SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgments
  8. REFERENCES

Objective

To determine the prevalence of anti–apoptotic cell (anti-AC) antibodies with the 9G4 idiotype (9G4+) and the relationship between this and other known 9G4+ specificities and disease activity in patients with systemic lupus erythematosus (SLE).

Methods

Serum samples from 60 SLE patients and 40 healthy donors were incubated with apoptotic Jurkat cells and assayed by flow cytometry for the binding of 9G4+ antibodies. The samples were also tested for 9G4+ reactivity against naive B cells and total IgG and IgM anti-AC antibody reactivity.

Results

The 9G4+ antibodies bound late ACs in sera from a majority of the SLE patients (60%) but in sera from only 2 healthy control subjects. Among samples with global IgM or IgG anti-AC antibodies, those with 9G4+ anti-AC antibodies predominated. Patients with high levels of 9G4+ anti-AC antibodies were more likely to have active disease. This was the case even in patients with IgG anti-AC antibodies or anti–double-stranded DNA antibodies. Patients with lupus nephritis were also more likely to have 9G4+ anti-AC antibodies. While 9G4+ reactivity to ACs often coincided with anti–B cell reactivity, some samples had distinct anti–AC or anti–B cell reactivity.

Conclusion

The 9G4+ antibody represents a major species of anti-AC antibody in SLE serum, and this autoreactivity is associated with disease activity. The anti-AC reactivity of 9G4+ antibodies can be separated from the germline VH4–34–encoded anti–B cell autoreactivity. Our results indicate that ACs are an important antigenic source in SLE that positively selects B cells with intrinsic autoreactivity against other self antigens. This selection of 9G4+ B cells by ACs may represent an important step in disease progression.

During homeostasis, billions of cells die through apoptosis each day, and as a potential source of autoantigens, these cells must be efficiently cleared in an immunologically silent manner to prevent pathologic autoimmune reactions ([1]). Defective clearance of apoptotic cells (ACs) has been demonstrated both in vitro and in vivo in systemic lupus erythematosus (SLE), an autoimmune disease characterized by the generation of antibodies against multiple nuclear antigens ([2-4]). This is demonstrated by the high incidence of SLE in patients with a genetic deficiency of C1q, a complement component involved in the opsonization and clearance of ACs. Like humans with SLE, transgenic mice deficient in complement C1q develop autoantibodies or a lupus-like disease ([5]), and mice deficient in tyrosine receptor kinases necessary for AC phagocytosis also develop severe autoimmunity ([6]). Furthermore, immunization of mice with ACs results in autoantibody production and autoimmune disease ([7]).

Anti-AC antibodies have previously been identified in serum samples from patients with SLE and have been noted to be bound to glomerular apoptotic nucleosomes in kidneys from patients with lupus nephritis ([8]). Anti-AC antibodies from SLE patients can exercise pathogenic functions by promoting phagocytosis of ACs ([9, 10]), resulting in the engagement of intracellular Toll-like receptors, which leads to the release of type I interferon and other proinflammatory cytokines ([11-13]). While studies of IgG anti-AC antibodies in SLE have been focused on their influence on phagocytosis ([3, 10, 14]), systematic studies of their prevalence and significance are lacking. Similarly, the nature of IgG anti-AC antibodies and the processes leading to their generation and selection in SLE remain unclear. Of note, IgM anti-AC antibodies have been associated with protection against renal disease in SLE ([15]).

In this study, we systematically investigated the presence of IgG and IgM antibody binding to ACs in SLE patients by use of a flow cytometry–based assay, and we determined the contribution of antibodies bearing the 9G4 idiotype (9G4+) to this autoreactivity. The study of intrinsically autoreactive 9G4+ antibodies encoded by the VH4–34 gene is informative in SLE, since due to defective germinal center editing, these antibodies represent 10–40% of all serum IgG ([16, 17]). The relevance of understanding the antigenic forces that underpin the expansion of 9G4+ antibodies in SLE is further illustrated by their high degree of specificity for SLE and their correlation with disease activity and specific clinical manifestations, including lupus nephritis ([18-21]). Our results indicated that the presence of 9G4+ anti-AC antibodies is common in SLE and showed that patients with elevated levels of 9G4+ anti-AC antibodies are more likely to have active disease. These findings demonstrate that reactivity to AC antigens contributes significantly to the expansion of a major autoreactive B cell population that is specifically expanded in SLE and provide the experimental basis for a better understanding of the antigenic forces involved in the pathogenesis of this disease.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgments
  8. REFERENCES

Patient samples and study design

Serum samples were obtained from 40 healthy adult donors and 60 adult patients with SLE (1 man and 59 women). SLE study participants met at least 3 American College of Rheumatology criteria for the diagnosis of SLE ([22]). These patients had a wide range of clinical disease activity, as defined by their Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) scores (range 0–20, median 4) ([23]). The ages of the patients ranged from 18 to 85 years (mean age 44 years); 57% of them were Caucasian, 40% African American, and 3% Hispanic. Serum samples from an additional 25 SLE patients, 9 SLE patients with lupus nephritis, and 12 healthy donors were studied in a separate experiment.

