Mast Cells: A Cellular Link Between Autoantibodies and Inflammatory Arthritis

See allHide authors and affiliations

Science  06 Sep 2002:
Vol. 297, Issue 5587, pp. 1689-1692
DOI: 10.1126/science.1073176


Previous studies have revealed that autoantibodies, complement components, and Fc receptors each participate in the pathogenesis of erosive arthritis in K/BxN mice. However, it is not known which cellular populations are responsive to these inflammatory signals. We find that two strains of mice deficient in mast cells, W/Wv and Sl/Sld, were resistant to development of joint inflammation and that susceptibility was restored in the W/Wv strain by mast cell engraftment. Thus, mast cells may function as a cellular link between autoantibodies, soluble mediators, and other effector populations in inflammatory arthritis.

The pathogenic mechanisms at play in inflammatory arthritis, such as rheumatoid arthritis, remain poorly understood both systemically and in the microenvironment of the diarthrodial joint. A large number of soluble inflammatory mediators and cellular effector populations have been implicated in arthritis; however, the early clinical events remain elusive. Recent studies using the serum of an engineered mouse model, K/BxN, have revealed that autoantibodies directed against a ubiquitously expressed antigen can selectively provoke inflammatory, hyperplastic, and erosive synovitis (1–3). It is known that members of the complement network (the alternative pathway and C5a), Fc receptors (FcγRIII), and cytokines [interleukin 1 (IL-1), tumor necrosis factor α (TNF-α)], as well as neutrophils, have essential roles (4–6). Yet, the pathways leading to the development of synovial pathology by means of these autoantibodies remain to be explained. Given their resident nature in the synovium and their functional capabilities, we hypothesized that mast cells might provide a critical cellular link between soluble factors (autoantibodies, complement, cytokines) and the synovial eruption.

We utilized two strains of mice deficient in mast cells: WCB6F1 KitlSl / KitlSl-d (Sl/Sld ) and WBB6F1-KitWKitW-v (W/Wv) (7, 8) to examine the functional role for mast cells in the effector phase of inflammatory arthritis. The Sl/Sld strain lacks the transmembrane form of stem cell factor (SCF), which results in a deficiency of tissue-resident mast cells (9, 10). After K/BxN serum transfer, control littermates exhibited typical clinical arthritis with synovial hyperplasia, pannus formation, and inflammatory infiltrates (Fig. 1, A, B, and F; and fig. S4A). In contrast, Sl/Sld mice demonstrated almost no evidence of clinical joint inflammation (Fig. 1, A and B, and fig. S4A) or histopathologic abnormalities (Fig. 1E). Correspondingly, mast cells, many displaying features of degranulation, were identified in inflamed control synovial tissues but not in joint tissues from Sl/Sld mice (11).

Figure 1

Attenuation of arthritis in mast cell–deficient Sl/Sld and W/Wv mice. Arthritogenic K/BxN serum was transferred into 4- to 6-week-old male Sl/Sld or W/Wv mice and control wild-type littermates as described (3). Clinical index and change in ankle thickness were recorded. (A and B) Wild type (black squares; mean clinical index = 9.5), Sl/Sld (white diamonds; mean clinical index = 0). (C and D) Wild type (black squares; mean clinical index = 7), W/Wv (white diamonds; mean clinical index = 0). Error bars, SEM. Clinical index (per paw): 0 = no evidence of inflammation; 1 = subtle inflammation (metatarsal phalanges joints, individual phalanx, or localized edema); 2 = easily identified swelling but localized to either dorsal or ventral surface of paw; and 3 = swelling on all aspects of paw. Maximum score = 12. (E andF) Ankle sections from Sl/Sld (E) and wild-type (F) mice 11 days after K/BxN serum transfer. Sl/Sld mice (E) demonstrate thin synovial lining with loose connective tissue in sublining (S), smooth cartilage (Ca), and intact bone (Bn). In contrast, wild-type mice (F) demonstrate synovial lining hyperplasia (S) with inflammatory cellular infiltrates and synovial erosion through cartilage (Ca) into bone (Bn) (arrow). Magnification, ×100. Results are representative of two independent experiments, n = 3 or 4 mice per group.

The W/Wv mouse strain also lacks significant numbers of tissue mast cells because of mutations in the SCF receptor, c-kit (12, 13). Similar to findings in the Sl/Sld strain, W/Wv mice displayed little or no clinical or histologic evidence of arthritis compared with control littermate mice when exposed to arthritogenic serum from K/BxN mice (Fig. 1, C and D; fig. S1, A and B; and fig. S4B).

