Electroporation outperforms in vivo-jetPEI for intratumoral DNA-based reporter gene transfer

Intratumoral delivery of drug-encoding plasmid DNA (pDNA) enables localised in vivo expression of biological drugs, offering an attractive alternative to conventional protein treatment. However, this requires physical or chemical methods to enhance the low transfection efficiency of naked pDNA. Electroporation and complexation with the polycation in vivo-jetPEI are both evaluated in the clinic for intratumoral pDNA delivery, but lack head-to-head comparison. This study therefore compared both methods for intratumoral DNA-based reporter gene transfer in a subcutaneous mouse tumour model. Intratumoral electroporation resulted in strong reporter expression that was restricted to the tumour area and persisted for at least ten days. Intratumoral expression after injection of pDNA-jetPEI complexes was two to three logs lower, did not exceed the background in most mice, and lasted less than five days even with repeated dosing. Remarkably, reporter expression was primarily detected in the lungs, presumably due to leakage of pDNA-jetPEI complexes into the systemic circulation. In conclusion, electroporation enabled more efficient, prolonged and tumour-specific reporter expression compared to intratumoral injection of pDNA complexed with in vivo-jetPEI. These results favour the use of electroporation for intratumoral DNA-based gene transfer, and suggest further optimisation of pDNA-jetPEI complexes is needed to improve their efficacy and biosafety.


Results
The efficiency of electroporation-and polymer-based transfection was evaluated by means of intratumoral reporter gene transfer in C57BL/6J mice bearing a subcutaneous MC38 tumour. A firefly luciferase (fluc)encoding plasmid (pFluc) was delivered, which allowed both localisation and quantification of the resulting reporter expression in vivo and ex vivo. Electroporation pulse parameters were extracted from the literature 11 and recently validated in-house 3,4 . The pDNA-jetPEI complexes were prepared and administered according to the instructions of the manufacturer (Polyplus-transfection).
Intratumoral electroporation of 20 µg pFluc led to strong reporter expression, which was restricted to the tumour area and did not decline for at least ten days, until the mice had to be sacrificed because of increased tumour volume. Intratumoral injection of the same pFluc dose complexed with in vivo-jetPEI gave fluc signals virtually similar to the background in untreated mice, and 200-to 1,800-fold lower compared to the signals after intratumoral electroporation (Fig. 1A, p = 0.0079). Both transfection methods had a similar effect on tumour growth (Fig. 1B).
To improve in vivo transfection, repeated intratumoral pFluc-jetPEI dosing was evaluated and compared to a single pFluc-jetPEI injection. In this experiment, pFluc-jetPEI did enable clear but low bioluminescence signals at the tumour in three out of six mice. However, the signal dropped after one day and reached background values by day five in both treatment groups ( Fig. 2A). One mouse exhibited aberrantly high off-target expression that interfered with the measurement of the intratumoral fluc signal, and was therefore excluded in the analyses. Remarkably, also in other mice, the highest fluc expression was detected outside the tumour area at the chest region (Fig. 2B). This expression pattern suggested transfection of the lungs, which was confirmed by ex vivo imaging of different organs (Fig. 2C). We hypothesise that this was caused by leakage of pFluc-jetPEI in the systemic circulation, since it has previously been shown that intravenous injection of cationic systems like in vivo-jetPEI lead to high intrapulmonary expression 12 . With electroporation, naked pFluc can also leak into the bloodstream, but transfection is restricted to the area of the applied electrical field 3,5,13 , as further confirmed in the current study.

Discussion
Intratumoral gene transfer has shown promising results for various biological drugs in both preclinical and clinical trials. Most results to date are based on the use of viral vectors 2 , despite recent progress with intratumoral pDNA delivery 3,7,8,14 . Comparison of some of the most used pDNA transfection methods may guide future research to further improve the efficacy of DNA-based intratumoral gene transfer and advance the non-viral field.
In this preclinical study, we report the comparison of two intratumoral pDNA transfection methods: electroporation and complexation with the polycation in vivo-jetPEI. Both methods were compared by means of bioluminescence imaging of fluc expression, which provides a straightforward and quantitative readout. Whereas the results with electroporation were in line with the literature 3,13 , pDNA-jetPEI complexes failed to efficiently transfect the tumour, even with repeated dosing. This is in contrast with the anti-tumour responses observed after jetPEI-driven intratumoral gene transfer in some preclinical and clinical trials 8,14 . Interestingly, Ohlfest www.nature.com/scientificreports/ et al. also reported very limited intratumoral reporter expression with pDNA-jetPEI complexes, but were able to improve transfection by reducing the speed of injection (from 100 µl in > 5 s to 100 µl in 60 s). However, delivery with a micropump at 10 µl/min again reduced pDNA transfection 15 . Coll et al., on the other hand, showed that intratumoral reporter expression could be increased up to tenfold by switching from manual to micropumpassisted (20 µl/min) intratumoral injection of pDNA complexed with a linear polyethylenimine, and by adapting the N/P ratio (from 5 or 20 to 10), which represents the number of nitrogen residues of the polyethylenimine per pDNA phosphate group 16 . In the current study, only one setup was evaluated for the pDNA-jetPEI complexes (N/P ratio of 8 and injection of 50 µl in less than 20 s). Still, it is uncertain that even with additional improvements, in vivo-jetPEI could match electroporation, as the latter outperformed in vivo-jetPEI by a factor of 200 to 1,800 in our study. A few studies describe the spatial distribution of reporter expression after intratumoral DNA-based gene transfer. Both electroporation and in vivo-jetPEI have been shown to enable expression restricted to the tumour site 13,15 . In the current study, however, clear transfection of the lungs was detected in multiple mice after intratumoral pDNA-jetPEI injection, whereas expression following electroporation was always limited to the tumour area. Even though based on a limited data set, these results suggest that considerable attention must be given to pDNA distribution when using pDNA-complexing agents. In addition to improving transfection, as suggested above, slower injection could increase the retention of the pDNA complexes in the tumour and reduce leakage www.nature.com/scientificreports/ into the circulation 16 . Modification of cationic carriers with e.g. tumour-targeting peptides or the use of tumourspecific promoters could further avoid off-target transgene expression, reducing the risk of systemic toxicity with polymer-based intratumoral gene therapy 12 .
The conclusions of this study need to be interpreted with caution as they are based on a restricted number of experiments in a single tumour model, with a single pDNA-jetPEI composition and a single gene expression readout. Nevertheless, despite the small sample sizes, clear and significant differences were observed between electroporation-and jetPEI-driven transfection, which were consistent across experiments. Future studies may be considered to evaluate if these conclusions can be extrapolated to a general context, e.g. by comparing both transfection methods in additional tumour models. In follow-up of the similar tumour growth observed with both transfection methods, a more thorough assessment of their impact on cell viability could be considered. For the pDNA-jetPEI complexes, the composition and way of administration may be further optimised. However, as mentioned earlier, it is uncertain that these improvements will be sufficient to match the performance of electroporation.
In summary, we demonstrated that electroporation resulted in stronger, longer-term and more localised expression compared to pDNA complexation with in vivo-jetPEI, when applied for intratumoral DNA-based reporter gene transfer. This confirms the potential of electroporation, and illustrates that further optimisation of the in vivo-jetPEI complexes is required to allow efficient tumour-specific pDNA transfection.

