Evidence for intranasal anti-nuclear autoantibodies in Chronic Rhinosinusitis with Nasal Polyps.

Introduction

Chronic rhinosinusitis (CRS) is a clinical syndrome associated with persistent inflammation of nasal and paranasal sinuses mucosa. This definition encompasses the two most common variants of this disease - a form with nasal polyps (CRSwNP) and a non-polypoid form (CRSsNP) that have clinically and morphologically different characteristics¹. Historically, CRSsNP was considered as an incompletely treated case of acute rhinosinusitis resulting in chronic infection. CRSwNP was considered a distinct, non-infectious disorder of unclear etiology, perhaps related to atopy². While both forms of disease utilize surgery as a modality for improving paranasal sinus drainage and relieving patient symptoms, medical therapy for both forms utilizes antibiotics and corticosteroids^3-5. Treatment success for either form of CRS is variable with no currently established molecular predictors to guide choice of therapy or predict outcome.

Emerging research from our laboratory highlights a potentially important pathogenic role for B-lymphocytes in the inflammation associated with CRS. We have shown that the B-cell activating factor of the TNF family (BAFF, also called BLys, or TNFSF13B) is highly elevated in nasal polyp tissue in CRSwNP in comparison to CRSsNP, control and unaffected tissue in CRSwNP⁶. We found that BAFF is produced by epithelial cells and could be induced by stimulation with several cytokines and innate immune activators⁷. BAFF is a potent stimulator of B-cell proliferation and class switching in B-cells, and mice over-expressing BAFF manifest systemic autoimmunity^{8, 9}. In addition to BAFF, we have also found nasal polyps contain elevated levels of the cytokine IL6 and chemokines such as BLC (CXCL13) and SDF-1a which are known to play a role in B-cell recruitment and plasma cell differentiation^{10, 11}. We have proposed that these findings may account for the increased levels of IgA and IgG present in nasal polyp tissue, germinal-center like pseudofollicles and consistently high numbers of B cells and plasma cells^{10, 12}. At present, the nature of the antigen specificity of these B cells and their roles in pathogenesis of nasal polyposis remain unclear¹³.

The nature and specificity of the immunoglobulins found in nasal polyps has not been explored in detail. In 1974, Bass et al. examined the distribution of the different immunoglobulin subtypes in nasal polyps and found no significant IgM, IgA was found in numerous plasma cells in the sub-epitelial and periglandular regions of nasal polyps while IgG was found diffusely deposited thoughout the stroma¹⁴. Due to the proposed atopic etiology of nasal polyposis, the vast majority of the research focuses on IgE. Gevaert et al demonstrated a polyclonal IgE hyperglobulinemia along with elevated levels of specific IgE against aeroallergens and staphylococcus enterotoxins compared to serum¹². Similarly, a more recent study by Sabirov et al. demonstrated the presence of elevated levels of IgE against Alternaria alternata relative to serum¹⁵. Tellingly, similar elevations in local specific IgG and IgA suggested that local excess production of immunoglobulin was not isolated to the IgE isotype. We hypothesized that the local mucosal inflammatory microenvironment and chronic inflammation associated with CRSwNP was conducive to the expansion of autoreactive B-cell clones that may play a role in perpetuating inflammation. In this study, we examined nasal tissue for the presence of class switched autoantibodies as evidence that such phenomena might exist in CRS.

Methods

Results

Clinical and demographic characteristics of the patients enrolled in this study are shown in Table 1. Twelve of the patients with CRSwNP and six of the patients with CRSsNP were undergoing revision nasal surgery. Total IgA levels and IgG levels were elevated in nasal polyp tissue extract relative to control nasal tissue extracts (See Figure 1A and 1B respectively).

Analysis of autoantibody microarray results

The results of the autoantibody microarray demonstrated measurable IgG and IgA autoantibody levels against 25 of the 62 antigens and 24 of the 62 antigens respectively. Statistical analysis revealed that relative to control nasal tissue, statistically significant elevations of several autoantibodies were found in nasal polyp tissue. There were no autoantibodies that were elevated in control nasal tissue extracts relative to nasal polyp extracts. Results of the statistically significant comparisons for both the IgG and IgA isotype antibodies are depicted in Table 2 along with several representative non-significant differences. There were more positive results for IgG autoantibodies than IgA. Revealingly, we observed that a large number of elevated autoantibodies were reactive against nuclear antigens. Interestingly, the elevated levels of autoantibodies were confined to nasal polyp tissue extract and were not observed in the inferior turbinates from patients with CRSwNP.