To confirm that the resistance to arthritis after transfer of autoantibody-containing serum was due to the absence of mast cells in the mutant mice, and not to defects in other cellular systems, we performed complementation analysis by assessing the ability of selective mast cell engraftment to confer arthritis susceptibility on these mice. Transfer of either whole bone marrow capable of forming mast cells or bone marrow mast cells (BMMCs) cultured in vitro allows selective engraftment of functional tissue-resident mast cells (7, 8, 14). Thus, we cultured wild-type bone marrow cells in IL-3 and SCF for 4 weeks (15) and confirmed expression of FcɛR and c-Kit and other histochemical properties of mast cells [fig. S2 and (11)]. These donor BMMCs from wild-type littermates or C57BL/6J mice were then transferred into W/Wv recipient mice, and engraftment was confirmed 10 weeks later (compare Fig. 2E and 2F). Analysis of a separate series of engrafted W/Wvmice after 22 weeks revealed no correction of anemia, which suggests that this engraftment protocol did not extend to all hematopoietic lineages [fig. S3 and (11)]. The K/BxN serum induced arthritis in these BMMC-engrafted W/Wv mice, whereas most sham-engrafted W/Wv mice remained free of signs of arthritis (Fig. 2, A and B). Similar to wild-type animals, BMMC-engrafted W/Wv mice had degranulated mast cells after K/BxN serum transfer. It is noteworthy that BMMC-engrafted W/Wv mice revealed synovial hyperplasia, cartilage and bone destruction, and inflammatory infiltrates, with neutrophil predominance similar to that typical of control littermates (Fig. 2, C and D; and fig. S4C). In contrast, most sham-engrafted W/Wv mice displayed normal-appearing ankle tissue with little evidence of synovitis and no discernible mast cells.

Figure 2

Restoration of arthritis in mast cell–engrafted W/Wv mice. Selective mast cell engraftment into 20-week-old W/Wv mice was performed by i.v. injection of 1 × 107 cultured BMMCs from wild-type WBB6 F1 littermates or C57BL/6J donors. BMMCs were derived in 4-week cultures as described (15), with the cytokine concentration modified as follows: 10 ng/ml recombinant IL-3 (rIL-3) and 12.5 ng/ml SCF. DMEM medium was injected as a sham treatment. After allowing 10 weeks for engraftment, arthritogenic K/BxN serum was transferred into sham-injected or BMMC-engrafted W/Wv and control wild-type littermates. Clinical index and change in ankle thickness were recorded. (A and B) Wild type (white diamonds; mean clinical index = 8.4), sham-engrafted W/Wv (white squares; mean clinical index = 0.8), BMMC-engrafted W/Wv (black triangles; mean clinical index = 7). Error bars, SEM. Note: One sham-injected mouse in each experiment developed arthritis. (C) Ankle sections from engrafted W/Wv mice, stained with toluidine blue, reveal synovial (S) hyperplasia and destruction of cartilage (Ca) overlying bone (Bn) 10 days after K/BxN serum transfer. Magnification, ×100. (D) High-power (×500) view of toluidine blue–stained engrafted synovium demonstrating engrafted mast cells (arrowheads). Separately, mast cell bone marrow engraftment is documented by chloroacetate esterase stain (arrowheads) in BMMC-engrafted (F) but not in sham-injected (E) W/Wv mice. Magnification, ×500. Results are representative of two independent experiments. n = 4 to 6 mice per group.

Implicit in the potential function of mast cells as a cellular link between soluble components and subsequent arthritogenic events is their activation and the rapid release of potent granule constituents after serum transfer. Therefore, we used a histologic approach to reveal mast cell degranulation and to delineate its time course of activation. Intact mast cells were easily identified in control mice (Fig. 3, A, C, and D), whereas significant mast cell degranulation was noted as early as 1 hour and even more strikingly at 2 and 24 hours after intraperitoneal transfer of K/BxN serum (Fig. 3A). This degranulation preceded the onset of any clinical evidence of inflammation in these animals. In contrast to synovial tissue, no enhanced degranulation was noted in mast cells at other anatomic locations (Fig. 3B). Because degranulation is the clearest histologic hallmark of mast cell activation, the presence of degranulated mast cells in joint tissue, and not other tissues, before overt clinical or histologic inflammation supports a proximal, synovium-specific role for mast cells in the effector phase of inflammatory arthritis. Moreover, because mast cells continue to demonstrate a degranulating phenotype during more chronic phases of arthritis (Fig. 3, E to H), it is also likely they play an ongoing role in the arthritic process.

