Histochem Cell Biol (2001) 116:69–77 DOI 10.1007/s004180100305
O R I G I N A L PA P E R
Robert Dinser · Florian Kreppel · Frank Zaucke Christoph Blank · Mats Paulsson · Stefan Kochanek Patrik Maurer
Comparison of long-term transgene expression after non-viral and adenoviral gene transfer into primary articular chondrocytes Accepted: 15 June 2001 / Published online: 5 July 2001 © Springer-Verlag 2001
Abstract Different gene transfer approaches to achieve long-term transgene expression in cultured primary bovine chondrocytes were compared using enhanced green fluorescent protein (EGFP) as a reporter. Transduction with a high-capacity adenoviral vector was 82% efficient when analysed by fluorescence microscopy, while up to 42% of plasmid-transfected cells were EGFP positive with FuGene as a transfection reagent. Rapid dominant marker selection of plasmid-transfected cells was achieved in monolayer culture. With either method of gene transfer, a high proportion of the chondrocytes remained transgene positive during prolonged alginate culture. Transgene transcription in single cells was quantified with a confocal laser scanning microscope. Detection of EGFP expression was more sensitive with this method, identifying more transgene-expressing cells than conventional fluorescence microscopy. Long-term EGFP expression was higher in adenovirally transduced chondrocytes embedded in alginate as compared to plasmidtransfected cells cultured in monolayer or in alginate. Both the adenoviral and the plasmid-based approach appear suited for studies of the molecular and cellular mechanisms by which mutations in cartilage matrix proteins cause disease. Keywords Transfection · Chondrocytes · Adenovirus · Plasmid · Alginate R. Dinser (✉) · F. Zaucke · C. Blank · M. Paulsson · P. Maurer Institute for Biochemistry II, Medical Faculty, University of Cologne, Joseph-Stelzmann-Strasse 52, 50931 Cologne, Germany e-mail:
[email protected] Tel.: +49-6841-1623088, Fax: +49-6841-1623069 F. Kreppel · S. Kochanek Center for Molecular Medicine (ZMMK), University of Cologne, Joseph-Stelzmann-Strasse 52, 50931 Cologne, Germany R. Dinser Medizinische Klinik I, Universitätskliniken des Saarlandes, 66421 Homburg, Germany
Introduction While many forms of skeletal malformations and osteoarthritis are genetically determined (Spector et al. 1996; Cicuttini and Spector 1997), we still lack insight into the pathogenic mechanisms causing defective morphogenesis of cartilage or its premature degeneration. Investigations using patient tissue are hampered by the fact that only small biopsies can be taken. In the case of osteoarthritis, samples are usually obtained only when the osteoarthritic process has already severely damaged the articular cartilage. Additionally, many single mutations causing skeletal disease are very rare (Cicuttini and Spector 1997; Briggs et al. 1998). Thus, there is a need for tissue culture models where the mechanisms by which mutant proteins cause cartilage abnormalities can be studied. Such models may be obtained by the transgenic expression of mutant proteins in otherwise phenotypically normal chondrocytes. In spite of the great interest, the optimal approach to long-term transgene expression in chondrocytes has not been established. Adenoviral gene transfer into primary chondrocytes has been reported by several groups (Baragi et al. 1997; Kang et al. 1997; Doherty et al. 1998), but it is unclear whether the parachromosomal adenoviral DNA is retained in long-term culture. An alternative, plasmid transfection, was until recently considered technically impossible (Johnson et al. 1999; Madry and Trippel 2000). Even if feasible, the lower efficiency of gene transfer achieved with plasmids usually necessitates the use of a selection marker to enrich transgenepositive cells, with the consequence that a longer period in monolayer culture is required relative to adenovirally transduced cells. A difficulty encountered with chondrocytes is that they dedifferentiate during monolayer culture, losing not only their typical morphology but also altering their biosynthetic repertoire. Three-dimensional matrices like alginate are known to stabilise the typical secretory phenotype of chondrocytes, but the effect of prolonged monolayer culture is only partially reversible (Benya and
70
Shaffer 1982; Zaucke et al. 2001). Repeated subcultures should therefore be avoided, if effects of transgenic protein expression on cartilage matrix composition are to be studied. For this reason, we have comparatively evaluated adenoviral and plasmid-based methods to achieve longterm transgene expression in primary chondrocytes using enhanced green fluorescent protein (EGFP) as a reporter. To assess long-term transgene expression, cells were transferred into alginate either directly after adenoviral gene transfer or after a short enrichment period following plasmid transfection.
