The contralateral effect conferred by intra-articular adenovirus ... - Nature

Gene Therapy (2003) 10, 2029–2035 & 2003 Nature Publishing Group All rights reserved 0969-7128/03 $25.00 www.nature.com/gt

RESEARCH ARTICLE

The contralateral effect conferred by intra-articular adenovirus-mediated gene transfer of viral IL-10 is specific to the immunizing antigen ER Lechman, A Keravala, J Nash, S-H Kim, Z Mi and PD Robbins Department of Molecular Genetics and Biochemistry, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261, USA

We have demonstrated previously that local, adenoviralmediated gene transfer of vIL-10 to a single joint of rabbits and mice with experimental arthritis can suppress disease in both the treated and untreated contralateral joints. These therapeutic effects observed in distant untreated joints following local intra-articular gene delivery have been termed the ‘contralateral effect’. To begin to understand the underlying immunologic mechanism that confers this effect, a dualantigen model of antigen-induced arthritis (AIA) in rabbit knee joints was utilized. Rabbits were immunized against two antigens, ovalbumin and keyhole limpet hemocyanin, and AIA generated by intra-articular injection of each antigen into contralateral knees. Intra-articular adenovirus-mediated

gene transfer of vIL-10 significantly reduced intra-articular leukocytosis and cartilage matrix degradation, while preserving near normal levels of cartilage matrix synthesis within treated joints. However, no antiarthritic effect was conferred in the contralateral control joints that received only a marker gene, in contrast to the results seen in a single-antigen AIA model. These results suggest that the distant antiarthritic effects associated with local gene delivery to joints are antigen-specific, and not due to vIL-10-induced generalized immunosuppression of the animal. Gene Therapy (2003) 10, 2029–2035. doi:10.1038/ sj.gt.3302109

Keywords: rheumatoid arthritis; autoimmunity; antigen-induced arthritis; distant therapeutic effects; contralateral effect

Introduction Rheumatoid arthritis (RA) is a debilitating systemic autoimmune disease distinguished by chronic inflammation of the distal diarthriodial joints. Affected joints exhibit inflammatory cell infiltration and synovial hyperplasia that contribute to the progressive degradation of cartilage and bone. The etiology of RA is unknown and there are currently no effective pharmacological treatments for the disease. Recently, biological agents that modulate the proinflammatory activities of TNF-a and IL-1b have shown efficacy as novel antiarthritic drugs.1–4 However, arthritis therapies that employ biological agents are currently limited by possible systemic side effects (ie the occurrence and reemergence of viral and bacterial infections) as well as their exorbitant expense. Moreover, upon the termination of systemic therapy, disease pathology returns unabated. Therefore, we have been evaluating local gene transfer as a way to circumvent the inherent impediments associated with delivery of these therapeutic protein agents. We recently reported that local adenovirus-mediated gene transfer of viral IL-10 to rabbit knee joints with antigen-induced arthritis (AIA) ameliorated disease not only within the treated joints, but also significantly reduced all associated pathologies in the untreated Correspondence: Dr PD Robbins, Department of Molecular Genetics and Biochemistry, E1240 Biomedical Science tower, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261, USA Received 22 July 2002; accepted 07 March 2003

contralateral joints.5 In addition, periarticular administration of adenovirus expressing vIL-10 into one paw of mice immunized with collagen type II prevented the development of collagen-induced arthritis (CIA) in both the injected paw and nontreated ipsilateral and contralateral paws.6 This phenomenon, termed the ‘contralateral effect’, was first observed by our group following adenoviral delivery of soluble receptors for IL-1b and TNF-a in the rabbit AIA model.7 However, recent studies suggest that the contralateral effect is not confined to adenovirus gene delivery. In fact, high-titer retrovirus expressing murine IL-4 injected directly into rat ankle joints was shown to confer a therapeutic effect in contralateral paws.8 In addition, the ex vivo gene transfer of IL-1Ra and soluble TNF-a receptor either alone or in combination using genetically modified autologous synovial fibroblasts transplanted intra-articularly was shown to confer a contralateral effect in contralateral control joints, demonstrating that the remedial effects conferred to distant joints do not require direct in vivo delivery of genes.9 In addition, administration of NF-kB deoxyoligonucleotides/liposome complexes into rat knee joints, able to block TNF and IL-1 expression by macrophages and dendritic cells (DC), conferred a contralateral therapeutic effect in untreated inflamed knee joints.10 Taken together, these results suggest that the contralateral effect is not restricted to any particular delivery vector, anti-inflammatory/immunosuppressive gene product or arthritis model. The mechanism involved in conferring these distant antiarthritic effects is poorly characterized. Results from