All samples were obtained after informed consent and in accordance with protocols approved by the institutional review boards at University of Rochester Medical Center (URMC) and Emory University.

Induction of cell death

The CD45-deficient human leukemic T cell line, Jurkat (J45.01), was maintained in complete medium at 37°C in an atmosphere containing 5% CO2. Apoptosis was induced by treatment for 16–18 hours with 20 μM camptothecin. After blocking with normal mouse serum, apoptotic J45.01 cells were resuspended at a density of 2.5 × 106 cells/ml and incubated for 30 minutes at 4°C with 25 μl of serum, which was diluted with phosphate buffered saline to yield equal amounts of IgM (0.4 mg/ml) or IgG (4 mg/ml). Cells were then incubated for 20 minutes at 4°C with fluorescein isothiocyanate (FITC)–conjugated anti-human IgG and phycoerythrin (PE)–conjugated anti-human IgM (BD Biosciences) or FITC-conjugated 9G4 (custom conjugated; SouthernBiotech), washed, and stained with 7-aminoactinomycin D (7-AAD) (SouthernBiotech) and, in some experiments, with PE–annexin V (SouthernBiotech). Cells were then analyzed by flow cytometry. Early ACs were defined as annexin V+ and 7-AAD–, and late ACs were defined as annexin V+ and 7-AAD+.

Binding for microscopy was done similarly, except that Alexa Fluor 488–labeled goat anti-rat secondary antibody (Molecular Probes) was used to detect 9G4+ binding, and the membrane-permeable dye DRAQ5 (Axxora) was used as a nuclear stain. After staining, cells were plated into poly-D-lysine–coated glass-bottomed petri dishes (MatTek). B cell binding was measured on freshly obtained peripheral blood lymphocytes from healthy donors, as previously described ([16]).

Flow cytometric analysis

Experiments were run on a LSRII 12-color or a FACSCanto II 8-color flow cytometry system (BD). Imaging cytometry experiments were carried out using an ImageStream system (Amnis) and were analyzed as previously described ([24]).

Quantification of serum IgG and 9G4+ IgG levels

Concentrations of total IgG, total IgM, and 9G4+ IgG in all serum samples were quantified by enzyme-linked immunosorbent assay (ELISA), as described previously ([16]).

Microscopic analysis

Images were obtained using an Olympus FV1000 confocal laser scanning microscope at the URMC Confocal and Conventional Microscopy Core. A 60× objective (1.42 numerical aperture) was used to obtain high-resolution images within the linear range, and parameters were kept identical between the incubated serum samples from the SLE patients and the healthy donors.

Monoclonal antibody generation

Monoclonal antibodies were generated from single-cell–sorted memory B cells by polymerase chain reaction amplification of the immunoglobulin heavy and light chains, followed by expression vector cloning and transient transfection of 293T cells, as described previously ([25, 26]).

Statistical analysis

Flow cytometry data were analyzed with FlowJo software (Tree Star), and all reported intensities are the median fluorescence intensity (MFI). GraphPad Prism version 5 software was used for statistical analysis. Correlation of nonparametric variables was determined using Spearman's rank correlation. Odds ratios (ORs) were analyzed by treating variables as dichotomous variables using the median value as the threshold value. For IgG and IgM anti-AC antibodies, the mean + 2SD of the values in the healthy donor samples was used as the threshold. Active disease was defined as a SLEDAI score above the median score of 4.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgments
  8. REFERENCES

Binding of late ACs by 9G4+ antibodies from SLE serum

Anti-AC antibodies in SLE patients have been reported to have different, sometimes opposing, functions ([3, 10, 14]). In addition, our preliminary results in SLE patients ([27, 28]) as well as patients with chronic lymphocytic leukemias ([29]) have demonstrated the ability of 9G4+ antibodies to bind ACs. In order to systematically study anti-AC antibodies in SLE and to understand the contribution and properties of 9G4+ anti-AC antibodies, we first established a flow cytometry assay to measure both 9G4+ anti-AC antibodies and global serum anti-AC antibodies of different isotypes. This assay is based on the detection of antibody bound to Jurkat cells after camptothecin-induced apoptosis, as illustrated in Figure 1 for 9G4+ anti-AC antibodies. These experiments were performed using healthy control sera or sera from SLE patients selected based on elevated levels of serum 9G4+ antibodies as determined by ELISA ([16]). Specificity for antigens exposed on ACs but not on viable cells was established by documenting a lack of reactivity with untreated cells. Given the demonstrated ability of 9G4+ antibodies to bind the B220 isoform of CD45 ([16, 30]), the CD45-deficient Jurkat clone J45.01 was used to avoid any global antilymphocyte reactivity.