Materials and methods Materials Low viscosity alginate (Keltone LV) was obtained from Kelco, Chicago, Ill., USA, collagenase P, FuGene from Roche, Mannheim, Germany, hygromycin and pronase from Calbiochem, Bad Soden, Germany, Effectene and Superfect from Qiagen, Hilden, Germany, geneticin (G418), puromycin, DMEM-F12, and fetal bovine serum from Gibco, Eggenstein, Germany. Other reagents were purchased from Sigma, Deisenhofen, Germany. Immunofluorescence chamber slides were obtained from Nalge Nunc, Naperville, Ill., USA. Isolation of primary chondrocytes Bovine shoulder joints of adult animals were obtained from the local slaughterhouse. Full thickness cartilage was collected from the head of the humerus and isolated overnight by sequential pronase and collagenase P digestion (Kuettner et al. 1982). Cartilage slices were treated at 37°C for 1.5 h with 0.1% pronase in DMEM/F-12 medium supplemented with 5% fetal calf serum. The cartilage was thoroughly washed and incubated at 37°C for 15–18 h with 0.01% collagenase in DMEM/F-12 supplemented as above. Alternatively, the concentrations of pronase and collagenase were increased to 0.4% and 0.025%, respectively. Cells were cultured in monolayer as described below until the day before transfection, when the cells were harvested by trypsinisation and plated at a density of 300,000 cells per well in six-well plates or 4.5×106 cells in 67-mm dishes. Vector constructs The cDNA encoding EGFP was cloned from the pEGFP-N1 plasmid (Clontech, Heidelberg, Germany) by HindIII/NotI digestion into the equivalent sites of the pRc-gen, pCEP-hyg and pCEP-pur plasmids (Fig. 1). The transcription thus remained under the control of the CMV immediate early promoter. The pRc-gen (pRc/CMV; Invitrogen, Groningen, The Netherlands) plasmid is a eukaryotic expression vector with a neomycin resistance similar to the pcDNA3 plasmid. The pCEP-hyg (pCEP-4; Invitrogen) plasmid is an episomal vector based on the EBV origin of replication and nuclear antigen (EBNA-1) and includes a hygromycin resistance gene. The pCEP-pur plasmid derives from the pCEP-hyg as described earlier (Kohfeldt et al. 1997), with the difference that the hygromycin resistance gene has been replaced by a puromycin resistance gene. Plasmid DNA was propagated in DH5α-competent cells. DNA was prepared using the Nucleobond AX100 kit (Macherey Nagel, Düren, Germany). The plasmid pFK7 used to construct the adenoviral vector AdFK7 (Fig. 1) was generated by inserting the AflIII/AflII fragment (blunt-ended with Klenow) from pEGFP-N1 into the bluntended NotI site of pSTK129 (Kochanek, unpublished observations), a shuttle plasmid that can be used to generate high-capacity
Fig. 1A–C Structure of plasmids. All plasmids contain the enhanced green fluorescent protein (EGFP)-expression cassette under the control of the CMV immediate early promoter and the SV40 or bovine growth hormone polyadenylation signal. Plasmid pRc-gen (A), 6.2 kb, contains the geneticin (gen) resistance cassette, plasmids pCEP-pur and pCEP-hyg (B), 10.4 kb, contain a puromycin (pur) or hygromycin (hyg) resistance cassette and the EBNA-1 gene with origin of replication (oriP) for episomal replication. The plasmid used to generate the adenoviral vector AdFK7, pFK7 (C), 31 kb, has a 29-kb insert in the multiple cloning site of pBluescript KSII, which contains, from left to right, the left inverted terminal repeat (ITR), the packaging signal (ψ), a 20-kb fragment from the human HPRT locus, the EGFP expression cassette, a 6.5-kb fragment of C346, and the right ITR adenoviral vectors. pSTK129 consists of the left terminus of the adenovirus type 5 (nt 1–440), a 20-kb DNA fragment derived from the human HPRT locus (Edwards et al. 1990; locus: HUMHPRTB, gene map positions 1777–21729), a NotI-cloningsite, a 6.5-kb human fragment of C346 (locus: HUMDXS455 A, cosmid map positions 10205–16750) and the right terminus of adenovirus type 5 (nt 35818–35935). The locations of the inverted terminal repeats (ITR) and the packaging signal (ψ) are indicated in Fig. 1. The insert of pFK7 is flanked by PmeI restriction sites. To rescue the high-capacity AdFK7 vector, 5 µg pFK7 were cleaved with PmeI and transfected into 293cre66 cells (Schiedner, unpublished observations) that were subsequently infected with helper virus AdLC8luc (Parks et al. 1996) at a multiplicity of infection (MOI) of five. After complete cytopathic effect, the medium and infected cells were harvested and freeze-thawed to release the virus. Aliquots of the crude vector lysate were serially passaged through 293cre66 cells as described (Parks et al. 1996; Parks and Graham 1997; Schiedner et al. 1998). The yield of AdFK7 after CsCl equilibrium density centrifugation was 2×1012 particles as determined by OD260. The infectious titre (1×107 green-forming units per µl) was determined in triplicate experiments by infecting HeLa cells with different numbers of particles and counting the resulting green cells using a fluorescence microscope. Transfection procedures Optimal gene transfer conditions were determined for each method using 300,000 