Induction of Apoptosis in Sonoporation and Ultrasonic Gene Transfer

Abstract

The role of apoptosis in sonoporation and ultrasound enhanced gene transfection of cell suspensions was examined in vitro. Suspensions of HL-60 and of CHO-K1 cells were exposed to 2.25 MHz continuous ultrasound for 1 min in a 60 rpm rotating-tube exposure system, with ultrasound contrast media added to ensure nucleation of cavitation. Cell necrosis was measured by trypan blue dye exclusion (using a hemacytometer) and by propidium iodide nuclear staining (using flow cytometry). Apoptosis was detected by the annexin V method with Alexa Fluor 350 as the fluorescent label, and confirmed by Hoechst 33342 nuclear staining. Sonoporation cell loading was assessed by uptake of large fluorescent-dextran molecules from the medium. Transfection was demonstrated by expression of green fluorescent protein (GFP) from plasmids transferred into the cells by the treatment. Cell scoring was performed by flow cytometry with necrotic cell events excluded. For HL-60 cells at 0.4 MPa, cell loading and transfection was significantly increased relative to shams at 2 h, 6 h and 24 h post exposure, peaking at 19.0 ± 5.5 % and 9.6 ± 4.2 % of non-necrotic cells, respectively, at 6 h. However, about one third of the treatment-positive cells were identified as apoptotic. The cell loading and gene transfer effects increased for increasing peak rarefactional pressure amplitude, reaching 24.4 ± 7.7 % and 12.7 ± 5.1 % of non-necrotic cells, respectively for 0.6 MPa exposure. However, the lethal cellular injury caused by cavitation in the rotating tube system reduced the overall apparent efficacy of cell loading and gene transfer to 5.1 ± 2.1 % and 2.1 ± 0.9 %, respectively, after accounting for necrosis and apoptosis. Similar tests with CHO cells showed increased sonoporation but mostly cell death by necrosis, rather than apoptosis. The induction of apoptosis by cavitation treatments should be considered as a possible confounding factor, in addition to necrosis, in sonoporation and ultrasonic gene transfer research.

Introduction

When acoustic cavitation occurs during ultrasound exposure of a biological medium, cell injury is likely to occur upon interaction of a cell with a microbubble in violent pulsation (NCRP, 2002). The mechanical bioeffects of cavitation can range from transient permeabilization of cell membranes, called sonoporation, to permanent permeabilization and necrosis. Inertial cavitation can generate free radicals, due to high temperatures reached upon collapse of the microbubbles, but bioeffects from free radicals and sonochemicls are difficult to detect (NCRP, 2002). Cavitational bioeffects can occur in vitro for cell suspensions or monolayers depending not only on the ultrasonic pressure amplitude and the presence of suitable cavitation nuclei in the medium, but also on a variety of other exposure parameters and culture conditions(Miller M et al, 1996).

Sonoporation has received significant research attention within the last decade because it can be utilized for delivery of molecules, which are normally excluded by the cell membrane, into surviving cells. This phenomenon is most intriguing as a means of gene transfer, which is enabled, for example, by the simple addition of plasmids to a cell suspension subject to acoustic cavitation (Bao et al. 1997; Greenleaf et al. 1998). This means of gene transfer is attractive for several reasons, such as the non viral nature of the transfer and the targeting of the gene transfer by aiming of ultrasound beams (Miller, 2006). One limitation of ultrasound for cell loading treatment is that mechanical cellular trauma accompanies the sonoporation phenomenon. Sonoporated cells showed reduced proliferation (Bao et al. 1997) and reduced plating efficiency (Miller et al. 1999) relative to unaffected cells. In addition, gene transfer and expression represents a challenge for sonoporation treatments, which appears to be more difficult to accomplish than loading of small marker molecules. In a comparison study, the fraction of treated cells, which were shown to exhibit gene expression in transfection tests with marker plasmids coding for green fluorescent protein (GFP), was shown to be much less than the fraction of treated cells, which were simply loaded with fluorescent marker molecules in sonoporation tests under the same conditions (Zarnitsyn and Prausnitz, 2004).

