA carrying haplogroup H had higher ratios than patients with OA carrying haplogroup J (p=0.003) (table 3). The analysis of the radiological groups showed statistical interactions between haplogroups and radiological groups on the Coll2-1NO2/Coll2-1 ratio (p=0.009). Carriers of haplogroup H within group B have higher ratios than carriers of haplogroup J (p=0.002) and U (p=0.010) (figure 3C).
MRA showed higher serum levels of C2C in healthy controls (p<0.001) compared to patients with OA. Interestingly, the results also found an influence of haplogroups on the serum levels of C2C (p<0.001). Participants carrying haplogroup J showed the lowest levels of serum C2C and those carrying haplogroup H showed the highest levels. This correlation was maintained when comparisons among the three haplogroups were performed (p<0.003 for all the comparisons) (figure 1C). No significant interaction between diagnosis and haplogroups was seen in serum levels of C2C. Statistical interactions between haplogroups and radiological groups were not significant.
A significant influence of the haplogroups in the serum levels of CPII was detected regardless of diagnosis, gender and smoking status (p<0.001). Carriers of the haplogroup H showed higher levels than non-H carriers (p<0.001) (figure 1D). Interestingly, a significant relationship between diagnosis and haplogroups was also detected (p=0.003). Patients with OA carrying the haplogroup H showed higher serum levels of CPII than patients with OA and healthy controls carrying haplogroup J (p<0.001 for OA J participants and p=0.01 for healthy J participants) and patients with OA carrying the haplogroup U (p<0.001). The healthy controls carrying haplogroup H showed a trend towards higher values than patients with OA with the haplogroup J (p=0.021) (figure 2C and table 3). Analysis of radiological group B also showed significant differences in CPII serum levels (p=0.001). Carriers of haplogroup H within radiological group B showed higher serum levels of CPII than carriers of haplogroups J and U (p<0.001) (figure 3D).
MRA showed that the C2C/CPII ratio was higher in healthy controls than in patients with OA regardless of gender and smoking status (p<0.001). A non-significant trend towards an increased ratio with ageing was seen. Interestingly, participants carrying haplogroup J showed the lowest ratio (p<0.001) (figure 1E). No significant interactions between diagnosis or radiological groups and haplogroups were found.
The results showed significantly higher serum levels of HA in patients with OA than in healthy controls (p<0.001), and in men (p<0.001) compared to women. Interestingly, these levels increased with age (p<0.001). A trend towards interactions between diagnosis and haplogroups was also detected (p=0.044). Patients with OA carrying the haplogroup H showed higher serum levels of HA than did healthy controls carrying haplogroups H or U, while patients with OA carrying haplogroup J have higher levels of HA than healthy controls with haplogroup H (table 3). No statistical interactions between haplogroups and radiological groups were found.
The MRA found no statistical differences in the serum levels of YKL-40. A significant (p<0.001) increase of YKL-40 concentration with age was observed. The analysis of radiological groups showed no statistical interactions between haplogroups and radiological group B.
A new role for the haplogroups in the pathogenesis of OA has emerged from this work. The results obtained provide clear evidence, for the first time, that haplogroups are related to serum levels of several molecular biomarkers of OA. The choice of the haplogroups J and U for this study was due to the relationship that our group found between the haplogroups J and U with the lower prevalence and severity of knee17 and hip OA.18 Regarding to the haplogroup H, it was also selected because is the most prevalent in European populations.23
Because of the relation found by our group between the haplogroup J and knee and hip OA,17 18 and in order to obtain an adequate sample size, we pooled knee and hip samples to perform the statistical analyses. Initially, a non-parametric statistical analysis, in which the clinical variables described in this study were not taken into account, was performed. From these results we conclude that HA appeared statistically increased in the serum of patients with OA compared to healthy controls, agreeing with previous studies.24 25 Collagen type II molecular markers, such as C2C and the C2C/CPII ratio were significantly higher in healthy controls compared to patients with OA.
The MRA also showed C2C and C2C/CPII ratio were significantly higher in healthy controls, while this analysis also showed HA levels to be higher in patients with OA. Interestingly, the MRA also showed a clear relationship of the haplogroups with the serum levels of all the collagen type II markers. Significant interactions between diagnosis and haplogroups were found in all the molecular markers except C2C and the C2C/CPII and Coll2-1NO2/Coll2-1 ratios. To summarise, participants carrying the haplogroup J, who are those with a lower prevalence and severity of knee OA17 and with lower prevalence of hip OA,18 showed the lowest serum levels of almost all collagen type II markers. Inversely, people carrying the haplogroup H, the most prevalent in European populations, had the highest levels of all the collagen type II markers.
