New humanized HLA–DR4–transgenic mice that mimic the sex bias of rheumatoid arthritis




To generate a mouse model that can mimic human rheumatoid arthritis (RA). A major difference between RA in humans and collagen-induced arthritis (CIA) in mice is the lack of sex bias and autoantibodies in the animal model. We used DRB1*0401-transgenic mice to understand the role of DR4 in susceptibility and sex bias in RA.


A transgenic mouse was generated that lacked all endogenous mouse class II genes (AEo) and expressed the RA susceptibility allele HLA–DRB1*0401. These transgenic mice were tested for incidence, severity, and sex distribution of CIA.


DRB1*0401.AEo mice developed CIA predominantly in females and produced rheumatoid factors, similar to the features of human RA. Another feature similar to human RA is the expression of class II molecules on antigen-presenting cells as well as T cells. Activated and sorted CD4+ T cells can present DR4-restricted type II collagen (CII)–derived peptide in vitro, but cannot process the antigen. This suggests a role for these cells in epitope presentation locally in joints, which affects disease severity. After challenge with CII, female mice had higher cellularity and increased T cell proliferation and produced higher levels of proinflammatory cytokines than did the male mice.


DR4.AEo mice expressed HLA similar to humans and displayed increased arthritis susceptibility in females, thus mimicking RA in humans. This model may be valuable for studying sex differences observed in humans and for understanding why autoimmunity is increased in women. These mice may also be useful for developing future therapeutic strategies.

Rheumatoid arthritis (RA) is an autoimmune disease that results from a complex interplay of both genetic and environmental factors. The major genetic contribution comes from genes located in the HLA class II region. In most populations, DRB1*0401, 0404, and 0101 are associated with RA. Gregersen and coworkers (1) suggested a “shared epitope” hypothesis to explain the association of various DR alleles with arthritis. According to the hypothesis, DRB1 alleles sharing a motif at positions 67, 70, 71, and 74 are implicated in the pathogenesis of RA by their ability to present similar antigens, leading to the development of arthritis. DRB1*0402 was found to be associated with non-susceptibility to RA; however, both DRB1*0401 and DRB1*0402 can present the DR4-restricted immunodominant peptide of type II collagen (CII) (2, 3).

Another hypothesis suggested that “shared epitopes” could be involved in T cell receptor (TCR) selection because peptides from DRB1 alleles are presented by HLA–DQ in the thymus (4). This is supported by the finding that naturally processed peptides presented by class II molecules are derived from endogenous class II molecules (5). However, it is difficult to experimentally assess the role of a single allele and its contribution to the shared epitope in order to confirm this hypothesis in humans.

RA is a chronic inflammatory disease characterized by synovial inflammation and erosion of bone and cartilage, which lead to the destruction of joints. Patients with RA have CII-reactive T cells and antibodies, which suggests that CII is a candidate autoantigen in disease pathogenesis (6). Collagen-induced arthritis (CIA) has been used for the last 2 decades as an experimental model with which to study RA. Our previous studies have shown that Aβo-transgenic mice expressing HLA–DQA1*0301/DQB1*0302 are highly susceptible to CIA (7) and can present multiple human CII peptides (8). Previous studies using DR4/H-2 mice showed that DR4 could modulate arthritis, thus suggesting an important role of this haplotype in RA (3, 9, 10).

In the present study, the DRB1*0401.Aβo-transgenic mice, which lack functional Aβ and Eα genes but which express Aα and Eβ genes, did not develop CIA. A potential problem with these mice is the expression of endogenous H-2E chain. To overcome this problem, we mated DRB1*0401.Aβo mice with MHCIIΔ/Δ mice, which lack all of the endogenous class II molecules (11), to generate DR4.AEo mice that lack all 4 classic murine chains (Aα, Aβ, Eα, and Eβ). The only class II molecules expressed in these transgenic mice are DR4 molecules (EαDRB1*0401). Unlike endogenous class II molecules, HLA–DR molecules in transgenic mice are expressed on a subset of CD3+ T cells that are also known to be present in humans. In vivo studies after induction of CIA showed an increased susceptibility to disease in female DR4 mice. A sex bias in terms of disease was not observed in the DQ8.AEo mice. The observations in DR4.AEo-transgenic mice are similar to those in humans with RA and may provide a model with which to study sex bias for arthritis in humans.


