Lipofection of Cultured Mouse Muscle Cells: a Direct Comparison of Lipofectamine and DOSPER

Gene Therapy (1998) 5, 542–551  1998 Stockton Press All rights reserved 0969-7128/98 $12.00 http://www.stockton-press.co.uk/gt Lipofection of cultured mouse muscle cells: a direct comparison of Lipofectamine and DOSPER

E Dodds1, MG Dunckley1, K Naujoks2, U Michaelis2 and G Dickson1 1Division of Biochemistry, School of Biological Sciences, Royal Holloway University of London, Egham, Surrey TW20 0EX, UK; and 2Division of Therapeutics, Boehringer Mannheim GmbH, Penzberg, Germany

Cationic lipid–DNA complexes (lipoplexes) have been (GFP) were used in order to evaluate transfection widely used as gene transfer vectors which avoid the efficiency by histochemical staining or FACS analysis, adverse immunogenicity and potential for viraemia of viral respectively. Both lipid formulations were able to promote vectors. With the long-term aim of gene transfer into skel- efficient, reproducible gene transfer in C2C12 cells, and etal muscle in vivo, we describe a direct in vitro comparison to transfect primary mouse myoblast cultures successfully. of two commercially available cationic lipid formulations, However, DOSPER exhibited the important advantage of Lipofectamine and DOSPER. Optimisation of transfection being able to transfect cells in the presence of serum of was performed in the C2C12 mouse muscle cell line, both bovine and murine origin. This feature allowed before further studies in primary mouse myoblasts and increased cell survival during in vitro transfections, and C2C12 myotubes. Reporter gene constructs expressing may be advantageous for direct in vivo gene transfer either E. coli ␤-galactosidase or green fluorescent protein efficacy.

Keywords: gene transfer; lipoplex; lipofection; skeletal muscle

Introduction biosafety issues and adverse immune responses have led to the development of a range of alternative, nonviral Skeletal muscle provides an excellent target for the systems, where repeated administration is possible.4 expression of recombinant genes as its structure and Some of the most promising results have come from the function are well understood. It represents 30–50% of use of cationic lipids which are able to complex with total body mass and its easy accessibility lends itself well nucleic acid to form a lipoplex and deliver genes to the to the study of gene transfer techniques. A further advan- nucleus via the cellular endocytosis pathway, without tageous characteristic is the post-mitotic state of mature significant toxicity, immunology or germ cell localis- myofibres, which may facilitate long-term persistence of ation.5–7 The process of lipoplex formation involves elec- episomal transgenes. This feature is especially relevant to trostatic interactions between the cationic lipid and liposome-mediated gene transfer where stable inte- anionic DNA, as opposed to traditional liposome techno- gration of a transgene into the host genome does not nor- logies which result in encapsulation of DNA.8 Since their mally occur at a high frequency, so genetic material can introduction 10 years ago,9 many cationic lipid formu- be rapidly lost from transfected cells during repeated lations have been tested in a variety of in vitro situations rounds of cell division. Gene transfer into muscle has and have often yielded highly efficient transfection,10,11 been proposed for the treatment of inherited and superior to that obtained with encapsulating liposomes.12 acquired disorders both of muscle and nonmuscle tissues. However, studies administering lipoplexes to a range of For example, pre-clinical studies are currently underway tissues in vivo have led to mixed results. It seems that to try and alleviate the symptoms in animal models of cationic lipids may be suited to certain tissue types such primary muscle myopathies such as Duchenne muscular as lung or tumour13–15 where clinical trials are now in 1 dystrophy, using viral and nonviral techniques. A progress.16,17 Muscle presents a special case, where direct second major therapeutic application is to utilise muscle injection of naked plasmid DNA has resulted in signifi- tissues as an environment to express recombinant, non- cant levels of expression.18 By complexing the DNA with muscle proteins actively. This approach has been used cationic lipid, it may be possible to produce a more as a DNA vaccination system against, for example, the efficient and reproducible system of gene transfer into 2 hepatitis B virus, and as a means to augment levels of muscle, perhaps by promoting an alternative route of 3 serum proteins such as cytokines and lipoproteins. uptake. The concept of using lipoplexes as a method of High efficiency gene transfer into many eukaryotic DNA vaccination via intramuscular injection has already tissues and cell types has previously been achieved using been demonstrated.19 viral vectors. However, associated toxicity problems, As a prelude to direct in vivo gene transfer studies, we have compared the ability of two commercially available cationic lipids (Lipofectamine and DOSPER) to transfect Correspondence: G Dickson C2C12 cells, a mouse muscle cell line in vitro. Two differ- Received 8 August 1997; accepted 17 November 1997 ent reporter plasmids were utilised, first