Antigen specificity of the contralateral effect ER Lechman et al

2030

marker gene studies suggest the involvement of trafficking leukocytes in the production of the contralateral effect. Specifically, the injection of adenoviruses encoding the luciferase gene into one knee joint of a bilaterally inflamed AIA rabbit led to luciferase expression in not only the injected knee joint and draining lymph node, but also in the noninjected contralateral joint.7 No other major organs were found to contain luciferase activity, suggesting that systemic virus dissemination was not the source of this peculiar effect. Additional experiments using an Ad.eGFP vector revealed that a morphologically distinct subset of cells infected in the injected joint could traffick to the contralateral joint.5 Preliminary work has identified these cells as macrophages and DC (E Lechman, unpublished observations). However, it remains uncertain whether these cells play a direct role in the contralateral effect. Indeed, if trafficking cells are involved in conferring therapeutic effects in distant joints, it is clear that they require no direct genetic modification, since the transplantation of ex vivo gene modified synovial fibroblasts can also generate therapeutic effects in untreated contralateral joints. Since contralateral therapeutic effects can be induced within untreated joints by both in vivo and ex vivo gene transfer, it has been postulated that circulating levels of transgene product may be responsible for the observed antiarthritic effects in distant joints. Indeed, several groups observe high systemic levels of transgene products after local delivery of genes to joints.8,11,12 However, in our experiments, systemic transgene levels were either below detectable levels or too low to confer a therapeutic effect in our model systems. The present study was undertaken to determine if the contralateral effect is directly attributable to systemic levels of vIL-10 or a more complex antigen-specific immunologic mechanism. To address this question, we utilized a dual-antigen (DA) model of AIA (DA-AIA) in the rabbit. In contrast to existing models of AIA, the DAAIA model rabbits were immunized against two different antigens (ovalbumin (OVA)) and keyhole limpet hemocyanin (KLH)) with disease induced by separate injections of OVA and KLH contralateral to each other into the rear knee joint of each rabbit. We demonstrate that administration of Ad.vIL-10 could effectively block the progression of experimental arthritis by reducing leukocytic infiltration into the joint space, normalizing cartilage metabolism and moderating the degree of synovitis within treated joints. However, in contrast to the results seen with monoantigenic AIA, the normally reproducible therapeutic effect in the untreated contralateral joint was abrogated. These results suggest that the distant antiarthritic effects observed with local intraarticular gene therapy using vIL-10 are not attributable to a systemic immunosuppression resulting from circulating levels of transgene product, but is conferred by an antigen-specific downregulation of inflammation. Furthermore, these results suggest that local intraarticular gene transfer of certain anti-inflammatory agents is able to block inflammation at distant sites in an antigen-specific manner.

cDNAs encoding EBV-derived viral IL-10 and b-galactosidase (lacZ) were inserted into the E1 region with gene expression driven by the human cytomegalovirus (CMV) early promotor. High-titer virus was produced by permissive replication in the 293 human embryonic kidney cell line (American Type Culture Collection, Manassas, VA, USA) as described previously.14 Viral titers were determined by optical density at 260 nm (OD260) where 1 OD unit¼1012 viral particles.15

Animals and experimental protocol Female New Zealand White rabbits weighing approximately 5–6 lb were obtained from Myrtles Rabbitry (Thompson Station, TN, USA) and housed at the Central Animal Facility (CAF) at the University of Pittsburgh (Pittsburgh, PA, USA). Animals were allowed to acclimate 3 days prior to experimentation and were fed water and chow ad libitum. Both the KLH monoantigen AIA and DA AIA models were set up in identical fashion with the only exception being that the DA model was induced by injection of OVA and KLH contralateral to each other, whereas the KLH model utilized only bilaterally injected KLH (Figure 1). Rabbits were sensitized to OVA and KLH by a series of two intradermal injections of 5 mg each of OVA (Sigma, St Louis, MO, USA) and KLH (Sigma, St Louis, MO, USA) emulsified in the first injection in Freund’s complete adjuvant (Pierce) and Freund’s incomplete adjuvant (Pierce) in the second.16 At 2 weeks following the second injection, an acute articular arthritis was initiated in both hind knees of the rabbits by the intra-articular administration of 5 mg of OVA or KLH dissolved in 0.5 ml of saline into the right and left knee joints, respectively. At 24 h postinduction of AIA, 5 109 particles of replication-defective adenovirus encoding either vIL-10 or lacZ were suspended in 0.2 mlof 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, USA) through the patellar tendon. After manipulation of the joint to allow for ample mixing, the needle was

Materials and methods Adenovirus vectors The vectors used in this study are first-generation E1/E3 deleted type 5 replication-defective adenoviruses.13 Gene Therapy

Figure 1 Experimental protocol. A schematic representation outlining the induction and treatment regimen of both single- and dual-antigen models of AIA in rabbits.


