Detection of integration of plasmid DNA into host genomic DNA following intramuscular injection and electroporation AbstractPlasmid vectors have been widely used for DNA vaccines and gene therapy. Following intramuscular injection, the plasmid that persists is extrachromosomal and integration into host DNA, if it occurs at all, is negligible. However, new technologies for improving DNA delivery could increase the frequency of integration. In the present study, we tested the effect of electroporation on plasmid uptake and potential integration following intramuscular injection in mice, using a plasmid containing the mouse erythropoietin gene. Electroporation increased plasmid tissue levels by approximately six- to 34-fold. Using a quantitative gel-purification assay for integration, electroporation was found to markedly increase the level of plasmid associated with high-molecular-weight genomic DNA. To confirm integration and identify the insertion sites, we developed a new assay 鈥?referred to as repeat-anchored integration capture (RAIC) PCR 鈥?that is capable of detecting rare integration events in a complex mixture in vivo. Using this assay, we identified four independent integration events. Sequencing of the insertion sites suggested a random integration process, but with short segments of homology between the vector breakpoint and the insertion site in three of the four cases. This is the first definitive demonstration of integration of plasmid DNA into genomic DNA following injection in vivo. IntroductionPlasmid expression vectors have represented an attractive approach for gene therapy and genetic vaccination because of their ease of preparation, stability, biochemical simplicity, and relative safety compared to viral vectors. 1,2 However, the cellular uptake of plasmids after administration in vivo is highly inefficient.3 One promising method for improving the effectiveness of plasmid DNA vectors is electroporation, which can increase the expression of the plasmid-encoded transgene by several orders of magnitude.4,5,6,7A major safety issue for plasmid DNA vectors is the potential for integration of the vector DNA into host genomic DNA 鈥?although, in the case of gene therapy, integration may be desirable for stable gene expression, particularly when targeting proliferating cells. To assess integration, we previously developed a quantitative, gel-purification assay for integration in vivo,8,9,10 in which high-molecular-weight (HMW) genomic DNA is purified away from extrachromosomal (free) plasmid vector by multiple rounds of agarose-gel electrophoresis and then assayed for potentially integrated vector using vector-specific PCR. Using this assay, we demonstrated that virtually all of the plasmid DNA that persisted after intramuscular administration was extrachromosomal, and, if it occurred at all, the frequency of integration was at least three orders of magnitude below the spontaneous rate of gene-inactivating mutations (using a worst case scenario).8,9,10,11 Similar results have been reported by others.12 However, new technologies such as electroporation that improve the efficiency of DNA delivery, and potentially alter qualitative aspects as well, could theoretically increase the frequency of integration. Depending on various benefit-risk considerations, enhanced integration could either be desirable (eg, for gene therapy targeted to a proliferating cell population that would lose episomal plasmids) or undesirable (eg, for prophylactic vaccines).Previously, in an intramuscular gene therapy study in Balb/c mice using a plasmid (pCMV/mEPO) that contained the mouse erythropoietin (EPO) gene, Rizzuto et al13 demonstrated that electroporation increased EPO transgene expression by at least 100-fold and greatly reduced the amount of plasmid necessary to affect an increase in hematocrit. The present report describes a related study in which we examined the effect of electroporation on the uptake of pCMV/mEPO plasmid, and on the frequency of integration of the plasmid into host genomic DNA. The gel-purification assay was used to assess the frequency of integration. However, since a positive signal in the gel-purification assay could be due to incomplete removal of free plasmid from HMW genomic DNA, confirmation of integration would require demonstration of the covalent junction of plasmid-to-genomic DNA sequences. To accomplish this, we developed a new integration assay, referred to as repeat-anchored integration capture (RAIC PCR), that is capable of detecting rare, single-copy, integration events in a complex mixture in vivo.ResultsElectroporation results in increased cellular uptake of plasmid DNATwo electroporation studies were performed. In the first study, Balb/c mice were injected with pCMV/mEPO plasmid, with and without subsequent electroporation, and the level of vector DNA in the injected muscle at 1, 6, and 16 weeks after dosing was examined using real-time fluorescence quantitative PCR (TaqManR庐 PCR). For mice treated with electroporation, the amount of plasmid DNA administered was 1, 3, or 50鈥壩糶. For nonelectroporated mice, 50鈥壩糶 of plasmid DNA was administered. Control mice received vehicle only. In the second study, mice were treated with 50鈥壩糶 of plasmid DNA, with and without electroporation, and examined at 6 weeks post-dose.With a dose of 50鈥壩糶 of plasmid, electroporation caused an increase of approximately six- to 34-fold in the level of plasmid DNA in muscle (Table 1). Also, at the later time points, a dose of 3鈥壩糶 of plasmid with electroporation generally resulted in higher plasmid levels in muscle than did the dose of 50鈥壩糶 of plasmid without electroporation.Table 1 Effect of electroporation on plasmid uptakeFull size tableGel-purification assay for integrated plasmid: increased association of plasmid DNA with HMW genomic DNAUsing a quantitative, gel-purification assay for integration, electroporation was found to markedly increase the level of the plasmid that was associated with HMW genomic DNA through repeated rounds of gel purification, suggestive of integration of the vector into host genomic DNA. Reproducible results were obtained in both studies.Representative results are shown in Table 2, comparing a 鈥?0鈥壩糶 (no electroporation)鈥?and 鈥?0鈥壩糶 + electroporation鈥?sample from the 6-week time point in