SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. REFERENCES

Objective

To analyze changes in autoantibodies occurring in patients with systemic lupus erythematosus (SLE) treated with 4 infusions of the chimeric anti–tumor necrosis factor α (TNFα) antibody infliximab.

Methods

In an open-label safety study, 7 patients with SLE were treated with infliximab at weeks 0, 2, 6, and 10 in combination with azathioprine or methotrexate. Antibodies to double-stranded DNA (dsDNA) were determined by radioimmunoassay and the Crithidia luciliae indirect immunofluorescence assay; anticardiolipin antibodies (aCL) and antibodies to histone and chromatin were measured by enzyme-linked immunosorbent assay. Antihistone antibodies were also analyzed by immunoblotting. Peripheral blood mononuclear cells from healthy individuals and SLE patients were incubated for 2 weeks with or without TNFα. TNFα was removed by washing and by the addition of infliximab. Apoptotic cells were stained with annexin V and analyzed by flow cytometry.

Results

Autoantibodies to dsDNA increased in 5 of 7 patients. Histone, chromatin, and IgM aCL levels were increased in 4 of 7, 6 of 7, and 4 of 7 patients, respectively, peaking 4–10 weeks after the last infliximab infusion, but falling to baseline levels or lower thereafter. In the in vitro experiments, TNF withdrawal after long-term incubation with recombinant human TNF led to increased percentages of apoptotic cells.

Conclusion

While TNF blockade was clinically effective, the majority of SLE patients treated with infliximab showed an increase in autoantibodies to nuclear antigens and phospholipids. These increases were transient and were not associated with disease flares. Increased availability of apoptotic antigens after TNF blockade may play a role in the autoantibody formation induced by TNF blockade.

Tumor necrosis factor (TNF) has long been known to function as an immunoregulatory cytokine. TNF has immunosuppressive properties, and Jacob and McDevitt (1) have demonstrated that a genetic deficiency in TNF plays a significant role in the onset of autoimmunity in (NZB × NZW)F1 mice, a finding that was more recently supported by the observation of severe lupus-like disease in TNF-deficient NZB mice (2). However, TNF is also a proinflammatory cytokine that is pivotal in the pathogenesis of a variety of inflammatory diseases.

When therapeutic TNF blockade became available, patients receiving infliximab were found to develop antinuclear and other autoantibodies and, rarely, lupus-like syndromes (3). Studies on autoantibody formation in patients with RA, spondylarthritis, or Crohn's disease suggest that TNF blocker therapy induces new-onset antinuclear antibodies (ANAs) in 25–70% of initially ANA-negative patients (3, 4), with occurrence being partially dependent on the underlying disease. Antibodies to double-stranded DNA (dsDNA) emerge in 10–30% of patients receiving infliximab therapy; the vast majority of these antibodies to dsDNA are of the IgM isotype and are usually nonpathogenic. In contrast, the formation of IgG anti-dsDNA, a relatively rare event, may be more closely associated with the development of a syndrome like systemic lupus erythematosus (SLE) (3). Lupus induced by TNF blockers is not only rare, but it is usually also benign, with resolution upon cessation of TNF-blocker therapy (5).

We have recently evaluated short-term TNF-blocker therapy with infliximab in an open-label safety trial in SLE patients (6). In this trial, proteinuria not only subsided rapidly, but this improvement was long-lasting; in contrast, remission of lupus arthritis lasted only a few weeks after cessation of TNF blockade. While transient increases in dsDNA autoantibodies were observed in the majority of patients, we did not observe disease flares during the initial observation period (6).

We performed detailed analyses of autoantibodies to nuclear antigens and to phospholipids in this cohort of patients. Moreover, since it has been suggested that apoptotic cells may fuel autoimmunity (7), we also performed in vitro experiments to ascertain if TNF inhibition induces an increase in apoptosis, which may constitute a possible mechanism of autoantibody induction (8, 9).