Transgenic mice.

The generation of HLA–DRB1*0401 (DR4)–transgenic mice has been described previously (12). DRB1*0401.Aβo mice were mated with MHCIIΔ/Δ (AEo) mice (11) to generate DR4.AEo mice. In a similar manner, we generated DQ8.AEo mice. DR4.Aβo and DQ8.Aβo mice (3, 7) were used as controls. Mice of both sexes between the ages of 8 and 12 weeks were used in this study and were bred and maintained in the pathogen-free Immunogenetics Mouse Colony at the Mayo Clinic in accordance with the Institutional Animal Care and Use Committee (IACUC). All experiments included littermate controls and were performed with the approval of the IACUC.

Flow cytometry.

The expression of DRβ, H-2E, and TCR Vβ-chains on peripheral blood lymphocytes from transgenic mice were analyzed by flow cytometry using the following monoclonal antibodies: L227 (anti-DR), IVD12 (anti-DQ), 14-4-4s (anti-Eα), and GK1.5 (anti–mouse CD4) (American Type Culture Collection, Manassas, VA), and B20.6 (anti-Vβ2), KT4-10 (anti-Vβ4), MR9-4 (anti-Vβ5.1.2), MR9-8 (anti-Vβ5.1), 44-22-1 (anti-Vβ6), TR310 (anti-Vβ7), KJ-16 (anti-Vβ8.1.2), F23.2 (anti-Vβ8.2), MR10-2 (anti-Vβ9), KT11 (anti-Vβ11), 14.2 (anti-Vβ14), and KJ23a (anti-Vβ17) (BD PharMingen, San Diego, CA). Conjugated antibodies against CD3, CD4, CD8, Mac-1, and B220 (BD PharMingen) were also used. For analysis of DR expression on activated cells, primed spleen cells were challenged in vitro with CII or staphylococcal enterotoxin B (SEB). All cell surface markers were analyzed in cells pooled from 2 mice per strain, and experiments were performed 2–3 times.

Induction and evaluation of CIA.

Pure native chick CII was obtained by multiple-step purification as described previously (13). CIA was induced in transgenic animals by immunization with CII (100 μg of CII emulsified in Freud's complete adjuvant [CFA]) according to the standard protocol described elsewhere (3). Mice were studied in 2 groups to ensure reproducibility of results. The onset and progression of CIA was monitored from week 3 through week 12 postimmunization. The severity of arthritis in each paw was graded on a scale of 0–3, as described previously (14). The mean arthritis score was determined in arthritic animals only.

Histopathologic analysis.

Mice were killed 10–12 weeks after immunization, and the paws were removed, decalcified, and fixed. Sections were stained with hematoxylin and eosin and examined for infiltrations and erosions. Spleen, thymus, liver, salivary gland, and kidney were removed from naive DR4 mice, then frozen and sectioned. Sections were stained with anti-DR antibody (L227) that had been conjugated with fluorescein isothiocyanate or phycoerythrin, and some sections were stained with CD4 that had been conjugated with phycoerythrin. Slides were also stained with 4′,6-diamidino-2-phenylindole.

T cell proliferation assay.

Mice were immunized intradermally at the base of the tail and 1 hind footpad with 200 μg of CII emulsified 1:1 in CFA (Difco, Detroit, MI). Ten days postimmunization, draining lymph nodes/spleen were removed and cultured in vitro. Lymph node cells (1 × 106) were cultured in HEPES buffered RPMI 1640 containing 5% heat-inactivated horse serum and antibiotics (streptomycin and penicillin) in 96-well flat-bottomed tissue culture plates. Cells were challenged by the addition of RPMI 1640 (100 μl; negative control), concanavalin A (20 μg/ml, positive control), or native collagen (50 μg/ml).

