A comparative study on intra-articular versus systemic gene electrotransfer in experimental arthritis

THE JOURNAL OF GENE MEDICINE RESEARCH ARTICLE J Gene Med 2006; 8: 1027–1036. Published online 30 May 2006 in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/jgm.922

A comparative study on intra-articular versus systemic gene electrotransfer in experimental arthritis M. Khoury1,2 P. Bigey3,4,5,6 P. Louis-Plence1,2 D. Noel1,2 H. Rhinn3,4,5,6 D. Scherman3,4,5,6 C. Jorgensen1,2,7 F. Apparailly1,2* 1

Inserm, U 475, F-34000 Montpellier, France 2

Universit´e Montpellier1, UFR de M´edecine, F-34000 Montpellier, France 3 Inserm, U 640, F-75006 Paris, France 4

CNRS, UMR8151, F-75006 Paris, France 5

Universit´e Paris Descartes, Facult´e de Pharmacie, Chemical and Genetic Pharmacology Laboratory, F-75270 Paris, France 6

Ecole Nationale Sup´erieure de Chimie de Paris, F-75005 Paris, France

7

Unit´e Clinique d’Immuno-rhumatologie: Th´erapeutique des maladies articulaires et osseuses, CHU Lapeyronie, F-34000 Montpellier, France *Correspondence to: F. Apparailly, INSERM U475, 99 rue Puech Villa, 34197 Montpellier cedex 5, France. E-mail: [email protected]

Abstract Background Electric pulse mediated gene transfer has been applied successfully in vivo for increasing naked DNA administration in various tissues. To achieve non-viral gene transfer into arthritic joint tissue, we investigated the use of electrotransfer (ET). Because anti-inflammatory cytokine strategies have proven efficient in experimental models of arthritis, we compared the therapeutic efficiency of local versus systemic delivery of the interleukin-10 (IL-10) using in vivo ET. Methods A plasmid vector expressing IL-10 was transferred into DBA/1 mouse knee joints by ET with 12 pulses of variable duration and voltage. The kinetics of transgene expression were analyzed by specific enzyme-linked immunosorbent assay (ELISA) in sera and knees. Optimal conditions were then used to deliver increasing amounts of IL-10 plasmid intra-articularly (i.a.) in the collagen-induced arthritis (CIA) mouse model. The therapeutic efficiency was compared with the potency of intra-muscular (i.m.) ET. Results Following i.a. ET, local IL-10 secretion peaked on day 7 and dropped 2 weeks after. A second ET produced the same kinetics without enhancing gene transfer efficiency, while transgene was still detected in injected muscles 4 weeks after ET. Only the i.m. ET of 25 µg of IL-10 significantly inhibited all the clinical and biological features of arthritis. The i.a. ET only showed mild improvement of arthritis when 100 µg of IL-10 plasmid were electrotransfered weekly from day 18 following arthritis induction. Conclusions The present results suggest that gene transfer into arthritic joints by ET is an effective means to deliver anti-inflammatory cytokines. However, short duration of transgene expression impedes a significant effect for the treatment of arthritis, making i.m. ET more potent than i.a. ET for clinical benefit in CIA. Copyright  2006 John Wiley & Sons, Ltd. Keywords rheumatoid arthritis; interleukin-10; electrotransfer; intra-articular; intra-muscular

Introduction

Received: 28 June 2005 Revised: 7 March 2006 Accepted: 8 March 2006

Copyright  2006 John Wiley & Sons, Ltd.

A large body of work has focused on using gene therapy in animal models of rheumatoid arthritis (RA), and results convincingly support the fact that this therapeutic approach can be an advantageous strategy in the treatment of inflammatory and destructive joint diseases such as RA [1]. Although RA is an autoimmune disease with unknown aetiology, gene therapy approaches appear to hold promise for specific suppression of the immune system

1028

targets and long-term expression of therapeutic genes, replacing the frequent administration of recombinant proteins as anti-TNF biotherapies. Specific aims for successful gene therapy in RA include the design of safe, efficient, long-lasting and specific techniques for therapeutic molecule delivery, with minimal adverse effects. Viral vectors are very efficient in transferring genes into mammalian cells both ex vivo and in vivo. The use of viruses is a powerful technique, because many of them have evolved a specific machinery to deliver DNA or RNA into cells. However, the immune response against viral proteins may reduce the efficacy of viral vectors. There are also several recurring issues that have led to a reconsideration of their use in human clinical trials such as the ability of some viral vectors to integrate the host genome and permanently alter its genetic structure, and the capability to self-replicate with the possibility of recombination and complement activation. On the other hand, gene therapy protocols for the treatment of RA are complicated by the perception that the disease is non-lethal. Taking these considerations into account, the design of non-viral vectors is receiving increasing interest since it could provide alternative tools with low toxicity and low immunogenicity that are inexpensive, easy to handle and have long-term effects. The simplest approach is naked DNA injection into the targeted tissue or systemic circulation. Physical (gene gun, hydrostatic pressure, ultrasound, electroporation) and chemical (polymers or cationic lipids) methods have been applied to improve the well-known poor efficacy of gene transfer by plasmid DNA [2]. Recently, in vivo electrotransfer (ET) has been shown to be an efficient method for increasing naked DNA administration into skeletal muscle [3,4], liver [5], skin [6], cornea [7], blood vessels [8], and tumors [9,10]. Thus, plasmid DNA again appears as an attractive technique for non-viral gene delivery of therapeutic genes. Recent studies have demonstrated that ET is a feasible approach to deliver reporter DNA into normal rat joint tissue [11,12], and small interfering RNA into arthritic mice joints [13]. Since we have previously reported that ET of the antiinflammatory interleukin-10 (IL-10) cytokine into the tibialis anterior muscle of arthritic mice resulted in a beneficial therapeutic outcome [14], we investigated whether the local production of a therapeutic gene using in vivo ET would have beneficial effects on experimental arthritis. The present study aimed at evaluating the therapeutic efficiency of such an approach for local non-viral gene transfer in a mouse model of arthritis, and at comparing its effect with that of systemic ET strategy.