the second study. For optimal separation of free plasmid from genomic DNA, rather than using a conventional preparation of DNA as starting material, tissue DNA was isolated using a 鈥楩ilter DNA prep鈥? which results in genomic DNA of very high molecular weight ( 200鈥塊b), and which removes some of the free plasmid DNA during the isolation procedure. The filter prep DNA was then purified by four successive rounds of preparative pulsed-field gel electrophoresis, using instrument parameters optimized to resolve 20鈥?00鈥塳b fragments to provide maximum separation of HMW genomic DNA from free plasmid vector. The band of HMW genomic DNA was electroeluted from the gel and assayed by TaqManR PCR for potentially integrated plasmid.Table 2 Gel purification assay for integrated plasmidFull size tableFor the muscle DNA samples from nonelectroporated mice, virtually all of the plasmid was removed by gel purification. In the example shown in Table 2, only 17 copies/渭g DNA were present following gel purification. These results are consistent with those previously published, suggesting that the frequency of integration following normal intramuscular injection is negligible.8,9,10,11,12In contrast, after identical filter DNA isolation and gel purification, the electroporated sample had approximately 980 plasmid copies per 渭g DNA (Table 2). Two additional rounds of gel purification had little impact, even utilizing digestion with SfiI restriction enzyme 鈥?which can digest plasmid concatemers and facilitate their removal via gel purification, but which also causes a selective loss in detecting integrated plasmid.10This increase in the level of plasmid associated with HMW genomic DNA caused by electroporation was observed consistently in repeated attempts at gel purification in both studies (not shown). Indeed, a 鈥?0鈥壩糶+EP鈥?quadriceps sample at 16-weeks post-dose (see Table 1) had approximately 250 copies/渭g of DNA remaining after gel purification (not shown). Note, in other plasmid studies using simple intramuscular injection, gel purification successfully reduced the plasmid level to 16 copies/渭g DNA even for samples that had 106 copies/渭g DNA (not shown). Together, these results suggest that electroporation caused an increase in plasmid integration frequency. However, confirmation of integration required development of an assay to unambiguously identify the plasmid-to-genomic DNA junctions, as described below.Description of the RAIC PCR assay for confirming plasmid integrationTo confirm plasmid integration and identify the insertion sites, we developed a new PCR assay capable of detecting rare, single-copy, integration events in a complex mixture in vivo. The assay, referred to as RAIC PCR, employs features from several different methods to provide the necessary sensitivity and specificity. First, the RAIC PCR assay uses 鈥榬epeat-anchored PCR,鈥?in which a repetitive DNA element is used for the 鈥榰nknown鈥?genomic DNA primer. For the present study in mice, the mouse B1 short interspersed nucleotide element (B1 SINE) was used, although GT-dinucleotide repeat primers have also been successfully used in the RAIC PCR assay to detect integrations (unpublished results). A similar approach, Alu-PCR, has been used in human cells.14 Second, the assay employs a biotinylated-vector primer, for 鈥榗apture PCR鈥? to allow for enrichment by streptavidin-bead precipitation of PCR products containing vector sequences away from a preponderance of repeat-to-repeat PCR products.15 Third, the assay employs the dUTP-labeled primer/UDG-digestion technique, described below, to reduce reamplification of repeat-to-repeat PCR products.14,16 Finally, the assay uses 鈥榣ong PCR鈥?conditions to permit amplification between an integrated vector and a B1 element that could be, on average, 5- to 10鈥塳b away. PCR products are sequenced to identify plasmid-to-genomic DNA junctions. Importantly, the RAIC PCR method has no ligation steps that could result in false positives (artefactual joining of plasmid-to-genomic DNA).The primary difficulty in using PCR to amplify vector-to-genomic DNA junctions of unknown integration sites is that only the vector sequence is known. Our strategy uses a vector-specific primer and a genomic primer based on repetitive DNA 鈥?the mouse B1-SINE sequence.17 Since there are approximately 564鈥?00 copies of the B1 element interspersed in the mouse 2.5-gigabase haploid genome,18 B1 elements are approximately 4.4鈥塳b apart on average. Therefore, a random integration event would fall within approximately 2.2鈥塳b on average of one of the two surrounding B1 elements. Therefore, it is possible to amplify a significant subset of integration junctions using the vector-specific primer and a mouse B1 primer together with long PCR conditions.Owing to their frequency in the mouse genome, the first round of RAIC PCR generates predominately B1-to-B1 products. Two techniques were employed to overcome this tremendous excess of undesired products and to enrich for B1-to-vector products. First, the B1 primers were designed with a 5鈥?Tag sequence, and with dUTPs in substitution for dTTPs to permit destruction of the B1-Tag primer sequence in PCR products by uracil DNA glycosylase (UDG) after the first round of PCR14,16鈥攖his approach has also been used with Alu-PCR.14 Second, the vector-specific primer was labeled with biotin, to permit enrichment for PCR products containing the vector using streptavidin-coated magnetic beads.15 In subsequent rounds of PCR, amplification was performed with a nested, biotinylated vector-specific primer and the Tag sequence within B1 primer. Further amplification of PCR products between two B1 elements was blocked since, after UDG treatment, the B1-Tag primer segments in the PCR products from the first round PCR products were destroyed. The overall strategy for the RAIC PCR assay is illustrated in Figure 1.Figure 1Strategy of the RAIC PCR assay. The left panel represents the vector-to-B1 amplification, and the right panel shows the B1-to-B1 amplification. The open box and line represent the integrated vector sequence and unknown mouse genomic sequence, respectively. The shadow arrow boxes represent B1 elements on the mouse genome in different orientations. The arrow with a bold bar shows the B1-Tag primer in which dTTPs were substituted with dUTPs, and where