To examine the role of dipeptidyl peptidase I (DPPI), a widely expressed lysosomal cysteine protease, in the development of collagen-induced arthritis (CIA) in mice.
To examine the role of dipeptidyl peptidase I (DPPI), a widely expressed lysosomal cysteine protease, in the development of collagen-induced arthritis (CIA) in mice.
Wild-type (WT) and DPPI-deficient (DPPI−/−) mice backcrossed to DBA/1J mice for 10 generations were immunized with bovine type II collagen (CII), and disease susceptibility and severity were assessed over time. Collagen-specific B cell and T cell responses and the production of proinflammatory cytokines (tumor necrosis factor α, interleukin-1, and interleukin-6) were measured. In addition, adoptive transfer of splenocytes from WT, CII-sensitized mice was performed to evaluate the specific role of DPPI−/− T lymphocytes.
The majority of DPPI−/− mice were resistant to CIA induction, although clinical disease (i.e., evidence of inflammation and bone erosions) did develop in a small number of DPPI−/− mice. The protection against disease development was not attributable to a defect in the B and T cell response to collagen immunization, because both anticollagen antibody production and T cell proliferation in response to CII were normal. Release of the proinflammatory cytokines was largely unaffected in CII-stimulated DPPI−/− splenocytes. In addition, when cells isolated from the joints of DPPI−/− mice were stimulated in vitro, they had no intrinsic defect in their ability to release inflammatory cytokines. Last, adoptive transfer of splenocytes from WT, CII-immunized mice into naive WT and DPPI−/− mice led to development of arthritis in WT mice but not in DPPI−/− mice.
These results indicate that DPPI regulates a critical step in the development of CIA that is independent of T cell and B cell functions.
It has long been thought that the major function of proteases in the inflammatory milieu is to cause tissue damage by degrading extracellular matrix proteins. However, it has become clear that extracellular proteases play an important regulatory role in controlling local inflammatory processes that extends beyond their established destructive capability (1, 2). Dipeptidyl peptidase I (DPPI; also known as cathepsin C) is a widely expressed lysosomal cysteine protease belonging to the papain gene family. DPPI is critical for the activation of multiple serine proteases, including cytotoxic lymphocyte–associated granzymes (3), neutrophil-derived elastase, cathepsin G, proteinase 3 (4), and mast cell chymase (5).
We previously demonstrated that DPPI−/− mice were protected against acute arthritis induced by passive transfer of monoclonal antibodies against type II collagen (CII) (4). In this model of acute arthritis, mice were injected intravenously with 4 mg of a cocktail of 4 separate monoclonal antibodies specific for the main arthrogenic determinant of CII, followed by the injection of lipopolysaccharide, a strong inducer of proinflammatory cytokines (6). Although this model provided insights into the role of DPPI in acute inflammation, the disease is transient and is distinct histologically from the chronic disease seen in human rheumatoid arthritis (RA). In addition, using the passive antibody transfer model, we were unable to evaluate the role of DPPI in cellular and humoral immunity to collagen.
In the current study, we extend our observations to a model of more chronic inflammation. Collagen-induced arthritis (CIA) is a model of inflammatory arthropathy that shares many features with human RA. Here, we established that despite having normal humoral and cellular immunity to bovine CII, the majority of DPPI−/− mice were highly resistant to the development of CIA. In addition, adoptive transfer of CII-sensitized splenocytes from immunized WT animals induced arthritis in wild-type (WT) mice but not in DPPI−/− mice. Taken together, these results suggest that DPPI regulates a critical step in the development of arthritis that is independent of T cell and B cell functions.
DPPI−/− mice were generated in a 129/SvJ strain, as previously described (3), and backcrossed to DBA/1J mice (The Jackson Laboratory, Bar Harbor, ME) for 10 generations, using the DPPI mutant allele as a marker for selection at each generation. WT and DPPI−/− mice were maintained in pathogen-free conditions at the Washington University Specialized Research Facility. All animal experiments were performed according to protocols approved by the Division of Comparative Medicine at Washington University.
