SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Objective

To identify genetic factors driving pathogenic autoantibody formation in collagen-induced arthritis (CIA), a mouse model of rheumatoid arthritis (RA), in order to better understand the etiology of RA and identify possible new avenues for therapeutic intervention.

Methods

We performed a genome-wide analysis of quantitative trait loci controlling autoantibody to type II collagen (anti-CII), anti–citrullinated protein antibody (ACPA), and rheumatoid factor (RF). To identify loci controlling autoantibody production, we induced CIA in a heterogeneous stock–derived mouse cohort, with contribution of 8 inbred mouse strains backcrossed to C57BL/10.Q. Serum samples were collected from 1,640 mice before arthritis onset and at the peak of the disease. Antibody concentrations were measured by standard enzyme-linked immunosorbent assay, and linkage analysis was performed using a linear regression–based method.

Results

We identified loci controlling formation of anti-CII of different IgG isotypes (IgG1, IgG3), antibodies to major CII epitopes (C1, J1, U1), antibodies to a citrullinated CII peptide (citC1), and RF. The anti-CII, ACPA, and RF responses were all found to be controlled by distinct genes, one of the most important loci being the immunoglobulin heavy chain locus.

Conclusion

This comprehensive genetic analysis of autoantibody formation in CIA demonstrates an association not only of anti-CII, but interestingly also of ACPA and RF, with arthritis development in mice. These results underscore the importance of non–major histocompatibility complex genes in controlling the formation of clinically relevant autoantibodies.

Rheumatoid arthritis (RA) is a complex autoimmune disease of unknown etiology involving chronic inflammation of the peripheral joints, which leads to cartilage destruction and bone erosion. The main genetic factor contributing to susceptibility to RA across species is the major histocompatibility complex (MHC), designated HLA in humans. Collagen-induced arthritis (CIA) is an animal model resembling human RA; in mice, CIA is strongly associated with the positionally identified class II MHC Aβ gene (1). Interestingly, the RA-associated class II MHC molecules expressed in the mouse system allow the development of CIA, thereby responding to the same CII peptide as Aq-expressing mouse strains (2). A hallmark of human RA is the development of autoantibodies directed against the Fc portion of immunoglobulins (rheumatoid factor [RF]), antibodies directed against citrullinated protein antigens (ACPA), and, in a subset of patients, anti–type II collagen (anti-CII) antibodies. In fact, the development of ACPA is included in RA classification criteria sets, and both ACPA positivity and RF positivity have a high predictive value for later onset of RA (3). A close association between ACPA and certain MHC alleles strongly indicates that specific T cell–driven autoimmune responses precede the development of RA. However, despite the high value of RF and ACPA as prognostic biomarkers for severe erosive arthritis, it has been difficult to demonstrate a direct link between autoantibody development and RA pathogenesis (4).

Anti-CII antibodies are less frequent, and there is no evidence that they occur well ahead of the development of clinical symptoms, as is the case with ACPA and RF (5). These antibodies can be detected around the time of disease onset, and several lines of evidence indicate their pathogenic importance. In CIA, anti-CII is a driving force in disease pathogenesis, illustrated by the fact that passive transfer of anti-CII from CII-immunized mice into naive hosts elicits a similar disease (6, 7). Of note, the same results have been obtained across species, after transfer of the anti-CII–containing Ig fraction from an RA patient into mice (8). Moreover, in CII-immunized mice, the B cell response is essentially directed toward conserved triple-helical CII epitopes (9). Similar fine specificity has been observed in human arthritis patients, strengthening the notion that CIA is a relevant model for pathogenic B cell response in RA (10, 11). Antibodies specific for these major epitopes induce arthritis in mice upon single transfer or in combination, further supporting the idea that anti-CII autoimmunity is pathogenic in RA (12). Anti-CII is also associated with the class II MHC region in humans, though not as strongly as either ACPA or RF. However, little is known about the contribution of non-MHC genes in controlling the formation of autoantibodies against CII in RA. In one study using a candidate gene approach, the R620W polymorphism in Ptpn22 in conjunction with HLA–DRB1 alleles was associated with C1-directed humoral immune responses in RA patients (13).

Animal models such as CIA provide a unique opportunity to study the genetics of antibody response systematically, as large cohorts of mice can be phenotyped under stable environmental conditions. Recently, heterogeneous stock (HS) mice have been successfully used to fine-map known quantitative trait loci (QTLs) and in high-resolution genome-wide mapping of a diverse range of quantitative phenotypes. Moreover, HS-derived mouse cohorts have been used to map the genetic control of arthritis severity in an antigen-dependent model as well as in a mouse model based on passive transfer of arthritogenic autoantibodies (14, 15). Currently the contribution of non-MHC alleles to genetic susceptibility to formation of autoantibodies to CII in human RA, as well as its murine counterpart, is incompletely understood. Resolution of QTLs controlling autoantibody production in CIA and identification of the underlying polymorphisms may result in identification of novel therapeutic targets for human RA. Herein we report on a detailed global genetic analysis of the antibody profile during CIA development, performed using our previously described HS-derived cohort of mice (14).

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Animals and arthritis induction.

