ARTHRITIS & RHEUMATISM Vol. 46, No. 9, September 2002, pp 2514–2523 DOI 10.1002/art.10527 © 2002, American College of Rheumatology
Adeno-Associated Virus–Mediated Osteoprotegerin Gene Transfer Protects Against Particulate Polyethylene–Induced Osteolysis in a Murine Model Shang-You Yang,1 Lois Mayton,1 Bin Wu,1 J. Jeffrey Goater,2 Edward M. Schwarz,2 and Paul H. Wooley1 Objective. Osteoprotegerin (OPG), a natural negative regulator of osteoclastogenesis and bone resorption, may be a potential therapeutic agent for treatment of osteolysis-associated prosthetic joint loosening. Using an in vivo adeno-associated virus (AAV)–mediated gene transfer technique, this study was designed to evaluate the protective effects of OPG transgene against orthopedic wear debris–induced bone loss in a murine model of osteolysis. Methods. Bone tissue was implanted into established pouches on BALB/c mice, followed by the introduction of ultra-high-molecular-weight polyethylene (UHMWPE) particles to provoke inflammation and osteolysis. The viruses encoding human OPG gene (rAAV-hOPG) or -galactosidase marker gene (rAAVLacZ) were injected into the air pouches, and the tissue was harvested 7 days after viral infection for histologic and molecular analyses. Results. Successful transgene expression was confirmed by the detection of OPG by enzyme-linked immunosorbent assay and positive X-Gal staining of pouch tissue (LacZ). Real-time polymerase chain reaction indicated significant diminishment of messenger RNA expression of osteoclast markers in OPG-
transduced pouches compared with rAAV-LacZ– transduced pouches. The transduction and expression of OPG also markedly decreased the gene copies of the biologic receptor activator of nuclear factor B. The expression of OPG in the bone-implanted pouch reduced bone calcium release by a mean of 39% compared with the calcium release in the other 2 groups. Computerized image analysis revealed that expression of OPG significantly protected against bone collagen loss. Conclusion. OPG gene transfer mediated by rAAV effectively protects against particulate polyethylene– induced bone resorption in this experimental model. Data suggest that gene transfer using rAAV-OPG may be a feasible and effective therapeutic candidate to treat or prevent wear debris–associated osteolysis and aseptic loosening. Aseptic loosening is the single most common complication of total joint prosthesis, occurring in up to 25% of implant recipients (1–4). Among the most important factors that may contribute to loosening is the adverse tissue response to particulate wear debris (5). Ultra-high-molecular-weight polyethylene (UHMWPE) has been broadly used in total joint prosthesis, and the UHMWPE components removed at revision surgery regularly show wear (6,7). Histologic evaluation of tissues from failed primary arthroplasties showed that particulate polyethylene is the most common debris found in periprosthetic tissue (8,9). It has been accepted that particles generated by mechanical wear of the prosthesis are phagocytosed by macrophages, resulting in cellular activation and release of proinflammatory mediators and cytokines, such as interleukin-1 (IL-1), tumor necrosis factor (TNF), and IL-6 (10,11). These mediators, in turn, induce local chronic inflammation with activation and recruitment of osteoclasts to the
Supported by grants from the Arthritis Foundation and the Department of Veterans Affairs (to Dr. Wooley). Dr. Yang is the recipient of a William D. Robinson Fellowship from the Arthritis Foundation, Michigan Chapter. Jeffrey Goater’s and Dr. Schwarz’s work was supported by PHS grants AR-45971 and AR-46545. 1 Shang-You Yang, MD, Lois Mayton, Bin Wu, MD, Paul H. Wooley, PhD: Wayne State University School of Medicine, Detroit, Michigan; 2J. Jeffrey Goater, MS, Edward M. Schwarz, PhD: University of Rochester, Rochester, New York. Address correspondence and reprint requests to Paul H. Wooley, PhD, Department of Orthopaedic Surgery, Wayne State University, 1 South, Hutzel Hospital, 4707 St. Antoine Boulevard, Detroit, MI 48201. E-mail:
[email protected]. Submitted for publication January 16, 2002; accepted in revised form June 5, 2002. 2514
OPG GENE TRANSFER PROTECTS AGAINST OSTEOLYSIS
2515
bone–implant interface (11,12). This affects bone remodeling around the implant and leads to osteolysis and aseptic loosening (13). While many different cytokines contribute to this process, studies have shown that the osteoclast differentiation factor (also called receptor activator of nuclear factor-B ligand [RANKL]) is 1 of the only 2 essential mediators to promote osteoclastogenesis. It binds to its membrane-bond signaling receptor, RANK, and stimulates osteoclast differentiation and maturation (14,15). Recently, a soluble protein osteoprotegerin (OPG) was identified in many types of cells and proved a natural “decoy” receptor that competed for RANKL with RANK and blunted its effects of osteoclastogenesis (15,16). Mice genetically deficient for RANKL or RANK suffered severe osteopetrosis (17,18), whereas OPG transgenic mice expressed the same pathology (19), demonstrating that RANKL and RANK are essential for osteoclast development, and OPG is a potent negative regulator for osteoclastogenesis. Based on the anti-osteolytic nature of OPG, it may be a potential therapeutic agent to treat debrisassociated periprosthetic bone resorption and aseptic loosening. While it is difficult, by conventional therapy, to administer sufficient OPG to osteolytic sites around the prosthetic joint, gene therapy provides an elegant solution to the delivery problem (20). In this study, we examined adeno-associated virus (AAV)–mediated OPG gene transfer to protect against particulate UHMWPE-induced osteolysis in a murine air pouch model of bone resorption.