containing the Lipofection of muscle cells E Dodds et al 543 lacZ gene (pCMV␤), and second, the gene for green flu- transfection efficiency, all further studies were performed orescent protein (GFP; pUCUF2). Cultures were then using the lacZ reporter system and lipoplexes formed in assessed using histochemical staining or FACS analysis, these established optimal lipid:DNA ratios. respectively. Parameters were completely optimised in Lipoplexes formed in these optimal ratios were C2C12 cells, and then transfections into primary mouse additionally observed by negative staining TEM. Aggre- myoblasts and myotubes were performed using these gates of multilamellar vesicles were seen for both Lipo- optimal conditions, as these are likely to represent a fectamine–DNA and DOSPER–DNA lipoplexes (Figure closer model of the in vivo situation. In addition, formu- 3). The size of these complexes was extremely varied, lations suitable for in vivo injection were determined. with Lipofectamine–DNA complexes being significantly bigger than DOSPER–DNA complexes (100–830 nm and Ͻ Results 125–350 nm, respectively; P 0.05). Discrete lipid vesicles approximately 100 nm in diameter were observed To establish maximal DNA transfection efficiency using in controls with lipid alone. the cationic lipids, a range of parameters were optimised. The optimisation studies were all performed in cultured Stability of expression C2C12 cells using either the lacZ or GFP reporter genes The stability of transgene expression was studied in a set in the plasmid expression vectors pCMV␤ and pUCUF2, of transfections in which wells were fixed and stained respectively. Direct comparison of transfection with the at varying time-points (20–72 h) following transfection. two vectors showed no significant difference in the opti- Maximal expression (25.6% ± 2.94 and 17.5% ± 2.32 for mal conditions found, indicating that the nature of the Lipofectamine and DOSPER, respectively; means ± s.d.) plasmid used is largely irrelevant. All experiments were was achieved by 18 h and remained at an equivalent level carried out in parallel for Lipofectamine and DOSPER up to 72 h (results not shown). These experiments in and each condition was repeated in triplicate in inde- C2C12 myoblasts could not be extended past 72 h as pendent cultures for statistical significance. Lipofectam- myoblast differentiation and fusion began to take place. ine-mediated transfections were performed in serum-free In all further studies expression was analysed 24 h medium, and DOSPER-mediated transfections in 5% FCS after transfection. medium unless stated otherwise. Time of incubation of complexes with cells Lipid:DNA ratio of complex An incubation time of 6 h is recommended by the manu- Lipoplex formation is dependent upon the charge ratio facturers for both DOSPER and Lipofectamine transfec- between the DNA and lipid.8 Therefore, a range of tions. This was re-examined in the myoblast system to weight to weight ratios was tested using the GFP reporter establish the effect of reduced incubation times on trans- system and FACS analysis (Figure 1). Transfections were fection efficiency and toxicity (Figure 4). It was found that performed in serum-free medium with Lipofectamine transfection efficiency increased up to 4 h (41.7% ± 3.5 for and 5% FCS medium with DOSPER. Cells were counted Lipofectamine and 37.0% ± 4.3 for DOSPER), and by FACS analysis 24 h after transfection. This revealed remained stable with longer incubation of the lipoplexes that a DOSPER:DNA ratio of 2:1 and a Lipofectami- with the cells up to 6 h. Nevertheless, an incubation time ne:DNA ratio of 7.5:1 led to maximal efficiency (Figure of 2 h may be preferable as significant efficiencies of 2). These are equivalent to charge ratios 30.5% ± 1.7 with Lipofectamine and 28.3% ± 2.4 with (positive:negative) of 2.73 and 11.2, respectively, showing DOSPER were achieved, with the advantage of increased that both lipoplexes have an overall positive charge (see levels of cell survival (56% ± 8.7 and 45.3% ± 9.3 for Lipo- Materials and methods). The amount of lipoplex required fectamine and DOSPER, respectively, at 2 h, as opposed for maximal transfection efficiency was detected by fixing to 17.7% ± 2.1 and 24.3% ± 4.6 at 4 h). As both Lipofect- the weight:weight ratio and altering the amount of DNA amine- and DOSPER-mediated transfections were perfor- and therefore also lipid added. For both Lipofectamine- med in serum-free medium for this set of experiments, and DOSPER-mediated transfections in 35 mm dishes, 4 and control cells exposed to serum-free medium for 6 h ␮g of DNA led to maximal efficiencies (33% ± 5.74 and showed notable levels of toxicity, the level of cell death 23% ± 0.68 of gated cells, respectively). Increasing the perceived is not entirely due to the transfection reagent, amount of DNA further had no significant increase on but in part to the incubation in serum-free medium. efficiency (Figure 2) and although a slight increase in cell Therefore it was necessary to study the exact effect of toxicity was observed, cell survival remained high serum on transfection efficiency and cell survival. (results not shown). Full cell survival data are not shown for this set of experiments using the GFP reporter system, Effect of serum on transfection as