EIA based analysis of anti-dsDNA levels in nasal tissue

We next attempted to confirm the elevations of both the IgA and IgG autoantibodies focusing on anti-dsDNA in light of the well-described pathogenic potential of anti-dsDNA antibodies in lupus and the high sensitivity of the commercially based assays. A quantitative EIA utilizing recombinant human DNA to specifically assay for dsDNA (Alpco Diagnostics, Salem, NH) binding was utilized. An expanded set of tissue comprising extracts of nasal polyps (P) and nasal tissue from the inferior turbinates (IT) and uncinate processes (U) obtained from patients with CRSwNP, CRSsNP and controls was used. These results demonstrate a 6 and 7-fold higher level of anti-dsDNA IgA (Figure 1C) and IgG (Figure 1D) antibodies respectively in nasal polyp tissue when compared with control inferior turbinate tissue. Multi-group comparisons revealed statistically significant elevations of anti-dsDNA IgG and IgA antibodies with significant post-hoc pairwise comparisons of nasal polyps to nasal tissue from patients with CRSsNP and controls. Within CRSwNP, statistically significant elevations were found in some nasal polyp tissue relative to the inferior turbinate but not the uncinate processes. The latter observation is consistent with clinical observations that the uncinate process is more closely involved with the inflammatory process in CRSwNP.

Correlation of IgG anti-dsDNA to Total IgG levels

The lack of measurable autoantibodies in over half the antigens on the microarray makes it unlikely that these observations are secondary to a non-specific binding from elevated levels of IgG and IgA in nasal polyp extract. To assess whether the elevated levels of anti-dsDNA antibodies were due to non-specific elevated immunoglobulin levels, the anti-dsDNA IgG results were normalized to total IgG levels (Figure 2A). This revealed a statistically significant multi-group comparison although post-hoc testing did not reveal a specific binary comparison that was statistically significant. Within nasal polyp tissue, there was no correlation between anti-dsDNA autoantibody levels and total IgG (Figure 2B). However, data derived from tissue samples from non-CRSwNP patients showed a positive correlation between anti-dsDNA IgG autoantibody levels and Total IgG (Figure 2C). These results suggest that the production of anti-dsDNA IgG antibodies in nasal polyps was independent of total IgG levels. There was a positive correlation between the anti-dsDNA IgG and anti-dsDNA IgA levels (Figure 2D) suggesting that the production of the two isotypes is driven by related processes.

Correlation of CRSwNP disease activity with locally elevated levels of anti-dsDNA

We then correlated anti-dsDNA autoantibody levels with the clinical parameters collected on our patients. We found that elevated levels of the IgG anti-dsDNA autoantibodies were found at higher levels (Figure 3A) and more frequently (Figure 3C) in patients with more aggressive clinical course requiring revision surgery. There was a trend toward higher levels of anti-dsDNA antibodies in higher-grade nasal polyps (Figure 3F). None of the patients who were receiving revision surgery for CRSsNP, had elevated levels of anti-dsDNA IgG antibodies (Figure 3C). Anti-dsDNA autoantibodies were not affected by asthma status or atopic status (Figure 3D and 3E respectively). In contrast, aggressive nasal polyps requiring revision surgery was not correlated with total IgA and IgG levels (Figure 3B).

Direct and Indirect Immunoflorescence

Using direct immunofluorescence of nasal polyps, we frequently observed extensive deposition of IgG within the stroma of the nasal polyps. Most IgG was deposited extracellularly with some regions enriched in IgG containing plasma cells (not shown). In some samples, there was intraepithelial IgG deposition seen (Figure 4A). In contrast, control nasal tissue showed little extravascular IgG and few perivascular IgG secreting plasma cells (Figure 4B, plasma cell denoted by red arrow). In contrast, IgA in nasal polyps was largely confined to intracellular IgA in subepithelial and stromal plasma cells and within secretory cells within the glandular tissue (data not shown). Control tissue stained with isotype-controlled rabbit IgG showed no specific staining (data not shown).

