Veterinary Pathology OnlineFirst, published on June 10, 2015 as doi:10.1177/0300985815588612
Review
Mouse Models of Rheumatoid Arthritis
Veterinary Pathology 1-9 ª The Author(s) 2015 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0300985815588612 vet.sagepub.com
P. Caplazi1, M. Baca1, K. Barck2, R. A. D. Carano2, J. DeVoss3, W. P. Lee3, B. Bolon4, and L. Diehl1
Abstract Rheumatoid arthritis (RA) is a chronic debilitating autoimmune disorder characterized by synovitis that leads to cartilage and bone erosion by invading fibrovascular tissue. Mouse models of RA recapitulate many features of the human disease. Despite the availability of medicines that are highly effective in many patient populations, autoimmune diseases (including RA) remain an area of active biomedical research, and consequently mouse models of RA are still extensively used for mechanistic studies and validation of therapeutic targets. This review aims to integrate morphologic features with model biology and cover the key characteristics of the most commonly used induced and spontaneous mouse models of RA. Induced models emphasized in this review include collagen-induced arthritis and antibody-induced arthritis. Collagen-induced arthritis is an example of an active immunization strategy, whereas antibody- induced arthritis models, such as collagen antibody–induced arthritis and K/BxN antibody transfer arthritis, represent examples of passive immunization strategies. The coverage of spontaneous models in this review is focused on the TNFDARE mouse, in which arthritis results from overexpression of TNF-a, a master proinflammatory cytokine that drives disease in many patients. Keywords arthritis, experimental, arthritis, rheumatoid, models, animal, mice, collagen-induced arthritis, collagen antibody–induced arthritis, k/bxn antibody transfer arthritis, TNFDARE mouse
Lesions in rheumatoid arthritis (RA), a debilitating systemic autoimmune disease, are characterized as chronic destructive synovitis with predilection for small diarthrodial joints, especially those of the hands and feet. The pathogenesis of RA is complex and involves genetic predispositions as well as environmental factors.21 Central to the pathogenesis is the activation of macrophages by autoreactive T cells, resulting in the release of key proinflammatory cytokines, such as tumor necrosis factor a (TNF-a) and interleukins 1, 6, and 17 (IL-1, IL-6, and IL-17).13 Immunotherapies targeting these master cytokines, their receptors, or their downstream signaling components have shown dramatic success and greatly improved disease management in patient populations that have access to these therapies.13 Yet, the events that trigger the generation and recruitment of autoreactive lymphocytes in the first place are still emerging. For example, the role of human leukocyte antigen variants in promoting reactivity to citrullinated antigens has only recently been recognized as important in the initial pathogenesis.6,12,46 Antibodies to citrullinated antigens are detectable in a variety of autoimmune diseases, including RA, indicating that citrullination plays a role in the process of epitope spreading52 and the emergence of autoepitopes.35,60,64 Likewise, citrullinated antigens are relevant in animal models of autoimmunity.9,24,58 Neutrophils represent another example of hitherto underappreciated contributors to the pathogenesis of autoimmune diseases. The role of neutrophils in initiating
and promoting autoimmunity to citrullinated nuclear antigens through the extrusion of neutrophil extracellular traps is being reevaluated.70 Mouse models that recapitulate aspects of the pathogenesis in humans are an important tool for investigating such mechanisms in vivo and have been in use for many years. The purpose of this brief review is to summarize the key features of some of the more widely used induced and spontaneous models of immune-mediated joint disease (Fig. 1). For the appropriate selection of models according to features of pathogenesis in human RA, we recommend consulting the review by Vincent et al on behalf of the Arthritis Research UK Animal Models Working Group.63 That article also provides useful information on features of osteoarthritis and briefly touches on important 1 Departments of Research Pathology, Genentech Inc, South San Francisco, CA, USA 2 Biomedical Imaging, Genentech Inc, South San Francisco, CA, USA 3 Translational Immunology, Genentech Inc, South San Francisco, CA, USA 4 Department of Veterinary Biosciences and the Comparative Pathology and Mouse Phenotyping Shared Resource, The Ohio State University, Columbus, OH, USA
Corresponding Author: P. Caplazi, Department of Research Pathology, Genentech Inc, 1 DNA Way, South San Francisco, CA 94080, USA. Email: [email protected]