For inhibition experiments, culture supernatant containing 25 μg/ml of monoclonal antibody GK1.5 (anti-CD4), anti-DR (L227), or Lyt2 (anti-CD8) was added to the cells and challenged in vitro with 50 μg/ml of CII. The cells were incubated for 48 hours at 37°C. During the last 18 hours, the cells were pulsed with 3H-thymidine (1 μCi/well). At the end of the assay, the cells were harvested using a plate harvester, and the incorporated radioactivity was determined using an automated counter (Microbeta; PerkinElmer Wallac, Gaithersburg, MD). Results are expressed as a stimulation index (SI).

Transgenic and transgene-negative littermates were tested for T cell responses to human CII–derived peptide 254-273, which was synthesized and purified at the Mayo Clinic Peptide Facility. The mice were primed with 200 μg of peptide and then challenged in vitro with 100 μg/ml of the peptide.

For presentation of CII and its derived human CII peptide 254–273, CD4+ cells were sorted with a fluorescence-activated cell sorter (FACS) from lymph node cells isolated from primed mice. CD4+ cells were cultured in vitro in the presence or absence of antigen-presenting cells (APCs). Up to 5 × 105 CD4+ sorted cells were used for culturing alone with CII (50 μg/ml). In the other cultures, 5 × 105 irradiated spleen cells were used as APCs. DR restriction was studied by adding anti-DR antibody (25 μg/ml) to the culture. Cells were cultured in RPMI 1640 medium as described above, and cell proliferation was determined by 3H-thymidine incorporation. An SI ≥2 was considered to be a positive response. The experiment was performed twice with cells pooled from 2 mice per experiment.

Measurement of cytokines.

The cytokines interferon-γ (IFNγ), interleukin-10 (IL-10), IL-18, transforming growth factor β (TGFβ), and tumor necrosis factor α (TNFα) were measured by capture enzyme-linked immunosorbent assay (ELISA), using kits obtained from BD PharMingen and following the manufacturer's instructions.

Measurement of anti-CII antibodies.

Levels of anti-chick and anti-mouse CII IgG antibodies were measured in sera obtained 35 days after CII immunization. A standard ELISA was used for the analyses, and data are reported as the optical density.

Measurement of rheumatoid factors (RFs).

IgG and IgM RFs were measured by ELISA in mouse sera obtained 35 days after priming, as previously described (15). Briefly, ELISA plates were coated overnight at 4°C with rabbit IgG. After washing, sera (1:40 dilution) were added, and the plates were incubated for 45 minutes at room temperature. Plates were then washed 5 times with phosphate buffered saline containing 0.05% Tween 20. Subsequently, wells were incubated for 1 hour with horseradish peroxidase–conjugated rabbit anti-mouse IgG (Fc specific) or rabbit anti-mouse IgM (μ-chain specific) (both from Pierce, Rockford, IL). After washing, 3,3′,5,5′-tetramethylbenzidine substrate (Sigma, St. Louis, MO) was added, and the absorbance spectrum was determined with an automated spectrophotometer (Bio-Rad, Hercules, CA). Sera from MRL/lpr and B10 mice were used as positive and negative controls, respectively.

Measurement of antibodies to cyclic citrullinated peptide (anti-CCP).

Serum levels of anti-CCP antibodies were measured by ELISA using a kit obtained from Axis-Shield (Kimbolton, UK).

Statistical analysis.

The difference in the incidence of arthritis between groups was analyzed using Fisher's exact test. Antibody levels, onset of arthritis, mean arthritis scores in arthritic mice, and numbers of CD4+ cells were compared using the nonparametric Mann-Whitney 2-tailed U test. The mean flow intensity of DR expression was compared by Kolmogorov-Smirnov test. All other significance assessments were calculated using Student's t-test or t-test with unequal variance.


Expression of class II molecules on T cells in AEo-transgenic mice.

The DR4.AEo-transgenic mice developed normally and had no gross phenotypic abnormalities. Immunostaining of various organs showed expression of DR4 in the thymus and spleen, whereas no expression in the kidney or liver and rare patchy expression in salivary gland was observed (Figures 1A–E).

Figure 1.