These and other research reports on sonoporation, gene transfer and gene expression show that translation of initial sonoporation of cell membranes into efficient gene expression was complicated by the concomitant cell injury and necrosis. However, the cell injury also appeared to induce a subtle impairment or progressive degeneration of the cells, as noted above, which is poorly understood. A possible explanation may be the induction of apoptosis in the treated cells. Apoptosis is an organized process of cell death, which occurs naturally for unneeded cells (e. g. in immune responsive cells or in embryonic development) (Kumar et al. 2005). Apoptosis also can be induced in cells by physical agents, such as ionizing radiation (Watters, 1999). Apoptotic cells have irreversible injury but do not immediately display signs of death such as failure to exclude trypan blue dye. In research, which was not directed toward sonoporation or gene transfer, exposure of cells to ultrasonic cavitation was shown to induce apoptosis in addition to the conventionally reported instantaneous cell lysis and necrotic disintegration (Ashush et al. 2000). The induction of apoptosis by 1 MHZ ultrasound exposure has been directly linked to inertial cavitation by its dependence on the presence of an ultrasound contrast agent, which provides cavitation nuclei, and by the influence of different dissolved gases (Honda et al. 2002; Feril et al. 2003). However, the induction of apoptosis was not correlated with the production of free radicals by the inertial cavitation activity, suggesting that the bioeffect mechanism was mechanical trauma. As for other cavitational bioeffects, addition of ultrasound contrast agents containing microbubbles, which can serve as cavitation nuclei, can enhance the apoptosis effect (Ando et al. 2006)

The induction of apoptosis by cavitation activity is of interest in sonoporation research, because induction of apoptosis might complicate the evaluation of the efficiency of gene transfer and expression by sonoporation due to brief gene expression before final apoptotic cell death. This present study was performed to examine the potential for apoptosis to accompany sonoporation and transfection of two cells lines: leukemic HL-60 cells, which have been shown to undergo apoptosis induced by ultrasound treatment (Ashush et al. 2000) and epithelial CHO-K1 cells, which have been used for sonoporation and gene transfer studies (Bao et al. 1997).

Materials and Methods

Cell suspension

The cell lines were obtained from the American Type Culture collection (ATCC, Rockville, MD, USA). The HL-60 human leukemia-derived cell line was used for this study, because of its previous use for apoptosis and ultrasound bioeffects research (Ashush et al. 2000). The cells were thawed from frozen storage and cultured in Iscove’s modified Dulbecco’s medium with glutamine and antibiotics (ATTC, Manassas VA USA) supplemented with 20 % fetal bovine serum in a humidified incubator at 37 C with 5% CO₂ in air. These cells did not attach to the culture flasks and trypsin was not needed for harvest. CHO-K1 cells, which is an epithelial-like cell line, were also used, because these cells were previously used for sonoporation and gene transfer research (Bao et al 1997; Miller et al 1999). These cells were cultured under the same conditions in F-12 medium with glutamine and antibiotics (GIBCO Invitrogen, Grand Ilse NY USA), which was supplemented with 10 % fetal bovine serum. The CHO cells attached to the flask surface, which required mild trypsinization for harvest. For exposure, cells were harvested and resuspended at 10⁶ cells/ml in culture medium. Definity® (perflutren lipid microsphere injectable suspension, Bristol Myers Squib Medical Imaging, N. Billerica, MA USA), which is an ultrasound contrast agent containing stabilized microbubbles suitable for cavitation nucleation, was added at 0.1 % of the volume. In addition, a fluorescent dextran (FD500) labeled with fluorescein isothiocyanate (FITC) (FD-500s, Sigma Chemical Co. St. Louis MO, USA) in saline solution at 100 mg/ml was added at 10% of the volume for tests of sonoporation. Alternatively, eGFP coding plasmids in solution, which was provided by the Vector Core Laboratory of the University of Michigan Center for Gene Therapy, was added to provide 40 μg/ml of DNA for tests of gene transfer. For exposure, the suspension mixture was loaded into UV sterilized transfer pipette bulbs (No. 241, Saint-Amand Mfg. Co. San Fernando, CA, USA) and the neck of the polyethylene pipette was heat-sealed. The bulb was 9 mm in diameter, about 2.5 cm long and contained 1.3 ml of liquid.

Ultrasound exposure

The ultrasound exposure system has been described previously (Bao, et al. 1997). Briefly, The suspension-filled pipette was clamped at the neck at the bottom of a motor shaft, which was immersed in a 37° C water bath. This allowed the bulb to be rotated at 60 rpm around its vertical axis during exposure. The polyethylene bulb allowed for well-characterized, nearly free-field exposure. The bulb rotation sustained and promoted the cavitation activity, which is known as the rotating-tube system for in vitro exposure of cell suspensions (Miller MW et al. 1996). This cell suspension exposure system was used because it allows for accurate field characterization, which gives reproducible exposure conditions.

An air-backed ultrasound transducer was aimed horizontally at the pipette bulb from a distance of 12.5 cm. The transducer was energized by a signal from a function generator (model 3314A, Hewlett-Packard Co., Palo Alto CA, USA) set to generate a 2.25 MHz continuous sine wave, which was amplified (A-500, Electronic Navigation Industries, Rochester NY, USA). The 1 min duration exposures were controlled by a manual switch. The ultrasound field from the 1.9 cm diameter transducer was measured at the bulb position with a membrane hydrophone (Model 805 PVDF Bilaminar Membrane hydrophone, Sonora Medical Systems, Longmont CO USA). A peak rarefaction pressure amplitude (PRPA) of 0.2, 0.4 or 0.6 MPa was set by the function generator voltage. The half-pressure diameter (-6 dB beam width) of the beam was 8.5 mm, about the same as the bulb.

