Symptomatic knee osteoarthritis (OA) is a chronic progressive joint disease that occurs in approximately 6% of adults 30 years of age or older.1 Because of its prevalence and the disability that accompanies the disease in a considerable proportion of patients, it has great impact and consequences in the aging population of the industrialised world. Joint destruction as a consequence of knee OA is therefore the most common reason for total knee joint replacement. The rate of knee OA progression varies between patients and is associated with several factors, most importantly local biomechanical factors2 3 and inflammatory factors as reflected by knee effusion and warmth.4 Recently, we showed that familial factors may also play a role in the progression of knee OA.5 Subsequently, it is of interest to search for risk factors for progression of knee OA with a genetic basis, since this will give more insight into the pathophysiological processes involved in the progression of this disease. This insight could lead to new therapeutic modalities for knee OA.

There has been increasing evidence for the involvement of cytokines as mediators in cartilage degradation and inflammation.6 Cytokines influence the metabolic state of chondrocytes and the balance between pro- and anti-inflammatory cytokines may determine the severity of cartilage damage. Prominent pro-inflammatory cytokines involved in OA include interleukin (IL)1β and tumour necrosis factor (TNF)α.7^–9 They have been shown to stimulate production of proteases and prostaglandin E₂ (PGE2), to induce synovial cells and chondrocytes to produce other cytokines such as IL8 and IL6, and to stimulate their own production, leading to accelerated damage of articular tissue.6

Increased levels of anti-inflammatory cytokines have also been found in the synovial fluid of patients with OA.10 Anti-inflammatory cytokines, such as IL4, IL10 and IL13 have been shown in vitro to decrease the production of IL1β, TNFα and matrix metalloproteinases (MMPs), upregulate IL1-receptor antagonist (Ra), as well as inhibit the release of PGE2.11 12

Innate immunity provides an important candidate genetic component of OA development and progression. Twin studies have shown that the ex vivo innate production of cytokines upon lipopolysaccharide (LPS) stimulation is influenced substantially by genetic factors and varies by 53–86% on the basis of inheritability alone.13 Inheritable differences in cytokine production upon stimulation seems to play a role in families with a predisposition for development of OA.14 Patients with a high innate ex vivo production of IL1β and IL1Ra and low innate ex vivo production of IL10 had an increased risk of having OA at multiple joint sites. Nothing is known, however, about the influence of innate cytokine production on the rate of OA progression, which may have more relevance from a clinical perspective than just the presence of OA.

Elucidating the role of cytokines in knee OA progression would improve our understanding of the complex processes involved in progression of knee OA. The inter-individual differences in the capacity to produce cytokines can be used as an endophenotype to enhance insight into the genetic risk factors for knee OA progression. Eventually this knowledge could offer a potential target for disease-modifying drugs to alter the course of this disease. Hence, we investigated whether variations in the innate ex vivo production of IL1β, TNFα, IL1Ra and IL10 in whole blood assays upon LPS stimulation is associated with the radiological progression of knee OA over 2 years in a prospective cohort of patients with symptomatic knee OA.

METHODS

Study design and patient selection

Patients are participants of the Genetics, Arthrosis and Progression (GARP) study, which is a prospective longitudinal cohort study, aimed at identifying determinants of OA susceptibility and progression. The study is based on sibling pairs of caucasian Dutch ancestry with OA at multiple sites. The GARP study was approved by the Medical Ethics Committee of the Leiden University Medical Center.

Details of recruitment, selection procedures and inclusion into the GARP study have been described previously.15 Briefly, probands (aged 40–70 years) and their siblings, with OA at multiple sites in the hands or at two or more of the following joints sites: hand, spine (cervical or lumbar), knee or hip, were included after giving informed consent. Of these sibling pairs, all with at least one subject with symptomatic hip or knee (but not radiological end stage disease) OA were eligible for the 2-year follow-up study. In the present study, only patients with symptomatic knee OA in at least one knee at baseline were included.