Characterization of DR4.AEo mice. Expression of DR4 in sections of A, spleen, B, thymus, C, kidney, D, liver, and E, salivary gland from DRB1*0401.AEo mice was studied by immunostaining with fluorescein isothiocyanate (FITC)–conjugated anti-DR antibody and with 4′,6-diamidino-2-phenylindole (DAPI). FI, Thymus sections showing staining with phycoerythrin-conjugated CD4 (F), FITC-conjugated DR (G), a superimposition of CD4- and DR-stained sections, showing some CD4+ cells that are also DR+ (H), and lack of staining with isotype control antibody (I). These sections were also stained with DAPI. (Original magnification × 10 in AE; × 40 in FI.) J, Dot plots of gated splenic CD3+ cells showing similar percentages of CD4+ and CD8+ cells in DR4.Aβo-transgenic and DR4.AEo-transgenic mice. Results are from 1 of 3 experiments performed with pooled cells from 2 mice per strain per experiment. Values are percentages.

Immunostaining of thymus sections with antibodies to CD4 and DR showed some DR+,CD4+ cells (Figures 1F–I). Subsets gated on splenic CD3+ cells showed a similar number of CD3+,CD4+ and CD3+,CD8+ cells in both transgenic mouse strains (Figure 1J). Expression of DR varied for transgenic mice, with 30–65% of cells staining positive for the transgene by flow cytometry in cells from the spleen and thymus (Figure 2A).

Figure 2.

Higher expression of DR4 in DR4.AEo-transgenic mice than in DR4.Aβo-transgenic mice. The percentage expression of DR4 on CD4+ cells from male and female AEo-transgenic mice is increased in activated cells. A, Expression of DR in naive DR4.Aβo (gray line) and DR4.AEo (black line) mice in thymus and spleen cells and in gated splenic B220+ and CD3+ cells. Values are percentages. Dotted line represents the isotype control. B, Expression of DR on gated CD3+,CD4+ and CD3+,CD8+ cells from naive mice and on activated T cells from mice primed with type II collagen (CII) or staphylococcal enterotoxin B (SEB). Values are percentages in cells from male (black line) and female (gray line) mice and in the isotype control (dotted line). The difference in mean fluorescence intensity between naive and activated CD4+ cells was significant (P < 0.05). Results are representative of 2–3 experiments performed with cells pooled from 2 mice per strain per experiment for all experiments.

The DR molecule is expressed at higher levels in AEo-transgenic mice than in Aβo-transgenic mice. To determine the expression of HLA transgene in mice as compared with humans, we studied the expression of DR4 on spleen or lymph node B220+ and CD3+ cells. As expected, B220+ cells stained positive for DR4 (Figure 2A). FACS analysis of CD3+, CD4+, and CD8+ cells demonstrated expression of DR on CD3+ cells as well as CD3+ subsets in these transgenic mice (Figure 2B). When activated with CII and SEB, the expression of DR4 on CD3+,CD4+ cells was increased. Activated antigen-specific CD4+ cells consistently showed significantly higher expression of DR, as determined by mean flow intensity, in female mice than in male mice, although the difference was not statistically significant. However DR expression was significantly higher in activated CD4+ cells compared with naive CD4+ cells (P < 0.05). The number of cells expressing DR was similar in CD8+ naive and SEB-activated cells.

To test whether the DR4 transgene was functional in the selection of various TCRs in transgenic mice, we studied the Vβ T cell repertoire profile. All DR4 mice showed partial deletion of Vβ5.1 and Vβ4 (data not shown) as compared with B10 mice (3). An increased number of Vβ6+ T cells were selected by DR4-transgenic mice as compared with B10 mice. A comparison of the Vβ profile in Aβo-transgenic mice and AEo-transgenic mice suggested that there was a similar profile, except for an increased frequency of Vβ8, which was not statistically significant. Findings in the DQ8.AEo mice were similar to those in the DQ8.Aβo described previously (7).

Presence of DR4 and disease susceptibility in female mice.

Our previous study showed that only 15% of DR4.Aβo mice develop CIA (unpublished observations). In the present study, we compared the incidence, severity, and sex distribution of arthritis in DR4.AEo and DR4.Aβo mice (Table 1). In DR4.AEo mice immunized with chick CII, 39% (14 of 36) developed arthritis. However, in DR4.AEo mice, CIA was observed predominantly in females (52% versus 15% of males; P = 0.03) (Figure 3A). There was no significant difference in the severity of arthritis in female versus male DR4.AEo mice, although arthritis onset occurred much earlier in the females (P = 0.03). A sex bias in the susceptibility to arthritis was not observed in DQ8-restricted arthritis (Table 1). None of the transgene-negative littermates (Aβo and AEo) developed arthritis.

