Adenoviral gene transfer in a rat fracture model - SAGE Journals

Adenoviral gene transfer in a rat fracture model M. van Griensven1 , P. Lobenhoffer1 , A. Barke2 , T. Tschernig3 , W. Lindenmaier4 , C. Krettek1 & T. G. Gerich1 1

Department of Trauma Surgery, 2 Department of Virology, 3 Department of Functional & Applied Anatomy, Hannover Medical School, Carl-Neuberg-Strasse 1, D-30625 Hannover, Germany and 4 Gesellschaft fu¨r Biotechnologische Forschung, Mascheroder Weg 1, D-38124 Braunschweig, Germany

Summary For the enhancement of fracture healing, either puri®ed proteins or vectors for expression of growth factors in situ may be used. Adenoviral vectors directly convert cells to express a transgene. However, the cell types which are preferentially infected and the time of expression during fracture healing are currently not known. T he adenoviral type 5 vectors used in this study are replication incompetent viruses, one encoding b-galac tosidase (b-GAL) and one green ¯uorescent protein. Fem ora of 35 Sprague±Dawley rats were fractured. T hree days after stabili zation with Kirschner wire, 10 12 pfu viral suspension were injected into the fracture zone. As a control, ®ve animals received injections of adenovirus type 2. Animals were sacri®ced after 3 days, 1, 2 and 4 weeks. Fract ures healed radiographically within 2±3 weeks. All specimens were examined for b-GAL and green ¯uorescent protein (GFP) expression. Fibrobl ast and osteoblast s within callus tissue displayed a high transgene expression (week 1). A decrease of expression was observed during the observation period. In this experimental study, we have demonstrated that all cells of the primary callus can be transfected using adenoviral vectors, which provide a tool to further investigate adenoviral transfer of growth factors such as bone morphogenetic protein-2 (BMP-2 ). Keywords

Growth factors; fracture; transfection; adenovirus

Post-traum atic defects in fractures still compromise operative stabili zation and functional outcome. Recombinant BM P-2 has the potential to enhance the native repair response and is currently used in clinical trials at the time of index surgery. However, presently these factors need repetitive application to exhibit a physiological effect and deposit new bone (Kolbeck e t a l. 1999 ). Expression of this cytokine by cells in situ is thought to have an advantage by a more physiological and prolonged release over time. For this purpose a number of vectors have been developed which transfer the speci®c DNA to encode the relevant factor.

C o rre spond e nc e to : Dr Ma rtijn va n G rie nsven MD Ph D E-m a il: grie nsven.m a rtijn.va n@m h-h a nno ve r.d e Accepted 7 February 2002

Vectors for a therapeutical purpose can be separated into two distinct groups, viral and non-viral vectors. Among the viral vectors, recombinant adenovirus is widely used as a gene delivery vehicle. T he vector is charact erized by its potency to infect different cell types with high ef®ciency. Dividing and non-dividing cells are targeted with equal ef®ciency. After infection the DNA is not incorporated in the host genome, but is lost after several cell cycles. Recent experiments have shown that an adenoviral BM P-2 vector can induce neo-bone-formation in a fracture zone in mice. T hese experiments, however, have used an e x vi vo approach with transfection and transplantat ion of omnipotent mesenchym al stem cells, thereby not using the bene®ts of a direct injection of the virus into the fract ure zone (Lieberman e t a l.

# Laboratory Animals Ltd. Laboratory Animals (2002) 36, 455–461

456

1998). Although it has been successfully demonstrated that this technique may be applicable, certain problems remain unsolved. After direct injection of a BMP vector, it is not known which cells are infected by adenoviral transduction or how the cells transform during fracture repair. To trace those infected cells it is necessary to visualize transduced cells. For this investigation, two types of adenoviral vectors were applied that encode detectable marker proteins. In a fracture model, we describe the time-dependent decrease of expression and the tropism of cells preferentially infected.