II and Eco RI sites of pSL301 (Invitrogen, Carlsbad, CA) to generate pSL301-CMV-mIB. Then, a 1.8-kb fragment containing the CMV promoter and mIB cDNA was generated by digesting pSL301-CMV-mIB with Xba I and Eco RI, to generate 5⬘-Xba I. The pIRES-EGFP was digested with Xho I and Eco RI to release the IRES-EGFP fragment, which was ligated into the Xho I and Eco RI sites of pcDNA1 (Invitrogen) to generate pcDNA-IRES-EGFP. This plasmid was then digested with Eco RI and Xba I to release the IRES-EGFP fragment to generate 3⬘-Xba I. Then, pSub201 (a gift from Dr. J. Samulski, University of North Carolina [UNC], Chapel Hill) was digested with Xba I and Hind III, and the 4-kb vector fragment containing the AAV inverted terminal repeat packaging sequences was purified, and a 3-fragment ligation was performed with 5⬘-Xba I and 3⬘-Xba I to generate pAAVCMV-mIB-IRES-EGFP. The mIB cDNA in this vector was then replaced with the human OPG cDNA (Amgen, Thousand Oaks, CA) in pRc/CMV-Hu CR1 by subcloning into the Not I and Eco RI sites to generate pAAV-CMV-OPG-IRES-EGFP. Extensive restriction digests were performed to confirm the authenticity of the vector, and double-stranded DNA sequencing of the OPG cDNA through the 5⬘ and 3⬘ junctions was performed to confirm that there were no mutations in this region of the vector. Then, 0.5 mg of purified pAAV-CMV-OPG-IRESEGFP (Qiagen, Valencia, CA) was sent to the Gene Core Facility (UNC, Chapel Hill), which prepared the purified rAAV-OPG-IRES-EGFP using type-2 rAAV, and also provided the rAAV-LacZ control vector using a helper virus–free method (21). The resulting rAAV-OPG-IRES-EGFP was titered on human 293 cells and fibroblast-like synoviocytes, as described previously (22), to determine the concentration of infectious units, which was ⬃1 ⫻ 109/ml. Murine air pouch model of bone resorption to evaluate OPG gene transfer. A modification of the air pouch model of inflammation described previously (23) was used for this study (24). Six days after air pouch formation on the backs of BALB/c mice (23), a 0.6-cm section of femur (from the head to the shaft), or a 0.5 ⫻ 0.3–cm fragment of calvarium, from genetically identical littermates was surgically placed inside the pouch. Then 0.3 ml of sterile PBS containing 1:100 penicillin: streptomycin (Gibco BRL, Gaithersburg, MD) was injected into the pouch, while the pouch layers and skin incisions were closed using 4-0 Prolene sutures. The procedure was performed under sterile conditions in a laminar flow hood. After 24 hours, 5 mg of particulate UHMWPE suspension in 0.5 ml of 10% FBS/saline (1.5 ⫻ 107 particles/ml) was injected into the pouch to provoke inflammatory and osteolytic responses. The mice were divided into 3 groups the following day, and 0.5 ml of culture medium containing 1 ⫻ 108 infectious units of AAVs coding hOPG gene fused to the EGFP gene (rAAVOPG-IRES-EGFP) or LacZ control gene (rAAV-LacZ) was injected into the air pouches, respectively. Culture medium without virus (0.5 ml) was injected into bone-implanted pouches of the third group as a nonviral control. Previous studies have indicated that UHMWPE particles induce significantly increased pouch inflammation and bone implant resorption compared with pouches and implanted bones stimulated with FBS/saline vehicle (24,25); hence, the particle-free pouch control was eliminated from this study. The mice were killed in a CO2 chamber 7 days after viral
MATERIALS AND METHODS UHMWPE particles. Pure UHMWPE particles were the generous gift of Dr. John Cuckler (University of Alabama, Birmingham). Particles were sterilized by gamma irradiation and resuspended in sterile phosphate buffered saline (PBS) containing 10% fetal bovine serum (FBS) at a concentration of 1.5 ⫻ 107 particles/ml. The particle suspension was determined to be endotoxin free by Limulus amebocyte cell lysate assay (Endosafe; Charles River, Charleston, SC). Particle analysis with a Coulter Multisizer II (Coulter, Hialeah, FL) demonstrated that ⬎95% of the particles were ⬍7 m in diameter, with a mean size of 2.6 m (range ⬍0.7 to 21 m). Recombinant AAV vectors. Generation of the rAAV transfer vector (pAAV-cytomegalovirus [CMV]–OPG– internal ribosome entry site [IRES]–enhanced green fluorescent protein [EGFP]) was carried out via multiple subcloning steps. First, pCMV-mIB-IRES-EGFP, which contains the mIB complementary DNA (cDNA) cloned into the Not I site of pIRES-EGFP (Clontech, Palo Alto, CA), was digested with Bgl II and Eco RI, and the 1.8-kb fragment containing the CMV promoter and mIB was purified and cloned into the Bgl
2516