Table 1. Incidence of CIA in DRB1*0401-transgenic mice*
Mouse strainCIA incidence, no. positive/no. tested (%)CIA severity, mean ± SDCIA onset, mean ± SD weeksAnti-CII antibodies, mean ± SD
FemalesMalesTotalChick CIIMouse CII
  • *

    Values represent all mice from 2 experiments. No significant differences were observed in the DQ8.Aβo and DQ8.AEo mouse strains. CIA = collagen-induced arthritis; anti-CII = anti–type II collagen; ND = not done.

  • P = 0.03 versus female DR4.AEo mice.

  • P = 0.06 versus DR4.Aβo mice.

  • §

    P = 0.04 versus DR4.Aβo mice.

DR4.AEo12/232/1314/36 (39)3.5 ± 1.25.2 ± 1.4§0.6 ± 0.10.4 ± 0.2
DR4.Aβo1/51/82/13 (15)1.5 ± 0.56.5 ± 0.7NDND
DQ8.AEo6/98/1114/20 (70)6 ± 1.25.3 ± 1.30.7 ± 0.50.7 ± 0.0
DQ8.Aβo5/84/79/15 (60)5.1 ± 1.85.6 ± 1.10.8 ± 0.00.7 ± 0.0
AEo0/50/30/8 (0)    
o0/30/30/6 (0)    
Figure 3.

Sex bias in susceptibility to collagen-induced arthritis in DR4-transgenic mice, with predominantly female mice developing disease. A, Incidence and onset of arthritis in male and female DR4.Aβo (AboM and AboF) and DR4.AEo (AEoM and AEoF) mice, showing increased susceptibility and earlier onset in female DR4.AEo mice. Arthritis was scored as described elsewhere (14). B, Hematoxylin and eosin staining of paw sections from an arthritic DR4.AEo mouse (top), showing infiltration of cells in the synovium and erosive arthritis in the joint of a digit, and from a normal mouse (bottom), without arthritis. C, Rheumatoid factor (RF) levels in sera collected 35 days after immunization. Group 1 represents type II collagen (CII)–primed DR4.AEo mice (n = 12), group 2 represents CII-primed DQ8.AEo mice (n = 15), group 3 represents naive DR4 mice (n = 6), group 4 represents Freund's complete adjuvant (CFA)–injected DR4 mice (n = 6), and group 5 represents AEo mice (n = 10). Both DR4.AEo and DQ8.AEo mice produced IgG and IgM RFs. DR4.AEo mice immunized with CFA alone and naive mice were used as controls. Bars show the mean for each strain. Upper and lower horizontal lines indicate the means in MRL-lpr (positive control) and CII-primed B10 (negative control) mice, respectively. Data points below the mean for the negative controls indicate mice that did not develop arthritis. OD = optical density.

The onset of disease was earlier (mean ± SD 5.2 ± 1.4 weeks in DR4.AEo and 6.5 ± 0.7 weeks in DR4.Aβo mice; P = 0.04) and the arthritis was more severe (P = 0.06) in DR4.AEo mice as compared with DR4.Aβo mice. The arthritis was more prominent in the digits of the paws in DR4.AEo mice as compared with DQ8.AEo mice. Histopathologic analysis of sections stained with hematoxylin and eosin showed severe infiltration and erosive arthritis of the digits of the paws in DR4 mice (Figure 3B).

Development of autoantibodies.