FACS analysis proved to be a poor indicator of the In an effort to eliminate the cell death resulting from true level of cell death following transfection, as dis- incubation in serum-free medium, a comparison of trans- cussed later. Therefore, a comparison with an alternative fections performed in serum-free 5% FCS and 10% FCS reporter system was made, transfection with pCMV␤ fol- containing medium was made (Figure 5). Lipoplexes lowed by X-gal histochemical staining for ␤-galactosidase were formed with 4 ␮g DNA in a 7.5:1 or 2:1 ratio for activity. C2C12 myoblasts were transfected with pCMV␤ Lipofectamine and DOSPER, respectively, and incubated lipoplexes in the optimal ratios described above. with C2C12 myoblasts for 6 h. Lipofectamine had a sig- Efficiencies were expressed as the percentage of positive nificantly reduced transfection efficiency in 5% and 10% cells in relation to the total surviving cell number at the FCS DMEM, relative to serum-free medium, resulting in time of analysis, and values of 41.3% ± 1.3 and 35.9% ± 4.0 only 1.2% ± 0.9 and 0.5% ± 0.3 positive cells, respectively. for Lipofectamine and DOSPER, respectively, were Transfection with DOSPER was, however, less adversely obtained. As this represents a more accurate measure of affected by the presence of serum, as an efficiency of Lipofection of muscle cells E Dodds et al 544

Figure 1 Green fluorescent protein FACS analysis following transfection of C2C12 myoblasts with lipoplexes. Cells were initially evaluated by size (1) in order to select a region representing single cells. This gated region was further analysed for fluorescence. A negative control (cells transfected with lipid complexed with an irrelevant DNA (pCMV␤)) was used to set the position of quadrants separating GFP-positive cells (green; lower right), GFP- negative cells (nonfluorescent; lower left) and dead cells (red; upper left) (2). These quadrants were then applied to the analysis of experimental wells (3) and percentage cell number/ total cell number in gated region calculated for each quadrant. These procedures were followed for cells transfected with either Lipofectamine or DOSPER. (a) Transfection with Lipofectamine–pUCUF2 lipoplexes resulted in: GFP-positive cells (lower right), 37.19%, GFP- negative cells (lower left), 55.43% and dead cells (upper left), 7.10%. (b) Transfection with DOSPER–pUCUF2 lipoplexes resulted in: GFP-positive cells (lower right), 23.81%, GFP-negative cells (lower left), 68.09% and dead cells (upper left), 7.08%.

9.9% ± 1.9 was achieved in 5% FCS DMEM compared tions with different lipid:DNA ratios, was not clearly with 33.1% ± 3.6 positive cells in serum-free medium. The seen in primary cultures. Transfection of 5-day-old advantage of reduced toxicity was evident during these C2C12 myotubes proved to be inefficient, resulting in DOSPER-mediated transfections, with 83% ± 8.7 of cells only a small number of positives per well. The total num- surviving in the presence of serum, relative to 35% ± 4.2 ber of positive ␤-gal myotubes was marginally higher in in serum-free medium. Similar studies in the presence DOSPER-transfected cultures than those transfected with and absence of mouse serum led to even more pro- Lipofectamine (results not shown). nounced differences. In serum-free medium, efficiencies of 42% ± 9.2 and 41.9% ± 2.1 were obtained for transfec- Complex formation time and volume tions with Lipofectamine and DOSPER, respectively. This One long-term aim of our work is to assess the ability compares with 0.9% ± 0.1 and 31.8% ± 3.2 in the presence of Lipofectamine–DNA and DOSPER–DNA lipoplexes at of 10% mouse serum. It is important to note that the mediating transfection of muscle cells following direct decrease in percentage efficiency seen in the presence of intramuscular delivery in mice. When performing intra- mouse serum or FCS, does not simply reflect an increase muscular injections in the mouse, a physical limit on in cell survival, but the absolute number of positively injection volume per muscle exists. Therefore we exam- transfected cells actually decreases. ined whether it was possible to produce the lipoplex in a concentrated form, in order to inject a larger amount of Transfection of primary cultures and C2C12 myotubes DNA. Experiments were performed on C2C12 myoblasts, Primary mouse muscle cultures were transfected with and lipoplexes were formed using fixed amounts of DNA Lipofectamine–DNA and DOSPER–DNA lipoplexes in (pCMV␤) and lipid in varying volumes of HBS (Figure serum-free medium using 4 ␮g DNA in varying lip- 6). Equal volumes of lipid and DNA were mixed to give id:DNA ratios. Both lipids were able to mediate transfec- total volumes varying from 10 to 120 ␮l. Transfections tion, but at a lower level than achieved in the C2C12 cell were performed in serum-free medium using 4 ␮g line (Table 1). Also, the significant variation noted in pCMV␤ and 30 ␮g Lipofectamine, or in 5% FCS medium C2C12 cells between efficiencies resulting from transfec- using 2 ␮g pCMV␤ and 4 ␮g DOSPER. In the case of Lipofection of muscle cells E Dodds et al 545 volume of 60 ␮l over 1 h. Lipofectamine complexes, however, retained their activity relative to DOSPER complexes and mediated transfection of similar efficiencies if formed over 15 min or 1 h (Figure 6).