Downloaded from vet.sagepub.com at CARLETON UNIV on June 13, 2015
2
Veterinary Pathology therapeutic agents selectively influence some lesion features. For the assessment of bone lesions, sophisticated quantitative methods based on histomorphometry are possible47 but may require more advanced processing techniques, such as embedding of undemineralized tissue in methylmethacrylate. While histology remains the most widely used method for assessment of lesions in mouse models of RA, sophisticated in vivo or ex vivo imaging modalities exist that can complement or altogether replace histologic evaluation.2,29,32,62,71
Systemically Induced Mouse Models of RA
Figure 1. Key concepts shared between human rheumatoid arthritis (RA) and mouse models of RA. j Autoreactive T cells providing help to B cells or releasing a variety of proinflammatory cytokines are an important component in the initiation and maintenance of lesions in RA as well as mouse models.9,48 An intact adaptive immune system is not necessary for some models, however. Notably, T cells are not required for the induction of collagen antibody–induced arthritis.37 The antigen-experienced T-effector/memory cell phenotype is prominent among T cells in the RA synovium.8 k Antibodies to citrullinated antigens (anticitrullinated peptide antibodies [ACPAs]) that are implicated in the pathogenesis of RA have been shown to be important in mouse models, including collagen-induced arthritis.24,28 More broadly, antibody- and complement-mediated effects are important drivers in RA as well as mouse models.33 l Neutrophils are abundant in murine autoimmune arthritis and contribute to the pathogenesis through the release of cytotoxic products and immunoregulatory mediators. In addition, neutrophils may promote autoimmunity by formation of neutrophil extracellular chromatin traps (NETs) and the associated promotion of ACPA.11,70 m Cartilage and bone injury is driven by the formation of an inflammatory pannus that classically invades the joint from the capsular angle. In mouse models, however, substantial amounts of pannus also extend from the bone marrow cavity subjacent to the articular cartilage (Figs. 3, 9, 13). n Macrophages (MF) infiltrating the synovium are central in the pathogenesis of RA, serving as an important source of key proinflammatory cytokines such as interleukin 1b (IL-1b) and tumor necrosis factor a (TNF-a). Furthermore, macrophage-derived cytokines and their effects on synovial fibroblasts are essential for the differentiation of osteoclasts and osteolysis. Bones are shown in gray, articular cartilage in light blue, and synovium and adjacent soft tissues in light pink.
considerations for animal welfare and statistical powering of groups, areas that are not covered here. General practical considerations in the use of various rodent joint disease models are reviewed by Bolon et al.3 Finally, for a general scoring scheme of histologic lesions, we refer to Caplazi and Diehl.4 Besides this general scoring scheme, other, more narrowly defined morphologic scoring systems may be suited in situations where
Based on the method of induction, systemically induced models can be divided into 3 groups: those elicited by active immunization, those elicited by passive immunization, and those elicited by administration of irritant chemicals resulting in chronic inflammation. Collagen-induced arthritis (CIA) is the archetypical model of RA induced by active immunization. First developed in rats,59 the model was later adapted to the mouse16 and is now commonly used in both species, with the choice of rodent dependent on the scientific questions and types of therapies to be tested. Rats or mice develop an acute to subacute monophasic erosive polyarthritis after immunization with collagen II (but not other collagens) from many species. Type II collagen is the major constituent collagen form of articular cartilage, and immunoreactivity to type II collagen can be identified in some RA patients.7,56 Many important features of human RA are recapitulated in mouse CIA, including the presence of rheumatoid factor or anticitrullinated peptide antibody.51 At least in the mouse, a Th17-driven response is essential for the production of lesions.41,50 The effects of interferon g are divergent51 with inhibitory as well as stimulatory effects on development of disease. Overall, however, interferon g attenuates CIA.22,51 Practical aspects of working with the mouse CIA model have been reviewed elsewhere66 and so are only briefly summarized here. A typical induction protocol might use heterologous (eg, bovine) collagen II for immunization of mice from a strain with permissive major histocompatibility