Introduction

The nucleosome is the fundamental chromatin unit. It consists of a core particle, composed of an octamer of two copies each of histones H2A, H2B, H3, and H4 around which two superhelical turns of approximately 146 base pairs of helical DNA are wrapped. Neighboring nucleosomes are joined together by linker DNA, which is associated with histone H1 located outside the nucleosome core. Its molecular weight is 262 kDa, and its crystal structure has recently been identified [1]. Nucleosomes are generated during cell apoptosis by chromatin cleavage carried out by endonucleases. Apoptosis-induced modifications of autoantigens may make them more immunogenic, particularly if an inadequate clearance of apoptotic cells coexists [2].

In lupus nephritis, nucleosomes, anti-nucleosome specific antibodies, and nucleosome/IgG complexes were identified in glomerular immune deposits. Nucleosomes mediate the binding of antinuclear antibodies to glomerular basement membranes via their cationic histone component [3].

In another paper, we found a correlation between anti-nucleosome antibodies and systemic lupus erythematosus (SLE) disease activity as expressed by the higher ECLAM score, by the significantly lower levels of C3 and C4 and by the lower levels of hemoglobin [4].

In another paper, in a large cohort of patients with connective tissue diseases, absence of significance for anti-nucleosome specific antibodies presence was detected in 36 patients with rheumatoid arthritis (RA) [5]. If patients with RA do not have circulating nucleosomes at the steady state, treatment with anti-tumor necrosis factor α (TNFα) could induce significantly high levels of nucleosomes in the plasma level. The accumulation of nucleosomes possibly can develop autoantibodies in subjects with appropriate genetic backgrounds [6].

The aim of our study was to detect in 91 patients with rheumatoid arthritis the positivity of anti-nucleosome antibodies during therapy with three different TNFα blocking agents and to underline the possible correlation with the development of antinuclear autoantibodies (ANA) and anti-dsDNA autoantibodies.

Materials and methods

Ninety-one consecutive outpatients with RA were recruited at the Rheumatology Unit of ASL 10 of Florence (Florence, Italy). At enrolment, all patients fulfilled the 1986 American Rheumatism Association criteria for classification of RA [7].

We observed 19 patients on infliximab treatment at 3 mg/kg every 8 weeks, 43 patients on etanercept treatment at 25 mg twice a week, and 29 patients on adalimumab treatment at 40 mg every other week.

Clinical features of the patients and concomitant therapy are summarized in Table 1.

Table 1 Clinical features of the patients and concomitant therapy
Full size table

We detected anti-nucleosome antibodies, ANA and anti-dsDNA autoantibodies at T-0 and after 12 (T-12) and 24 weeks (T-24) of treatment assay, respectively.

Anti-nucleosome antibodies (IgG class) were analyzed by ELISA technique (Orgentec Diagnostika GmbH, Mainz, Germany), a quantitative immunoenzymatic test for the detection of anti-nucleosome antibodies, in which antigens are represented by human nucleosomes. During the first incubation, the serum to be tested is added to the antigen-coated microtiter plate wells. If anti-nucleosome antibodies are present in the sample, they will bind to the antigen-coated solid phase. After washing the unbound material, peroxidase-polyclonal antibody is added to bind the newly formed antigen–antibody complexes. After washing, specific antibodies are revealed during incubation with a substrate (tetramethylbenzydine solution). After adding the blocking solution, enzymatic reaction is stopped and the concentration of anti-nucleosome antibodies is directly correlated to the intensity of the color developed during the enzymatic reaction. In the assay calibration, human sera showing seven different concentrations ranging from 0 to 200 U/ml and a positive and a negative control are used.

Anti-nucleosome antibodies are measured by arbitrary units (U/ml): the sample is considered positive if the titre of anti-nucleosome antibodies is ≥20 U/ml and negative if the titre is <20 U/ml.

ANA were detected both by indirect immunofluorescence (IF) technique on Hep2 (Scimedx, USA) and by ELISA (Autoimmune EIA ANA screening test Bio-Rad Laboratories, CA, USA), and anti-dsDNA (IgG and IgM classes) by ELISA (Kallestad, Bio-Rad Laboratories, CA, USA) and confirmed for specificity by IIF technique on Crithidia luciliae (Immunoconcepts N.A., Sacramento, CA, USA).