Discussion The efficacy of Lipofectamine as a mediator of transfection in vitro has been extensively documented.20,21 How- ever, reports of associated toxicity22 have led to the development of newer alternative cationic lipids. One such novel formulation is DOSPER, and here we describe a detailed comparison of Lipofectamine and DOSPER in their ability to transfect C2C12 and primary mouse muscle cells in culture. These studies have highlighted several differences between the two lipids, which have implications for the design of future large-scale transfections and in vivo applications. While there has been much debate concerning the mechanism of uptake of lipoplexes into the cell, it is now accepted that endocytosis is the main process involved.5 However, the initial interaction between lipoplex and cell membrane is electrostatic,8 hence the overall charge of the lipoplex is critical. For this reason, the first parameter to be optimised was the weight:weight ratio of positively charged lipid to negatively charged DNA. With both lipids tested, the optimal lipid:DNA ratio resulted in lipoplexes that bear an overall positive charge, enabling association with the negative charge of the cell membrane. Increasing the ratio further, by addition of more lipid, led to a decrease in resultant transfection efficiency, possibly due to the toxicity of excess lipid to cells. The Figure 2 Optimisation of lipid:DNA ratio (weight:weight) for (a) Lipofec- two different plasmids used, pCMV␤ and pUCUF2, gave tamine–pUCUF2 and (b) DOSPER–pUCUF2 lipoplexes. Transfections the same optimal lipid:DNA ratios, as a given quantity of were performed on cultured C2C12 myoblasts, using varying amounts of DNA, irrelevant of sequence, will have the same charge. lipid and DNA to investigate the amount of complex required for optimal Electron micrographs of lipoplexes formed in these opti- expression of the transgene. Efficiencies are expressed as percentage positive cells/total cell number in gated region. mal ratios show populations of lipid–DNA aggregates with significant variation in size. We found DOSPER lipoplexes to be generally smaller than Lipofectamine lipoplexes, which may be significant during cell uptake Lipofectamine–DNA lipoplexes, the volume of formation and in future in vivo administrations where large lipo- had little effect as resultant efficiencies were similar plexes may be physically restricted. regardless of whether a volume of 20 ␮l or 120 ␮l was Toxicity parameters have as much value as transfection adopted. However, DOSPER–DNA lipoplexes formed efficiency when assessing the effectiveness of a lipid in more efficiently and led to higher transfection efficiencies transfecting maximum numbers of cells. Toxicity data if a volume of 10 ␮l was used as opposed to 100 ␮l also provide valuable information required for the design (13.8% ± 0.6 and 6.9% ± 1.4, respectively; P Ͻ 0.01). In of improved lipid formulations. Use of the GFP reporter both cases these results are beneficial for in vivo experi- system23 allowed rapid analysis of transfection efficiency ments as a higher dose of lipoplex can be administered to as no additional histochemical staining was required. each muscle. It is important to note that the transfection However, although this method of FACS analysis gave efficiencies obtained with DOSPER–DNA lipoplexes in quantitative data for live, positive cells, it proved to be a this particular set of studies is lower than previously poor indicator of the true level of toxicity when com- seen. This is due to less DNA being used (2 ␮gas pared with direct microscopic examination. With FACS opposed to 4 ␮g), and the transfections being performed analysis, estimates of toxicity are inaccurate, as only the in the presence of 5% serum which has a slight inhibi- cells surviving the transfection procedure and trypsinis- tory effect. ation are measured. Those cells which detach during In order to ensure consistency during the injection pro- transfection are immediately discarded when the cedure, the stability of lipoplexes before addition to the medium containing the lipoplexes is replaced. It was cells was investigated. Lipoplexes were formed in HBS found that microscopic observation of the cultures pro- and held at room temperature for 15 min, 30 min and vided a better measurement of toxicity as the number of 1 h. This revealed that DOSPER–DNA complexes were cells remaining on a treated well in proportion to the less active at certain concentrations if left to form for number of cells on an untreated well can be calculated. longer than the 15 min suggested by the manufacturers. In this way, a realistic estimate of percentage cell survival The resultant transfection efficiency was reduced to as enabled true comparisons to be made. little as 2.8% ± 1.0 if complexes were formed in a total Our data revealed that there appear to be two aspects Lipofection of muscle cells E Dodds et al 546

Figure 3 Electron micrographs of lipoplexes. Lipoplexes were formed for 15 min with (a) 5 ␮g Lipofectamine and 0.66 ␮g pCMV␤ and (b) 5 ␮g DOSPER and 2.5 ␮g pCMV␤. Adsorbed lipoplexes were stained for 2 min with a 1% ammonium molybdate solution. Magniﬁcation × 120 000. Scale bar represents 100 nm. The range of complex sizes observed was 100–830 nm for Lipofectamine–DNA and 125–350 nm for DOSPER–DNA.

