Experimental endostatin-GFP gene transfection into human retinal vascular endothelial cells using ultrasound-targeted cationic microbubble destruction

Cell culture

Human retinal vascular endothelial cells (HRECs) were purchased from TongPai Biotechnology Co., Ltd (Shanghai, China) and grown in RPMI-1640 medium (Gibco, San Jose, CA) supplemented with 10% fetal bovine serum (FBS; v/v), 100 μg/ml streptomycin (Sigma, St. Louis, MO), and 100 U/ml penicillin (Sigma) at 37 °C in a 5% CO₂ humidified atmosphere. The cell numbers for all experiments were determined with a hemocytometer.

Plasmid preparation

A eukaryotic coexpression vector (pEZ-M46-ES) encoding green fluorescent protein (GFP) was obtained from GeneCopoeia^TM Co. (Guangzhou, China). The GFP sequence was directly amplified via PCR using the upstream and downstream primers 5′-ATC GTT CGA ACC ATG GCG CCG AGG TGC CCC TGG CC-3′ and 5′-ATC GTG CGG CCG CAC TCG AGG TAC TTG GAG GCA GTC ATG AAG CTG T-3′, respectively. The recombinant plasmid was designated ES-GFP. The plasmids were grown under kanamycin selection in the host strain DH5α and purified through alkaline lysis and chromatographic methods using an EndoFree kit (Qiagen, Valencia, CA) at a concentration of 500 ng/μl. The concentration of the plasmid DNA was determined via spectrophotometry (NanoDrop 2000/2000C, Thermo Scientific, Waltham, MA).

CMB and NMB preparation and characterization

The neutral microbubbles (NMBs) consisted of 5 mg dipalmitoyl phosphatidylcholine (DPPC) and 2 mg 1,2-dipalmitoyl-sn-glycero-3-phosphoethanol-amineN-[methoxy(polyethylene glycol)-2000] (DSPE-PEG2000). The CMBs consisted of 5 mg DPPC, 2 mg 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[biotinyl (polyethylene glycol)-2000] (DSPE-PEG2000-Biotin), and 1 mg 3-(N-(N',N'-dimethylaminoethane)carbamoyl)cholesterol (DC-Chol). All chemicals were obtained from Avanti Polar Lipids (Alabaster, AL). They were dissolved in PBS (1X; 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, pH 7.4) at a final volume of 0.5 ml in 1.5-ml vials. The vials were incubated at 37°C for 30 min. The headspace in each vial was filled with perfluoropropane gas, and the vial was then mechanically shaken for 45 s using a dental amalgamator (Medical Apparatus and Instrument, YJT, Shanghai, China). This solution was diluted with 0.5 ml PBS and sterilized through Co60 irradiation. The density of the cationic microbubbles was 1×10⁹/ml.

Both types of microbubbles were stable in suspension for up to 1 month at 4 °C. The size distribution and zeta potential were analyzed in Opti-MEM (Invitrogen, Carlsbad, CA) using a laser particle size analyzer system (Zetasizer 3000HS; Malvern, UK).

Assessment of plasmid binding to microbubbles

The strategy for evaluating the amount of plasmid DNA that can bind to microbubbles followed the method described by Panje et al. [22]. Briefly, 5×10⁸ NMBs and CMBs were incubated with varying doses of ES-GFP plasmids (10, 20, and 40 μg) in Opti-MEM for 15 min to allow spontaneous electrostatic charge-coupling between the DNA and the microbubbles. Then the samples were centrifuged at 400 ×g for 4 min to form two phases: The upper layer contained microbubbles with bound DNA, and the lower layer contained unbound DNA. The lower layer was collected and centrifuged at 14,000 ×g for 5 min using 0.45-μm centrifugation filters to remove MB shell contaminants. The concentration of unbound DNA was determined via spectrophotometry (NanoDrop 2000/2000C, Thermo Scientific) and by multiplying the total sample volume obtained by the total amount of unbound DNA. The amount of bound DNA was calculated by subtracting the amount of unbound DNA from the total amount of DNA added initially. By increasing the initial amount of DNA, the maximum amount of bound DNA was obtained, which was regarded as the loading capacity for 5×10⁸ microbubbles. All experiments were performed in triplicate.