All arthritic mice developed anti-CII antibodies in response to immunizing and self collagen (Table 1), and all developed IgG and IgM RFs (Figure 3C). Levels of IgG-RF were significantly higher in DR4 mice than in either their transgene-negative littermates (P < 0.0001), CFA-immunized DR4-transgenic mice (P < 0.005), or naive DR4-transgenic mice (P < 0.005). There was no difference in IgG-RF and IgM-RF levels between DR4-transgenic and DQ8-transgenic mice. The levels of RF correlated with the severity of disease (data not shown). Very few mice produced RF levels as high as those in the MRL/lpr mice (positive controls). RF levels in B10 mice (negative controls) and in naive mice were similar to those in mice immunized with CFA only. IgM-RF levels were higher in DR4 mice compared with CFA-immunized and transgene-negative CFA-immunized mice, although the difference did not reach statistical significance (P < 0.08).

We also tested mice with and without CIA for the presence of anti-CCP antibodies. Mice with CIA had very high levels of anti-CCP antibodies (mean ± SD 36 ± 3 units/ml), whereas mice without CIA had very low levels (8 ± 4 units/ml) of these antibodies (P < 0.005).

HLA-restricted in vitro response to CII in transgenic mice.

Cells isolated from CII-primed DR4.AEo mice responded to CII when challenged in vitro (Figure 4A). The response was CD4-mediated and was restricted by the transgene, as indicated by the results of the inhibition assay. To determine if the response to DR4-restricted peptide in DR4.AEo mice is similar to that in DR4.Aβo mice, we studied both transgenic mouse strains for in vitro responses to human CII–derived peptide 254–273. Both transgenic mouse strains mounted a CD4-mediated and major histocompatibility complex (MHC)–restricted response to the peptide, although a higher response was observed in DR4.AEo mice (P < 0.004) (Figure 4B). DR4.AEo mice produced Th1 cytokines (IFNγ, IL-18, and TNFα) as well as regulatory cytokines (IL-10 and TGFβ) when challenged in vitro with CII or with peptide (Figure 4C).

Figure 4.

In vitro response to type II collagen (CII) and human CII–derived peptide 254–273. Results suggest that CD4 cells from DR4.AEo mice can present antigen in vitro. A, DR4.AEo mice were immunized with CII emulsified in Freund's complete adjuvant. For the inhibition assay, spleen cells were cultured in the presence of CII, either alone or with one of the following antibodies: GK1.5 (anti-CD4), Lyt2 (anti-CD8), or L227 (anti-DR). In addition, CD4+ T cells from CII-primed mice were sorted by fluorescence-activated cell sorting (FACS) and cultured in the presence and absence of antigen-presenting cells (APCs) as described in Materials and Methods. SI = stimulation index; nt = not tested. B, Comparison of the response of lymph node cells to human CII–derived peptide 254–273 (HII 254–273) from peptide-primed DR4.Aβo and DR4.AEo mice, showing a significantly higher response in the latter group (P < 0.004). C, Cytokine profile (tumor necrosis factor α [TNFα], interleukin-18 [IL-18], IL-10, transforming growth factor β [TGFβ], and interferon-γ [IFNγ]) in DR4.AEo mice in response to in vitro challenge with CII and human CII–derived peptide 254–273. D, CD4+ T cells from DR4.AEo mice primed with human CII–derived peptide 254–273 were sorted by FACS and cultured in the presence and absence APCs. The results in A and D represent a comparison of collagen and its derived peptide using sorted CD4+ cells from DR4.AEo mice. Values are the mean ± SD of 2–3 experiments using pooled cells from 2–3 mice per experiment.

Antigen presentation, but not processing, by CD4+ T cells from DR4.AEo mice.

We tested the ability of CD4+ T cells to process and present antigen in vitro. CD4+ cells from DR4.AEo mice that had been primed with CII and with human CII–derived peptide 254–274 were sorted by flow cytometry after staining with specific antibody. The rest of the cells were designated as CD4−. In vitro assay was performed using spleen cells, sorted CD4+ cells only, CD4+ cells in the presence of irradiated APCs, and CD4− cells in the presence of irradiated APCs. Sorted CD4+ cells mounted an in vitro response, with an SI of 3, when challenged with the peptide (Figure 4D), but not when challenged with CII (Figure 4A). In the presence of irradiated APCs, sorted CD4+ cells proliferated to an extent similar to that of whole spleen cells when challenged with peptide (SI of 5), but the response to CII was much lower. The CD4− cells responded poorly to the peptide (data not shown).

Higher cellularity and IFNγ production by female mice than by male mice.