Statistics

Data were analyzed using SPSS 10.0 for Windows. Data were analyzed using χ 2 test with Yates’s factor correction for percentage differences and McNemar test for concordance evaluation. P values less than 0.05 were considered to be of statistical significance.

Results

At baseline, we observed in the group of patients treated with infliximab the following positivity: anti-nucleosome 1/19 (5.26%), ANA 3/19 (15.7%), anti-dsDNA 1/19 (5.26%); in the group of patients treated with etanercept: anti-nucleosome 2/43 (4.65%), ANA 1/43 (2.32%), anti-dsDNA 0/43; and in the group treated with adalimumab: anti-nucleosome 2/29 (6.89%), ANA 1/29 (3.44%), anti-dsDNA 0/29. The results at 12 weeks were as follow—in the group of patients treated with infliximab: anti-nucleosome positivity 3/19 (15.7%), ANA positivity 10/19 (52.6%), anti-dsDNA positivity 2/19 (10.5%); in the group of patients treated with etanercept: anti-nucleosome positivity 3/43 (6.9%), ANA positivity 10/43 (23.2%), anti-dsDNA positivity 1/43 (2.32%); in the group of patients treated with adalimumab: anti-nucleosome positivity 3/29 (10.3%), ANA positivity 4/29 (13.7%), anti-dsDNA positivity 1/29 (3.4%); while at 24 weeks—in the group of patients treated with infliximab: anti-nucleosome positivity 6/19 (31.5%), ANA positivity 12/19 (63.1%), anti-dsDNA positivity 2/19 (10.5%); in the group of patients treated with etanercept: anti-nucleosome positivity 11/43 (25.5%), ANA positivity 22/43 (51.1%), anti-dsDNA positivity 2/43 (4.65%); in the group of patients treated with adalimumab anti-nucleosome positivity 4/29 (13.7%), ANA positivity 13/29 (44.8%), anti-dsDNA positivity 1/29 (3.4%). At 24 weeks, we had the following positivity percentages for anti-nucleosome antibodies: infliximab 31.5%, etanercept 25.5%, adalimumab 13.7%; for ANA: infliximab 63.1%, etanercept 54.1%, adalimumab 44.8%. The statistical analysis between the three drugs showed the following results at baseline: for anti-nucleosome antibodies p = 0.918 (ns); for ANA, p = 0.844 (ns); and for anti-dsDNA, p = 0.239 (ns). At 12 weeks, these were respectively p = 0.560 (ns), p = 0.090, p = 0.932 (ns), while at 24 weeks were p = 0.038, p = 0.044, and p = 0.46 (ns), respectively. The results are summarized in Tables 2 and 3. We observed a positivity concordance of anti-nucleosome/ANA antibodies at 85.5% (p < 0.001). No concordance in positivity between anti-nucleo some antibodies and anti-dsDNA was observed. No drug-induced SLE disease was observed.

Table 2 Percentage difference of different autoantibodies at baseline and during therapy
Full size table
Table 3 Statistical analysis between three groups at baseline and at T-12 and T-24
Full size table

Discussion

The aim of our work was to analyze the clinical and biologic correlation of autoantibody induction during long-term TNFα blocking therapy with either the monoclonal antibody infliximab and adalimumab or the soluble receptor etanercept. We know that the induction of autoantibodies is different for the different drugs. In particular, data from literature showed that infliximab, a chimeric-murine monoclonal antibody, can induce in 53–78% of patients treated ANA positivity, 7–10% can develop anti-dsDNA, but drug-induced SLE was present only in 0.22% [8]. Data about etanercept, a soluble receptor, showed a development of ANA positivity in 8–11% of patients in treatment and anti-dsDNA in 3–15%, but with only 16 case reports by US FDA Adverse Reporting from 1998 to 2002 [9].

Data about the last TNFα blocking agent, adalimumab, a fully human monoclonal antibody, showed development of ANA positivity in 12.6% of patients in treatment and anti-dsDNA in 3.9% of patients [10], but drug-induced SLE disease is rare [11].