Evaluation of arthritis To measure rates of proteoglycan synthesis, articular cartilage was shaved from the femoral condyles and weighed. Approximately 30 mg of cartilage was then incubated in 1 ml Neuman and Tytell (Life Technologies, Rockville, MD, USA) serumless medium with 40 mCi of 35 SO2 for 24 h at 371C. Afterward, the medium was 4 recovered and stored at 201C. Proteoglycans were extracted from the cartilage shavings by incubation for 48 h in 1 ml of 0.5 M NaOH at 41C with gentle agitation. Following chromatographic separation of unincorporated 35SO2 using PD-10 columns (Pharmacia, Piscat4 away, NJ, USA), radiolabeled glycosaminoglycans (GAGs) released into the culture media or recovered by alkaline extraction were quantified using scintillation counting as described previously.17 To measure 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 ml were treated with papain. Papain suspension (type III, 20 ml, 19 U/mg protein, Sigma) was added to 1 ml of buffer containing 10 mM sodium EDTA and 0.4 M sodium acetate, pH 5.2. The papain solution (100 ml) was added to lavage fluid (100 ml) and incubated overnight at 601C. Papain was inactivated by the addition of iodoactetic acid to a final concentration of 4 mM. The samples were then centrifuged at 12 000 g for 10 min. Afterward, 2 U of hyalurate lyase was added and the samples incubated at 371C overnight. Determination of sulfated GAG levels was performed by a colorimetric dye binding assay using 1,9-dimethylene blue as previously described.18 Histology and statistical analysis Rabbit knees were removed from euthanized rabbits at day 7 and tissues were fixed in 10% buffered formalin for 24 h. The fixed tissues were then imbedded in paraffin, sectioned at 5 mm and stained with hematoxylin and eosin. Sections were examined by light microscopy. All data were analyzed using the Microsoft Excel software program. Group comparisons were performed using both Student’s t-test and ANOVA.

Results Intra-articular injection of recombinant adenovirus expressing vIL-10 can induce a contralateral effect in a KLH-induced AIA model We have previously shown that administration of adenovirus expressing vIL-10 to arthritic AIA rabbit joints, induced bilaterally with OVA, resulted in a therapeutic effect in both the injected joints as well as in the contralateral untreated joints.5 Since the DA-AIA model utilizes both OVA and KLH antigens, it was first necessary to assess whether vIL-10 could also induce a contralateral effect in a monoantigenic KLH-induced AIA model of arthritis. To test the ability of vIL-10 to

inhibit the inflammatory effects of bilaterally induced KLH-AIA in the rabbit knee joint, KLH-AIA was induced in both knees of 12 rabbits. At 24 h postinduction of the model, approximately 5 109 particles of adenovirus encoding vIL-10 were injected into the left knee of six rabbits and 5 109 particles of adenovirus encoding LacZ were injected into the right knee of the same six rabbits. A control group of six rabbits received 5 109 particles of adenovirus encoding LacZ into both inflamed rear knee joints. At 3 days after adenovirus injection, both knees of each rabbit were lavaged. At 7 days postinfection, the rabbits were killed, the knees lavaged, dissected and analyzed for effects of vIL-10 transgene expression. Leukocytes in lavage fluids from each group of rabbits were counted and compared as a quantitative measure of inflammation. As shown in Figure 2, at day 3, the control group of rabbits, which were injected with Ad.lacZ into both knees, was inflamed with a mean level of infiltrating leukocytes of more than 43.7 106 per ml of recovered lavage fluid. By day 7, this level had increased to 67 106 leukocytes per ml of recovered lavage fluid. In comparison, the rabbit group injected with Ad.vIL-10 in the left knee and Ad.lacZ in the right knee showed levels of infiltrating cells into the left joint (vIL-10) that were lower than that of the control knees. Leukocyte numbers in the vIL-10 knees at day 3 averaged 26 106 infiltrating cells, a 32% reduction in infiltration. At day 7, this reduction significantly increased to 63% as the model progressed. The contralateral knees receiving Ad.lacZ also displayed a reduction in the amount of infiltration into the joint space. Rabbits injected with vIL-10 in the left knees and lacZ into the right knee displayed a 11% reduction in the mean infiltration into the contralateral joint, and by day 7 this difference increased to 44%. Expression levels of vIL-10 averaged 2.25 and 1.81 ng/ml at days 3 and 7 post adenovirus administration, respectively (data not shown). Examination of other sequelae of disease such as cartilage metabolism, TNF-a expression and histology also displayed a discernable Leukocytic Infiltration (106 cells/ml)

reinserted and the fluid aspirated. Leukocytes in recovered lavage fluids were counted by hemocytometer. Levels of vIL-10 expression in recovered lavage fluids and serum were measured by a cytokine ELISA kit (Pierce Endogen, Rockford, IL, USA).

