SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

Rheumatoid arthritis (RA) is an autoimmune joint disease associated with chronic inflammation of the synovium that causes profound damage of joints. Inflammation results in part from the influx of immune cells secreting inflammatory cytokines and the reduction in the number of Treg cells. We undertook this study to assess the effect of furin, a proteinase implicated in the proteolytic activity of various precursor proteins and involved in the regulation of both proteinase maturation and immune cells, in an experimental model of RA.

Methods

The effect of furin and its inhibitor α1-PDX was tested in mice with collagen-induced arthritis (CIA). Joints were processed for histology and protein expression. Levels of cytokines were measured in joint tissue, and Treg cell numbers were measured in spleens.

Results

Furin expression and activity were high in the synovial pannus in RA patients and mice with CIA. Systemic administration of furin prevented increases in the arthritis score, joint destruction, and bone loss, in contrast to systemic administration of the furin inhibitor α1-PDX, which enhanced these parameters. By preventing the development of synovial pannus, furin reduced the expression of metalloproteinases in the joints. In contrast, α1-PDX enhanced synovial proliferation and the expression and activity of matrix metalloproteinases. Furthermore, furin reversed the local Th1/Th2 balance and restored the number of Treg cells in the spleen, indicating mediation by immune cells.

Conclusion

These findings show the protective role of exogenous furin against RA, mediated by an immune response. The data suggest the potential therapeutic use of furin or its derivatives in autoimmune diseases including RA.

In rheumatoid arthritis (RA), joint destruction is the consequence of an imbalance between the catabolic and anabolic pathways. Associated with chronic inflammation of the synovium, these processes involve cytokines, growth factors, and matrix-degrading enzymes that are produced by effector cells. Similarly, the disequilibrium of the Th1/Th2 balance affects the production of cytokines that orchestrate the immune response (1, 2). Indeed, the shift toward Th1 promotes the production of proinflammatory cytokines (interleukin-1 [IL-1], tumor necrosis factor α [TNFα]) and reduces the production of antiinflammatory cytokines (IL-10, IL-4). Although recent therapies designed to inhibit inflammatory cytokines do provide significant protection against structural damage in RA, the persistence of synovitis and the high rate of nonresponse to biologic therapies still result in joint destruction (3), which means that there is still a need to develop an alternative approach to controlling the cytokines secreted by effector cells.

Furin is a ubiquitous proprotein convertase involved in the proteolytic processing of a wide range of precursor proteins, including growth factors and their receptors, adhesion molecules, and various metalloproteinases (4). The cleavage of furin substrates, such as insulin-like growth factor 1 and its receptor and several proteinases, including stromelysin 3 (5), MT1-MMP (6), and ADAMTS (7, 8), is crucial for the mediation of their functions. Unprocessed forms of some of these molecules are biologically active, and in certain cases, such as endothelial lipase (9), matrix metalloproteinase 2 (MMP-2) proforms (10), and fibroblast growth factor 23 (11), their processed forms are inactive or even mediate the opposite biologic action via specific receptors (12). In addition to its implication in the activation of molecules involved in several diseases such as cancer and infections, furin has been shown to mediate the proteolytic activation of several proteins involved in cartilage remodeling (13). Indeed, it was previously proposed that furin could play a role in joint remodeling by its ability to generate active forms of MMPs (14) and to reduce cartilage degradation ex vivo (13). On the other hand, furin was also found to inhibit the activity of active MMP-2 (10), a well-established extracellular matrix (ECM)–degrading proteinase.

Recently, the loss of furin was found to reduce peripheral tolerance in mice in vivo (15). Therefore, in vivo, furin exerts different local effects in tissue, which may influence the outcome of the disease. Indeed, furin might affect the development of arthritis as a result of positive regulation of peripheral tolerance and, in contrast, by enhancing the intracellular maturation of proteolytic enzymes that may lead to degradation of the cartilage matrix. We undertook this study to evaluate the final in vivo effect of furin during inflammation and joint degradation in arthritis.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Immunohistochemical analysis of human samples.