Methods
Plasmid DNA. pFluc was obtained from Icosagen (Tartu, Estonia). This pDNA construct encodes firefly luciferase 2 under control of a CAG promoter, and contains a backbone comprising an ampicillin resistance gene and pUC origin of replication. pFluc production and purification was performed as described previously 17 , except that pFluc was eluted with sterile milliQ water instead of D-PBS.
Mouse tumour model. The MC38 murine colon cancer cell line was purchased from Kerafast (ENH204-FP, Boston, MA, USA) in March 2017 and was shown to be free of Mycoplasma. Cells were grown in Dulbecco's Modified Eagle Medium, supplemented with 10% heat-inactivated foetal bovine serum, 0.1 mM non-essential amino acids, 1 mM sodium pyruvate, 10 mM HEPES and 50 U/ml penicillin/streptomycin (Thermo Fischer Scientific, Waltham, MA, USA), in a humidified incubator at 37 °C and 5% CO 2 . Before tumour injections, cells were harvested with 0.05% trypsin-EDTA (25300-054, Thermo Fischer Scientific) and resuspended in D-PBS (14190-094, Thermo Fischer Scientific). 1 × 10 6 MC38 cells in 100 µl were injected subcutaneously in the right flank of 6-to 7-week-old female C57BL/6J mice. Tumour volumes were evaluated with an electronic calliper (500-712-20, Mitutoyo, Kawasaki, Japan), and calculated with the formula length x width 2 × 0.5. Mice were sacrificed when tumour volume exceeded 1500 mm 3 , or when reporter expression was comparable to background. C57BL/6J mice were purchased from Charles River Laboratories (Saint Germain Nuelles, France) or bred at the KU Leuven Animal Research Center. All animal experiments were approved by the KU Leuven Animal Ethics Committee (P130/2017) and were performed in accordance with the regulations of the European Union and Belgium concerning the protection of laboratory animals.
Intratumoral DNA transfection. Intratumoral electroporation was performed according to a previously described protocol 3,11 . 20 µg pDNA in 50 µl sterile milliQ water was injected intratumorally, immediately followed by two series of four 5-ms square-wave pulses of 600 V/cm in perpendicular directions at a frequency of 1 Hz. Electrical pulses were delivered by the NEPA21 Electroporator (Sonidel Limited, Dublin, Ireland) with CUY650P5 tweezer electrodes (Sonidel Limited) at a fixed width of 5 mm and covered with Eco Ultrasound Transmission Gel (G0066, Fiab, Vicchio, Italy). Pulse current and total energy were verified with the NEPA21 readout.
For electroporation-and jetPEI-driven transfection, pDNA injection was performed manually with a syringe without thorough control of the injection speed. On average, injection of 50 µl lasted 10-20 s. Bioluminescence imaging. Fluc expression was visualised and quantified by bioluminescence imaging (IVIS Spectrum, PerkinElmer, Waltham, MA, USA) at the Molecular Small Animal Imaging Center (MoSAIC) at KU Leuven. For in vivo imaging, mice were subcutaneously injected with 126 mg/kg D-luciferin substrate (E6551, Promega, Madison, WI, USA) at 15 mg/ml in D-PBS, after which bioluminescence intensity was measured every two minutes. Intratumoral fluc expression was defined as the maximal total radiance (in photons per second) measured in a specified region of interest covering the tumour area. Mice were anesthetised by isoflurane inhalation during the whole procedure. For ex vivo imaging, mice received a second subcutaneous injection with D-luciferin after in vivo bioluminescence imaging. Five minutes later, mice were sacrificed. Different organs were excised and analysed. Fluc expression in the individual organs was calculated as the total bioluminescent signal measured in a region of interest covering the organ.
Statistics. At the start of experiments, mice were randomised based on tumour volume and weight. Data of different groups were compared at different time points with Mann-Whitney tests. Outliers were detected with the Grubb's test. Statistical analyses were performed with Graphpad Prism 8.4.3 (Graphpad Software, San Diego, CA, USA), with two-sided P values below 0.05 considered as significant.

Data availability
The datasets generated or analysed during the current study are available from the corresponding authors on reasonable request.