image

Figure 1. Binding of 9G4+ antibodies to late apoptotic cells (ACs) and apoptotic bodies in sera from patients with systemic lupus erythematosus (SLE). A, Separation of Jurkat cells into live (annexin V– and 7-aminoactinomycin D [7-AAD]–negative), early apoptotic (annexin V+ and 7-AAD−), or late apoptotic (annexin V+ and 7-AAD+) populations after incubation with serum from healthy control donors (HCD) or SLE patients and labeling with phycoerythrin (PE)–annexin V and 7-AAD. Histograms show fluorescein isothiocyanate (FITC)–labeled 9G4 early apoptotic (left) and late apoptotic (right) populations. Shaded histograms represent healthy donor serum–treated cells; open histograms represent SLE serum–treated cells. B, Separation of treated Jurkat cells into live cells (yellow) and ACs (orange) based on nuclear morphology, as determined by plotting nuclear size (DRAQ5 area) (left) versus nuclear fragmentation (DRAQ5 bright detail intensity) (right). ACs could be further separated into early ACs (blue), late ACs (magenta), and apoptotic bodies (red) based on membrane integrity (7-AAD intensity) and cell size (brightfield area). C, Representative images of each cell population, showing brightfield (BF), 9G4+ binding (green), 7-AAD (red), and DRAQ5 (DQ5; magenta). D, Frequency (top) and intensity (bottom) of 9G4+ antibody binding. Only 7-AAD+ cells stained brightly for 9G4+, providing further evidence that 9G4+ anti-AC antibodies bind to late ACs (magenta) and some apoptotic bodies (red), but not to early ACs (blue).

Download figure to PowerPoint

The percentages of viable cells, early ACs, and late ACs were similar after incubation with SLE serum or healthy donor serum, demonstrating that under these conditions, SLE serum was not cytotoxic (Figure 1A). The degree of 9G4 staining of early ACs was low and did not differ between healthy donor serum and SLE serum. In contrast, 9G4 staining was much higher in late ACs incubated with SLE, but not healthy donor, serum. This observation was confirmed using flow-based single-cell microscopy. Using combined staining with 7-AAD and DRAQ5, Amnis ImageStream analysis provided more accurate morphometric separation of live, early apoptotic, and late apoptotic populations and allowed the direct visualization of apoptotic bodies and localization of nuclear material (Figures 1B and C) ([24]). This approach demonstrated that 9G4+ antibodies bound strongly to 7-AAD+ late ACs and more weakly to apoptotic bodies, but not to live cells or early ACs (Figure 1D). Additionally, small DRAQ5+ particles consisting primarily of nucleic acids did not show 9G4+ staining.

Frequency of 9G4+ anti-AC antibodies in SLE

The frequency of 9G4+ anti-AC antibodies in serum samples obtained from a group of 60 unselected SLE patients and 9 healthy donors was determined by flow cytometry of camptothecin-treated Jurkat cells. Consistent with the results described above, only 7-AAD+ late ACs incubated with SLE serum bound 9G4+ antibodies; 7-AAD– ACs or cells incubated with healthy donor serum did not (Figure 2A). When positivity was defined as an MFI greater than the mean + 2SD of the value in healthy donors (MFI >328), binding for 9G4+ anti-AC antibodies was noted in 60% of the SLE patients, with varying degrees of reactivity (Figure 2B). One healthy donor sample showed a low level of 9G4+ anti-AC antibodies at this threshold, with an MFI of 331. These results were validated in a larger group of 31 healthy donors in which 1 sample had 9G4+ anti-AC antibodies (data not shown). This frequency is consistent with the observed occurrence of autoreactivity in the general population ([31]).

image

Figure 2. Differential prevalence of 9G4+ anti-AC antibodies in SLE patients and healthy donors. A, Representative histograms from late apoptotic (low FSC, high 7-AAD) or viable (high FSC, low 7-AAD) Jurkat cells after incubation with sera from healthy donors or SLE patients (left). Histograms show FITC-labeled 9G4 early apoptotic (left) and late apoptotic (right) populations. Shaded histograms represent healthy donor serum–treated cells; open histograms represent SLE serum–treated cells. B, Median fluorescence intensity (MFI) of 9G4+ anti-AC antibodies after incubation with serum from SLE patients or healthy donors. Each data point represents a single subject; broken line indicates the cutoff for positivity (mean + 2SD of the MFI in healthy donors). C, Correlation between serum concentrations of 9G4+ IgG and 9G4+ anti-AC antibody levels in SLE patients. Not all samples with high levels of serum 9G4+ IgG had high levels of 9G4+ anti-AC antibodies. Broken lines indicate the threshold for positivity of the respective parameters. See Figure 1 for other definitions.