The animals used for the CIA experiments, the experimental procedures regarding CIA induction, and the details concerning production of the HS × C57BL/10.Q (BQ) cohort, including a detailed breeding scheme, have been described by us previously (14). Briefly, HS mice known as the Northport stock, derived from 8 common inbred strains (A/J, AKR/J, BALB/cJ, CBA/J, C3H/HeJ, C57BL/6J, DBA/2J, and LP/J), originating in the laboratory of Dr. R. Hitzemann (Oregon Health and Science University, Portland), were backcrossed in an F3 cross with BQ mice, and animals expressing the CIA susceptibility class II MHC allele Aq were selected. Three batches comprising a total of 1,764 mice (approximately equal numbers of male and female animals) were produced, immunized, and monitored for CIA development. All mice were immunized on day 0 with 100 μg rat CII (rCII) emulsified in 50 μl Freund's incomplete adjuvant (Difco) containing 25 μg Mycobacterium butyricum. Thereafter, they were monitored for 7 weeks for development of clinical arthritis. Of the 1,640 mice from which serial serum samples were obtained (see below), arthritis developed in 43% (30% of the female mice and 56% of the male mice). It was first observed a mean ± SD of 30.6 ± 9.0 days after immunization, and the mean ± SD maximum severity score was 25.3 ± 12.9 (scale of 0–60). In a verification experiment, Cia9-congenic mice, harboring a 10.1-Mb NOD mouse–derived fragment on a BQ background, were immunized with 100 μg rCII emulsified in 50 μl Freund's complete adjuvant (Difco) on day 0 and boosted 35 days after initial immunization with 50 μg rCII emulsified in 25 μl Freund's incomplete adjuvant. HS × BQ mice were genotyped using a custom-designed Illumina panel (for details, see refs.14 and16). All experimental animal procedures were approved by the local ethics committees (no. M107-07).

Serum sampling.

Of the 1,764 animals, 1,640 were bled from the retroorbital venous sinus into heparin-containing tubes (10 units/ml). Blood samples were collected on day 14 (prior to CIA onset), as well as at the peak of the disease 7 weeks post–rCII immunization (day 50), upon termination of the experiment. Cia9-congenic mice were bled 35 days after initial immunization. Plasma was separated from erythrocytes after centrifugation for 20 minutes at 4,000 revolutions per minute and was frozen at −20°C until assayed.

Anti-CII enzyme-linked immunosorbent assay (ELISA).

For detection of anti-CII, MaxiSorp plates (Nunc) were coated with CII (10 μg/ml) in carbonate/bicarbonate buffer (pH 9.6) overnight at 4°C and subsequently blocked with 2% bovine serum albumin (BSA; Sigma). Diluted serum samples were added and incubated overnight at 4°C. Ig, IgG1, and IgG3 antibodies were measured using horseradish peroxidase (HRP)–conjugated secondary polyclonal antibodies (goat anti-mouse Ig, IgG1, and IgG3; SouthernBiotech).

Anti-CII and ACPA peptide ELISA.

Collagen peptides were synthesized as previously described (10). The citC1 peptide was obtained from GL Biochem. For measurement of peptide-directed antibody responses, MaxiSorp plates were coated with triple-helical peptides (U1, J1, C1, citC1) at 5 μg/ml in phosphate buffered saline overnight at 4°C and subsequently blocked with 2% BSA. Plates were washed, and diluted serum samples were added and incubated overnight at 4°C. Antibody titers were measured using HRP-conjugated anti–κ light chain–specific secondary antibody (clone 187.1; SouthernBiotech). All peptides were investigated at both time points (day 14 and day 50) except the J1 peptide, which was assessed on day 50 only.

RF ELISA.

RF was detected using anti–κ light chain–specific antibodies as described previously (17). Briefly, plates were coated with rabbit IgG (10 μg/ml; Sigma). After blocking with 2% BSA, plates were washed and diluted serum samples added and incubated overnight at 4°C. Antibody titers were measured using HRP-conjugated anti–κ light chain–specific secondary antibody (clone 187.1).

Plate development.

ELISA plates were developed after addition of ABTS (Roche). A pool of serum samples was used as standard, and all antibody titers were measured as arbitrary concentrations. Anti-CII IgG isotype (IgG1, IgG3) and anti-CII peptide (C1, U1, J1) antibody titers were quantified using defined affinity-purified antibodies from CII-immunized DBA/1J mouse sera. RF and citC1 titers were not quantified due to the lack of a similarly defined standard.

Statistical analysis.

The statistical analysis of the HS population for CIA has been extensively outlined previously (14). Before analysis all antibody titers were normalized using Box-Cox transformations.

Single-locus mapping.

The haplotype structure of the mice was inferred via the haplotype reconstruction method Happy, utilizing founder strain genotypes to calculate probability, FLi(s,t), that a locus interval originates from founder strains s,t (18). Evidence of a QTL was calculated for each interval by fitting the regression model E(yi) = ΣstFLi(s,t)TL(s,t), where TL(s,t) is the phenotype effect due to strains s,t at the locus T. A 5% genome-wide significance threshold was determined by 200 permutations of the data set. Age, sex, and batch were used as covariates in all analyses.

Resample model inclusion probability (RMIP).