To assess the expression of furin in humans, synovial tissue was harvested from 3 patients in whom RA was diagnosed on the basis of joint erosions, elevated erythrocyte sedimentation rate (ESR), and positivity for anti–cyclic citrullinated peptide (anti-CCP) antibodies. Control synovial tissue samples were obtained from 3 patients with posttraumatic knee pain who were undergoing arthroscopy. These patients had normal ESRs and were anti-CCP negative. The samples were collected in accordance with the French National Authority Legislation for the collection of human tissue, and the collection was approved by the Institutional Review Board (IRB) (approval no. 0000383). As required by French bioethics law and the local IRB, the need for informed consent was waived since these tissues were surgical waste of routine joint replacement surgery and since there was no private patient information being collected. The participants received an information note explaining the purpose of the study and were asked for their nonopposition to participation. Furin expression was investigated by immunohistochemistry as described below for mouse samples.

Induction of collagen-induced arthritis (CIA) and treatment.

In vivo experiments on mice complied with the guidelines for animal experimentation issued by our local Ethics Committee on Animal Care and Experimentation. Male DBA/1 mice, which are susceptible to developing CIA, were purchased from Janvier at ages 5–7 weeks. After 1 week in their housing, mice were immunized as described previously (16). The mice received a subcutaneous injection, at the base of the tail, of 100 μg of native bovine type II collagen (Chondrex) emulsified in Freund's complete adjuvant. A subcutaneous boost of 100 μg of type II collagen in Freund's incomplete adjuvant was given 21 days later, and mice were then injected intraperitoneally (IP) 3 times a week with 1 unit of furin (no. P8077L; New England Biolabs), 14 μg α1-PDX, a furin inhibitor (no. RP-070; Thermo Scientific ABR), or phosphate buffered saline as vehicle. Another group of mice was left untreated without immunization. Preliminary experiments showed no significant dose effect using 1 or 2 units of furin and 14 or 28 μg of α1-PDX (data not shown). Therefore, only 1 dose was used in the subsequent experiments, which were performed in quadruplicate with 5–10 mice per group.

Assessment of arthritis.

From 30 days following the injections to the end of the experiment, paws of mice with CIA were scored 3 times a week in a blinded manner. The clinical severity of the arthritis was scored as follows: 0 = normal; 1 = erythema; 2 = swelling; 3 = deformity or swelling of the entire paw; and 4 = ischemia for each limb (i.e., the toes, the tarsus, and the ankles). The joint scores were then summed to obtain the total arthritis score of the 4 limbs, yielding a score of 0–48 per animal. In each group, the mean of the total arthritis score was calculated 3 times a week to evaluate the severity of CIA. The animals were killed 46 days after immunization, when the mean arthritis score had reached a plateau.

Histologic score and immunohistochemical analysis.

After the mice were killed, their legs were dissected free of muscle and processed for histologic studies. One paw was collected in a solution of 4% paraformaldehyde (PFA) and then decalcified in 1% PFA/0.2M EDTA for 3 weeks. The solution was replaced 3 times a week. The samples were embedded in paraffin, and at least 4 serial sections were cut from each paw to ensure extensive evaluation of the arthritic joints using hematoxylin–eosin–saffron staining. Evaluation of each slice was done in a blinded manner using 4 parameters (0–3 scale for each parameter): synovial inflammation and thickness, synovial invasion into the joint, cartilage unevenness, and bone erosion. The total histologic score represented the sum of the 4 parameters.

Immunohistochemical analysis was performed on sections following their incubation at 90°C in citrate buffer (pH 6.5) and treatment with 1 mg/ml hyaluronidase for 30 minutes. Rabbit polyclonal and ABC kits were used according to the manufacturer's protocol (Vector). The primary antibodies used were sc-20801 at 1:500 dilution for antifurin (Santa Cruz Biotechnology) and ab7032, ab38898, and ab73879 (at 1:1,000 dilution for each) for anti–MMP-2, anti–MMP-9, and anti–MMP-14, respectively (all from Abcam). Sections were sequentially incubated with secondary biotinylated antibody, with the biotin–avidin amplification system, and, finally, with diaminobenzidine substrate. Sections were then counterstained with toluidine blue. Results were analyzed in 3 sections in each of 2 separate experiments.