cells per well in six-well plates. EGFP positivity was evaluated 72 h after gene transfer. FuGene is described by the manufacturer as a non-liposomal blend of lipids. The reagent was added to 100 µl serum-free culture medium and incubated for 5 min at room temperature. The mixture was transferred to another tube containing the vector DNA and incubated for 20 min at room temperature. The reagent–DNA mixture was then applied dropwise to the cell cultures. DNA amounts varied from 0.5 to 2 µg and the quantity of reagent from 3 to 9 µl. In one set of experiments, hyaluronidase was added to the culture medium 12 h before transfection until 24 h after transfection at a concentration of 4 U/ml medium (Madry and Trippel 2000). Effectene is described by the manufacturer as a non-liposomal lipid mixture which is provided with a DNA-condensing “enhancer” solution. From 0.1 to 2 µg DNA were used with 0.4–16 µl “enhancer” solution and 5–10 µl reagent. The vector DNA was diluted in 100 µl of the in-
71 cubation buffer provided by the manufacturer. The “enhancer” solution included in the transfection kit was added, followed by 5 min of incubation at room temperature. The transfection reagent was then applied to the mixture, which was vortexed for 10 s and incubated for 10 min at room temperature. Afterwards, the transfection mixture was added in a dropwise fashion to the culture medium. Superfect is a reagent based on activated polyamidoamine dendrimers (Tang et al. 1996). It was tested with 0.5–2.5 µg DNA and 5–10 µl reagent solution. The vector DNA was diluted with 100 µl serum-free medium before the reagent was added. The mixture was vortexed for 10 s followed by 10 min of incubation at room temperature. The DNA–reagent complexes were then added dropwise to the culture supernatants. The DNA/reagent/medium mixture was replaced by normal medium 4 h after transfection for Superfect or the following day for FuGene and Effectene. Electroporation was performed with the GenePulser from Biorad (Hercules, Calif., USA), using 3–10 µg DNA, and varying the capacitance from 500 to 950 µF and the voltage from 200 to 500 V. Time constants were between 6 and 12 ms. A ballistic (“gene gun”) approach (Sanford et al. 1993; Yang and Sun 1995) was tested using 35 µg DNA on a gold microcarrier 0.8–1.6 µm in diameter and applying 2,200 psi rupture pressure per 67-mm dish with the PDS1000/He-system (Biorad). Initial experiments with the high-capacity adenoviral vector AdFK7 were performed with 4–100 MOI of vector. Cells were rinsed with PBS prior to infection. Infections were performed in 0.75 ml fresh culture medium at 37°C. The amount of medium was increased to 2 ml after 2 h of incubation at 37°C. Medium was replaced 16–24 h after infection. Optimal results using FuGene were obtained using 1 µg DNA per 9 µl reagent for 300,000 cells per well in six-well plates. For electroporation, 10 µg DNA was used with a capacitance of 950 µF and a voltage of 400 V. Best conditions for Effectene were 2 µg DNA, 16 µl enhancer and 10 µl reagent. AdFK7 was used at 100 MOI for long-term experiments. For long-term experiments with transfer of plasmid-transfected cells into alginate, cell number, amount of DNA and FuGene were scaled up in proportion to the culture dish. Cell culture and selection conditions Cells were cultured in DMEM/F-12 including HEPES, pyridoxine-HCl and glutamate, and supplemented with 10% fetal bovine serum, 100 µM ascorbic acid and 50 mg/l gentamycin. Medium was replaced twice weekly, and cells in monolayer were split as necessary, usually once per week. Embedding of cells in alginate was performed as described (Häuselmann et al. 1994). Briefly, cells were suspended in 0.15 M NaCl containing 1.2% low-viscosity alginate at a density of 4×106 cells/ml. This suspension was added dropwise to a 0.1 M CaCl2 solution through a 22-gauge needle. After allowing the beads to gel for 10 min, they were washed several times in 0.15 M NaCl and DMEM/F-12. To determine the adequate concentrations of the selection antibiotics, untransfected chondrocytes were cultured in monolayer for 1, 5 or 9 weeks before being transferred to a 48-well plate at a density of 25,000 cells/well. The cells were allowed to adhere and then subjected to different concentrations of the antibiotics, ranging from 0 to 2 µg/ml puromycin, 0–500 µg/ml geneticin and 0–300 µg/ml of hygromycin. Medium was changed twice weekly for 2 weeks. The proportion of viable cells was assessed using a tetrazolium salt (MTT) colorimetric assay (Mosmann 1983). For selection of transfected cells in monolayer, treatment was started 72 h after transfection, using 500 µg/ml geneticin, 1–2 µg/ml puromycin or 100–150 µg/ml hygromycin. Upon transfer into alginate, no selection antibiotics were used. Fluorescence microscopy Cells were trypsinised 72 h after transfection and counted in a Neubauer chamber with a fluorescence microscope (Axiophot; Zeiss, Jena, Germany) using an excitation range from 450 to 490 nm and an emission range from 515 to 565 nm. Total cell numbers were assessed using differential interference contrast.