Using a modified indirect immunofluorescence technique, we incubated sections of control uncinate process tissue with nasal tissue extract from a series of two control uncinate processes and eight nasal polyps with a range of anti-dsDNA IgG autoantibodies (range= 65 to 6332 U anti-dsDNA IgG /mg Total Protein) and examined the resulting sections using direct immunofluorescence. In this assay, if the autoreactive antibodies present in the nasal extract were predominantly anti-nuclear, a nuclear staining pattern would be anticipated, and the staining intensity would parallel the anti-dsDNA antibodies found in the tissue. The staining pattern when control uncinate tissue was incubated with control tissue extract was similar to those obtained from direct immunofluorescence of control uncinate tissue (Compare Figure 4B-direct immunofluorescence and Figure 4C-indirect immunofluorescence with control extract). In contrast, there was a specific nuclear staining pattern seen in some samples incubated with polyp extracts containing high levels of anti-dsDNA antibodies (Figure 4D). The Hep-2 based indirect immunofluorescence assay (Figure 4E) demonstrated anti-nuclear antibodies in some nasal polyp extracts. In this assay, heavy non-nuclear “spindle” staining was seen in addition to the nuclear staining being scored. This resulted in some difficulty in directly using the assay criteria established for scoring serum indirect immunofluorescence. However, examining only nuclear staining, nasal tissue extracts from samples with elevated anti-dsDNA had more intense nuclear staining (Figure 4F).

Discussion

Chronic rhinosinusitis (CRS) is a prevalent chronic inflammatory condition of the paranasal sinuses that affects approximately eight percent of the United States population. Recently, our laboratory demonstrated evidence for overproduction of BAFF and IL6 in nasal polyp tissue. Since both these cytokines are associated with autoimmunity, we sought to examine nasal polyps for evidence of autoreactive B-cells^{9, 19, 20}. The recent development of an autoantigen microarray enabled us to simultaneously query nasal polyp tissue for the presence of multiple autoantibodies¹⁸. As our microarray results suggest, a polyclonal, class switched autoantibody response against nuclear components, thyroid antigens and some epithelial antigens is found in nasal polyps relative to control nasal tissue and inflamed tissue from CRSsNP (Table 2).

Given their central pathogenic role in systemic lupus erythematosus (SLE), we chose to confirm the positive result for the anti-dsDNA antibodies on commercially available EIAs. In our analysis of nasal tissues from patients with CRSwNP, highly elevated levels of autoantibodies are found locally within some nasal polyps and more modestly in the uncinate process when compared with the clinically unaffected inferior turbinates (Figure 1). Uncinate process tissue in CRSsNP, which is frequently involved in the chronic inflammation, showed no elevation in the levels of anti-dsDNA antibodies suggesting that chronic inflammation alone does not result in elevated autoantibodies. Futhermore, the anti-dsDNA antibody production in nasal polyps was disproportionately elevated in comparison with total immunoglobulin levels (Figure 2). Clinically, nasal polyps obtained from patients undergoing revision surgery for recurrence frequently had elevated anti-dsDNA IgG antibodies compared with nasal polyps obtained from primary surgery. A similar trend was seen for the anti-dsDNA IgA antibodies although this did not achieve statistical significance (Figure 3A). In contrast, there was no evidence for increased autoantibodies in patients with CRSsNP who had multiple revision surgeries, suggesting that multiple surgeries themselves are not the trigger of this response (Figure 3C). Elevated levels of anti-dsDNA antibodies were found in some polyp samples obtained at initial surgery and it would be interesting to evaluate if they had clinical evidence for recurrence on longitudinal follow up. It is known that 25-75% of CRSwNP patients experience recurrence depending on the technique and length of follow up^{21, 22} and a biomarker to predict outcome following surgery would be clinically valuable. Given that there are surgery naïve CRSwNP patients with autoantibody elevations and several CRSsNP patients who had previous nasal surgery without evidence of elevated levels of autoantibodies, the possible interpretation that surgical trauma triggers autoantibody production is unlikely.