Cytokine measurements.

At the time the mice were killed, 1 paw from each mouse was collected in order to measure the cytokine secretion in the whole joint. After dissection of the soft tissue, the whole joint was ground up in buffer and assayed for cytokine concentration. Proinflammatory cytokines (TNFα, IL-1, IL-17) and antiinflammatory cytokines (IL-4, IL-10) were measured in the whole paws using enzyme-linked immunosorbent assays (R&D Systems). Levels were expressed in pg/mg of protein.

Flow cytometric analysis.

Treg cell profiles were assessed in the spleen cells of mice from each treatment group and from the naive mice. Fourteen days after the second boost, the spleens were collected and cells were isolated for fluorescence-activated cell sorting analysis. Spleens were collected on ice and prepared for cell suspension in complete medium containing RPMI 1640 and 20% serum (Gibco) supplemented with 20 units/ml penicillin, 0.02 mg/ml streptomycin, and 2 mM L-glutamine (Sigma) by mechanical mashing through a 40-μm cell strainer (Falcon). Red blood cells were lysed using Gey's balanced salt solution and washed out with RPMI 1640 with 20% serum.

Splenocytes were labeled with PerCP–Cy5.5–conjugated anti-CD4 and fluorescein isothiocyanate–conjugated anti-CD25. Intracellular FoxP3 staining was performed with phycoerythrin-conjugated anti-mouse/rat FoxP3 according to the BD Biosciences protocol. Isotype-matched antibodies were used as controls, and IgG blocking was used to avoid nonspecific binding. Cells were analyzed using a FACSCalibur (BD PharMingen), and data were analyzed with CellQuest software (Becton Dickinson). Results are expressed as the percentage of gated events. At this time point (35 days after immunization), the mean arthritis score was 8.2 for mice treated with furin, 13.7 for mice treated with α1-PDX, and 11.8 for mice treated with vehicle.

Evaluation of bone parameters.

To assess the effects of furin and α1-PDX on systemic bone loss, we measured the whole-body bone mineral density (BMD; in gm/cm2) at baseline before immunization and at euthanasia. BMD was measured using a Lunar Piximus device, and the BMD change from baseline to euthanasia (ΔBMD) was calculated as (BMD at euthanasia—BMD at baseline)/BMD at baseline and expressed as a percentage.

Measurement of furin activity.

Proprotein convertase activity in the paws was assessed by evaluating its ability to digest the universal proprotein convertase substrate, the fluorogenic peptide pERTKR-MCA, as previously described (17). Briefly, tissue extracts were incubated with pERTKR-MCA (100 μmoles/liter) for various time periods in the presence of 25 mmoles/liter Tris, 25 mmoles/liter methylethanesulfonic acid, and 2.5 mmoles/liter CaCl2 (pH 7.4) at 37°C, and the fluorometric measurements were performed using a spectrofluorometer (FLUOstar Optima; BMG Lab Technologies). The general proprotein convertase inhibitor decanoyl Arg-Lys-Val-Arg chloromethylketone was obtained from Calbiochem, and recombinant furin was obtained from Sigma.

Real-time polymerase chain reaction (PCR).

Total RNA was subjected to complementary DNA (cDNA) synthesis using the SuperScript First-Strand cDNA Synthesis system (Invitrogen). The relative quantification of specific messenger RNA (mRNA) was performed by real-time PCR using the StepOnePlus Real-Time PCR System and Power SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer's instructions. Briefly, the mixture of the reaction consists of a 20-μl total volume of a 2-μl of cDNA, 2× QuantiTect SYBR Green PCR Master Mix, and 0.5 μM of the forward and reverse primers. Primers used for amplification were 5′-TGA-GCC-ATT-CGT-ATG-GCT-ACG-3′ (sense) and 5′-TGC-GCA-CCT-CTA-GCC-GTT-3′ (antisense). Forty cycles of PCR were performed at 94°C for 15 seconds and at 60°C for 1 minute. Transcription of hypoxanthine guanine phosphoribosyltransferase was evaluated in each sample as an endogenous control.

MMP arrays.