The complex family structure of HS-derived mouse populations has been shown to increase the risk of false-positive findings. To address this, a resampling-based multilocus modeling method, RMIP, has been developed (19). RMIP for loci were calculated by subsampling (n = 60; subsample size 80%) with forward selection, adding loci to the model as long as the addition generated a P < 0.05 genome-wide significance threshold (adjusted to the subsample size), as described previously (19). All loci were included in the model selection. Closely spaced inclusion signals were assumed to originate from the same locus and, as in our earlier study (14), inclusion signals within a 5-Mb window were added to yield a range probability (range RMIP). Single-marker associations were calculated with Plink, version 1.07 (http://pngu.mgh.harvard.edu/∼purcell/plink/) (20). Correlations between antibody titers and disease parameters were determined by logistic or linear regression. The antibody titers were log-transformed before analysis, and sex, age, and experiment batch were included as covariates.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

When we investigated the Ig response against CII and against CII peptides, we found that at all time points, antibody titers were significantly higher in animals with CIA than in healthy animals, regardless of which serologic parameter was assessed (Figure 1A and Supplementary Figure 1 [on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131]). In addition, antibody titers increased significantly over time during CIA for all serologic parameters except anti-CII of IgG3 isotype, which exhibited significantly higher titers on day 14 than on day 50 in mice with CIA as well as healthy mice.

thumbnail image

Figure 1. Distribution and genetic regulation of anti–type II collagen (anti-CII) responses in heterogeneous stock mice. A, Anticollagen antibody titers in mice with clinical arthritis and in healthy mice. The titers of all antibodies were significantly higher in mice with collagen-induced arthritis (CIA) than in healthy mice (all P < 0.0001). In addition, titers of all antibodies were significantly higher on day 50 compared to day 14, with the exception that IgG3 antibody titers were higher on day 14 compared to day 50 both in healthy mice (∗ = P < 0.05) and in mice with CIA (# = P < 0.0001). Only male animals are represented, with similar responses observed in females. Values are the mean ± SEM. B, Genome-wide associated loci (range resample model inclusion probability [RMIP] >0.5) for anti-CII isotypes, anti-CII main epitopes, rheumatoid factor (RF), and citC1 responses. Potentially immunologically relevant genes/loci within the analyzed region are shown. The peak range includes the positions with a −logP ≥ 0.7 × max(−logP). Chr = chromosome; f = data obtained with the full model.

Download figure to PowerPoint

QTL mapping of anticollagen IgG antibody response.

As in our previously reported study on QTL mapping of CIA in HS × BQ mice (14), we initially used the Happy algorithm to infer the founder haplotype structure in any given interval based on the genotype of the adjacent markers. We investigated the genetic control of anti-CII Ig and IgG1, which represents a T cell–dependent response against protein antigens constituting the major component of the Ig fraction, and IgG3, representing a T cell–independent response against sugar moieties of polysaccharide antigens. We identified 58 sometimes-overlapping loci for these 3 phenotypes (anti-CII Ig, IgG1, IgG3) that reached a range RMIP of >0.25. Of these, 15 loci displayed a range RMIP of >0.50 and are included in Figure 1B. (For the 43 remaining loci see Supplementary Table 1, http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131.)

The most strongly associated locus controlled total Ig response on day 50 and mapped to the telomeric portion of chromosome 12 at the immunoglobulin heavy chain (Igh) locus in the mouse at ∼119 Mb, with a −logP value of 29.3 and a range RMIP of 1.00 (Figure 1B). This locus also associated with anti-CII response of the IgG1 isotype during the late phase of CIA, displaying a range RMIP of 1.00 and a −logP value of 10.7 (Figure 1B). When we investigated the shared genetic component controlling the T cell–dependent and –independent anti-CII IgG isotype responses, we found loci that were uniquely associated with one isoform (e.g., IgG1 or IgG3) as well as loci that were simultaneously associated with both phenotypes (Figure 1B and Supplementary Table 1). Interestingly, anti-CII IgG1 levels on day 50 were uniquely associated with a locus on chromosome 1 at ∼173 Mb, with the peak marker located near the Fc receptor (FcR) gene cluster (−logP = 12.6, range RMIP 0.68). On the other hand, anti-CII IgG3 was associated with loci distinct from the FcR locus, with the peak markers located at 155 Mb (−logP = 7.3, range RMIP 1.0) and 169 Mb (−logP = 8.1, range RMIP = 0.30) on the same chromosome (Figure 1B and Supplementary Table 1).

QTL mapping of anticollagen peptide antibody and RF response.

We mapped the antibody response to the most relevant CII epitopes (C1, J1, U1), ACPA (citC1), and RF. For these 5 phenotypes, a total of 57 loci reached a range RMIP of >0.25, of which 20 showed a range RMIP of >0.50 (Figure 1B and Supplementary Table 2 [http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131]). A locus on chromosome 4 at 136–140 Mb was strongly associated with RF production and displayed a range RMIP of 1.0 at both time points (Figures 1B and 2A). When we examined the genes in the associated interval, we identified C1q as a likely candidate. This notion was supported by the strain distribution of coding nonsynonymous single-nucleotide polymorphisms (SNPs) within the C1q gene cluster, which followed the same pattern as the haplotypes that correlated with low RF levels (Figure 2B and Supplementary Table 3 [http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131]). For RF and CII-specific epitope (C1, U1, J1) responses, we found an association with the Igh locus on chromosome 12 between 109 Mb and 120 Mb at both time points during CIA development (−logP = 6.2–156.3, range RMIP 0.73–1.00) (Figure 1B). The Igh locus was the only locus significantly associated with J1 response in HS × BQ mice on day 50.

thumbnail image

Figure 2. Fine-mapping of the main quantitative trait locus regulating RF production. A, Association of RF with a locus on chromosome 4. RF was measured on days 14 and 50 after administration of CII, and severity of CIA was assessed on day 49. B, Haplotype reconstruction of the RF response, mapping to the C1q locus. Black bars indicate the strains that differ from the C57BL/6J sequence. See Figure 1 for definitions.