to the problem of cell survival, the toxicity of the lipo- DOSPER was able to mediate transfection efficiently in plexes themselves and the adverse effect of the cells being the presence of 5% serum, and was able to transfect cells exposed for various periods of time to serum-free at a low level even in the presence of 10% serum. This medium. We have demonstrated that treatment with contrasted with Lipofectamine which, unless used in complexes containing Lipofectamine or DOSPER in serum-free medium, was unable to induce any gene serum-free conditions is equally toxic to cells. When the transfer. This phenomenon of inactivation in the presence effect of serum deprivation was eliminated by transfect- of serum has previously been reported for other cationic ing cells in 10% FCS DMEM, only low levels of cell death lipids24 and is likely to be due to the neutralising effect were observed, at similar levels, for both Lipofectamine of anionic components of the serum. Although Lipofecta- and DOSPER. Thus it seems that the majority of cell mine promotes transfection at a higher efficiency relative death observed is due to the absence of serum from the to DOSPER in serum-free medium, this can be considered medium which may affect the toxicity of transfection in an artificial environment. Transfection efficiencies in the a number of ways. First, we found that cells incubated in presence of serum are more likely to mimic the in vivo serum-free medium containing no transfection reagents situation. In these circumstances, DOSPER proved to be showed significant levels of cell death as a result of them the superior transfection reagent, so should be pursued being deprived of growth factors vital for their survival. as a potential mediator of gene transfer in vivo. Second, charge neutralisation of lipoplexes and reduced Importantly, similar results were obtained following cellular uptake in the presence of serum, may also result transfection in the presence and absence of mouse serum in decreased apparent cell toxicity. Therefore it was desir- (results not shown). able to attempt to reduce the length of time in which cells As in vivo gene transfer conditions are generally more were subjected to lipoplexes in serum-free medium. We limited than in cell culture, other relevant parameters showed that the incubation time for both Lipofectamine- were examined. Lipoplex formation volumes and forma- and DOSPER-mediated transfections could be reduced to tion times are two such parameters requiring optimis- 2 h, as opposed to the 6 h recommended by the manufac- ation before intramuscular administration, where it is turers. This led to increased levels of cell survival, with desirable to produce the lipoplex in a more concentrated only minimal loss of transfection efficiency. form without loss of function. In a series of transfections In the light of these data, the ideal situation would be we varied the volume during complex formation. We to avoid serum-free transfection conditions altogether. found that smaller volumes for DOSPER–DNA com- Studies in serum-containing medium revealed a major plexes promoted transfection at a significantly higher difference between Lipofectamine and DOSPER. efficiency than that with larger volumes. However, vari- Lipofection of muscle cells E Dodds et al 547 to fuse into myotubes before transfection proved to take up DNA at a much lower efficiency than in myoblasts. However, the use of lipid for ex vivo myoblast gene transfer, followed by transplantation into muscle would avoid this problem.26 It is also possible to take advantage of the regenerative capacity of muscle, where new populations of myoblasts, termed satellite cells, could be transfected. This process of regeneration occurs naturally in some muscle disorders such as the muscular dystrophies. Transfection of primary mouse muscle cultures also yielded lower efficiencies, with no notable difference between Lipofectamine and DOSPER in serum-free medium as observed in C2C12 myoblasts. Unlike in C2C12 cells, variation of lipid:DNA ratios revealed no clear optimal ratio, but a more extensive programme of optimisation experiments in primary cultures is now in progress. This lack of variation is possibly due to the low efficiencies seen. In conclusion, muscle is an attractive gene therapy target tissue for expression of both endogenous and secreted products. Transgenic myofibres exhibit high stability characteristics and long-term persistence of gene expression. Gene transfer can be mediated via ex vivo or in vivo transfection of proliferating myoblasts which fuse into existing myofibres. Our studies have served to high- light the similarities and differences in the ability of Lipo- fectamine, and the less well tested lipid, DOSPER, to transfect muscle myoblasts in vitro. We have demonstrated them both to be capable of efficiently transfecting cultured myoblasts in serum-free medium. However, Lipofectamine is dramatically inhibited by the presence of serum, unlike DOSPER which is able to continue to mediate transfection of a high efficiency. DOSPER may therefore prove to be better suited to gene transfer in vivo, Figure 4 Effect of the time of incubation with DNA lipoplexes on transfection of C2C12 myoblasts in culture. Transfections were performed with and because the requirement for serum-free medium can Lipofectamine–DNA complexes formed with 4 ␮g pCMV␤ in a 7.5:1 lip- be avoided, large-scale transfections in vitro may be per- id:DNA