IL-10, a cytokine produced primarily by macrophages, B lymphocytes, and Th2 cells, has both immunostimulatory and immunosuppressive properties. A homologue of IL-10 encoded by EBV, known as viral IL-10 (vIL-10), is also able to suppress the immune response, but may lack some of the immunostimulatory properties of IL-10. To evaluate the potential of vIL-10 to block the progression of rheumatoid arthritis, we have utilized a replication-defective adenovirus vector to deliver the gene encoding vIL-10 to the knee joints of rabbits with Ag-induced arthritis. Intraarticular expression of vIL-10 significantly reduced leukocytosis, cartilage matrix degradation, and levels of endogenous rabbit TNF-α, as well as the degree of synovitis, while maintaining high levels of cartilage matrix synthesis. Interestingly, an antiarthritic effect was also observed in opposing contralateral control knee joints that received only a marker gene. An adenoviral vector carrying the enhanced green fluorescent protein marker gene was used to demonstrate that a morphologically similar subset of cells infected in the injected knee joint are able to traffic to the uninjected contralateral knee joint. Our results suggest that direct, local intraarticular delivery of the vIL-10 gene may have polyarticular therapeutic effects.

Rheumatoid arthritis (RA)3 is a systemic autoimmune disease characterized by chronic erosive inflammation of the peripheral joints. The etiology of RA is largely unknown, although current evidence suggests contributions from both environmental and genetic components. Currently, the symptoms of RA are managed with a variety of pharmacological agents, no combination of which has proven efficacious in halting the progression of the disease. However, recent research has identified a number of cytokines and other biological agents that show promise as novel antiarthritic drugs (1). Among the most promising therapeutic agents are those that modulate the activities of the proinflammatory cytokines TNF-α and IL-1 (2, 3, 4, 5), which are thought to be important mediators that drive the pathophysiology of RA (6, 7). Most of the recently described biologic agents are proteins, such as the soluble receptors for TNF-α and IL-1, anti-TNF-α Abs, and IL-1 receptor antagonist protein. Gene transfer can potentially circumvent the inherent delivery problems associated with proteins by the transfer of genes encoding the therapeutic agent directly to the synovial lining of the joint (8).

One cytokine of particular interest as a therapeutic for RA is IL-10. Originally termed cytokine synthesis inhibitory factor, IL-10 (9) is a 35-kDa homodimeric cytokine product of Th2 cells, B cells, and macrophages. The actions of IL-10 are diverse in that IL-10 can be antiinflammatory, immunosuppressive, or immunostimulatory, depending upon the target cell. Primarily, IL-10 can act as an antiinflammatory by inhibiting synthesis of macrophage-derived proinflammatory cytokines, such as TNF-α, IL-1α, IL-1β, IL-6, and IL-8 (10, 11, 12, 13). The classification of IL-10 as an immunosuppressive agent is due to its ability to inhibit the Ag-presenting functions of macrophages and dendritic cells through the down-regulation of MHC class II molecules and the costimulatory molecules ICAM-1 and B7.1 and B7.2 (14, 15, 16, 17). Aside from its suppressive activities, IL-10 also retains several stimulatory properties. Most notable are the potent effects that IL-10 has on B cell proliferation and differentiation (18) and its ability to act as a chemoattractant for CD8+ T cells (19). IL-10 is also able to block the apoptosis of both CD4+ and CD8+ T cells, as well as germinal center B cells (20, 21). IL-10 is elevated in the serum of RA patients; the significance of this for disease pathology, however, is currently under debate (22, 23). While IL-10 may be acting in a suppressive capacity in RA, high levels of IL-10 also show strong correlation with rheumatoid factor titers, as well as spontaneous IgM-rheumatoid factor production (23). Thus, therapy with exogenous IL-10 has the potential to enhance humoral immunity and undermine its potential effects as a therapeutic.

Examination of human IL-10 and murine IL-10 cDNA sequences revealed extensive homology to an open reading frame product from the EBV (24). Originally termed BCRF1, this viral protein product is now considered a viral form of IL-10 (vIL-10) (24, 25). The human IL-10 and vIL-10 mature proteins are 84% identical, with most of the divergence found at the amino terminus. vIL-10 shares many of the antiinflammatory and immunosuppressive properties of IL-10, but seems to lack certain immunostimulatory functions (18, 26). Like its human homologue, vIL-10 has demonstrated suppressive functions in several inflammatory settings. It has been shown to increase the survival of mice with experimental endotoxemia (27), prolong the survival of cardiac allografts in mice (28), and suppress the rejection of allogeneic and syngeneic tumors in mice, while, in contrast, cellular IL-10 stimulated tumor rejection (29). The results of the mouse tumor model, in particular, suggest a potential superiority of vIL-10 as an immunosuppressive agent over that of its cellular homologue.

In the present study, we have tested the ability of adenovirus-mediated gene delivery of vIL-10 to protect rabbit knee joints from Ag-induced arthritis (a.i.a). We found that administration of vIL-10 could effectively halt the progression of experimental arthritis by blocking leukocytic infiltration into the joint, while normalizing cartilage metabolism and reducing the degree of synovitis. Most interestingly, therapeutic effects were also observed in opposing, contralateral arthritic knees that had received only the LacZ marker gene. A morphologically similar subset of adenovirally infected cells was shown to traffic from the treated joint to the untreated, inflamed joint. These results suggest that the therapeutic effects of local gene therapy using vIL-10 may not be limited to the treated joint, but may also affect additional joints within the same animal.

Female New Zealand White rabbits weighing ∼5–6 lb were obtained from Myrtles Rabbitry (Thompson Station, TN) and housed at the Central Animal Facility at the University of Pittsburgh (Pittsburgh, PA). Animals were allowed to acclimate 3 days before experimentation and were fed water and chow ad libitum.