Figure 3

Histologic analysis of synovial mast cells in arthritis. (A and B) For kinetic analyses of mast cell degranulation, 4-week-old C57BL/6J male mice were injected intraperitoneally with 200 μl of either normal mouse serum (Ctl) or K/BxN serum, and tissues were harvested at the time points indicated. Consecutive mast cells in ankle sections (A) or gastric mucosa (B) were visually assessed for intact versus degranulating phenotype in a blinded fashion. Shown are results from individual mice pooled from three separate experiments. Horizontal bars represent mean percent degranulation. Ankle tissues from normal (C andD) and arthritic C57BL/6J (E and F) and K/BxN (G and H) mice were fixed and stained with hematoxylin and eosin [(C), (E), and (G)] or toluidine blue [(D), (F), and (H)]. (C) Normal mice demonstrate thin synovial lining and loose connective tissue sublining (S), smooth cartilage (Ca), and intact bone (Bn). Magnification, ×100. (D) View (×400) of tissue outlined in (C) demonstrating intact mast cells, with compact metachromatically staining granules (inset), which are present sparsely throughout the synovium. Ankle sections obtained 12 days after transfer of arthritogenic K/BxN serum (E) and in 8-week-old transgenic K/BxN mice (G) demonstrate synovial hyperplasia and pannus formation (S/P) with evidence of invasion and destruction of cartilage (Ca) and bone (Bn), as well as inflammatory cellular infiltrates. Magnification, ×100. (F and H) Higher magnification (×400) views of toluidine blue–stained tissue outlined in (E) and (G), which demonstrate degranulating and intact mast cells. Insets show high magnification (×1000) of degranulating mast cells.

Initial results from the mast cell–deficient mouse strains left open the possibility that the SCF/c-kit signaling pathway played some role in the induction of arthritis other than via mast cells. However, the restoration of arthritis susceptibility by mast cell engraftment defined mast cells as the element that prevents disease development in Sl/Sld and W/Wv mice. The striking requirement for mast cells, coupled with evidence for their rapid degranulation within the first hours after serum transfer, leads us to suggest that mast cells may provide the cellular target of autoantibodies, the complement network, and Fc receptors in the subsequent development of inflammatory arthritis. Because C5a and FcγR ligation are potent activators of mast cell function in vitro and in vivo (16–22), it is likely that synovial mast cells are activated by articular autoantibody immune complexes suggestive of an immune complex hypersensitivity (Arthus) reaction in the synovium (23, 24).

Mast cells themselves produce a series of effector molecules that mediate permeability, inflammation, chemotaxis, and tissue destruction. They are the only cells that contain preformed TNF-α in granules and they also display an ability to rapidly produce large amounts of both TNF-α and IL-1 (25,26), cytokines that play a critical role in K/BxN arthritis (6), as well as in human rheumatoid arthritis. Mast cell granules also contain an abundance of proteases capable of activating matrix metalloproteinases (27) and mMCP-6, a potent indirect neutrophil chemoattractant (28). These cells also produce large quantities of other inflammatory molecules including histamine, eicosanoids, fibroblast growth factor, and angiogenesis factors (VEGF), which may contribute further to the arthritic process.

Historically, mast cells have been implicated in two contrasting types of immune responses. First, they can be activated by immunoglobulin IgE receptors to mediate immediate hypersensitivity reactions associated with allergic phenomena. Second, their acute activation by microbial products, as in bacterial peritonitis models, underscores their role in infection (29, 30). Here, we present evidence of a direct role for mast cells in the pathogenesis of inflammatory arthritis. Our findings in the K/BxN model of destructive arthritis point to a likely role for mast cells in human arthritis associated with immune complex formation, namely, cryoglobulin-associated synovitis in hepatitis C infection, postinfectious arthritis, and perhaps others. Moreover, histologic analyses of synovial sections from humans have documented the presence of mast cells in abundance (22) and immune complexes, complement fragments, and SCF are present in synovial fluid and tissue in rheumatoid arthritis (31–33). Together, these findings illustrate that mast cells can contribute to the pathogenic mechanisms in the synovium that result in erosive arthritis.

Supporting Online Material

Materials and Methods

Figs. S1 to S4

  • * To whom correspondence should be addressed. E-mail: mbrenner{at}


View Abstract

Navigate This Article