ratio or DOSPER–DNA complexes formed with 4 ␮g pCMV␤ formed with reduced levels of cell death. In addition, in in a 2:1 lipid:DNA ratio. combination with virus vector or vector components, application to post-mitotic target cells such as myofibres may be feasible.27 ation of volume during the formation of Lipofectamine– DNA complexes had no notable effect on transfection Materials and methods efficiency. Our studies, in agreement with work pub- lished by Staggs et al,25 suggested that it is only the con- Cell culture centration during lipoplex formation which is critical, C2C12 cells are an adherent cell line initially cloned from and the final concentration of the mixture presented to mouse skeletal muscle.28 They were grown at 37°C and the cells is less important. When lipoplex formed in a 8% CO2 in Dulbecco’s modified Eagle’s medium (DMEM) fixed volume was added to cells in 1 ml, 2 ml, 3 ml or supplemented with 10% (v/v) foetal calf serum (FCS) 4 ml medium, there was little resultant variation in trans- and 2 mm glutamine. For efficient formation of myotubes, fection efficiency (results not shown). culture medium was switched to 5% heat inactivated Varying the time of complex formation revealed horse serum to promote cell differentiation and fusion. another main difference between Lipofectamine and Primary cell cultures were prepared by dissociation of DOSPER. At certain concentrations, DOSPER–DNA com- hind and fore limb muscle tissues of new born C57/B10 plexes lost some of their ability to transfect C2C12 cells mice as previously described.29 These cells were plated in if left to form for longer than 15 min, whereas Lipofecta- 15% FCS DMEM plus 2 mm glutamine, on 35 mm dishes mine–DNA complexes in general could be left without (Nunclon, from Life Technologies, Paisley, UK) that had significant effect for up to 1 h. One explanation is that been pre-coated in Matrigel (Becton Dickinson, Oxford, increased complex aggregation with time occurs in com- UK). Cultures were grown to 50–80% confluence and plexes containing DOSPER, to a greater extent than in transfected without further passage. those containing Lipofectamine, and that this increased clumping leads to a decrease in transfection efficiency. Plasmid constructs Further EM studies observing lipoplex sizes, may serve The pCMV␤ plasmid (Clontech, Palo Alto, California, to resolve these questions. USA) is a 7.2 kb eukaryotic expression construct contain- As observed previously,21,22 cultures of C2C12 cells left ing a lacZ gene and a CMV promoter. The pUCUF2 plas- Lipofection of muscle cells E Dodds et al 548

Figure 5 The effect of serum on the efﬁciency and toxicity of transfection of C2C12 myoblasts. Lipoplexes were formed with either Lipofectamine– pCMV␤ (4 ␮g DNA, 7.5:1 lipid:DNA ratio) or DOSPER–pCMV␤ (4 ␮g DNA, 2:1 lipid:DNA ratio). Transfections were performed in serum-free DMEM, or DMEM supplemented with 5% or 10% FCS and the effect on transfection efﬁciency (a) and cell survival (b) following transfection was studied.

shifting the excitation peak to a higher wavelength of Table 1 Transfections into primary mouse myoblast cultures 31 with Lipofectamine–pCMV␤ and DOSPER–pCMV␤ lipoplexes 490 nm. Cationic lipid formulations Ratio (wt:wt) % Efficiency Range of lipid:DNA (positive cell efficiencies Both lipids used in these studies are commercially avail- number/total cell (%) able. Lipofectamine (Life Technologies) is a 3:1 (wt:wt) number on well) formulation of the polycationic lipid 2,3-dioleoyloxy-N- (2(sperminecarboxamido)ethyl)-N,N-dimethyl-1-propan- Lipofectamine aminium trifluoroacetate (DOSPA) and the neutral lipid 6:1 2.9 ± 1.1 1.9–4.1 dioleoyl phosphatidylethanolamine (DOPE) in mem- 7.5:1 3.1 ± 1.4 1.9–4.7 ␮ ␮ ± brane filtered water; stock concentration 2 g/ l. 9:1 2.2 1.0 1.6–3.3 DOSPER (Boehringer Mannheim, Mannheim, Germany) DOSPER is a polycationic lipid (1,3-dioleoyloxy-2-(6-carboxy- 1:1 0.6 ± 0.2 0.4–0.7 ± spermyl)-propylamide) without any neutral lipid in 50 2:1 2.9 1.0 2.2–4.0 mm MES buffer, pH 5.6; stock concentration 0.5 ␮g/␮l. 3:1 3.7 ± 0.8 2.9–4.4 All stocks were stored at 4°C. ␮ Lipoplexes were formed using 4 g DNA in varying lipid:DNA Transfection procedure ratios in optimal complex formation volumes, and efficiencies × 5 2 expressed as positive cell number/total cell number on well. Cells (1.5 10 ) were plated in each well (8.8 cm )ofa six-well plate (Nunclon) in DMEM plus 10% FCS. Trans- fections were performed after an overnight incubation. The following describes a typical example of lipoplex for- mid is a 4.4 kb eukaryotic construct containing the GFP mation, although volumes and amounts used were gene under the action of a CMV early promoter. This varied. DNA (usually 4 ␮g) was diluted in sterile HEPES plasmid was constructed by inserting the GFP gene from buffered saline (20 mm Hepes, 150 mm NaCl, pH 7.4) pTRUF230 into pUC18. The GFP gene is gfph2 which is a (HBS) to give a total volume of 50 ␮l. The required humanised version with a further ser-thr(65) mutation, amount of stock lipid for a specific lipid:DNA ratio, was Lipofection of muscle cells E Dodds et al 549

Figure 6 Optimisation of the complex formation step of transfection. C2C12 myoblasts were transfected with lipoplexes formed in varying total volumes of lipid plus DNA (10–120 ␮l), over varying times (15–60 min). Here the effect on transfection efﬁciency is shown for complexes made with (a) 4 ␮g pCMV␤,30␮g Lipofectamine and (b) 2 ␮g pCMV␤,4␮g DOSPER.