Material and methods C e lls a nd culture m e d ia T he adenoviral-derived, replicat ion-de®cient vectors Ad5C MVlacz (Fig 1) and Ad5cos45gfp (Fig 2) were propagated in human embryonic kidney cells HEK-293. T he cell line is stable transformed with the adenoviral E1A and E1B genes (Graham e t a l. 1977). Cells were grown in 90% Dulbecco’s minimal essential medium (DMEM , GibcoBRL, Rockville, Maryland, USA) supplemented with 10% foetal calf serum (Cytogen, Berlin, Germany). Further cell lines were employed for in vit ro assays monitoring the expression of the reporter genes b-galac tosidase (b-GAL) and the green ¯uorescent protein (GFP) in nonreplicative cells: NIH 3T 3 (EAC C Ref. No. 930615 24 ) and RCJ3.1 (C5.18 ): a rat osteogenic sarcoma cell line. T he 3T 3 cell line was maintain ed in growth medium (DMEM, Biochrom , Berlin, Germany)

Fig 1 Physical map of Ad5CMVlaczDE1=DE3: LacZ vector containing a polycloning site. XbaI and ClaI are used for subcloning the green uorescent protein (GFP) cassette with XbaI and the isoschizomeric PspI. Selection was carried out by the introduced ampicillin resistance gene Laboratory Animals (2002) 36

van Griensven et al.

Fig 2 Physical map of Ad5cos45gfp: green uorescent protein (GFP) cassette containing sticky ends of XbaI and PspI for subcloning in the polycloning site of the adenoviral cosmid vector. The cassette contains the coding sequence for GFP, driven by a CMV promotor. See text for further details

supplemented with 10% foetal calf serum. T he RCJ3 cell line was propagated in DMEM with Glutamax-1 (GibcoBRL) supplemented with 15% foetal calf serum. In addition, the culture medium of all three cell lines contained 100 units =ml of penicillin and 100 mg=ml streptomycin. Con¯uent cultures of cells in 75 cm 2 ¯asks contai ned between 2 to 4610 6 cells and were split 1:4 to 1:5 every 3±4 days. T he cells were incubated in a humidi®ed environment at 37 C in 5% CO 2.

Ad e no vira l ve c to rs Two different adenoviral vectors carrying reporter genes were used (Ad5C MVlacZDE1 =DE3, Quantum Biotechnologies Inc, Montreal, Canada and Ad5cos45gfp). T he Ad5C MVlacZDE1 =DE3 was part of the basic construction kit for adenoviral vectors (QBIInfect ‡ Adeno-Quest TM). T he viral particles of Ad5C MVlacZDE1 =DE3 had an initial concentration of approximately 10 3 pfu =ml. T hey were propagated in 293 cells to a ®nal concentration of approxim at ely 10 12 pfu =ml. Ad5cos45gfp was constructed using a cosmid cloning procedure, which allows the direct assembly of recombinant adenovirus genomes by cloning in Esc h e ric h ia co li (W. Lindenmaier, unpublished results). T he cosmid vector pAdcos45, which contains an E1 and E3 deleted adenoviral genome derived

Gene transfer for fracture healing

from pBHG11 and pDE1sp1A (Bet t e t a l. 1994 ), was digest ed with ClaI and XbaI. ClaI and XbaI were single cloning sites in the E1 region that allowed introduction of the GFP cassette. T he expression cassette for GFP was derived from pGreenlantern. It contained the immediate early gene promoter of CMV, the gene encoding an optimized GFP (Zolotukhin e t a l. 1996 ), and 3 0 regulatory sequences. It was adapt ed for cloning in the cosmid vector by introducing restriction sites for restriction enzym es Psp1406 I (Cla I-compatible) and Xba I at the 5 0 and 3 0 ends, respectively. Cosmid cloning yielded pAdcosgfp with the expected structure after in vitro packaging and propagation in E. c o li. Transfection of circular pAdcosgfp DNA into 293 helper cells resulted in excision of the recombinant genome from the circular cosmid DNA. GFP expression allows virus production and the infection of target cells to be followed directly.