Biological endpoints

All fluorescent labels and markers were measured by flow cytometry. Sonoporation was assessed by measuring the green cellular fluorescence resulting from the loading of the FD500 into cells. The samples with the FD500 were washed twice by centrifugation and resuspension in fresh media to reduce the background amount of fluorescent material in the medium surrounding the cells during flow cytometry. Centrifugation was at 1200 rpm for 10 min for HL-60 cells and 5 min for CHO cells. GFP expression in the cells was also assessed using the green fluorescence and included washing the cells twice for consistency. Other endpoints required addition of fluorescent stains to the samples prior to flow cytometry. Apoptosis was evaluated by using annexin V conjugated with a blue fluorescent marker (Alexa Fluor® 350, Molecular probes, Eugene OR USA) as detailed by the manufacturer. A similar stain method using annexin V conjugated to biotin plus Alexa Fluor® 350 conjugated to streptavidin, which required two steps, was also tried, but the one step procedure was easier to process and was used for most samples. The annexin V binds to phosphatidylserine exposed on the surface of apoptotic cells, which is a reliable method for in vitro apoptosis research (Koopman et al. 1994; Del Bino et al. 1999). The annexin V method can yield false positive results in necrotic cells, due to penetration and binding inside the cells. For confirmation of the detection of apoptosis, another blue fluorescent marker (Hoechst 33342, Molecular Probes, Eugene OR USA) also was used according to the manufactures directions in some samples, which were divided in half to accommodate the two tests. This marker stains the nuclei of all cells, but is brighter in apoptotic cells which have condensed chromatin.

Separate tests were conducted to assess the efficiency of sonoporation at about 2 h, 6 h and 24 h, and of gene expression at 6 h and 24 h post-exposure. Two hours was approximately the shortest time needed for exposure, counting, staining and set up at the flow cytometry facility, which was in a separate building. Within 30 min after exposure, a trypan blue dye-exclusion test was performed on each sample using a hemacytometer in order to assess the immediate cell death resulting from the ultrasound treatment. This test was probably a reasonable accounting of the necrosis induced by the treatment in the 2 h samples, but might have underestimated the cell necrosis existing after 6 h and 24 h time periods for those samples. For shams, the necrosis was too low for accurate counting by hemacytometer, and was not measured. All samples were processed by two washing steps, which tended to reduce the number of necrotic cells remaining in the samples for flow cytometry. The necrotic cells tended to swell, loose density or fragment and many did not reach the pellet during centrifugation. Therefore, the trypan blue results immediately after exposure were taken to be the best available measure of the necrosis. For both apoptosis detection methods, red fluorescent propidium iodide (PI) stain was also added to the samples to label necrotic cells. The PI results were difficult to interpret in terms of overall necrosis in exposed samples, because some of the necrotic cells were preferentially lost in centrifugation (washing) steps, and were not present at the flow cytometer. However, the PI results were valuable for the accurate assessment of apoptosis by flow cytometry. The final flow cytometry evaluation of apoptosis was made on the non-necrotic fraction with PI positive cells gated out, because the annexin V stains can enter necrotic cells and yield false positive apoptosis counts.

Flow cytometry

Control samples were run each flow cytometry session in order to aid in setting regions by placing cut-off points in two dimensional dot-plots which approximately delineate the categories of cells. One control contained unstained normal cells to indicate autofluorescence. Normal cells also were used for negative controls for the stains, which were essentially the same as the shams (as expected). A positive apoptosis control was prepared by incubating normal cells with 1 μM in camptothecin (BD Biosciences, Franklin Lakes NJ, USA) for 24 h, which is a chemical inducer of apoptosis in HL-60 cells and other cells (Wolbers et al. 2004). This positive control method did not perform as well for the CHO cells as for the HL-60 cells, even for a higher dose of 10 μM, as discussed below (Table 3). A control was also prepared using FD500 to aid in judging the amount of uptake by pinocytosis, which was not attributable to sonoporation, and the amount of fluorescent material remaining in the medium after washing. The pinocytosis was minimal for the short time the cells were in medium with fluorescent dextran, and few cells indicated a positive result in shams. Comparisons between this control, shams and exposed samples were used for setting region cut-off points for FD500 cell loading. A positive control for GFP expression was prepared by treatment of control cells with a commercial transfection reagent (Lipofectamine™, or Lipofectamine™ LTX, Invitrogen Corp., Carlsbad CA USA) to aid in gauging the GFP signal. This method did not perform well for the HL-60 cells, and comparisons between sham and exposed samples was used to set the gene expression regions for flow cytometry. All sham and exposed (I. e. non-control) samples had fluorescent markers added.