Download figure to PowerPoint

We next investigated the correlation between the levels of total 9G4+ antibodies, which we previously reported to account for 10–40% of all IgG in SLE serum ([17]), and the MFI of the binding of 9G4+ anti-AC antibodies. Levels of serum 9G4+ IgG antibodies >0.2 mg/ml were present in 75% of patients, and the 9G4+ IgG antibody titers correlated with the MFI of 9G4+ anti-AC antibodies (r = 0.313, P = 0.014) (Figure 2C). However, approximately one-third of patients with elevated levels of 9G4+ IgG antibodies did not demonstrate anti-AC reactivity, indicating that binding to ACs is not an intrinsic, universal property of antibodies expressing the 9G4+ idiotype encoded by the germline VH4–34 heavy chain. Rather, our data showed that this property is present in only a fraction of 9G4+ antibodies that appear to be selected for in a significant portion of the SLE population.

Presence of 9G4+ anti-AC antibodies in samples with global IgM or IgG anti-AC antibodies

To better understand the participation of 9G4+ antibodies in the anti-AC response in SLE, we determined the overall anti-AC antibody binding of total IgM and IgG antibodies. Anti-AC antibodies of both isotypes were detected in SLE sera, with 32% of samples exhibiting IgM anti-AC antibodies and 41% IgG anti-AC antibodies (Figure 3A). In contrast, both isotypes were rare in healthy donors. In a large group of healthy donors, 10% had low levels of IgG anti-AC antibodies and 3% had IgM (n = 31) (data not shown).

image

Figure 3. Presence of 9G4+ anti-AC antibodies in a majority of SLE sera with global IgM and IgG anti-AC antibodies. A, IgM (left) and IgG (right) median fluorescence intensity (MFI) of 7-AAD+ Jurkat cells incubated with sera from SLE patients with or without 9G4+ anti-AC antibodies or sera from healthy donors. Each data point represents a single subject; broken line indicates the cutoff for positivity (mean + 2SD of the MFI in healthy donors). B, Percentage of IgM+ and/or IgG+ samples from SLE patients with anti-AC antibodies that did (solid area) or did not (open area) also contain 9G4+ anti-AC antibodies. C, Relationship between the MFI of IgG, IgM, and 9G4+ anti-AC antibodies in SLE samples with (darkly shaded circles) and without (lightly shaded circles) 9G4+ anti-AC antibodies (left). Broken lines indicate the threshold for positivity of the respective parameters. The percentages of samples with neither IgM nor IgG anti-AC antibodies (open area), only IgM anti-AC antibodies (lightly shaded area), only IgG anti-AC antibodies (darkly shaded area), or both IgM and IgG (solid area) for SLE samples with and without 9G4+ anti-AC antibodies are also shown (right). See Figure 1 for other definitions.

Download figure to PowerPoint

In the SLE samples with IgM or IgG anti-AC antibodies, samples with 9G4+ anti-AC antibodies predominated, making up more than two-thirds (71%) of the samples (Figure 3A). Furthermore, sera with 9G4+ anti-AC antibodies had the highest levels of global IgM or IgG anti-AC antibodies. Thus, when samples with IgG anti-AC antibodies were ranked by IgG MFI, 90% of the top one-third of samples also had 9G4+ anti-AC antibodies, and the top one-third of samples with IgM anti-AC antibodies all had 9G4+ anti-AC antibodies. When IgM and IgG anti-AC antibodies were analyzed together, 22% of samples had both IgM and IgG activity (Figure 3B). The distribution of IgG and IgM differed between samples with and those without 9G4+ anti-AC antibodies (P = 0.008 by chi-square test). Samples with 9G4+ anti-AC antibodies were more likely to have IgM anti-AC antibodies, in particular, IgM and IgG anti-AC antibodies together, than samples without (Figure 3C). In fact, no samples that lacked 9G4+ anti-AC antibodies had anti-AC antibodies of both isotypes.

Some samples with 9G4+ anti-AC antibodies had neither IgM nor IgG anti-AC antibodies. This discrepancy could be explained by differences in the detection sensitivity of the assays we used, or it may indicate the presence of 9G4+ IgA anti-AC antibodies. The latter possibility is supported by a separate analysis of SLE 9G4+ IgA monoclonal antibodies performed in our laboratory that demonstrated significant anti-AC activity among 9G4+ IgA antibodies ([26]).