The induction of anti-nucleosome antibodies during therapy with TNFα blocking agents has produced different results in different papers. A study on 34 patients with spondiloarthropathy and on 59 patients with rheumatoid arthritis treated with infliximab for 2 years and on 20 more patients with spondiloarthropathy treated with etanercept for 1 year showed a high number of patients with SpA or RA developing newly induced ANA (61.8 and 40.7%, respectively) and anti-dsDNA antibodies (70.6 and 49.2%, respectively). By contrast, induction was observed only occasionally in etanercept-treated patients with SpA (10% in each of the patients) and neither infliximab nor etanercept induced anti-nucleosome antibodies [12]. On the contrary, a study of 53 patients with rheumatoid arthritis treated with infliximab and followed for 54 weeks and of 6 patients treated with etanercept showed a high prevalence of IgG anti-dsDNA with a frequency of 66% at 30 weeks and 45% at 54 weeks, and of IgM anti-dsDNA at 85 and 70%, respectively. Moreover, in this study the titer and the number of patients developing antibodies against nucleosomes and ANA also increased significantly [13]. In another paper, an increase of anti-histone antibodies was observed with a frequency of 32% for IgG class and 79% for IgM class [14].

In a recent paper on 54 patients with rheumatoid arthritis treated with infliximab (43 patients) and with etanercept (11 patients) for 54 weeks, a different incidence of ANA positivity was found in infliximab- and etanercept-treated RA patients; this finding might be due to the partially different method of inhibition of TNFα between the two drugs [15]. In the sera from 58 RA patients, treated for 1 to 3 years with infliximab, serum ANA, anti-dsDNA, and anti-dsDNA IgG antibodies provided predictor factors of lupus-like symptoms and/or anaphylactoid reactions in patients with RA [16].

Our paper is the first that compares the three different TNFα blocking agents from the point of view of autoantibodies induction during treatment, showing in particular a lower induction of anti-nucleosome antibodies and ANA positivity for adalimumab. Similar frequency was detected in another paper using the same methods to determine the anti-nucleosome antibodies on 26 patients treated only with adalimumab [17]. A reduction of induction of anti-nucleosome antibodies can explain the different induction of ANA positivity and anti-dsDNA positivity for different drugs.

In 11 healthy controls, 51 patients with rheumatoid arthritis, and 33 patients with systemic lupus erythematosus, circulating nucleosomes, detected by enzyme-linked immunosorbent assay, were observed at baseline in the last group, but 8 of 11 patients with rheumatoid arthritis had significantly higher levels of plasma nucleosomes after infliximab treatment [6]. Recently, the human–mouse chimeric TNFα neutralizing antibody, infliximab, has been shown to cause apoptosis of leukocytes in the lamina propria of patients with Crohn’s disease [18, 19], but this mechanism does not involve etanercept [20]. Apoptotic cells might represent a reservoir of autoantigen targeted in susceptible hosts [21], as a consequence of defective phagocytosis [22], and the failure of complement fractions [23] or pentraxins [24] can be involved in nucleosome accumulation. It is possible that during infliximab therapy the p55 and p75 membrane receptors interact by oligomerization and were bound with TNFα that is required for triggering of signal pathways involved in apoptosis with recruitment of upstream caspase [6]. The notable down-regulation of CRP that occurs after infliximab treatment could further potentiate autoimmunity by reducing the clearance of nuclear material by CRP [8].

Another important observation is that TNFa could up-regulate cellular expression of the adhesion molecule CD44, which has a role in the clearance of apoptotic neutrophils by phagocytes [25]. An impaired clearance of apoptotic cells and a reduced leukocyte CD44 have been described in SLE [26]. Many hypotheses have been reported, but the exact molecular mechanisms by which TNFα inhibitors induce autoantibodies or autoimmune responses remain to be elucidated. Our data showed a concordance between anti-nucleosome antibodies and ANA positivity in patients with RA during therapy with TNFα blocking agents. The induction of positivity for autoantibodies is different for each TNFα blocking agent and adalimumab seems to show the lowest immunogenicity.