Materials and methods Plasmid preparation The pVax2 plasmid consists of the pVax1 plasmid (InVitrogen) in which the promoter has been replaced by Copyright  2006 John Wiley & Sons, Ltd.

M. Khoury et al.

the cytomegalovirus (CMV) promoter contained in the pCMVβ (Clontech). The pCMVβ plasmid was digested with EcoRI, then blunt ended by the Klenow fragment, and finally digested by BamHI. A resulting 629 bp fragment was purified after agarose gel electrophoresis. This promoter was ligated into the HincII–BamHI pVax1 fragment to give pVax2. The pGEM-7Zf containing the human IL-10 cDNA, kindly provided by Dr. C. Verwaerde (Institut Pasteur, Lille, France), was digested by XbaI and EcoRI. The resulting 642 bp fragment was purified after agarose gel electrophoresis and ligated into a SpeI-EcoRI digested pVax2 to give rise to the recombinant pVax2hIL10. The pVax2-hIL10 was amplified in DH5α E. coli strain, purified by caesium chloride (CsCl) ultracentrifugation, and dialyzed three times against 4 l 150 mM NaCl.

Animal studies Male DBA/1 mice (Harlan, France) were bred in our facilities and used at the age of 8–10 weeks. Bovine type II collagen (bCII) (Sigma-Aldrich, l’Isle d’Abeau, France) was dissolved in 0.05 M acetic acid at the concentration of 2 mg/ml by stirring overnight at 4 ◦ C and emulsified with an equal volume of Freund’s complete adjuvant (Perbio Science, Bezons, France). Collagen-induced arthritis (CIA) was induced by intra-dermal injection at the base of the tail with 100 µg bCII at day 0. On day 21 after priming, the mice received a second intra-dermal booster injection with 100 µg bCII in Freund’s incomplete adjuvant (Perbio Science), as previously reported [15]. Paw thickness was measured over time with a Mitutoyo micrometer (Sigma) and arthritis was scored by macroscopic examination. Clinical score was assessed using the following system: grade 0, no swelling; grade 1, slight swelling and erythema with >0.1 mm increase of paw swelling; grade 2, swelling and erythema with >0.15 mm paw swelling increase; grade 3, extensive swelling (>0.25 mm) and erythema with severe joint deformity or ankylosis; grade 4, pronounced swelling (>0.45 mm) with pronounced joint deformity or ankylosis. Each limb was graded, resulting in a maximal score of 16 per animal. Serum samples were obtained by retro-orbital puncture on anesthetized animals at various time points before and after pVax injection, and stored at −20 ◦ C until tested. The hind paws were collected for radiography, then fixed in 4% paraformaldehyde (Sigma), decalcified (Eurobio France), embedded in paraffin, and 5 µm sections were stained with hematoxylin/eosin/safranin O. X-rays were performed on Kodak films and paw sections were scored for overall histopathological analysis as previously described [16]. Severe cases of arthritis were considered for radiological or histological scores over 2. Experiments with animals were conducted following the recommendations of the NIH and European Convention for the Protection of Vertebrate Animals used for Experimentation. J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm

1029

Gene Electrotransfer in Arthritis

Electric pulses delivery When indicated, mice were randomized and anaesthetized by intra-peritoneal injection of a ketamine (30 mg/kg) and xylazine (10 mg/kg) solution. The knees were shaved, and indicated doses of pCMVβ expressing β-galactosidase, pVax2-hIL10 or empty plasmid were injected intra-articularly (i.a.) in 10 µl of 0.9% NaCl into both knee joints, or intra-muscularly (i.m.) in 30 µl into the tibialis anterior (TA) muscle, by using a Hamilton syringe with a 30-gauge needle (NH-BIO, Massy, France). Following i.a. injection, transcutaneous electric pulses were applied through two custom-made stainless steel tweezer electrodes placed on either side of the knee. Square wave electric pulses were generated by an ECM830 electropulsator (BTX, San Diego, CA, USA) with indicated output voltages, pulse lengths and numbers, and a frequency of pulse delivery of 1 Hz. Electric-field strengths (in V/cm) are reported as the ratio of applied voltage to the distance between electrodes. After i.m. injection, transcutaneous electric pulses were applied using two stainless steel plate electrodes placed on either side of the hind limb as previously described [14,17]. Briefly, eight square wave electric pulses of 20 ms length, and an output voltage of 200 V/cm, were generated.

Measurement of gene expression by enzyme-linked immunosorbent assay (ELISA) At the indicated times after DNA electroinjection, mice were sacrificed under anesthesia, the TA muscles were dissected out and rapidly frozen in liquid nitrogen for further analysis. Muscle extracts were prepared by homogenizing the muscle tissue with an Ultra-Turrax homogenizer (Bioblock Scientific, Illkirch, France) in 500 µl of lysis buffer (Tris-phosphate, 1 mM dithiothreitol, 1 mM EDTA, 15% glycerol, 8 mM MgCl2 , 0.2% Triton X-100) containing protease inhibitors (Sigma). Extracts were frozen/thawed twice and immediately cleared by centrifugation (5000 g for 15 min at 4 ◦ C). The total protein concentration was determined for each homogenate using bovine serum albumin (BSA) as a standard in the BCA protein kit (Pierce, Bezons, France). For i.a. gene expression, the left and right knee joints were collected at the indicated time and incubated for 24 h in 200 µl RPMI. Supernatants and knees were stored at −20 ◦ C until tested. The level of transgene expression was measured in supernatants from muscle or knee extracts and knee-conditioned medium by a specific hIL-10 Ready-setgo ELISA kit (CliniSciences, Montrouge, France) with a detection limit of 2 pg/ml. Data are expressed as mean of hIL-10 levels measured in pVax2-hIL10-electrotransfered mice, normalized by subtracting the background noise from the corresponding tissue of NaCl-electrotransfered animals. To achieve prokaryotic β-galactosidase expression following 25 µg of pCMVβ i.a. ET, injected and Copyright  2006 John Wiley & Sons, Ltd.

control knee joints were processed as previously described [18].

RNA isolation and quantitative reverse-transcription polymer chain reaction (Q-PCR) Total RNA was extracted using the RNeasy mini kit (Qiagen) from homogenized knee joints using a UltraTurrax homogenizer and quantified by spectrophotometry (Eppendorf, France). After treatment with DNAse, 1 µg of total RNA was reverse-transcribed using the Multiscribe reverse transcriptase (Applied Biosystems, Courtaboeuf, France). The assays-on-demand primers specific for human IL-10 and mouse GAPDH, and the Taq Man Universal Master Mix, were used according to the manufacturer’s recommendations (Applied Biosystems). Measurement and analysis of gene expression were performed using the ABI Prism 7000 sequence detection system software. Content of cDNA samples was normalized by subtracting the number of copies of the endogenous gylceraldehyde-3-phosphate dehydrogenase (GAPDH) reference gene to the target gene (Ct = Ct of target gene – Ct of GAPDH).

Proliferation assay and cytokine production profiles Spleens and draining lymph nodes were collected at sacrifice and cultured at 2 × 106 cells/well as previously published [19]. Supernatants were harvested for murine IFN-γ and IL-4 quantification using specific ELISA (CliniSciences).

Statistics Statistical analysis was performed using Fisher’s exact test for contingency group comparisons, and unpaired t test or Mann-Whitney test as appropriate according to data distribution. The Chi-2 test was used for percentage comparisons (disease incidence and severity). All data were analyzed by the program Instat2.1.

Results Optimizing parameters for ET into mice joints The feasibility of using ET for gene delivery into joint tissues has been previously reported [11–13]. The optimum efficiencies were achieved at 200–250 V/cm. Thus, in order to determine the optimum parameters in our conditions, experiments were designed using different voltages and pulse durations, starting from the parameters previously published. Electric pulses were applied to J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm

1030

M. Khoury et al. A

**

8

40

6 4 2 0 200

250 300 Voltage (V/cm)

350

25

30 20 10

B 30 hIL-10 (pg/ml)

50

10

hIL-10 (pg/ml)

hIL-10 (pg/ml)

A 12

250 V/cm 300 V/cm

20

***

0 Arthritic

**

15

B

10

Control

Non arthritic pCMVß

5 0 10

20 Pulse duration (ms)

30

Figure 1. Effect of voltage and pulse duration on intra-articular electrotransfer efficiency. Mice were injected in the knee joints with 25 µg of pVax-hIL10 and electroporated as indicated in Materials and methods. The cytokine production was measured in the conditioned medium from patellas by hIL-10-specific enzyme-linked immunosorbent assay (ELISA) and levels are normalized with the corresponding saline-electrotransfered controls: (A) 12 pulses of 20 ms and various voltages (200–350 V/cm) were applied; (B) 12 pulses of 10 to 30 ms and 250 (black bars) or 300 (hatch bars) V/cm were applied to the injected knees. Bars show the mean and SEM results representative of three separate experiments. ∗∗ p < 0.01; ∗∗∗ p < 0.001 (n = 8 mice/group)

knee joints of DBA/1 mice whose articular cavities were injected with 25 µg of plasmid expressing the human IL-10 (pVax2-hIL10). Figure 1 shows the level of hIL10 secretion by knee joints collected 3 days after gene transfer in all ET conditions tested. Although the amounts of hIL-10 detected in knee-conditioned media were low, the sensitivity of the ELISA kit allowed detection of a significantly higher transgene expression secreted by the joint tissue when applying pulses of 250 V/cm and 20 ms, compared with other tested conditions (Figure 1A) (p < 0.0001). Pulses of longer duration (30 ms) and higher voltage (300 V/cm) achieved similar efficiency, with no significant improvement (Figure 1B). Thus, the conditions chosen for the following experiments were the less stressful ones (12 pulses of 20 ms and 250 V/cm).

ET efficiency in arthritic joints We then wanted to investigate the ET efficiency on joints of arthritic mice. We injected normal and full-blown arthritic knee joints (day 32 post-induction) with pVax2hIL10, and electroporated both groups with the optimal experimental conditions determined above. Significant cytokine expression was detected in the joints of both groups when measured 3 days after ET. Although arthritic knee joints secreted three times more hIL-10 than Copyright  2006 John Wiley & Sons, Ltd.

Figure 2. Efficiency of the intra-articular electrotransfer (ET) in arthritic joints. (A) The pVax2-hIL10 was electrotransfered to knee joints of na¨ıve or arthritic DBA/1 mice (day 32) and the local IL-10 production measured on day 3 as described in Figure 1. Error bars represent SEM (n = 5 mice/group). The results are representative of three different experiments. (B) Full-blown arthritic knee joints of DBA/1 mice were injected with phosphate-buffered saline (PBS) (control) or pCMVβ (25 µg) and the β-galactosidase expression revealed 1 day after ET. Representative macroscopic views embedded in paraffin of the two electrotransfered groups are presented

non-arthritic joints, the difference was not statistically significant due to high individual variability (Figure 2A). Macroscopic examination of pCMVβ-electrotransfered knee joints revealed that the reporter gene expression was limited to the intra- and peri-articular electroporated sites (Figure 2B). Histological analysis of the electro-injected arthritic joints showed a synovial lining tissue and bone marrow staining, while no LacZ expression was observed in cartilage and bone (data not shown).

Kinetics of transgene expression For efficiency of the ET in arthritic mice, the persistence of the therapeutic transgene expression is a prerequisite for a fair and lasting protection in a chronic disease such as RA. Thus, mice with CIA were sacrificed at different time points following ET, and like the previous experiment, hIL10 was dosed in the knee-conditioned J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm

1031

Gene Electrotransfer in Arthritis

Table 1. Clinical efficiency of local and systemic hIL-10 DNA delivery in collagen-induced arthritisa

hIL-10 (pg/ml)

A 50 40

Histological scores

30

Treat- Incidenceb ment (%)

20

Arthritic Radiological severityd scoree Hind (%) (%) paws Knees

10 0

3

7 Days following ET

CT IA IM

100 86 57∗∗

32.8 ± 1.31 34.6 ± 1.66 29.9 ± 1.25

66 66 43∗∗

37.5 57 7∗∗∗

50 47 43 52 13∗∗∗ 19∗∗∗

a Male

B 50 hIL-10 (pg/ml)

Onsetc (days)

40

**

30 20 10 0 1

7 15 Days following second ET

Figure 3. Kinetics of IL-10 secretion by patellas following one (A) or two (B) i.a. ET. The cytokine production was measured in the conditioned medium from patellas by ELISA at the indicated days after gene transfer. (A) the pVax2-hIL10 (25 µg) plasmid was electrotransfered in arthritic knees when clinical signs appeared (0 = day 32 post-induction), and IL-10 secretion was measured in the conditioned medium from patellas by ELISA on days 3 and 7. (B) The pVax2-hIL10 (25 µg) plasmid was electrotransfered in arthritic knees on days 0 and 10 after arthritis onset and cytokine production was measured 1, 7 and 15 days after the second ET. Bars represent SEM (n = 6 mice/group) and ∗∗ p < 0.01. Results are representative of two separate experiments

medium. Figure 3A shows an expression of hIL-10 that gradually increased on days 3 and 7, and was undetectable at day 14 (data not shown). We then wondered if a second ET could enhance gene transfer performances (Figure 3B). When knees were injected with pVax2hIL10 and electroporated a second time, 10 days after the first ET, the hIL-10 secretion by transduced knee joints reached similar levels 7 days after the second ET. However, transgene secretion dropped again to nearly undetectable levels 2 weeks after the second ET.

Therapeutical application of i.a. ET to experimental arthritis Aiming at a complete validation of the proposed non-viral gene transfer system for local delivery of a therapeutic gene to arthritic joints, we evaluated the efficacy of i.a. ET in delivering clinically relevant levels of therapeutic protein in the CIA mouse model of RA, and compared it with the i.m. ET performances. DBA/1 mice were injected either in both knee joints or TA muscles with the same amount of plasmid (25 µg), and ET was performed on days 18 and 28 post-immunization as it was previously optimized [20]. Clinical severity of arthritis is classically evaluated by macroscopic examination of Copyright  2006 John Wiley & Sons, Ltd.

DBA1 mice were immunized with bovine collagen II and boosted 21 days after as described in Materials and methods. On days 18 and 28 following arthritis induction, mice were injected i.a. or i.m. with 25 µg of pVax-hIL 10 or an empty pVax, and electroporated as indicated in M&M. b Percentage of affected animals at the day of sacrifice in each group. c Day of clinical occurence of arthritis observed in any of the hind paws. Mean ± SEM. d Arthritic severity in each group was determined by macroscopic examination. An arthritic score for each mouse was calculated by summing the scores for each paw as described in Materials and methods. e Radiological and histological examinations were performed on day 50 of arthritis induction and scored as described in Materials and methods. Data are expressed as percentage of animals with severe arthritis features in each group. ∗ p < 0.05, ∗∗ p < 0.01, and ∗∗∗ p < 0.001 compared with the control group.

disease features and measurement of the ankle joint swelling during disease course. ET of pVax-hIL10 given twice showed a statistically significant lower incidence and a decreased severity of CIA only when delivered i.m., but no delay of the disease onset (Table 1). By the end of the experiment, 100% and 86% of paws were affected in the groups electrotransfered i.a. with empty plasmid or pVax2-hIL-10, respectively, while only 57% of the paws developed clinical signs for arthritis in the i.m. treated group (p < 0.0001). Mean paw widths were only significantly decreased in the long term for the group of mice electrotransfered i.m. (p < 0.01). At no time during the study was severity of paw swelling significantly different between the control group and the i.a. treated group (Figure 4). To investigate more precisely the erosion and demineralization criteria of the arthritic ankles, we evaluated by radiological examination the paws, recovered from each group of mice after sacrifice (Table 1). On day 50, the ankles of control group showed radiological features of severe joint destruction. The X-ray scores were significantly lower in the i.m. electrotransfered group: 7% of animals showed severe radiological score, against 37.5% in the control group and 57% following i.a. ET (p < 0.0001). Histological examination of arthritic joints from these two groups revealed a significant increased inflammatory synovitis, pannus formation and massive inflammatory cell infiltration (50 and 43%, respectively) compared with the group electrotransfered with IL-10 i.m. (13%). Thus, clinical observations were confirmed by the radiological and histological observations. In order to evaluate a potential therapeutic effect of pVax-hIL10 i.a. ET restricted to the injected tissue, we performed a histological examination of knees in all three groups following hematoxylin and eosin (H&E) staining, J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm

1032

M. Khoury et al.

2.30

IM

IA

CT

Paw swelling

2.20 2.10

ET1

ET2

2.00

* *

**

1.90 1.80 15

20

25

30

35

40

45

50

Days following arthritis induction Figure 4. Effect of i.a. and i.m. electrotransfer (ET) of pVax2-hIL10 on CIA. DBA/1 mice were immunized with bovine type II collagen, boosted on day 21 and synchronized with LPS on day 28. The ET was delivered on days 18 and 28 (arrows) post-arthritis induction with 25 µg of pVax2-hIL10 in 10 µl, injected either in both knee joints (
) or in the tibialis anterior muscle ( ). The control group was ET with empty plasmid (). Paw swellings were measured until day 50 and the mean value was represented for each group (n = 15 mice/group). ∗ p < 0.05; ∗∗ p < 0.01. Results are representative of two separate experiments

ž

and scored tissue sections for pathology (Table 1 and Figure 5). Scores showed no difference between groups electrotransfered i.a. with empty plasmid or pVax2-hIL-10 (47 and 52% of severe arthritic scores), and improved arthritic scores in the knees of the group electrotransfered i.m. (19% of severe scores), confirming the clinical, radiological and histological analysis of hind paws. As an attempt to find conditions for therapeutically efficient i.a. ET of hIL-10, we increased the amount of both the injected DNA and the number of ET. DBA/1 mice were injected in both knee joints with either 30 or 100 µg of the pVax2-hIL10, and ET was performed on days 18, 28 and 35 following arthritis induction. In both groups, the clinical severity of arthritis was comparable to the control electrotransfered group (data not shown). On the day of euthanasia, transgene expression was measured by ELISA in knee-conditioned medium and by Q-PCR in homogenized knee joints. Similar amounts were detected in both treated groups, at both RNA (32.5 ± 7.8 and 30.6 ± 6.9 AU) and protein levels (18.2 ± 3.4 and 17.3 ± 3.2 pg/ml). Interestingly, the percentage of severe histopathological scores was significantly decreased in the high-dose pVax2-hIL10-treated group (75% compared with 100% in the control group). Despite the lack of difference in hyperplasia of synovial tissue, the degree of cartilage and bone erosion was significantly reduced (p < 0.001), as evidenced by more conserved intraarticular spaces (Figure 5).

Transgene expression in electrotransfered tissues and serum We quantified the levels of hIL-10 in both serum and electrotransfered tissues by ELISA at euthanasia. On the one hand, treated TA muscles and knee joints Copyright  2006 John Wiley & Sons, Ltd.

were dissected, homogenized and tissue extracts were prepared, and on the other culture media conditioned by the injected knee joints were obtained, as described in Materials and methods. Knee joint supernatants or knee homogenates from mice electrotransfered i.a. twice did not reveal detectable cytokine secretion on day 50, 3 weeks after the last ET. When knee joints were treated by ET three times and collected 2 days after the last ET, 18.34 ± 2.93 pg/ml of hIL-10 were measured in the pVax2-hIL10-injected group. In contrast, injected TA muscles expressed an average of 7.48 ± 0.49 pg/mg of muscle proteins 3 weeks after the last ET. In another experiment, mice were sacrificed on days 25, 35 and 50 following arthritis induction. One week after the first ET, IL-10 protein expression was 131.46 ± 18.46 pg/mg of muscle proteins in the injected TA muscles, 64.13 ± 10.73 pg/mg of muscle proteins 1 week after the second ET, and still detectable 3 weeks after the last ET (2.13 ± 1.66 pg/mg of muscle proteins). IL-10 could not be detected at any time in the serum of treated groups.

Th1/Th2 balance in CIA Since CIA is a Th1-mediated autoimmune disorder and Th2-derived cytokines such as IL-10 ameliorate the disease, we investigated whether the IL-10 ET had an effect on Th1 and Th2 cytokine production by bCII-specific effector T cells in spleens and draining lymph nodes. As previously described, CIA resulted in the development of T cells that produce high levels of IFNγ and low levels of IL-4 (Figure 6). However, ET with IL-10 led to a decrease of IFN-γ , that was higher for the i.a. delivery route, and to a similar increase in IL-4 secretion in both groups, compared with controls. There was no difference in the T cell proliferative response to bCII according to the treatment (data not shown). These data indicate that, no matter what the used route for IL-10 gene delivery during CIA development was, the ET efficiency was not due to inhibition of antigen-specific T cells. The IL-10 ET led to an attenuation of the Th1 response and to an increase of the Th2 response during CIA development, as indicated by the inversion of the Th1/Th2 cytokine balance.

Discussion Although RA gene transfer strategies mediated by viral vectors have proven to be highly efficient in delivering therapeutic genes [1], the design of non-viral vectors is receiving increasing interest. Indeed, since joint diseases are non-lethal, the main purpose of using gene therapy is to improve the disorder with safe delivery systems. In experimental models of RA, the i.m. injection of naked DNA alone, encoding potential therapeutic genes such as TGF-β1 [21] or IL-10 [22], showed significant protection. However, direct i.a. injection of plasmid DNA alone, or complexed with cationic liposome, was J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm

1033

Gene Electrotransfer in Arthritis

Knees

Paws

A

B

C

D

Figure 5. Histological analysis of electrotransfered mice. Representative slides hematoxylin and eosin staining (×100) showing the knees (left column) and hind paws (right column) of arthritic mice electrotransfered with 25 µg of pVax2-hIL10, injected either in both tibialis anterior muscles (A) or knee joints (B), or with empty plasmid (C): (A) intra-muscular ET shows aspect of normal synovial and cartilage; (B, C) intra-articular ET with therapeutic and control plasmids shows synovitis and cartilage destruction; (D) intra-articular ET with 100 µg of pVax2-hIL10 shows synovitis but conserved intra-articular joint space

reported to transduce between 1 and 5% of the synovial lining cells, and to produce a very short local expression (2–3 days) of the reporter gene [23]. Thus, plasmid DNA Copyright  2006 John Wiley & Sons, Ltd.

is poorly efficient for local gene transfer and could only be considered when an amplification of the therapeutic gene effect is expected, such as the bystander effect observed J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm

1034

A

M. Khoury et al. 4500 *

4000

mIFNγ (pg/ml)

3500

CT IM IA

*

3000 **

2500 2000 1500 1000 500 0

B

70

**

mIL-4 (pg/ml)

60 50

**

CT IM IA

40 30 20 10 0

Figure 6. Effect of electrotransfer on Th1/Th2 balance in CIA. Spleen cells were collected from control group (white bar), or mice electroporated with pVax2-hIL10 either intra-articularly (hatch bar) or intra-muscularly (black bar), and stimulated with ConA. The IL-4 and IFNγ levels were assayed by ELISA in culture supernatants. Results represent the mean ± SEM for each group of treated mice (n = 15)

with HSV-tk [24] or in TNF-based vaccination protocols [25]. For some time now, in vivo ET has been shown to be an efficient method to increase naked DNA administration into tissues and organs. Four groups showed an improvement in arthritis in the CIA mice model following the i.m. ET of plasmid DNA encoding IL-10 [14,26] or TNF-α inhibitors [27,28], and two studies demonstrated the successful transduction of normal rat articular tissues using the local gene transfer by i.a. ET of a reporter DNA [11,12]. More recently, a group successfully used local ET for a direct delivery of siRNA targeting TNF-α into the joint tissues of arthritic mice [13], showing for the first time that i.a. ET can efficiently inhibit CIA in mice following local delivery of therapeutic compounds. Since we previously showed that systemic delivery of IL-10 by i.m. ET protected mice from CIA [14], we wanted to explore the potential use of i.a. ET for therapeutic gene delivery in experimental arthritis, as a way to compare the local versus systemic delivery of IL-10. The present report confirmed the feasibility of using a non-viral strategy for delivering a therapeutic gene Copyright  2006 John Wiley & Sons, Ltd.

in mice joints, and is the first to investigate the i.a. injection of plasmid DNA encoding hIL-10 as a biologically relevant transgene, followed by ET using non-invasive electrodes, in the pathological context of experimental arthritis. The experimental parameters determined for IL10 gene transfer (12 pulses of 250 V/cm and 20 ms) were similar to that reported for marker gene ET in normal rat joints [11,12], and comparable to that reported for siRNA in arthritic mouse joints [13]. One study precisely addressed the safety concern and showed that ET, with or without DNA, produced no tissue damage, nor influenced the physiological morphology and metabolism of the articular tissues [11]. Since our experimental conditions are similar to the parameters they used, local response to the procedure is likely to be the same in healthy mice. For arthritic mice joints, such a question cannot be addressed due to the high inflammatory infiltrates in pathological joint tissues. We showed that electrotransfered joints are able to secrete the anti-inflammatory cytokine IL-10, in slightly higher levels when knees are arthritic. This observation has to be related to the highest cellular content in arthritic joints compared with normal joints, due to synovium hyperplasia and immune cell infiltrate. The histological analysis of arthritic mice joints revealed β-gal staining of both synovial lining cells and bone marrow cells, but never cartilage or bone. This is in accordance with the observation made by Ohashi et al. [12] when using a reporter gene i.a. ET of normal rat joints. We cannot exclude that immune cells infiltrating the pannus could also be targeted by ET. The staining in the bone marrow compartment was observed near the growth plate, mostly in the zone connecting the bone marrow to the synovium. This observation might reveal either that the DNA is able to diffuse from the electro-injected joint space to the bone marrow compartment to transduce bone marrow cells, or that cells transduced within the synovial space migrated from the synovial hyperplasia to the bone marrow. Both hypotheses could be considered since (1) enlarged bone canals link the bone marrow and the synovium, and (2) cells morphologically similar were identified in bone marrow and in rheumatoid synovial tissue [29]. The therapeutic protein was transiently expressed within arthritic joints, reaching undetectable levels at 2 weeks, and never detectable in circulation. This kinetic profile is comparable with results obtained using a reporter gene [12], and different from that observed following i.m. IL-10 ET. Indeed, i.m. ET resulted in a local drop in IL-10 secretion after the first week, as already reported [26], but was still detectable for at least 3 weeks following gene ET. Although we could extend the duration of i.a. transgene expression when ET was performed twice or thrice, the expression was still transient since it decreased to basal level 2 weeks after the last ET. The significance of this short transgene expression within the first weeks following i.a. ET is not clear. Moreover, increasing the amount of electrotransfered plasmid above 30 µg does not confer a higher expression level within the joint, as shown in the ELISA and Q-PCR data. This might J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm

1035

Gene Electrotransfer in Arthritis

suggest that i.a. saturation is reached for the injected dose. This difference between i.a. and i.m. kinetics could be due to better availability of target tissue for i.m. than i.a. transduction. It could also be related to the attenuation of the CMV promoter activity due to high amounts of inflammatory cytokines in arthritic joints compared with muscle tissue [30]. Although the transgene is of human origin, the hypothesis of possible immune response against IL-10 can be excluded since none of the studies using injection of plasmid hIL-10 to immunocompetent mice, even arthritic, revealed an immune response against it. Its anti-inflammatory properties might predominate and explain the long-lasting expression after i.m. ET. It is possible that local inflammation due to the arthritic status of the joint, plasmid sequences, and/or ET procedure induce an immune response against the transduced cells. Although the pVax backbone used contains minimal vector sequences, minimizing extraneous genetic elements, we cannot exclude that plasmid-specific sequences might activate innate immunity through toll-like receptors. The major purpose of our study was to assess the feasibility and efficiency of such ET as a local therapeutic gene delivery in the CIA mouse model of RA, and to compare it with the therapeutic efficiency of i.m. ET. Many studies on IL-10 in CIA using adenoviral vectors [15,16,31–34], adeno-associated virus [20], or plasmid DNA [14,26,35] have already demonstrated the therapeutic efficacy of the cytokine. In previous works using i.m. ET of IL-10, we and others showed that 0.1–0.5 ng of IL-10/mg muscle proteins along the disease course were enough to achieve a therapeutic effect on CIA [20,36], even in the absence of detectable serum IL-10. In this study, we measured ten times less IL-10 within the treated joints. We showed that i.a. ET of 25 µg of IL-10 is ineffective in decreasing the clinical features of CIA, both locally in the injected knees, and systemically in the hind paws. These results reflect a relationship between the levels of IL-10 locally detected and its inhibitory effect on clinical features of the disease, suggesting that i.a. ET clinical efficiency might be increased over time if the kinetics of transgene secretion could be prolonged by repeating ET along the disease course. Indeed, increasing the number of ET induced a significant protection against cartilage and bone erosion, as assessed by a decreased histopathological severity, although it was insufficient to show an effect on inflammation, leading to unchanged macroscopic features of the disease. Thus, although our data indicate that i.a. ET of an antiinflammatory cytokine is feasible in vivo in the CIA model of RA, only the i.m. ET was able to demonstrate a strong therapeutic effect, due to a higher transduction efficiency of muscle compared with joint tissues, providing longer kinetics of transgene expression. The kinetics of IL-10 expression obtained after i.a. ET do not fulfill the main requirement of a chronic disease such as RA which is a long-term production of the therapeutic protein. In conclusion, the potential practical application of such a gene transfer strategy is not yet reached, although an appreciable therapeutic effect was observed on CIA mice Copyright  2006 John Wiley & Sons, Ltd.

receiving repetitive electrotransfer of high IL-10 doses once a week. Intra-articular ET appears to be safe and efficient for anti-inflammatory cytokine gene delivery, and opens perspectives for non-viral gene therapy applications to RA, but further work is needed to improve its efficacy.

Acknowledgements This work was supported by the INSERM, the ENSCP, the CNRS and the University of Montpellier I. We would like to thank D. Greuet and C. Cantos for animal experience assistance, and Mich`ele Radal from the CRLC Val d’Aurelle in Montpellier for the preparation of histological sections. We are grateful to the service of the breast radiography group from Pr. Taourel at the Lapeyronie Hospital in Montpellier who performed the hind paw radiographies. This work was supported by the European consortium Stemgenos QLRT-2001-02039.

References 1. Apparailly F, Plence P, Noel D, et al. Recent advances for cell/gene therapy in rheumatoid arthritis. Curr Med Chem 2004; 3: 119–138. 2. Nishikawa M, Huang L. Nonviral vectors in the new millennium: delivery barriers in gene transfer. Hum Gene Ther 2001; 12: 861–870. 3. Bureau MF, Gehl J, Deleuze V, et al. Importance of association between permeabilization and electrophoretic forces for intramuscular DNA electrotransfer. Biochim Biophys Acta 2000; 1474: 353–359. 4. Miklavcic D, Semrov D, Mekid H, et al. A validated model of in vivo electric field distribution in tissues for electrochemotherapy and for DNA electrotransfer for gene therapy. Biochim Biophys Acta 2000; 1523: 73–83. 5. Heller L, Jaroszeski MJ, Coppola D, et al. Electrically mediated plasmid DNA delivery to hepatocellular carcinomas in vivo. Gene Ther 2000; 7: 826–829. 6. Titomirov AV, Sukharev S, Kistanova E. In vivo electroporation and stable transformation of skin cells of newborn mice by plasmid DNA. Biochim Biophys Acta 1991; 1088: 131–134. 7. Oshima Y, Sakamoto T, Hisatomi T, et al. Targeted gene transfer to corneal stroma in vivo by electric pulses. Exp Eye Res 2002; 74: 191–198. 8. Matsumoto T, Komori K, Shoji T, et al. Successful and optimized in vivo gene transfer to rabbit carotid artery mediated by electronic pulse. Gene Ther 2001; 8: 1174–1179. 9. Heller LC, Coppola D. Electrically mediated delivery of vector plasmid DNA elicits an antitumor effect. Gene Ther 2002; 9: 1321–1325. 10. Lohr F, Lo DY, Zaharoff DA, et al. Effective tumor therapy with plasmid-encoded cytokines combined with in vivo electroporation. Cancer Res 2001; 61: 3281–3284. 11. Grossin L, Cournil-Henrionnet C, Mir LM, et al. Direct gene transfer into rat articular cartilage by in vivo electroporation. FASEB J 2003; 17: 829–835. 12. Ohashi S, Kubo T, Kishida T, et al. Successful genetic transduction in vivo into synovium by means of electroporation. Biochem Biophys Res Commun 2002; 293: 1530–1535. 13. Schiffelers RM, Xu J, Storm G, et al. Effects of treatment with small interfering RNA on joint inflammation in mice with collagen-induced arthritis. Arthritis Rheum 2005; 52: 1314–1318. 14. Perez N, Plence P, Millet V, et al. Tetracycline transcriptional silencer tightly controls transgene expression after in vivo intramuscular electrotransfer: application to interleukin 10 therapy in experimental arthritis. Hum Gene Ther 2002; 13: 2161–2172. 15. Joosten LA, Lubberts E, Durez P, et al. Role of interleukin-4 and interleukin-10 in murine collagen-induced arthritis. Protective effect of interleukin-4 and interleukin-10 treatment on cartilage destruction. Arthritis Rheum 1997; 40: 249–260. J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm

1036 16. Apparailly F, Verwaerde C, Jacquet C, et al. Adenovirusmediated transfer of viral IL-10 gene inhibits murine collageninduced arthritis. J Immunol 1998; 160: 5213–5220. 17. Mir LM, Bureau MF, Gehl J, et al. High-efficiency gene transfer into skeletal muscle mediated by electric pulses. Proc Natl Acad Sci U S A 1999; 96: 4262–4267. 18. Devine MJ, Mierisch CM, Jang E, et al. Transplanted bone marrow cells localize to fracture callus in a mouse model. J Orthop Res 2002; 20: 1232–1239. 19. Djouad F, Fritz V, Apparailly F, et al. Reversal of the immunosuppressive properties of mesenchymal stem cells by tumor necrosis factor alpha in collagen-induced arthritis. Arthritis Rheum 2005; 52: 1595–1603. 20. Apparailly F, Millet V, Noel D, et al. Tetracycline-inducible interleukin-10 gene transfer mediated by an adeno-associated virus: application to experimental arthritis. Hum Gene Ther 2002; 13: 1179–1188. 21. Song XY, Gu M, Jin WW, et al. Plasmid DNA encoding transforming growth factor-beta1 suppresses chronic disease in a streptococcal cell wall-induced arthritis model. J Clin Invest 1998; 101: 2615–2621. 22. Miyata M, Sasajima T, Sato H, et al. Suppression of collagen induced arthritis in mice utilizing plasmid DNA encoding interleukin 10. J Rheumatol 2000; 27: 1601–1605. 23. Nita I, Ghivizzani SC, Galea-Lauri J, et al. Direct gene delivery to synovium. An evaluation of potential vectors in vitro and in vivo. Arthritis Rheum 1996; 39: 820–828. 24. Sant SM, Suarez TM, Moalli MR, et al. Molecular lysis of synovial lining cells by in vivo herpes simplex virus-thymidine kinase gene transfer. Hum Gene Ther 1998; 9: 2735–2743. 25. Wildbaum G, Youssef S, Karin N. A targeted DNA vaccine augments the natural immune response to self TNF-alpha and suppresses ongoing adjuvant arthritis. J Immunol 2000; 165: 5860–5866. 26. Saidenberg-Kermanac’h N, Bessis N, Deleuze V, et al. Efficacy of interleukin-10 gene electrotransfer into skeletal muscle in mice with collagen-induced arthritis. J Gene Med 2003; 5: 164–171.

Copyright  2006 John Wiley & Sons, Ltd.

M. Khoury et al. 27. Bloquel C, Bessis N, Boissier MC, et al. Gene therapy of collageninduced arthritis by electrotransfer of human tumor necrosis factor-alpha soluble receptor I variants. Hum Gene Ther 2004; 15: 189–201. 28. Gould DJ, Bright C, Chernajovsky Y. Inhibition of established collagen-induced arthritis with a tumour necrosis factor-alpha inhibitor expressed from a self-contained doxycycline regulated plasmid. Arthritis Res Ther 2004; 6: R103–113. 29. Marinova-Mutafchieva L, Williams RO, Funa K, et al. Inflammation is preceded by tumor necrosis factor-dependent infiltration of mesenchymal cells in experimental arthritis. Arthritis Rheum 2002; 46: 507–513. 30. Qin L, Ding Y, Pahud DR, et al. Promoter attenuation in gene therapy: interferon-gamma and tumor necrosis factoralpha inhibit transgene expression. Hum Gene Ther 1997; 8: 2019–2029. 31. Kim KN, Watanabe S, Ma Y, et al. Viral IL-10 and soluble TNF receptor act synergistically to inhibit collagen-induced arthritis following adenovirus-mediated gene transfer. J Immunol 2000; 164: 1576–1581. 32. Quattrocchi E, Dallman MJ, Dhillon AP, et al. Murine IL-10 gene transfer inhibits established collagen-induced arthritis and reduces adenovirus-mediated inflammatory responses in mouse liver. J Immunol 2001; 166: 5970–5978. 33. Ma Y, Thornton S, Duwel LE, et al. Inhibition of collageninduced arthritis in mice by viral IL-10 gene transfer. J Immunol 1998; 161: 1516–1524. 34. Whalen JD, Lechman EL, Carlos CA, 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. 35. Fellowes R, Etheridge CJ, Coade S, et al. Amelioration of established collagen induced arthritis by systemic IL-10 gene delivery. Gene Ther 2000; 7: 967–977. 36. Deleuze V, Scherman D, Bureau MF. Interleukin-10 expression after intramuscular DNA electrotransfer: kinetic studies. Biochem Biophys Res Commun 2002; 299: 29–34.

J Gene Med 2006; 8: 1027–1036. DOI: 10.1002/jgm