The vectors used in this study were E1- and E3-deleted type 5 replication-defective adenoviruses (30). cDNAs encoding vIL-10, β-galactosidase (LacZ), and enhanced green fluorescent protein (eGFP) were inserted into the E1 region of the vector with gene expression driven by the human CMV early promoter. High titer virus was produced by permissive replication in the 293 human embryonic kidney cell line (American Type Culture Collection, Manassas, VA), as described previously (31). Viral titers were determined by optical density at 260 nm (OD260) where 1 OD unit = 1012 viral particles (32).

Rabbits were sensitized to OVA by a series of two intradermal injections of 5 mg OVA emulsified in CFA in the first injection, and in IFA in the second (33). Two weeks following the second injection, an acute articular arthritis was initiated in both hind knees of the rabbits by the intraarticular administration of 5 mg OVA dissolved in 0.5 ml of saline. Twenty-four hours after induction of a.i.a., 5 × 109 particles of replication-defective adenovirus encoding either vIL-10 or LacZ was suspended in 0.2 ml of sterile saline and injected into the joint space via the patellar tendon.

Rabbit knee joints were lavaged on days 3 and 7 post adenovirus administration, by injection of 1 ml Gey’s balanced salt solution (Life Technologies, Rockville, MD) through the patellar tendon. After manipulation of the joint to allow for ample mixing, the needle was reinserted and the fluid aspirated. Leukocytes in recovered lavage fluids were counted with a hemocytometer. Levels of vIL-10 expression in recovered lavage fluids and serum were measured using a cytokine ELISA kit (R&D Systems, Minneapolis, MN).

To measure rates of proteoglycan synthesis, articular cartilage was shaved from the femoral chondyles and weighed. Approximately 30 mg of cartilage were then incubated in 1 ml Neuman & Tyell (Life Technologies) serumless medium with 40 μCi of 35SO4−2 for 24 h at 37°C. Subsequently, medium was recovered and stored at −20°C. Proteoglycans were extracted from the cartilage shavings by incubation for 48 h in 1 ml of 0.5 M NaOH at 4°C with gentle agitation. Following chromatographic separation of unincorporated 35SO4−2 using PD-10 columns (Pharmacia, Piscataway, NJ), radiolabeled glycosaminoglycans (GAGs) released into the culture media or recovered by alkaline extraction from the cartilage were quantitated using scintillation counting, as described previously (34).

To quantitate GAGs released into the joint space as a result of cartilage proteoglycan breakdown, recovered lavage fluids were first centrifuged at 12,000 × g for 10 min to remove debris and the supernatants recovered. Aliquots of 100 μl of lavage fluid were treated with papain: 20 μl of papain suspension (type III, 19 U/mg protein; Sigma, St. Louis, MO) was added to 1 ml of buffer containing 10 mM EDTA and 0.4 M sodium acetate (pH 5.2). The papain solution (100 μl) was added to the lavage fluid (100 μl) and incubated overnight at 60°C. Papain was inactivated by the addition of iodoacetic acid to a final concentration of 4 mM. The samples were then centrifuged at 12,000 × g for 10 min. Afterward, 2 U of hyaluronate lyase (Sigma) were added and the samples incubated at 37°C overnight. Sulfated GAG concentrations were measured by a colorimetric dye binding assay using 1,9-dimethylene blue, as previously described (35).

Bilateral arthritis was induced in rabbit knees as described above. Twenty-four hours after induction of a.i.a., 5 × 1010 particles of adenovirus encoding the cDNA for eGFP (Ad.eGFP) were injected into one knee joint, while the opposing joint received no treatment. Three days postvirus administration, both knee joints were lavaged and the recovered cells divided into two equal portions. One-half of each recovered lavage fluid was cultured in Ham’s F12 media (Life Technologies) supplemented with 10% FCS. After 24 h, cultures derived from both the injected knee and the contralateral joint were examined by fluorescent microscopy (Phase Contrast-2; Nikon, Melville, NY) for the presence of cells that were expressing eGFP. The other half of each recovered lavage fluid was reserved for analysis by FACS. One-color flow cytometry was performed using an EPICS ELITE flow cytometer (Coulter, Hialeah, FL). As a control, the gate for flow cytometry was set using recovered cells aspirated from an inflamed nontransduced control rabbit knee. Logarithmically amplified fluorescent data were collected on 20,000 cells from each recovered lavage fluid, which had been extensively washed and then resuspended in flow cytometry buffer (0.01% NaN3 and 1% BSA in PBS).

Endogenous TNF-α levels in serum and lavage fluids were measured using a sandwich ELISA with specific anti-TNF-α polyclonal Abs (PharMingen, San Diego, CA). Briefly, microtiter plates were coated with 50 μl anti-rabbit TNF-α capture Abs (4 μg/ml) overnight at 4°C, then washed twice with PBS containing 0.05% Tween 20 and blocked overnight at 4°C with 10% FCS in PBS. After washing the plate four times, 100 μl of standards and samples were incubated in duplicate overnight at 4°C. Plates were again washed and a biotin-conjugated anti-TNF-α secondary Ab (2 μg/ml) was added for 1 h at room temperature. A 30-min incubation with a 1:400 dilution of avidin-peroxidase (Sigma) followed an extensive washing of the plate. Finally, TMBlue (Intergen, Milford, MA) substrate was added (100 μl/well) and incubated at room temperature for 30 min. The reaction was stopped by the addition of 0.5 N H2SO4 to each well. The absorbance was read at 450 nm with a UVmax microplate reader (Molecular Devices, Menlo Park, CA). Standards were plotted to form a linear standard curve, and the unknown concentrations were determined by linear regression analysis.