Table 3.

A comparison of the negative and positive control results for the two cell lines: the apoptosis positive control was much more effective in the HL-60 cell line but the gene transfer positive control was much more effective in the CHO cell line.

	Apoptosis		GFP Expression
	NC	PC	NC	PC
HL-60	3.5 ± 1.3	49.8 ± 7.4	0.0	0.5 ± 0.6
CHO	0.7 ± 0.4	5.2 ± 3.1	0.0	37.6 ± 7.6

Flow cytometry was performed by the University of Michigan Flow Cytometer Core facility using a FACSVantage SE Cell Sorter with argon laser (BD Biosciences, San Jose CA, USA) by standard methods (e. g. see Givan, 2001). The Flow Cytometry Standard data was analyzed with the aid of WEASEL software (V. 2.4, Walter and Eliza Hall Institute of Medical Research, Parkville, AU). The prepared samples were kept on ice until evaluated using standard flow cytometry methods for counting 10,000 cell events. A two dimensional dot-plot of forward and side scattering was used to set a gate for including events only from intact cells (I. e. excluding very low forward scattering events). The intact cells were evaluated for red (cell necrosis), blue (apoptosis assays) and green (sonoporation or gene expression) fluorescence. The wavelengths of excitation/emission for the fluorochromes were 535/617 nm for propidium iodide, 346/442 nm for Alexa Fluor® 350, 350/461 nm for Hoechst 33342, 494/518 nm for FD500, and 488/507 nm for GFP. The red, blue and green fluorescent stain spectra were well separated, which minimized cross-over for the flow cytometry channels and allowed the three tests in a the same sample. The negative and positive control samples were used to set cutoff points for the division of negatively and positively scored cells. The PI results were used to set a gate region, by comparison of normal viable cells with exposed cells, for excluding necrotic cells from the evaluation of apoptosis. The final result for each sample was the cell count dot plot, with necrotic cells excluded, for apoptosis versus FD500 or GFP. These plots identified the fractions of the cells which were normal, apoptosis positive only, sonoporation positive only and sonoporation plus apoptosis positive. The percentages of the events for each of these quadrants subsequently were averaged over 5-8 test repetitions (experiments repeated on different days) and used for evaluation of the results.

Experimental plan and statistics

The study was conducted in three separate experiments to evaluate cell apoptosis induced by sonoporation treatment. The first experiment was planned to determine if apoptosis occurred as a consequence of 0.4 MPa ultrasound treatment of HL-60 cells using the Annexin V assay, and to confirm apoptosis using the Hoechst assay (together with the various negative and positive controls) 6 h post exposure. A treatment sample was divided in half with the Annexin V assay performed on one half and the Hoechst assay on the other half. These tests were done for the sonoporation and gene transfer treatments, compared to control cells and repeated five times (excluding initial preliminary results). The second experiment was performed with HL-60 cells using the Annexin-V assay for sonoporation and gene transfer treatments for 2, 6 and 24 h post exposure at 0.4 MPa, and for 6 hr post exposure at 0.0 (sham), 0.2 and 0.6 MPa. These treatment tests were repeated seven times (excluding one run for which the cell culture appeared to be unhealthy). The third experiment involved the same tests for apoptosis and apoptosis confirmation, and the different exposure pressure amplitudes and post-exposure times for CHO cells, which were repeated eight times.

Final results are given as the mean and standard deviation, or plotted as the mean with standard error bars. Statistical comparisons were made using the Student’s t-test or Mann-Whitney rank sum test as appropriate.

Results

Figure 1 shows examples of dot plots of PI fluorescence versus the Annexin V conjugated with Alex Fluor 350 for a sham and an exposed (0.4 MPa) HL-60 sample at 6 h post exposure. A horizontal line was drawn to identify the region (the upper half of the plot) with necrotic cell events, which were 9.7% of the total in the exposed sample. As noted above, this flow cytometry count underestimated the actual cell necrosis after exposure, which was found to be 29.4% in the trypan blue dye exclusion test performed before the cell washing steps. The vertical line divides the cell events into non-apoptotic and apoptotic events. There are very few events in the upper left quadrant (PI positive but not staining for annexin V), because dead cells with highly permeable membranes will allow the annexin V to penetrate and label the phosphatidylserine within the cells (rather than exposed on the surface as for non-permeable apoptotic cells). There were many more apoptotic cell events for the exposed samples relative to the shams, which was induced by the ultrasonic cavitation treatment. Figure 2 shows examples of the final result, which has the PI positive events identified in Fig. 1 gated out, with annexin V conjugated with Alex Fluor 350 versus FD500 fluorescent dextran for the same sham and exposed (0.4 MPa) HL-60 samples used for Fig. 1. Vertical and horizontal lines set four quadrants of the plot, which approximately identifies each event dot as normal (lower left), apoptosis positive only (upper left), FD500 positive only (lower right) and FD500 plus apoptosis positive (upper right). The numerous sonoporated cells, indicated by high FD500 fluorescence, included both normal and apoptotic cells. That is, the apoptotic cells, although irreversibly injured, retain the fluorescent dextran taken up during exposure (and also GFP expression occurring before final necrosis).