Protein antibody arrays were performed to assess the expression of MMPs in mouse paws (Protein antibody arrays Biotin Label-based Mouse Antibody Array I, AAM-BLM-1-4; RayBiotech) according to the manufacturer's instructions. Briefly, 50 μg of protein from each paw extract was labeled with biotin and loaded onto an antibody-coated membrane. Horseradish peroxidase–conjugated streptavidin–based detection was then performed with a FujiFilm LAS-3000 Intelligent dark box. The integrated optical density dot was assessed with ImageJ software (National Institutes of Health) after retrieving the nonspecific optical signal density. Each measurement is a ratio of the mean of the dot of interest to the mean of 6 membrane internal control dots.

Zymography.

The gelatinase activities of MMP-2 and MMP-9 were assayed in culture medium and in the whole paw by gelatin zymography assay. From each sample, 6 μg of protein normalized by the bicinchoninic acid method was separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis in a 10% (weight/volume) gel containing 1 mg/ml gelatin. The gel was washed for 30 minutes in 2.5% (w/v) Triton X-100, 40 mM Tris HCl (pH 7.8) buffer at room temperature. The zymography gel was then incubated overnight at 37°C in activation buffer containing 50 mM Tris HCl (pH 8), 10 mM CaCl2. Finally, staining in 0.5% Coomassie blue was performed for 1 hour, followed by 30 minutes of destaining in 20% (w/v) ethanol, 10% (w/v) glacial acetic acid. Pictures of the gels were taken with a FujiFilm LAS-3000 Intelligent dark box.

Statistical analysis.

Results are expressed as the mean ± SEM. Comparison tests were performed using analysis of variance for repeated measurements followed by Fisher's protected least significant difference test when appropriate.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

High expression of furin in human and mouse arthritic joints.

We first investigated the expression of furin in the synovial pannus of patients with RA and in controls. Real-time PCR analysis revealed higher levels of furin transcripts in RA patients than in controls (mean ± SEM mRNA fold change 4.9 ± 0.2 versus 1 ± 0.1; P < 0.01) (Figure 1A). Immunohistochemistry analysis revealed dramatic levels of furin, which was found in both the cells and the ECM of the synovial pannus of RA patients. In contrast, furin was not detected in human control synovium.

thumbnail image

Figure 1. Furin is highly expressed in the synovial pannus, and enzymatic activity is enhanced in mice with arthritis. A, Synovial tissue was obtained from patients with rheumatoid arthritis (RA) and from control patients. Furin was highly expressed in synovial pannus in RA patients but was absent from synovial tissue in control patients, as assessed by real-time polymerase chain reaction and immunostaining. Values are the mean ± SEM. ## = P < 0.01 versus control patients. B, Immunohistochemistry revealed that furin was expressed in the synovium of arthritic mice but not in that of naive mice. Furin activity was significantly higher in the paws of arthritic mice than in those of naive mice. Values are the mean ± SEM. # = P < 0.05 versus naive mice. C, Zymography of matrix metalloproteinase 2 (MMP-2) and MMP-9 was performed in the supernatants of paw cultures from naive mice and arthritic mice (n = 3 experiments). Catalytic activity of culture supernatants showed that the maturation of MMP-2 and MMP-9 was higher in paws of mice with arthritis than in those of naive mice. Values are the mean ± SEM. ### = P < 0.0001 versus naive mice. Original magnification × 20 in A; × 10 in B.

Download figure to PowerPoint

Similarly, high levels of furin were found in the synovial pannus of arthritic mice, but furin was restricted to a very few cells in the synovium of naive mice (Figure 1B). Moreover, furin activity was significantly higher in the paws of arthritic mice than in those of naive mice (mean ± SEM 29.36 ± 20.59 × 103 μg protein versus 13.98 ± 5.80 × 103 μg protein). Similarly, analysis of the supernatants derived from paw cultures showed higher levels of active forms of MMP-2 and MMP-9 in arthritic mice than in naive mice (Figure 1C). These findings indicated that increased furin expression/activity is associated with increased cleavage of immature forms of MMP-2 and MMP-9, previously reported as furin substrates (18).

Systemic furin protects against arthritis and bone loss in mice.