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. REFERENCES

Patients and protocol.

As previously described (6), patients who fulfilled the American College of Rheumatology criteria for SLE (10) were included in an open-label infliximab safety protocol. A total of 7 patients received 4 doses of 300 mg infliximab on day 0 and at weeks 2, 6, and 10, in combination with their baseline immunosuppressive therapy with azathioprine (AZA) or methotrexate (MTX) (6).

Determination of autoantibodies.

Anti-dsDNA antibodies were assessed by radioimmunoassay (RIA; Ortho-Clinical Diagnostics, Rochester, NY) and by an IgG-specific Crithidia luciliae indirect immunofluorescence test (Bios, Gräfelfing, Germany). Antibodies against Ro/SSA, La/SSB, Sm, U1 RNP, and other ANA subsets were detected by line immunoassay (Innogenetics, Ghent, Belgium) and by immunoblotting using HeLa cell nuclear extracts, as previously described (11). Enzyme-linked immunosorbent assay (ELISA) was used to measure antihistone (Imtec, Berlin, Germany), antichromatin (Inova Diagnostics, San Diego, CA), and IgG and IgM anticardiolipin (aCL) (Pharmacia, Uppsala, Sweden) autoantibodies.

Assessment of apoptosis.

Peripheral blood mononuclear cells (PBMCs) derived from the peripheral venous blood of consenting healthy donors or SLE patients were purified over Ficoll-Paque density gradients (Amersham Pharmacia Biotech, Uppsala, Sweden), washed twice in phosphate buffered saline, and incubated at 37°C in RPMI 1640 medium with or without increasing concentrations of recombinant human TNF (Strathmann Biotec, Hannover, Germany). RPMI 1640 medium with or without TNF was replaced every 3 days, and cultures were maintained until day 12.

On day 12, cells were washed and the medium was replaced with fresh RPMI 1640 medium without TNF, but with the addition of infliximab (0.5 mg/ml). After 24 hours of incubation, cells were stained with fluorescein isothiocyanate–conjugated annexin V (Alexis Biochemicals, Lausen, Switzerland) to detect apoptotic cells, and analyzed using a Becton Dickinson (San Jose, CA) FACScan flow cytometer (Becton Dickinson, Mountain View, CA), as previously described (12).

Likewise, fresh PBMCs from 20 SLE patients were incubated overnight in medium supplemented with 100 ng/ml of recombinant human TNF or with 0.5 mg/ml of infliximab. In a subset of patients, a third portion of PBMCs was incubated with 0.5 mg/ml of the chimeric anti-CD20 antibody rituximab for control purposes. Cells were prestained with antibodies to CD4, CD8, CD14, and CD19 conjugated to phycoerythrin or peridinin chlorophyll A protein (all from Becton Dickinson), stained with annexin V, and analyzed as described above.

Statistical analysis.

To account for the significant differences in autoantibody levels among different individuals, autoantibodies were normalized for baseline levels. Values are expressed as the mean ± SEM percentage of baseline. Likewise, apoptosis values were normalized for the percentage of apoptotic cells in medium. Descriptive statistics were performed using Wilcoxon's signed rank test. Results of the 24-hour apoptosis experiments were not normalized, and paired t-tests were used for statistical analysis. P values less than 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. REFERENCES

Transient increase in autoantibodies to dsDNA.

Treatment with 4 infusions of infliximab led to an increase in high-affinity autoantibodies to dsDNA, as measured by RIA, in 5 of 6 patients with preexisting anti-dsDNA antibodies, but did not lead to the new formation of such antibodies in the 1 patient without preformed anti-dsDNA. The kinetics of the increase were comparable. Increases were first observed at approximately week 10, the time point of the last infliximab infusion, with peak levels observed 4 weeks later (Figure 1A). After this peak, which was >100% above baseline levels, anti-dsDNA antibodies decreased if infliximab was not reinitiated. However, in 2 patients who were treated with additional infliximab infusions after the end of the study, these autoantibody titers increased between weeks 20 and 36 (from 73 to 317 IU/ml and from 155 to 352 IU/ml, respectively), further confirming the relationship between TNF blockade and autoantibody formation.