the bold bar indicates the Tag sequence for the successive rounds of PCR. The closed circles represent biotin coupled to the 5鈥?end of the vector primer. The short dotted lines indicate the B1-Tag primer where dUTP was digested by UDG. The long dotted line represents the initial product from Tag primer extension. P1 is the vector-specific primer in first round PCR, and P2 is the nested vector-specific primer in second round PCR. The assay consists of the following procedures. The first round of PCR is carried out between a biotinylated specific primer and the B1-Tag primer in which dTTPs are replaced with dUTPs. PCR products are treated with UDG. Then, the biotinylated DNA fragments that originate from the vector primer are isolated, and the biotinylated strands are purified using streptavidin-coated magnetic beads. The second round of PCR is performed using the purified biotinylated strands as templates and the nested biotinylated vector primer together with a Tag primer. Usually, a third round of PCR using the purified biotinylated strands from the second round PCR as template is needed. Finally, the PCR products are directly sequenced and analyzed.Full size imageDevelopment of the RAIC PCR assayTo validate the RAIC PCR assay for detecting insertion of foreign DNA at unknown sites, and to determine the sensitivity of the method, a transfected mouse cell line that contains a single integrated copy of the V1Jp-Hygro plasmid10 was used. The location of the integration in the genomic DNA was not known. A 5-kb PCR product was obtained in the second round of RAIC PCR (Figure 2a,鈥塨). Covalent linkage of the vector-to-mouse DNA was confirmed by DNA sequencing. Furthermore, using the newly discovered vector-flanking sequence and B1-flanking sequence as primers for PCR (see Figure 2a), it was confirmed that the junction sequence was indeed mouse genomic DNA (Figure 2c).Figure 2Validation of the RAIC PCR assay using DNA from a transfected mouse cell line that contains a single integrated copy of the V1Jp-Hygro plasmid. (a) Structure of the junction fragment. The 5鈥?end includes the vector sequence and the 3鈥?end is a mouse sequence that contains a B1 element. The shadow arrow box indicates the mouse B1 element. The bold line represents confirmed mouse DNA, and the dotted line indicates DNA that was not sequenced. The arrows represent the primer set for PCR of the panel (c). (b) The 5-kb junction fragment detected in the second round of PCR. Lane M, molecular weight standard; lane 1, PCR products from 100鈥塶g of the transfected cell line genomic DNA. (c) PCR confirmation of the mouse genomic DNA sequence within the 5-kb junction fragment. PCR was performed using the primer set shown in panel (a) with various DNA templates. Lane M, molecular weight standard; lane 2, water as template (negative control); lane 3, junction fragment as template (positive control); lane 4, human genomic DNA as template (negative control); and lane 5, mouse DNA as template.Full size imageUsing serial dilutions of the V1Jp-Hygro cell line DNA into control mouse DNA, the sensitivity of the method was approximately 1.5 copies/渭g of DNA (or per 150鈥?00 diploid cells) when using three rounds of PCR amplification. While that sensitivity is specific to a particular integration event in a clonal population, it does suggest that the method is capable of detecting single-copy integration events.Confirmation of plasmid integration in electroporated quadriceps DNAAfter removal of extrachromosomal plasmid by preparative gel electrophoresis, quadriceps DNA samples treated with 50鈥壩糶 plasmid plus electroporation (from 6 weeks post-dose in the second study) were assayed by RAIC PCR. The vector primers that were used are illustrated in Figure 3. Four independent integration events were identified. For integration 5427 (an arbitrary name based on plasmid breakpoint), a 6-kb product was detected after three rounds of PCR (Figure 4a,鈥塨). DNA sequencing confirmed that the PCR product contained vector sequences (nucleotides 3213鈥?427) covalently linked to mouse genomic DNA (containing a B1 element) 鈥?localized to an intergenic region on mouse chromosome 16 by comparison to the mouse genome database. Furthermore, using the newly discovered vector-flanking sequence and B1-flanking sequence as primers for PCR (see Figure 4a), it was confirmed that the junction sequence was indeed mouse genomic DNA (Figure 4c).Figure 3Primer design for pV1JnsBmEpo (pCMV/mEPO) vector. The coding sequence for erythropoietin spans nucleotide 59鈥?37. Three sets of primers were designed. Each primer set includes four primers to allow for two rounds of nested PCR and direct sequencing of the third round PCR products. The closed circles represent biotin coupled to the 5鈥?end of the first and second round primers.Full size imageFigure 4Identification of the integration 5427 in the EPO-treated mouse quadriceps DNA by RAIC PCR. (a) Structure of the junction fragment. The 5鈥?end includes vector sequence from 3213 to 5427, and the 3鈥?end is a nonvector sequence that contains a mouse B1 element. The shadow arrow box indicates the mouse B1 element. The bold line represents confirmed mouse sequences, and the dotted line indicates DNA that was not sequenced. The arrows represent the primers for the PCR that is shown in panel (c). (b) A 6-kb fragment is identified in the third round of RAIC PCR. Lane M, molecular weight standard; lane 1, 4000 copies of pV1JnsBmEpo plasmid spiked into mouse DNA were used as a nonintegrated spiked control; lane 2, 0.4鈥壩糶 of the EPO-treated mouse quadriceps DNA as template. (c) PCR confirmation of the mouse sequence within the junction fragment. PCR was performed using the mouse genomic primer set shown in panel (a) (Table 4) and various DNA templates. Lane M, molecular weight standard; lane 1, water as template (negative control); lane 2, junction fragment as template (positive control); lane 3, human genomic DNA as template (negative control); lanes 4 and 5, mouse DNA samples as template.Full size imageTo improve PCR efficiency, a double-tagged B1 primer was designed to allow for two rounds of seminested PCR (see Materials and methods). Using this primer, three additional integration events were detected. For integration 1005, a 2-kb fragment was obtained after three rounds of PCR (see Figures 5a,鈥塩, lane 2), and DNA sequencing confirmed that the PCR product contained rearranged vector sequences (nucleotides 3053鈥?380 and 751鈥?005) covalently linked to mouse DNA (containing a B1 element) 鈥?localized to an intergenic region on mouse chromosome 6. For integration 2217, a 1.3-kb fragment was obtained after three rounds of PCR (see Figures 5b,鈥塩, lane 3), and DNA sequencing confirmed that the PCR product contained vector sequences (nucleotides 3053鈥?217) covalently linked to mouse DNA (containing a B1 element) 鈥?localized to an intron in the Protein Phosphatase 1 Regulatory Subunit 9A gene on mouse chromosome 6. For integration 370, a 1.5-kb fragment was obtained after three rounds of PCR (see Figure 6a,鈥塨), and DNA sequencing confirmed that the PCR product contained vector sequences (nucleotides 1586鈥?70) covalently linked to mouse DNA (containing a B1 element) 鈥?localized to an intron in the Hook Homolog 2 (Hook-2) gene on mouse chromosome 8.Figure 5Identification of integration 1005 and integration 2217 in the EPO-treated mouse quadriceps DNA by RAIC PCR. (a) Structure of the junction fragment identified in integration 1005. The 5鈥?end includes rearranged vector sequence (from nucleotides 3053 to 2380 and 751 to 1005), and the 3鈥?end is a mouse sequence that contains a B1 element. The shadow arrow box indicates the mouse B1 element. The bold line represents confirmed mouse sequence. (b) Structure of the junction fragment identified in integration 2217. The 5鈥?end includes vector sequence from nucleotides 3053 to 2217, and the 3鈥?end is a mouse sequence that contains a B1 element. The shadow arrow box indicates the mouse B1 element. The bold line represents confirmed mouse sequence. (c) 2- and 1.3-kb junction fragments are identified in the third round of RAIC PCR. Lane M, molecular weight standard; lane 1; 4000 copies spiked in 0.4鈥壩糶 of mouse DNA as template; and lanes 2 and 3, 0.4鈥壩糶 of the EPO-treated mouse quadriceps DNA as template.Full size imageFigure 6Identification of integration 370 in the EPO-treated mouse quadriceps DNA by RAIC PCR. (a) Structure of the junction fragment identified in the integration 370. The 5鈥?end includes vector sequence from nucleotides 1586 to 370, and the 3鈥?end is a mouse sequence that contains a B1 element. The shadow arrow box indicates the mouse B1 element. The bold line represents the mouse sequence. (b) A 1.5-kb junction fragment is identified in the third round of RAIC PCR. Lane M, molecular weight standard and lane 1; 0.4鈥壩糶 of the EPO-treated mouse quadriceps DNA.Full size imageAs a control for false positives via intermolecular PCR jumping 鈥?a remote but theoretical possibility 鈥?the RAIC PCR assay was routinely applied to control mouse DNA samples spiked with free plasmid DNA. After exhaustive attempts with DNA sequencing of all products, only relatively short, nonspecific PCR products were obtained (eg, see Figure 4b, lane 1) 鈥?none of which were found to contain plasmid linked to mouse DNA.Integration events occurred at random sites in genomic DNAThe insertion sites for integration events can either be targeted or random, from a chromosomal locus perspective. Sequence analyses revealed that the four integration events detected in this study occurred at different insertion sites. None of the integrations involved homologous recombination into the endogenous mouse EPO locus. To investigate whether the insertion sites represented hot spots for 鈥榯argeted鈥?integration, PCR assays using a specific primer for the vector and a specific primer for the genomic DNA insertion site were developed, based on integrations 5427 and 1005. When these PCR assays were applied to the original quadriceps DNA sample, the results were negative to a sensitivity of detection of one copy/渭g DNA (demonstrated by serial dilution of the cloned junction fragment). Thus, there were no other detectable DNA molecules in the original sample in which the vector was in proximity to the genomic sequence of either insertion site, indicating that the detected integration events were random and not targeted to a particular hot spot of the genome.The integration events tended to occur in genomic regions that are slightly rich in AT base pairs. The frequency of AT base pairs at the genomic insertion sites (surrounding 300鈥塨p) of integration 5427, integration 1005, integration 2217, and integration 370 were 61, 64.7, 65.7, and 54.3, respectively. The average AT content in the mouse genome is 58%.18 Also, as shown in Table 3, sequence analyses revealed a 4-, 6-, or 12-bp homology between the vector and genomic DNA sequences at the breakpoints in three of the four integrations.Table 3 DNA sequences of the junction sitesFull size tableDiscussionPlasmid DNA vectors have advantages over viral vectors in tolerability, in the relative absence of pre-existing or induced antivector neutralizing antibodies, and in the simplicity of manufacture and characterization. However, the cellular uptake of injected plasmids is very inefficient. Therefore, simple injection procedures have often been supplemented or replaced by other approaches. Examples of these approaches (reviewed in Herweijer and Wolff1 and Nishikawa and Huang3) include the use of electroporation, formulation of DNA within cationic lipids or other particles, delivery Biojector needleless jet injection, delivery of DNA bound to gold particles using a gene gun, and use of various routes of administration 鈥?for example, intra-arterial (femoral) delivery of naked plasmid DNA with application of an external tourniquet has been shown to lead to high levels of transgene expression in the hindlimb muscle.19 The present paper describes the consequences of using one of these approaches, electroporation, on the state of plasmid DNA.Using simple intramuscular injection, the vast majority of plasmid DNA that persists is extrachromosomal, and the frequency of integration, if it occurs at all, is negligible. Generally, the level of plasmid remaining associated with HMW genomic DNA after gel purification of the injected muscle DNA is between 1 and 30 copies/渭g DNA (or per 150鈥?00 diploid cells).8,9,10,11,12 Similar results were obtained in the nonelectroporated sample in the present study. The residual plasmid that remains with HMW genomic DNA following gel purification