Anti-topoisomerase II α autoantibodies in systemic sclerosis—association with pulmonary hypertension and HLA-B35

Abstract

We have previously detected autoantibodies against topoisomerase II α (anti-topo II α) in sera from patients with idiopathic pulmonary fibrosis. To determine whether anti-topo II α is also present in systemic sclerosis (SSc) patients with pulmonary involvement, we screened sera from 92 patients and 34 healthy controls. Presence of anti-topo II α was investigated with respect to clinical and serological features, including the frequencies of HLA class I and II alleles. Anti-topo II α was detected in 20/92 (21.7%) patients. No association was found with either anti-topoisomerase I (Scl-70 or anti-topo I) or anti-centromere antibodies. However, anti-topo II α was associated with the presence of pulmonary hypertension (PHT) (as opposed to pulmonary fibrosis), and with a decrease of carbon monoxide diffusing capacity. Anti-topo II α was strongly associated with the presence of the class I antigen HLA-B35. No significant association was found with HLA class II antigens. HLA-B35 also turned out to be associated with the presence of PHT. These results indicate that in SSc patients, the presence of anti-topo II α is associated with PHT, and that the simultaneous presence of HLA-B35 seems to add to the risk of developing PHT.

INTRODUCTION

Systemic sclerosis (SSc), otherwise known as scleroderma, is a systemic autoimmune disease characterized by microvascular damage and cutaneous and visceral fibrosis [1]. A prominent immunological abnormality found in SSc is the presence of serum antinuclear antibodies, which are detectable in the majority of cases [1]. The autoantigens to which these antibodies bind include DNA topoisomerase I (formerly termed Scl-70) [2, 3] centromere/kinetochore [4], RNA polymerase I, II and III [5, 6], heterogeneous nuclear ribonucleoprotein [7] and histones [8]. Anti-fibrillarin and anti-Th RNP are also specific for SSc, but are rarely detected [9, 10]. Each distinct SSc-related antinuclear antibody is closely associated with its own unique combination of clinical features, suggesting that the production of these types of antibody may be related to fundamental disease processes [11].

Pulmonary involvement is a common clinical feature of SSc. A high percentage of patients with the diffuse cutaneous form of SSc (dcSSc), and a rather lower percentage of patients with limited cutaneous SSc (lcSSc), develop a pulmonary interstitial lung disease with histological and physiological characteristics similar to those found in idiopathic pulmonary fibrosis (IPF) [12].

In a previous study, we detected autoantibodies against the DNA topoisomerase II α (anti-topo II α) in the sera of about one-third of IPF patients [13]. Although anti-topo II α may be useful for the diagnosis of IPF, its presence did not correlate with indices of disease severity [14]. Anti-topo II α is not confined to IPF, since it has also been detected in rather lower percentages of patients with other pathologies, such as juvenile rheumatoid arthritis [15], insulin-dependent diabetes mellitus [16], and systemic lupus erythematosus [17].

In the present study, we examined the occurrence of anti-topo II α in SSc patients, about 20% of whom turned out to be positive for anti-topo II α. Associations were identified with certain clinical and immunogenetic features, including the presence of pulmonary hypertension (PHT) and the class I antigen HLA-B35.

PATIENTS AND METHODS

Patients

Sera were obtained from 92 Italian SSc patients, all of whom satisfied the preliminary classification criteria of the American College of Rheumatology [18]. Disease duration was calculated from the time of diagnosis of SSc. According to the extent of skin involvement, patients were classified as having lcSSc or dcSSc [19]. Demographic and clinical features of the patients are listed in Table 1. The presence of pulmonary fibrosis was assessed by chest high-resolution computed tomography (CT). Pulmonary function tests were performed on a body plethismograph Jaeger Body Screen while testing for single breath carbon monoxide diffusing capacity (DLco) on a Transfer Screen Jaeger. The results were calculated as percentages of the predicted normal value [20]. Diagnosis of PHT was based on the presence of tricuspid regurgitation, which allowed reliable measurement of pulmonary pressure by continuous wave Doppler [21]: PHT was considered to be present when the systolic pressure gradient across the tricuspid valve increased to ≥ 30 mmHg. Sera were also obtained from 34 healthy donors (28 females and six males, age 40 ± 1·5 years (mean ± s.e.m.)). All patients gave informed consent to participate in the study, which was approved by the institutional Review Board.

Table 1.