To investigate the role of furin in arthritis, we administered furin or its inhibitor α1-PDX IP in mice with CIA. At the time that the mice were killed, the arthritis score was significantly reduced following injection of furin (mean ± SEM 12.61 ± 1.67) and increased following injection of α1-PDX (20.41 ± 1.58) as compared with injection of vehicle (17.10 ± 1.31) (both P < 0.05) (Figure 2A). Histologic analysis showed a thick synovial pannus invading the joint in vehicle-treated arthritic mice, which was markedly reduced in furin-treated mice. In contrast, the synovial pannus of α1-PDX–treated mice was thicker and more invasive into the joint. The total histologic score was lower in arthritic mice receiving furin (4.67 ± 1.38), whereas damage was worse in mice treated with α1- PDX (11.67 ± 0.17) than in vehicle-treated mice (9.44 ± 1.05) (P < 0.01 for furin versus vehicle and P < 0.05 for α1-PDX versus vehicle) (Figure 2A). Synovial inflammation and invasion as well as the parameters of cartilage and bone destruction were reduced in furin-treated mice (Table 1). The protective effects of furin were partially reversed by α1-PDX, suggesting that the prevention of arthritic lesions might be mediated by blockade of furin activity.

thumbnail image

Figure 2. Furin protects against arthritis and joint destruction and diminishes the expression of matrix metalloproteinases (MMPs) in the synovial pannus in arthritic mice. A, The mean ± SEM arthritis score was measured in the paws of mice (n = 7–10 mice per group) from the onset of arthritis until mice were killed. The administration of furin reduced the arthritis score, while α1-PDX increased it. Mean ± SEM histologic scores showed that furin significantly reduced histologic lesions, while these lesions were increased by α1-PDX. ### = P < 0.0001 versus naive mice. ∗ = P < 0.05; ∗∗ = P < 0.01 versus vehicle (Veh)–treated mice. § = P < 0.05; §§ = P < 0.01; §§§ = P < 0.0001 versus α1-PDX–treated mice. B, The expression of MMP-2, MMP-9, and MMP-14 was investigated by immunohistochemistry. MMP expression was mainly observed in the synovial pannus and was reduced in mice treated with furin and enhanced in the presence of α1-PDX. Original magnification × 10.

Download figure to PowerPoint

Table 1. Inflammation and bone parameters in mice at the time that they were killed*
 Naive miceMice with CIA
Vehicle-treatedFurin-treatedα1-PDX–treated
  • *

    Values are the mean ± SEM. Weight and histologic parameters were measured at the time that mice were killed. Bone mineral density (BMD) was measured at baseline and at the time that mice were killed. Bone gain (as a percent of BMD) was lower in vehicle-treated mice with collagen-induced arthritis (CIA) than in naive mice; this effect was reversed in furin-treated mice but enhanced in α1-PDX–treated mice (n = 10–17 mice per group).

  • P < 0.001 versus naive mice.

  • P < 0.01 versus vehicle-treated mice; P < 0.001 versus naive mice and versus α1-PDX–treated mice.

  • §

    P < 0.001 versus naive mice; P < 0.05 versus vehicle-treated mice.

  • P < 0.01 versus naive mice, versus vehicle-treated mice, and versus α1-PDX–treated mice.

  • #

    P < 0.01 versus naive mice; P < 0.001 versus vehicle-treated mice and versus α1-PDX–treated mice.

  • **

    P < 0.01 versus naive mice.

  • ††

    P < 0.01 versus vehicle-treated mice and versus α1-PDX–treated mice.

Weight, grams22.2 ± 0.319.8 ± 0.420.7 ± 0.619.74 ± 0.52
Synovial inflammation, 0–3 scale02.5 ± 0.271.42 ± 0.443.00 ± 0
Synovial invasion, 0–3 scale02.25 ± 0.331.25 ± 0.253 ± 0§
Bone histologic score, 0–3 scale02.31 ± 0.361.08 ± 0.422.75 ± 0.1
Cartilage destruction, 0–3 scale02.37 ± 0.260.92 ± 0.35#2.92 ± 0.08
Change in BMD, gm/cm20.258 ± 0.0230.152 ± 0.018**0.243 ± 0.017††0.105 ± 0.025

Because systemic bone loss related to RA causes fractures and disability, we investigated the effect of furin on bone loss. The gain in BMD was significantly higher in furin-treated mice (0.243 ± 0.017 gm/cm2) than in mice treated with vehicle (0.152 ± 0.018 gm/cm2) (P < 0.01), but gain in BMD was not affected when the mice were treated with α1-PDX (0.105 ± 0.025 gm/cm2) (P = 0.15) (Table 1).