thumbnail image

Figure 1. Levels of autoantibodies against chromatin and its components in 7 infliximab-treated systemic lupus erythematosus patients during and after anti–tumor necrosis factor therapy. A–D, Mean ± SEM levels of autoantibodies to double-stranded DNA (anti-dsDNA) as measured by A, radioimmunoassay (RIA) and B,Crithidia luciliae indirect immunofluorescence test (Crith), and levels of autoantibodies to C, chromatin and D, histones, both measured by enzyme-linked immunosorbent assay. Results are normalized to baseline levels. Patients reinfused with infliximab after week 20 (n = 2) were not included in the analyses after this time point. ∗ = P < 0.05; ∗∗ = P < 0.01 versus baseline, by Wilcoxon's signed rank test. E, Representative histone immunoblots from patients 1, 2, 3, and 5. Each arrow in A–E indicates the time of one 300-mg infusion of infliximab.

Download figure to PowerPoint

Since most anti-dsDNA autoantibodies occurring in infliximab-treated patients have been reported to be of the IgM isotype, but not of the potentially more pathogenic IgG isotype (3), we assessed sera for the presence of IgG anti-dsDNA autoantibodies by Crithidia luciliae assay. An increase in mean IgG anti-dsDNA antibodies was observed, which peaked at week 14 and decreased after cessation of infliximab infusions (Figure 1B).

Comparison of anti-dsDNA with autoantibodies to chromatin and histones.

Infliximab therapy also led to an increase in autoantibodies to chromatin (Figure 1C), which occurred in 6 of the 7 patients. It is interesting that the initiation of this response preceded the initiation of the IgG anti-dsDNA response, peaked somewhat later, and declined more sharply than anti-dsDNA thereafter. These data suggest that the time course of antichromatin antibodies preceded that of anti-dsDNA antibodies. Total antihistone autoantibodies, as measured by ELISA, were similarly increased, but in 4 of 7 patients only, and in a less pronounced manner (Figure 1D).

Additional detailed analyses, aimed at determining autoantibody responses directed at specific histone classes, were performed by immunoblotting (Figure 1E). These showed that some antihistone response occurred in all patients, although with significant variability both with regard to the amounts of autoantibodies and the histone classes targeted. While antihistone antibodies were detected at baseline in all patients, these antibody levels increased during the course of infliximab therapy and returned to pretreatment response levels after cessation of such therapy. However, in 2 patients, new specificities emerged that had not been visible at baseline, and these were primarily directed against histone H4 (Figure 1E, patients 2 and 5). Thus, even after the first administration of TNF-blocker therapy, novel reactivity appeared that might have been due to an increase in initially undetectable autoantibody levels or to intermolecular epitope spreading. Interestingly, however, in a majority of the patients, the antihistone responses subsided below baseline levels after cessation of therapy, as assessed by both immunoblotting (Figure 1E, patients 1, 2, and 5) and ELISA (Figure 1D). The emergence of these antihistone specificities even preceded that of antichromatin, suggesting that it is the first immune response that occurs or increases.

In contrast to the increases in antichromatin antibodies, no changes in autoantibodies to other ANA subsets, such as Ro or Sm, were observed. In particular, no autoantibodies with new specificity were found in any patient (data not shown).

Transient increase in aCL.