Demographic and clinical characteristics of the systemic sclerosis patients

Characteristic
No. of patients	92
Sex, M:F	9:83
Age (years) at diagnosis, mean ± s.e.m.	42·9 ± 1·3
Disease duration (years), mean ± s.e.m.	9 ± 0·66
Raynaud's phenomenon, duration (years), mean ± s.e.m.	34·7 ± 1·6
Teleangectasias	82·6%
Calcinosis	46·7%
Distal bone resorption	39·1%
Oesophageal involvement	87·0%
Intestinal involvement	39·1%
Restrictive lung disease	33·7%
DLco reduction	75·0%
Pulmonary hypertension (PHT)	27·2%
Serositis	59·8%
Myositis	8·7%
Anti-centromere antibodies (ACA)	27·2%
Anti-topoisomerase I (Scl-70) antibodies	50·0%
Anti-topoisomerase II α antibodies	21·7%

DLco, Carbon monoxide diffusing capacity.

Autoantibody analysis

Antibodies to extractable nuclear antigens Scl-70 (anti-topo I) were determined in all sera samples by ELISA kit (Biochem Immunosystems, Bologna, Italy) according to the manufacturer's instructions. Seventy out of the 92 SSc samples were also analysed using counter immunoelectrophoresis, confirming the very good overall agreement previously reported between the two assays [22]. Anti-centromere antibodies (ACA) were tested on HEp-2 cells (Antibodies Inc. Davis, CA), as previously described [23].

Anti-topo II α detection

Anti-topo II α was detected by ELISA, as previously reported [12, 13]. We employed a purified topo II α (Topogen, Columbus, OH) completely free of contamination from topoisomerase II β or topoisomerase I or from other nucleases, as evaluated by specific enzymatic activity assay, fast performance liquid chromatography and SDS–PAGE. The latter showed the presence of a single band at 170 kD. The optimal concentration for the coating of microtitre plates with purified topo II α was evaluated by chess board titrations, as were appropriate dilutions of patient sera and conjugates. ELISA microtitre plates (EIA Microplate; ICN, Costa Mesa, CA) were coated with the purified topoisomerase II adjusted to a concentration of 1·0 μg/ml in 0·05 m sodium carbonate buffer pH 9·6 and incubated at 4°C overnight. Non-adsorbed antigen was removed by washing the plates four times with PBS containing 0·1% (v/v) Tween 20. After blocking with 5% casein in PBS–Tween, human sera were applied, as was a polyclonal antibody (Topogen) against a 16-residue oligopeptide derived from the carboxyl terminal region of human topo II α. The purity of the peptide was 95%, as evaluated by high performance liquid chromatography (HPLC) and mass spectral analysis. As indicated by the manufacturer, the antibody titre was determined by ELISA to be ≥ 1:10⁴ and immunoaffinity gel chromatography; antibody purification steps were carried out to obtain a homogeneously purified IgG. Patient and normal control sera were diluted 1:100 in PBS–Tween–casein and incubated for 2 h at room temperature. The polyclonal antibody was diluted 1:1000. After four washes with PBS–Tween 20, the appropriate horseradish peroxidase-labelled antibodies (Dako, Glostrup, Denmark) were added and incubated for 2 h at room temperature. The bound antibodies were detected by adding 1,2 o-phenylenediamine (Sigma, St Louis, MO) and H₂O₂ in 100 mm phosphate buffer, pH 6·0. The reaction with the substrate solution was stopped with 2 m H₂SO₄ and absorbance was measured at a wavelength of 492 nm, using a multichannel photometer (BioRad, Hercules, CA). The cut-off point of the assay (upper limit of the normal range) was determined by the mean + 3 s.d. of the readings obtained from the 34 healthy controls.

HLA typing

HLA typing was performed on 85/92 patients. HLA class I typing was done with a standard complement-dependent microlymphocytotoxicity test and read with a fluorescence microscope [24]. Blood samples were collected in acid-citrate dextrose. Typing was carried out on peripheral blood mononuclear cells (PBMC) obtained by density gradient centrifugation on Lymphoprep (Nicomed AS, Oslo, Norway). Selected monospecific or oligospecific antisera were obtained from commercial sources (Bio Test, Dreyeich, Germany; Behering, Marburg, Germany; Fresenius, Bad Homburg, Germany; Merieux, Lyon, France). HLA-DRB, DQA1, DQB1, and DPB1 alleles were typed by the polymerase chain reaction (PCR) sequence-specific oligonucleotide probes method, which was carried out using digoxigenin-labelled probes [25]. The primers and oligonucleotide probes were those validated by the 12th International Histocompatibility Workshop.

Statistical analysis

Statistical computations were performed using CSS Statistica-Statistical software (Statsoft Inc., Tulsa, OK). Quantitative parameters were compared by means of the Mann–Whitney U-test for independent groups. Categorical data were analysed by the χ² test using Yate's correction for continuity, or else, when appropriate, by Fisher's exact test. Odds ratios (OR) were calculated and referred to as relative risk (RR).