Patients with secondary OA, familial syndromes with a Mendelian inheritance pattern, or a very limited life expectancy were excluded from the GARP study. Crystal deposition arthropathies (unless in the case of severe polyarticular gout) and diabetes mellitus or thyroid conditions were not considered as exclusion criteria.

Diagnosis of OA

Symptomatic OA in the knee was defined according to the American College of Rheumatology (ACR) recommendations for knee OA16 as a combination of knee pain or stiffness on most days of the preceding month and osteophytes at the joint margins of the tibiofemoral joint. OA in the hand and hip was also defined according to the ACR recommendations.17 18 Prosthetic joints in hips or knees as a result of end stage OA were defined as OA in that particular joint. Degeneration of the spine (cervical and lumbar) was defined as pain or stiffness on most days of the preceding month in the spine in addition to a Kellgren–Lawrence score of 2 in at least one disc or one apophyseal joint.

Knee radiographs

The standardised non-fluoroscopic fixed-flexion protocol was used to obtain physical activity weight bearing radiographs of right and left knees at baseline and after 24 months.19 Uniform anatomical alignment of the knees was facilitated by the use of a specifically designed positioning frame that places the patient’s feet at 5 degrees external rotation and the knees and thighs in contact with the cassette and coplanar with the hips and tips of the great toes, resulting in a fixed knee angulation of approximately 20 degrees flexion.20 The x ray tube was angled to point 10 degrees downwards and the x ray beam centred to the knee joint line. A single experienced radiographer took all the radiographs using a fixed film-focus distance (1.30 m).

Assessment of radiological progression of knee OA

Medial and lateral tibiofemoral joint space narrowing (JSN) in both knees was graded (0–3) with the Altman atlas.21 A total score for JSN in the knees (range 0–12) was calculated by adding scores of right and left medial and lateral tibiofemoral compartments. All baseline and 24 month radiographs were scored by consensus opinion of two experienced readers (SB, IW). In cases of disagreement the lower, more conservative score was recorded. Films were blinded for patient characteristics and read with the radiographs of the same patient at both timepoints being assessed together without knowledge of the chronological order. The reproducibility for JSN grading, depicted by the intra-class correlation coefficient, was very good (0.88). The reproducibility was based on the repeat readings of a random sample of 20 radiographs, which were selected throughout the duration of the period of observation.

Radiological progression was predefined by at least a 1-score increase in JSN total scores after 2 years. Knees without end stage disease (grade 3) in terms of JSN at baseline, who received a total knee joint replacement during the follow-up period, were considered to have progression of JSN in that particular joint.

Whole blood stimulation system

Whole blood sample stimulation was performed as previously described.22 Briefly, blood samples were collected at baseline in pyrogen-free heparinised tubes (Endotube; Chromogenix, Mölndal, Sweden). Whole blood samples (8 ml) were diluted 1:1 with RPMI 1640 (Gibco Life Technologies, Paisley, UK) and stimulated with 10 ng/ml Escherichia coli LPS (Difco, Detroit, Michigan, USA). To minimise the influence of circadian rhythms and measurement errors, blood samples were obtained between 8:00 am and 11:00 am, the time frame between blood collection and stimulation was <1.5 h, and all stimuli were performed in the same endotoxin batch. One medium-diluted blood sample without LPS was used as a negative control. After 4-h and 24-h incubations, samples were centrifuged twice (at 600 g) and the supernatants were stored at −70°C. TNFα was measured in 4-h samples and IL1β, IL1Ra and IL10 production was measured in 24-h samples in one batch, by ELISA according to the manufacturer’s guidelines (Central Laboratory of the Blood Transfusion Service, Amsterdam, The Netherlands).

Statistical analyses

The difference in the innate ex vivo production (mean levels) of IL1β, TNFα, IL1Ra and IL10 between the patients with and without progression of JSN was calculated by linear mixed models in SPSS, V.11.0 (SPSS Inc, Chicago, Illinois, USA) with a random intercept to adjust for the familial effect within the sibling pairs/families. Estimates of fixed effects were reported with 95% CIs. The estimates represent the magnitude of the difference in the mean innate ex vivo production of IL1β, TNFα, IL1Ra and IL10 between the patients with and without progression of JSN.