Figure 1.

Flow cytometry dot-plots showing the necrosis indicator versus the apoptosis indicator for sham and exposed HL-60 samples 6 h post exposure at 0.4 MPa. Events located above the horizontal bar, which was set as a gate for exclusion of necrotic cells from subsequent analysis, included 2.3% of sham events and 9.7 % of exposed events. Events located to the right of the vertical bar were considered to be apoptotic or necrotic cells, which were greatly increased in the exposed sample.

Figure 2.

Flow cytometry dot-plots showing the apoptosis indicator versus the FD500 sonoporation indicator for sham and exposed HL-60 samples with necrotic events from Fig. 1 gated out. Events located above the horizontal bar were considered to be apoptotic cells, and events located to the right of the vertical bar were considered to be sonoporated cells loaded with FD500. The sonoporation was greatly increased in the exposed sample, from 2.5 % to 24.3 % of events, but 43 % of these cells were positive for apoptosis.

The first experiment examined the occurrence of apoptosis as a consequence of 0.4 MPa ultrasound treatment of HL-60 cells using the annexin V assay, and confirmed apoptosis using the Hoechst assay at 6 h post exposure. The HL-60 cell line had a clear fraction of cells which were positive for apoptosis in the negative controls in addition to the normal small fraction of necrotic cells. However the positive controls had substantially higher apoptosis (P<0.001). The two apoptosis tests gave similar results (with independent adjustment of the cutoff for annexin V-Alexa Fluor 350 and Hoechst 33342), which were not significantly different (P>0.1). In addition, the necrosis indicated by the post exposure trypan blue dye exclusion tests was not significantly different for the two sets of exposures: 72 ± 5.5 % and 69.8 ± 3.8 % of cells were non-necrotic in the FD500 and GFP exposures, respectively. The flow cytometry data from the exposures is plotted in Fig. 3. Both the sonoporation and gene expression increases were statistically significant relative to the negative control (P=0.008 for both). The total apoptotic cell counts were not significantly different for the FD500 and the GFP exposures nor for the Hoechst test relative to the annexin V test. The total apoptosis was significantly greater than the result for negative controls (P<0.005) for both of the exposures. In Fig. 3, the bars are stacked to more clearly indicate the fraction of the normal or sonoporated cells which were labeled apoptosis positive. For the FD500, 33 % and 45 % of the total was apoptotic for the Hoeschst and annexin V tests, respectively. For the GFP expression, 36 % and 42 % of the total was apoptotic for the Hoeschst and annexin V tests, respectively. This initial experiment demonstrated that ultrasound exposure of cells in the rotating tube system induced substantial apoptosis in addition to the well known necrosis effect. Furthermore, the apoptosis occurred in a large fraction of the cells showing either molecular loading or gene transfer and expression.

Figure 3.

A bar chart of flow cytometry results for HL-60 exposures at 0.4 MPa with either FD500 cell loading or GFP gene expression evaluated (Sono), and apoptosis (Apop) detected with the aid of either the annexin V or Hoechst 33342 method at 6 h post exposure. Each group of four bars represents mean values with standard error bars for independent samples of the four quadrants of dot plots as in Fig. 2. The two apoptosis detection methods gave similar results, confirming the induction of apoptosis by ultrasonic cavitation.

The second experiment was conducted to follow these bioeffect indicators in time for 2, 6 and 24 h, and to assess the exposure-response trends at 0, 0.2, 0.4 and 0.6 MPa. The flow cytometry data for the 0.4 MPa exposures at different post-exposure times is shown in Fig. 4. For each time point, the numbers of cells scored as positive for FD500 or GFP were significantly greater than the sham result and the apoptosis-positive cells were also significantly increased. For the exposed samples, the fraction of the total of FD500 positive cells was the same for 2 h and 6 h post exposure, but declined after 24 h (P = 0.001), which may reflect the loss of sonoporated cells through apoptosis. This trend was also evident for the GFP expression at 6 h and 24 h (P<0.05). This finding indicates that many cells undergoing plasmid gene transfer in the rotating tube system become injured sufficiently to progress to apoptosis.

Figure 4.

A bar chart of the HL-60 cell counts for FD500 cell loading or GFP expression evaluated (Sono) and apoptosis (Apop) detected by the annexin V method at different times post exposure at 0.4 MPa. Each group of four bars represents mean values with standard error bars for independent samples of the four quadrants of dot plots as in Fig. 2. For each time point, the FD500 or GFP positive cells were statistically significantly larger than the sham value. However, apoptosis became evident for 6 h post exposure incubation, and represented a substantial fraction of the cells having positive FD500 loading or gene expression.