could represent integrated plasmid, but it is also possible that the residual plasmid is extrachromosomal. Purification procedures almost never yield 100% pure material, and one copy ( 1 脳 10鈭?7鈥塯) of plasmid/渭g genomic DNA represents an impurity level of 100 billionth by weight.10 However, even if the residual plasmid in the gel-purified genomic DNA did represent integrated plasmid, one copy of integrated plasmid/渭g of genomic DNA (representing 150鈥?00 diploid cells) would be at least three orders of magnitude below the frequency of spontaneous gene-inactivating mutations, using a worst-case analysis (see Ledwith et al10 for a discussion of this calculation).New technologies for improving the efficiency of DNA delivery could conceivably increase the frequency of integration. Factors that could theoretically affect integration include the plasmid sequence (by virtue of recombination signals), mode and route of administration, use of adjuvants, and the nature of the transgene protein product (see Manam et al11 for a thorough discussion of these factors). However, studies exploring such variables showed no evidence of plasmid integration.11 These and other studies have supported a view that plasmid integration in vivo is essentially a theoretical concern.In the present study, we demonstrated that electroporation increased the level of DNA uptake, generally between six- to 34-fold. This correlated well with the increased expression of the EPO transgene reported earlier.13 While the precise mechanism is unclear, electroporation is thought to promote cellular uptake of DNA through permeabilizing cell membranes and driving DNA entry via an electrophoretic process.4,5 Using the gel-purification assay for integration, we observed that electroporation markedly increased the level of the plasmid that was associated with HMW genomic DNA. To confirm integration and identify the insertion sites, we developed a new assay, referred to as RAIC PCR, that is capable of detecting rare, single-copy, integration events in a complex mixture in vivo. Using this assay, we identified four independent integration events in DNA isolated from electroporated muscle. This is the first definitive demonstration of integration of plasmid DNA into genomic DNA following injection in vivo.A number of methods have been successfully used to isolate and characterize the genomic insertion sites of integrated foreign DNA. However, in most cases, the foreign DNA insertion had been present in a large fraction of the cells, or present in a clonal or partially clonal cell population such that there were many cells with the same integration event/structure. The challenge with the present study of plasmid integration in muscle cells are (1) the random nature of the integration event, (2) the low frequency of integration, and (3) the low degree of cell replication and the nonclonal nature of the muscle cell population affected 鈥?suggesting that any particular integration event (junction fragment) would be present at or near single-copy levels. For detecting these rare, unique integration events, the RAIC assay offers several advantages over other methods. The RAIC assay involves direct PCR amplification (utilizing the repeat-anchor primer) to provide sensitivity, includes a simple biotin-capture step to enrich for vector-containing PCR products and enhance sensitivity, and contains no cloning or ligation steps that could produce false positives through artefactual joining of genomic DNA to vector DNA (this is particularly important for characterizing unique integration events where the junctions cannot be confirmed). In contrast, inverse PCR, one widely used method, involves restriction enzyme digestion and circularization (intramolecular ligation) of the cellular DNA prior to PCR, followed by PCR in which opposite-facing vector primers are used to amplify the circularized DNA. As discussed in a recent review, a relatively high copy number of integrations is necessary to compensate for the inefficiency of this method.20 For example, in one successful example of inverse PCR, highly efficient retroviral integration was examined and cells containing integrations were selected using thymidine kinase/HAT selection.21 Owing to the direct-PCR approach, lack of ligation steps and high sensitivity, repeat-anchored (eg, Alu) PCR is generally preferred over inverse PCR.20 The RAIC method described in the present report simply augments the sensitivity of repeat-anchored PCR by also employing the advantages of biotin-capture PCR.15,20Sequencing of the genomic insertion sites suggested a random integration process, but with short segments (4鈥?2 nucleotides) of homology between the vector breakpoint and the insertional site of cellular DNA in three of the four integration events, consistent with microhomologies observed in other models of illegitimate (or nonhomologous) integration.22,23 The integration events tended to occur in regions slightly rich in AT base pairs, consistent with a model for recombination that involves AT-rich, nuclear scaffold/matrix-attached regions that are thought to provide a propensity for local strand-unpairing and torsional strain.24 Finally, two of the four integrations occurred within an intron of a gene, consistent with observed propensities of foreign DNA to integrate near transcriptionally active DNA.25,26 Of course, the insertion sites detected by the RAIC assay were biased by the choice of B1-SINE as the anchor. However, B1-SINE elements in mice, like the related Alu elements in humans, are retroposons and hence located in regions that are favorable to recombination,18,24 which in addition to their repetitive nature make these elements attractive for use in the RAIC assay for detecting integration of foreign DNA.The low frequency and random nature of the integration events, and the nonclonal nature of the cellular material, limit the ability to characterize each integration event. For example, only one of the two vector-to-genomic DNA junctions for a given integration event is revealed in this assay, and the structure of the overall plasmid and insertion site are not determinable. Furthermore, the effect of integration on transgene expression cannot be determined for such rare integration events. However, it is clear from the gel purification results