also diluted in HBS to give a final volume of 50 ␮l. The and expression analysed after 24 h unless otherwise lipid was then added to the DNA and mixed by gentle stated. pipetting. With Lipofectamine transfections, the medium was removed from the wells and replaced with 2 ml FACS analysis and fluorescence microscopy serum-free DMEM for the duration of the transfection. Analyses were performed 24–48 h after transfection of With DOSPER transfections, DMEM containing 5% FCS cultures. Medium containing any dead or detached cells was used in place of the serum-free medium, unless was removed and retained. Adherent cells were released otherwise stated. Following a 15 min incubation at room from the dish using 0.1% (w/v) trypsin in PBS, pooled temperature, the 100 ␮l of lipoplex formed was added with the medium, spun and resuspended in 1 ml directly into the wells. Negative controls for each experi- medium. These cells were stored on ice to prevent clump- ment consisted of untreated cultures, or cultures treated ing. Immediately before measurement, 1 ␮l of propidium with DNA or lipid only. For FACS analysis of pUCUF2 iodide was added to mark the dead cells fluorescently. transfections, the additional control of cultures trans- In the FACS, cells were initially analysed by size, and a fected with lipid complexed with an irrelevant DNA region representing single cells gated off, thereby elimin- (pCMVCAT) was necessary. After a standard incubation ating clumps and cell debris from further analysis. The time of 6 h, the medium containing the lipoplexes was gated region was then selected and analysed for green removed and replaced with 2 ml DMEM plus 10% FCS, and red fluorescence. Quadrants were set using the irrel- Lipofection of muscle cells E Dodds et al 550 evant DNA plus lipid control, so experimental samples References could be measured and separated into dead cells (red), live positive cells (green) and live negative cells 1 Dickson G, Brown SC. Duchenne muscular dystrophy. In: (nonfluorescent) (Figure 1). Efficiency of transfection was Dickson G (ed). Molecular and Cell Biology of Human Gene Thera- expressed as the percentage of positive cells over total peutics. Chapman and Hall: London, 1995, pp 261–280. cells in the gated region. Where appropriate, cultures 2 Davis H, Michel M, Whalen RG. DNA based immunisation for were visualised by fluorescence microscopy. hepatitis B induces continuous secretion of antigen and high levels of circulating antibody. Hum Mol Genet 1993; 2: 1847–1851. 3 Bohl D, Naffakh N, Heard JM. Long-term control of erythropo- Calculation of lipid:DNA charge ratio ietin secretion by doxycycline in mice transplanted with engin- The charge ratio of the lipoplex was calculated by posi- eered primary myoblasts. Nature Med 1997; 3: 299–305. tive charge equivalents of the cationic compo- 4 Schofield JP, Caskey CT. Non viral approaches to gene therapy. nent/negative charge equivalents of the nucleic acid Br Med Bull 1995; 51: 56–71. component.32 Each nucleotide monomer of DNA carries 5 Zhou XH, Huang L. DNA transfection mediated by cationic liposomes containing lipopolylysine – characterization and one negative charge and has an average molecular mechanism of action. Biochim Biophys Acta 1994; 1189: 195–203. weight of 330. Both DOSPER and DOSPA (cationic lipid 6 Nabel EG et al. Gene transfer in vivo with DNA–liposome com- present in Lipofectamine) have a maximum of four plexes: lack of autoimmunity and gonadal localisation. Hum positive charges per molecule, and molecular weights of Gene Ther 1992; 3: 649–656. 968 and 884, respectively. 7 San H et al. Safety and short-term toxicity of a novel cationic lipid formulation for human gene therapy. Hum Gene Ther 1993; 4: 781–788. Electron microscopy 8 Felgner PL, Ringold GM. Cationic liposome-mediated transfec- Lipoplexes were formed using either 5 ␮g of Lipofectam- tion. Nature 1989; 337: 387–388. ine and 0.66 ␮g DNA, or 5 ␮g DOSPER and 2.5 ␮g DNA, 9 Felgner PL. Lipofection: a highly efficient, lipid-mediated DNA in a total volume of 50 ␮l HBS. This solution was applied transfection procedure. Proc Natl Acad Sci USA 1987; 84: 7413– to formvar coated copper grids and left for 2 min, before 7417. drying with filter paper. The adsorbed samples were then 10 Gao X, Huang L. A novel cationic liposome reagent for efficient negatively stained for 2 min using a 1% ammonium mol- transfection in mammalian cells. Biochim Biophys Res Commun ybdate solution. Micrographs were taken on a TEM Hita- 1991; 179: 280–285. chi H600 (75 kV) (Wokingham, UK). 