80

60

2031

Naïve control Ad.vIL-10 Ad.LacZ contralateral Ad.LacZ control

40

20

0 Day 3

Day 7

Figure 2 Leukocytic infiltration in KLH-AIA rabbit joints after intraarticular Ad.vIL-10 delivery. At 1 day postinduction of KLH-AIA, one group of six rabbits was injected in the left knee (KLH-induced) with 5 109 particles of Ad.vIL-10 and in the right knee (KLH-induced) with 5 109 particles of Ad.lacZ. A control group of six rabbits was injected with 5 109 particles of Ad.lacZ into both the right and left knee joints (both joints induced with KLH). At days 3 and 7 after injection of the virus, the knees were lavaged and numbers of leukocytes determined with a hemocytometer. Values shown represent the mean7s.e.m. of six rabbit joints. Naı¨ve control leukocyte levels averaged in the range of 104 cells/ml and are not represented on the graph. Asterisks denote values that differ at Po0.05 (t-test and ANOVA). Gene Therapy


contralateral effect (data not shown). These results demonstrate that although KLH-induced disease is more severe than OVA-induced disease, the intra-articular injection of Ad.vIL-10 is able to confer a contralateral therapeutic effect, similar to previous results in the monoantigenic OVA model of AIA.5

vIL-10 expression in the DA-AIA model Since administration of adenovirus expressing vIL-10 could induce a contralateral effect in both single-antigen OVA and KLH models of AIA, we were interested in determining whether Ad.vIL-10 could induce this same contralateral effect when bilateral joint inflammation was induced with dissimilar antigens. To test the ability of vIL-10 to inhibit the destructive pathology of DA-AIA in the rabbit knee joint, arthritis was induced by injection of OVA and KLH into the right and left knee joints, respectively, of 14 rabbits. At 24 h post model induction, approximately 5 109 particles of adenovirus encoding vIL-10 were injected into the left knee of seven rabbits and 5 109 particles of adenovirus encoding LacZ were injected into the right knee of the same seven rabbits. A control group of seven rabbits received 5 109 particles of adenovirus encoding LacZ into both knees. At 3 days after injection of the adenovirus, both knees of each rabbit were lavaged with 1 ml saline. At 7 days postinfection, the rabbits were killed and evaluated. ELISA measurements of vIL-10 levels in recovered lavage fluids detected approximately 1.95 and 5.75 ng/ ml at days 3 and 7, respectively, in knees that received Ad.vIL-10 (Figure 3). Levels of vIL-10 were not detectable in sera or knees contralateral to those knees receiving the adenovirus encoding vIL-10 (data not shown). Effect of vIL-10 expression on DA-AIA joint inflammation As a measure of inflammation, leukocytes in recovered lavage fluids from each rabbit group were counted. As

vIL-10 ng/ml Lavage Fluid

10

8

OVA/Ad.Lac Z KLH/Ad.Lac Z OVA/Ad.Lac Z KLH/Ad.vIL-10

6

4

2

0 Day 3

Day 7

Figure 3 In vivo expression of viral IL-10 after direct adenoviral delivery to rabbit knees with DA-AIA. At 24 h after induction of DA-AIA, a group of seven rabbits was injected in the left (KLH-induced) knee with 5 109 particles of Ad.vIL-10 and in the right (OVA-induced) knee with 5 109 particles of Ad.lacZ. A control group of seven rabbits was administered 5 109 particles of Ad.lacZ into both the left (KLH-induced) and right (OVA-induced) joints. 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 viral IL-10. Gene Therapy