The results for the exposure-response tests are plotted in Fig. 5a and 5b for the FD500 and GFP gene expression tests, respectively. The fraction of the non-necrotic cells showing positive for FD500 or GFP increased with increasing PRPA relative to shams. Numerical results are also given in Table 1, with adjustments for the loss of cells through necrosis and apoptosis. After the adjustments, the actual net yield of sonoporated or gene transfected cells can be seen to peak at the middle 0.4 MPa exposure, which appears to be the optimum exposure for this system. For 0.6 MPa exposure, the 24.4 % FD500 positive and 12.7 % GFP positive cells obtained for the non-necrotic fraction analyzed by flow cytometry represent actual yields of only 5.1% and 2.1%, respectively, for positive and surviving (non-necrotic and non-apoptotic) cells.

Figure 5.

Bar charts of the HL-60 cell counts for the variation of the PRPA, with FD500 cell loading or GFP expression evaluated and apoptosis (Apop) detected by the annexin V method at 6 h post exposure. Each group of four bars represents mean values with standard error bars for independent samples of the four quadrants of dot plots as in Fig. 2. The results for FD500 cell loading are plotted in a, while the GFP gene expression results are plotted in b. Both sonoporation indicators and apoptosis increased with the PRPA of the exposure for these plots, which exclude necrotic cells.

Table 1.

An approximate accounting of the impact of lethal cell injury of HL-60 cells, which leads to immediate necrosis or delayed apoptosis, on the results for FD500 cell loading or gene transfer and expression. The non-necrotic percentage (NNP) is listed from the trypan blue counts soon after treatment. The flow cytometry results for cell loading and gene expression in the non-necrotic cells at 6 hr post exposure (I. e. from Fig. 5) are listed as the flow cytometry positive percentage (FCP). This FCP, reduced by the NNP (I. e., FCP × NNP/100), is listed as the necrosis-corrected percentage. However, many of the FCP cells were apoptotic (see Fig. 5), and a further adjustment for the non-apoptotic percentage is listed as the apoptosis adjusted percentage, which reflects the actual positive percentage in surviving cells from the entire pre-exposure suspension. Statistical P values are for the final apoptosis adjusted result relative to the sham result. After accounting for the lethal cell injury, the sonoporation effects peaked at 0.4 MPa and declined at 0.6 Mpa.

Exposure	MPa	Non-Necrotic %	Flow Cytometry positive %	Necrosis - corrected %	Apoptosis - adjusted %	P value <
FD500	sham	-	2.0 ± 0.9	2.0 ± 0.9	1.5 ± 0.7	-
	0.2	76.4±6.7	4.1 ± 1.4	3.2 ± 1.1	2.3 ± 0.8	0.06 NS
	0.4	58.6 ± 11.4	19.0 ± 5.5	11.1 ± 3.3	7.3 ± 2.9	0.001
	0.6	33.1 ± 5.5	24.4 ± 7.7	8.0 ± 2.0	5.1 ± 2.1	0.001
pGFP	sham	-	0.6 ± 0.4	0.6 ± 0.4	0.3 ± 0.2	-
	0.2	81.1 ± 11.7	1.9 ± 1.0	1.6 ± 0.9	1.1 ± 0.8	0.05
	0.4	48.2 ±10.2	9.6 ± 4.2	5.0 ± 3.0	3.4 ± 2.3	0.001
	0.6	26.6 ± 3.7	12.7 ± 5.1	3.5 ± 1.4	2.1 ± 0.9	0.001

For the third experiment, essentially the same tests were carried out for the CHO cells. The results for the comparison of apoptosis assessment by annexin V-Alexa Fluor 350 and Hoechst 33342 are plotted in Fig. 6. The fraction of cells counted as apoptosis positive was much less for the CHO cells than for the HL-60 cells for all tests. The apoptosis induction by ultrasound treatment (comparing sham and exposed samples) was small but statistically significant for the annexin V test for both the FD500 and GFP exposures (P<0.001), but not for the Hoechst 33342 test (which may have been less sensitive for the 6 h time point tested). Both the fraction of FD500 positive and GFP positive cells declined for 24 h, relative to the 6 h time point, but the apoptosis remained a small part of the total, as shown in Fig. 7. The results for the exposure-response tests are plotted in Fig. 8a and 8b for the FD500 and GFP gene expression tests, respectively. The fractions of FD500 and GFP positive cells increased to substantial levels of 64.7 % and 20.1 % at 0.6 MPa; however, the net yields after adjustment for necrosis and apoptosis as listed in Table 2, were 15.3 % and 4.3 %, respectively. As for the HL-60 cells, the optimum exposure for overall treatment yield was 0.4 MPa.