We suggest that the collagen type II biomarker serum level results obtained in this work that are not in agreement with those obtained by other authors26,–,28 are clearly due to the influence of haplogroups, so an explanation is possible. It must be taken into account that the frequency of the most prevalent haplogroup in European populations, haplogroup H, is about 40% to 47%, while haplogroup J is about 7% to 12% and haplogroup U is about 12% to 22%,18 23 29 so these are the expected frequencies in the studies carried out by all the other authors. On the contrary, as we want to test the influence of the haplogroups in the serum levels of OA-related molecular biomarkers, in our study the frequency of each of the haplogroups was equal (approximately 33.3% for each haplogroup); If we look only at the results obtained from the most prevalent haplogroup H, a trend towards agreement with the results obtained by other authors, in which cohort of samples there will be a clear prevalence of this haplogroup, can be seen: patients with OA that carry haplogroup H have significantly higher serum levels of Coll2-1NO2 and elevated levels of Coll2-1 and Coll2-1NO2/Coll2-1 ratio than healthy controls carrying any of the three haplogroups (table 3). Besides, healthy controls that carry the haplogroup H showed a trend (bordering the statistical significance) towards higher serum levels of CPII than patients with OA that carry the haplogroup J. Analysing the serum levels of these molecular markers after adjusting the sample population to reflect the frequency of these three haplogroups in a randomised subgroup of the cohort analysed in this study, the results trend towards those described by other authors (data not shown). Hence, the whole population analysed in the present study is reflecting a clear relationship among haplogroups and the serum levels of OA-related molecular biomarkers.
When radiological group B, consisting of 47 patients with OA with radiological K/L grades II and III, was analysed separately, a pronounced influence of the haplogroups on serum marker levels was observed too. Carriers of the haplogroup H tended to have higher serum levels of most collagen type II molecular markers, while those carrying haplogroup J had lower values. Because the typical randomised cohort in most clinical trials has consisted of patients with OA with radiological K/L grades II and III, these findings are very interesting and highly relevant.
A possible explanation for the influence of the haplogroups in the serum levels of OA-related molecular markers may be that haplogroup H carriers have more active collagen type II metabolism than haplogroup J carriers do, perhaps because of differences in OXPHOS performance.30 This indicates that an altered mitochondrial function plays an important role in the development of OA, which is reflected by the serum levels of collagen type II biomarkers. Our group has previously demonstrated a significant decrease in complex II and III activity in OA chondrocytes, compared with normal chondrocytes,11 and that tumour necrosis factor α (TNFα) and interleukin 1 (IL1), which are the two most important cytokines involved in the OA process, regulate mitochondrial function in articular chondrocytes.31 Interestingly, patients with OA carrying haplogroup J have lower serum levels of Coll2-1NO2 than do patients with OA carrying the haplogroup H. Haplogroup J carriers also showed the lowest levels of C2C and the C2C/CPII ratio, indicating less destruction of collagen type II molecules. Because the effect of nitric oxide (NO) on chondrocyte survival has been shown to be mediated by its effect on the MRC,32 and Coll2-1NO2 is an indicator of oxidative stress status of the chondrocyte,26 these findings suggest that haplogroup J carriers having OA may have less oxidative chondrocyte stress. This would be in accordance with other results obtained by our group showing that people carrying haplogroup J may be at lower risk for developing knee OA, and those carrying this haplogroup who do develop knee OA may have decreased severity in disease progression, probably due to reduced ROS production.17 Our group has also recently found that the inhibition of complexes in the MRC of chondrocytes causes a ROS-mediated inflammatory response and increased matrix metalloproteinase 1 (MMP1) and MMP3 production.13
HA is widely distributed throughout many tissues; thus, its presence in the serum can be caused by conditions other than joint damage.8 However, increased serum levels of HA have been reported in patients with OA24 and in patients with OA with faster radiological progression.25 In this work we observed a trend towards an interaction between haplogroups and diagnosis on serum levels of HA. Patients with OA with haplogroup H showed a trend towards higher levels of HA than healthy controls with the same haplogroup. Overall, patients with OA showed higher levels of serum HA than healthy controls, and patients in radiological groups B and C showed proportionally increased serum HA values than those in group A.
Serum levels of YKL-40 did not show significant differences between healthy controls and patients with OA, unlike previous studies that showed patients with severe knee OA have higher serum levels of YKL-40 and suggested YKL-40 may be a useful marker for OA.33 34 MRA did not show significant association of YKL-40 with haplogroups.
In summary, the main target of this study was to test the influence of the haplogroups in the serum levels of several OA-related molecular biomarkers, for this reason, we selected patients with OA and healthy controls that carry the haplogroup J, U or H in a similar frequency. This study found a relationship between the haplogroups and serum levels of several OA-related biomarkers. Based on these results we suggest that the effect of haplogroup on serum level of markers be considered when levels of collagen type II markers are used for diagnosis or prognosis of OA. Clinical trials might be best advised to use haplogroup H. Our results support the important role of haplogroups in the complex OA process.
We are grateful to Ms Pilar Cal Purriños for her expert secretarial assistance. The authors express appreciation to Lourdes Sanjurjo and Maria Dolores Velo and to the Department of Orthopaedic from the CHU A Coruña for providing cartilage samples. This study was supported by grants from Secretaria I+D+I Xunta Galicia (PGIDIT06PXIC916175PN); Fundación Española de Reumatologia (programa GEN-SER) and from Fondo Investigación Sanitaria (CIBER-CB06/01/0040)-Spain, with participation of funds from FEDER (European Community). IR-P was supported by Contrato de Apoyo a la Investigación-Fondo Investigación Sanitaria (CA06/01102).
Competing interests None.
Patient consent Obtained.
Ethics approval This study was conducted with the approval of the Ethics Committee from Galicia-Spain.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.