shown in Figure 4, at day 3, the control group of rabbits injected with Ad.lacZ into both knees were inflamed with a mean level of infiltrating leukocytes exceeding 26 106 per ml of recovered lavage fluid for the left (KLH) induced joints and 27 106 per ml of lavage fluid for the right (OVA) induced joints. In contrast to the OVA-induced joints, it was interesting to note a more rapid disease progression in the KLH joints. By day 7, this level had increased to 28 106 leukocytes per ml of recovered lavage fluid in the left (KLH) induced joints and 22.9 106 in the OVA-induced right joints. In comparison, in the rabbit group injected with Ad.vIL10 in the left (KLH) induced knee and Ad.lacZ in the right (OVA) induced knee, the levels of infiltrating cells into the right knee (vIL-10) were significantly lower than that of the control knees. The level of leukocytic infiltration in the vIL-10 knees at day 3 averaged 32.3 106 cells. By day 7, the therapeutic effect became evident with an 80% reduction in cell number compared to the control KLH joints. Perhaps the most interesting observation was seen in the contralateral knees receiving Ad.lacZ. Rabbits injected with vIL-10 in the right knees and lacZ into the left displayed no transfer of the therapeutic effects to contralateral untreated joints. It is important to note that analysis of the cell types in the cytospins of lavage fluids from the Ad.lacZ-treated knees did not reveal a reduction in any specific cell types. Similarly, the reduction in leukocytic infiltrate observed in the Ad.vIL-10-treated joint was not due to a reduction in a specific cell type.

Effect of vIL-10 expression on rates of cartilage breakdown In order to monitor cartilage matrix degradation in the rabbit knees due to disease progression, GAGs released into synovial fluid as a result of proteoglycan breakdown were measured in recovered lavage fluids (Figure 5a). The control group of rabbits receiving injections of Ad.lacZ in both knees had very high levels of GAGs in the lavage fluids of the KLH-induced knees at both days 3 and 7. Knees induced with OVA showed moderate levels of GAGs released at day 3. By day 7, GAG levels in OVA control joints increased to levels comparable with Leukocytic Infiltration (106 cells/ml)

2032

60 50 40

Naïve Control OVA/Ad.Lac Z KLH/Ad.Lac Z OVA/Ad.Lac Z KLH/Ad.vIL-10

*

30 20 10 0

Day 3

Day 7

Figure 4 Leukocytic infiltration in DA-AIA knees after intra-articular injection of Ad.vIL-10 and Ad.lacZ. Experimental conditions were as described in Figure 2. At days 3 and 7 after adenovirus injection, the joints were lavaged and numbers of leukocytes determined with a hemocytometer. Values shown represent the mean7s.e.m. of six rabbit joints for the naı¨ve controls, seven rabbit knees per experimental group. Converging bars represent joints that are contralateral to each other. Asterisks denote values that differ at Po0.05 (t-test and ANOVA).


2033

a 40 GAG Release (µg/ml)

* 30

20

10

0 Day 3

Day 7

GAG Synthesis (35S counts/mg)

b *

8000

Naïve Control OVA/Ad.Lac Z KLH/Ad.Lac Z OVA/Ad.Lac Z KLH/Ad.IL-10

6000

4000

2000

0 Day 7

Figure 5 Effect of vIL-10 on cartilage matrix metabolism. Experimental conditions were as described in Figure 2. Concentrations of GAG released into the lavage fluids (a) were measured spectrophotometrically as an index of proteoglycan breakdown. To measure rates of proteoglycan synthesis (b) pieces of articular cartilage were shaved from the femoral condyles and into macromolecular material their in vitro incorporation of 35SO2 4 measured. Values given are the means þ s.e.m. Numbers of knees per group are as described in Figure 2. Converging bars represent joints that are contralateral to each other. Asterisks denote values that differ at Po0.05 (t-test and ANOVA).

those of the control KLH knees. Rabbits injected with Ad.vIL-10 in the left knee (KLH) and Ad.lacZ in the opposite joint (OVA) showed reductions in the levels of GAGs released in only the knees receiving vIL-10. At day 3, knees receiving Ad.vIL-10 displayed no reduction in the amount of GAGs released into the joint space. This changed to a 43% reduction by day 7. Once again, the Ad.lacZ control joints contralateral to those that received vIL-10 displayed no reduction in GAGs released at days 3 and 7, respectively. The rates of GAG synthesis by condylar cartilage (Figure 5b) were strongly suppressed after 7 days of DAAIA in KLH-induced control rabbit knees receiving Ad.lacZ. Compared to naı¨ve controls, the rates of GAG synthesis were depressed by 48% in KLH-induced knees of arthritic control rabbits and 26% in OVA-induced joints that received only the Ad.lacZ virus into both joints. In contrast, within KLH-induced knees that received Ad.vIL-10, GAG synthesis rates were protected to a moderate degree with levels only 19% lower than those of naı¨ve joints. Contralateral joints treated with Ad.lacZ showed levels of GAG synthesis 16% lower than naı¨ve control joints.