De te c tio n o f re po rte r gene e xpre ssio n Approximately 1.5610 6 cells (293, 3T 3, or RCJ3) were plated in 25 cm 2 ¯asks 48 h before transfection, by which time they reached 80% con¯uency. Cells were washed once with 2 ml of pre-warmed phosphate-buffered saline (PBS) (37 C ) and were infected at a multiplicity of infection (moi) of 10 of the corresponding adenoviral vector. After 24 h of adsorpti on, 2 ml of culture medium supplemented with 2% foetal calf serum was added. T he expression of the reporter genes was tested for 24, 48 and 72 h aft er transfect ion. G FP e xpre ssio n Fluorescent cells were monitored directly at 580 nm wavelength. X-gal sta ining At the given times the cells were washed twice with PBS and ®xed for 30 min at 20 C with 0.5% glutaraldeh yde in PBS. T he ®xative was removed and the cells were washed twice at 20 C with 1 mM MgCl 2 in PBS and incubated with the X-gal reagent 5-brom o4-c hloro-3-indolyl-b-D-galac topyranoside (Boehringer, Mannheim, Germany) at 37 C for at least one hour. Staining was stopped by washing with PBS.

457

Am pli® c a tio n o f vira l pa rtic le s To generate high-titre viral stocks two different methods were employed. C e ll pa c k m e th od Six by 10 8 293 cells were infected at a moi of 1 to 10 with the corresponding adenoviral vector and incubated for 48 h at 37 C. Infected cells were collected by low-speed centrifugati on, washed once with PBS and resuspended in PBS. T he virus particles were extracted by ®ve freeze =thaw cycles at ¡ 80 C. T he cell debris was pelleted by lowspeed centrifugation at 1000 g. T he virus particles in the supernatant were concentrated by centrifugati on (100 000 g, 4 C, 1 h) and resuspended in the appropriate buffer giving a ®nal 10 12 pfu =ml. C e sium ch lo rid e grad ie nt c e ntrifugatio n m e th od About 10 9 infected production cells (293 cells) were washed with PBS and resuspended in 0.1 M Tris-saline buffer, pH 8.0. T he virus particles were extracted by ®ve freeze =thaw cycles at ¡ 80 C. T he cell debris was pelleted by low-speed centrifugati on at 1000 g. T he supernatant was collected and loaded on top of a discontinuous (step) cesium chloride gradient (2.5 ml CsCl, d ˆ 1.4 g=ml gently overlaid with 2.5 ml CsCl, d ˆ 1.2 g=ml, SW40 tubes, Beckm an ) and centrifuged at 100 000 g for 4 h at 15 C. T he infectious viral particles appeared as a bluish white band, and were collected by puncturing the bottom of the centrifuge tubes. If necessary the virions were further puri®ed on a continuous self forming CsCl gradient. For this purpose the viral band was loaded into a centrifuge tube (SW50 tubes, Beckman) and overlaid with CsCl, d ˆ 0.5 g=ml. After ultracentrifugat ion at 100 000 g at 4 C for 20 h, a bluish white viral band was visible and again this was collected by puncturing the bott om of the centrifuge tube. T he cesium chloride was removed by dialysis in a buffer. T he ®nal amount of infectious virions was about 10 12 pfu =ml. In vivo inve stiga tio n T he study was approved by the legislative animal welfare committee of the state of Laboratory Animals (2002) 36

458

Lower Saxony, Germany. All anim als were handled at room temperature for 14 days before treat ment. T hroughout the study period, pelleted rodent chow (Altromin 1324 ) and water were available a d lib itum . T he lighting was maintai ned on a 12 h cycle. T hirty-®ve male Sprague±Dawley rats (Central Animal Laboratory, Hannover Medical School) weighing 400 g were used for the in vivo experiment. Fifteen rats were used for the Lacz vector. Fifteen rats were used for the GFP vector, and ®ve rats served as controls and received an injection with wild-t ype adenovirus type 2. Anaesthesia was performed using an i.p. injection of ketam ine (100 mg=kg) and xylazine (16 mg=kg). T he fem ur was exposed through the lateral approach. T hrough the lat eral condyle, a 1.2 mm Kirschner wire was inserted into the trochanteric region. T he femur was then fractured in the diaphysis using a clamp. Analgesic treatm ent was provided daily for the ®rst week after the fracture in all anim als (200 mg/g metam izol-sodium i.m. (Novalgin 1 , Hoechst, Unterschleiûheim, Germany)). On the 3rd post-operati ve day, one millilitre of virus suspension (approximat ely 10 12 pfu =ml) was injected into the fracture zone. Previous experiments have

