High prevalence of anti-CD74 antibodies specific for the HLA class II-associated invariant chain peptide (CLIP) in patients with axial spondyloarthritis
- 1Rheumazentrum Ruhrgebiet Herne, Herne, Germany
- 2Department of Immunology and Rheumatology, Medical University Hannover, Hannover, Germany
- Correspondence to Dr J Braun, Rheumazentrum Ruhrgebiet, Ruhr-University Bochum, Landgrafenstr 15, Herne 44652, Germany;
- Accepted 14 April 2013
- Published Online First 3 May 2013
Objectives The pathogenesis of axial spondyloarthritis (axSpA) is still unclear. There is a strong association with HLA-B27 and other genes. Recently, anti-CD74 antibodies with specificity to a class II-associated invariant chain peptide (anti-CLIP-ABs) were found in axSpA patients. We examined the prevalence, sensitivity and specificity of anti-CLIP-ABs in axSpA in comparison with controls.
Methods Sera of axSpA and non-SpA patients were analysed for IgG-antibodies against CD74 by ELISA with specificity for CLIP developed in cooperation with AESKU Diagnostics (Germany). A cut-off of ≥4 SDs of arbitrary units (AU) from mean serum levels was used to differentiate the results. The laboratory workers were completely blinded for clinical data.
Results We analysed 145 sera from 94 axSpA and 51 non-SpA patients. AxSpA patients were more often male and younger. HLA-B27 status was available in 72 patients. Anti-CLIP-ABs were detected in 85.1% in axSpA but in only 7.8% in non-SpA patients (p≤0.0001). AxSpA patients showed higher levels of anti-CLIP-ABs versus non-SpA: mean 14.5 versus 0.8 AU (p≤0.0001). The sensitivity of anti-CLIP-ABs for diagnosing axSpA was 85.1%, specificity 92.2%, likelihood ratio (LR) LR+ was 10.8 and LR− was 0.08. Anti-CLIP-ABs and HLA-B27 were positive in 87.5% patients with axSpA, but only 14.9% were anti-CLIP-negative, while 23.6% were HLA-B27-negative.
Conclusions Anti-CLIP antibodies were strongly associated with axSpA. The LR for confirming axSpA by using anti-CLIP was even higher than by using HLA-B27. More studies using this promising new method in patients with non-radiographic axial SpA or peripheral SpA are needed to establish its usefulness in clinical practice.
The partly heterogeneous group of patients with spondyloarthritis (SpA) consists of axial SpA (axSpA) including ankylosing spondylitis (AS) but also non-radiographic axSpA, which takes into account different SpA features such as psoriatic arthritis, reactive arthritis, SpA associated with chronic inflammatory bowel disease and other cases with uncertain symptoms.1 ,2 AxSpA, including its major subtype AS, affects the axial skeleton and is associated with characteristic clinical features such as sacroiliitis, uveitis and enthesitis. The prevalence of SpA has been estimated between 0.5% and 2%.3 ,4
Radiographs of the spine that may show characteristic structural changes and sacroiliitis as detected by MRI are important tools for diagnosis and classification.5 It may take several years after the onset of the first symptoms until definite x-ray changes will occur.6 Since chronic back pain is a rather frequent symptom and the sensitivity and specificity of inflammatory back pain are limited, the early diagnosis of axSpA may be a clinical challenge, and different strategies for early referral have been proposed.7 As a matter of fact, a considerable delay of 7–10 years between the onset of symptoms and the diagnosis of AS has been reported.8 Furthermore, there are several other diseases which may be relevant for the differential diagnosis of axSpA including osteoarthritis, osteoporosis or mechanical back pain.9 In conclusion, another test or biomarker that could potentially identify patients with SpA and serve as a reliable tool for differential diagnosis may be useful.
The main laboratory test to support a diagnosis of SpA is HLA-B27, which has a good sensitivity and specificity, but it is also found in up to 10% of healthy individuals.2 The likelihood ratio (LR) for a positive HLA-B27 for classification to AS has been calculated at 9.0,10 similar to the LR for a positive MRI. This was the reason why these two features were considered as main entry criteria for classification to axSpA.5
CD74 was shown to be involved in the assembly of major histocompatibility complex II and in preventing premature binding of peptides to it.11 ,12 The role of CD74 in B cell differentiation is less clear.13 The human CD74 is also known as HLA class II γ-chain or invariant chain. The extracellular part (aa 73–296) has two different domains, thyroglobulin type-1 and class II-associated invariant chain peptide (CLIP). Binding of CLIP antibodies to CD74 may lead to activation of cells and production of proinflammatory cytokines such as tumour necrosis factor α (TNFα).14 Baerlecken et al have described the function of these peptides as potential factors in the pathogenesis of axSpA in an accompanying paper.15 Indeed, anti-CD74 antibodies are not the first autoantibodies described in SpA: the first report on potential autoantibodies in AS appeared in 197916 but their existence was denied in other studies.17 Later, several autoantibodies were claimed to be associated with SpA but no test was finally found useful for diagnostic purposes because of low sensitivity and specificity.18–21
The aim of this study was to confirm the usefulness of anti-CLIP antibodies to detect patients with axSpA in patients seen in daily clinical routine.