Figure 6.

A bar chart of flow cytometry results for CHO cell exposures at 0.4 MPa with either FD500 cell loading or GFP gene expression evaluated (Sono), and apoptosis (Apop) detected with the aid of either the annexin V or Hoechst 33342 method at 6 h post exposure. Each group of four bars represents mean values with standard error bars for independent samples of the four quadrants of dot plots as in Fig. 2. The two apoptosis detection methods gave similar results, which were very low for this cell line (compared to HL-60 cells in Fig. 3).

Figure 7.

A bar chart of the CHO cell counts for FD500 cell loading or GFP expression evaluated (Sono) and apoptosis (Apop) detected by the annexin V method at different times post exposure at 0.4 MPa. Each group of four bars represents mean values with standard error bars for independent samples of the four quadrants of dot plots as in Fig. 2. For each time point, the FD500 or GFP positive cells were statistically significantly larger than the sham value. For this cell line the cell loading and gene expression were increased, and the apoptosis induction decreased, relative to the HL-60 cell results (Fig. 4).

Figure 8.

Bar charts of the CHO cell counts for the variation of the PRPA, with FD500 cell loading or GFP expression evaluated and apoptosis (Apop) detected by the annexin V method at 6 h post exposure. Each group of four bars represents mean values with standard error bars for independent samples of the four quadrants of dot plots as in Fig. 2. The results for FD500 cell loading are plotted in a, while the gene expression results are plotted in b. Both sonoporation indicators increased with the PRPA and were larger than for the HL-60 cells (Fig 5), but the apoptosis was greatly reduced for this cell line.

Table 2.

An approximate accounting of the impact of lethal cell injury of CHO cells, which leads to immediate necrosis or delayed apoptosis, on the results for FD500 cell loading or gene transfer and expression. The non-necrotic percentage (NNP) is listed from the trypan blue counts soon after treatment. The flow cytometry results for cell loading and gene expression in the non-necrotic cells at 6 hr post exposure (I. e. from Fig. 8) are listed as the flow cytometry positive percentage (FCP). This FCP, reduced by the NNP (I. e., FCP × NNP/100), is listed as the necrosis-corrected percentage. However, many of the FCP cells were apoptotic (see Fig. 8), and a further adjustment for the non-apoptotic percentage is listed as the apoptosis adjusted percentage, which reflects the actual positive percentage in surviving cells from the entire pre-exposure suspension. Statistical P values are for the final apoptosis adjusted result relative to the sham result. After accounting for the lethal cell injury, the sonoporation effects peaked at 0.4 MPa and declined at 0.6 MPa, as for the HL-60 cells (Table 1).

Exposure	MPa	Non-Necrotic %	Flow Cytometry positive %	Necrosis - corrected %	Apoptosis - adjusted %	P value <
FD500	sham	-	1.3 ± 0.6	1.3 ± 0.5	1.1 ± 0.6	-
	0.2	81.8 ± 15.5	14.3± 17.3	12.4 ± 8.0	11.8 ± 7.4	0.001
	0.4	36.1 ± 5.2	50.6 ± 13.1	18.2 ± 5.4	17.4 ± 5.3	0.001
	0.6	24.0 ± 5.5	64.7 ± 9.8	15.7 ± 5.3	15.3 ± 5.1	0.001
pGFP	sham	-	0.5 ± 0.4	0.5 ± 0.4	0.4 ± 0.3	-
	0.2	89.2 ± 11.6	3.4 ± 4.1	2.9 ± 2.9	2.8 ± 2.9	0.01
	0.4	33.7 ± 8.3	21.4 ± 5.5	7.3 ± 3.0	7.2 ± 2.9	0.001
	0.6	21.0 ± 5.6	20.1 ± 4.3	4.4 ± 1.9	4.3 ± 1.8	0.001

Discussion

When a cell suspension is exposed to ultrasound with cavitation activity promoted by tube rotation, many cells have severe membrane damage leading to immediate necrosis. However, a fraction of effected cells due to milder interaction with cavitation have membrane sonoporation and survive. This phenomenon affords the possibility of transferring large molecules into these cells, including transfer of DNA thus allowing expression of the foreign gene inside surviving cells. The sonoporation also represents a form of injury, since undesirable molecules could leak into the cell and desirable molecules could also leak out. Recently, induction of apoptosis has also been reported to be induced in cells exposed to cavitation, but this cell death pathway is not accounted for by simple membrane dye exclusion tests for necrosis (until very late in the process). The goal of this study was to assess the possible role of apoptosis in sonoporation and gene expression treatments. Three experiments were conducted: the first searched for apoptosis in HL-60 cells by two methods, the second expanded this testing to different times after exposure and different exposure PRPAs, and the third repeated the first two experiments with CHO cells.