Co-occurrence of AC binding and B cell reactivity of 9G4+ antibodies in some SLE patients

As shown in Figure 2, some SLE patients had no anti-AC reactivity despite having high levels of serum 9G4+ IgG antibodies. The 9G4+ antibodies have been reported to react with naive B cells (B cell binding) through their ability to bind surface B220, and this reactivity is common in patients with SLE ([16]). Therefore, we determined the relationship between these two types of 9G4+ autoreactivity in the same cohort of SLE patients. While AC and B cell specificities were simultaneously present in many samples, other samples had only one or the other (Figure 4A). Because of this, 9G4+ anti-AC antibodies and 9G4+ B cell binding were not correlated. Of the samples with 9G4+ AC or B cell reactivity, 25% had only AC binding, and a few samples (12%) had only B cell reactivity (Figure 4B).

image

Figure 4. Distinct 9G4+ AC and B cell binding in sera from patients with SLE. A, Representative histograms of AC (top) or naive B cell (bottom) staining after incubation with serum from a healthy donor (open histograms) or sera from 3 different SLE patients (shaded histograms). B, Median fluorescence intensity (MFI) of 9G4+ in naive B cells (IgD+CD27–) and ACs incubated with serum from a healthy donor (○) and an SLE patient (•) (left). Broken lines indicate the threshold for positivity of the respective parameters. The percentage of each type of 9G4+ binding for the samples with reactivity is also shown (right). C, Monoclonal antibody staining of ACs (top) or naive B cells (bottom). Open histograms represent 152B5, an antibody negative for both specificities; shaded histograms represent the 3 monoclonal antibodies indicated. See Figure 1 for other definitions.

Download figure to PowerPoint

Serum is a mixture of antibodies with differing specificities, and samples with both specificities could consist of a mixture of AC and B cell reactivities rather than antibodies with shared specificity against both targets. To clarify how individual antibodies contributed to B cell and apoptotic reactivities, we generated recombinant 9G4+ monoclonal antibodies from single-cell–sorted memory B cells (IgD–CD27+) from an SLE patient. As illustrated by the representative examples in Figure 4C, serum binding profiles (either AC binding, B cell binding, or both) were recapitulated by 9G4+ monoclonal antibodies. These distinct autoreactivity profiles have been confirmed with a large panel of monoclonal antibodies whose binding properties and structural correlates will be published separately ([26]).

Association of 9G4+ anti-AC antibody binding with high levels of disease activity, lupus nephritis, and anti–double-stranded DNA (anti-dsDNA) antibodies

To evaluate disease correlates and to begin to understand their pathologic implications, we determined the association between disease activity, as measured by the SLEDAI, and the MFI of 9G4+ anti-AC antibody binding. We also determined associations between the SLEDAI scores and the concentrations of 9G4+ IgG antibodies, 9G4+ B cell–binding antibodies, IgG anti-AC antibodies, and anti-dsDNA antibodies.

The MFI of 9G4+ anti-AC antibodies correlated positively with disease activity (r = 0.322, P = 0.012) (Figure 5A). In contrast, 9G4+ binding to B cells correlated only weakly with the SLEDAI scores (r = 0.201, P = 0.123) (data not shown). When 9G4+ anti-AC antibodies and SLEDAI scores were analyzed as dichotomous variables, patients with high levels of 9G4+ anti-AC antibody binding were more likely to exhibit active disease than were those with low levels of binding (OR 4.76, P = 0.008). As previously reported, serum concentrations of 9G4+ IgG antibodies also correlated positively with the SLEDAI score (r = 0.266, P = 0.041) (Figure 5B). However, almost all samples from patients with high serum concentrations of 9G4+ IgG antibodies and active disease (85%) also had high levels of 9G4+ anti-AC antibodies. In the group of patients with high serum levels of 9G4+ IgG antibodies, those who also had high levels of 9G4+ anti-AC antibodies were more likely to have active disease (OR 7.7, P = 0.032) than were those with low levels.