Next, we compared the risk of having radiological progression over 2 years in the patients according to quartiles of the ex vivo production levels, based on the distribution in the patients. Logistic regression analyses were used to estimate crude and adjusted odds ratios (ORs) with corresponding 95% CIs. Adjustments were made for age, sex and body mass index. To take into account the intrafamily effect, robust standard errors were computed using the statistical program Stata, V.7.0 (Stata, College Station, Texas, USA). The adjusted ORs and 95% CIs were subsequently transformed to risk ratios (RRs) and corresponding 95% CIs using the approximation formula described by Zhang et al.23 This was performed because the ORs for common outcomes in a closed cohort are not good approximations of RRs.

RESULTS

Patient characteristics

Between August 2000 and March 2003, 191 sibling pairs with OA at multiple sites were included in the GARP study. Of these patients, 104 sibling pairs (208 patients) were included in the 2-year follow-up study. At baseline, 89 (42.8%) patients had symptomatic OA in at least one knee, of which 86 (96.6%) completed the 24-month follow-up and were included in the present study. Baseline characteristics of these patients are shown in table 1. The four patients with incomplete follow-up did not differ from those with complete follow-up (data not shown). The median age at baseline was 60.1 years and the majority of the patients (80.2%) were women. During the 24 months follow-up, four knees from four patients received a knee joint replacement. This resulted in progression of JSN in 29 (33.7%) of the 86 followed patients after 2 years. Baseline characteristics did not differ between patients with progression compared to patients without progression (table 1).

Innate ex vivo production of cytokines in patients with and without progression

Whole blood samples with complete data on levels of IL1β, TNFα, IL1Ra and IL10 were obtained at baseline in 85 patients with knee OA. The innate ex vivo production was higher in patients with than without progression for TNFα (respectively, median, IQR 9222 pg/ml, 6233–11468 vs 6662 pg/ml, 4586–8844) and for IL10 (respectively, median, IQR 867 pg/ml, 620–1148 vs 701 pg/ml, 505–878) (fig 1). The mean differences between patients with and without progression in ex vivo production of TNFα was 1992 pg/ml (95% CI 403 to 3583) (p = 0.012) and of IL10 was 166 pg/ml (95% CI 21 to 310) (p = 0.025).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Ex vivo production of interleukin (IL)1β, tumour necrosis factor (TNF)α, IL1 receptor antagonist (IL1Ra) and IL10 upon lipopolysaccharide stimulation in patients with non-progressive and progressive knee osteoarthritis (OA). Data are presented as box plots, where the boxes represent the 25th to 75th percentiles, the lines within the boxes represent the means and the lines outside the boxes represent the 5th and 95th percentiles. Asterisks indicate outliers.

The innate ex vivo production of IL1β and IL1Ra was not different in patients with progression of JSN than patients without (IL1β respectively, median, IQR 3582 pg/ml, 1905-4998 vs 2899 pg/ml, 2051-3791 and IL1Ra respectively, median, IQR 25034 pg/ml, 21215-33245 vs 22822 pg/ml, 18076–28118) (fig 1). The mean differences of innate ex vivo production of IL1β and IL1Ra were respectively 734 pg/ml (95% CI −301 to 1768) (p = 0.452) and 3956 pg/ml (95% CI −94 to 8005) (p = 0.062).

Risk of JSN progression according to innate ex vivo cytokine production

Table 2 shows the RRs for having JSN progression according to quartiles of cytokine production. Patients in the highest quartile of TNFα production had a sixfold increased risk of having JSN progression (adjusted RR 6.4, 95% CI 1.5 to 9.8) and patients in the highest quartile of IL10 production had a fourfold increased risk of JSN progression (adjusted RR 4.3, 95% CI 1.7 to 6.3), both in comparison with those patients in the lowest quartile. No significant associations were found between variations in IL1β and IL1Ra production and JSN progression.

In additional analyses (data not shown), adjustment for baseline JSN radiological score and baseline WOMAC score did not substantially change the results.