Histological analysis The histological analyses of tissue recovered from the knees of each group of rabbits are shown in Figure 6. Compared to tissue recovered from normal nondiseased rabbits (Figure 6a), sections from the Ad.lacZ (KLHinduced) control group displayed a severe synovitis typical of that seen with AIA (Figure 6c). The synovium appeared thickened, fibrous, hyperplastic and hypertrophic due to resident synovial cell proliferation and infiltration by mononuclear leukocytes. Although somewhat variable within this group, Ad.vIL-10-treated KLHinduced knees (Figure 6b) displayed a considerable reduction in the amount of synovitis. OVA-induced control joints (Figure 6e) that received Ad.lacZ were histologically indistinguishable from OVA-induced joints contralateral to Ad.vIL-10 joints (Figure 6d).

Discussion In this study, we investigated whether the distal therapeutic effects observed in local arthritis gene therapy are directly attributable to systemic levels of circulating transgene product. To address this issue, we utilized a DA model of AIA (DA-AIA). In contrast to existing models of Ag-induced arthritis, the animals in the DA-AIA model are immunized against two different antigens (OVA and KLH) with disease induced by injections of OVA and KLH contralateral to each other into the rear knee joint of each rabbit. The administration of 5 109 particles of Ad.vIL-10 resulted in a significant reduction of several indices of DA-AIA disease in the treated KLH-induced joints. However, in contrast to several previous results within monoantigenic AIA models, the therapeutic effect in the contralateral joint was not observed. As a control, we further demonstrated that local gene transfer of Ad.vIL-10 was able to confer a contralateral effect in the monoantigen KLH model, similar to our previous monoantigen study with OVA.5 These results strongly suggest that polyarticular therapeutic effects seen with local gene therapy are not due to circulating levels of transgene product conferring systemic, nonspecific immunosuppression, but instead argue for a more complex antigen-specific mechanism. The contralateral effect was first observed following injection of adenoviral vectors expressing IL-1and sTNF soluble receptors7 and has been more recently observed following intra-articular gene transfer of IL-4,19 FasL,20 IL-1Ra,9 and NF-kB decoy oligonucleotides.10 Hypothetically, the contralateral effect could be conferred by several mechanisms including leakage of virus from the injected joint, generalized immunosuppression due to overexpression of transgene product, migration of genetically modified or functionally altered cells or through a neurogenic pathway. Previous studies examining adenoviral distribution following intra-articular injection of Ad.luciferase and Ad.eGFP vectors suggest that little or no virus leaves the joint after intra-articular injection, effectively negating systemic viral dispersal as a likely mechanism5,7 However, we did observe that a distinct subset of genetically modified cells could be detected in the draining lymph node of the injected joint as well as in the contralateral joint at 3 and, in particular, 7 days postinjection.5 Analysis of the cells within the synovial fluid infected following adenoviral injection Gene Therapy


2034

Figure 6 Histological analysis of synovial tissue recovered from rabbit knees. AIA was induced in both knees of two groups of seven rabbits. At 24 h after induction of DA-AIA, a group of seven rabbits was injected in the left (KLH-induced) knee with 5 109 particles of Ad.vIL-10 and in the right (OVA-induced) knee with 5 109 particles of Ad.lacZ. A control group of seven rabbits was administered 5 109 particles of Ad.lacZ into both the left (KLH-induced) and right (OVA-induced) joints. At 7 days after virus injection, synovial tissue was harvested, sectioned and stained with hematoxylin and eosin. (a) Synovium from a normal naı¨ve rabbit joint. (c) Synovium from KLH-induced rabbit joints that received Ad.lacZ. Relative to naı¨ve joints (a) the synovium is thicker, more fibrous, hyperplastic and infiltrated with mononuclear cells. (b) Tissue recovered from KLH-induced rabbit knees treated with Ad.vIL-10. (d) Synovial tissue from OVA-induced knees injected with Ad.lacZ and contralateral to those that were treated with vIL-10. (e) Synovium from control OVAinduced joints receiving only an adenovirus expressing the b-galactosidase marker gene.