Download figure to PowerPoint

Next, we compared the genetic control of antibody response against the major B cell epitope C1 (CII359–370) and its citrullinated form citC1 (ACPA), which are both also recognized in human RA (21). We found a distinct genetic regulation pattern for the citC1 response compared with C1. This was shown both by the low correlation of antibody responses against C1 and the citC1 epitope (r2 = 0.06 [day 14] and r2 = 0.15 [day 50], by linear regression) and by the genetic mapping results (Figure 1B and Figures 3C and D). In fact at both time points we detected several loci that uniquely controlled either C1 or citC1 response. For example, the C1 response mapped to the Igh locus, whereas only a weak association (range RMIP <0.25) was found for citC1 (data not shown).

thumbnail image

Figure 3. Genetic regulation of anti-CII response in immunized HS × BQ mice. A and B, Genome-wide marker trait association with clinical score of CIA on day 49, and anti-CII IgG1 response on day 14 (A) and day 50 (B). C and D, Anti-C1 and anti-citC1 epitope responses on day 14 (C) and day 50 (D). See Figure 1 for definitions.

Download figure to PowerPoint

Correlation between CIA and anticollagen antibody levels.

All antibody isotypes measured on day 14 were associated with an increased risk of CIA development (Figure 4A). Of the epitope-specific antibodies, antibodies to the C1 epitope were the strongest predictor of disease. ACPAs directed toward the citC1 epitope also showed a significant association with incidence of the disease, but the increase in risk was comparatively small (Figures 4A and B). This was also the case for RF (Figure 4A and Supplementary Table 4 [on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131]). Time of disease onset was significantly associated with all antibody isotypes, as were antibodies against the C1 and U1 peptides. Because RF, ACPA, and anti-CII are all predictors of disease severity in human RA, we analyzed the association between day 14 antibody titers and CIA scores on day 49 (adjusted for time of onset). In contrast to what is found in humans, we found that only total Ig, IgG1, and IgG3 response against CII were associated with increased disease severity on day 49 (Supplementary Table 4). No correlation between RF concentrations on day 14 and disease severity was observed in our experiments. At peak of the disease (day 50), all antibody isotypes and epitope specificities, as well as RF, correlated positively with disease score on day 49 (Supplementary Table 4).

thumbnail image

Figure 4. Correlation of anti-CII antibody titers with incidence and severity of CIA. A, Odds ratios (with 95% confidence intervals) for the correlation between antibody titers on day 14 and observed incidence of CIA on day 50. (Detailed information is presented in Supplementary Table 4, on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131.) B, Identified loci that were strongly correlated with CIA severity (sum score), anti-CII IgG1 titers, and anti-C1 response. ∗ = locus identified as correlating with anti-CII IgG1 response on day 14; ∗∗ = locus identified as correlating with anti-C1 response on day 14. All other loci were identified as correlating with CIA severity. See Figure 1 for definitions.

Download figure to PowerPoint

Comparison between QTL mapping of antibody response and CIA.

Finally, we assessed the shared genetic control of anti-CII antibody response and CIA development, by comparing the loci that were most strongly associated with CIA in our previous study (14). Investigating levels of C1 and IgG1, which were the best predictors of disease, we found that 2 CIA loci overlapped directly with loci controlling these phenotypes (Figures 3A and B). One of these loci mapped to the telomeric region of murine chromosome 1 at a locus previously described as Cia9, with the peak marker for CIA severity, onset, and incidence being located close to the FcR gene cluster (22) (Figures 1B and 5A). In order to confirm these findings we performed a CIA experiment in a 10-Mb NOD mouse–derived congenic fragment, which included the previously described Cia9 region (23). (A genetic map of this region is available on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131.) We found that upon immunization with rCII, CIA severity, as well as levels of anti-CII of the IgG1 and IgG3 isotypes, were strongly increased in these mice compared to BQ controls (Figures 5A, C, and D). This provides further evidence of the usefulness of heterogeneous stock mice for fine-mapping of QTLs.

thumbnail image

Figure 5. Fine-mapping of the main quantitative trait loci regulating anti-CII IgG1 production. A, CIA and IgG1 associations with a locus on chromosome 1, assessed using the Happy algorithm. Genome-wide significance levels for the displayed phenotypes were between 4.8 and 5.0. B, Haplotype distribution of the anti-CII IgG1 response mapping to the Fcgr locus. (Detailed information is presented in Supplementary Table 5, on the Arthritis & Rheumatism web site at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)1529-0131.) Black bars indicate the strains that differ from the reference C57BL/6J strain. ∗ = single-nucleotide polymorphism also associated with CIA sum score. C and D, Regulation of CIA (C) and anti-CII antibody (D) phenotype by the congenic fragment Cia9. Values in C are the mean ± SEM. Values in D are shown as box plots, where boxes represent the 25th to 75th percentiles, lines inside the boxes represent the median, lines outside the boxes represent the 10th and the 90th percentiles, and circles represent outliers. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001. WT = wild-type (see Figure 1 for other definitions).