Rabbit knees were dissected from euthanized animals at day 7, and tissues were fixed in 10% buffered formalin for several days. The fixed tissues were then embedded in paraffin, sectioned at 5 μm, and stained with hematoxylin and eosin. Sections were examined by light microscopy at ×40 magnification.

Data were analyzed using the Macintosh Statworks software program. Group comparisons were performed using both paired and unpaired Student’s t tests where appropriate.

To test the ability of vIL-10 to inhibit the inflammatory and chondrodestructive effects of a.i.a. in the rabbit knee joint, arthritis was induced in both knees of 16 rabbits. Twenty-four hours post induction, 5 × 109 particles of adenovirus encoding vIL-10 were injected into the right knee of eight rabbits, and 5 × 109 particles of adenovirus encoding LacZ were injected into the left knee of the same eight rabbits. An arthritic control group of eight rabbits received 5 × 109 particles of Ad.LacZ into both knees. A naïve control group of two rabbits was neither induced with arthritis nor injected with virus. Three days after injection of the adenovirus, both knees of each rabbit were lavaged with saline. At 7 days postinfection, the rabbits were sacrificed, the knees lavaged, dissected, and analyzed for effects of transgene expression. It should be noted that we have previously shown through extensive adenovirus titrations that adenoviral doses of 7 × 109 particles or less produce no substantial leukocytic infiltration in synovial fluid for up to 14 days after injection, while maintaining high levels of transgene expression (36).

ELISA measurements of vIL-10 levels in recovered lavage fluids detected ∼17 ng/ml and 22 ng/ml at days 3 and 7, respectively, in knees that received the adenovirus encoding vIL-10 (Fig. 1). vIL-10 was not detectable in sera (data not shown) or lavage fluid recovered from knees contralateral to those knees receiving the adenovirus encoding vIL-10.

FIGURE 1.

In vivo expression of vIL-10 after direct adenoviral delivery to rabbit knees with a.i.a. Twenty-four hours after induction of a.i.a., a group of eight rabbits was injected in the right knee with 5 × 109 particles of Ad.vIL-10 and in the left knee with 5 × 109 particles of Ad.LacZ. At days 3 and 7 after injection of the virus, the knees were lavaged and the recovered fluids were analyzed by ELISA for levels of vIL-10. Values shown represent the mean ± SEM of eight rabbit knee joints each from the vIL-10 and LacZ contralateral groups. Values shown represent the mean + SEM of eight rabbit joints from the vIL-10 and LacZ contralateral groups.

FIGURE 1.

In vivo expression of vIL-10 after direct adenoviral delivery to rabbit knees with a.i.a. Twenty-four hours after induction of a.i.a., a group of eight rabbits was injected in the right knee with 5 × 109 particles of Ad.vIL-10 and in the left knee with 5 × 109 particles of Ad.LacZ. At days 3 and 7 after injection of the virus, the knees were lavaged and the recovered fluids were analyzed by ELISA for levels of vIL-10. Values shown represent the mean ± SEM of eight rabbit knee joints each from the vIL-10 and LacZ contralateral groups. Values shown represent the mean + SEM of eight rabbit joints from the vIL-10 and LacZ contralateral groups.

Close modal

Leukocytes in the recovered lavage fluids from each group of rabbits were counted and compared as a quantitative measure of inflammation. As shown in Fig. 2, at day 3, the control group of rabbits, which were injected with Ad.LacZ into both knees, exhibited severe joint inflammation with a mean level of infiltrating leukocytes exceeding 7.7 × 106 per ml of recovered lavage fluid. By day 7, this level had risen to 9.3 × 106 leukocytes per ml of recovered lavage fluid. In comparison, the group of rabbits that was injected with Ad.vIL-10 in the right knee and Ad.LacZ in the left knee had significantly lower white blood cell infiltrates in the right knees than in control knees. Levels in the vIL-10 knees at day 3 averaged 3.2 × 106 infiltrating cells, a 72% reduction. By day 7, this reduction further increased to 81%. Interestingly, the knees receiving Ad.LacZ, contralateral to the Ad.vIL-10 treated knees, also displayed a reduction in the amount of cellular infiltration into the joint space. Rabbits injected with Ad.vIL-10 into the right knees and Ad.LacZ into the left, displayed a 23% reduction in the mean infiltration into the contralateral joint, and, by day 7, this difference increased to 60%, compared with the knees of rabbits receiving only Ad.LacZ.

FIGURE 2.

Leukocytic infiltration in a.i.a. knees after intraarticular injection of Ad.vIL-10 or Ad.LacZ. Twenty-four hours postinduction of a.i.a. in both knees of two groups of eight rabbits, one group received Ad.vIL-10 in the right knee and Ad.LacZ into the left, while the second (arthritic control) group of rabbits received Ad.LacZ in both knees. At days 3 and 7 after adenovirus injection, the joints were lavaged and numbers of leukocytes determined with a hemocytometer. Values shown represent the mean ± SEM of 4 rabbit joints from naïve controls, 8 knees each from the vIL-10 and contralateral joints, and 16 joints from the Ad.LacZ control group. Asterisks denote values which differ at p < 0.05 (Student’s t test and paired t test, where appropriate).

FIGURE 2.

Leukocytic infiltration in a.i.a. knees after intraarticular injection of Ad.vIL-10 or Ad.LacZ. Twenty-four hours postinduction of a.i.a. in both knees of two groups of eight rabbits, one group received Ad.vIL-10 in the right knee and Ad.LacZ into the left, while the second (arthritic control) group of rabbits received Ad.LacZ in both knees. At days 3 and 7 after adenovirus injection, the joints were lavaged and numbers of leukocytes determined with a hemocytometer. Values shown represent the mean ± SEM of 4 rabbit joints from naïve controls, 8 knees each from the vIL-10 and contralateral joints, and 16 joints from the Ad.LacZ control group. Asterisks denote values which differ at p < 0.05 (Student’s t test and paired t test, where appropriate).