11 Jarnagin WR. Cationic lipid-mediated transfection of liver cells in primary culture. Nucleic Acids Res 1992; 20: 4205–4211. 12 Legendre JY, Szoka FC. Delivery of plasmid DNA into mam- ␤-Galactosidase histochemistry malian-cell lines using pH-sensitive liposomes – comparison Cells were washed once in 100 mm phosphate buffer (PB, with cationic liposomes. Pharma Res 1992; 9: 1235–1242. pH 7) and then fixed in 0.5% glutaraldehyde (Agar Scien- 13 Plautz GE et al. Immunotherapy of malignancy by in vivo gene transfer into tumours. Proc Natl Acad Sci USA 1993; 90: 4645– tific, Stansted, UK) in PB containing 1 mm MgCl2. Cells 4649. were then given three 10-min washes in Nonidet P-40 sol- 14 Stephan DJ et al. A new cationic liposome DNA complex ution (1 mm MgCl2, 0.02% Nonidet P-40 (Merck, Leices- enhances the efficiency of arterial gene transfer in vivo. Hum ter, UK) in PB) to permeabilise the cells further. The stain- Gene Ther 1996; 7: 1803–1812. ing procedure was performed overnight at 37°C using 1 15 Lui Y et al. Factors influencing the efficiency of cationic lipo- mm MgCl2, 0.15 mm NaCl, 33 mm K4Fe(CN)6,33mm some-mediated intravenous gene delivery. Nature Biotechnol ␤ K3Fe(CN)6, and 0.1% 5-bromochloro-3-indolyl- -d-galac- 1997; 15: 167–173. topyranoside (X-gal; Life Technologies) in PB. Transfec- 16 Nabel GL et al. Direct gene transfer with DNA liposome com- tion efficiency and toxicity were then assessed by coun- plexes in melanoma: expression, biological activity, lack of tox- ting ␤-gal-positive cells and total cell numbers in three icity in humans. Proc Natl Acad Sci USA 1993; 90: 11307–11311. random fields of vision per well, through an eye piece 17 Porteous DJ et al. Evidence for safety and efficacy of DOTAP graticule. Efficiencies were expressed as the percentage cationic liposome-mediated CFTR gene transfer to the nasal epi- thelium of patients with cystic fibrosis. Gene Therapy 1997; 4: of positive cells over the total cell number on an experi- 210–218. mental well. Cell survival was expressed as the percent- 18 Wolff JA et al. Direct gene transfer into mouse muscle in vivo. age of total cell number on the experimental well over Science 1990; 260: 926–932. the total cell number on an untreated control well. 19 Gregoriadis G, Saffie R, Brian de Souza J. Liposome-mediated DNA vaccination. FEBS Lett 1997; 402: 107–110. 20 Ciccarone V, Hawley-Nelson P, Gebeyehu G, Jessee J. Cationic Statistical analysis lipid-mediated transfection of eukaryotic cells: high efficiency All transfection efficiencies and cell survivals are nucleic acid delivery with Lipofectin and Lipofectamine presented as mean percentage ± standard deviation. reagents. FASEB J 1993; 7: 454. Where appropriate, Student’s t tests were performed, to 21 Trivedi RA, Dickson G. Liposome-mediated gene transfer into give significance values. normal and dystrophin-deficient mouse myoblasts. J Neurochem 1995; 64: 2230–2238. 22 Vitiello L et al. Condensation of plasmid DNA with polylysine improves liposome-mediated gene transfer into established and Acknowledgements primary muscle cells. Gene Therapy 1996; 3: 396–404. 23 Chalfie M et al. GFP as a marker for gene expression. Science This work was supported by the Muscular Dystrophy 1996; 263: 802–805. Group of Great Britain and Northern Ireland and the 24 Farhood H, Bottega R, Epand RM, Huang L. Effect of cationic Medical Research Council of Great Britain. We thank cholesterol derivatives on gene transfer and protein kinase C Andreas Machl for providing pUCUF2. activity. Biochim Biophys Acta 1992; 1111: 239–246. Lipofection of muscle cells E Dodds et al 551 25 Staggs DR, Burton DW, Deftos LJ. Importance of liposome com- 29 Yasmin R et al. A quantitative technique for growing human plexing volume in transfection optimisation. Biotechniques 1997; adult skeletal muscle in culture starting fron mononucleated 21: 792–798. cells. J Neurol Sci 1977; 32: 347–360. 26 Partridge T, Davies KE. Myoblast based gene therapies. Br Med 30 Zolotukhin S et al. A ‘humanised’ green fluorescent protein Bull 1995; 51: 123–137. cDNA adapted for high level expression in mammalian cells. 27 Fasbender A et al. Complexes of adenovirus with polycationic JVirol1996; 70: 4646–4654. polymers and cationic lipids increase the efficiency of gene 31 Heim R, Cubitt AB, Tsien RY. Improved green fluorescence. transfer in vitro and in vivo. J Biol Chem 1997; 272: 6479–6489. Nature 1995; 373: 663–664. 28 Yaffe D, Saxel O. Serial passaging and differentiation of myo- 32 Felgner PL et al. Nomenclature for synthetic gene delivery sys- genic cells isolated from dystrophic mouse muscle. Nature 1977; tems. Hum Gene Ther 1997; 8: 511–512. 270: 725–727.