A total of 145 sera obtained from 94 patients with axSpA (89 with AS, 85 of which were participants in the European AS Infliximab Cohort (EASIC)22 and additional five with non-radiographic ax-SpA) and 51 from patients with other diseases (13 with rheumatoid arthritis, 17 with fibromyalgia and 17 with degenerative spinal disease, three with psoriatic arthritis without axial involvement, and one with polymyalgia rheumatic) were analysed to detect IgG antibodies against CD74 using an ELISA with specificity for CLIP. All patients had given written informed consent for use of their sera in the study. Patients from the EASIC study had already agreed at the time point of participation (by agreeing to the use of their sera for future genetic examinations) and consecutive patients agreed after information about the type and purpose of the current examination.
The ELISA was developed by the medical university of Hannover in cooperation with AESKU Diagnostics (Wendelsheim, Germany), as described by Baerlecken et al15 and according to the manufacturer's protocol. Samples were stored for at least 6 months in –80° temperature. Due to the nature of the study, EASIC samples were stored longer than the samples from non-EASIC patients; the latter include patients with axSpA and controls. A cut-off of ≥4 SDs of arbitrary units (AU) from the mean serum level of autoantibodies was used to qualitatively differentiate between positive and negative results. The collaborators in the laboratory (NB and TW) were blinded to the clinical diagnosis of the patients. The study was approved by the local ethical committee of the medical university of Hannover, Germany (approval number 1315–2012).
Clinical and laboratory measures for disease activity, such as the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI23) and erythrocyte sedimentation rate or C reactive protein (CRP), and for disease function Bath Ankylosing Spondylitis Functional Index24 were measured in axSpA patients. For non-SpA patients, only information on CRP was available.
Comparisons between groups were evaluated using the Mann–Whitney U test. For correlations we used the Spearman test. A p<0.05 was regarded as statistically significant. Statistical analysis was performed with SPSS 20.0.
The mean age was 43.0±11.0 in patients with axSpA and 57.7±14.5 in patients with non-SpA, while there were 81.9% male patients in the axSpA and 27.5% male patients in the non-SpA group (both p<0.001). CRP values were 0.6±0.6 mg/dl in patients with axSpA and 0.5±1.3 mg/dl in patients with non-SpA (p<0.001). The mean BASDAI in patients with axSpA was 6.4±1.4. Information on HLA-B27 was available in 72/94 patients (76.6%) with axSpA, 56 of which were positive (77.8%). There was no difference in the baseline characteristics between patients with and without knowledge of the HLA-B27 status.
There were significant differences between HLA-B27-positive and HLA-B27-negative patients: the mean age at diagnosis was 29.0±9.4 years in HLA-B27-positive patients and 39.8±10.2 years in HLA-B27-negative patients (p=0.003). Similarly, the mean age at inclusion into the study was 42.5±11.1 years in HLA-B27-positive and 49.5±9.4 years in HLA-B27-negative patients (p=0.013). No other differences between HLA-B27-positive and -negative patients were found. In contrast, no such differences were seen when comparing anti-CLIP-positive with -negative patients.
Prevalence of anti-CLIP antibodies in patients with axSpA in comparison with patients with other diagnoses
Significantly increased levels of anti-CLIP antibodies were detected in patients with axSpA (n=80/94, 85.1%) versus non-SpA patients (4/51, 7.8%), p≤0.0001. The diagnoses of non-SpA patients with significant anti-CLIP levels were two with degenerative spinal disease and one for rheumatoid arthritis and psoriatic arthritis each. Furthermore, higher levels of anti-CLIP antibodies were detected in patients with axSpA (mean 14.5 AU) as compared with non-SpA patients (0.8 AU), p≤0.0001.
A slightly higher proportion of male patients with axSpA had anti-CLIP antibodies (67/77, 87%) in comparison with female patients (13/17, 76.5%). The mean anti-CLIP levels were also somewhat higher in male versus female patients (14.9 vs 12.9 AU). However, both differences were not significant.
Sensitivity, specificity of anti-CLIP and calculation of the LRs and predictive values for anti-CLIP in the diagnosis of axSpA
The sensitivity of anti-CLIP for a diagnosis of axSpA was calculated at 85.1% with a specificity of 92.2% (table 1) while the positive LR for anti-CLIP antibodies to diagnose axSpA was 10.9 and the negative LR was 0.08. In comparison, the sensitivity of HLA-B27 was 77.8%.
Overall, a similar proportion of axSpA patients (49/72, 68.1%) was positive for HLA-B27 and anti-CLIP antibodies in parallel, while 9/72 (12.5%) patients were negative for anti-CLIP antibodies and 17/72 (23.6%) patients were negative for HLA-B27.