DISCUSSION

This prospective cohort study, among middle-aged patients with symptomatic OA of the knee, provides the first evidence that genetic differences in TNFα and IL10 production are associated with radiological progression of knee OA. A sixfold increase in risk for radiological progression of JSN, reflecting articular cartilage loss, was found in patients in the highest quartile of ex vivo production of TNFα compared to patients in the lowest quartile of TNFα production. Patients in the highest quartile of ex vivo IL10 production had a fourfold increased risk of radiological progression of JSN. These associations were not caused by differences in age, sex or BMI. Innate ex vivo production of IL1β and IL1Ra was not associated with progression of knee OA over 2 years.

In the present study, all patients included in the GARP study with knee OA at baseline were followed after 2 years. This follow-up study differs from a previous case control study comprising 305 patients from the GARP study at baseline with OA at multiple sites and 137 healthy controls.14 In the latter study, the role of inheritable differences in cytokine production upon stimulation in the development of OA was shown. Patients with a high innate ex vivo production of IL1β and IL1Ra and low innate ex vivo production of IL10 had an increased risk of having OA at multiple joint sites. The innate ex vivo production of TNFα was not different in the patients with OA in comparison to the healthy controls. These results suggest that the role of cytokines in OA susceptibility could be different from their role in OA progression, which is in line with other risk factors that are associated with OA.2

TNFα plays an important role in cartilage damage in rheumatoid arthritis, where competitive inhibition of TNFα with anti-TNFα antibodies has been found to improve symptoms and prevent progression of joint destruction.24 The role of TNFα in OA is less clear, but in vitro and animal studies suggest similar actions.25 26 Since the pathophysiological pathways that contribute to the rate of cartilage loss in OA are not clear, one could only speculate on the mechanisms by which TNFα may influence the rate of progression of OA. TNFα led to upregulation of MMP synthesis27 and production of NO by chondrocytes, inducing cartilage catabolism and chondrocyte apoptosis. Furthermore, it may lead to synovial membrane inflammation, which have been associated with an increased risk for progression in patients with knee OA.28 29 How the innate ex vivo production of TNFα translates to concentrations of TNFα in the serum and synovial fluid of patients with knee OA is not clear.

The observation that a high innate ex vivo production of IL10 is associated with a increased risk for progression of knee OA is in line with results from a longitudinal study in patients with rheumatoid arthritis, showing that high levels of IL10 are associated with a high rate of progression of joint damage.30 In rheumatoid arthritis, this could be explained by increased autoantibody titres in patients with high IL10 production genotypes. In OA, it is difficult to indicate the pathophysiological mechanism involved. IL10 has an inhibitory function, downregulating the production of IL1β, TNFα and MMPs as well as inhibiting the release of PGE2. This would suggest a downregulatory effect on joint destruction in OA, which makes it an unlikely explanation. Further investigation of the role of IL10 in OA progression is required.

A potential limitation of the present study was the possibility to misclassify patients for the presence or absence of progression of OA. Currently, there are no recommendations for the assessment of progression of knee OA on an individual basis. We defined radiological progression of knee OA in a patient as a 1-score increase in JSN total scores in the right and left knees. The appearance of JSN in a knee that were unaffected at baseline as well as progression of JSN in a baseline knee with OA were both considered progression in a patient with knee OA at baseline. Any misclassification of progression, however, will be non-differential because it was assessed blinded to and therefore unrelated to the ex vivo cytokine production. Such non-differential misclassification will tend to dilute the effect, producing estimates of the effect that are closer to a no-effect value than the actual effect.

In conclusion, our study provides the first evidence that the innate capacity to produce TNFα and IL10 associates with progression of knee OA, even over a relatively short follow-up period of 2 years. These findings are in line with results from rheumatoid arthritis studies that show that TNFα and IL10 play an important role in progression of cartilage damage. These results, if confirmed, imply that cytokine pathways might be appropriate pathophysiological targets for OA therapy. To increase our understanding of the complex mechanisms driving OA progression, further research is required on the role of cytokines in cartilage degradation in OA.