Fig 3 -galactosidase and eosin staining. The histology represents granulation tissue one week after injection. Numerous b-galactosidase positive-stained cells are incorporated in this tissue. Fragments of adherent muscle bres fail to react (magni cation 100x) Laboratory Animals (2002) 36


shown that at this time of fracture healing a highly cellular granulati on tissue has developed. After surgery, free cage activity was allowed. T he animals were then sacri®ced after 1, 2, 3, and 4 weeks by intracardial injection of potassium chloride after the animals had been anaesthetized with ketamine (100 mg=kg) and xylazine (16 mg=kg). T he fractures were healed radiographically without complication within 2 weeks. A dislocation of the Kirschner wire was observed in four cases. T he bone and adjacent soft tissue was harvested and prepared for histological evaluation. Specimens were decalci®ed in 45% formic acid and 10% sodium citrate for 5 days. Staining and counterstaining for the Lac z vector was performed using X-gal and eosin. After embedding in paraf®n, slices of 3 mm thickness were prepared. Specim ens transfected by the GFP vector were examined under ¯uorescent light at a wavelength of 580 nm.

Results In vivo inve stiga tio n T he analysi s of regional and temporal expression of transgenes in this fract ure model is semiquantitative, since no exact count of transfect ed cells is possible. For the early time period we focused on the expression within the granulation tissue. At this tim e, we observed fracture fragm ents embedded in a homogenous but not organized highly cellular tissue. Cells expressing Lac z and GFP were found in high numbers. Remnants of thigh muscle were adherent to the fragm ents. In the muscle tissue however, no expression could be demonstrated (Fig 3). Transfected cells were interpreted to be mostly of mesenchym al origin. At 1 and 2 weeks after surgery ossi®cati on and remodelling of original bone was in progress. Chondral progenitor cells with a hyaline cartil age-like appearance were found to express both of the transgenes. Adjacent to this cart ilage-like tissue, transduced osteoblasts and osteoclasts were found as lining cells (Fig 4). T hese cells were active in resorption of fract ure fragm ents and deposi-


459

Fig 4 -galactosidase and eosin staining. Cartilagelike appearance (magni cation 100x)

tion of osteoid. During ossi®cati on transduced cells were stepwise incorporated in the newly mineralized bone. T he cells were found in small clusters surrounded by solid bone (Fig 5). At the tim e of fract ure healing newly formatted bone was found to solidly bridge the fragm ents (Fig 6). Still a high number of cells expressed the transgenes. T he ®nding of transduced cells in the subchondral area. T his may have occurred by forced injection of the virus supension in the fracture region. However, the subchondral

Fig 6 Green uorescent protein (GFP) uorescence. At 3 weeks the fracture zone is bridged by a callus formation with high activity of the transgene, demonstrating the transfer from progenitor cells to differentiated osteoblasts (magni cation 100x)

Fig 7 Green uorescent protein (GFP) uorescence. Extrusion of the adenovirus into the subchondral area with uorescence below the cartilage (magni cation 100x)

Fig 5 -galactosidase and eosin staining. After 3 weeks newly developed bone incorporates blue staining osteoblasts. The margin is covered by a dense connective tissue (magni cation 100x)

lamella hindered a further extrusion of the virus, since no chondrocytes of the artic ular cartil age have been transformed to a GFP or Lac z expressing phenotype (Fig 7). T he sham-operated side using an adenovirus wild type did not display a staining related to transgene expression. Laboratory Animals (2002) 36