CIA was induced in 8–12–week-old mice by intradermal injection of 200 μg of bovine CII (Elastin Products, Owensville, MO) in Freund's complete adjuvant (CFA) supplemented with 1 mg/ml of Mycobacterium tuberculosis on day 0, followed by a booster injection of 200 μg of bovine CII in Freund's incomplete adjuvant (IFA) on day 21. Between day 21 and day 42, mice were examined 3 times weekly, and the severity of arthritis in each paw was scored as previously described (4), using a 3-point scale for each paw (maximum possible score per mouse = 12). Paw thickness was measured across the ankles at the peak of arthritis (day 42), using calipers.
At least 4 mice per genotype were killed on day 44, and all 4 paws of each mouse were excised, processed, and examined as previously described (4). Joint inflammation (the degree of synovitis and joint exudates) and bone erosions were scored on a scale of 0 (normal) to 3 (severe).
Ninety-six–well plates were coated with 5 μg/ml of CII overnight at 4°C and blocked with 1% bovine serum albumin. After thorough washing, 100-μl aliquots of each serum diluted in phosphate buffered saline (PBS) were added, and the plates were incubated at 4°C overnight. After thorough washing, 100 μl of horseradish peroxidase–conjugated goat anti-mouse IgG1 or IgG2a antibodies (1:8,000 dilution; Southern Biotechnology, Birmingham, AL) was added, and color development was obtained using 1-Step Turbo TMB-ELISA (enzyme-linked immunosorbent assay) (Pierce, Rockford, IL). Each serum sample was analyzed in triplicate, and the results were expressed as the fold increase in levels of anticollagen antibodies between day 0 (naive) and day 28 (immunized).
Cells obtained from the paws of the mice were prepared as previously described (7). Briefly, WT and DPPI−/− mice (n = 4 per genotype) were killed on day 44, and their paws, including bones, were minced and digested for 4 hours in 500 units/ml of Dispase (Fisher Scientific, Pittsburgh, PA) and 1 mg/ml of type I collagenase (Sigma, St. Louis, MO) at 37°C. The dispersed cells were seeded in complete media (106 cells/ml). After 24 hours, the supernatant was collected, and spontaneous cytokine release was measured by standard ELISA.
On day 44, draining inguinal lymph node cells from immunized mice were pooled and plated at 2.5 × 106 cells/well (tested in triplicates per point) in 96-well plates in the presence of 50 μg/ml of heat-denatured CII. Lymph node cells (5 × 105 cells/well) were also stimulated with concanavalin A (Con A) (5 μg/ml; Calbiochem, San Diego, CA) as a nonspecific stimulus. After 48 hours, the plates were pulsed with 3H-thymidine for an additional 24 hours and harvested for the measurement of 3H-thymidine incorporation.
Cytokine production was determined in cells harvested on day 42. Single-cell suspensions (107 cells/well in 24-well plates) were cultured in triplicate with 50 μg/ml of CII or 5 μg/ml of Con A for 48 hours, and cytokines released into the supernatant were measured using a specific ELISA (R&D Systems, Minneapolis, MN), according to the manufacturer's protocol. The lower limit of sensitivity for these ELISAs was ∼10 pg/ml for the cytokines tested (tumor necrosis factor α [TNFα], interleukin-1β [IL-1β], and IL-6).
WT mice were immunized on day 0 with CII in CFA, as described above, and on day 21 were given intraperitoneal booster injections of 100 μg of CII in PBS. Five weeks after the initial injection, splenocytes from mice with manifestations of arthritis were harvested, pooled, and transferred to naive WT or DPPI−/− mice (2 × 107 splenocytes per mouse, injected intraperitoneally); transfer was followed by 200-μg peritoneal injections of CII in PBS. Arthritis was scored as described above. Anti-CII antibody levels were measured 2 weeks after the adoptive transfer.
Wilcoxon's 2-tailed rank sum test for unpaired variables was used to compare differences between groups. P values less than 0.05 were considered significant.