Download figure to PowerPoint

The other disease locus was located on chromosome 12 at ∼77 Mb, with the peak marker mapping close to the Esr1 gene (estrogen receptor α) (Figure 4B). The strongest CIA locus, which was found on chromosome 2 mapping to the Hc gene (complement component 5), did not show an association with either IgG1 or C1 response (Figure 4B).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The present study was a comprehensive genetic analysis of clinically relevant autoantibody responses during CIA development, using a mixed inbred-outbred HS-derived mouse cohort. Antibodies against the Fc portion of Ig (RF) were described 60 years ago and have been used as a serologic marker to classify human RA according to the American College of Rheumatology (ACR) criteria (24). Their role in RA susceptibility has long been debated, and despite the fact that RF can be detected prior to disease onset, evidence of a clear role in RA pathogenesis is lacking. We found that circulating RF prior to disease onset slightly increases the risk of developing CIA.

The locus with the highest association for RF was detected on chromosome 4. We examined the genes in the associated interval and identified C1q as a likely candidate gene, which is supported by single-marker association of closely located SNPs. The importance of the classical pathway of complement activation for antibody production is illustrated by the fact that mice deficient either in one of its components (C1q, C2, C3) or in one of its receptors (CR1, CR2) display severely impaired primary and secondary antibody responses upon immunization (25–27). Experiments with CR1/2-knockout animals demonstrated a critical requirement for IgM–antigen complexes for efficient primary antibody responses (28). Furthermore, studies of C1q-deficient mice indicated that this molecule plays a role in the negative selection of B cells, as these animals exhibited higher levels of B1 cells and increased secretion of low-affinity IgM autoantibodies (29). On the other hand, conventional B cells in these mice, which generate autoantibodies of the IgG isotype and require T cell–dependent maturation, are anergized in the periphery (30).

In addition to the RF response we investigated the ACPA response in our mice. ACPAs directed against cyclic citrullinated peptides (CCPs) are a highly RA-specific biomarker and have been successfully introduced into the clinic within the last decade (31). They are also part of the revised ACR/European League Against Rheumatism criteria for RA classification (32, 33). However, in contrast to human RA, specific anti-CCP antibodies are not present in murine CIA (34, 35). When we analyzed ACPA response against the citC1 epitope, which is also present in RA patients, we found that in CIA, similar to findings in humans, the presence of these antibodies predicts disease. Moreover, we showed that at the peak of the disease, citC1 antibody levels correlate with disease severity. This is consistent with our previous observations in studies using the ACC4 monoclonal antibody, which is specific for the citC1 epitope. This antibody is not arthritogenic itself, but aggravates arthritis severity after coadministration with the M2139 antibody in subarthritogenic doses in the collagen antibody–induced arthritis model, which mimics the effector phase of RA (36). Nevertheless, C1 antibodies have a much higher predictive value for CIA development and correlate better with disease severity during the late stage of CIA than do citC1-specific ACPAs, probably reflecting the fact that CII autoimmunity is the driving factor in CIA following immunization, whereas in human RA, CII autoimmunity occurs only in a subset of patients reacting to modified cartilage CII (2).

Mapping of autoantibody responses against CII in HS × BQ mice strongly highlighted the critical importance of the Igh locus. Not surprisingly, we found strong associations of all anti-CII isotypes with Igh during early and late CIA, indicating that in certain mouse strains B cells preferably use a specific CH gene or are differentially responsive to T cell–assisted IgG class-switch recombination. More interestingly, we found that CII epitope responses against U1, C1, and J1 also map to the Igh locus in HS × BQ mice. It has been shown that these responses vary greatly during development of chronic CIA, with both genetic and environmental factors contributing (37). Studies on genetic control of CII epitope–specific responses in BQ × (BALB/c × BQ)N2 mice previously demonstrated association of J1 and U1 with Igh during chronic CIA (38). Our findings underscore the importance of Igh, indicating that certain mouse strains preferably use certain VH genes to mount a response against specific pathogenic CII epitopes. It is thus easy to imagine that SNPs in a particular VH gene segment in this locus select for certain CII epitope specificities. An example is the response against J1, which is predominantly controlled by the Igh locus and its production strongly limited to a pair of VH genes in the Igh locus (Raposo B, et al: unpublished observations).

When we investigated the direct overlap of loci controlling CIA and antibody phenotypes, we found that the most prominent locus linking both phenotypes is located on chromosome 1 mapping to the FcR locus. This locus controls CIA onset, incidence, and severity as well as IgG1 and anti-C1 response, but, surprisingly, none of the other epitopes or IgG3. Multiple studies with genetically targeted mice have implicated FcR genes in susceptibility to arthritis in mice, using various models of disease (39, 40). Further, a SNP in the promoter of the FcR-like 3 gene has been associated with human RA in Asians (41, 42). In addition, we reproduced our findings regarding CIA and IgG1 in the HS mice in a NOD mouse–derived congenic fragment that partially spans over the original Cia9 interval. By single-marker association, we found that SNPs located between Fcgr4 and Fcgr3 yielded the highest association with IgG1.