image

Figure 5. Likelihood of active disease in SLE patients with high levels of 9G4+ anti-AC antibodies. A, Correlation between the 9G4+ anti-AC antibody median fluorescence intensity (MFI) and the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) scores. B, Correlation between the intensity of 9G4+ IgG and the SLEDAI scores for samples with (shaded circles) and without (open circles) 9G4+ anti-AC antibodies. C, Correlation between the 9G4+ anti-AC antibody MFI and the anti–double-stranded DNA (anti-dsDNA) titer (left) and between the anti-dsDNA titer and the SLEDAI scores for samples with (shaded circles) and without (open circles) 9G4+ anti-AC antibodies (right). Broken lines in A, B, and C indicate the median value of the measured parameter or a SLEDAI score greater than the median score of 4. D, Correlation between the IgG anti-AC antibody MFI and the SLEDAI scores for samples with (shaded circles) and without (open circles) 9G4+ anti-AC antibodies (left). Broken lines indicate the threshold for positivity of the respective parameters. Percentages of patients with IgG anti-AC antibodies who have (shaded area) and do not have (open area) active disease, according to the presence and absence of 9G4+ anti-AC antibodies and IgM anti-AC antibodies, are also shown (right). See Figure 1 for other definitions.

Download figure to PowerPoint

Because of the clinical significance of lupus nephritis and its association with high serum levels of 9G4+ antibodies ([20]), we also examined 9G4+ anti-AC antibodies in patients with lupus nephritis. Seven of the 60 SLE patients had lupus nephritis, and in a separate experiment, we recruited an additional 9 patients with lupus nephritis. In this analysis, we found that 9G4+ anti-AC antibodies were much more prevalent in patients with lupus nephritis (88%) than in those without (57%) (P = 0.021 by Fisher's exact test). The 9G4+ anti-AC antibodies were found both in patients with active nephritis and in those with stable disease. However, those with active nephritis had some of the highest levels of 9G4+ anti-AC antibodies, and the 2 lupus nephritis patients without 9G4+ anti-AC antibodies both had stable disease.

Titers of anti-dsDNA antibodies determined at the same time as the levels of 9G4+ anti-AC antibody binding were available for 37 patients. Anti-dsDNA titers were highly correlated with the MFI of the 9G4+ anti-AC antibodies (r = 0.759, P < 0.001) (Figure 5C). Consistent with previous studies ([32]), we found that serum anti-dsDNA titers correlated with the level of disease activity (r = 0.433, P = 0.008). However, patients with both high levels of 9G4+ anti-AC antibodies and anti-dsDNA antibodies were more likely to have active disease (80%; OR 9.5, P = 0.012) than were patients grouped only for the presence of high titers of anti-dsDNA (67%; OR 7.5, P = 0.007).

As was the case for the 9G4+ anti-AC antibody levels, the levels of IgG anti-AC antibodies correlated with disease activity, but to a lower degree (r = 0.271, P = 0.034) (Figure 5D). Patients with IgG anti-AC antibody–positive serum that also had 9G4+ anti-AC antibodies were much more likely to have active disease (75%) than were those without 9G4+ anti-AC antibodies (25%) (OR 9, P = 0.020) (Figure 5D). IgM anti-AC antibodies were neither negatively nor positively correlated with the SLEDAI scores and were not protective in patients with IgG anti-AC antibodies. Patients with both IgM and IgG anti-AC antibodies were actually more likely to have active disease than were those with IgG anti-AC antibodies alone (62% versus 43%), although this difference was not statistically significant.

Perinuclear localization of 9G4+ antibody binding to ACs

Several distinct antigens, including oxidized lipids ([33]), nucleic acids ([9]), ribonucleoproteins ([34]), and nucleosomes ([35]), have been defined as targets of antiapoptotic antibodies. During apoptosis, these antigens cluster in apoptotic blebs on the cell surface ([36]). The regions of Jurkat ACs recognized by 9G4+ anti-AC antibodies were examined by confocal microscopy of two different SLE samples with different anti-AC antibody–binding strengths (intermediate [SLE1] and strong [SLE2]). In contrast to healthy donor sera, which did not generate significant staining, strong 9G4+ antibody staining was clearly detected with both SLE serum samples (Figure 6), and 2 different staining patterns were observed. While 9G4 staining was predominantly perinuclear in both cases, serum SLE1 created a spotted pattern, with small dots outlining the nucleus (Figure 6A); serum SLE2 displayed 9G4 staining in a much more homogeneous perinuclear ring pattern (Figure 6B). In this sample, 9G4+ antibodies also clearly bound apoptotic blebs.

image

Figure 6. Localization of 9G4+ antibodies to the perinuclear space and apoptotic blebs in systemic lupus erythematosus (SLE) patients. Confocal microscopy was performed on apoptotic Jurkat cells incubated with sera from healthy donors (HCD1 and HCD2) or SLE patients (SLE1 and SLE2) and then stained with 9G4 (green) and with the nuclear stain DRAQ5 (red). A, Localization of 9G4+ staining to perinuclear spots (white arrowheads) in cells incubated with serum from patient SLE1. Image at the right is a digital enlargement of the image at the left. B, Localization of 9G4+ staining to thick perinuclear rings (blue arrowheads) and apoptotic blebs (yellow arrowheads) in cells incubated with serum from patient SLE2. Some nuclear fragments have no associated 9G4+ staining (orange arrowheads). Original magnification × 60.