Close modal

To examine the effect of intraarticular expression of vIL-10 on cartilage matrix degradation, the levels of GAGs released into synovial fluid as a result of proteoglycan breakdown were measured in recovered lavage fluids. As shown in Fig. 3 A, expression of vIL-10 inhibited cartilage breakdown. The control group of rabbits receiving injections of Ad.LacZ in both knees had very high levels of GAGs in the lavage fluids of both knees at both days 3 and 7. Rabbits injected with Ad.vIL-10 in the right knee and Ad.LacZ in the opposite joint showed reductions in the levels of released GAGs. By day 3, knees receiving Ad.vIL-10 displayed a 56% reduction in the amount of GAGs released into the joint space. This difference increased to 72% by day 7. Once again, the joints contralateral to those that received vIL-10 displayed a moderate reduction in GAG release at day 3 (38%), with a more significant reduction at day 7 (53%).

FIGURE 3.

Effect of vIL-10 on cartilage matrix metabolism. Experimental conditions were as described for Fig. 2. As an index of proteoglycan breakdown, concentrations of GAG released into the lavage fluids were measured spectrophotometrically (A). To measure rates of proteoglycan synthesis (B), pieces of articular cartilage were shaved from the femoral chondyles and their in vitro incorporation of 35SO4−2 into macromolecular material measured. Values given are the means ± SEM. Asterisks denote values which differ at p < 0.05 (Student’s t test and paired t test, where appropriate).

FIGURE 3.

Effect of vIL-10 on cartilage matrix metabolism. Experimental conditions were as described for Fig. 2. As an index of proteoglycan breakdown, concentrations of GAG released into the lavage fluids were measured spectrophotometrically (A). To measure rates of proteoglycan synthesis (B), pieces of articular cartilage were shaved from the femoral chondyles and their in vitro incorporation of 35SO4−2 into macromolecular material measured. Values given are the means ± SEM. Asterisks denote values which differ at p < 0.05 (Student’s t test and paired t test, where appropriate).

Close modal

The ability of vIL-10 to overcome the inhibition of new matrix synthesis was also examined by measuring 35S incorporation into proteoglycans in articular cartilage shavings from the femoral chondyles at day 7 (Fig. 3 B). Compared with naïve controls, the rates of GAG synthesis were depressed by 60% in a.i.a. knees of arthritic control rabbits that received only the Ad.LacZ virus into both joints. In contrast, GAG synthesis rates in knees that received Ad.vIL-10 were moderately protected with synthesis levels only 30% lower than those of naïve joints. Joints receiving Ad.LacZ contralateral to those that received Ad.vIL-10 exhibited only a 40% decline in synthesis rates, compared with naïve knee joints. However, due to interrabbit variability, the effect of vIL-10 on matrix synthesis in the injected and contralateral joints was not statistically significant.

The histological analysis of tissue recovered from the knees of each group of rabbits is shown in Fig. 4. Compared with tissue recovered from normal naïve rabbits (Fig. 4,A), sections from the Ad.LacZ control group displayed the severe synovitis typically observed with a.i.a. (Fig. 4,B). The synovium was thickened, fibrous, hyperplastic, and hypertrophic due to synovial cell proliferation and infiltration by mononuclear leukocytes. Treatment of a.i.a. knees with Ad.vIL-10 (Fig. 4,C) blocked disease to such a degree that these knees were largely indistinguishable from naïve control knees (Fig. 4,A). Opposing contralateral joints (Fig. 4 D) also displayed a large reduction in the amount of synovitis.

FIGURE 4.

Histological analysis of synovial tissue recovered from rabbit knees. Experimental conditions were the same as described for Fig. 2. Seven days after virus injection, synovial tissue was harvested, sectioned, and stained with hematoxylin and eosin. A, Synovium from a normal naïve rabbit (magnification, ×40). B, Representative section of synovium from a rabbit that received 5 × 109 particles of Ad.LacZ after a.i.a. induction (magnification, ×40). Relative to a naïve joint (A), the synovium is thicker, more fibrous, and hyperplastic. C, Representative tissue recovered from rabbit knee receiving 5 × 109 particles of Ad.vIL-10 after a.i.a. induction (magnification, ×40). D, Representative section of synovial tissue from knees injected with Ad.LacZ contralateral to those that were treated with vIL-10 (magnification, ×40).

FIGURE 4.

Histological analysis of synovial tissue recovered from rabbit knees. Experimental conditions were the same as described for Fig. 2. Seven days after virus injection, synovial tissue was harvested, sectioned, and stained with hematoxylin and eosin. A, Synovium from a normal naïve rabbit (magnification, ×40). B, Representative section of synovium from a rabbit that received 5 × 109 particles of Ad.LacZ after a.i.a. induction (magnification, ×40). Relative to a naïve joint (A), the synovium is thicker, more fibrous, and hyperplastic. C, Representative tissue recovered from rabbit knee receiving 5 × 109 particles of Ad.vIL-10 after a.i.a. induction (magnification, ×40). D, Representative section of synovial tissue from knees injected with Ad.LacZ contralateral to those that were treated with vIL-10 (magnification, ×40).