460

Discussion Since the ®rst hint to the existence of bone morphogenetic proteins by Urist and co-workers, there has been a constant effort to further investigat e growth fact ors (Johnson e t a l. 1988, Urist 1965 ). During the last years, we have increased our knowledge on the biology and use of BMP (Rosen & T hies 1992, Cook e t a l. 1995 ). Recombinant proteins are currently used in a number of clinical trials. However, it is theorized that a single application of growth factors, either as free molecules or as molecules associat ed with slow release devices, may not provide suf®ciently high local concentrations of the factors for suf®cient periods of time (King 2001 ). T herefore, novel approaches for growth factor delivery using an adenoviral vector are being investigated. T hese experiments have revealed that in situ production of BM P-2 by adenoviral-m ediated transduction may result in new bone formation that is of superior qualit y than that provided by the delivery of the recombinant protein alone (Lieberman e t a l. 1999). However, this technique evokes a number of questions that cannot be answered using the BM P vector, namely the type of cell that is transduced and the temporal patt ern of transgene expression. We therefore created a fracture model in the rat in an attem pt to answer these questions. In a preliminary study, the reporter genes Lac z and GFP were used for visualizing transfection. No adverse events occurred regarding fract ure consolidation. Histologically we found a rich expression of the transgene at all the time points studied, which is closely related to the time of injection. It is known that an adenovirus vector is rapidly inactivat ed after introduction into tissue (Dedieu e t a l. 1997, Ji e t a l. 1999 ). T herefore, an im mediat e contact to a target cell is crucial. T he tissue has to be cellular and penetrable for liquids. T his situati on is usually given when the haematoma is resorbed and the fracture site is in®ltrated by mesenchymal stem cells. Since this was a preliminary marker study, our primary objective was to describe the continuation of the transgene expression. Our ®nding was a persistent expression in Laboratory Animals (2002) 36


higher differentiat ed cells, demonstrati ng that the transgene is not inactivated by im munological reactions as suggested in previous studies (Dedieu e t a l. 1997, Ji e t a l. 1999 ). At the time of fracture healing there was still a signi®cant expression. T his ®nding correlates with the literature describing an expression lasting up to 6 weeks (Nishida e t a l. 1998). Presently, no explanation can be given for our ®nding that there was no expression of the transgene within the adjacent muscle ®bres. It is known that adenovirus easily infects muscle tissue (Kirshenbaum e t a l. 1993 ). Okubo and co-workers describe in vitro expression of BMP-2 in myoblasts, and a subsequent conversion to osteoblasts with production of alkaline phosphatase and osteocalcin (Okubo e t a l. 1999 ). Balt zer and co-workers share this observat ion in an in vivo study (Balt zer e t a l. 1999 ). We hypothesize that our observation depends on the fracture type model and the tim e of infection. T he viral suspension was injected on the 3rd day after surgery. At this tim e a granulati on tissue had developed, creating a barrier to the muscle tissue. T herefore, our observation seems not to be a principal difference but a technical difference. If this observation holds true, the risk of heterotopic ossi®cation seems to be reduced. It remains to be further investigated whether or not this was merely an accidental ®nding. With this experimental study we have demonstrated that cells of the granulation tissue can be transduced using an adenoviral vector. Transgene expression can be found in differentiat ed cells of the callus such as chondrocytes, osteoblasts, and osteoclasts. In conclusion, this study can be regarded as a basis for further investigat ion into the feasibility of introducing a cytokin e vector such as BMP in fracture sites. References Baltzer AW, Lattermann C, Whalen JD, Braunstein S, Robbins PD, Evans CH (1999) A gene therapy approach to accelerating bone healing. Evaluation of gene expression in a New Zealand white rabbit model. Knee Surge ry a nd Sports Tra um a to lo gy a nd Arth ro sc o py 7, 197±202