During this open-label safety study, none of the patients experienced thrombotic or thromboembolic events (6). However, in 4 patients, changes in aCL of the IgM isotype were detected (Figure 2B). Similar to the pattern observed with the autoantibodies against chromatin, these aCL titers also tended to decrease after a peak no later than 10 weeks after therapy was stopped (Figure 2).

thumbnail image

Figure 2. Anticardiolipin antibody (aCL) levels during and following anti–tumor necrosis factor therapy. Values are the mean ± SEM levels of A, IgGaCL and B, IgM aCL, as measured by enzyme-linked immunosorbent assay. Results are normalized to baseline levels. Patients reinfused with infliximab after week 20 (n = 2) were not included in the analyses after this time point. Each arrow indicates the time of one 300-mg infusion of infliximab. ∗ = P < 0.05 versus baseline, by Wilcoxon's signed rank test.

Download figure to PowerPoint

Apoptosis rates and long-term TNF exposure.

We mimicked the situation of PBMCs in the peripheral blood of SLE patients, in which TNF is highly increased and bioactive (6), by long-term incubation of the PBMCs in medium containing TNF. After 12 days of constant TNF exposure, TNF was withdrawn and blocked with infliximab, and the percentage of cells undergoing programmed cell death was analyzed. Control experiments were performed using medium without TNF. Two weeks of in vitro culture with TNF was sufficient to increase the rates of apoptosis upon withdrawal of TNF (Figure 3A).

thumbnail image

Figure 3. Induction of apoptosis by sudden withdrawal of tumor necrosis factor (TNF) after long-term stimulation with TNF. A, Peripheral blood mononuclear cells (PBMCs) from healthy individuals were washed and treated with infliximab after 12 days of incubation without TNF or with various concentrations of TNF. After 18 hours, the percentage of apoptotic cells was determined by flow cytometry, after staining with fluorescein isothiocyanate–labeled annexin V. B, The same experiment was conducted with PBMCs from systemic lupus erythematosus patients. Cells from patients with a normal rate of in vitro apoptosis after long-term culture (<50%) behaved differently from those with very high baseline rates of apoptosis (>50%). Values are the mean and SEM of 7 independent experiments. ∗ = P < 0.05 versus treatment with medium alone, by Wilcoxon's signed rank test.

Download figure to PowerPoint

When the same experiment was repeated with PBMCs from SLE patients, the cells showed a similar increase in apoptosis rates when preincubated with TNF, if spontaneous apoptosis rates with medium were <50% (as for healthy controls). However, this effect could not be distinguished if apoptosis rates were excessive (open triangles in Figure 3B).

To further analyze the effect of TNF withdrawal, fresh PBMCs from 20 SLE patients were incubated overnight, either with medium, maintaining the influence of TNF, or with infliximab, thus completely blocking the influence of TNF. Apoptosis was determined separately for CD4+ and CD8+ T cells, CD19+ B cells, and CD14+ monocytes. Increased apoptosis upon blocking TNF was seen for CD4+ cells (mean ± SD 19.1 ± 12.8% versus 12.5 ± 5.5%; P < 0.05), but not for CD8+ T cells (20.1 ± 10.1% versus 18.7 ± 10.5%; P not significant [NS]), B cells (25.6 ± 22.9% versus 22.9 ± 21.1%; P NS), or monocytes (83.9 ± 9.5% versus 84.8 ± 17.8%; P NS). In contrast, the anti–B cell antibody rituximab led to B cell apoptosis (71.4% versus 30.0 ± 18.8%; P < 0.01), but not to T cell apoptosis (15.2 ± 6.2% versus 15.2 ± 5.6%, P NS for CD4+ cells), as expected.

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. REFERENCES

It is known that autoantibodies develop with TNF blockade therapy. Accordingly, organ damage potentially induced by such antibodies is a significant concern, particularly in patients with SLE. In our study, findings from the first series of SLE patients treated with a short course of 4 infusions of infliximab over 10 weeks revealed that titers of some autoantibodies, namely, those directed against dsDNA and DNA-binding proteins and against phospholipids, increased. In fact, all major autoantibody types peaked 4–10 weeks after the last infliximab infusion, well within the time frame of detectable infliximab concentrations (12). Despite these increases in autoantibody titers, all patients had a good clinical response to infliximab treatment.