CIA in DPPI−/−mice. WT and DPPI−/− mice backcrossed to a DBA/1J strain for 10 generations were immunized with 200 μg of bovine CII in CFA on day 0, followed by 200 μg of bovine CII in IFA on day 21. In WT mice, clinical arthritis was usually detected 5–7 days after the second immunization. The cumulative incidence of disease was 93% in WT mice (n = 14) compared with 26% in DPPI−/− mice (n = 14) (Figure 1A). In a few DPPI−/− mice, mild redness and swelling developed following the booster injection, but the inflammation resolved without further progression. Some DPPI−/− mice had delayed onset of clinically evident arthritis manifested by redness and swelling in 1 or 2 digits (Figure 1D, bottom left panel). On day 44, the mean ± SD arthritis index in the DPPI−/− group was 0.43 ± 0.85, compared with 7.37 ± 3.5 in WT mice (P < 0.001) (Figure 1B). In addition, DPPI−/− mice had significantly reduced paw thickness (mean ± SD 3.08 ± 0.35 mm versus 3.71 ± 0.56 mm in WT mice; P < 0.01), as reflected by the measurements obtained on day 42 (Figure 1C).
Histologic analysis confirmed that DPPI−/− mice were more protected against severe inflammation compared with WT mice. Evaluation of diseased joints from WT mice revealed extensive inflammatory infiltrates accompanied by severe bone and joint erosions (Figure 1D, top right panel). In contrast, most of the DPPI−/− mice exhibited normal joint morphology, with no inflammatory infiltrates (Figure 1D, middle right panel). DPPI−/− mice in which clinical disease developed exhibited synovitis with focal bone erosions; however, there was some preservation of the joint architecture (Figure 1D, bottom right panel). Overall, 87% of the joints from WT mice showed erosive changes (n = 31 joints; mean ± SD severity score 2.1 ± 1.2) compared with 11% of the joints from DPPI−/− mice (n = 27 joints; mean ± SD severity score 0.2 ± 0.7 [P < 0.001 versus WT]). In addition, 90% of the joints from WT mice showed inflammation (mean ± SD severity score 2.4 ± 1.0) compared with 20% of the joints from DPPI−/− mice (mean ± SD severity score 0.4 ± 0.9; P < 0.001). Taken together, these results suggest that the majority of DPPI−/− mice are resistant to the development of CIA; however, once disease is initiated in a joint, DPPI−/− mice are not entirely protected against synovial inflammation and bony erosions.
Humoral immunity to collagen in DPPI−/−mice. Humoral immunity to collagen, specifically IgG2a production, has been shown to play an essential role in the initiation of disease in CIA. In the model of passive transfer of anti-CII antibody, the response of DPPI−/− mice to CII immunization could not be evaluated. Due to the lysosomal localization of DPPI, it has been proposed that DPPI might play a role in antigen processing that is very similar to the action of cathepsin S and cathepsin L (8, 9). To evaluate whether DPPI−/− mice responded normally to collagen immunization, serum levels of IgG1 and IgG2a anticollagen antibodies were measured 7 days after the booster injection was administered. The levels of IgG1 and IgG2a anticollagen antibodies were equivalent in WT and DPPI−/− mice (Figure 1E). These results suggest that DPPI does not participate in the presentation of CII, and that the mechanism that confers resistance to the development of CIA in DPPI−/− mice is not attributable to diminished humoral immunity.
Cellular immunity to collagen in DPPI−/−mice. To assess cellular immunity to collagen in DPPI−/− mice, in vitro T cell proliferation was measured. WT and DPPI−/− mice were killed on day 44, and their draining inguinal lymph nodes were harvested and pooled. T cell proliferation was measured in lymph node cells cultured in the presence of 50 μg/ml of collagen or 5 μg/ml of Con A. The values for CII- or Con A–induced proliferation were similar for both WT and DPPI−/− cells (Figure 2A). These results suggest that cellular immunity to collagen was not altered in the absence of DPPI.
