The Ku antigen is a heterodimer of 70 kDa (p70) and 80 kDa (p80) protein subunits involved in DNA repair, immunoglobulin gene recombination, transcription regulation, modification of steroid action and development of the brain.1^–4

Anti-Ku autoantibodies were originally described in nine individuals with various connective tissue diseases, studied with a classic immunodiffusion assay. Of these patients, 6 belonged to a subgroup of 11 individuals with polymyositis/systemic sclerosis (PM/SSc) overlap syndrome.5 Hirakata et al6 reported on 21 consecutive anti-Ku Japanese patients, all of whom were positive for human leukocyte antigen (HLA) DPB1*0501, whereas DRB1*0901-DQA1*0302-DQB1*0303 haplotype also correlated with PM/SSc overlap syndrome. O’Hanlon et al7 found DRB1*11 and DQA1*06 to be risk factors for patients that are anti-Ku positive associated with idiopathic inflammatory myositis. In the study of Wang et al8 the prevalence of anti-Ku antibodies showed racial differences, being relatively common in African–American patients with systemic lupus erythaematosus (SLE), but rare among Caucasian patients.

Ho et al9 described 159 anti-Ku positive patients so far with the following diagnoses: overlap syndrome in 34%, SLE in 28%, idiopathic inflammatory myositis in 4%, SSc in 14% and other diseases in 20%. Among the latter, primary pulmonary hypertension10 and Graves disease were especially emphasised, in which p70 was simultaneously identified to be identical to a thyroid related autoantigen.11 Of the patients with overlap syndromes, nearly 65% had clinical features of SSc.12 Common features of anti-Ku positive patients described by Cooley et al13 and Francescini et al12 (reported in undifferentiated connective tissue diseases, PM/SSc, SLE/SSc/PM and SLE/PM overlaps; SSc, primary Sjögren syndrome, SLE and rheumatoid arthritis) were skin thickening, Raynaud phenomenon, muscular and joint involvement and oesophageal reflux. Rebora and Parody14 suggested that anti-Ku antibodies define a group of patients with the same clinical symptoms as mixed connective tissue disease and good overall prognosis. Anti-Ku positive patients often have steroid sensitive myositis4 and Troyanov et al15 recently suggested that anti-Ku antibodies are markers of monophasic myositis in overlap myositis syndromes.

In patients with SSc a large variety of different autoantibodies (such as Scl70, anti-centromere antibodies (ACA), PM-scleroderma (Scl), anti-RNA polymerase antibodies) may be found, each one being useful in the diagnosis and prognosis of affected patients. Previous studies on the clinical association of anti-Ku antibodies in patients with SSc are limited by small sample numbers.16^–18 Several case reports have also been published,12 13 19^–25 which have contributed to further understanding the clinical aspects of the association of anti-Ku antibodies with a variety of connective tissue diseases, including SSc overlap syndromes, but to date a large scale survey has not been conducted.

Therefore, in the present work we aimed to determine the prevalence of anti-Ku antibodies in patients with SSc and to evaluate their clinical and serological characteristics in a large cohort of patients from six European rheumatological centres initiated by EUSTAR (European League Against Rheumatism (EULAR) Scleroderma Trials and Research). A case–control study of anti-Ku positive and anti-Ku negative patients with SSc was performed.

MATERIALS AND METHODS

Sera from 625 consecutive patients who fulfilled the American College of Rheumatology (ACR) criteria for SSc26 and LeRoy et al27 classification criteria for limited cutaneous (lc) and diffuse cutaneous (dc) SSc were collected from 6 European centres (Vienna, Austria; Pécs, Hungary; Florence, Italy; Niš, Serbia; Ljubljana, Slovenia; and Zurich, Switzerland), all of which are members of EUSTAR (the material is derived from the EUSTAR initiative and from EUSTAR centres). However, the large number of samples and patients could only be achieved by including all stored samples in biobanks from the participating centres. These included samples obtained before EUSTAR was initiated; the sera collection was obtained with a timespan ranging up to 15 years. Unfortunately, it was not feasible to obtain the full set of baseline data from all centre patients due to the large sample number and due to the fact that part of the patient group was assessed before EUSTAR was initiated.

The centres contributed between 69 and 145 sera. Stored sera were sent on dry ice to the reference centre (Department of Rheumatology, Ljubljana, Slovenia) where they were tested. Antibodies against soluble extractable nuclear antigens (anti-ENA) including anti-Ku antibodies were screened by counterimmunoelectrophoresis (CIE) as described previously.28

After the laboratory evaluation of sera, anti-Ku positive patients with SSc (n = 14) and all age- and sex-matched anti-Ku negative control patients with SSc (n = 43) from the same cohort were selected. Clinical data were derived from the databases of each centre. The EUSTAR minimal essential data set was used including definitions for each clinical parameter given there.29 Sera from both groups were additionally tested in the same laboratory for anti-nuclear antibodies (ANA) by indirect immunofluorescence on Hep-2 cell line substrate (Immunoconcepts, Sacramento, California, USA) and by line immunoassay (LIA) for autoantibodies present in systemic myositis, which besides anti-Ku also includes antibodies against Jo-1, Mi-2, PM/Scl and U1-small nuclear ribonucleoprotein (snRNP) (Imtec-Myositis LIA, Imtec, Immunodiagnostika, Berlin, Germany).