The alteration of the anti-CII IgG1 response by the Fc locus provides evidence of altered Th1/Th2 differentiation, since it is well known that IgG1 production is strongly associated with skewing toward a Th2 response in mice (43). It has been demonstrated that Fcgr3 signaling is absolutely essential for an optimal Th2 T cell response (44). Previous studies have suggested that IgG can, in an isotype-dependent manner, enhance antibody responses to large protein antigens such as collagen by shuttling immune complexes to antigen-presenting cells via activating Fc receptors, subsequently leading to more efficient CD4 T cell responses (45). In addition, Fcgr2b, to date the only described inhibitory Fc receptor, has been demonstrated to indirectly participate in IgG isotype-specific–mediated enhancement of antibody responses, being a critical checkpoint allowing signaling via the activating Fc receptors (46). Thereby, the T cell–independent IgG3 response does not seem to be controlled by the FcR locus, probably due to the fact that this isotype binds only weakly, if at all, to Fcgr1 (47, 48).

On the other hand, the Traf1-C5 (complement component) locus, which is highly associated with arthritis development across species, was not associated with antibody development (22, 49). This finding supports experimental data from rodent models of arthritis, in which C5 deficiency was associated with protection not only against CIA but also against different forms of experimental arthritis induced using models based on passive transfer of arthritogenic antibodies, highlighting its importance for mediating effector mechanisms in the pathogenesis of murine arthritis (50, 51).

In conclusion, our data provide evidence of complex genetic control of anti-CII of different isotypes, ACPA, and RF, indicating that non-MHC genes determine antibody specificity, with the Igh locus being particularly important. This study demonstrates the potential of HS stock mice for high-resolution mapping of autoantibody-controlling QTLs, thereby facilitating more rapid identification of candidate genes as well as candidate polymorphisms that can be further exploited to define novel targets for therapeutic intervention. We also found that some CIA- and anti-CII antibody–controlling loci, such as the FcR or the Esr1 loci, overlap and thereby cosegregate within the mouse genome, indicating that the underlying genes were coselected during the evolution of murine inbred strains.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

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. Holmdahl 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. Förster, Raposo, Ekman, Klaczkowska, Popovic, Nandakumar, Lindvall, Hultqvist, Teneva, Johannesson, Ahlqvist, Holmdahl.

Acquisition of data. Förster, Raposo, Klaczkowska, Popovic, Lindvall, Hultqvist, Teneva, Ahlqvist, Holmdahl.