Bett AJ, Haddara W, Prevec L, Graham FL (1994) An ef®cient and ¯exible system for construc tion of adenovirus vectors with insertions or deletions in early regions 1 and 3. Pro c ee d ings o f th e Na tio na l Aca d em y o f Scie nc es o f th e U SA 91, 8802±6 Cook SD, Wolfe MW, Salkeld SL, Rueger DC (1995) Effect of recombinant hum an osteogenic protein-1 on healing of segmental defects in non-human primates. Jo urna l o f Bo ne a nd Jo int Surge ry (Am ) 77, 734±50 Dedieu JF, Vigne E, Torrent C, Jullien C, Mahfouz I, Caillaud JM, Aubailly N, Orsini C, Guillaume JM, Opolon P, Delaere P, Perri caudet M, Yeh P (1997) Long-term gene delivery into the livers of immunocompetent mice with E1=E4-defective adenoviruses. Jo urna l o f Viro lo gy 71, 4626±37 Graham FL, Smiley J, Russell WC, Nairn R (1977) Characteristics of a hum an cell line transformed by DNA from human adenovirus type 5. Jo urna l o f G e ne tic Viro lo gy 36, 59±74 Ji L, Bouvet M, Price RE, Roth JA, Fang B (1999) Reduced toxicity, attenuated immunogenicity and ef®cient mediation of hum an p53 gene expression in vivo by an adenovirus vector with deleted E1-E3 and inactivated E4 by GAL4-TATA promoter replacement. G e ne Th e ra py 6, 393±402 Johnson EE, Urist MR, Finerman GA (1988) Bone morphogenetic protein augmentation grafting of resistant femoral nonunions. A preliminary report. C linic a l O rth o pae d ic s 230, 257±65 King GN (2001) T he importanc e of drug delivery to optim ize the effects of bone morphogenetic proteins during periodontal regeneration. C urre nt Ph a rm a c o lo gy a nd Bio te c h nolo gy 2, 131±42 Kirshenbaum LA, MacLellan WR, Mazur W, French BA, Schneider MD (1993) Highly ef®cient gene transfer into adult ventricular myocytes by recombinant adenovirus. Jo urna l o f C linica l Inve stiga tio n 92, 381±7

461

Kolbeck S, Bail H, Schmidmaier G, Remmler G, Raun K (1999) Rekombinantes Wachstumshorm on bewirkt eine Beschleunigung der Heilung von tibialen Knochendefekten beim Minischwein. In: He fte zu d e r U nfa llch irurg (Hertel P, Rehm KE, eds). Berlin: Springer, pp 141±2 Lieberman JR, Daluiski A, Stevenson S, Wu L, McAllister P, Lee YP, Kabo JM , Finerman GA, Berk AJ, Witte ON (1999) The effect of regional gene therapy with bone morphogenetic protein-2 -producing bone-marrow cells on the repair of segmental femoral defects in rats. Jo urna l o f Bo ne a nd Jo int Surge ry (Am ) 81, 905±17 Lieberman JR, Le LQ, Wu L, Finerman GA, Berk A, Witte ON, Stevenson S (1998) Regional gene therapy with a BMP-2 -producing murine strom al cell line induces heterotopic and orthotopic bone form ation in rodents. Jo urna l o f O rth o ped ic Re se a rc h 16, 330±9 Nishida K, Kang JD, Suh JK, Robbins PD, Evans CH, Gilbertson LG (1998) Adenovirus-m ediated gene transfer to nucleus pulposus cells. Implications for the treatment of intervertebral disc degeneration. Spine 23, 2437±42 Okubo Y, Bessho K, Fujimura K, Iizuka T, Miyatake S (1999) Expression of bone morphogenetic protein-2 via adenoviral vector in C2C12 myoblasts induces differentiation into the osteoblast lineage. Bio c h e m istry a nd Bio ph ysic s Re se a rch C o m m unic a tio ns 262, 739±43 Rosen V, Thies RS (1992) T he BMP proteins in bone form ation and repair. Tre nd s in G e ne tic s 8, 97±102 Urist MR (1965) Bone: form ation by autoinduction. Scie nc e 150, 893±9 Zolotukhin S, Potter M, Hauswirth WW, Guy J, Muzyczka N (1996) A `humanized’ green ¯uorescent protein cDNA adapted for high-level expression in mammalian cells. Jo urna l o f Viro lo gy 70, 4646±54

Laboratory Animals (2002) 36