Cytokine production in CII-stimulated spleen cells. It is well recognized that the production of proinflammatory cytokines is critical for the development and perpetuation of arthritis. To evaluate whether DPPI−/− splenocytes have a defect in their ability to produce cytokines, spleen cells from WT and DPPI−/− mice were harvested on day 44 after immunization and cultured in the presence of 50 μg/ml of CII or 5 μg/ml of Con A. The levels of TNFα, IL-1β, and IL-6 released from splenocytes were measured after 48 hours in culture. WT and DPPI−/− mouse splenocytes cultured in the presence of CII or Con A produced equivalent levels of TNFα (Figure 2B). IL-1β production by sensitized splenocytes in response to CII stimulation was minimal in both genotypes. Although we detected a 50% reduction in IL-1β production by DPPI−/− mouse splenocytes stimulated with Con A (Figure 2C), this decrease was not reproducible in subsequent experiments (data not shown). In contrast, IL-6 production was increased by 50% and 35% in DPPI−/− mouse splenocytes cultured in the presence of CII and Con A, respectively (Figure 2D). Taken together, these results indicate that DPPI−/− mouse splenocytes have no significant intrinsic defect in their ability to produce and release proinflammatory cytokines.
Cytokine production by joint mononuclear cells. We next isolated mononuclear cells from the paws of WT and DPPI−/− mice in order to determine the spontaneous release of TNFα, IL-1β, and IL-6 after 24 hours of in vitro culture. Similar to the results of the histologic analysis, we recovered 3-fold more mononuclear cells from the paws of WT mice than from the paws of DPPI−/− mice (n = 16 paws per genotype; P < 0.01) (Figure 2E). However, the spontaneous release of TNFα and IL-6 from WT and DPPI−/− mouse mononuclear cells was not different on a single-cell basis (Figures 2F and H). In these in vitro cell cultures, IL-1β was barely detectable (Figure 2G). The normal production of proinflammatory cytokines by these mononuclear cells is consistent with the inflammation observed in DPPI−/− mice with clinically manifest arthropathy.
Adoptive transfer of arthritis using splenocytes from CII-sensitized WT mice. Previous studies have shown that the transfer of splenocytes from arthritic mice into naive recipients induces arthritis. Depletion of CD4+ T cells inhibits the transfer of arthritis, while depletion of CD8+ T cells enhances the onset of arthritis (10). To directly evaluate whether the resistance of DPPI−/− mice to CIA depends on subtle defects in T lymphocyte functions, we performed adoptive transfer of splenocytes from CII-immunized, arthritic WT mice into naive WT and DPPI−/− mice, followed by immunization with CII. In naive WT mice that received 107 splenocytes from CII-immunized WT DBA/1J mice, moderate arthritis developed after 6–8 days (mean ± SD arthritis severity score 5 ± 2) (Table 1). In contrast, DPPI−/− mice were completely resistant to arthritis development after splenocyte transfer, although the increases from baseline in anti-CII antibody titers were equivalent in both genotypes (Table 1). These results indicate that resistance to CIA in DPPI−/− mice was independent of T cell and B cell functions.
|Mean ± SD arthritis severity score†||5 ± 2||NA|
|Day at onset of arthritis||6–8||NA|
|Mean ± SD anti-CII antibody titer‡||34 ± 13||59 ± 25|
In this study we have established that the absence of DPPI protects mice against the development of CIA, even when humoral and T cell immunity are intact. These results also suggest an essential role for innate immunity in the development of CIA. Consistent with this notion, a recent report and results of our own studies suggest that neutrophil-derived serine proteases CG and NE (both of which are activated by DPPI) play a nonredundant role in the generation of CXC chemokines in response to inflammatory stimuli in vitro and in vivo (11, 12). CXC chemokines are small chemotactic cytokines that play an important role in recruiting leukocytes to sites of inflammation. By modulating the production of CXC chemokines, DPPI and the serine proteases it activates may profoundly influence the inflammatory response to and development of CIA.