Gene transfer

ES-GFP plasmids grown in Escherichia coli were purified and suspended in 2.5 mM Tris-HCl (pH 8.5) at a concentration of 1.0 μg/μl. Five micrograms of plasmids (5 μl with 5×10⁸ of a CMB or NMB suspension were gently blended, and the mixture was gently incubated for a few minutes at 4 °C to achieve adhesion. Cells were transfected with the ES-GFP vector using Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Five micrograms of plasmids were gently blended with 45 μl of RPMI-1640 medium or 3 μl of liposome plus 47 μl of RPMI-1640 medium, and these two components were mixed at room temperature for 20 min. Gene transfer (UGT 1025 type, Chongqing, China) was performed by the Ultrasonographic Image Research Institute of Chongqing Medical University. The following parameters were used for the ultrasound treatment of the HRECs: continuous wave, 1 MHz, 1 W/cm², 20% duty cycle, 1 min, and a 10% concentration of microbubbles. Cultured HRECs were divided into four groups: (1) control (no treatment), (2) ES-GFP plasmid with NMBs and ultrasound (hereafter referred to as the NMB group), (3) ES-GFP plasmids with liposomes (hereafter referred to as the liposome group), and (4) ES-GFP plasmids with CMBs and ultrasound (hereafter referred to as the CMB group). The second and fourth groups of cells were grown in an Opticell for 24 h before transfection. The Opticell contained two layers of parallel breathable polystyrene (PS) film with a special surface treatment, making it suitable for cell growth. The total size of the training device was the same as a traditional microporous plate. After the microbubble composite plasmid was injected, the Opticell was inverted to allow full contact between the microbubbles and the cells to maximize the effect. After incubation for 6 h at 37 °C with 5% CO₂, the compound solution was replaced with 500 μl of fresh media containing 10% fetal bovine serum and 1% penicillin/streptomycin. The cells were cultured for another 24 h, and the reporter ES-GFP expression was analyzed. All transfection experiments were performed in triplicate.

MTT assay

The influence of all treatment factors on the cell viability of the HRECs was assessed using the 3-(4,5-dimthylthiazol-2-yl)-2,5 diphenyl-tetrazolium bromide (MTT) assay. After transfection, all groups of HRECs were collected and immediately seeded into 96-well plates at a density of 1×10⁴ cells per well. When the cells had grown to approximately 60–80% confluency, the old medium was replaced with 200 μl of RPMI-1640 without FBS and penicillin/streptomycin. The cells were then incubated at 37 °C and inspected daily for 5 days. Cell growth was determined by adding 20 μl of MTT (5 μg/ml) to each well, followed by incubation for an additional 4 h at 37 °C. The supernatants were discarded, and 150 μl of dimethyl sulfoxide was added. The absorbance was determined via spectrophotometry using a wavelength of 570 nm, with 630 nm as a reference. All experiments were performed in triplicate.

qPCR

The total RNA was extracted via phenol-chloroform extraction with ethanol precipitation using TRIzol (Life Technologies, Gaithersburg, MD). The integrity of the RNA was determined on 1.5% agarose gels containing ethidium bromide. Reverse transcription to cDNA was performed using the Takara PrimeScript™ RT Reagent Kit (AMV; Takara, Tokyo, Japan) with Bio-Rad iQ5. The reaction conditions were as follows: 25 °C for 10 min, 42 °C for 50 min, and 85 °C for 5 min. The quantitative real-time PCR (qPCR) amplification conditions were as follows: 94 °C for 4 min; 35 cycles of 94 °C for 20 s, 60 °C for 30 s, and 72 °C for 30 s; and finally, the signal was measured at 72 °C. The bands were normalized to β-actin. The primer sequences are provided in Table 1.

Western blot analysis

All HREC groups were collected 72 h after transfer. Briefly, the proteins were separated via sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) in 12% (w/v) polyacrylamide gels and electrotransferred to Millipore polyvinylidene difluoride membranes (Millipore, Bedford, MA). The completion of the protein transfer from the gels to the membranes was verified by staining the gels with Coomassie Blue R-250. The membranes were blocked in Tris-buffered saline containing 0.05% Tween-20 and 5% nonfat dry milk for 60 min at room temperature. Then, the blots were incubated with a 1:1,000 dilution of a rabbit anti-human endostatin antibody (Abcam, London, UK), rabbit anti-human VEGF antibody (Abcam), or rabbit anti-human ES antibody (Abcam) for 1.5 h at room temperature in blocking solution. The membranes were washed in Tris-buffered saline and incubated with horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (IgG; 1:3,000) in blocking solution for 1.5 h. The blots were examined using a chemiluminescent substrate (Santa Cruz Biotechnology, Santa Cruz, CA) according to the manufacturer’s instructions. The protein bands were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and all of the blots were quantified using Quantity One software (Bio-Rad, Hercules, CA).