Close modal

Since TNF-α is thought to be a major proinflammatory contributor to the pathophysiology of RA and is the target of several novel therapies, the effect of Ad.vIL-10 therapy on endogenous levels of rabbit TNF-α in lavage fluids was determined (Fig. 5). At days 3 and 7 in naïve knees, levels of TNF-α averaged 30 pg/ml and 15 pg/ml, respectively. Interestingly, 3 days postadenovirus injection, the arthritic knees that received Ad.vIL-10 exhibited the highest average TNF-α levels, 175 pg/ml, followed by the opposing contralateral knees that received Ad.LacZ, which averaged 140 pg/ml. Control rabbits that received Ad.LacZ in both joints had the lowest levels of TNF-α, which averaged 100 pg/ml. By day 7, a dramatically different pattern of expression was observed. TNF-α levels in control animals that received Ad.LacZ in both joints remained high, averaging 160 pg/ml. In stark contrast, knees that received Ad.vIL-10, as well as the opposing contralateral Ad.LacZ knees, saw a significant reduction in TNF-α levels to near normal levels of 50 pg/ml and 40 pg/ml, respectively.

FIGURE 5.

.Levels of endogenous rabbit TNF-α levels following vIL-10 treatment. Experimental conditions and numbers of knees were as described for Fig. 2. Lavage fluids from both days 3 and 7 were assayed for rabbit TNF-α levels by ELISA. Values given are means ± SEM. Asterisks denote values that differ at p < 0.05 (Student’s t test and paired t test, where appropriate).

FIGURE 5.

.Levels of endogenous rabbit TNF-α levels following vIL-10 treatment. Experimental conditions and numbers of knees were as described for Fig. 2. Lavage fluids from both days 3 and 7 were assayed for rabbit TNF-α levels by ELISA. Values given are means ± SEM. Asterisks denote values that differ at p < 0.05 (Student’s t test and paired t test, where appropriate).

Close modal

The results of the above experiments suggested that intraarticular injection of a replication-defective adenovirus vector encoding vIL-10 could have antiarthritic effects at sites distal to the site of injection. Cell trafficking provides a possible mechanism by which adenovirally transduced cells may be involved in conferring the therapeutic effect seen in the contralateral joint. Previous experiments with an adenoviral vector encoding the luciferase marker gene suggested that cells from the injected joint were able to traffic to the contralateral joint and its draining lymph node (36). In an attempt to visualize the type(s) of cells trafficking to the contralateral joint, three rabbits were induced bilaterally with a.i.a., as described in the above experiments. Twenty-four hours postinduction, 5 × 1010 particles of Ad.eGFP were injected into one knee of each rabbit. Three days post injection, each joint was lavaged and the recovered fluids divided for either tissue culture or FACS analysis. FACS analysis of the lavage fluid recovered from the injected joint showed that ∼20% of the cells infiltrating the joint space were GFP-positive, whereas 3% of infiltrating cells from the contralateral joint were positive for GFP (Fig. 6, A and B). Fluorescent microscopy detected GFP-positive cells of various morphologies in cultures established from the ipsilateral joint (Fig. 6,C). In contrast, GFP-expressing cells cultured from the contralateral joint were of only the single morphology shown in Fig. 6 D. These results suggest that a certain subset of infected cells is able to traffic from the injected knee joint to the contralateral joint.

FIGURE 6.

Expression and distribution of eGFP-positive cells after intraarticular injection of Ad.eGFP into a.i.a. knees. Twenty-four hours postinduction of a.i.a in both knees of three rabbits, 5 × 1010 particles of Ad.eGFP were injected into the right joint of each rabbit. The left knee joint was not treated. Three days after virus administration, each joint was lavaged and the recovered fluids divided for analysis either by fluorescent microscopy or FACS. A, FACS analysis of infiltrating cells recovered in lavage fluids from the Ad.eGFP-injected joints show that 20% are positive for GFP expression. B, Three percent of cells recovered from the contralateral joint are positive for GFP expression, as determined by flow cytometry. C, A representative view of the several distinct cellular morphologies of eGFP-positive cells recovered from the ipsilateral joint. D, A representative view of eGFP positive cells from the contralateral joint display a single cellular morphology.

FIGURE 6.

Expression and distribution of eGFP-positive cells after intraarticular injection of Ad.eGFP into a.i.a. knees. Twenty-four hours postinduction of a.i.a in both knees of three rabbits, 5 × 1010 particles of Ad.eGFP were injected into the right joint of each rabbit. The left knee joint was not treated. Three days after virus administration, each joint was lavaged and the recovered fluids divided for analysis either by fluorescent microscopy or FACS. A, FACS analysis of infiltrating cells recovered in lavage fluids from the Ad.eGFP-injected joints show that 20% are positive for GFP expression. B, Three percent of cells recovered from the contralateral joint are positive for GFP expression, as determined by flow cytometry. C, A representative view of the several distinct cellular morphologies of eGFP-positive cells recovered from the ipsilateral joint. D, A representative view of eGFP positive cells from the contralateral joint display a single cellular morphology.

Close modal

In this study, we have examined the ability of vIL-10, when expressed intraarticularly by gene transfer, to inhibit the pathologies associated with a.i.a. in a rabbit knee model. We have demonstrated that direct intraarticular injection of Ad.vIL-10 can ameliorate ongoing a.i.a. in rabbit knee joints. The administration of 5 × 109 particles of Ad.vIL-10 resulted in significant decreases in several indices of disease, including intraarticular leukocytic infiltration, TNF-α levels, and the degree of synovitis in each joint. Furthermore, vIL-10 displayed a strong chondroprotective effect in blocking cartilage degradation while maintaining new matrix synthesis. This antiarthritic effect was not only apparent in treated joints, but also in contralateral joints that received only the LacZ marker gene, suggesting that local intraarticular gene therapy may confer polyarticular therapeutic effects.