infection, and both the pouch membrane and implanted bone tissue were harvested. A portion of air pouch with the bone implant intact was either snap-frozen for crystallized sectioning or fixed in 10% buffered formalin (Fisher Scientific, Fair Lawn, NJ) for paraffin embedding. Pouch fluid, combined with a 0.2-ml saline rinse of the remaining portion of pouch tissue, was collected for determination of free calcium and cytokine production before the remainder of the tissue was stored for molecular analysis. Institutional approval was granted for these experiments by the Animal Investigation Committee, Wayne State University. Molecular and immunologic analysis. Enzyme-linked immunosorbent assays (ELISAs) were conducted on the supernatants of the pouch homogenates and the pouch fluids to examine OPG-transgene production and cytokine levels. All capture and detection monoclonal antibody pairs against different epitopes of various cytokine molecules were purchased from PharMingen (San Diego, CA), and antibodies for human OPG were obtained from R&D Systems (Minneapolis, MN). Tests were performed using the standardized protocol previously described (26). Real-time reverse transcriptase–polymerase chain reaction (RT-PCR) was performed to assess the influence of gene transfer on osteoclastogenesis. Gene expression of RANK, calcitonin receptor (CTR), and cathepsin K in pouch membranes was determined, and changes in expression of proinflammatory cytokines, including IL-1, TNF, and IL-6, were also examined. Total RNA from pouch homogenates was extracted following the manufacturer’s instructions (Tel-Test, Friendswood, TX). The cDNA was reverse transcribed from 0.5 g of total RNA in a 20-l reaction mixture containing 1⫻ PCR buffer, 500 M each of deoxynucleotide triphosphates, 0.5 units/l of RNase inhibitor, 2.5 M random hexamers, 5.5 mM MgCl2, and 1.25 units/l of RT (Perkin-Elmer Cetus, Norwalk, CT). The reaction mixture was incubated in a Thermal Cycler (Perkin-Elmer Cetus) at 25°C for 10 minutes, 48°C for 25 minutes, followed by 95°C for 5 minutes. Real-time PCR was performed according to the manufacturer’s instructions. To standardize target gene level with respect to variability in quality of RNA and cDNA, we used GAPDH transcripts, a housekeeping gene, as an internal control. Reaction mixtures of 25 l included 12.5 l of 2⫻ SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) and target gene primer pairs (at 400 nM final concentration) and 2 l cDNA. The primer pairs for target osteoclast marker genes were picked up in sequences from GenBank, NIH (http:// www.ncbi.nlm.nih.gov), and constructed by Genosys (The Woodlands, TX). The primers for mouse CTR were 5⬘ACAACTGCTGGCTGAGTG and 5⬘-GAAGCAGTAGATAGTCGCCAC; the primers for mouse cathepsin K were 5⬘-TATGTATAACGCCACGGCAA and 5⬘-CCGAGCCAAGAGAGCATATC; the primers for murine RANK were 5⬘CGAGGAAGATTCCCACAGAG and 5⬘-CAGTGAAGTCACAGCCCTCA. The primer pairs for the proinflammatory cytokines (IL-1, TNF, and IL-6) were purchased from Clontech. All reagents were from Applied Biosystems. The reactions were run in MicroAmp optical 96-well reaction plates with MicroAmp optical caps for 40 cycles (95°C for 15 seconds, 60°C for 1 minute) in the ABI Prism 7700 Sequence Detector (Applied Biosystems) and the fluorescent signals were recorded dynamically. Normalization and analysis of the reporter
YANG ET AL
signals (⌬Rn) at the threshold cycle was recorded by the machine’s built-in software, and target gene copies were calculated against the regression of the standard curve. The reaction mixture after real-time PCR was electrophoresed on 1.8% agarose gels containing ethidium bromide to verify the correct amplification of the target gene. Mobilized calcium determination. Calcium concentration in air pouch fluid was determined as a measure of bone demineralization, using an automatic fluorometric titration method in a fluorescent spectrophotometer (Model 810 Photomultiplier Detection System; Photon Technology, Lawrenceville, NJ) with fluorescent probe Fura-2 (Molecular Probes, Eugene, OR). The concentration of calcium released from implanted bone was calculated according to the method of Grynkiewicz et al (27) and expressed as nmoles/liter. Histologic analysis. Formalin-fixed pouches with implanted bone were decalcified with formic acid/sodium citrate before being embedded in paraffin and mounted with a consistent orientation and cut into 6-m sections. The sections were routinely stained with hematoxylin and eosin to examine bone erosion and changes in inflammation parameters in pouch membranes. Modified Masson’s trichrome staining (28) was performed to quantify the bone collagen content as a parameter of bone osteolysis. Briefly, the sections were deparaffinized and hydrated before equilibrating in Bouin’s solution (70% picric acid, 5% glacial acetic acid, and 10% formaldehyde) at 56°C for 1 hour. The sections were then incubated in phosphomolybdic (0.21% weight/volume)–phosphotungstic acid (0.21% [w/v]) for 10 minutes, followed by aniline blue solution (2.5% aniline blue in 2% acetic acid) staining for 5 minutes. The sections were then incubated in 1% acetic acid for 4 minutes before dehydration in graded alcohol. The bone collagen acquired a blue stain, with the stain density proportional to the collagen content. Histochemical tartrate-resistant acid phosphatase (TRAP) staining was performed to localize the osteoclast-like cells in the pouch tissue (29). Cryosections at 8-m thickness were prepared and fixed in buffered acetone for 30 seconds. Sections were incubated at 37°C for 1 hour in 100 mM acetate buffer (pH 5.2) containing 0.5 mM of naphthol AS-BI phosphoric acid, 2.2 mM of fast garnet GBC base, and 8 mM of sodium tartrate (Sigma, St. Louis, MO). The presence of dark purple staining granules in the cytoplasm was a specific criterion for counting TRAP-positive cells. X-Gal staining was performed to confirm LacZ gene transduction, as previously described (30). Briefly, frozen sections of bone-containing pouches were cut at 8 m and fixed in 0.5% glutaraldehyde for 30 minutes, followed by incubating overnight in a mixture containing 1 mg/ml X-Gal, 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, 2 mM MgCl2, and 0.02% of Nonidet P40. The sections were washed the next day and evaluated for the intensity and distribution of the blue coloration under a light microscope. Image analysis. The images of various stained tissue sections were digitally photographed with a Toshiba color camera (IK-TU40A; Toshiba, Tokyo, Japan) under a Zeiss light microscope (07740 Jena; Zeiss, Wetzlar, Germany). Air pouch membrane thickness and cellular infiltration were quantified by a computerized image analysis system with Image-Pro
OPG GENE TRANSFER PROTECTS AGAINST OSTEOLYSIS
2517
Plus software (Media Cybernetics, Silver Spring, MD) as described previously (23). To determine the changes of bone collagen content, the 100⫻ magnified images of modified trichrome blue (TRB)–stained pouch sections containing femur implants were captured by the software package. Integrated optical densities (IOD) of the areas at the bone surface that interacted with particles containing inflammatory pouch membranes were recorded and normalized with IODs obtained at inner parts of bone areas away from pouch membranes. Six pairs of IOD readings at different regions of the each bone section were averaged; at least 8 mice per group were studied. Due to the blurred boundaries of TRAP-positive osteoclasts in frozen sections, and the difficulty in separating individual cells, the amount of TRAP-positive osteoclast-like cells on the bone–membrane interface was quantified and compared by the image analysis system according to area of the dark purple coloration per millimeter of bone surface. Statistical analysis. Two separate experiments using 46 mice were conducted for this study, plus a preliminary experiment (with 10 mice) designed to select the optimum viral dosage to yield sufficient transgene (hOPG) product. In each of the 2 experiments, mice were randomly assigned to experimental groups with at least 7 mice per group. Data were combined, and statistical analysis among groups was performed by analysis of variance, with the Scheffe formula for post hoc multiple comparisons, using the SPSS software package (SPSS, Chicago, IL). Values are expressed as the mean ⫾ SEM. P values less than 0.05 were considered significant.
RESULTS Gene transduction and expression. ELISA was performed to verify hOPG-gene transduction and expression in the bone-implanted pouch tissues and subsequent secretion of OPG protein. Table 1 shows the OPG protein production in the fluids and homogenates of pouches transduced with rAAV-OPG-IRES-EGFP for 7 days at different virus concentrations. Viral infection at 108 infectious units gave the optimum OPG transgene production (mean ⫾ SEM 70.1 ⫾ 10.8 and
Table 1. Human osteoprotegerin (hOPG) protein levels in pouch lavages and pouch homogenates measured by enzyme-linked immunosorbent assay* Study condition Lavage fluid 106 AAV 108 AAV 109 AAV Pouch homogenate 106 AAV 108 AAV 109 AAV
hOPG gene– transduced
Virus-free control
21.9 ⫾ 9.3 70.1 ⫾ 10.8 71.8 ⫾ 10.6
ND ND ND
20.2 ⫾ 4.7 87.8 ⫾ 16.4 76.2 ⫾ 20.2
ND ND ND
* Values are the mean ⫾ SEM ng/mg total protein. AAV ⫽ adenoassociated virus; ND ⫽ not detectable.
Figure 1. Confirmation of successful LacZ gene transduction with X-Gal stains. A, A bone-implanted pouch infected with rAAV-LacZ for 7 days. B, A control pouch without the viral vector transduction.
87.8 ⫾ 16.4 ng/mg total protein in the pouch fluid and homogenates, respectively), and all the following experiments used this viral concentration. Auto green fluorescence that resulted from the gene production of the EGFP marker was captured on fresh-cut frozen sections of air pouches infected with rAAV-OPG-IRES-EGFP and further supported the success of the viral vector transduction (results not shown). X-Gal staining on rAAV-LacZ–transduced pouches revealed strong blue colors compared with the negative staining of rAAV-
2518
YANG ET AL
Figure 2. Gene expression changes of cathepsin K (CPK), calcitonin receptor (CTR), and receptor activator of nuclear factor B (RANK) in bone-implanted pouches. Changes were determined by real-time polymerase chain reaction. Data are converted as target gene copies reverse-transcribed from 1 ng of total RNA. Values are the mean and SEM. P values are versus the medium- and LacZ-treated groups.