After reaching their peak, all autoantibody levels steadily decreased to pretreatment concentrations or lower, as long as TNF blockade was not reinstituted. Overall, despite their preformed antibodies, our findings in the infliximab-treated SLE patients were consistent with the published data on infliximab-induced autoantibodies in patients with other diseases (4, 5). However, presumably due to the high prevalence of preformed antichromatin antibodies in SLE, anti-dsDNA antibodies occurred more frequently in our study population than is seen in other diseases, and were of the IgG isotype in all affected patients.

While all infliximab-treated patients had some increase in their autoantibody response to chromatin (Figure 1), these transient increases were without pathologic consequences. The pattern observed here, namely, a rapid decrease in autoantibodies after short-term infliximab treatment, may render disease flares less likely, as long as TNF blockade is used for a limited time period only.

Induction of apoptosis could constitute a pathway to the induction of new antibodies under TNF blockade (3). In fact, autoantibodies against chromatin and against phospholipids are known to target antigens that are accessible on apoptotic cells and are in part modified during programmed cell death (13, 14). Moreover, long-term TNF exposure also increases antiapoptotic proteins (15). In SLE, TNF is present in high concentrations both in tissue and in serum, and the latter was demonstrated to be bioactive (6). We therefore hypothesized that constant TNF exposure may provide a survival signal (15) that, upon the rapid withdrawal of TNF resulting from infusion of infliximab, is suddenly lost, thus leading to enhanced cell death.

Indeed, when mimicking the in vivo situation by long-term incubation of PBMCs with TNF, its sudden withdrawal led to an increase in the number of apoptotic lymphocytes. Thus, sudden TNF blockade with infliximab may lead to increased rates of apoptosis in patients with SLE, thereby possibly stimulating autoantibody formation. This is also consistent with the reduction of autoantibodies observed in patients who stop infliximab therapy, but not in those who continue infliximab therapy.

Another interesting observation is the antihistone reactivity during TNF blockade in patients with SLE. Not only did levels of antihistone antibodies increase, but reactivities to histone classes and subtypes not seen prior to infliximab treatment emerged. Thus, infliximab may have induced intermolecular epitope spreading with respect to the nucleosome-directed response, while other specificities, such as anti-Ro or anti-Sm, neither occurred de novo nor spread to additional moieties of the respective molecular complexes, such as the spliceosome (results not shown). Interestingly, after cessation of infliximab treatment, the antihistone responses assessed by both ELISA and immunoblotting rapidly decreased to levels below baseline; similar behavior was also observed for antichromatin antibodies but not for anti-dsDNA antibodies.

Taken together, our findings indicate that, while all SLE patients improved clinically following 4 infusions of infliximab (administered in conjunction with either AZA or MTX and low-dose corticosteroids), all patients also had an increase in levels of autoantibodies directed to one or more chromatin components. In addition, increases in aCL were detected in 4 patients. All of these increases were transient and in close temporal association with anti-TNF therapy. The finding that TNF blockade induces PBMCs that have adapted to long-term TNF exposure to undergo apoptosis may suggest increased antigen presentation under these circumstances, driving autoantibody formation.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. REFERENCES

Dr. Aringer had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Drs. Aringer, Steiner, Graninger, and Smolen.

Acquisition of data. Drs. Aringer and Steiner, Ms Höfler, and Dr. Steiner.

Analysis and interpretation of data. Drs. Aringer, Steiner, Graninger, and Smolen.

Manuscript preparation. Drs. Aringer, Steiner, and Smolen.

Statistical analysis. Dr. Aringer.

ROLE OF THE STUDY SPONSOR

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. REFERENCES

Centocor proofread the manuscript for language, but otherwise had no role in the study design or in the collection, analysis, or interpretation of the data, the writing of the manuscript, or the decision to submit the manuscript for publication.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. ROLE OF THE STUDY SPONSOR
  8. REFERENCES