The potent antiinflammatory effects of vIL-10 observed in the rabbit a.i.a. model are consistent with the in vitro actions of vIL-10 on T cells and macrophages. vIL-10 has been shown to block the expression of certain cytokines, such as IL-2, IL-12, IFN-γ, and GM-CSF by T cells (9), as well as proinflammatory cytokines, such as TNF-α and IL-1β by macrophages (10, 11, 12), that are thought to be important mediators in the pathogenesis of RA (6, 7). We have demonstrated that levels of TNF-α in the arthritic knees were reduced at day 7 following vIL-10 treatment. Thus, it is likely that the therapeutic effects of vIL-10 seen in this system are due, in part, to the suppression of TNF-α by vIL-10. However, why TNF-α is elevated in lavage fluids from day 3 knees is unclear. In addition to blocking the production of proinflammatory cytokines and T cell growth factors, vIL-10 is able to block Ag presentation through the down-regulation of MHC class II, ICAM-1, and B7 expression on macrophages and dendritic cells (14, 15, 16, 17).

We observed significant effects of intraarticular vIL-10 expression on cartilage metabolism. The chondroprotective effects demonstrated in this model are unlikely to reflect a direct effect on articular chondrocytes. In vitro transduction of chondrocytes with Ad.vIL-10 shows no effect on normal cartilage metabolism, nor can vIL-10 protect against IL-1β-induced suppression of matrix synthesis (E. Lechman and P. Robbins unpublished observations). Instead, it is quite possible that the chondroprotective effects observed with vIL-10 treatment are a consequence of the powerful antiinflammatory actions of vIL-10. In agreement with this conclusion, vIL-10 expression has been shown to inhibit synovial cell invasion of articular cartilage in a human SCID model of arthritis, but did not influence chondrocytic chondrolysis (37).

Recently, several groups have demonstrated that adenoviral administration of vIL-10 is efficacious for the treatment of murine collagen-induced arthritis (CIA) (38, 39, 40). However, only a mild therapeutic effect was seen when vIL-10 was introduced in established CIA (38). Nevertheless, in this report, we have demonstrated a more potent therapeutic effect following intraarticular expression of vIL-10 in established rabbit a.i.a. This discrepancy may be due to the inherent differences in the models of disease. In addition, we were able to deliver the genes directly to the joint, whereas others have delivered the adenovirus systemically (38, 39). Direct intraarticular administration and concentrated expression of vIL-10 at the site of Ag presentation may be critical to its ability to suppress disease. It is important to note that other groups have reported an Ad.vIL-10-associated inflammation upon local vector administration to mouse knee joints (38). However, we have not observed this effect in naïve rabbit joints with the doses of virus used in the vIL-10 studies presented here. In fact, we have demonstrated, via extensive titrations of adenovirus in naïve rabbit knees, that adenoviral doses of less than 7 × 109 particles generate little or no measurable inflammation within the synovial fluid (36). This apparent disparity in observations may be due to titers of virus, site of injection, purity of virus preparation, and virus storage buffer conditions. Furthermore, we saw no increase in the persistence of gene expression due to expression of vIL-10 (data not shown) as other groups have reported (41).

The experiments described here suggest that intraarticular delivery of the vIL-10 gene to a single joint has polyarticular antiinflammatory effects. This phenomenon was previously reported as a “contralateral effect” following the administration of soluble receptors for TNF-α and IL-1 by replication-defective adenovirus (36). Levels of these gene products in the serum and contralateral joint in treated animals were too low to confer a therapeutic effect. Marker gene trafficking studies indicated that virus was retained within the injected joint space. It was also determined that cells that had been transduced in one joint had migrated to the lymph nodes and synovial space of the contralateral joint. In this report, we have further examined the apparent trafficking of cells to the contralateral joint. Although we cannot completely discount that the presence of eGFP-positive cells in the contralateral joint may be due to the leakage of virus into the systemic circulation, the combined results of our trafficking studies with Ad.eGFP and Ad.Luc suggest that a population(s) of leukocytes exists that does indeed migrate from the treated joint to the contralateral joint. These cells may be capable of carrying therapeutic genes to other sites of inflammation through either the systemic circulation or the lymphatic system. Interestingly, it seems that these cells represent a single morphological cell type. From preliminary morphologic inspection, it appears that the cells resemble either macrophages or immature dendritic cells. However, further characterization of the trafficking cells has been hindered by the lack of appropriate immunological reagents with which to identify cells of the rabbit immune system. It is interesting to note that cell trafficking appeared to depend on several parameters, including the initial amount of inflammation in the Ad.eGFP-infected joint, the numbers of adenovirally transduced cells within the injected joint, and the amount of inflammation in the contralateral joint (data not shown). Whether these cells actually contribute to the contralateral effect, or are mere spectators, remains unclear.

We also have observed a similar “contralateral” joint effect with vIL-10 in the murine CIA model, in that periarticular injection of Ad.vIL-10 into an ankle joint at the onset of disease protects the remaining three uninjected paws (39). Expression of vIL-10 was shown in the injected paw as well as the draining lymph node (39). In a delayed-type hypersensitivity (DTH) model, where the mouse is first immunized to Ag and then challenged by injection of Ag into the footpad, prior injection of Ad.vIL-10 into that footpad blocks the DTH response. Interestingly, local footpad injection of Ad.vIL-10 also reduces the DTH response in the contralateral footpad injected with Ag (J. Whaler and C. Evans, unpublished observations). Taken together with our data in the rabbit model of a.i.a., these results are consistent with a model in which local injection of Ad.vIL-10 results in the genetic modification of cells, possibly APCs, that are able to traffic to spleen, lymph nodes, and sites of inflammation. Exactly how and where the therapeutic effect is conferred is still unclear. However, given that a similar effect has been observed in the rabbit a.i.a. model following injection of adenoviral vectors expressing inhibitors of TNF and IL-1, vIL-10 might be working in part through the down-regulation of proinflammatory cytokines by APCs. Clearly, a better understanding of the cell types important for conferring the therapeutic effects following local intraarticular injection of Ad.vIL-10 could provide insights into how to more effectively treat autoimmune and inflammatory diseases by gene therapy.

We thank Jennifer Siegert for her help with the manuscript.