OPG–inserted and nonviral pouches, indicating the successful transfer and expression of -galactosidase (LacZ) gene (Figure 1). Suppressive effects of gene transfer on osteoclastogenesis. Real-time PCR revealed marked changes in gene expression of the osteoclast markers (CTR and cathepsin K) following in vivo gene transfer. A dimin-
ishment of CTR and cathepsin K gene expression was observed due to rAAV-OPG transduction. This reduction was significant in comparison with the expression of osteoclast markers in either rAAV-LacZ transduction or virus-free controls (P ⬍ 0.05; Figure 2). In addition, OPG transduction significantly reduced the gene expression of RANK, the biologic receptor for RANKL, after therapeutic transgene incorporation (P ⬍ 0.05; Figure 2). Histochemical TRAP staining was conducted to identify osteoclast-like cells in the model. Figure 3 shows a typical TRAP-stained photomicrograph revealing that dark brown TRAP-positive cells accumulate along the bone–membrane interface. While Figure 3A illustrates heavy TRAP-positive cell deposition along the bone surface in the LacZ-transduced pouch, there were only small numbers of TRAP-positive cells on the hOPGgene–transduced sample (Figure 3B). Figure 4 summarizes the gene transfer effects on TRAP-positive cell infiltration analyzed by the image analysis system. OPG transduction significantly decreased the amount of TRAP-positive cellular deposits in comparison with the other 2 groups (P ⬍ 0.05). Inhibitory effects of OPG transfer on osteolysis. Protection against the bone resorption by rAAV-OPG transduction in the bone-implanted pouch was confirmed by the inhibition of bone calcium release into pouch fluid. Figure 5 indicates that concentration of free calcium in the pouch fluid collected from rAAV-OPG–
Figure 3. A, Typical tartrate-resistant acid phosphatase (TRAP)–stained frozen section of polyethylene particle–interacted, bone-implanted air pouch with LacZ reporter gene transfer. B, TRAP-stained section of bone pouch with human osteoprotegerin gene transduction. Osteoclastlike cells, that were TRAP-positive stained dark purple. (Original magnification ⫻100.)
OPG GENE TRANSFER PROTECTS AGAINST OSTEOLYSIS
2519
Figure 6. Typical bone pouch histologic sections with modified trichrome stain, showing bone collagen content (blue coloration). A, Virus-free control. B, Transduced with LacZ marker gene only. C, Transduced with rAAV-OPG. Figure 4. Quantification of the amount of tartrate-resistant acid phosphatase (TRAP)–positive stained cells in pouches following the gene transfer. A computerized image analysis system was used to compare the differences among groups according to the mean areas of dark purple color in sections. OPG ⫽ osteoprotegerin. Values are the mean and SEM. P value is versus the other 2 groups.
infected pouches was 18 nM, significantly less than the mean concentration of ⬎30 nM in the fluid from the other 2 groups (P ⫽ 0.03). The change in bone collagen content also indicated a protective effect due to the gene insertion. Modified TRB staining clearly revealed a reduction of blue staining (bone collagen content) at bone surfaces in contact with inflammatory pouch membranes. Figure 6 shows examples of typical TRB-stained histologic sections. Whereas nonvirus control (Figure 6A) and LacZcontrol (Figure 6B) showed dramatic bone collagen loss
Figure 5. Free calcium ion concentration in the pouch fluid as a measure of bone demineralization. Values are the mean and SEM. P value is versus the other 2 groups.
at the bone surface compared with the center of the section, OPG gene transfer resulted in marked preservation of bone collagen content (Figure 6C). A computerized image analysis system was used to quantify the collagen changes by comparing the integrated color density difference on modified TRB-stained sections. The measurements revealed that expression of human OPG significantly protected against collagen loss from bone areas in close contact with inflamed membranes (9.8% collagen loss), in comparison with similar regions in LacZ-transduced sections (22.1% loss; P ⬍ 0.04) and in nonviral controls (27.7% collagen loss; P ⬍ 0.01) (Figure 7). There was no significant difference in bone collagen loss between LacZ-transduced and nonviral medium groups. OPG gene transfer also dramatically attenuated
Figure 7. Comparison of integrated optical density of bone sections stained with modified trichrome blue (see Materials and Methods) to examine bone collagen loss. Values are the mean and SEM. P value is versus the other 2 groups.
2520
YANG ET AL
stimulated inflammation, bone erosion was essentially undetectable at 7 days post–OPG transduction. Assessment of particle-induced membrane inflammation. Human OPG gene transfer did not affect most parameters of particle-induced membrane inflammation. There was no significant difference between the groups in either pouch membrane thickness or cellular infiltration on day 7 following the gene transfer treatment. However, real-time PCR confirmed that cDNA transcribed from OPG-inserted pouch RNA contained significantly fewer gene copies of IL-1 in comparison with cDNA from the other 2 groups (P ⬍ 0.05; Figure 9A). There was no significant difference among the groups in the messenger RNA expression of the TNF or IL-6 genes (Figure 9A). ELISA detected substantial levels of IL-1, TNF, and IL-6 in pouch fluid and tissue
Figure 8. A, Model with human osteoprotegerin gene transduction for 7 days. Bone surface was intact though exposed to inflammatory pouch membrane. B, Typical histologic appearance of ultra-high-molecularweight polyethylene particle–induced bone resorption model without gene transfer. (Hematoxylin and eosin stained; original magnification ⫻200.)
the particle-associated bone erosion and macrophage/ osteoclast invasion. Figure 8A shows an image of pouch transduced with therapeutic OPG gene. Figure 8B shows a typical example of a bone-implanted air pouch stimulated with polyethylene particles for 7 days without gene transfer. Localized bone erosions were found at areas in close contact with the particle-stimulated inflammatory membrane. Despite the exposure to the polyethylene-
Figure 9. Expression of proinflammatory cytokines (interleukin-1 [IL-1], tumor necrosis factor [TNF], and IL-6) examined by A, real-time reverse transcriptase–polymerase chain reaction and B, enzyme-linked immunosorbent assay (ELISA). hOPG ⫽ human osteoprotegerin. Values are the mean and SEM. P values are versus the nonviral medium– and LacZ-treated groups.