RESULTS

Table 1 lists the demographic and clinical features of the 92 SSc patients studied. The dcSSc variety was present in 37/92 (40%) patients. Anti-topo II α was found in 20/92 (21·7%) patients. No reactivity against topo II α was observed in the sera from the 34 healthy donors, which showed binding values under the upper limit of the normal range (mean optical density (OD) value = 0·247, s.d. = 0·092, cut-off value = 0·523).

Table 2 lists the epidemiological and clinical characteristics of the SSc patients with respect to positivity or negativity for anti-topo II α. No association was found with anti-topo I or with ACA. Among the 20 patients who were positive for anti-topo II α, five (25%) were negative for both anti-topo I and ACA. Anti-topo II α was not associated with the dcSSc clinical subset. Disease duration was significantly shorter among patients with anti-topo II α (P = 0·04).

Table 2.

Epidemiological, clinical and functional features of the systemic sclerosis patients grouped with respect to positivity or negativity to anti-topoisomerase II α (anti-topo II α) autoantibodies

	Anti-topo II α+ (n = 20)	Anti-topo II α− (n = 72)	P*
Age (years) at investigation, mean ± s.e.m.	50·6 ± 2·8	52·4 ± 1·5	NS
Disease duration (years), mean ± s.e.m.	6·2 ± 0·9	9·8 ± 0·8	0·04
Sex, F:M	19:1	64:8	NS
dcSSc	50·0%	37·5%	NS
Anti-centromere antibodies (ACA)	30·0%	25·0%	NS
Anti-topo I (Scl-70) antibodies	45·0%	51·4%	NS
Pulmonary hypertension (PHT)	55·0%	19·4%	0·004
Restrictive lung disease	45·0%	30·6%	NS
DLco (percent predicted)	48·6 ± 4·4	63·8 ± 2·8	0·01

dcSSc, Diffuse cutaneous systemic sclerosis; anti-topo I, anti-topoisomerase I; DLco, carbon monoxide diffusing capacity; NS, not significant.

^*Significance of the difference between the anti-topo II α–positive (+) and -negative (−) patients as calculated by Mann—Whitney U-test, or χ² tests with Yate's correction or, when appropriate, Fisher's exact test.

As regards functional features, we observed that patients positive for anti-topo II α showed significantly decreased DLco values (P = 0·01). No association was found with forced vital capacity (FVC), arterial oxygen partial pressure or arterial oxygen saturation (Table 2). Since DLco values decrease both in the presence of restrictive lung disease and in the presence of PHT, we searched for any associations with anti-topo II α. Whereas no association was found among the SSc patients between presence of anti-topo II α and restrictive lung disease, a highly significant association was observed with PHT (11/20, 55%; P = 0·004). However, among the 25 patients who also presented PHT, the values of pulmonary arterial pressure (PAP) did not significantly differ between those subjects who were positive or negative for anti-topo II α (11/25, mean ± s.d. = 53·3 ± 15·1 mmHg versus 14/25, mean ± s.d. = 46·5 ± 13·2 mmHg). We also searched for any possible differences between primary PHT patients and those with PHT secondary to restrictive lung disease. The prevalence of anti-topo II α was 33% (4/12) among primary PHT patients, and 54% (7/13) in those with secondary PHT (P > 0·05). Therefore, a strong association emerged between anti-topo II α and PHT, independently of its specific form (primary or secondary to restrictive lung disease). We also analysed the prevalence of anti-topo II α in the subgroup of SSc patients with restrictive lung disease. Among the 31 patients with restrictive lung disease, positivity for anti-topo II α was found in only 2/18 (11%) patients without PHT, as against 7/13 (54%) with PHT (P = 0·02). This finding confirms that anti-topo II α positivity is associated with presence of PHT, but not with restrictive lung disease.

We looked for an association between anti-topo II α and HLA alleles. The frequencies of HLA-B35 were significantly different among patients positive and negative for anti-topo II α (66·7% versus 23·9%, respectively; OR = 6·4; P < 0·002). The presence of HLA-B35 also turned out to be significantly associated with the risk of developing PHT (RR = 4; P < 0·02). We found that the association of anti-topo II α with HLA-B35 was not merely secondary to that with PHT: anti-topo II α was also significantly associated with PHT in HLA-B35-negative subjects (OR = 7·5; P < 0·05) (Table 3). Nevertheless, the highest risk of developing PHT seemed to derive from an interaction between anti-topo II α and HLA-B35, since the value of the OR of patients who were positive for both anti-topo II α and HLA-B35 was much higher (OR = 21; P < 0·0001). No significant association was found with any other class I or class II antigen.

Table 3.