Analysis and interpretation of data. Förster, Raposo, Ekman, Ahlqvist, Holmdahl.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The authors would like to thank Sandy Liedholm for excellent animal care.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information
  • 1
    Brunsberg U, Gustafsson K, Jansson L, Michaelsson E, Ahrlund-Richter L, Pettersson S, et al. Expression of a transgenic class II Ab gene confers susceptibility to collagen-induced arthritis. Eur J Immunol 1994; 24: 1698702.
  • 2
    Backlund J, Carlsen S, Hoger T, Holm B, Fugger L, Kihlberg J, et al. Predominant selection of T cells specific for the glycosylated collagen type II epitope (263–270) in humanized transgenic mice and in rheumatoid arthritis. Proc Natl Acad Sci U S A 2002; 99: 99605.
  • 3
    Rantapaa-Dahlqvist S, de Jong BA, Berglin E, Hallmans G, Wadell G, Stenlund H, et al. Antibodies against cyclic citrullinated peptide and IgA rheumatoid factor predict the development of rheumatoid arthritis. Arthritis Rheum 2003; 48: 27419.
  • 4
    Rowley MJ, Nandakumar KS, Holmdahl R. The role of collagen antibodies in mediating arthritis. Mod Rheumatol 2008; 18: 42941.
  • 5
    Morgan K, Clague RB, Collins I, Ayad S, Phinn SD, Holt PJ. Incidence of antibodies to native and denatured cartilage collagens (types II, IX, and XI) and to type I collagen in rheumatoid arthritis. Ann Rheum Dis 1987; 46: 9027.
  • 6
    Stuart JM, Dixon FJ. Serum transfer of collagen-induced arthritis in mice. J Exp Med 1983; 158: 37892.
  • 7
    Wooley PH, Luthra HS, Krco CJ, Stuart JM, David CS. Type II collagen-induced arthritis in mice. II. Passive transfer and suppression by intravenous injection of anti–type II collagen antibody or free native type II collagen. Arthritis Rheum 1984; 27: 10107.
  • 8
    Wooley PH, Luthra HS, Singh SK, Huse AR, Stuart JM, David CS. Passive transfer of arthritis to mice by injection of human anti-type II collagen antibody. Mayo Clin Proc 1984; 59: 73743.
  • 9
    Nandakumar KS, Svensson L, Holmdahl R. Collagen type II-specific monoclonal antibody-induced arthritis in mice: description of the disease and the influence of age, sex, and genes. Am J Pathol 2003; 163: 182737.
  • 10
    Burkhardt H, Koller T, Engstrom A, Nandakumar KS, Turnay J, Kraetsch HG, et al. Epitope-specific recognition of type II collagen by rheumatoid arthritis antibodies is shared with recognition by antibodies that are arthritogenic in collagen-induced arthritis in the mouse. Arthritis Rheum 2002; 46: 233948.
  • 11
    Kraetsch HG, Unger C, Wernhoff P, Schneider C, Kalden JR, Holmdahl R, et al. Cartilage-specific autoimmunity in rheumatoid arthritis: characterization of a triple helical B cell epitope in the integrin-binding-domain of collagen type II. Eur J Immunol 2001; 31: 166673.
  • 12
    Nandakumar KS, Holmdahl R. Efficient promotion of collagen antibody induced arthritis (CAIA) using four monoclonal antibodies specific for the major epitopes recognized in both collagen induced arthritis and rheumatoid arthritis. J Immunol Methods 2005; 304: 12636.
  • 13
    Burkhardt H, Huffmeier U, Spriewald B, Bohm B, Rau R, Kallert S, et al. Association between protein tyrosine phosphatase 22 variant R620W in conjunction with the HLA–DRB1 shared epitope and humoral autoimmunity to an immunodominant epitope of cartilage-specific type II collagen in early rheumatoid arthritis. Arthritis Rheum 2006; 54: 829.
  • 14
    Ahlqvist E, Ekman D, Lindvall T, Popovic M, Forster M, Hultqvist M, et al. High-resolution mapping of a complex disease, a model for rheumatoid arthritis, using heterogeneous stock mice. Hum Mol Genet 2011; 20: 303141.
  • 15
    Johnsen AK, Valdar W, Golden L, Ortiz-Lopez A, Hitzemann R, Flint J, et al. Genome-wide and species-wide dissection of the genetics of arthritis severity in heterogeneous stock mice. Arthritis Rheum 2011; 63: 263040.
  • 16
    Keane TM, Goodstadt L, Danecek P, White MA, Wong K, Yalcin B, et al. Mouse genomic variation and its effect on phenotypes and gene regulation. Nature 2011; 477: 28994.
  • 17
    Wolfowicz CB, Sakorafas P, Rothstein TL, Marshak-Rothstein A. Oligoclonality of rheumatoid factors arising spontaneously in lpr/lpr mice. Clin Immunol Immunopathol 1988; 46: 38295.
  • 18
    Mott R, Talbot CJ, Turri MG, Collins AC, Flint J. A method for fine mapping quantitative trait loci in outbred animal stocks. Proc Natl Acad Sci U S A 2000; 97: 1264954.
  • 19
    Valdar W, Holmes CC, Mott R, Flint J. Mapping in structured populations by resample model averaging. Genetics 2009; 182: 126377.
  • 20
    Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet 2007; 81: 55975.
  • 21
    Burkhardt H, Sehnert B, Bockermann R, Engstrom A, Kalden JR, Holmdahl R. Humoral immune response to citrullinated collagen type II determinants in early rheumatoid arthritis. Eur J Immunol 2005; 35: 164352.
  • 22
    Lindqvist AK, Johannesson M, Johansson AC, Nandakumar KS, Blom AM, Holmdahl R. Backcross and partial advanced intercross analysis of nonobese diabetic gene-mediated effects on collagen-induced arthritis reveals an interactive effect by two major loci. J Immunol 2006; 177: 39529.
  • 23
    Johansson AC, Sundler M, Kjellen P, Johannesson M, Cook A, Lindqvist AK, et al. Genetic control of collagen-induced arthritis in a cross with NOD and C57BL/10 mice is dependent on gene regions encoding complement factor 5 and FcγRIIb and is not associated with loci controlling diabetes. Eur J Immunol 2001; 31: 184756.
  • 24
    Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988; 31: 31524.
  • 25
    Molina H, Holers VM, Li B, Fung Y, Mariathasan S, Goellner J, et al. Markedly impaired humoral immune response in mice deficient in complement receptors 1 and 2. Proc Natl Acad Sci U S A 1996; 93: 335761.
  • 26