OPG GENE TRANSFER PROTECTS AGAINST OSTEOLYSIS
2521
homogenates of all groups (Figure 9B). There was a trend toward declining IL-1 protein levels 7 days after the OPG gene transfer; which did not quite achieve statistical significance (P ⫽ 0.059).
animal arthritis (30,35,36); however, retroviral vectors have the disadvantage of only transducing dividing cells. AAV vectors have shown some distinct advantages over other viral vectors; they can infect both dividing and nondividing host cells and provoke a very limited host immune response, unlike the adenoviral vectors (22,34). AAV vectors also have better long-term transgene expression. Therefore, we have used this vector with the OPG gene to study potential anti-osteolytic effects. The data indicate that the exogenous gene was successfully incorporated and subsequently expressed with the secretion of high levels of therapeutic protein that actively inhibited polyethylene particle–induced osteoclastogenesis and osteolysis. In our preliminary study, we examined the dose of OPG-encoding AAV required to effectively deliver sufficient OPG to the pouch. Whereas infection with 106 IU of AAV produced only a trivial amount of OPG gene transduction, infection with 108 ml dramatically increased therapeutic protein production at local sites and subsequently protected against bone resorption, indicating the importance of viral vector dose in the therapeutic gene delivery system. Previous studies have suggested that a systemic concentration of 1 mg/kg is an effective dose of OPG (37). In this study, the OPG level at the local osteolytic site reached 70–87 ng/mg protein, i.e., 1.4–1.8 mg/kg body weight provided an effective dosage of the protein. In our previous studies, we have used the murine air pouch model of inflammation to examine tissue responses to orthopedic materials and potential therapeutic strategies to control the particle-associated inflammation (23,30,35). The major limitation of the murine air pouch, however, is the lack of bone tissue to evaluate wear debris–induced osteolysis. The development of surgical implantation of a certain-sized femur section or calvarium from a syngeneic donor into an established air pouch, followed by the injection of particulate UHMWPE or other orthopedic biomaterials, has led to a model that mimics many aspects of the failed prosthetic joint environment (24,25). The model provides a synovial-like membrane stimulated by particles that closely surround and interact with the implanted bone. The particle-stimulated membrane releases proinflammatory cytokines and contains infiltrating inflammatory cells, including macrophages and multinucleated foreign-body giant cells. With specific controls of the size of formed pouch and implanting bone, amount of particle introduction, and virus injection, we are able to assess the wear debris–associated inflammation, monitor changes in osteoclastogenesis and bone resorption, and evaluate ther-
DISCUSSION This study examined the protective effects of OPG, using in vivo AAV vector–mediated OPG gene transfer in a murine model of inflammatory osteolysis due to stimulation by UHMWPE particles. Loosening of an orthopedic prosthesis due to osteolysis represents the major long-term complication associated with joint arthroplasty, and wear debris–associated inflammation leading to bone resorption is believed to be the cause of this complication (31). Attenuation of the debrisassociated inflammation and retardation of implantsurrounding bone resorption would thus appear to hold considerable promise as a therapeutic strategy. As first noted in 1990, the in vitro maturation of macrophages into osteoclasts requires the presence of marrow stromal cells or their osteoblast progeny (32). It is now clear that these accessory cells express the 2 molecules that are essential and sufficient to promote osteoclastogenesis. One of these 2 molecules is recognized as the osteoclast differentiation factor (RANKL). It binds its physiologic signaling receptor, RANK, on membranes of macrophages and osteoclast precursors, thereby providing signals required for their survival, maturation, and activation (14). OPG has been identified as a potent negative regulator of the osteoclast differentiation factor, which is also naturally produced in stromal and other cells. Indeed, it is the balance between the expression of the stimulator of osteoclastogenesis, RANKL, and the inhibitor, OPG, that dictates the quantity of bone resorbed (33). Thus, OPG may be an effective therapeutic candidate to retard the process of debris-associated inflammatory osteoclastic bone resorption associated with loosening of the prosthetic joint. The major problems in the approach to treating a localized chronic inflammatory disorder such as debrisassociated aseptic loosening include the lack of adequate suppressive agents and effective specific therapeutic delivery systems. Viral vector–mediated gene transfer provides a novel approach to delivering the antiinflammatory genes to the site of disease to produce therapeutic proteins in a persistent and localized manner (34). Using retroviral vectors to code the IL-1 receptor antagonist, we and others have successfully attenuated orthopedic particle–induced inflammation and experimental
2522