Download figure to PowerPoint

Consistent with the morphometric analysis shown in Figure 1, only limited binding to small apoptotic bodies was observed; many particles consisting of nucleic acid had no 9G4 staining. Because 9G4+ antibodies did not colocalize with nucleic acids, we used a separate group of 25 patients to examine the association between 9G4+ anti-AC antibodies and other common SLE autoreactivities. We found that patients with high levels of 9G4+ anti-AC antibodies were more likely to have anti-RNP, anti-Sm, and anti-Ro antibodies. The difference in anti-Ro was most evident, with 63% of patients with high levels of 9G4+ anti-AC antibodies also having antiRo, as compared to 21% of patients with low levels of 9G4+ anti-AC antibodies (P = 0.042 by Fisher's exact test).

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgments
  8. REFERENCES

Antinuclear antibodies are not uncommon in otherwise healthy individuals ([31]), which demonstrates that a breakdown in tolerance is insufficient for the occurrence of autoimmune disease and that tolerance breakdown alone cannot explain why a person develops a specific autoimmune disease. Instead, selection by disease-specific self antigens is an important factor in determining how autoimmunity is manifested. These considerations are of central significance in SLE, a systemic autoimmune disease characterized by the presence of multiple autoantibodies, some of which (anti-dsDNA, anti-Sm, and anti–ribosomal P antibodies) are highly specific for the disease. Anti-AC antibodies constitute another relevant autoantibody system, as SLE is characterized by defective clearance of ACs, which express a high density of SLE immunogens.

Anti-AC antibodies can mediate multiple pathogenic mechanisms, including tolerance breakdown, epitope spreading, and further induction of autoantibodies capable of both amplifying inflammation and directly inducing tissue damage ([8, 12, 37-39]). Elevated levels of 9G4+ antibodies are also specific for SLE and are prevalent in patients with active SLE ([16]). The canonical intrinsic autoreactivity of 9G4+ antibodies is imparted by a germline-encoded hydrophobic patch that mediates binding to N-acetyl-lactosamine (NAL) sugar and accounts for the striking anti–B cell autoreactivity of 9G4+ antibodies in vivo and in vitro ([16]). NAL sugar chains have been shown to be exposed on the surface of ACs ([40, 41]), and furthermore, our preliminary results ([27, 28]) indicated the potential of 9G4+ antibodies to bind ACs. Additionally, defective censoring leads to the accumulation of 9G4+ B cells in SLE germinal centers ([17]), sites of accumulation of uncleared ACs ([4]).

In this study, we tested the hypothesis that defective tolerance and the localization of 9G4+ B cells in germinal centers would result in selection of 9G4+ antibodies with anti-AC antibody reactivity that would contribute to the global anti-AC antibody response in SLE. The results presented here support this hypothesis. Elevated levels of 9G4+ anti-AC antibodies were present in the majority of our unselected patients with SLE (60%). Moreover, the 9G4+ anti-AC antibody MFI correlated with the degree of disease activity as measured by the SLEDAI, and patients with high levels of 9G4+ anti-AC antibodies were significantly more likely to have active disease and lupus nephritis. While the concentrations of IgG anti-AC antibodies, serum 9G4+ IgG, and anti-dsDNA antibodies also correlated with disease activity, patients with high values for these correlates who also had high levels of 9G4+ anti-AC antibodies were more likely to have active disease than were those with lower levels of 9G4+ anti-AC antibodies. Taken together, our results identify 9G4+ anti-AC antibodies as an important marker of disease activity and possibly of disease severity, an intriguing finding whose clinical implications need to be conclusively validated by ongoing longitudinal studies.

During normal homeostasis, the phagocytosis of ACs is generally antiinflammatory ([42]). This tolerogenic antiinflammatory response is altered in SLE patients because of both changes in SLE monocytes ([43]) and a shift to Fc receptor internalization through anti-AC antibodies ([44]). SLE anti-AC antibodies alone are sufficient to promote inflammation, as healthy donor monocytes incubated with necrotic cell material and phagocytosis-promoting SLE antibodies produce large amounts of inflammatory cytokines ([12]). The abundance of 9G4+ antibodies in SLE means 9G4+ anti-AC antibodies are likely an important part of this process. Additionally, because 9G4+ antibodies can recognize a diverse array of self antigens ([26, 45]), 9G4+ B cells presenting AC-derived antigens may result in epitope spreading that expands and perpetuates anti-AC responses and enhance antigen presentation of AC-derived T cell epitopes, such as histone peptides from nucleosomes ([46]).