For long-term assessment, cells were observed and photographed while remaining in the culture dish, using an inverted fluorescence microscope (Olympus IX 70; Olympus, Hamburg, Germany) with a 460- to 490-nm excitation filter and a 515- to 550-nm emission filter. For exact determination of cell numbers, monolayer cells were counted in a Neubauer chamber. Cells cultured in alginate were isolated by dissolution of the beads with 55 mM sodium citrate, 0.15 M NaCl, pH 6 at 37°C for 20 min (Häuselmann et al. 1994) followed by treatment with trypsin/ EDTA (0.05/0.02% w/v) solution for a minimum of 30 min before the cells were counted in a Neubauer chamber. Alternatively, trypsinised cells were recovered by centrifugation, plated on 0.1% gelatine-treated glass chamber slides and allowed to adhere for 3 days in serum-free medium. Slides were fixed with 2% paraformaldehyde for 10 min and a nuclear counterstain with the bisbenzimide dye was performed (1 µg/ml in methanol for 15 min at 37°C). Confocal laser scanning microscopy Fixed cells were evaluated with the spectral confocal microscope Leica TCS SP (Leica, Heidelberg, Germany) using an argon laser with an excitation wavelength of 488 nm and an emission range of 490–550 nm. Data were analysed with the TCS NT software. Different sensitivity settings were employed, allowing in one setting the evaluation of the total cell number from the autofluorescence of chondrocytes and in the other setting the evaluation of EGFPpositive cells. Semiquantitative fluorescence intensity assessments were made with the help of the TCS NT software.
Results Efficiency of gene transfer Three plasmid-based and one adenoviral vector with identical expression cassettes encoding the gene for the EGFP from the CMV promoter were used for gene transfer into primary bovine chondrocytes (Fig. 1). The plasmids were pRc-gen, a vector designed to achieve stable transfection, and the episomal vectors pCEP-pur and pCEP-hyg. AdFK7 is a high-capacity (“gutless”) adenoviral vector. After an initial optimisation of transfection procedures (see Materials and methods), efficiencies for the different constructs and transfection methods were compared at an early timepoint (72 h) after transfection (Table 1). The highest efficiency, with up to 82% of all cells being EGFP positive, was obtained using the high-capacity adenoviral vector AdFK7 (Table 1). Transduction with this vector showed a clear dose dependence. Plasmid vectors could also be efficiently transfected using different approaches. Most efficient, with respect to both the number of EGFP-positive cells and fluorescence intensity, was the non-liposomal reagent FuGene. Other commercially available reagents were inferior in performance to conventional electroporation. The efficiency using FuGene was almost doubled by using higher concentrations of pronase and collagenase for the cell isolation from cartilage (0.4% instead of 0.1% pronase, 0.025% instead of 0.01% collagenase). Additional digestion by hyaluronidase before and during transfection (Madry and Trippel 2000) did not further improve efficiency (results not shown). Chondrocytes were usually transfected 7 days after isolation. Transfec-
72 Table 1 Comparison of the efficiency of transfection using different vectors and transfection methods. Primary bovine chondrocytes were cultured in monolayers for 1 week after isolation and then transfected with vectors expressing EGFP from the CMV promoter. pCEP-pur denotes an episomal plasmid vector, pRc-gen a plasmid designed for stable transfection, AdFK7 an adenoviral vector. (MOI Multiplicity of infection) Construct
Transfection method
Efficiency Number of (Percent ± SD) experiments
FuGene6a 32±1.5 17±1.2 5 6±1.2 5 6±3.3 3