    Ahearn JM, Fischer MB, Croix D, Goerg S, Ma M, Xia J, et al. Disruption of the Cr2 locus results in a reduction in B-1a cells and in an impaired B cell response to T-dependent antigen. Immunity 1996; 4: 25162.
  • 27
    Carroll MC. The role of complement and complement receptors in induction and regulation of immunity. Annu Rev Immunol 1998; 16: 54568.
  • 28
    Applequist SE, Dahlstrom J, Jiang N, Molina H, Heyman B. Antibody production in mice deficient for complement receptors 1 and 2 can be induced by IgG/Ag and IgE/Ag, but not IgM/Ag complexes. J Immunol 2000; 165: 2398403.
  • 29
    Ferry H, Potter PK, Crockford TL, Nijnik A, Ehrenstein MR, Walport MJ, et al. Increased positive selection of B1 cells and reduced B cell tolerance to intracellular antigens in C1q-deficient mice. J Immunol 2007; 178: 291622.
  • 30
    Cutler AJ, Cornall RJ, Ferry H, Manderson AP, Botto M, Walport MJ. Intact B cell tolerance in the absence of the first component of the classical complement pathway. Eur J Immunol 2001; 31: 208793.
  • 31
    Schellekens GA, Visser H, de Jong BA, van den Hoogen FH, Hazes JM, Breedveld FC, et al. The diagnostic properties of rheumatoid arthritis antibodies recognizing a cyclic citrullinated peptide. Arthritis Rheum 2000; 43: 15563.
  • 32
    Aletaha D, Neogi T, Silman AJ, Funovits J, Felson DT, Bingham CO III, et al. 2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Ann Rheum Dis 2010; 69: 15808.
  • 33
    Aletaha D, Neogi T, Silman AJ, Funovits J, Felson DT, Bingham CO III, et al. 2010 rheumatoid arthritis classification criteria: an American College of Rheumatology/European League Against Rheumatism collaborative initiative. Arthritis Rheum 2010; 62: 256981.
  • 34
    Lopez-Hoyos M, Marquina R, Tamayo E, Gonzalez-Rojas J, Izui S, Merino R, et al. Defects in the regulation of B cell apoptosis are required for the production of citrullinated peptide autoantibodies in mice. Arthritis Rheum 2003; 48: 235361.
  • 35
    Vossenaar ER, Nijenhuis S, Helsen MM, van der Heijden A, Senshu T, van den Berg WB, et al. Citrullination of synovial proteins in murine models of rheumatoid arthritis. Arthritis Rheum 2003; 48: 2489500.
  • 36
    Uysal H, Bockermann R, Nandakumar KS, Sehnert B, Bajtner E, Engstrom A, et al. Structure and pathogenicity of antibodies specific for citrullinated collagen type II in experimental arthritis. J Exp Med 2009; 206: 44962.
  • 37
    Bajtner E, Nandakumar KS, Engstrom A, Holmdahl R. Chronic development of collagen-induced arthritis is associated with arthritogenic antibodies against specific epitopes on type II collagen. Arthritis Res Ther 2005; 7: R114857.
  • 38
    Nandakumar KS, Lindqvist AK, Holmdahl R. A dominant suppressive MHC class II haplotype interacting with autosomal genes controls autoantibody production and chronicity of arthritis. Ann Rheum Dis 2011; 70: 166470.
  • 39
    Diaz de Stahl T, Andren M, Martinsson P, Verbeek JS, Kleinau S. Expression of FcγRIII is required for development of collagen-induced arthritis. Eur J Immunol 2002; 32: 291522.
  • 40
    Van Lent P, Nabbe KC, Boross P, Blom AB, Roth J, Holthuysen A, et al. The inhibitory receptor FcγRII reduces joint inflammation and destruction in experimental immune complex-mediated arthritides not only by inhibition of FcγRI/III but also by efficient clearance and endocytosis of immune complexes. Am J Pathol 2003; 163: 183948.
  • 41
    Kochi Y, Yamada R, Suzuki A, Harley JB, Shirasawa S, Sawada T, et al. A functional variant in FCRL3, encoding Fc receptor-like 3, is associated with rheumatoid arthritis and several autoimmunities. Nat Genet 2005; 37: 47885.
  • 42
    Lee YH, Woo JH, Choi SJ, Ji JD, Song GG. Fc receptor-like 3 −169 C/T polymorphism and RA susceptibility: a meta-analysis. Rheumatol Int 2010; 30: 94753.
  • 43
    Stevens TL, Bossie A, Sanders VM, Fernandez-Botran R, Coffman RL, Mosmann TR, et al. Regulation of antibody isotype secretion by subsets of antigen-specific helper T cells. Nature 1988; 334: 2558.
  • 44
    Bandukwala HS, Clay BS, Tong J, Mody PD, Cannon JL, Shilling RA, et al. Signaling through FcγRIII is required for optimal T helper type (Th)2 responses and Th2-mediated airway inflammation. J Exp Med 2007; 204: 187589.
  • 45
    Hjelm F, Carlsson F, Getahun A, Heyman B. Antibody-mediated regulation of the immune response. Scand J Immunol 2006; 64: 17784.
  • 46
    Nimmerjahn F, Ravetch JV. Divergent immunoglobulin G subclass activity through selective Fc receptor binding. Science 2005; 310: 15102.
  • 47
    Gavin AL, Barnes N, Dijstelbloem HM, Hogarth PM. Identification of the mouse IgG3 receptor: implications for antibody effector function at the interface between innate and adaptive immunity. J Immunol 1998; 160: 203.
  • 48
    Hjelm F, Carlsson F, Verbeek S, Heyman B. IgG3-mediated enhancement of the antibody response is normal in FcγRI-deficient mice. Scand J Immunol 2005; 62: 45361.
  • 49
    Plenge RM, Seielstad M, Padyukov L, Lee AT, Remmers EF, Ding B, et al. TRAF1-C5 as a risk locus for rheumatoid arthritis—a genomewide study. N Engl J Med 2007; 357: 1199209.
  • 50
    Grant EP, Picarella D, Burwell T, Delaney T, Croci A, Avitahl N, et al. Essential role for the C5a receptor in regulating the effector phase of synovial infiltration and joint destruction in experimental arthritis. J Exp Med 2002; 196: 146171.
  • 51
    Hietala MA, Nandakumar KS, Persson L, Fahlen S, Holmdahl R, Pekna M. Complement activation by both classical and alternative pathways is critical for the effector phase of arthritis. Eur J Immunol 2004; 34: 120816.

Supporting Information

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ART_34658_sm_SupplTable1.tif9650KSupplementary Table 1
ART_34658_sm_SupplTable2.tif9641KSupplementary Table 2
ART_34658_sm_SupplTable3.tif10426KSupplementary Table 3
ART_34658_sm_SupplTable4.tif10426KSupplementary Table 4
ART_34658_sm_SupplTable5.tif10431KSupplementary Table 5

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.