Levels of 9G4+ anti-AC antibodies and anti-dsDNA antibodies correlated with each other and with disease activity but were nonetheless dissociated in a significant proportion of patients. This indicates that these two autoantibody species, while often produced concurrently by patients with active disease, recognize separate antigens. This conclusion is further supported by the colocalization results of the ImageStream and confocal microscopy studies (Figures 1 and 5, respectively), in which 9G4+ antibody binding to ACs did not merge with nucleic acid staining, and many DRAQ5-staining nuclear bodies had no associated 9G4 staining. This suggests that 9G4+ antibody binding to ACs is not mediated by the ability of a fraction of these serum antibodies to recognize DNA. Although it is only correlative, the higher incidence of anti-Ro antibodies in patients with 9G4+ anti-AC antibodies is consistent with our microscopy findings, as Ro localizes to apoptotic blebs ([47]) and to the perinuclear space during apoptosis ([48]).

While the precise antigenic targets recognized by 9G4+ antibodies in ACs remains to be determined, their reactivity with other relevant antigens and established structure–function correlations provide important clues. Since 9G4+ idiotype expression, NAL binding, and B cell binding activity are dependent on conservation of the VH4–34 germline–encoded hydrophobic patch in framework region 1 ([26, 49]), the anti-AC antibody reactivity of 9G4+ antibodies could potentially also have similar structural requirements and would therefore segregate with B cell binding activity. This model would point to recognition of antigens shared between the two cellular targets, NAL in particular. Several lines of evidence, however, do not support this possibility. We observed 9G4+ anti-AC antibodies in the absence of B cell binding activity in 23% of patients, and a smaller percentage of patients displayed only 9G4+ B cell binding. Moreover, separate studies of monoclonal antibodies have indicated that the apoptotic and B cell reactivities of 9G4+ antibodies are dependent on different regions of the VH4–34 heavy chain and are likely determined by distinct antigens on the corresponding cellular targets ([26]). That 9G4+ anti-AC antibody reactivity is not largely due to the canonical anti-NAL binding is also supported by the observation that 9G4+ antibodies in the sera of patients infected with the human immunodeficiency virus bind to B cells ([50]) but not to ACs (Jenks SA, et al: unpublished observations).

It is possible that 9G4+ anti-AC antibodies could originate from 9G4+ B cells that initially recognize canonical NAL antigens on B cells, red blood cells, or other tissues that are subsequently diversified by somatic hypermutation and selection by ACs. Yet, given the continuous presence of B cells that could exert selective pressure in the germinal centers, this model would also predict the persistence of 9G4+ B cell binding autoreactivity concurrently with the development of anti-AC antibody reactivity. Our results are more consistent with independent triggering and selection of different 9G4+ B cell clones by either B cells or ACs, with the provision that retention of the VH4–34 germline–encoded hydrophobic patch in framework region 1 would endow most AC-reactive 9G4+ antibodies with B cell binding activity as well.

Irrespective of the molecular underpinnings and temporal events of selection, these results demonstrate that like other lupus autoantibodies, while tolerance against AC antigens in a significant proportion of lupus patients is broken, this tolerance breakdown is not a universal phenomenon in SLE. Approximately 20% of patients with elevated 9G4+ IgG levels did not demonstrate anti-AC antibody reactivity, indicating that binding to ACs is not an intrinsic property of antibodies expressing the 9G4+ idiotype. Rather, this property is present in only some 9G4+ antibodies that appear to be selected in a segment of the SLE population.

Taken together, the data presented here strongly suggest that measurement of anti-AC antibodies in general and 9G4+ anti-AC antibodies in particular provide a useful tool with which to assess disease activity in SLE. Moreover, the presence of these antibodies may provide a way to segment lupus patients with different genetic and immunologic defects. Thus, a better understanding of the origin and consequences 9G4+ antibodies against ACs has the potential to increase our understanding of the pathogenesis of SLE and provide new approaches to treatment.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgments
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Sanz 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 conception and design. Jenks, Palmer, Sanz.

Acquisition of data. Jenks, Palmer, Marin, Hartson, Chida, Richardson.

Analysis and interpretation of data. Jenks, Palmer, Marin, Sanz.

Acknowledgments

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgments
  8. REFERENCES

We thank Dr. Linda Callahan (URMC Confocal and Conventional Microscopy Core) for microscopy assistance, Dr. Timothy Bushnell and the University of Rochester Medical Center Flow Cytometry Core for flow cytometry assistance, and Bridget Neary and Dr. Emily Blalock for editing assistance.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgments
  8. REFERENCES