Statistical evaluation

SPSS V.14.0.2 (SPSS, Chicago, Illinois) was used for statistical analysis. Regarding the age/sex matching, we determined unbalanced numbers of control patients (table 1). To compare clinical data between anti-Ku positive and control patients using all information available for control patients, a biostatistician (DH) calculated bootstrap CIs for the clinical data of controls. Bootstrap 95% CIs were computed with 1000 drawings of 14 age/sex-matched controls. Rates of clinical parameters of anti-Ku positive patients not lying in the confidence interval of control patients were considered significant.

RESULTS

The study was divided into two phases: in the first phase, screening of all sera from patients having either lcSSc or dcSSc for anti-Ku and other anti-ENA antibodies by CIE was performed. In the second phase, case–control evaluation of anti-Ku positive vs negative patients to identify clinical associations of anti-Ku antibodies in this large cohort of clinically well defined patients with SSc was undertaken.

Prevalence of anti-ENA antibodies in patients with SSc

Autoantibodies against ENA of all 625 patients are summarised in table 2 and are specified for each participating centre. Anti-ENA were positive in the range from 36.0 to 64.8% of all patients. Anti-Scl70 antibodies ranged from 9.5% to 30%, anti-U1-snRNP from 1.4% to 15.9% and anti-PM/Scl from 0 to 6.1% of all patients. The differences between the minimum and maximum frequencies of anti-Scl70 and anti-U1-snRNP autoantibodies were statistically significant in individual centres.

Prevalence of anti-Ku antibodies in patients with SSc

Overall, anti-Ku antibodies were found in 14/625 patients with SSc. This corresponds to a prevalence of 2.2%, ranging from 0 to 7% in individual centres in this large and unselected population of patients with SSc. Anti-Ku antibodies were the only anti-ENA antibodies found in 10 patients with SSc. In the remaining four patients with SSc, anti-Ku antibodies were combined with anti-Ro/La in two patients, and with anti-U1-snRNP and anti-Scl70 in one patient each.

Longitudinal follow-up of anti-Ku antibody titres

Follow-up investigations on a subset of four anti-Ku positive patients of long disease duration (5 years and above) revealed changes in anti-Ku titres, which fluctuated without any evident association to clinical events or therapy (data not shown).

Clinical and laboratory characteristics of anti-Ku positive patients with SSc

To identify clinical characteristics associated with the presence of anti-Ku antibodies in patients with SSc, a case–control strategy was chosen. All available age- and sex-matched controls and patients with SSc positive for anti-Ku antibodies were chosen to calculate bootstrap confidence intervals (table 2). Those clinical features lying outside the confidence intervals of the control patients were considered to be significantly associated with anti-Ku antibodies.

The clinical and laboratory features of 14 anti-Ku positive patients and 43 SSc controls are summarised in table 3. All but two anti-Ku positive patients were of the lcSSc subset. Musculoskeletal manifestations were significantly more common in the anti-Ku positive subset. In particular, synovitis and joint contractures were significantly associated with the presence of anti-Ku antibodies. A particularly strong and significant association of anti-Ku antibodies was found with the presence of muscle weakness/atrophy as well as with the elevation of muscle enzymes. Pathological electromyogram (EMG) results were obtained in 40% of anti-Ku positive patients, while no EMG analysis was performed in the control patients. Similarly, muscle biopsies confirmed the diagnosis of myositis in 40% of anti-Ku antibodies, while no biopsies needed to be performed in the control group. There may have been a selection bias in the way EMG and muscle biopsy were performed in different centres. However, usually they are performed for either/both of pronounced muscle weakness and elevated creatine phosphokinase (CK). Taken together, the presence of anti-Ku antibodies in patients with SSc was strongly associated with synovitis, joint contractures and clinical features of myositis. In two of the anti-Ku positive cases myositis was evidently monophasic, whereas it appeared more chronic in others, two cases of which were very severe, demanding high dose glucocorticoids and immunosuppressive therapy.

By contrast, anti-Ku positive patients were negatively associated with vascular manifestations of the disease (table 3). Significantly fewer patients with anti-Ku antibodies had teleangiectasias and fingertip ulcers than control patients. No difference was found for the frequency of pulmonary hypertension as diagnosed by echocardiography. Arterial hypertension was less frequent in anti-Ku positive patients, although no renal crisis was evident in any of the controls and anti-Ku positive patients.