that 99.5% of the vector in the electroporated sample remained extrachromosomal (ie, 1000/220鈥?00 copies at the 6 week time-point were integrated). Thus, the enhanced EPO expression observed with electroporation13 is unlikely to be due to the integration, but rather to the increased level of cellular uptake of the plasmid DNA.While observing plasmid integration in vivo is unusual, the level of integration is relatively low. Using a worst-case calculation,10 an integration frequency of 1000 copies/渭g DNA is still below the spontaneous rate of gene-inactivating mutations. Depending on other risk-benefit considerations, the enhanced gene expression afforded by electroporation may outweigh the increased risk of mutation. This is particularly true for certain gene therapy applications where integration may be desirable. Furthermore, just as electroporation has resulted in a qualitative difference in integration risk, it is possible that different electroporation procedures could exhibit significantly different effects on integration frequency. Modulation of electroporation conditions4,5,6 and electrode designs6,27,28 could affect DNA uptake and expression, and thus could potentially affect the state of the DNA as well.Materials and methodsPlasmidpCMV/mEPO, also known as pV1JnsBmEPO, is a 5527-bp plasmid that contains the coding sequence of the mouse EPO gene under the control of the human cytomegalovirus (CMV) immediate/early region promoter and enhancer with intron A, followed by a polyadenylation signal from the bovine growth hormone gene.13Mouse treatment and electroporation conditionsBalb/c mice were injected with pCMV/mEPO, with and without subsequent electroporation of the injected muscle. pCMV/mEPO was given in the left quadriceps at a dose volume of 50鈥壩糽 on day 1. The vehicle control groups were dosed similarly with 50鈥壩糽 of 0.9% saline. Animals receiving electroporation were anesthetized with an intraperitoneal injection of a solution containing 100鈥塵g/kg of ketamine, 8鈥塵g/kg of xylazine, and 2.5鈥塵g/kg of acepromazine. The left quadriceps muscle was surgically exposed, intramuscularly dosed, and electroporated. Electrodes were introduced to each side of the exposed quadriceps muscle. Stainless-steel electrodes in the form of parallel 0.2-mm wires about 1.2-cm long and approximately 5鈥塵m apart were brought into contact with the muscle in parallel orientation with respect to the muscle fibers. The electric field was applied in a pulsed form through a Pulsar 6bp-a/s bipolar stimulator (FHC, Bowdoinham, ME, USA), and each cycle of stimulation comprised a 1-s pulse train of square bipolar pulses delivered every other second (2-s cycle time). Each train consisted of 103 pulses of 200鈥壩約 length and approximately 40鈥撯€?5鈥塚 amplitude (when the contact was closed around the muscle). Pulses were monitored by using a digital oscilloscope (Tektronix Model TDS210). The muscle was stimulated for 10 trains (20鈥塻) and then the electrodes were removed. The incision was sutured and each anesthetized mouse was placed on a heating pack to enhance recovery.Necropsy was limited to removal of the left quadriceps (injection site), which was frozen in liquid nitrogen and stored at 鈭?0掳C.Total DNA Isolation from quadricepsDNA was isolated from frozen quadriceps by lysis in sodium dodecyl sulfate (SDS), digestion with Proteinase K and RNase, extraction with phenol:chloroform:isoamyl alcohol and chloroform:isoamyl alcohol, and precipitation with isopropanol. DNA samples were quantitated by spectrophotometry.Filter DNA prepFor the integration assays, DNA was also isolated from tissue using a filter DNA prep. This procedure keeps the genomic DNA at a larger molecular weight such that higher resolution electrophoresis parameters can be used for gel purification. The filter prep also removes some of the free vector during the DNA isolation procedure itself. Homogenized or minced tissue was placed on a 2-渭m filter and gently lysed in a solution containing SDS and proteinase K. Following digestion, the solution was pulled through the filter using a peristaltic pump. HMW DNA remained on top of the filter and was rinsed with several changes of buffer.Gel purificationHMW genomic DNA was separated from free vector using pulsed-field gel electrophoresis (PFGE). All PFGE procedures used the BioRad CHEF庐 (Clamped Homogeneous Electric Fields) System. The electrophoresis parameters were determined by the instrument using an autoalgorithm for resolution of 20鈥?00鈥塳b DNA fragments. Usually, between 2 and 5鈥壩糶 of DNA was loaded per well. Different samples were always run in different gels. A negative control DNA sample was usually run in each experiment to control for contamination. For multiple rounds of purification, the excised band of HMW genomic DNA from one round was directly subjected to the next round without elution from the gel. After gel purification, DNA was electroeluted from gel slices and then concentrated by centrifugation in Centricon-30 filtration units (Millipore, Inc.). In some cases, gel-purified DNA was digested with SfiI restriction enzyme prior to further gel purification.10 SfiI was used to digest potential concatemers of the vector. After SfiI digestion, conventional agarose gel purification procedures, referred to as TAE and TBE gels, were used as previously described.10TaqManR PCRSamples (0.5鈥壩糶 of DNA per reaction) were assayed in triplicate by 鈥楾aqManR鈥?real-time quantitative fluorescence PCR using the ABI Prism庐 7700 Sequence Detection System (from Applied Biosystems) for the following segments: CMV segment The CMV segment is targeted to the CMV promoter and the BGH segment is targeted to the bovine growth hormone terminator, both located within the EPO plasmid.Prior to PCR, DNA samples were heat denatured at approximately 95掳C for approximately 15鈥塵in. TaqManR PCR reactions were carried out in a final volume of 50鈥壩糽 containing 1 脳 TaqManR Universal Master Mix (Applied Biosystems), 100鈥塶M probe, 180鈥塶M of each primer, and 0.5鈥壩糶 of sample DNA. In each TaqManR PCR experiment, a positive control titration curve consisting of 0, 10, 102, 103, 104, 105, and 106 copies of EPO plasmid/渭g of control DNA was assayed along with the test samples. The level of vector DNA in a sample was determined by comparison of the sample\'s Ct value with