In systemic lupus erythermatosus (SLE), the presence of B cell and T cell autoimmunity to the nucleosome, and its individual components namely native dsDNA and histones, are important in establishing a diagnosis and correlate with the severity of clinical parameters such as lupus nephritis and lupus associated cognitive impairment^23-26. The proposed mechanisms by which anti-dsDNA antibodies induce nephritis include the formation of immune complexes with DNA/nucleosome components released from apoptotic cells and cross reactivity with components of the basement membrane^{27, 28}. While the pathogenicity of the anti-dsDNA antibodies in CRSwNP has still not been established, our data suggest that the presence of anti-dsDNA antibodies has implications on clinical parameters independent of total immunoglobulin levels. Whether these autoantibodies are directly pathogenic and bind to nucleosomal components or represent an epiphenomenon of a more severe and persistent form of sinonasal inflammation remains to be determined. These results also raise the possibility that memory B-cells reactive to self-antigens may persist after surgical removal of the inflamed tissue and potentially play a role in triggering recurrent inflammation.

While anti-nuclear antibody response is considered a dominant feature in SLE³¹, other organ-specific autoimmune diseases such as Sjogren’s syndrome and psoriasis also have elevated anti-dsDNA antibodies^{32, 33}. Experimental models suggest autoimmunity results from the initial activation of naïve T-cells through microbial molecular mimicry or superantigens and are subsequently amplified by the enhanced processing and presentation of autoantigens and bystander activation of lymphocytes in an inflamed site^{35, 36}. In CRSwNP, there is evidence for increased local viral and bacterial colonization providing the context by which pathogen derived peptides may cross activate autoreactive T or B cells^37-39. Furthermore, evidence for staphylococcus superantigen expansion with Vβ skewing of T cell populations exists in CRSwNP, providing another mechanism for non-selective activation of autoreactive T-cells^39-41. Following activation of these autoreactive B and T cells, epitope spreading and proliferation of these autoreactive clones can occur, thus expanding B-cells producing specific immunoglobulins to foreign antigens such as aeroallergens and colonizing microbial bacteria. The generation of autoantibodies and specific antibodies to extrinsic antigens is frequently found in other chronic inflammatory diseases of the epithelial interface. For example, in IBD, elevated levels of autoantibodies to perinuclear anti-neutrophil cytoplasmic antibodies (p-ANCA) along with antibodies to gut microflora such as Escherichia coli and Gram-positive anaerobic rods are found^42-44.

A limitation of this study was that we focused analysis to tissue specific autoantibodies largely due to limited availability of matching sets of serum and nasal lavage particularly in the revision cases in which the autoantibodies appear most elevated. It will be of value to perform a matched tissue, nasal lavage and serum study to clarify whether circulating systemic autoantibodies can be detected within a subset of patients with nasal polyps and whether these correlate with intranasal autoantibody levels. Given the lack of reports of associated autoimmune conditions⁴⁶ and the relatively limited mucosal surface area involved in CRS, it is probably unlikely that circulating autoantibodes will be significantly elevated. In addition to self-reactive IgG autoantibodies, our data suggest that IgA autoantibodies are present within nasal polyps. Since IgA is a potent stimulator of eosinophil degranulation, it raises the possibility of secreted autoreactive IgA antibodies propagating inflammation to more distal portions of the airway⁴⁷.

Currently, B-cells are increasingly recognized as important therapeutic targets in many autoimmune diseases. The importance of B-cells in linking innate and adaptive immunity, and their contribution toward long term immune memory is evident due to the success of rituximab in treating multiple autoimmune diseases previously thought to be T-cell driven processes⁴⁸. Interestingly, clinical experience with rituximab therapy in autoimmune disease demonstrates that levels of autoantibodies such as anti-dsDNA decrease more than the total immunoglobulin levels following therapy⁴⁹. These data suggest that autoantibodies are produced by shorter-lived CD20+ B-cell dependent plasma cells or plasmablasts. Additionally, an anti-BAFF agent, Belimumab, was recently approved as a novel treatment for SLE. Data from the clinical trials of Belimumab demonstrates that selective targeting of BAFF has clinical activity and reduces anti-dsDNA antibodies secondary to depletion of both naïve and transitional B-cells without affecting traditional memory B-cells. While we have much to learn about the anti-dsDNA producing B-cells in CRSwNP, their presence within the most therapeutically refractory patients provides new potential avenues for targeted therapy in this disease.