apeutic approaches to debris-induced inflammation/ osteolysis, such as gene therapy. We have previously found that the peak of bone resorption responses to polyethylene particles occurs 6–10 days following particle stimulation in this model; hence, the selection of 7 days post–viral infection (8 days after UHMWPE particle injection in the bone-pouch model) in the current study to evaluate the gene transfer effects on particleinduced osteolysis. As stated elsewhere (25), this model does have several limitations. One of the most notable is the lack of blood supply to implanted bone. This limits the model to the study of acute osteolysis, rather than the chronic osteolysis seen in aseptic loosening. Different types of bone implants were used in this model, and while femoral implants provide assessment of a typical articulating bone, they cannot be sectioned without decalcification, which affects some immunologic or histologic determinations such as TRAP staining. While calvarium implantation can be adopted for the latter purposes, this bone is not anatomically typical of bony surfaces involved in aseptic loosening. Despite these limitations, this model, under strictly controlled experimental conditions, appears to be useful in the basic in vivo screening evaluation of gene transfer for wear debris– associated adverse tissue response and bony resorption. The data presented here clearly indicate that transduced OPG efficiently inhibits inflammatory bone resorption by blocking osteoclast maturation and infiltration. It results in significantly decreased gene expression of CTR, cathepsin K, and RANK, all of which are located on the osteoclast membranes (38), indicating a decrease in intensity and speed of osteoclastogenesis. The reduced number of TRAP-positive cells present in pouches treated with OPG also supports this observation. As a result, bone degeneration and resorption are significantly attenuated (less bone calcium release and bone collagen loss), and bone pit formation is protected. Although this study did not generate conclusive information on whether overexpression of OPG influences the initial particle-induced tissue inflammatory responses, transduction of hOPG in this model appears to inhibit IL-1 expression. This finding is interesting and novel, and might be due to OPG-related apoptosis of macrophage/preosteoclasts and osteoclasts that produce IL-1 (37,39). Overall, this study suggests that the murine air pouch model of bone resorption is a useful tool to screen therapeutic approaches to debris-associated osteolysis, and that gene transfer using rAAV-hOPG appears to be a feasible and effective candidate to treat or prevent
YANG ET AL
wear debris–associated osteolysis and aseptic loosening. Further studies are under way to help us understand the transducive efficacy and long-term effects of the AAVOPG gene transfer and related therapeutic mechanisms and safety concerns. ACKNOWLEDGMENTS The authors would like to thank Drs. William Boyle and Colin Dunstan for their thoughtful insights.
REFERENCES 1. Kawamoto K, Hasegawa Y, Iwase T, Iwasada S, Kanamono T, Iwata H. Failed cementless total hip arthroplasty for osteoarthrosis due to hip dysplasia: a minimum five-year follow-up study. Bull Hosp Jt Dis 1998;57:130–5. 2. Kobayashi S, Eftekhar NS, Terayama K, Joshi RP. Comparative study of total hip arthroplasty between younger and older patients. Clin Orthop 1997;339:140–51. 3. Kim YH, Kim VE. Cementless porous-coated anatomic medullary locking total hip prostheses. J Arthroplasty 1994;9:243–52. 4. Clohisy JC, Harris WH. The Harris-Galante uncemented femoral component in primary total hip replacement at 10 years. J Arthroplasty 1999;14:915–7. 5. Toumbis CA, Kronick JL, Wooley PH, Nasser S. Total joint arthroplasty and the immune response. Semin Arthritis Rheum 1997;27:44–7. 6. McGee MA, Howie DW, Neale SD, Haynes DR, Pearcy MJ. The role of polyethylene wear in joint replacement failure. Proc Inst Mech Eng [H] 1997;211:65–72. 7. Boynton E, Waddell JP, Morton J, Gardiner GW. Aseptic loosening in total hip implants: the role of polyethylene wear debris. Can J Surg 1991;34:599–605. 8. Buly RL, Huo MH, Salvati E, Brien W, Bansal M. Titanium wear debris in failed cemented total hip arthroplasty: an analysis of 71 cases. J Arthroplasty 1992;7:315–23. 9. Goodman SB, Fornasier VL, Kei J. Quantitative comparison of the histological effects of particulate polymethylmethacrylate versus polyethylene in the rabbit tibia. Arch Orthop Trauma Surg 1991;110:123–6. 10. Goodman SB, Huie P, Song Y, Schurman D, Maloney W, Woolson S, et al. Cellular profile and cytokine production at prosthetic interfaces: study of tissues retrieved from revised hip and knee replacements. J Bone Joint Surg Br 1998;80:531–9. 11. Shanbhag AS, Jacobs JJ, Black J, Galante JO, Glant TT. Cellular mediators secreted by interfacial membranes obtained at revision total hip arthroplasty. J Arthroplasty 1995;10:498–506. 12. Goldring SR, Schiller AL, Roelke M, Rourke CM, O’Neil DA, Harris WH. The synovial-like membrane at the bone-cement interface in loose total hip replacements and its proposed role in bone lysis. J Bone Joint Surg Am 1983;65:575–84. 13. Cooper RA, McAllister CM, Borden LS, Bauer TW. Polyethylene debris-induced osteolysis and loosening in uncemented total hip arthroplasty: a cause of late failure. J Arthroplasty 1992;7:285–90. 14. Takayanagi H, Ogasawara K, Hida S, Chiba T, Murata S, Sato K, et al. T-cell-mediated regulation of osteoclastogenesis by signalling cross-talk between RANKL and IFN-␥. Nature 2000;408:600–5. 15. Teitelbaum SL. Bone resorption by osteoclasts. Science 2000;289: 1504–8. 16. Takahashi N, Udagawa N, Suda T. A new member of tumor necrosis factor ligand family, ODF/OPGL/TRANCE/RANKL, regulates osteoclast differentiation and function. Biochem Biophys Res Commun 1999;256:449–55.