those of the titration curve.RAIC PCRPrimersPrimers used for the RAIC PCR assay are listed in Table 4 and were purchased from Operon Technologies. The mouse B1 primer was based on the mouse B1 sequence (Genebank accession number, J00631). For the first round of PCR, the mouse B1 primer contained a 5鈥?Tag sequence and dUTPs were substituted for dTTPs. The primers for the V1Jp-Hygro clone DNA were based on the sequence of the V1Jp-Hygro plasmid near the plasmid breakpoint. The primers for the EPO sample were based on sequence of pV1JnsBmEpo (pCMV/mEPO) plasmid. Since the plasmid breakpoints were unknown, three sets of primers were designed to amplify different portions of the vector (see Figure 3). Each primer set included four primers that allowed up to three rounds of nested PCR followed by sequencing. The 5鈥?ends of the first two primers were labeled with biotin.Table 4 Primers for the RAIC PCR assayFull size tablePCRReactions for the first round of PCR contained 20鈥壩糽 of reaction mix-1 and 30鈥壩糽 of reaction mix-2. Mix-1 contained 0.1鈥?鈥壩糶 of target genomic DNA, 4鈥壩糽 of 10鈥塵M blend dNTPs, 250鈥塶g of the vector-specific primer, and 250鈥塶g of B1 primer. Mix-2 contained 3鈥壩糽 of buffer A, 7鈥壩糽 of buffer B from the ELONGase amplification system (Invitrogen Life Technologies), and 5鈥塙 of TaqPlus Long polymerase mixture (Stratagene). PCR was initiated at 70掳C by adding the mix-2 to mix-1, followed by preamplification denaturation for 1鈥塵in at 94掳C, 10 cycles of denaturation for 15鈥塻 at 94掳C, annealing for 30鈥塻 at 65掳C, extension for 18鈥塵in at 68掳C, and a final extension for 12鈥塵in at 72掳C. For the second and third rounds of PCR, the nested vector-specific primer and Tag primer were used. PCR conditions were the same as the first round PCR except 40 cycles of amplification were used and 1.5鈥塙 of ELONGase was used instead of the TaqPlus enzyme. All PCR was performed in the GeneAmp PCR System 9700 (Applied Biosystems).UDG treatmentFirst round PCR products were initially purified using a Microcon-PCR centrifugal filter device (Millipore). To a final collection volume of 26鈥壩糽, 3鈥壩糽 of 10 脳 UDG buffer and 1鈥塙 of UDG (New England Bioloabs) were added, followed by incubation for 30鈥塵in at 37掳C.Isolation of the biotinylated DNA strandsBiotinylated PCR products were isolated with streptavidin-coated magnetic Dynabeads M270 (Dynal Biotech) according to instructions from the manufacturer with some modifications. For each UDG-digested PCR reaction, 30鈥壩糶 of Dynabeads were washed and suspended in 30鈥壩糽 of binding solution. Then, 30鈥壩糽 of the prewashed Dynabeads were added to 30鈥壩糽 of the UDG-treated PCR products, and the mixture was incubated at room temperature for 1鈥塰 on a rocker platform to keep the Dynabeads in suspension. After an initial wash with buffer (10鈥塵M Tris-HCl (pH 7.5), 1鈥塵M EDTA, 2.0鈥塎 NaCl), the Dynabeads-bound DNA fragments were denatured in 0.1鈥塏 NaOH for 10鈥塵in, and the supernatant containing the nonbiotinylated strands was removed. The biotinylated DNA strands were neutralized with 0.2鈥塎 Tris-HCl (pH 7.5) and resuspended in 10鈥壩糽 of water. The entire 10鈥壩糽 of the biotinylated DNA was used as a template in the second round of PCR. The same procedure was used for isolation of the biotinylated DNA strands from the second round of PCR with three exceptions: 40鈥壩糶 of the Dynabeads were used; the Dynabeads-bound biotinylated DNA strands were suspended in 30鈥壩糽 of TE; then only 1鈥壩糽 was used as a template in the third round PCR.Sequencing of the PCR productsThe PCR products from the second or third round were purified using Microcon-PCR. A nested vector-specific primer and Tag primer were used to sequence both ends of the fragment. When direct sequencing failed to generate acceptable reads (usually owing to the Tag primer), the PCR products were isolated by agarose gel electrophoresis and purified by a glass beads procedure. The purified DNA fragments were cloned into TOPO pCR2.1 vector using the TOPO TA Cloning Kit (InVitrogen), then each end of the insert was sequenced using forward and reverse M13 primers. All sequencing was performed on an ABI 377 DNA sequencer with the ABI Prism BigDye Terminator v3.0 (Applied Biosystems). When cloning was used to facilitate DNA sequencing, the sequence was confirmed by reamplification of the uncloned PCR products using the newly identified sequences. References1Herweijer H, Wolff JA . Progress and prospects: naked DNA gene transfer and therapy. Gene Therapy 2003; 10: 453鈥?58.CAS聽 Article聽Google Scholar聽 2Donnelly JJ, Ulmer JB, Shiver JW, Liu MA . DNA vaccines. Annu Rev Immunol 1997; 15: 617鈥?48.CAS聽 Article聽Google Scholar聽 3Nishikawa M, Huang L . Nonviral vectors in the new millennium: delivery barriers in gene transfer. Hum Gene Ther 2001; 12: 861鈥?70.CAS聽 Article聽Google Scholar聽 4Bigey P, Bureau MF, Scherman D . In vivo plasmid DNA electrotransfer. Curr Opin Biotechnol 2002; 13: 443鈥?47.CAS聽 Article聽Google Scholar聽 5Gehl J . Electroporation: theory and methods, perspectives for drug delivery, gene therapy and research. Acta Physiol Scand 2003; 177: 437鈥?47.CAS聽 Article聽Google Scholar聽 6Trezise AEO . In vivo DNA electrotransfer. DNA Cell Biol 2002; 21: 869鈥?77.CAS聽 Article聽Google Scholar聽 7Fattori E, La-Monica N, Ciliberto G, Toniatti C . Electro-gene-transfer: a new approach for muscle gene delivery. Somat Cell Mol Genet 2002; 27: 75鈥?3.CAS聽 Article聽Google Scholar聽 8Nichols WW, Ledwith BJ, Manam SV, Troilo PJ . Potential DNA vaccine integration into host cell genome. Ann NY Acad Sci 1995; 772: 30鈥?9.CAS聽 Article聽Google Scholar聽 9Ledwith BJ et al. Plasmid DNA vaccines: assay for integration into host genomic DNA. Dev Biol (Basel) 2000; 104: 33鈥?3.CAS聽Google Scholar聽 10Ledwith BJ et al. Plasmid DNA vaccines: investigation of integration into host cellular DNA following intramuscular injection in mice. Intervirology 2000; 43: 258鈥?72.CAS聽 Article聽Google Scholar聽 11Manam S et al. Plasmid DNA vaccines: tissue distribution and effects of DNA sequence, adjuvants and delivery method on integration into host DNA. Intervirology 2000; 43: 273鈥?81.CAS聽 Article聽Google Scholar聽 12Martin T et al. Plasmid DNA malaria vaccine: the potential for genomic integration after intramuscular injection. Hum Gene Ther 1999; 10: 759鈥?68.CAS聽 Article聽Google Scholar聽 13Rizzuto G et al. Efficient and regulated