complex (MHC haplotype q, H-2q; eg, DBA/1). For induction, a subcutaneous injection of type II collagen suspended in complete Freund adjuvant on day 0 is followed by a booster of type II collagen in incomplete Freund adjuvant on day 21. Effects of experimental treatment can then be tested by applying a prophylactic design (treatment initiated before the onset of clinical signs at around day 20 to 30) or a therapeutic design (treatment initiated after the onset of clinical signs on animals selected for presence of clinical signs and then randomized into treatment groups). Effects can be studied longitudinally in vivo using noninvasive assessments, such as clinical arthritis score, paw volume, caliper measurements of ankle thickness, or imaging modalities (eg, micro–computed tomography). Parameters determined in vivo can be complemented by terminal assessments, which include imaging (micro–computed tomography, benchtop radiography, magnetic resonance imaging), histology, or biochemical
Downloaded from vet.sagepub.com at CARLETON UNIV on June 13, 2015
Caplazi et al
3
Figures 2–7. Collagen-induced arthritis (CIA), day 63 postinduction, tarsal region, sagittal section, mouse (DBA/1J). Hematoxylin and eosin. Figure 2. In CIA, complex joints are usually affected globally, with inflammatory cells and fibroplasia spanning several nearby individual joints leading to locally extensive expansion of the affected soft tissues. Synovitis of tendons and tendon sheaths is common and may precede arthritis proper in some instances. Ta, talus; Tar, central tarsal bone; Te, common digital flexor tendon; Ti, distal tibial condyle. The dashed rectangle denotes the area shown enlarged in Figure 3. Figure 3. Pannus (black arrows) extends from the inflamed periarticular soft tissue into the subcartilaginous bone and onto the proximal cartilage surface of the central tarsal bone. In contrast to human RA, pannus-like tissue in mice also erodes from the bone marrow cavity into the subcartilaginous bone. This process of subcartilaginous osteolysis is facilitated by the close proximity (
Review
Mouse Models of Rheumatoid Arthritis
Veterinary Pathology 1-9 ª The Author(s) 2015 Reprints and permission: sagepub.com/journalsPermissions.nav DOI: 10.1177/0300985815588612 vet.sagepub.com
P. Caplazi1, M. Baca1, K. Barck2, R. A. D. Carano2, J. DeVoss3, W. P. Lee3, B. Bolon4, and L. Diehl1
Abstract Rheumatoid arthritis (RA) is a chronic debilitating autoimmune disorder characterized by synovitis that leads to cartilage and bone erosion by invading fibrovascular tissue. Mouse models of RA recapitulate many features of the human disease. Despite the availability of medicines that are highly effective in many patient populations, autoimmune diseases (including RA) remain an area of active biomedical research, and consequently mouse models of RA are still extensively used for mechanistic studies and validation of therapeutic targets. This review aims to integrate morphologic features with model biology and cover the key characteristics of the most commonly used induced and spontaneous mouse models of RA. Induced models emphasized in this review include collagen-induced arthritis and antibody-induced arthritis. Collagen-induced arthritis is an example of an active immunization strategy, whereas antibody- induced arthritis models, such as collagen antibody–induced arthritis and K/BxN antibody transfer arthritis, represent examples of passive immunization strategies. The coverage of spontaneous models in this review is focused on the TNFDARE mouse, in which arthritis results from overexpression of TNF-a, a master proinflammatory cytokine that drives disease in many patients. Keywords arthritis, experimental, arthritis, rheumatoid, models, animal, mice, collagen-induced arthritis, collagen antibody–induced arthritis, k/bxn antibody transfer arthritis, TNFDARE mouse
Lesions in rheumatoid arthritis (RA), a debilitating systemic autoimmune disease, are characterized as chronic destructive synovitis with predilection for small diarthrodial joints, especially those of the hands and feet. The pathogenesis of RA is complex and involves genetic predispositions as well as environmental factors.21 Central to the pathogenesis is the activation of macrophages by autoreactive T cells, resulting in the release of key proinflammatory cytokines, such as tumor necrosis factor a (TNF-a) and interleukins 1, 6, and 17 (IL-1, IL-6, and IL-17).13 Immunotherapies targeting these master cytokines, their receptors, or their downstream signaling components have shown dramatic success and greatly improved disease management in patient populations that have access to these therapies.13 Yet, the events that trigger the generation and