A comparison of female and male DR4.AEo mice showed a higher number of splenic CD3+ and CD4+ cells in the females than in the males (P < 0.05) (Figure 5A). The high cellularity in female mice was associated with higher spleen weight as compared with that in male mice (P = 0.01). In vitro, lymph node cells isolated from mice primed with immunodominant human CII–derived peptide and CII mounted a stronger response and produced higher levels of proinflammatory cytokines in cells from females than in cells from males, with a significant increase in IFNα levels (P < 0.05) (Figure 5B). Characterization of activated cells after immunization consistently showed a higher number of activated CD4+ cells that were also positive for CD25+ and CD69+ in female mice, although the difference did not reach statistical significance. Even though the total number of Vβ8 cells were not different between female and male mice (results not shown), expansion of Vβ8+ CD4 T cells was higher in female than in male mice after immunization (P = 0.05). In addition, analysis of sera from primed mice showed that female mice produced higher levels of anti-CII antibodies in vivo than did male mice (Figure 5D), but the difference did not reach statistical significance (P = 0.06).

Figure 5.

Higher cellularity and higher production of inflammatory cytokines and antibodies in female DR4.AEo mice than in male DR4.AEo mice. Values are the mean ± SD of 2–3 experiments using pooled cells from 2–3 mice per experiment. ∗ = P = 0.05 for females versus males. A, Numbers of spleen cells and cell subsets in male and female DR4.AEo mice. B, Lymph node cells isolated from mice primed with human type II collagen (CII)–derived peptide 254–273 (HII 254-273) and with CII, showing higher levels of proinflammatory cytokines (interleukin-18 [IL-18] and interferon-γ [IFN-γ]) in females than in males. Cytokines were measured in culture supernatants by enzyme-linked immunosorbent assay (see Materials and Methods for details). C, Levels of anti-CII antibodies (Abs) in sera obtained 35 days after immunization from DR4.AEo mice, showing higher levels in females than in males. Only data from mice that developed collagen-induced arthritis are shown. OD = optical density.


Mouse models of RA have often been critiqued as having inherent immunologic shortcomings: expression of endogenous mouse class II chains obfuscates interpretation of the data; mice do not express class II molecules on T cells; and mice models fall short in generating endocrinologic conditions that reproduce the female-biased susceptibility to RA in humans. The model presented here is the first to share similarities with humans for the major differences observed in human RA and mouse models of arthritis. Mice used in the present study lack all the classic murine endogenous class II chains.

One major difference between human and mouse class II molecules is that human class II molecules are expressed on T cells, especially activated T cells, whereas mouse class II molecules are not (16, 17). Human T cell clones can present antigen (18). Thus, expression of class II molecules on T cells could potentially play a role in human autoimmune diseases. During the onset of disease, activated CD4 T cells infiltrate the target tissue, which results in inflammation. The class II+ CD4 T cells could present locally available peptides and activate other cells. DR4.AEo mice express class II molecules on T cells, a feature similar to that in humans. Although our study does not directly implicate these T cell–expressing class II molecules in disease pathogenesis, a probable scenario can be envisaged. We observed increased expression of DR in activated T cells. CD8 cells showed a higher number of DR+ cells when activated by CII as compared with SEB, although the difference was not statistically significant. One possibility could be that compared with CII, SEB leads to vigorous expansion of CD8 cells, resulting in a lower percentage of cells that are positive for DR.

The other major difference between RA and models of CIA is that whereas there is a significant sex bias and production of autoantibodies such as RFs and anti-CCP in RA (19), neither of these occurs in CIA. Studies in humans have shown that anti-CCP antibody production precedes the onset of RA and are associated with the presence of DR4 (20, 21). The DR4-transgenic mice described herein were shown to produce RFs as well as anti-CCP antibodies. Thus, we have generated a new humanized mouse model of RA that mimics the phenotype of the human disease and shows a sex bias.