Risk of developing pulmonary hypertension in presence of anti-topoisomerase II α (anti-topo II α) autoantibodies and the class I HLA-B35 antigen in 85 systemic sclerosis patients

Factor	Patients	No. of patients	Odds ratio	P*
Anti-topo II α	All	85	6·36	< 0·001
Anti-topo II α	HLA-B35-negative patients	57	7·50	< 0·05
HLA-B35	All	85	6·25	< 0·001
HLA-B35	Anti-topo II α-negative patients	67	6·80	< 0·02
Combined HLA-B35 and anti-topo II α	All	85	21·00	< 0·0001

^*P values were calculated using χ² tests with Yate's correction.

DISCUSSION

Topoisomerases are ubiquitous enzymes that introduce transient single (topoisomerase I) or double (topoisomerase II) stranded breaks into DNA molecules [26]. Many antibacterial and antiblastic drugs target topoisomerases and influence key steps in their catalytic cycle [27–30]. They also form a target for autoantibodies and may be involved in the pathogenesis of certain genetic disorders [31]. Whereas DNA topoisomerase I does not fluctuate during the cell cycle [32], the two forms of topoisomerase II present different characteristics. In particular, the 170-kD α isoform varies during the cell cycle and among different cell types: it is present in proliferating cells, where it is limited to the nucleoplasm, and it has a dual enzymatic/structural role [32]. On the other hand, the 180-kD β isoform, which has been isolated from mouse and human cells, is present mainly in cells that have reached the plateau phase of growth, where it is confined to the nucleolus [33, 34]. Anti-topo I antibodies are regarded as a marker of SSc, and they have also been reported to correlate closely with pulmonary fibrosis [35]. As regards the two main clinical subtypes of SSc, anti-topo I is known to be particularly associated with dcSSc, while presence of ACA is associated with lcSSc [36, 37].

We previously reported the presence of anti-topo II α in IPF patients: using recombinant proteins, we were able to map different regions of reactivity within the topo II α molecule [13, 38]. In the present study, we found that anti-topo II α could also be detected in about 22% of our SSc patients. Its presence was not confined to one or other of the two clinical subtypes (dcSSc/lcSSc). Furthermore, no association was found with either anti-topo I or ACA. Therefore, the first major finding of the present study is the definition of a new subset of SSc patients who are positive for anti-topo II α but can be negative for anti-topo I and anti-centromere. Patients positive for anti-topo II α had a more recent presentation of SSc and worse DLco values, suggesting more aggressive and rapidly evolving disease.

IPF and SSc are both progressive fibrogenic diseases that appear to be associated with antibodies to enzymes of the topoisomerase family. However, among our SSc patients, positivity for anti-topo II α turned out to be associated with reduced DLco values and also with PHT, but not with restrictive lung disease or decreased FVC. The association between anti-topo II α and PHT is also supported by the observation that the presence of anti-topo II α was uniformly distributed among patients with either of the main forms of PHT (primary or secondary to fibrosis). Moreover, among the SSc patients with restrictive lung disease, we found that the frequency of anti-topo II α positivity was significantly higher in those who also presented PHT.

Although SSc is not primarily a genetic disorder, it is generally agreed that the disease has a genetic component [39]. Several investigators have tried to associate individual HLA class II antigens to the production of some autoantibody groups. In our patients, anti-topo II α was associated with class I but not class II HLA antigens. Furthermore, HLA-B35 correlated with PHT, independently of the presence of anti-topo II α. We are not currently able to provide an explanation for the association found between anti-topo II α, HLA-B35 and PHT. Primary PHT can be associated with autoimmune disorders [40], drug therapy or HIV infection [41, 42]. Furthermore, associations have been observed between autoantibodies and HLA in several subsets of PHT patients [43]. One possible explanation for our findings is the existence of molecular mimicry involving homology between target autoantigens and infectious agents, initiating certain autoimmune responses. An infective agent could cause immunocompetent and non-specific inflammatory cells to produce soluble mediators. In certain HLA class I subjects, this could activate and damage connective and vascular structures, as seems to occur during HIV infection [44]. Therefore, the production of autoantibodies could simply be secondary to tissue damage.

In SSc, the mimicry between sequences of retroviral p30^gag protein and topoisomerase I has been proposed as a basis for such a mechanism [45]. In this scenario, the production of anti-topo I autoantibodies could represent a footprint of the retrovirus. In an earlier study on IPF, we were unable to find any significant homology between selected epitopes of topoisomerase II α and foreign pathogens [38]. However, the epitopes recognized by anti-topo II α in SSc could be different from those of IPF.

In conclusion, our results indicate that even though the presence of anti-topo II α and prevalence of HLA-B35 alleles are mutually independent factors, their simultaneous presence drastically increases the risk of developing PHT. This observation allows identification of a new serological subtype of SSc, corresponding to a separate clinical entity showing a strict association with particular HLA alleles. These concepts provide further support for the hypothesis that in SSc, multiple etiologic processes can give rise to a similar pathogenic mechanism.