Statistical analysis

All of the data are expressed as the mean ± standard deviation and were analyzed using SPSS 17.0 software (SPSS Inc., Chicago, IL). Means showing p<0.05 were considered to be significantly different from one another. Comparisons between groups were performed using one-way ANOVA test. The least significant difference (LSD) and Student t tests were used for further comparisons.

Characterization of microbubbles

The characteristics of the CMBs and the NMBs are listed in Table 2. There were no significant differences in the mean diameters or concentrations of the NMBs and the CMBs, but the zeta potentials of the CMBs (+25.62±4.38 mV) were significantly higher than those of the NMBs (−2.78±1.56 mV; p<0.001). Inverted microscopy images of the NMBs and CMBs are shown in Figure 1.

Plasmid binding ability of microbubbles

Quantitative assessment of the DNA loading capacity of the NMBs and the CMBs revealed significantly higher plasmid binding abilities for the CMBs compared with the NMBs at every dose. As the number of added plasma molecules increased, the amount of plasmids bound to the CMBs increased as well. However, no further increases in the amount of plasmids bound to the CMBs was observed when 40 μg of plasmids was added to the microbubble solution. Therefore, the amount of plasmids that bound to the microbubbles when 40 μg of plasmid DNA was added to the microbubble solution was regarded as the saturated DNA loading capacity of 5×10⁸ CMBs. The saturated DNA loading capacities of the NMBs and the CMBs were 19.02±0.76 μg and 5.01±0.74 μg, respectively (Figure 2, p<0.05). The zeta potential of the CMBs decreased slightly (+16.32±3.28) after the bound DNA, but the size of the CMBs has no obvious change, and the size and the zeta potential of NMBs had no obvious change after the bound DNA.

MTT assay

The influence of all treatment factors on HREC cell viability was evaluated using the MTT assay. The cells treated with the ES-GFP plasmid together with NMBs and ultrasound, or with the ES-GFP plasmid with liposomes, or the ES-GFP plasmid with CMBs and ultrasound all underwent apoptotic-like cytological changes characterized by a loss of anchorage dependence and the assumption of a small spherical shape. The cells were collected and counted from day 1 through day 5, and the cell viability rate of the control group was 100%; the viability rate of the cells in each group was as shown in Figure 3. In multiple trials, the cell viability rate was observed to decrease significantly in the CMBs compared with the other groups.

qPCR

To evaluate the transcription of the ES-GFP gene after transfection into HRECs in vitro, the relative ratio of ES RNA in the treated and control HRECs was determined via qPCR at 48 h post-transfection (Figure 4A). To evaluate the effect of endostatin on VEGF, Bcl-2, and Bcl-xl expression, the relative ratios of VEGF, Bcl-2, and Bcl-xl RNA were also determined (Figure 4B–D). Following transfer, the transcription levels of the ES gene were highest in the CMB group (p<0.05 relative to the NMB and liposome groups, p<0.01 relative to the control group), while there was almost no difference between the NMB and liposome groups (p>0.05). With the increasing levels of ES RNA, the level of VEGF, Bcl-2, and Bcl-xl mRNA decreased (Figure 4). Thus, these effects may be due to the therapeutic effect induced by endostatin.

Western blotting

To evaluate the delivery of the ES-GFP gene into HRECs in vitro, the total amount of endostatin protein in the treated and control HRECs was determined with western blotting at 48 h post-transfection (Figure 5A). The endostatin protein level following transfer using CMBs and ultrasound was increased compared with the other groups. The difference in the endostatin protein levels between the CMB group and the other groups was significant (p<0.05 relative to the NMB and liposome groups, p<0.01 relative to the control group). However, the difference in the endostatin protein levels between the NMB group and the liposome group was not significant (p>0.05). To better evaluate the therapeutic effect following ES-GFP gene transfer, the protein levels of VEGF, Bcl-2, and Bcl-xl were examined. As the levels of the endostatin protein increased, the VEGF, Bcl-2, and Bcl-xl protein level decreased (Figure 5B-D), confirming the antiangiogenesis and apoptosis effects of endostatin.