All anti-Ku positive patients and 37 of the 43 controls were additionally tested by Imtec-Myositis LIA (see Materials and methods). Of 14 anti-Ku positive patients in the CIE, 11 were also positive by LIA. In addition, by LIA, one anti-Ku positive serum was also positive for autoantibodies against Mi-2 and one serum showed antibodies against PM/Scl in the control group (data not shown). The slightly different results for anti-Ku detected by both methods could be explained by different sources of antigen (native vs recombinant Ku antigen), so we considered all anti-Ku positive patients suitable for clinical comparisons (table 3).

Strong differences were found between anti-Ku positive and anti-Ku negative patients with SSc regarding the presence of autoantibodies. Anti-Ku positive patients had a strikingly lower percentage of punctate nucleolar immunofluorescence patterns on Hep-2 cells. Most interestingly, anti-Scl70 antibodies were rarely detectable in anti-Ku positive patients and ACA were not found in a single anti-Ku positive patient with SSc. The anti-Ku positive, Scl70 positive patient was a female with lcSSc, with diffuse lung fibrosis, oesophageal involvement, malabsorbtion, digital ulcers and mild myositis as the major clinical manifestations of disease in that case.

The serological differences indicated above were confirmed by the statistical analysis using the 95% bootstrap CIs derived from the control group. The anti-Ku positive patient group had a lower prevalence of anti-Scl70 antibodies, ACA and nucleolar/punctuate nucleolar immunofluorescence patterns than their anti-Ku negative counterparts.

DISCUSSION

The function of Ku antigen was extensively studied.30^–35 Ku antigen could even provide an explanation for the pathological immunity associated with parvovirus B19 infection.36

The specific objective of the present study was to evaluate the frequency of anti-Ku antibodies in a large cohort (n = 625) of patients with SSc from six different European academic rheumatology centres, and to compare the features of anti-Ku positive patients with SSc with those of an age- and sex-matched Ku-negative group. As a limitation of the study, baseline data could not be obtained for all patients. However, all patients were derived from scleroderma reference centres and, thus, the overall population is very likely to reflect the standard population known from other SSc studies performed at university centres in Europe.

In this large population, anti-Ku autoantibodies were found in the sera of 14 patients (2.2%). It is difficult to compare the results of the present study with those of other investigators, due to different study designs and patient selection criteria. However, the closest comparison would be represented by the study of Bunn et al,37 who found 9/735 anti-Ku positive UK patients with SSc (1.2%), even though their study was not focused on anti-Ku antibodies in particular and lacked the clinical details of anti-Ku positive patients. In a small series of patients with SSc and overlap syndromes, reported frequencies of anti-Ku positive patients varied from 0 to 40%.16^–19 22 37 38 There is no established standard laboratory test for anti-Ku; immunodiffusion, immunoprecipitation, antigen capture ELISA, fusion protein ELISA and Western blot have been used in the past, which have been extensively discussed by Reeves et al.39 So far, no consensus has been reached about a preferred particular method of detection. In the present study, we used two different methods (CIE and LIA) enabling detection of antibodies to native and/or recombinant antigens.

In our study, the frequency of anti-Ku positive patients differed from 0% to 7% among individual centres. As essentially all patients are Caucasian, it therefore appears that anti-Ku antibodies may be more frequent in Central and Eastern European countries.

This is the first report identifying clinical associations of anti-Ku antibodies in a large SSc population. A strength of the study was the defined statistical approach using bootstrap confidence intervals derived from an anti-Ku negative control population.

Our data add important information for the clinical features of anti-Ku positive patients with SSc and may play a role in the care of patients with SSc. First, patients with SSc with anti-Ku antibodies have an increased risk to develop musculoskeletal manifestations, such as arthritis, joint contractions and most notably the signs of myositis, and need to be screened accordingly. Second, by contrast, vascular manifestations such as fingertip ulcers are less likely to occur and expensive preventive treatments such as endothelin receptor antagonists might be less frequently necessary than in anti-Ku negative patients. Notably, however, the negative association of anti-Ku antibodies with vascular manifestations was not maintained for echocardiographic evidence of pulmonary hypertension, which was equally frequent between anti-Ku negative and anti-Ku positive patients. Third, as anti-Ku antibodies were largely exclusive with other SSc-specific autoantibodies such as anti-Scl70 autoantibodies and ACA,40 screening for anti-Ku antibodies might be particularly useful in patients who are positive for ANA, but negative for the above mentioned SSc-specific antibodies. However, for daily clinical care of patients it has to be considered that this study showed that the prevalence of anti-Ku antibodies is rather small among patients with SSc and can more easily be followed according to associated conditions such as myositis and arthritis.

In conclusion, this study of 625 patients with SSc is the largest cohort to date (due to a multi-centre approach of centres specialised in SSc) focusing on the prevalence of anti-Ku antibodies. The case–control approach was able to demonstrate a clinically distinct subset of anti-Ku positive patients with SSc with prominent signs of myositis, while vascular manifestations such as fingertip ulcers were less frequent. The absence of SSc-specific autoantibodies such as ACA and anti-Scl70 antibodies in anti-Ku positive patients support the concept of a clinically and serologically distinct subset of patients with SSc.