1

This work was supported in part by National Institutes of Health Contract no. N01-AR-6-2225.

3

Abbreviations used in this paper: RA, rheumatoid arthritis; Ad., adenovirus; a.i.a., Ag-induced arthritis; vIL-10, viral IL-10; CIA, collagen-induced arthritis; GAG, glycosaminoglycan; LacZ, β-galactosidase gene; eGFP, enhanced green fluorescent protein.

1
Moreland, L W., L. W. Heck, Jr, W. J. Koopman.
1997
. Biologic agents for treating rheumatoid arthritis: concepts and progress.
Arthritis Rheum.
40
:
397
2
Moreland, L. W., G. R. Margolies, L. W. Heck, P. A. Saway, C. Blosch, R. Hanna, W. J. Koopman.
1996
. Recombinant soluble tumor necrosis factor receptor (p80) fusion protein: toxicity and dose finding trial in refractory rheumatoid arthritis.
J. Rheumatol.
23
:
1849
3
Arend, W. P., M. Malyak, M. F. Smith, Jr, T. D. Whisenand, J. L. Slack, J. E. Sims, J. G. Giri, S. K. Dower.
1994
. Binding of IL-1α, IL-1β and IL-1 receptor antagonist by soluble IL-1 receptors and levels of soluble IL-1 receptors in synovial fluids.
J. Immunol.
153
:
4766
4
Arend, W.P..
1993
. Interleukin-1 receptor antagonist.
Adv. Immunol.
54
:
167
5
Elliot, M. J., R. N. Maini, M. Feldmann, J. R. Kalden, C. Antoni, J. S. Smolen, B. Leeb, F. C. Breeveld, J. D. Macfarlane, H. Bijl.
1994
. Randomized double-blind comparison of chimeric monoclonal antibody to tumor necrosis factor α (cA2) versus placebo in rheumatoid arthritis.
Lancet
344
:
1105
6
Arend, W. P., J. M. Dayer.
1995
. Inhibition of the production and effects of interleukin-1 and tumor necrosis factor in rheumatoid arthritis.
Arthritis Rheum.
38
:
151
7
Feldmann, M., M. J. Elliot, J. N. Woody, R. N. Maini.
1997
. Anti-TNF-α therapy of rheumatoid arthritis.
Adv. Immunol.
64
:
283
8
Evans, C. H., S. C. Ghivizzani, R. Kang, T. Muzzonigro, M. C. Wasko, J. H. Herndon, P. D. Robbins.
1999
. Gene therapy for rheumatic diseases.
Arthritis Rheum.
43
:
1
9
Fiorentino, D. F., M. W. Bond, T. R. Mosmann.
1989
. Two types of mouse helper T cell. IV. Th2 clones secrete a factor that inhibits cytokine production by Th1 clones.
J. Exp. Med.
170
:
2081
10
Moore, K. W., A. O’Garra, R. de Waal Malafyt, P. Vieri, T. R. Mosmann.
1993
. Interleukin-10.
Annu. Rev. Immunol.
11
:
165
11
de Waal Malefyt, R., J. Abrams, B. Bennett, C. G. Figden, J. E. DeVries.
1991
. IL-10 inhibits cytokine synthesis by human monocytes: an autoregulatory role of IL-10 produced by monocytes.
J. Exp. Med.
174
:
1209
12
Bogdon, C., Y. Vodovotz, C. Nathan.
1991
. Macrophage deactivation by interleukin-10.
J. Exp. Med.
174
:
1549
13
Wang, P., P. Wu, M. I. Siegel, R. W. Egan, M. M. Billah.
1994
. IL-10 inhibits transcription of cytokine genes in human peripheral blood.
J. Immunol.
153
:
811
14
de Waal Malefyt, R. D., J. Haanen, H. Spits, M. G. Roncarolo, A. te Velde, C. Figdor, K. Johnson, R. Kastelein, H. Yssel, J. E. deVries.
1991
. Interleukin-10 (IL-10) and viral IL-10 strongly reduce antigen-specific human T cell proliferation by diminishing the antigen-presenting capacity of monocytes via downregulation of class II major histocompatibility complex expression.
J. Exp. Med.
174
:
915
15
Ding, L., P. S. Linsley, L. Y. Huang, R. N. Germain, E. M. Shevach.
1993
. IL-10 inhibits macrophage costimulatory activity by selectively inhibiting the up-regulation of B7 expression.
J. Immunol.
151
:
1224
16
Willems, F., A. Marchant, J. P. Delville, C. Gerard, A. Delvaux, T. Velu, M. de Boer, M. Goldman.
1994
. Interleukin-10 inhibits B7 and intercellular adhesion molecule-1 expression on human monocytes.
Eur. J. Immunol.
24
:
1007
17
Buelens, C., F. Williams, A. Delvaux, G. Peirand, J. P. Delville, T. Velu, M. Goldman.
1995
. Interleukin-10 differentially regulates B7-1 (CD80) and B7-2 (CD86) expression on human peripheral blood dendritic cells.
Eur. J. Immunol.
25
:
2668
18
Go, N. F., B. E. Castle, R. Barrett, R. Kastelein, W. Dang, T. R. Mosmann, K. W. Moore, M. Howard.
1990
. Interleukin-10, a novel B cell stimulatory factor: unresponsiveness of X chromosome-linked immunodeficiency B cells.
J. Exp. Med.
172
:
1625
19
Jinquan, T., C. G. Larson, B. Gressor, K. Matsushima, K. T. Pederson.
1993
. Human IL-10 is a chemoattractant for CD8+ T lymphocytes and an inhibitor of IL-8-induced CD4+ lymphocyte migration.
J. Immunol.
151
:
4545
20
Taga, K., B. Cherney, G. Tosato.
1993