Data from the present study of DRB1*0401-transgenic mice show that DR4 renders female mice susceptible to the development of CIA in the absence of all of the endogenous class II molecules. DR4 occurs in linkage disequilibrium with DQ8 in humans and is inherited en bloc as a haplotype. DQ8-transgenic mice did not develop sex-biased CIA, which lends support to our observations that the sex bias could be due to DR4. Previous studies using the CIA model showed a contribution of DR4, but all the features of human RA were not presented in those models (3, 9, 10). This could have been due to the presence of endogenous class II chains that made it difficult to interpret the observations.

As in humans with RA, the reason why female mice are more susceptible to the development of CIA than male mice is not yet clear. Hormones have long been predicted to be one of the major reasons for this bias in autoimmunity in humans (22). Another important sex difference that can contribute to the immune response is the higher number of absolute CD4+ cells in women than in men (23). The data presented here reproduce the findings in humans. We found that the female mice responded with much stronger T cell responses to the DR4-restricted peptide than did the male mice. One possible reason for this stronger response could be the high cellularity and increased number of CD4+ cells in the female mice. A stronger in vitro response to the immunodominant peptide, increased amounts of inflammatory cytokines, IFNγ, and autoantibodies in female mice may be other reasons for the sex bias and earlier onset of disease, since there was no difference in the severity of arthritis between the two sexes. A recent study showed that deprivation of androgen in mice increases the cellularity of primary and peripheral lymphoid organs and T cell proliferation (24). This suggests that androgen might play a role in protecting male mice from the development of CIA. Estrogen has been shown to influence the immune response by influencing B cell survival and modulating dendritic cell function (25, 26). We observed a higher number of B cells in female mice, even though the difference did not reach statistical significance.

To understand the sex differences in autoimmunity, the basic immune response needs to be studied. Since the innate immune response drives the adaptive response, both the MHC and the cells of innate immune responses may be important in the development of autoimmunity. While studies in humans are difficult to undertake and interpret, the model described here provides a very good choice with which to study the mechanism of sex bias in autoimmunity.

MHC molecules have been shown to be important in the selection of the T cell repertoire in the thymus (27). Our data show that DR4 can positively and negatively select the TCR Vβ profile, thus suggesting that it can present peptides in the thymus. A significant observation in the present study was that CD3+ cells expressed class II molecules, a feature in humans, but not in mice. To determine if the HLA transgene expressed on T cells of transgenic mice plays a role in disease pathogenesis, we tested the ability of CD4+ T cells to process and present antigen. Sorted CD4+ cells mounted an in vitro response when challenged with the peptide, but not when challenged with CII, which suggests that they can present peptides but cannot process whole CII. This could indicate that in humans, activated CD4+ T cells could present antigen locally, leading to exacerbation of disease.

Our data suggest that DR4 can present antigen in a MHC-restricted manner and mount an immune response in vivo. DR4.AEo mice mounted a stronger response than DR4.Aβo mice to DR4-restricted collagen-derived peptide and produced the proinflammatory cytokines IFNγ, IL-18, and TNFα and the regulatory cytokines IL-10 and TGFβ in response to in vitro challenge with the immunizing CII and peptide. IL-10 has been associated with protection from arthritis in most of the experimental models studied (28). However, it can also be produced by CD4+ T cells in response to autoantigen presented by B cells (29). TGFβ has been shown to promote bone destruction in animal models of CIA and in humans with RA, which is related to an increase in the expression of proinflammatory cytokines and metalloproteinases (30). IL-18 induced by TNFα and IFNγ are important in the development of CIA (31).

The new humanized mice provide a better model for studying arthritis since they mimic the major features of the disease in humans. In conclusion, we have shown that the new transgenic mice are similar to humans in their expression of class II molecules on CD4 T cells, that activated CD4+ T cells can present peptides but do not process the antigen, and that the new AEo-transgenic mice develop autoimmunity in a sex-biased manner (a female to male ratio of ∼3:1) and produce autoantibodies similar to humans with RA.


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

Study design. Drs. Taneja and David.

Acquisition of data. Mr. Behrens and Dr. Mangalam.

Analysis and interpretation of data. Drs. Taneja, Griffiths, Luthra, and David.

Manuscript preparation. Drs. Taneja and David.

Statistical analysis. Dr. Taneja.


We thank Julie Hanson and Tad Trejo for maintaining the transgenic mice and Michele Smart for screening the transgenic mice.