erythropoietin production by naked DNA injection and muscle electroporation. Proc Natl Acad Sci USA 1999; 96: 6417鈥?422.CAS聽 Article聽Google Scholar聽 14Minami M, Poussin K, Brechot C, Paterlini P . A novel PCR technique using Alu-specific primers to identify unknown flanking sequences from the human genome. Genomics 1995; 29: 403鈥?08.CAS聽 Article聽Google Scholar聽 15Sorensen AB, Duch M, Jorgensen P, Pedersen FS . Amplification and sequence analysis of DNA flanking integrated proviruses by a simple two-step polymerase chain reaction method. J Virol 1993; 67: 7118鈥?124.CAS聽 PubMed聽 PubMed Central聽Google Scholar聽 16Longo MC, Berninger MS, Hartley JL . Use of uracil DNA glycosylase to control carry-over contamination in polymerase chain reactions. Gene 1990; 93: 125鈥?28.CAS聽 Article聽Google Scholar聽 17Krayev AS et al. The nucleotide sequence of the ubiquitous repetitive DNA sequence B1 complementary to the most abundant class of mouse fold-back RNA. Nucleic Acids Res 1980; 8: 1201鈥?215.CAS聽 Article聽Google Scholar聽 18Waterston RH, et al. Mouse Genome Sequencing Consortium. Initial sequencing and comparative analysis of the mouse genome. Nature 2002; 420: 520鈥?62.CAS聽 Article聽Google Scholar聽 19Zhang G et al. Efficient expression of naked DNA delivered intraarterially to limb muscles of nonhuman primates. Hum Gene Ther 2001; 12: 427鈥?38.CAS聽 Article聽Google Scholar聽 20Hui EK-W, Wang P-C, Lu SJ . Strategies for cloning unknown cellular flanking DNA sequences from foreign integrants. Cell Mol Life Sci 1998; 54: 1403鈥?411.CAS聽 Article聽Google Scholar聽 21Jin YF, Ishibashi T, Nomoto A, Masuda M . Isolation and analysis of retroviral integration targets by solo long terminal repeat inverse PCR. J Virol 2002; 76: 5540鈥?547.CAS聽 Article聽Google Scholar聽 22W眉rtele H, Little KCE, Chartrand P . Illegitimate DNA integration in mammalian cells. Gene Therapy 2003; 10: 1791鈥?799.Article聽Google Scholar聽 23Merrihew RV et al. High-frequency illegitimate integration of transfected DNA at preintegrated target sites in a mammalian genome. Mol Cell Biol 1996; 16: 10鈥?8.CAS聽 Article聽Google Scholar聽 24Bode J et al. Fatal connections: when DNA ends meet on the nuclear matrix. J Cell Biochem 2000; Suppl 35: 3鈥?2.Article聽Google Scholar聽 25Wu X, Li Y, Crise B, Burgess SM . Transcription start regions in the human genome are favored targets for MLV integration. Science 2003; 300: 1749鈥?751.CAS聽 Article聽Google Scholar聽 26Nakai H et al. AAV serotype 2 vectors preferentially integrate into active genes in mice. Nat Genet 2003; 34: 297鈥?02.CAS聽 Article聽Google Scholar聽 27Zhang L, Nolan E, Kreitschitz S, Rabussay DP . Enhanced delivery of naked DNA to the skin by non-invasive in vivo electroporation. Biochim Biophy Acta 2002; 1572: 1鈥?.CAS聽 Article聽Google Scholar聽 28Gilbert RA, Jaroszeski MJ, Heller R . Novel electrode designs for electrochemotherapy. Biochim Biophys Acta 1997; 1334: 9鈥?4.CAS聽 Article聽Google Scholar聽 Download referencesAcknowledgementsWe thank Carolann Beare, Steven Hall, Donna Lynch, and Brenda Givler for assistance with the antemortem portion of the study, and Yuan Liu for assistance with analysis of the integration sites using Bioinformatics. We also thank Warren Nichols for his continued interest and ever insightful discussions.Author informationAffiliationsDepartment of Biologics Safety Assessment, Merck Research Laboratories, West Point, PA, USAZ Wang,聽P J Troilo,聽X Wang,聽T G Griffiths II,聽S J Pacchione,聽A B Barnum,聽L B Harper,聽C J Pauley,聽Z Niu,聽L Denisova,聽T T Follmer聽 聽B J LedwithIstituto Di Ricerche Di Biologia Molecolare, 00040 Pomezie, Rome, ItalyG Rizzuto,聽G Ciliberto,聽E Fattori聽 聽N L MonicaDepartment of Regulatory Affairs, Merck Research Laboratories, Blue Bell, PA, USAS ManamAuthorsZ WangView author publicationsYou can also search for this author in PubMed聽Google ScholarP J TroiloView author publicationsYou can also search for this author in PubMed聽Google ScholarX WangView author publicationsYou can also search for this author in PubMed聽Google ScholarT G Griffiths IIView author publicationsYou can also search for this author in PubMed聽Google ScholarS J PacchioneView author publicationsYou can also search for this author in PubMed聽Google ScholarA B BarnumView author publicationsYou can also search for this author in PubMed聽Google ScholarL B HarperView author publicationsYou can also search for this author in PubMed聽Google ScholarC J PauleyView author publicationsYou can also search for this author in PubMed聽Google ScholarZ NiuView author publicationsYou can also search for this author in PubMed聽Google ScholarL DenisovaView author publicationsYou can also search for this author in PubMed聽Google ScholarT T FollmerView author publicationsYou can also search for this author in PubMed聽Google ScholarG RizzutoView author publicationsYou can also search for this author in PubMed聽Google ScholarG CilibertoView author publicationsYou can also search for this author in PubMed聽Google ScholarE FattoriView author publicationsYou can also search for this author in PubMed聽Google ScholarN L MonicaView author publicationsYou can also search for this author in PubMed聽Google ScholarS ManamView author publicationsYou can also search for this author in PubMed聽Google ScholarB J LedwithView author publicationsYou can also search for this author in PubMed聽Google ScholarCorresponding authorCorrespondence to B J Ledwith.Rights and permissionsReprints and PermissionsAbout this articleCite this articleWang, Z., Troilo, P., Wang, X. et al. Detection of integration of plasmid DNA into host genomic DNA following intramuscular injection and electroporation. Gene Ther 11, 711鈥?21 (2004). https://doi.org/10.1038/sj.gt.3302213Download citationReceived: 27 August 2003Accepted: 07 November 2003Published: 15 January 2004Issue Date: 01 April 2004DOI: https://doi.org/10.1038/sj.gt.3302213KeywordsintegrationelectroporationplasmidDNA vaccine Yen-Der Li, Wei-Yu Chi, Jun-Han Su, Louise Ferrall, Chien-Fu Hung T.-C. Wu Journal of Biomedical Science (2020) Lei Yu, Jun Pan, Guangli Cao, Mengsheng Jiang, Yunshan Zhang, Min Zhu, Zi Liang, Xing Zhang, Xiaolong Hu, Renyu Xue Chengliang Gong Virology Journal (2020) Patrick Tjok Joe, Ioanna Christopoulou, Lien van Hoecke, Bert Schepens, Tine Ysenbaert, Carlo Heirman, Kris Thielemans, Xavier Saelens Joeri L. Aerts Journal of Translational Medicine (2019)