Systemic furin reduces MMP expression and activity in mouse paws.

The expression and activity of MMPs involved in joint destruction were investigated after systemic administration of furin and α1-PDX in mice. Immunohistochemistry revealed a dramatic thickening of the synovium and a higher expression of MMPs in cells and ECM in the pannus of arthritic mice receiving vehicle (Figure 2B). Moreover, the administration of furin reduced the levels of MMP expression, which was linked to the reduction of the synovial pannus. In contrast, the expression of MMPs was enhanced in the presence of α1-PDX; this was related to greater proliferation and invasiveness of the synovium.

The catalytic functions of proteinases obtained from the paws were assessed by zymography. The analysis of gelatinolytic activity in the joint revealed that most of the MMP-2 and MMP-9 was present as precursor forms in naive mice (Figure 3A). Analysis of samples derived from arthritic mice revealed the accumulation of the processed active forms of these MMPs, indicating the existence of a proteolytic maturation process during inflammation. The administration of furin reduced the cleavage of MMP-2 and MMP-9, whereas the administration of α1-PDX induced greater cleavage of both MMPs. Accordingly, the activities of MMP-2, MMP-9, and MMP-14 were up-regulated under inflammatory conditions, whereas furin and α1-PDX decreased and enhanced their activities, respectively (Figure 3B).

thumbnail image

Figure 3. Furin reduces catalytic activity and MMP expression in the paws of arthritic mice. A, Shown is zymography of paw tissue from naive mice and from arthritic mice treated with vehicle, furin, or α1-PDX. The maturation of MMPs was greater in mice with arthritis than in naive mice. Furin administration inhibited the accumulation both of MMP precursors and of their processed forms. Both proMMPs and processed MMPs were enhanced by the administration of α1-PDX. Values are the mean ± SEM. ### = P < 0.0001 versus naive mice. ∗∗∗ = P < 0.0001 versus vehicle-treated mice. §§ = P < 0.01; §§§ = P < 0.0001 versus α1-PDX–treated mice. B, Protein MMP arrays were obtained using proteins derived from the paws of naive mice and paws of arthritic mice treated with vehicle, furin, or α1-PDX. MMP activity was higher in paws of vehicle-treated mice, reduced by furin, and enhanced by α1-PDX. Values are the mean ± SEM. # = P < 0.05 versus naive mice. ∗ = P < 0.05 versus vehicle-treated mice. § = P < 0.05 versus α1-PDX–treated mice. SU = standard unit (see Figure 2 for other definitions).

Download figure to PowerPoint

Systemic furin decreases local production of proinflammatory cytokines and enhances numbers of Treg cells.

Since systemic Treg cells modulate systemic inflammatory responses, Treg cells were assessed in the spleen. The induction of arthritis in mice resulted in a diminished number of Treg cells (CD4+CD25+ FoxP3+) compared with numbers of Treg cells in naive mice (mean ± SEM 2.71 ± 0.07% versus 3.74 ± 0.02%; P < 0.01) (Figure 4A). Furin administration restored the percentage of Treg cells (3.39 ± 0.05%; P < 0.05 versus vehicle-treated mice), while α1-PDX further depressed the number of Treg cells (2.11 ± 0.32%; P < 0.05 versus vehicle treatment).