This study shows that antibodies against CD74 with specificity against CLIP have a high prevalence in patients with axSpA but not in controls. Thus, this paper, which is an accompanying paper to the publication by Baerlecken et al,15 clearly suggests that this test may be of clinical relevance providing promising values for sensitivity and specificity, and positive and negative LRs in the diagnosis of axSpA. The question whether it may contribute to a diagnosis especially in HLA-B27-negative patients must be definitely answered in future prospective studies. In this study, there was some evidence that anti-CLIP antibodies were detected in HLA-B27-negative patients. When confirmed, this could improve an early diagnosis of axSpA. Furthermore, especially related to the results of first studies on this topic,15 there is also a need to study the possible differences between early and more advanced disease stages. In the present study, the prevalence of anti-CLIP antibodies was clearly not low in more advanced disease stages. A prospective study on this topic should give a clear answer on this, as yet, unclear observation.
The main laboratory test used to support a diagnosis of axSpA is very well known for some decades and is frequently used nowadays: it is usually performed to show the presence of the major histocompatibility complex class I molecule HLA-B27. This can be performed by the classical cytotoxicity method, flow cytometry or PCR.25 ,26 The costs for the test vary but are in general relatively high between €20 and €70. The sensitivity of HLA-B27 in the diagnosis of axSpA has been reported between 60% and 95%, largely depending on the selection of the population.27 ,28 Concentrating on patients with classical AS usually gives the highest proportion of HLA-B27+ patients. Naturally, HLA-B27 is found in up to 10% of healthy individuals largely depending on the geographic location.2 However, this does not have a major influence on the specificity of the test (only 10% false positive). Thus, the LR for a positive HLA-B27 for classification to AS has been calculated at 9.0.10 In our cohort, which mainly consisted of AS patients, only a few patients had been diagnosed as non-radiographic axSpA, and the sensitivity of anti-CLIP-ABs was 86% versus 78% for HLA-B27, with a specificity of 92.5%. Overall, while 88% of axSpA patients were positive for anti-CLIP antibodies or HLA-B27, about 25% patients were found to be HLA-B27-negative but only 13% were negative for anti-CLIP-ABs. These data suggest a very good performance of this test for supporting a clinical diagnosis of axSpA. Whether the combination of anti-CLIP-ABs with other clinical parameters of axSpA, such as a positive MRI of the sacroiliac joints, or other parameters will lead to even higher LRs and whether it can even be used as a diagnostic tool for early stages of axSpA needs to be evaluated in future studies with patients of all axSpA subgroups. This will be particularly interesting in HLA-B27-negative SpA patients.
Overall, anti-CLIP antibodies were detected in the majority of axSpA patients (85%) but in only <10% of non-SpA patients. In addition, the frequency and the serum levels of anti-CLIP antibodies were found to be significantly higher in patients with axSpA as compared with non-SpA.
There were no correlations of anti-CLIP antibodies with demographic or clinical data in our study. This finding could have been expected because of the overall high prevalence of the antibodies. Of interest, in the French cohort DESIR some clinical differences between HLA-B27-positive and -negative patients have been reported.29 This pattern of differentiation of subgroups of patients was similar in our study—with patients with positive HLA-B27 being younger at the time point of diagnosis. It remains to be seen whether similar differences can be found in anti-CLIP-positive and -negative patients with axSpA. Furthermore, it will be interesting to study the potential influence of disease activity on these antibodies since this is known to fluctuate in some patients.30 Many AS patients in our cohort had been treated with anti-TNFα therapy; thus, we cannot exclude that this may have influenced the results. Whether anti-CLIP antibodies may predict clinical response to anti-TNFα-treatment, as it is has been reported for HLA-B27,31 needs further study. Further study is also needed to determine the occurrence and prevalence of anti-CLIP antibodies in patients with peripheral SpA and, especially, psoriatic arthritis.
The occurrence of autoantibodies suggests a potential involvement of B cells in the pathogenesis of axSpA. There is limited evidence from experimental studies32–34 that B cells are present at the site of inflammation in SpA. However, there are two studies that suggest a limited efficacy of the anti-CD20-antibody rituximab in patients with AS.35 ,36 Whether anti-CLIP antibodies can predict a therapeutic response to B cell depleting therapies is not known.
In conclusion, anti-CLIP antibodies seem to be strongly associated with axSpA including AS. Of interest, the LR for a diagnosis of AS calculated for anti-CLIP antibodies was even higher than for HLA-B27. More studies using this promising new method in patients with non-radiographic axSpA or peripheral SpA are needed to establish its usefulness in clinical practice. The test may also be relevant for improving the rates of early diagnosis and identification of SpA patients in primary care, in particular in the differentiation of SpA from non-inflammatory causes of back pain.
Handling editor Tore K Kvien
Contributors XB, FH and JB: Coordination and collection of blood samples, statistical analysis, and writing of the manuscript. NB and TW: Blinded analysis of the samples in the laboratory and review of the manuscript.
Competing interests None.
Patient consent Obtained.
Ethics approval Ethical Committee of the Medical University of Hannover, Germany.
Provenance and peer review Not commissioned; externally peer reviewed.