suggests that primarily MHCII/CD11b þ and to a lesser extent CD11c þ cells, representing macrophages and DC respectively, are infected (E Lechman, unpublished observations). These observations suggested that the Gene Therapy

genetically modified cells are able to traffick to sites of inflammation as well as the lymph node to modulate the immune response. However, we have recently demonstrated that intra-articular injection of ex vivo modified synovial fibroblasts expressing IL-1Ra and sTNF-R also conferred a contralateral effect.9 Since synovial fibroblasts show little propensity to traffick, this result suggested that either low systemic levels of IL-1Ra and sTNF-R are able to confer the contralateral effect or that local gene expression could functionally alter a subset of cells that are able to traffick to distant joints. In this report, we have demonstrated using the DA-AIA model that the contralateral effect is antigen dependent, eliminating global immunosuppression by systemic levels of proteins as the major mediator of these distant antiarthritic effects. However, it is important to note that low levels of anti-inflammatory proteins present in the circulation may nevertheless contribute to the overall magnitude of the contralateral effect. Based on our previous results and the results reported here, we propose a model where local intra-articular gene delivery is able to functionally alter the activity of a subset of immunoregulatory cells that in turn can downmodulate the immune response at distant sites in an antigen-specific manner. In support of this model, we have recently demonstrated that local injection of Ad.vIL10 into inflamed mouse paws could generate antigenpresenting cells (APC) that could inhibit local and distal DTH reactions in recipient mice following adoptive transfer.21 The anti-inflammatory effect of the adoptively transferred APC was only mediated when the donor mice had been sensitized to the antigen used to incite the DTH response in the recipient mice. Moreover, bone marrow derived DC cultured in the presence of recombinant IL-10 or infected ex vivo with Ad.vIL-10 were shown to suppress DTH reactions by adoptive transfer, but only when pulsed with the inciting antigen.21 Together, these data support the role of antigen-primed APC in the downregulation of both local and distant joint inflammation. In addition, we have recently demonstrated that DCderived exosomes, 60–90 nm particles originating from the multivesicular compartment of cells, were able to suppress DTH in both injected paws and contralateral saline-injected paws (Lechman et al, manuscript in preparation). Together, these data support a role for APC and APC-derived vesicles in the suppression of contralateral inflammation. In support of the role of APC in conferring the therapeutic, antiarthritic effect is the observation that i.v. injection of DC genetically engineered to express IL-422 or FasL23 is able to suppress established collagen-induced arthritis in the mouse. Moreover, in preliminary experiments, the systemic injection of DC genetically engineered to express IL-1 and TNF inhibitors or treated with NF-kB decoys also are able to reduce established CIA in the mouse, although not to the same extent as DC-IL-4. These data strongly argue for an antigen-specific, APC-mediated mechanism in the systemic dispersal of antiarthritic effects following intraarticular gene transfer. Little is known about the role of trafficking cells in the systemic spread of therapeutic effects. Recently, however, a role for trafficking cells in the pathogenesis of RA was observed in the human/SCID mouse arthritis model. The human/SCID model of arthritis is induced by unilateral implantation of human RA synovium into the knee joints


of SCID mice. Unilateral implantation was found to develop into a bilateral knee joint disorder, often involving homo- and contralateral hips.24 No migration of the human cells was observed in secondarily affected joints suggesting that bilateral joint disease was not due to the spreading of human cells.24 This effect was also induced in RAG/ mice suggesting no involvement of T and B cells.24 Finally, extra-articular engraftment of human RA synovial tissue under the skin also induced an oligoarticular pattern of arthritis, suggesting no direct neurological transmission of disease.24 Large numbers of murine DC were found in the arthritic joint with murine macrophages near the front of cartilage destruction, suggesting a role for APC in both the polyarticular dispersal and destructive nature of the disease.24 Consistent with these results is our observation that intraarticular adenoviral-mediated gene transfer of TNF was able to initiate disease pathology in not only the treated joint, but also to a low level in distal joints (Lechman et al, manuscript in preparation). RA is an autoimmune disease with disease pathologies manifested in multiple joints. Although certain candidate autoantigens have been identified, there is no clear predominant common antigen that could be used to specifically tolerize against in order to eliminate disease. Results reported here suggest that local expression of certain anti-inflammatory agents within one inflamed joint is able to confer an antigen-specific therapeutic effect to distant untreated joints. Thus it may be possible to inhibit effectively autoantigen-specific immune responses by intra-articular gene transfer of certain antiinflammatory agents without the need to identify the autoantigen. Moreover, results in murine models of arthritis suggest that intra-articular expression of vIL-10 is required for only a transient period to permanently block disease progression.6

Acknowledgements We would like to thank Steve Ghivizzani, Tom Oligino and Chris Evans for helpful discussions. We would especially like to thank Jeff Mai for his technical expertise. The work was supported in part by Grants AR-6-2225 and DK44935 from the National Institutes of Health.