. IL-10 inhibits apoptotic cell death in human T cells starved of IL-2.
Int. Immunol.
12
:
1599
21
Levy, Y., J. C. Brouet.
1994
. Interleukin-10 prevents spontaneous death of germinal center B cells by induction of the bcl-2 protein.
J. Clin. Invest.
93
:
424
22
Katsikis, P. D., C. Q. Chu, F. M. Brennan, R. N. Miani, M. Feldmann.
1994
. Immunoregulatory role of interleukin-10 in rheumatoid arthritis.
J. Exp. Med.
179
:
1517
23
Cush, J. J., J. B. Splawski, R. Thomas, J. E. McFarlin, H. Schulze-Koops, L. S. Davis, K. Fujita, P. E. Lipsky.
1995
. Elevated interleukin-10 levels in patients with rheumatoid arthritis.
Arthritis Rheum.
38
:
96
24
Moore, K. W., P. Vieira, D. F. Fiorentino, M. L. Trounstine, T. A. Khan, T. R. Mosmann.
1990
. Homology of cytokine synthesis inhibitory factor (IL-10) to the Epstein-Barr virus gene BCRF1.
Science
248
:
1230
25
Baer, R., A. T. Bankier, M. D. Biggin, P. L. Deminger, P. J. Farrell, T. J. Gibson, G. Hatfull, G. S. Hudson, S. C. Satchwell, C. Seguin, P. S. Tuffnell, B. G. Barrell.
1984
. DNA sequence and expression of the B95–8 Epstein-Barr virus genome.
Nature
310
:
1207
26
DeVries, J. E..
1995
. Immunosuppressive and anti-inflammatory properties of interleukin-10.
Ann. Med.
27
:
537
27
Drazen, K. E., L. Wu, D. Bullington, A. Shaken.
1996
. Viral IL-10 gene therapy inhibits TNF-α and IL-1β, not IL-6, in the newborn endotoxemic mouse.
J. Pediatr. Surg.
31
:
411
28
Qin, L., K. D. Chevin, Y. Ding, H. Tahara, J. P. Favarro, J. E. Woodward, T. Suzuki, P. D. Robbins, M. Y. Lotze, J. S. Bromberg.
1996
. Retrovirus-mediated transfer of viral IL-10 gene prolongs murine cardiac allograft survival.
J. Immunol.
156
:
2316
29
Suzuki, T., H. Tahara, S. Narula, K. W. Moore, P. D. Robbins, and M. T. Lotze. Viral interleukin-10 (vIL-10), the human herpes virus-4 cellular IL-10 (cIL-10) homologue, induces local anergy to allogeneic and syngeneic tumors. J. Exp. Med. 102:477.
30
Yeh, P., M. Perricaudet.
1997
. Advances in adenoviral vectors: from genetic engineering to their biology.
FASEB J.
11
:
615
31
Graham, F. L., A. J. Van der Eb.
1973
. A new technique for the assay of infectivity of human adenovirus 5 DNA.
Virology
52
:
456
32
Mittereder, N., K. L. March, B. C. Trapnell.
1996
. Evaluation of the concentration and bioactivity of adenovirus vectors for gene therapy.
J. Virol.
70
:
7498
33
Glynn, L. E..
1968
. The chronicity of inflammation and its significance in rheumatoid arthritis.
Ann. Rheum. Dis.
27
:
105
34
Taskiran, D., M. Stefanovic-Racic, H. I. Georgescu, C. H. Evans.
1994
. Nitric-oxide mediates suppression of cartilage proteoglycan synthesis by interleukin-1.
Biochem. Biophys. Res. Commun.
200
:
142
35
Farndale, R. W., D. J. Buttle, A. J. Barnette.
1986
. Improved quantitation and discrimination of sulphated glycosaminoglycans by use of dimethylmethylene blue.
Biochim. Biophys. Acta
883
:
173
36
Ghivizzani, S. C., E. R. Lechman, R. Kang, C. Tio, J. Kolls, C. H. Evans, P. D. Robbins.
1998
. Direct adenovirus-mediated gene transfer of interleukin 1 and tumor necrosis factor α soluble receptors to rabbit knees with experimental arthritis has local and distal anti-arthritic effects.
Proc. Natl. Acad. Sci. USA
95
:
4613
37
Müller-Ladner, U., C. H. Evans, B. N. Franklin, C. R. Robets, R. E. Gay, P. D. Robbins, and S. Gay. 1999. Gene transfer of cytokine inhibitors into human synovial fibroblasts in the SCID mouse model. Arthritis Rheum. In press.
38
Ma, Y., S. Thornton, L. E. Duwel, G. P. Biovin, E. H. Giannini, J. M. Leiden, J. A. Bluestone, R. Hirsch.
1998
. Inhibition of collagen-induced arthritis in mice by viral IL-10 gene transfer.
J. Immunol.
161
:
1516
39
Apparailly, F., C. Verwaerde, C. Jacquet, C. Auriault, J. Sany, C. Jorgenson.
1998
. Adenovirus-mediated transfer of viral IL-10 gene inhibits murine collagen-induced arthritis.
J. Immunol.
160
:
5213
40
Whalen, J. D., E. R. Lechman, C. A. Carlos, K. Weiss, I. Kovesdi, P. D. Robbins, C. H. Evans.
1999
. Adenoviral transfer of the viral IL-10 gene peri-articularly to mouse paws suppresses development of collagen-induced arthritis in both injected and uninjected paws.
J. Immunol.
162
:
3625
41
Qin, L., Y. Ding, D. R. Pahud, N. D. Robson, A. Shaked, J. S. Bromeberg.
1997
. Adenovirus-mediated gene transfer of viral interleukin-10 inhibits the immune response to both alloantigen and adenoviral antigen.
Hum. Gene Ther.
8
:
1365