thumbnail image

Figure 4. Furin promotes the generation of Treg cells and restores the Th1/Th2 balance. A, Treg cells (CD4+CD25+FoxP3+) were counted in the spleens in each group of mice. The number of Treg cells was lower in immunized mice; this reduction was reversed with furin and worsened with α1-PDX (n = 3–4 mice per group). Values are the mean ± SEM. ## = P < 0.01 versus naive mice. ∗ = P < 0.05 versus vehicle (Veh)–treated mice. §§ = P < 0.01 versus α1-PDX–treated mice. B, Concentrations of antiinflammatory and proinflammatory cytokines were measured in the paws at the time that mice were killed. Local levels of cytokines showed the reversal of the Th1/Th2 balance (n = 3–9 mice per group) in furin-treated mice, and this effect was reduced in arthritic mice treated with α1-PDX. Interleukin-17 (IL-17) levels were increased in vehicle-treated mice and were unchanged after furin administration. Values are the mean ± SEM. # = P < 0.05; ## = P < 0.01; ### = P < 0.0001 versus naive mice. ∗∗ = P < 0.01; ∗∗∗ = P < 0.0001 versus vehicle-treated mice. §§§ = P < 0.0001 versus α1-PDX–treated mice. Prot = protein; TNFα = tumor necrosis factor α.

Download figure to PowerPoint

Therefore, we further assessed the regulation of the Th1/Th2 balance by furin in the paws. Furin administration decreased local concentrations of proinflammatory cytokines (TNFα [P < 0.01 versus vehicle], IL-1) and increased the production of antiinflammatory cytokines (IL-4 [P < 0.01 versus vehicle], IL-10 [P < 0.0001 versus vehicle]) (Figure 4B). In contrast, α1-PDX resulted in significantly higher levels of IL-1β and TNFα and lower levels of IL-10 and IL-4. However, local levels of IL-17 were not driven by the presence of furin or its inhibition, indicating that the effect of furin is not mediated via Th17 cells.

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We found that the systemic administration of furin reduced arthritis in conjunction with a reduction of the synovial pannus. The prevention of arthritic lesions by furin is mediated through the restoration of the Th1/Th2 balance in the joint, and is also associated with an enhanced proportion of Treg cells in the spleen, indicating systemic regulation of proliferation of the local pannus.

Initially, we observed that furin was highly expressed in samples from RA patients and from mice with arthritis. Furin expression in these tissues suggests that furin might promote joint inflammation or act as a compensatory mechanism in alleviating the inflammatory process. Therefore, we used exogenous furin to investigate the role of furin in the development of arthritis. We found that furin did not affect the general health or the behavior of the mice throughout these short-term experiments. No macroscopic alterations in other organs were noted at the time of death. The weight loss observed in mice treated with vehicle or furin inhibitor might be related to the inflammation conditions. However, ascertainment of information regarding toxicity due to prolonged or repeated administration of furin such as cell transformation requires additional investigations.

Systemic administration of furin was found to reduce local joint inflammation and damage and to promote the restoration of Treg cell numbers. In contrast, α1-PDX was found to enhance the severity of arthritis and reduce the proportion of Treg cells. This systemic effect is further shown by the fact that bone loss was prevented by furin and worsened by α1-PDX administration. Treg cells are involved in the suppression of both Th1 and Th2 pathogenic immune responses and control T cell homeostasis (2, 19–22) as well as the maturation of T effector cells (23). Furin has recently been found to play an important role in immunity through Treg cell activation, as revealed in studies using conditional deletion of furin in T cells, which resulted in impairment of the function of Treg cells as well as that of effector cells (15). Deletion of furin in CD4+ cells in mice resulted in the development of inflammatory bowel disease and lymph node hypertrophy, leading to Th1 activation as confirmed by serum cytokine patterns.

Here we show that in a context of systemic inflammation, exogenous furin enhanced the generation of Treg cells in a peripheral lymphoid organ such as the spleen, indicating that this enzyme plays an active role in immune tolerance, which is illustrated by an improvement of joint lesions. Our results provide further evidence that Treg cells have a role in regulating the immune tolerance that prevents the expansion of other T cells and the activation and functioning of autoreactive T cells (24). Th1 cells constitute a group of CD4+ T helper cells that are actively involved in the inflammatory immune responses that occur during intracellular infections, organ transplantation, and autoimmune responses (25), and they drive the immune responses that lead to joint damage. In these responses, furin reversed the local Th1/Th2 balance, as demonstrated by the reduced local levels of inflammatory cytokines, while the levels of antiinflammatory cytokines were enhanced. The shift toward Th1 with α1-PDX observed here further showed that blockade of enzymatic activity is one of the mechanisms driven by furin and is consistent with the invalidation of the furin gene in lymphocytes. IL-17, which is produced by Th17 cells, is a dominant cytokine involved in the development of arthritis (22, 26). However, we failed to demonstrate any regulation of IL-17 production by furin or its inhibition, suggesting that IL-17 is not involved in immune mediation in this particular condition.