The first experiment examined sonoporation with a FD500 molecules or with GFP gene transfer and expression with simultaneous assessment of apoptosis using the annexin V method for living cells (I. e. excluding necrotic cells by flow cytometry gating). Sonoporation was significantly increased for exposed relative to controls 6 hr after exposure at 0.4 MPa, see Fig. 3. However, a large fraction of these cells were identified as apoptotic for both simple cell loading and gene expression. The induction of apoptosis was confirmed using the independent method of Hoeschst 33342 staining. The second experiment involved only the annexin V assay for different time points and PRPAs. The fraction of cells which were positive for fluorescent dextran or GFP declined at 24 h relative to 6 hr post exposure incubation, see Fig. 4. The GFP expression might have been expected to increase due to a continued production and accumulation of GFP in the cells. The observation that many of the GFP positive cells were also apoptotic (about 36% at 6 h) indicated that the induction of apoptosis produced a progressive reduction in gene expression and reduced apparent treatment efficacy with time. The cell loading and gene expression results increased with increasing PRPA, see Fig. 5a and Fig. 5b respectively. However, when the necrosis and apoptosis of the cells was taken into account, as in Table 1, the overall results declined for 0.6 MPa relative to 0.4 MPa exposure. An additional observation from this data is that the fraction of cells positive for GFP was less than the fraction positive for FD500 loading. This probably reflects an increased level of sonoporation injury needed for effective gene transfer and a potential reduction of gene expression which would likely result from this injury.

The third experiment involved essentially the same tests for the CHO cells as for the HL-60 cells in the first two experiments. The results were similar or somewhat increased for cell necrosis, FD500 cell loading and gene transfer and expression, see Figs. 6, 7, 8a and 8b. For example, the percentage of non-necrotic and non-apoptotic cells expressing GFP was up to 3.4 % for HL-60 cells (Table 1), compared to 7.2 % in CHO cells (Table 2) (P<0.02)). The induction of apoptosis was greatly reduced for the CHO cells relative to the HL-60 cells, which may help to explain the improved transfection results.

The purpose of using these two cell lines was to assess the potential influence of cell type on the results. Leukemic cells, such as the HL-60 cell line, have apoptosis as a normal part of the cell function (e. g. leukocytes raised in response to inflammation undergo apoptosis when no longer needed). However, the CHO cells, an epithelial-like cell line, have a lower tendency for apoptosis. This difference in the cell lines was evident in the results for the negative and positive control samples for apoptosis listed in Table 3. The chemical induction of apoptosis was statistically significant for both cell lines (P<0.005) but clearly insubstantial for the CHO cells (even though the dosage of camptothecin was increased 10 fold for these cells). An additional observation, which was indicative of the differences in these two cell lines, was that the fraction of cells expressing detectable GFP was very different in the positive controls for transfection. The results are listed in Table 3, which show a much lower transfection efficiency much less for the HL-60 cells. This relative resistance to gene transfection was also evident in the ultrasound gene transfer results (see Figs. 5b and 8b). However, the relatively low (up to 3.4 % for non-necrotic and non-apoptotic cells, see Table 1) ultrasonic gene transfer for HL-60 cells actually was much larger than the clearly identifiable gene transfer for the positive controls (0.5%, see Table 3). This suggests that the ultrasonic method for transfection performed much better than the reagent method for these cells (although we did not pursue extensive testing and optimization with the commercial reagent, which might have improved the results).

The process of apoptosis induction is uncertain for mechanical injury from ultrasonic cavitation. Apparently, even relatively low energy exposure can induce apoptosis in the HL-60 cell line (Lagneaux et al. 2002). In other research on ultrasound induced apoptosis, a non-specific influx of intracellular calcium was detected (Honda et al. 2004). In addition, a role for reactive oxygen species was indicated by a reduction of apoptosis, which was accomplished by addition of an anti-oxidant to the medium (Honda et al. 2004). Ultrasonic cavitation is not the only gene transfer method to lead to apoptosis; other methods, such as use of antibody particles, have been shown to induce apoptosis in lymphocytes (Ebert et al. 1997).

Sonoporation of HL-60 cells by ultrasonic cavitation in the rotating tube system led to a substantial cell loading effect and moderate gene transfection efficiency. However, the sonoporated cells apparently suffered injury, which led to apoptosis in addition to the initially observed cell killing. The relative effectiveness of ultrasound gene transfer and the role of apoptosis induction was dependent on the cell line, with the CHO cells less susceptible to apoptosis and more amenable to gene transfer (by either the ultrasonic or chemical methods). Ultrasound mediated gene transfer is a promising approach for gene therapy in a wide range of exposure methods, cells and tissues (Miller, 2006). This study indicates that the possible induction of subtle cellular injury leading to apoptosis should be taken into account in ultrasound gene therapy research in addition to the well known necrosis effect.