Masato Nose, MD, PhD, Department of Pathology, Division of Pathogenomics, Ehime University Graduate School of Medicine, Shitsukawa, Toon, Ehime 791-0295, Japan. Email: email@example.com
Based on the hypothesis that the complex pathological and immunological manifestations of rheumatoid arthritis (RA) and the related diseases are under the control of multiple gene loci with allelic polymorphism, a recombinant congenic mouse strain was prepared between an MRL/Mp-lpr/lpr (MRL/lpr) strain, which develops arthritis resembling RA, and a non-arthritic strain C3H/HeJ-lpr/lpr (C3H/lpr). In MRL/lpr × (MRL/lpr × C3H/lpr) F1 mice, the mice developing severe arthritis were selected based on joint swelling to further continue intercrosses, and then an McH-lpr/lpr-RA1 (McH/lpr-RA1) strain was established and its histopathological phenotypes of joints and autoimmune traits were analyzed. Arthritis in McH/lpr-RA1 mice developed at a higher incidence by 20 weeks of age, compared with that in the MRL/lpr mice, who had severe synovitis (ankle, 60.3%; knee, 65.1%), and also fibrous and fibrocartilaginous lesions of articular ligamenta resembling enthesopathy (ankle, 79.4%; knee, 38.1%), resulting in ankylosis. The lymphoproliferative disorder was less, and serum levels of IgG and IgG autoantibodies including anti-dsDNA and rheumatoid factor were lower than those of both MRL/lpr and C3H/lpr strains. McH/lpr-RA1 mice may provide a new insight into the study of RA regarding the common genomic spectrum of seronegative RA and enthesopathy.
Rheumatoid arthritis and the related diseases such as spondyloarthropathy and juvenile idiopathic arthritis show complex pathological manifestations of articular lesions, involving synovitis, enthesitis, and/or abnormal proliferation of bone and cartilage.1,2 In general, disease categories with complex pathological manifestations have been defined based on clinical criteria, but pathology of the critical lesions in disease could reflect the specific phenotypes regarding pathogenesis. Recent studies of mouse genomics to dissect pathological phenotypes could provide a new insight into this subject,3–8 leading to the hypothesis that complex pathological phenotypes of diseases are under the control of multiple gene loci with allelic polymorphism.
According to this hypothesis, we examined remodeling of articular lesions in an MRL strain of mice bearing a Fas antigen deletion mutant, MRL/MpJ-lpr/lpr (MRL/lpr)9 with genome recombination. MRL/lpr mice spontaneously develop collagen diseases involving arthritis, which resembles rheumatoid arthritis (RA) in that they exhibit many of the serological characteristics of RA, including increased serum levels of antinuclear antibodies, circulating immune complexes, and IgM- and IgG- rheumatoid factors (RF).10–12 Moreover, the histopathological features of the articular lesions in these mice are characterized by synovial cell proliferation, inflammatory cell infiltration, granulomatous inflammation in synovial sublining tissue and pannus formation.11,13,14 But the incidence of arthritis in these mice is <30%, and the severity is limited according to microscopic observations, moreover, the incidence of arthritis is not associated with joint swelling.
In the present study we first prepared backcross generation mice using a non-arthritis strain of mice, C3H/HeJ-lpr/lpr (C3H/lpr), MRL/lpr × (MRL/lpr × C3H/lpr) F1. Among these N2 mice, we observed mice developing arthritis of the ankle joints with macroscopic swelling. We then began to intercross the N2 mice by means of selection based on swelling of the ankle joints. Finally, we established a novel recombinant strain of mice, designated McH-lpr/lpr-RA1, which developed arthritis at a high incidence and with enthesopathy.
MATERIALS AND METHODS
MRL/lpr, C3H/lpr and MRL/Mp-+/+(MRL/+) mice were purchased from Jackson Laboratory, Bar Harbor, Maine, USA. Using the former two strains, we prepared backcross generation mice, MRL/lpr × (MRL/lpr × C3H/lpr) F1. Among the N2 mice, we observed the mice that developed arthritis, macroscopically manifested by ankle joint swelling. We then established a recombinant congenic strain of mice that developed arthritis by means of selection based on the swelling of ankle joints. From the generation F54, this recombinant congenic strain of mice was designated McH-lpr/lpr-RA1 (McH/lpr-RA1). These mice were housed in the Animal Research Institute of Tohoku University Graduate School of Medicine under specific pathogen-free and climate-controlled conditions with 12 h light–dark cycles, and each female and male littermate at 5 weeks of age was removed to cages (fewer than six per cage) separated from their parents to avoid fighting as much as possible.
At the indicated weeks of age, each mouse was bled under ether anesthesia. Sections of kidneys, heart, lungs, spleen, pancreas, salivary glands and lymph nodes were then fixed in 10% formalin in 0.01 mol/L phosphate buffer (pH 7.2) and embedded in paraffin. They were stained with HE, elastica-Masson (EM) and PAS for histological examination on light microscopy.
For ankle joints, serial sections were taken sagittally through the talus. The lesions, including calcaneus bone and anterior and posterior synovial tissue were histopathologically evaluated with the following grading system: normal, grade 0; thickening and proliferation of synovial lining, grade 1; grade 1 with granulomatous and/or fibrous lesions in synovial sublining tissue, grade 2; and grade 2 with pannus formation against bone cortex and/or bone marrow, grade 3. For knee joints, serial sections were prepared sagittally through the patella. The articular lesions including the tibia, femur, meniscus, plica alaris and capsula articularis were similarly evaluated as described in the previous section. We categorized grade 0 or 1 individuals as synovitis negative, and grade 2 or 3 individuals as synovitis positive.
The presence of ankylosis of the ankle and knee joints was also evaluated as follows: normal, grade 0; fibrous and fibrocartilage proliferative lesions in joints involving articular ligamenta and periostea, slight, grade 1; lesions similar to grade 1, moderate, grade 2; and lesions similar to grade 2, associated with bone destruction, grade 3. We categorized individuals with a score of higher than grade 1 as ankylosis positive. The incidence of vasculitis or glomerulonephritis was calculated as described previously.15
Lymphoid cells from the lymph nodes and spleen were examined for surface markers on flow cytometry. The cells were treated with fluorescein isothiocyanate (FITC)-conjugated rat monoclonal antibodies to Thy1.2 (clone 3D-H12; Becton Dickinson, Bedford, CA, USA), and phycoerythrin (PE)-conjugated rat monoclonal antibodies to B220 (clone RA3.6B2; PharMingen, San Diego, CA, USA), with FITC-conjugated rat monoclonal antibodies to CD4 (clone GK1.5) and PE-conjugated rat monoclonal antibodies to CD8 (clone 53–6.7; PharMingen), or with FITC-conjugated goat polyclonal antibodies to IgG or IgM (Zymed, San Francisco, CA, USA). The Fas-expressing cells were analyzed using FITC-conjugated hamster monoclonal antibodies to Fas (clone Jo2; PharMingen). These cells were analyzed on FACSCalibur using the CellQuest program (Becton Dickinson, Franklin Lakes, NJ, USA).
Immunoglobulin and autoantibody serum level
The serum levels of immunoglobulins involving IgG and IgM were quantified on single radial immunodiffusion, using goat polyclonal antimouse IgG and IgM antibodies (Zymed), respectively. The amounts of each in the serum samples were estimated from a standard curve using mouse Ig standards (Bethyl Laboratories, Montgomery, TX, USA).
The relative amounts in the serum of double-stranded DNA (dsDNA) antibodies were determined on ELISA using calf thymus DNA (type I, Sigma Chemical, St. Louis, MO, USA) and alkaline phosphatase-conjugated rabbit polyclonal antibodies to mouse IgG (AP-antimouse IgG; Sigma) according to a method previously described.16 Rheumatoid factor was measured as IgG antibodies that specifically bound to the recombinant Fc fragments derived from human IgG1.17
The association of each strain with the incidence of synovitis or ankylosis was evaluated using Fisher's exact test. The difference between each strain in the weight of spleen, axillary lymph nodes, and serum levels of IgG, IgM and autoantibodies was evaluated on Student's t-test. P < 0.01 was considered to be significant.
Higher incidence of arthritis and enthesopathy
Ankle joint swelling of McH/lpr-RA1 mice was significantly observed at later than 16 weeks of age ( Fig. 1). At 20 weeks of age, these mice microscopically showed severe inflammatory synovitis with pannus formation resembling that in MRL/lpr mice (Fig. 2a), but the incidence of synovitis in the ankle joints was remarkably higher (60.3%) than in the MRL/lpr mice (8.0%; Table 1). Similar lesions were frequently observed in the knee joints (Fig. 2b). The incidence of synovitis of knee joints in McH/lpr-RA1 mice was not significantly higher than that in the MRL/lpr mice (Table 1).
Table 1. Incidence of synovitis in McH/lpr-RA1, MRL/lpr and C3H/lpr mice
Among McH/lpr-RA1 mice, many individuals developed ankylosis in their ankle and knee joints. Ankylosis in this strain of mice was characterized by fibrous and fibrocartilage proliferative lesions in the joints involving articular ligamenta and periostea, partly occupying the joint space, resembling enthesopathy (Fig. 2c–f). These lesions seemed to be specific for McH/lpr-RA1 mice because no MRL/lpr and C3H/lpr mice developed these lesions (Table 2). Sex difference in ankylosis of McH/lpr-RA1 mice was not a significant factor in the ankle or knee joints (P = 0.073 and P = 0.093, respectively), thus indicating that ankylosis in these mice might not result from an aggressive behavior as is generally considered in male mice. It seemed that synovitis, at least in ankle joints, developed prior to ankylosis at 8 week of age because ankylosis in ankle joints was never observed in any McH/lpr-RA1 mice at 8 weeks of age (Tables 1,2). But the correlation between the synovitis grades and ankylosis grades, at least in the ankle joints at 20 weeks of age, was not significant (female, correlation coefficient: r = −0.167; male; r = 0.404; total, r = 0.183).
Table 2. Incidence of ankylosis in McH/lpr-RA1, MRL/lpr and C3H/lpr mice
Development of vasculitis, but no glomerulonephritis
McH/lpr-RA1 mice coincidentally developed vasculitis, but not glomerulonephritis. The incidence of vasculitis in kidney was similar to that in MRL/lpr mice at 20 weeks of age (McH/lpr-RA1, 63/63; MRL/lpr, 17/25; P = 0.186). The histopathological features of vasculitis were characteristic of granulomatous arteritis, similar to those of MRL/lpr mice (data not shown). In contrast, the incidence of glomerulonephritis was significantly different (McH/lpr-RA1, 12/63; MRL/lpr, 17/25; P < 0.004). This may be the reason why the McH/lpr-RA1 mice lived remarkably longer than the MRL/lpr mice.
Dissociation from lymphoproliferative disorder
Importantly, all mice of the McH/lpr-RA1 strain had only mild splenomegaly and lymphadenopathy (spleen, 0.275 ± 0.145 g; lymph nodes, 0.174 ± 0.164 g), different from the original strains of MRL/lpr (spleen, 0.531 ± 0.199 g, P < 0.001; lymph nodes, 0.730 ± 0.298 g, P < 0.001) and C3H/lpr mice (spleen, 0.562 ± 0.189 g, P < 0.001; lymph nodes, 0.875 ± 0.288 g, P < 0.001), although there was no difference between MRL/lpr and C3H/lpr mice in spleen and lymph nodes (P > 0.01).
The accumulation of Thy1+B220+ T cells into the spleen and lymph nodes, which is a characteristic phenotype of the lpr gene,18 was not remarkable, in comparison to that in MRL/lpr and C3H/lpr mice (Fig. 3a). This seemed to reflect the decrease of the double negative (CD4–CD8–) T cells in the spleen and lymph nodes and, in contrast, the increase of CD4+ T cells and especially CD8+ T cells (Fig. 3b). As a result, the CD4/CD8 ratio of spleen cells and lymph node cells was found to be remarkably lower in the McH/lpr-RA1 mice than in the MRL/lpr or C3H/lpr mice, although the high CD4/CD8 ratio in lpr-bearing mice is thought to result from the absence of Fas-mediated apoptosis in CD4+ T cells.15 But the expression of Fas in splenic cells and lymph node cells was remarkably lower than that in MRL/+mice: almost at the same level as that in MRL/lpr and C3H/lpr mice (Fig. 3c).
Dissociation from autoimmune traits
The serum levels of IgG in McH/lpr-RA1 mice were significantly lower than those of MRL/lpr and C3H/lpr mice, whereas the IgM level was similar to that in MRL/lpr mice (Fig. 4) although the number of IgG or IgM positive cells in the spleen and lymph node in this strain were relatively higher than those of MRL/lpr and C3H/lpr mice (Fig. 3d,e).
The serum levels of both IgG anti-dsDNA antibodies and IgG rheumatoid factors in McH/lpr-RA1 mice were remarkably low in comparison to those of both MRL/lpr and C3H/lpr mice (Fig. 5). These titers were almost the same in the McH/lpr-RA1 mice at 8 weeks of age (data not shown).
In the present study we established a novel recombinant congenic strain of mice, McH/lpr-RA1, which was started from the MRL/lpr × (MRL/lpr × C3H/lpr) F1 generation, by means of selection based on swelling in the ankle joints. In this strain, synovitis developed in the ankle joints at a higher incidence than that in the MRL/lpr mice, and ankylosis manifested by enthesopathy developed in the ankle and knee joints. This supports the idea that arthritis is controlled by polygenic inheritance. Namely, a particular genome combination with C3H/lpr alleles accelerates the incidence of synovitis and produces enthesopathy.
Such a genome recombination was also useful for evaluating several serological parameters that might be responsible for the development of diseases. McH/lpr-RA1 mice had a lower level of both IgM and IgG despite the presence of the lpr gene (Fig. 4a,b), and their IgG autoantibodies were fewer than in MRL/lpr and C3H/lpr mice (Fig. 5a,c). In previous studies in MRL/lpr mice, high serum levels of IgG rheumatoid factor seemed to play a critical role in the development of arthritis.11,19 But it became clear in the present study that high serum levels of IgG rheumatoid factor as found in MRL/lpr mice are not necessary for the development of arthritis in McH/lpr-RA1 mice. As a result, the pathogenetic role of IgG rheumatoid factor on arthritis in MRL/lpr mice therefore needs to be re-evaluated.
Moreover, lymphoproliferative disorders manifested by splenomegaly and lymphadenopathy decreased in McH/lpr-RA1 mice in comparison to those in MRL/lpr and C3H/lpr mice, although the Fas deletion levels on the cell surface were almost the same among these three lpr strains (Fig. 3c). Throughout the repeated intercrossing for the establishment of the recombinant congenic strain, we observed that the lymph node weights seemed to decrease for each generation. Correspondingly, the accumulation of Thy1+B220+ T cells into spleen and lymph nodes decreased (Fig. 3a) while the ratio of CD4/CD8 decreased (Fig. 3b). This indicates that T-cell subsets in the periphery might be controlled by an allelic combination in the lpr mice, and that a mechanism besides Fas-mediated apoptosis has become accelerated in this strain. The expression levels of tumor necrosis factor-α should be further studied as one of the candidates inducing apoptosis,20 which might also correlate with the higher incidence of synovitis in this strain.
It is important to elucidate whether ankylosis in the McH/lpr-RA1 mice might be the result of synovitis. In general, synovitis associated with pannus formation is though to lead ankylosis in RA, mainly characterized by the replacement of the joint surface by fibrous tissue derived from the invading pannus. Ankylosis in this strain of mice, however, seems to be due to enthesopathy, which was histopathologically manifested by the fibrous and fibrocartilaginous proliferation of enthesis, the site of insertion of a tendon, ligamenta or articular capsule into bone, similar to that previously described.21 If that is the case, then synovitis in this strain of mice might be a result of secondary synovitis following enthesopathy.22 But this is unlikely considering that the correlation between the synovitis grades and the ankylosis grades in the ankle joints of McH/lpr-RA mice at 20 weeks of age was not significant, and also that MRL/lpr mice developed synovitis in knee joints as often as McH/lpr-RA1 mice, but never developed ankylosis (Tables 1,2). As a result, the development of enthesopathy may be genetically independent from synovitis. Similar lesions to enthesopathy in McH/lpr-RA1 mice can be observed in male DBA/1 mice that spontaneously develop the lesions in their ankles or lower extremities at older than 4 months of age, but these mice are not associated with severe synovitis.23,24 We recently found that the severity and early onset of enthesopathy in DBA/1 mice were accelerated by a gene locus in an MRL allele, which was localized on chromosome 10 and designated Adm1.25 This locus in McH/lpr-RA1 mice was an MRL homozygote, which involves several candidate genes mediating cellular proliferation and differentiation such as β2 integrin, basigin and macrophage migration inhibitory factor (MIF). In humans, MIF and ITGB2 (β2 integrin) are associated with juvenile idiopathic arthritis.26,27
The present findings thus suggest that a diversity of pathological manifestations of arthritic lesions might be involved in a common genomic spectrum of collagen disease including RA, spondylarthropathy and juvenile idiopathic arthritis, especially enthesitis-related arthritis.2
The authors are indebted to Ms S. Terasaki for technical help in animal care, and Ms E. Kondo and M. Terada for assistance with histopathology. This study was supported by Grant-in Aid for the Scientific Research Funds of the Ministry of Education, Science and Culture of Japan (S.M. 15390607, M.N. 18390123).