Our finding of immune-mediated response to furin in Treg cells and in the Th1/Th2 balance is related to its enzymatic activity, which could be reversed by α1-PDX. Indeed, in studies using various in vitro and cellular models, this inhibitor has previously been reported to inhibit the activity of furin, leading to the inability of the enzyme to induce the proteolytic cleavage of furin substrates or to induce its downstream effectors (27). The inhibition of furin was found to reduce the expression and activation of various molecules, including cytokines, adhesion molecules, and growth factors (28–30). The in vivo integrative approach used here showed that exogenous furin affects immune regulation more than local tissue regulation.

The joint destruction in RA is a consequence of an imbalance between the catabolic and anabolic pathways, involving local factors and matrix-degrading enzymes induced by effector cells (2, 19). Produced in the form of precursors, most of these molecules are proteolytically activated by furin (31, 32), whereas the processing of others inhibits their activity and functions (4). Previously, in chondrocytes, furin was found to interact with the proADAMTS-4 precursor form and to mediate its intracellular activation (33), whereas its inhibition resulted in reduced release of transforming growth factor β (TGFβ) (34). Therefore, endogenous furin promotes the maturation of some catalytic enzymes that might impair the balance of cartilage homeostasis. Similarly, metalloproteinases are central enzymes in matrix degradation and remodeling, which are key events in joint destruction and inflammation. Immunohistochemistry analysis performed on sections from arthritic mice revealed that the expression of MMP-2, MMP-9, and MMP-14 was increased, but was reduced in arthritic mice receiving systemic furin. Further analysis of tissue derived from the joints of mice revealed an association between increased immunoreactivity toward MMP-2 and MMP-9 and the level of MMP-2 and MMP-9 activity as assessed by zymography. In arthritic mice the injection of furin significantly reduced the enzymatic activity of both MMP-2 and MMP-9.

Although the mechanism of action of the inhibitory effect of systemic furin on MMP-2 expression and activity is unknown, the presence of furin in the circulation might inhibit its well-established local function on MMP-2 and MMP-9 processing and activation (13). These differences might be related to the ability of furin to activate the Treg cells that subsequently release mediators involved in the inhibition of MMP expression/activation, such as IL-4 and IL-10, or that reduce the expression of MMP inducers, such as TNFα. Therefore, although furin is involved in the activation of molecules that mediate matrix degradation, its protective effect might be related to its ability to inhibit the expression of these matrix degradation molecules. In addition, a previous study indicated that deletion of furin in CD4+ cells was found to impair T cell function by altering the production of TGFβ (15). Indeed, TGFβ promotes naive T cells to a Treg cell phenotype (35), and TGFβ-secreting cells suppress the development and severity of CIA (36). Therefore, our observations raise the hypothesis that exogenous furin might also affect the processing of TGFβ, resulting in protection against arthritis. We await further investigations into this.

In conclusion, these data indicate that systemic administration of furin promotes Treg cells and restores the Th1/Th2 balance through the inhibition of the conversion of MMP-9 and MMP-2 precursors into their active forms. Furin could therefore play a major role in reducing autoimmunity in RA, which raises the possibility of the potential use of furin or its derivatives in inflammation, thus providing a new therapeutic agent for autoimmune diseases such as RA.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Cohen-Solal 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 conception and design. Lin, Khatib, Cohen-Solal.

Acquisition of data. Lin, Ah Kioon, Lalou, Larghero, Launay, Khatib, Cohen-Solal.

Analysis and interpretation of data. Lin, Khatib, Cohen-Solal.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

The authors are grateful to M. Zouali for helpful advice.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES