A comparison of in vivo gene delivery methods for antisense ... - Nature

Gene Therapy (1998) 5, 1455–1461  1998 Stockton Press All rights reserved 0969-7128/98 $12.00 http://www.stockton-press.co.uk/gt

A comparison of in vivo gene delivery methods for antisense therapy in ligament healing N Nakamura1, SA Timmermann1, DA Hart1, Y Kaneda2, NG Shrive1, K Shino3, T Ochi3 and CB Frank1 1

McCaig Centre for Joint Injury and Arthritis Research, University of Calgary, Calgary, Alberta, Canada; 2Institute for Molecular and Cellular Biology, Osaka University; and 3Department of Orthopaedic Surgery, Osaka University Medical School, Osaka, Japan

To determine the most efficient in vivo delivery method of oligonucleotides for antisense therapy in ligament healing, fluorescence-labelled phosphorothioate oligodeoxynuleotides (ODN) were introduced into 12 rabbit ligament scars 2 weeks after injury using haemagglutinating virus of Japan (Sendai virus; HVJ)-conjugated liposomes. We compared the efficiency of cellular uptake of fluorescence as a percentage of all cells in each scar using three delivery procedures: (1) direct free-hand injection into the ligament scar using a conventional syringe; (2) systematic direct scar injection using a repeating 10 ␮l dispenser and a square mesh grid system; and (3) injection into the feeding (femoral) artery. Results showed that there was a significant difference in fluorescence uptake by scar cells on day 1 after injection between the three delivery methods: (1) direct free-hand, 9.7 ± 7.6% (average ± s.d.); (2) systematic direct, 58.4 ± 15.9%; and (3) intra-arterial, 0.2 ± 0.1%. Systematic direct injection was most efficient and it resulted in 25.9 ± 13.0% of scar cells being labeled

at 7 days after transfection. We then introduced antisense ODN for the rabbit proteoglycan, decorin, into ligament scars with this delivery method and confirmed a significant inhibition of decorin mRNA expression in antisense-treated scar tissues in vivo both at 2 days (42.3 ± 14.7% of sense control ± s.d.; P ⬍ 0.0025) and 3 weeks (60.5 ± 28.2% of sense control ± s.d.; P ⬍ 0.024) after treatment, compared with sense ODN-treated scars. Decorin was significantly suppressed also at protein level in antisense-treated scars at 4 weeks (66.6 ± 35.7% of sense control ± s.d.; P ⬍ 0.045) after treatment. These results demonstrate that in vivo transfection efficiency in ligament scars is ‘delivery system dependent’ and that introduction of antisense ODN for the small proteoglycan, decorin, with this delivery method can lead to significant suppression of its expression over 3 weeks both at mRNA and protein levels. Thus, an effective model for the potential manipulation of scar composition and quality in ligament healing has been established.

Keywords: gene therapy; antisense; Sendai virus; liposome; ligament; wound healing

Introduction Skeletal ligaments are discrete bands of dense connective tissue which connect bones across a joint, serving important roles in stabilizing joints and guiding joint motion.1 Ligament sprains are one of the most common injuries to joints.2 At present, however, despite a variety of treatments, the optimal conditions for promoting ligament healing still remain unclear. Previous research has shown that ligaments heal with scar-like tissue, which is inferior to normal ligament tissue both biologically and biomechanically.3–6 Whereas the tensile strength of injured skin recovers by 10 weeks following injury,7 gaphealing medial collateral ligament of the knee (MCL) reach only 30–40% of normal material strength of ligament at 1 year.3 Since a major mechanical role of ligaments is to stabilize joints and since loose ligaments can lead to significant joint instability, which may predispose the joint to the development of degenerative arthritis,8,9

Correspondence: CB Frank, McCaig Centre for Joint Injury and Arthritis Research, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta, Canada, T2N 4N1 Received 17 November 1997; accepted 30 June 1998

this inferior biomechanical quality of scar material gives rise to clinical concern. For the purpose of designing new therapies, it is important to note that ligament scar tissue is different from normal ligament tissue in many aspects, which collectively appear to contribute to the inferior biomechanical properties; elevated proteoglycan content such as decorin and biglycan;10 elevated adhesion molecule content such as fibronectin and tenascin;11 decreased type I collagen and elevated type III collagen; different collagen organization; and abnormal collagen cross-linking.3–6 Therefore, manipulation of these (or other) scar components toward those in normal tissue may improve scar quality. Gene transfer has received attention for its potential biological manipulation of scarring.12 As one possibility, the antisense method appears to be a feasible method for this purpose which blocks the transcription or translation of specific genes by binding of antisense phosphorothioate oligodeoxynucleotides (ODN) to target mRNA.13 To increase the efficiency of cellular uptake of ODN in vivo, we have developed a highly efficient Sendai virus (haemagglutinating virus of Japan; HVJ)-liposomemediated gene transfer method.14–17 HVJ promotes fusion of the liposome with target cells and delivers the ODN

Antisense therapy in ligament healing N Nakamura et al

1456

into the cells.18 We have previously reported that this gene delivery method can be applied to healing ligament by a direct injection of HVJ-liposomes into the ligament wound.19,20 However, the success of gene delivery into the glomerulus of kidney21 and the cardiac muscle22 via intra-arterial injection made us revisit the idea of using such potentially less invasive intra-arterial delivery of HVJ-liposome complexes into hypervascular ligament scars.23 The purpose of this study was two-fold. First, we wanted to optimize the number of cells within early ligament scars into which ODN can be transferred. To do this, we wanted to compare the three most promising delivery techniques into MCL scars; specifically comparing cellular uptake of fluorescence (6-carboxyfluorescein; 6-FAM)-labeled ODN in vivo. The others involved direct injection into scars and intra-arterial injection into feeding arteries of the scar. Second, we wanted to use the most efficient of these three methods to introduce antisense decorin ODN into MCL scars in vivo, since decorin has been implicated in the control of fibrillogenesis of collagen type I in vitro24 and in vivo.25 Therefore, excess decorin expression potentially inhibits normal collagen fibril assembly in early MCL healing and its transient inhibition could improve scar quality. Specifically in this study, we wanted to determine whether or not inhibition of decorin mRNA and protein expression could be achieved and whether its inhibition would persist in a ligament scar for a period of weeks, increasing the likelihood of generating a functional antisense effect in vivo.

Results Cellular uptake of 6-FAM-labeled ODN in ligament scar When scar cells contained 6-carboxyfluorescein (FAM) fluorescence-labeled ODN, their fluorescent areas corresponded to 4′, 6-diamidino-2-phenylindole (DAPI)stained areas. This suggests that most ODN introduced into scar cells accumulates within the nucleus (Figure 1). Comparison of three ODN delivery methods With lower power magnification, a clear difference was observed in the distribution of 6-FAM fluorescence in MCL scar tissue among the three delivery methods (Figure 2). By freehand direct injection, fluorescence was seen in a patchy distribution in the scar matrix. Fluorescence areas were mostly ovoid shapes with a longer radius up to 250 ␮m from the center, where the fluorescence intensity was strongest (Figure 2a). Needle holes were seen within the most fluorescent areas by viewing serial sections (not shown), suggesting that the most fluorescent area corresponds to the position of the needle tip at the time of injection. Using a 10-␮l syringe dispenser and a square mesh grid system, we also systematically injected every 0.5 mm within the ligament scar from the surface. With this method, fluorescence was diffuse throughout the scar (Figure 2b). In contrast, by intraarterial injection delivery, fluorescence was localized only around blood vessels. Spread of fluorescence from vascular areas was barely detectable, demonstrating that this method is unlikely to be effective in delivering any liposome-mediated products to ligament scars (Figure 2c). Quantitative comparisons showed that the percentage of fluorescence-labeled cells in MCL scar at 1 day after injec-

tion was 9.7 ± 7.6% (average ± s.d.) by direct free-hand injection, 58.4 ± 15.9% by systematic direct injection, and 0.2 ± 0.1% by intra-arterial injection (Figure 3). With the systematic injection method, 25.9 ± 13.0% of scar cells were still labelled 7 days after injection.

Antisense decorin therapy in vivo Results of testing the in vivo introduction of antisense ODN for the small proteoglycan, decorin, into MCL scars using systematic direct injection confirmed a significant suppression of decorin mRNA expression in antisense ODN-treated scar tissues in vivo both at 2 days (42.3 ± 14.7% of sense control ± s.d.; P ⬍ 0.0025) and 3 weeks (60.5 ± 28.2% of sense control ± s.d.; P ⬍ 0.024) after injection, compared with sense ODN-treated scars (Figure 4). Additional experiments have revealed that levels of decorin mRNA expression are similar, following injection of HVJ-liposomes alone, and injection of sense ODN in HVJ-liposomes, to uninjected controls (data not shown). Furthermore, we compared decorin protein expression in antisense and sense-ODN-treated scars at 4 weeks following injection. We first analysed equal volumes of tissue extract by SDS-PAGE followed by Coomassie blue staining to confirm that the amount of the loaded protein was similar (Figure 5a), and then performed immunoblotting using a specific antibody to decorin which recognizes the C-terminal half of the core protein. We observed that decorin levels in the tissues were significantly suppressed in antisense-treated scars compared with sense ODN-treated scars (33.4 ± 35.7% of sense control ± s.d.; P ⬍ 0.045) (Figure 5b and c).

Discussion Scar tissue is a difficult target for gene transfer because of rapid proliferation and high turnover of wound cells.6 Therefore, to address the difficulty in obtaining maximum gene introduction into scar cells, an optimization study of in vivo gene transfection was required. There have been two studies reporting the in vivo introduction of marker gene into ligament and ligament scar.19,26 These studies were done by a free-hand direct injection technique, where the importance of vector selection was emphasized. To date, no study has considered the significance of delivery mode to maximize gene transfer. This study is the first demonstration that optimal in vivo transfection efficiency in scars is definitely ‘delivery system dependent’. The delivery of ODN into ligament scar tissue was much more efficient by direct injection than by intra-arterial injection from the feeding artery. While HVJ-liposomes are known for their excellent diffusion capacity compared with other cationic liposomes or viral vectors,27 and despite the fact that early wound vasculature is highly permeable,28 the HVJ-liposome complex did not appear to penetrate well into scar matrix from the vasculature. This strongly suggests that vessels in the ligament scar matrix are resistant to the diffusion of a HVJ-liposome suspension. Poor diffusion of HVJliposomes in the scar matrix is further suggested by the restricted fluorescent area by free-hand direct injection. Cellular transfection occurred within only a small radius of 250 ␮m around the needle site. We observed that the fluorescence was even more restricted to a smaller radius when 3-week ligament scars were injected, suggesting that as the ligament scar matures, it becomes even more


1457

Figure 1 Fluorescence for DAPI and 6-FAM-labelled ODN in MCL scar. (a) Fluorescence microscopy of DAPI, nuclear stain dye in the MCL scar tissue 1 day after transfection. (b) The corresponding field showing the fluorescence of 6-FAM-labelled ODN in scar cells (×200).

Figure 2 Fluorescence of 6-FAM-labelled ODN in MCL scar on day 1 after injection. Fluorescence microscopy of 6-FAM-labelled ODN in the MCL scar tissue 1 day after transfection. (a) Free-hand direct injection; (b) systematic direct injection using a dispenser and a grid system; (c) Intra-arterial injection (×100).


1458

Figure 3 Transfection efficiency on day 1 after injection. Transfection efficiency (% total cells labelled ± s.d.) in MCL scars at 1 day after HVJliposomes containing 6-FAM-labelled scramble ODN by three different delivery methods (n = 4 each): free-hand direct injection, systematic direct injection using a dispenser and intra-arterial injection.

resistant to the diffusion of HVJ-liposome complexes (data not shown). Such poor diffusion of HVJ-liposome complexes in scar matrix demonstrates that systematic injection using a syringe dispenser and a square mesh grid system can improve transfection efficiency by up to six-fold, compared with that by free-hand injection. This allowed a transfection efficiency which was far higher than previous gene transfection studies in ligaments and tendons.19,27,29 We did not investigate the cell types of transfected cells in this study. We previously characterized the expression of a ␤-galactosidase gene introduced into cells in early ligament scar by direct injection of the HVJ-liposome vector,19 similar to the method performed in this study. With labelling for marker antigens, we identified that approximately one-third of the transfected cells were macrophage and monocyte lineage cells and two-thirds were fibroblastic cells. Considering that we used the same vector and injection technique (directly into scar tissue) in the present study, we presume that the population of transfected (ODN-introduced) cells resulting from the direct systematic injection method is similar to that in our former study. Finally, we succeeded, using the systematic direct injection method, in suppressing one specific matrix molecule for over 3 weeks at both the mRNA and protein levels. Notably, inhibition of decorin mRNA was comparable or even stronger at 3 weeks after injection, compared with that at 2 days. However, our 6-FAM-ODN introduction study showed that the percentage of fluorescent (transfected) cells in scar decreased by 50% in a week. Such discrepancy between the duration of the antisense effect and that of fluorescence survival in scar suggests that even ‘nonfluorescent’ scar cells may contain sufficient ODN to exert an antisense action, and that the actual percentage of transfected (antisense-influenced) cells may be higher than that estimated from our fluorescence uptake study. Collectively, these results demonstrate that this system may have promise as a model for future testing of antisense therapy in ligament healing and may provide a new therapeutic approach for manipulating scar properties and soft tissue healing.

Figure 4 Effect of antisense treatment on decorin mRNA expression. Semiquantiative RT-PCR analysis of total RNA from scar tissues for GAPDH (293 bp) and decorin (419 bp) mRNA levels at 2 days (a) and 3 weeks (b) after injection of HVJ-liposomes containing sense or antisense decorin ODN. The numbers (1–6) refer to individual animals. For each rabbit the left MCL was injected with sense ODN and the right leg the antisense ODN. Lane designations: S, sense-treated scar, AS, = antisensetreated scar. The results were analysed using image analysis and the effect of in vivo antisense therapy is presented as the mean percentage of sense control ± s.d. on mRNA levels in MCL scars at 2 days and 3 weeks after injection of HVJ-liposomes containing sense (n = 3) or antisense (n = 3) decorin ODN (c). Differences between groups were assessed using a paired t test.

Materials and methods Synthesis of oligonucleotides and selection of sequence targets The sequences of ODN against rabbit decorin were: antisense, 5′-GGA-TGA-GAG-TTG-CCG-TCA-TG-3′, sense, 5′-CAT-GAC-GGC-AAC-TCT-CAT-CC-3′ (−1 to +19 of rabbit sequence). This antisense ODN specifically inhibits decorin mRNA in primary cultured rabbit MCL scar cells (unpublished data). The sequences for 6-carboxyfluorescein (FAM) labelled ODN was 5′-CTT-CGT-CGGTAC-CGT-CTT-C-3′, with scramble sequence (19 bp). 6-FAM was labelled on the 3′ and 5′ ends of the ODN. All the ODNs were synthesized and purified by the University of Calgary regional DNA synthesis laboratory.


just before use), to allow formation of HVJ-liposomes. The mixture was incubated at 4°C for 10 min and, after adjusting total volume to 4 ml, incubated for 60 min with gentle shaking at 37°C. Free HVJ was removed from the HVJ-liposomes by sucrose density gradient centrifugation.

In vivo gene transfer of fluorescence ODN Twelve, 12-month-old female New Zealand White rabbits had the central 4-mm portion of their MCL removed surgically. After 2 weeks of healing, HVJ-liposome complex containing 50 ␮g of 6-FAM-labelled ODN was injected via one of three procedures (four MCLs for each): (1) direct free-hand injection into the ligament scar using a conventional syringe with a 30-gauge needle (n = 4); (2) direct scar injection using a repeating 10 ␮l dispenser with a 30-gauge needle (Hamilton, Reno, NV, USA) and a square mesh grid system which allowed us to inject systematically every 0.5 mm within the ligament gap (n = 4); (3) single shot injection into the feeding (femoral) artery with clamping of the femoral vein and pooling of the HVJ-liposome solution in the MCL scar area for 10 min (n = 4).

Figure 5 Effect of antisense treatment on decorin protein expression. Equal volume (20 ␮l) of scar tissue extracts (each from 50 mg of sample) as analysed by SDS-PAGE with Coomassie blue staining (a) and by immunoblotting using a monoclonal antibody for decorin which recognizes the C-terminal half of the core protein (b) at 4 weeks after injection of HVJ-liposomes containing sense or antisense decorin ODN. The numbers (1–4) refer to individual animals. For each rabbit the left MCL was injected with sense ODN and the right leg the antisense ODN. Lane designations: S, sense-treated scar; AS, antisense-treated scar. The results were analysed using image analysis and the effect of in vivo antisense therapy is presented as the mean percentage of sense control ± s.d. for decorin levels in MCL scars at 4 weeks after injection of HVJ-liposomes containing sense (n = 4) or antisense (n = 4) decorin ODN (c). Differences between groups were assessed using a paired t test.

Preparation of HVJ-liposomes HVJ-liposomes were prepared as described previously.16,19 Briefly, phosphatidylcholine, phosphatidylserine and cholesterol were mixed in a weight ratio of 1:4.8:2. Dried lipid was hydrated in balanced salt solution (BSS; 140 mM NaCl, 5.4 mM KCl, 10 mM Tris-HCl, pH 7.5) containing sense or antisense ODN. The mixture was agitated and sonicated for preparation of unilamellar liposomes. The liposome suspension (0.5 ml, containing 10 mg of lipids) as then incubated with HVJ (10 000 haemagglutinating units), which had been inactivated with ultra-violet irradiation (100 erg/mm2 per s for 3 min

Evaluation of transfection efficiency All 12 MCL scars were then harvested 1 day after transfection and fixed with 4% formalin. Three sections (from superficial, mid and deep portions) were examined from each specimen by fluorescence microscopy, after nuclear staining with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma, St Louis, MO, USA), and then the ratio of FAM-6-fluorescent area to DAPI-stained area were calculated using an image analysis system (Kontron Elektronik, Munich, Germany). Efficiency of transfection in scar tissue was determined using 12 sections in each injection group. Data were analyzed with an unpaired Student’s t test. With the most efficient procedure, the ODN was further introduced into another four 2-week MCL scars. On day 7 after transfection, the scars were harvested and the transfection efficiency was measured. Transfer of antisense decorin ODN Ten, 12-month-old female New Zealand White rabbits had bilateral MCL gaps as described. After 2 weeks, HVJliposome complex containing 20 ␮g of antisense rabbit decorin ODN was injected into MCL scars. Into the 10 contralateral MCL scars, HVJ-liposomes containing 20 ␮g of sense ODN were introduced as a control. Reverse transcription (RT)-PCR Two days and 3 weeks after injection, semiquantitative RT-PCR was performed on RNA extracted from scar tissues (n = 3 for each period) as previously reported30 using primer sets specific for rabbit decorin and a constituitively expressed housekeeping gene glyceraldehyde-3phosphate-dehydrogenase (GAPDH). The rabbit decorin 5′ primer (nucleotide 780–800) was 5′-TGT-GGA-CAATGG-TTC-TCT-GG-3′; the 3⬘ primer (nucleotides 1239– 1250) was 5′-CCA-CAT-TGC-AGT-TAG-GTT-CC-3′. The rabbit GAPDH 5′ primer (nucleotides 211–230 of rabbit sequence) was 5′-TCA-CCA-TCT-TCC-AGG-AGC-GA-3′; the 3′ primer (nucleotides 488–507) was 5′-CAC-AATGCC-GAA-GTG-GTC-GT-3′. The rabbit-specific decorin primers were synthesized and purified by the University of Calgary regional DNA synthesis laboratory. Cloning

1459


1460

and sequencing of the amplified cDNA fragment confirmed the identity of the product. Reaction mixtures were assessed by 2% agarose gel electrophoresis, ethidium bromide staining and then quantified using a Scanalytics Image Analysis system (Billerica, MA, USA). The ratio of decorin to GAPDH in each treatment group was calculated from the three specimens and compared with each other using a paired t test.

Proteoglycan extraction and immunoblotting Proteoglycan extraction from tissue was performed using a modification of the method reported previously.24,25 Four weeks after injection, proteoglycans were extracted from 50 mg of ligament scar tissue (n = 4 for each group) in 4 m guanidine HCl, 100 mm sodium acetate buffer, pH 6.0, containing 1 mm PMSF, 5 mm iodoacetamide, 5 ␮g/ml leupeptin, 5 ␮g/ml pepstatin and 10 mm EDTA. After a 72-h extraction at 4°C with gentle stirring, the unextracted residue was removed by centrifugation. Total extracted protein was precipitated with 10 volumes of cold 100% ethanol and collected by centrifugation. The resulting pellet was air-dried and resuspended in 200 ␮l SDS-PAGE buffer. Twenty microlitres of protein samples were separated on 10% polyacrylamide gels and the gels were stained with Coomassie blue or transferred to nitrocellulose (PROTRAN, Schleicher & Schuell, Keene, NH, USA) for 1 h at 250 mA. The nitrocellulose membrane was incubated with 0.01 U/ml chondroitinase ABC (Seikagaku, Tokyo, Japan) in 0.1 m Tris-acetate, pH 7.3, containing 1 mm PMSF, 5 mm iodoaccetamide, 5 ␮g/ml leupeptin, 5 ␮g/ml pepstatin and 10 mm EDTA, for 3 h at 37°C.31 The membrane was then incubated overnight at 4°C in a blocking solution containing TBS-T (20 mm Tris-HCl, pH 7.6, 137 mm NaCl, 0.1% Tween 20) and 5% skim milk. After washing with TBS-T, the membrane was incubated in a 1:10 dilution of a monoclonal antibody to bovine decorin core protein (6D6), which recognizes the C-terminal half of the core protein,32,33 overnight at 4°C. The membrane was extensively washed with TBS-T, and then incubated for 1 h with horse anti-mouse IgG conjugated to HRP (Amersham, Arlington Heights, IL, USA) diluted to 1:2500 in TBS-T containing 5% skim milk. After washing, the antigen–antibody complexes were detected with ECL detection reagents (Amersham) according to the manufacturer’s instructions. Exposure of the immunoblot to XAR5 film (Eastman Kodak, Rochester, NY, USA) was for 10–15 s. Protein expression in each treatment group from the four specimens was quantified using a Scanalytics Image Analysis system and levels were compared using a paired t test. Preliminary experiments with purified decorin and serial dilutions of the primary antibody confirmed the linearity of both the detection system and the image analysis assessment.

Acknowledgements We would like to thank Ms Linda Marchuk, Ms Leona Barclay and Mr Craig Sutherland for their technical assistance. We also would like to thank Professor Paul G Scott (Department of Biochemistry, the University of Alberta) for providing us with purified decorin, the decorin (6D6) antibody and critical discussion of the protein analysis. This work is supported by the London Life Insurance Company. The Canadian Arthritis Society, and the Alberta Heritage Foundation for Medical Research.

DAH is the Grace Glaum Calgary Foundation Professor in Arthritis Research. CBF is the McCaig Professor in Joint Injury.

References 1 Viidik A. Tendons and Ligaments. In: Comper WD (ed). Extracellular Matrix, Vol. 1. Harwood Academic: Amsterdam, 1996, pp 303–327. 2 Miyasaka KC, Daniel DM, Stone ML. The incidence of knee ligament injuries in the general population. Am J Knee Surg 1991; 4: 3–8. 3 Frank CB et al. Medial collateral ligament healing: a multidisciplinary assessment in rabbits. Am J Sports Med 1983; 11: 379–389. 4 Woo SL-Y, Horibe S, Ohland KJ, Amiel D. The response of ligaments to injury: healing of the collateral ligaments. In: Daniel DM, Akeson WH, O’Connor JJ (eds): Knee Ligaments: Structure, Function, Injury, and Repair. Raven Press: New York, 1991, pp 351–364. 5 Frank C, Schachar N, Dittrich D. Natural history of healing in the repaired medial collateral ligament. J Orthop Res 1983; 1: 179–188. 6 Frank CB et al. Soft Tissue Healing. In: Fu F, Harner CD, Vince KG (eds). Knee Surgery, Williams & Wilkins: Baltimore, 1994, pp 189–230. 7 Deuel TF, Kawahara RS, Mustoe TA, Pierce GF. Growth factors and wound healing: platelet-derived growth factor as a model cytokine. Ann Rev Med 1991; 42: 567–584. 8 Hirshman HP, Daniel DM, Miyasaka K. The fate of unoperated knee ligament injuries. In: Daniel DM, Akeson WH, O’Connor JJ (eds). Knee Ligaments: Structure, Function, Injury, and Repair. Raven Press: New York, 1991, pp 351–364. 9 Daniel DM et al. Fate of the ACL-injured patients. A prospective outcome study. Am J Sports Med 1995; 23: 632–644. 10 Koob TJ et al. Proteoglycan metabolism in the healing medial collateral ligament. Trans Orthop Res Soc 1994; 19: 404 (Abstr.). 11 Neurath M. Expression of tenascin, laminin and fibronectin following traumatic rupture of the anterior cruciate ligament. Zeitschrift fur Orthopadie und Ihre Grenzgebiete. 1993; 131: 168–172. 12 Evans Ch et al. Gene therapy for ligament healing. In: Jackson DW et al. (eds). The Anterior Cruciate Ligament. Current and Future Concepts. Raven Press: New York, 1993; pp 419–422. 13 Crooke ST, Bennett CF. Progress in antisense oligonucleotide therapeutics. Ann Rev Pharmacol Toxicol 1996; 36: 107–129. 14 Kaneda Y, Iwai K, Uchida T. Increased expression of DNA cointroduced with nuclear protein in adult rat liver. Science 1989; 243: 375–379. 15 Kato et al. Expression of hepatitis B virus surface antigen in adult rat liver. J Biol Chem 1991; 266: 3361–3364. 16 Kaneda Y, Mortishita R, Tomita N. Increased expression of DNA cointroduced with nuclear protein in adult rat liver. J Mol Med 1995; 73: 289–297. 17 Dzau VJ, Mann MJ, Morishita R, Kaneeda Y. Fusigenic viral liposome for gene therapy in cardiovascular disease. Proc Natl Acad Sci USA 1996; 93: 11421–114225. 18 Okada Y et al. Modification of cell membranes with viral envelopes during fusion of cells with HVJ (Sendai virus). Exp Cell Res 1975; 93: 368–378. 19 Nakamura N et al. Transient introduction of a foreign gene into healing rat patellar ligament. J Clin Invest 1996; 97: 226–231. 20 Nakamura N et al. Early biological effect of in vivo gene transfer of platelet-derived growth factor (PDGF)-␤ into healing patellar ligament. Gene Therapy 1998; 5: 1165–1170. 21 Isaka Y et al. Glomerulosclerosis induced by in vivo transfection with TGF-␤ or PDGF gene into rat kidney. J Clin Invest 1993; 92: 2597–2601. 22 Morishita et al. In vivo transfection of cis-element ‘decoy’ against nuclear factor-␬␤ binding site prevents myocardial infarction. Nature Med 1997; 3: 894–899. 23 Eng K et al. Quantitative analysis of the vascular structures of ligaments. IEEE Trans Biomed Eng 1992; 39: 296–306.


24 Vogel KG, Paulsson M, Heinegard D. Specific inhibition of type I and type II collagen fibrillogenesis by the small proteoglycans from tendon. Biochem J 1984; 223: 587–597. 25 Danielson KG et al. Target disruption of decorin leads to abnormal collagen fibril morphology and skin fragility. J Cell Biol 1997; 136: 729–743. 26 Gerich TG et al. Gene transfer to rabbit patellar tendon: potential for genetic enhancement of tendon and ligament healing. Gene Therapy 1996; 3: 1089–1093. 27 Yonemitsu Y et al. Characterization of in vivo gene transfer into arterial wall mediated by the Sendai virus (hemagglutinating virus of Japan) liposomes: an effective tool for the in vivo study of arterial diseases. Lab Invest 1996; 75: 313–323. 28 Wuepper KD. Biochemistry and biology of components of the plasma kinin-forming system. In: Lepow IH, Ward PA (eds). Inflammation: Mechanisms and Control. Academic Press: New York, 1972, pp 93–117. 29 Lou J, Manske PR, Aoki M, Joyce ME. Adeno-virus mediated

30

31

32

33

gene transfers into tendon and tendon sheath. J Orthop Res 1996; 14: 513–517. Reno C et al. Rapid isolation of total RNA from small samples of hypcellular dense connective tissues. Biotechniques 1997; 22: 1082–1086. Westergren-Thorsson G et al. Differential expression of mRNA for proteoglycans, collagens and transforming growth factor-␤ in the human cervix during pregnancy and involution. Biochem Biophys Acta 1998; 1406: 203–213. Scott PG, Dodd CM, Pringle GA. Mapping the locations of the epitopes of five monoclonal antibodies to the core protein of dermatan sulfate proteoglycan II (decorin). J Biol Chem 1993; 268: 11558–11564. Scott PG et al. Immunohistochemical localization of the proteoglycan decorin, biglycan and versican and transforming growth factor-␤ in human post-burn hypertrophic and mature scars. Histopathology 1995; 26: 423–431.

1461