References 1 Bresnihan B et al. Treatment of rheumatoid arthritis with recombinant human interelukin-1 receptor antagonist. Arthritis Rheum 1998; 41: 2196–2204. 2 Jiang Y et al. A multicenter, double-blind, dose-ranging, randomized, placebo-controlled study of recombinant human interleukin-1 receptor antagonist in patients with rheumatoid arthritis: radiologic progression and correlation of Genant and Larson scores. Arthritis Rheum 2000; 43: 1001–1009. 3 Moreland LW et al. Treatment of rheumatoid arthritis with a recombinant human tumor necrosis factor receptor (p75)-Fc fusion protein. N Engl J Med 1997; 337: 141–147. 4 Elliott MJ et al. Randomized double-blind comparison of chimeric monoclonal antibody to tumour necrosis factor [alpha](cA2) versus placebo in rheumatoid arthritis. Lancet 1994; 244: 1105–1110.

2035 5 Lechman ER et al. Direct adenoviral gene transfer of viral IL-10 to rabbit knees with experimental arthritis ameliorates disease in both injected and contralateral control knees. J Immunol 1999; 163: 2202–2208. 6 Whalen JD et al. Adenoviral transfer of the viral IL-10 gene periarticularly to mouse paws suppresses development of collagen-induced arthritis in both injected and uninjected paws. J Immunol 1999; 162: 3625–3632. 7 Ghivizzani SC et al . Direct adenovirus-mediated gene transfer of interleukin 1 and tumor necrosis factor a soluble receptors to rabbit knees with experimental arthritis has local and distal antiarthritic effects. Proc Natl Acad Sci USA 1998; 95: 4613–4618. 8 Boyle DL et al. Intra-articular IL-4 gene therapy in arthritis: antiinflammatory effect and enhanced th2 activity. Gene Therapy 1999; 6: 1911–1918. 9 Kim SH et al. Ex vivo gene delivery of IL-1Ra and soluble TNF receptor confers a distal synergistic therapeutic effect in antigeninduced arthritis. Mol Ther 2002; 6: 591–600. 10 Miagkov AV et al. NF-kappaB activation provides the potential link between inflammation and hyperplasia in the arthritic joint. Proc Natl Acad Sci USA 1998; 95: 13859–13864. 11 Watanabe S et al. On the mechanism of protection of distal joints after local gene transfer in collagen-induced arthritis. Hum Gene Ther 2000; 11: 751–758. 12 Lubbert E et al. Intra-articular IL-10 gene transfer regulates the expression of collagen-induced arthritis (CIA) in the knee and ipsilateral paw. Clin Exp Immunol 2000; 120: 375–383. 13 Yeh P, Perricaudet M. Advances in adenoviral vectors: from genetic engineering to their biology. FASEB J 1997; 11: 615–623. 14 Graham FL, Van der Eb AJ. A new technique for the assay of infectivity of human adenovirus 5 DNA. Virology 1973; 52: 456–467. 15 Mittereder N, March KL, Trapnell BC. Evaluation of the concentration and bioactivity of adenovirus vectors for gene therapy. J Virol 1996; 70: 7498–7509. 16 Glynn LE. The chronicity of inflammation and its significance in rheumatoid arthritis. Ann Rheum Dis 1968; 27: 105–121. 17 Taskiran D, Stefanovic-Racic M, Georgescu HI, Evans CH. Nitric-oxide mediates suppression of cartilage proteoglycan synthesis by interelukin-1. Biochem Biophys Res Commun 1994; 200: 142–148. 18 Farndale RW, Buttle DJ, Barnette AJ. Improved quantitation and discrimination of sulphated glycosaminoglycans by use of dimethylene blue. Biochem Biophys Acta 1986; 883: 173–177. 19 Kim SH et al. Gene therapy for established murine collageninduced arthritis by local and systemic adenovirus-mediated delivery of interleukin-4. Arthritis Res 2000; 2: 293–302. 20 Zhang H et al. Amelioration of collagen-induced arthritis by CD95 (Apo-1/Fas)-ligand gene transfer. J Clin Invest 1997; 100: 1951–1957. 21 Whalen JD et al. Viral IL-10 gene transfer inhibits DTH responses to soluble antigens: evidence for involvement of genetically modified dendritic cells and macrophages. Mol Ther 2001; 4: 543–550. 22 Kim S-H et al. Effective treatment of established murine collagen-induced arthritis by systemic administration of dendritic cells genetically modified to express IL-4. J Immunol 2001; 166: 3499–3505. 23 Kim SH, Kim S, Oligino TJ, Robbins PD. Effective treatment of established mouse collagen-induced arthritis by systemic administration of dendritic cells genetically modified to express FasL. 2002; 6: 584–590. 24 Sack U et al. Systemic characteristics of chronic arthritis induced by transfer of human rheumatoid synovial membrane into SCID mice (human/murine SCID arthritis). J Autoimmun 1999; 13: 335–346.

Gene Therapy