recruitment of autoreactive lymphocytes in the first place are still emerging. For example, the role of human leukocyte antigen variants in promoting reactivity to citrullinated antigens has only recently been recognized as important in the initial pathogenesis.6,12,46 Antibodies to citrullinated antigens are detectable in a variety of autoimmune diseases, including RA, indicating that citrullination plays a role in the process of epitope spreading52 and the emergence of autoepitopes.35,60,64 Likewise, citrullinated antigens are relevant in animal models of autoimmunity.9,24,58 Neutrophils represent another example of hitherto underappreciated contributors to the pathogenesis of autoimmune diseases. The role of neutrophils in initiating
and promoting autoimmunity to citrullinated nuclear antigens through the extrusion of neutrophil extracellular traps is being reevaluated.70 Mouse models that recapitulate aspects of the pathogenesis in humans are an important tool for investigating such mechanisms in vivo and have been in use for many years. The purpose of this brief review is to summarize the key features of some of the more widely used induced and spontaneous models of immune-mediated joint disease (Fig. 1). For the appropriate selection of models according to features of pathogenesis in human RA, we recommend consulting the review by Vincent et al on behalf of the Arthritis Research UK Animal Models Working Group.63 That article also provides useful information on features of osteoarthritis and briefly touches on important 1 Departments of Research Pathology, Genentech Inc, South San Francisco, CA, USA 2 Biomedical Imaging, Genentech Inc, South San Francisco, CA, USA 3 Translational Immunology, Genentech Inc, South San Francisco, CA, USA 4 Department of Veterinary Biosciences and the Comparative Pathology and Mouse Phenotyping Shared Resource, The Ohio State University, Columbus, OH, USA
Corresponding Author: P. Caplazi, Department of Research Pathology, Genentech Inc, 1 DNA Way, South San Francisco, CA 94080, USA. Email: [email protected]
Downloaded from vet.sagepub.com at CARLETON UNIV on June 13, 2015
2
Veterinary Pathology therapeutic agents selectively influence some lesion features. For the assessment of bone lesions, sophisticated quantitative methods based on histomorphometry are possible47 but may require more advanced processing techniques, such as embedding of undemineralized tissue in methylmethacrylate. While histology remains the most widely used method for assessment of lesions in mouse models of RA, sophisticated in vivo or ex vivo imaging modalities exist that can complement or altogether replace histologic evaluation.2,29,32,62,71
Systemically Induced Mouse Models of RA
Figure 1. Key concepts shared between human rheumatoid arthritis (RA) and mouse models of RA. j Autoreactive T cells providing help to B cells or releasing a variety of proinflammatory cytokines are an important component in the initiation and maintenance of lesions in RA as well as mouse models.9,48 An intact adaptive immune system is not necessary for some models, however. Notably, T cells are not required for the induction of collagen antibody–induced arthritis.37 The antigen-experienced T-effector/memory cell phenotype is prominent among T cells in the RA synovium.8 k Antibodies to citrullinated antigens (anticitrullinated peptide antibodies [ACPAs]) that are implicated in the pathogenesis of RA have been shown to be important in mouse models, including collagen-induced arthritis.24,28 More broadly, antibody- and complement-mediated effects are important drivers in RA as well as mouse models.33 l Neutrophils are abundant in murine autoimmune arthritis and contribute to the pathogenesis through the release of cytotoxic products and immunoregulatory mediators. In addition, neutrophils may promote autoimmunity by formation of neutrophil extracellular chromatin traps (NETs) and the associated promotion of ACPA.11,70 m Cartilage and bone injury is driven by the formation of an inflammatory pannus that classically invades the joint from the capsular angle. In mouse models, however, substantial amounts of pannus also extend from the bone marrow cavity subjacent to the articular cartilage (Figs. 3, 9, 13). n Macrophages (MF) infiltrating the synovium are central in the pathogenesis of RA, serving as an important source of key proinflammatory cytokines such as interleukin 1b (IL-1b) and tumor necrosis factor a (TNF-a). Furthermore, macrophage-derived cytokines and their effects on synovial fibroblasts are essential for the differentiation of osteoclasts and osteolysis. Bones are shown in gray, articular cartilage in light blue, and synovium and adjacent soft tissues in light pink.
considerations for animal welfare and statistical powering of groups, areas that are not covered here. General practical considerations in the use of various rodent joint disease models are reviewed by Bolon et al.3 Finally, for a general scoring scheme of histologic lesions, we refer to Caplazi and Diehl.4 Besides this general scoring scheme, other, more narrowly defined morphologic scoring systems may be suited in situations where
Based on the method of induction, systemically induced models can be divided into 3 groups: those elicited by active immunization, those elicited by passive immunization, and those elicited by administration of irritant chemicals resulting in chronic inflammation. Collagen-induced arthritis (CIA) is the archetypical model of RA induced by active immunization. First developed in rats,59 the model was later adapted to the mouse16 and is now commonly used in both species, with the choice of rodent dependent on the scientific questions and types of therapies to be tested. Rats or mice develop an acute to subacute monophasic erosive polyarthritis after immunization with collagen II (but not other collagens) from many species. Type II collagen is the major constituent collagen form of articular cartilage, and immunoreactivity to type II collagen can be identified in some RA patients.7,56 Many important features of human RA are recapitulated in mouse CIA, including the presence of rheumatoid factor or anticitrullinated peptide antibody.51 At least in the mouse, a Th17-driven response is essential for the production of lesions.41,50 The effects of interferon g are divergent51 with inhibitory as well as stimulatory effects on development of disease. Overall, however, interferon g attenuates CIA.22,51 Practical aspects of working with the mouse CIA model have been reviewed elsewhere66 and so are only briefly summarized here. A typical induction protocol might use heterologous (eg, bovine) collagen II for immunization of mice from a strain with permissive major histocompatibility complex (MHC haplotype q, H-2q; eg, DBA/1). For induction, a subcutaneous injection of type II collagen suspended in complete Freund adjuvant on day 0 is followed by a booster of type II collagen in incomplete Freund adjuvant on day 21. Effects of experimental treatment can then be tested by applying a prophylactic design (treatment initiated before the onset of clinical signs at around day 20 to 30) or a therapeutic design (treatment initiated after the onset of clinical signs on animals selected for presence of clinical signs and then randomized into treatment groups). Effects can be studied longitudinally in vivo using noninvasive assessments, such as clinical arthritis score, paw volume, caliper measurements of ankle thickness, or imaging modalities (eg, micro–computed tomography). Parameters determined in vivo can be complemented by terminal assessments, which include imaging (micro–computed tomography, benchtop radiography, magnetic resonance imaging), histology, or biochemical
Downloaded from vet.sagepub.com at CARLETON UNIV on June 13, 2015
Caplazi et al
3
Figures 2–7. Collagen-induced arthritis (CIA), day 63 postinduction, tarsal region, sagittal section, mouse (DBA/1J). Hematoxylin and eosin. Figure 2. In CIA, complex joints are usually affected globally, with inflammatory cells and fibroplasia spanning several nearby individual joints leading to locally extensive expansion of the affected soft tissues. Synovitis of tendons and tendon sheaths is common and may precede arthritis proper in some instances. Ta, talus; Tar, central tarsal bone; Te, common digital flexor tendon; Ti, distal tibial condyle. The dashed rectangle denotes the area shown enlarged in Figure 3. Figure 3. Pannus (black arrows) extends from the inflamed periarticular soft tissue into the subcartilaginous bone and onto the proximal cartilage surface of the central tarsal bone. In contrast to human RA, pannus-like tissue in mice also erodes from the bone marrow cavity into the subcartilaginous bone. This process of subcartilaginous osteolysis is facilitated by the close proximity (