OPG GENE TRANSFER PROTECTS AGAINST OSTEOLYSIS
2523
17. Li J, Sarosi I, Yan XQ, Morony S, Capparelli C, Tan HL, et al. RANK is the intrinsic hematopoietic cell surface receptor that controls osteoclastogenesis and regulation of bone mass and calcium metabolism. Proc Natl Acad Sci U S A 2000;97:1566–71. 18. Kim N, Odgren PR, Kim DK, Marks SC Jr, Choi Y. Diverse roles of the tumor necrosis factor family member TRANCE in skeletal physiology revealed by TRANCE deficiency and partial rescue by a lymphocyte-expressed TRANCE transgene. Proc Natl Acad Sci U S A 2000;97:10905–10. 19. Simonet WS, Lacey DL, Dunstan CR, Kelley M, Chang MS, Luthy R, et al. Osteoprotegerin: a novel secreted protein involved in the regulation of bone density. Cell 1997;89:309–19. 20. Evans CH, Robbins PD. Possible orthopaedic applications of gene therapy. J Bone Joint Surg Am 1995;77:1103–14. 21. Xiao X, Li J, Samulski RJ. Production of high-titer recombinant adeno-associated virus vectors in the absence of helper adenovirus. J Virol 1998;72:2224–32. 22. Goater J, Muller R, Kollias G, Firestein GS, Sanz I, O’Keefe RJ, et al. Empirical advantages of adeno associated viral vectors for in vivo gene therapy for arthritis. J Rheumatol 2000;27:983–9. 23. Wooley PH, Morren R, Andary J, Sud S, Yang SY, Mayton L, et al. Inflammatory responses to orthopaedic biomaterials in the murine air pouch. Biomaterials 2002;23:517–26. 24. Yang SY, Ren W, Mayton L, Nasser S, Wooley PH. A novel model for osteolysis in aseptic loosening: femoral implants within the inflammatory murine air pouch [abstract]. Arthritis Rheum 2000;43 Suppl 9:S214. 25. Ren W, Yang SY, Wooley PH. A novel model of orthopaedic wear debris-associated osteolysis. Submitted for publication. 26. Wooley PH, Petersen S, Song Z, Nasser S. Cellular immune responses to orthopaedic implant materials following cemented total joint replacement. J Orthop Res 1997;15:874–80. 27. Grynkiewicz G, Poenie M, Tsien RY. A new generation of Ca2⫹ indicators with greatly improved fluorescence properties. J Biol Chem 1985;260:3440–50. 28. Carson FL. Connective and muscle tissues. In: Carson FL, editor. Histotechnology: a self-instructional text. Chicago: ASCP Press; 1990. p. 139–65. 29. Suzuki M, Suzuki M, Uetsuka K, Shinozuka J, Nakayama H, Doi K. Changes in location and number of tartrate-resistant acid phosphatase (TRAP)-positive cells during the development of
type II collagen-induced arthritis in DBA/1J mice. Exp Anim 1998;47:211–4. Sud S, Yang SY, Evans CH, Robbins PD, Wooley PH. Effects of cytokine gene therapy on particulate-induced inflammation in the murine air pouch. Inflammation 2001;25:361–72. Goodman SB. Does the immune system play a role in loosening and osteolysis of total joint replacements? J Long Term Eff Med Implants 1996;6:91–101. Udagawa N, Takahashi N, Akatsu T, Tanaka H, Sasaki T, Nishihara T, et al. Origin of osteoclasts: mature monocytes and macrophages are capable of differentiating into osteoclasts under a suitable microenvironment prepared by bone marrow-derived stromal cells. Proc Natl Acad Sci U S A 1990;87:7260–4. Hofbauer LC. Osteoprotegerin ligand and osteoprotegerin: novel implications for osteoclast biology and bone metabolism. Eur J Endocrinol 1999;141:195–210. Evans CH, Ghivizzani SC, Kang R, Muzzonigro T, Wasko MC, Herndon JH, et al. Gene therapy for rheumatic diseases. Arthritis Rheum 1999;42:1–16. Yang SY, Song Z, Robbins PD, Evans CH, Wu B, Mayton L, et al. Anti-inflammatory cytokine gene therapy in a murine model of aseptic loosening [abstract]. Arthritis Rheum 1999;42 Suppl 9:S284. Makarov SS, Olsen JC, Johnston WN, Anderle SK, Brown RR, Baldwin AS Jr, et al. Suppression of experimental arthritis by gene transfer of interleukin 1 receptor antagonist cDNA. Proc Natl Acad Sci U S A 1996;93:402–6. Hofbauer LC, Neubauer A, Heufelder AE. Receptor activator of nuclear factor-B ligand and osteoprotegerin: potential implications for the pathogenesis and treatment of malignant bone diseases. Cancer 2001;92:460–70. Tsurukai T, Udagawa N, Matsuzaki K, Takahashi N, Suda T. Roles of macrophage-colony stimulating factor and osteoclast differentiation factor in osteoclastogenesis. J Bone Miner Metab 2000;18:177–84. Murakami T, Yamamoto M, Ono K, Nishikawa M, Nagata N, Motoyoshi K, et al. Transforming growth factor-1 increases mRNA levels of osteoclastogenesis inhibitory factor in osteoblastic/stromal cells and inhibits the survival of murine osteoclast-like cells. Biochem Biophys Res Commun 1998;252:747–52.
30. 31. 32.
33. 34. 35.
36.
37.
38.
39.
