2.1.

Microscopy Imaging Platform

A Zeiss 200M epifluorescence microscope with a $63\times$ water-immersion objective [numerical aperture $\left(\mathrm{NA}\right)=1.2$ ] and a five-channel filter turret (Carl Zeiss MicroImaging, Inc., Thornwood, New York) is the main observation platform. A motorized platform enables $z$ -stack imaging and deconvolution to achieve high-contrast fluorescence images and also enables time-lapse imaging for monitoring dynamic processes with a time resolution in seconds. Two CCD cameras—AxioCam MRm (Carl Zeiss MicroImaging, Inc., Thornwood, New York) and Imager QE (LaVision, Goettingen, Germany)—are installed on the top and base ports of the microscope, respectively. The micropulser (compact pulse generator) mated with the electrode microchamber (both built by our group) is mounted on the microscope stage in a standard frame insert [Fig. 1b]. Two computers control the microscope, cameras, and micropulser for pulse-synchronized image acquisition. Cells stained with fluorescent dyes are observed and monitored before, during, and after the application of nanoelectropulses.

2.2.

Micropulser and Microchamber

2.3.

CCD Cameras and Software

Two CCD cameras were used for fluorescence imaging. An AxioCam MRm [CCD monochrome, $1300\times 1030$ pixels, 1-ms to 20-s shutter time, UV-near-IR (NIR) spectral range], was controlled by the Zeiss image acquisition software AxioVision (Carl Zeiss MicroImaging, Inc., Thornwood, New York) through a PCI interface by computer I, shown in Fig. 1. The AxioVision program controls the microscope, including multichannel filters, excitation shutter, and AxioCam camera for multichannel, $z$ -stack, time-lapse image acquisition, and a deconvolution module for reconstructing 3-D images. The second camera, Imager QE from LaVision ( $1376\times 1040$ pixels, seven-electron sensitivity, 1 to 1000-ms shutter) is controlled by DaVis software (LaVision, Goettingten, Germany). This camera utilizes a two-stage Peltier system to operate at $-11°\mathrm{C}$ . The detection range is from 200 to 700 nm with highest quantum efficiency (65%) at 550 nm. In this paper, all image sequences (movies) were taken by this camera with time intervals from 15 to 500 ms. Through a PCI interface the camera can be controlled with either computer software or external hardware signals. It can also provide output triggers to other instruments, providing additional flexibility in experimental design. To enable fast data collection and analysis, customized functions for specific intensity and time export were added using macros composed in DaVis Command Language. LabView (National Instruments, Austin, Texas) was used to trigger the micropulser and to modify electric pulse characteristics such as pulse width and repetition rate. In this work, two personal computers were used to control the various instruments with different software and hardware links. The productivity of the system could be substantially improved by integration of all of these functions into a single computing system and instrument network.

3.1.

Cell Lines and Culture Conditions

Human Jurkat T lymphocytes (ATCC TIB-152) were grown in RPMI (Roswell Park Memorial Institute) 1640 (Irvine Scientific) containing 10% heat-inactivated fetal bovine serum (Irvine Scientific), $2\text{‐}\mathrm{mM}$ L-glutamine (Gibco-BRL), $50\text{units}∕\mathrm{mL}$ penicillin (Gibco-BRL), and $50\mu \mathrm{g}∕\mathrm{mL}$ streptomycin.

3.2.

Fluorescent Molecular Probes

The absorption and emission peak wavelengths and filter sets for these fluorescent probes are summarized in Table 1 .

Table 1

Excitation and emission peak wavelength of fluorescence probes and the corresponding filters.

3.3.

Experimental Conditions

Although the main effort for regular fluorescence imaging is to improve the emission intensity of stained cells, the priorities for living-cell observation are cell viability and the photostability of the staining. Strong excitation and high dye concentration are effective methods to increase the fluorescence intensity but they can induce serious cell damage and photobleaching. To balance imaging rate, image quality, photostability, and cell viability, photobleaching of Calcium Green staining was monitored with neutral density filters (ND) in the excitation light path. Figure 2 shows bleaching curves for Calcium Green in Jurkat cells using ND filters from 0 to 1.5. Calcium Green fluorescence decayed by 45% in $15\mathrm{s}$ without any filter (0). Here, stained cells were exposed to the full strength of the light source, a $100\text{‐}\mathrm{W}$ mercury lamp ( $700\mu \mathrm{W}∕{\mathrm{cm}}^2$ at the $63\times$ objective). The rapid bleaching makes it difficult to monitor fluorescence changes during dynamic observations, and the intense illumination induces cell damage, such as membrane blebs. The photobleaching data indicated that 1.0 ND could satisfy our requirements for limited cell damages, emission stability, good image quality, and proper acquisition rate. Data reported in this paper were recorded with 1.0 ND, except when total exposure time for sequential images was less than $5\mathrm{s}$ .

Fig. 2

Photobleaching of Calcium Green fluorescence with different neutral density filters.

4.1.

Calcium Bursts of Jurkat Cells under Nanoelectropulses

Sequential imaging of cells before, during, and after pulse exposure reveals a sharp increase in the fluorescence emission intensity of Calcium Green-loaded cells within seconds of pulse delivery, indicating a rise in the level of intracellular free calcium. The intensification occurs uniformly across the area of the whole cell [Fig. 3a ]. These images demonstrate a rapid intracellular calcium release triggered by nanoelectropulses. A plot of the quantitative intensity changes extracted from 14 cells [Fig. 3b] shows that the fluorescence intensity starts to rise within hundreds of milliseconds, continues to increase for approximately $1\mathrm{s}$ , then remains constant for at least several seconds. We have reported elsewhere¹¹ that pulse-induced Calcium Green fluorescence intensification increases with pulse number, electric field, and repetition rate, and White also described the similar behavior in HL-60 cells.²¹ In our experiments, the imaging rate typically was $2\text{to}10\text{frames}∕\mathrm{s}$ , limited by the data transfer rate and camera exposure time, but localizing the source of calcium ions requires much faster speed. For this purpose, the image acquisition parameters were modified. First, the imaging area was reduced to a single cell or partial cell to reduce the data transfer time. Second, the exposure time was decreased to $5\mathrm{ms}$ by removing the ND filter. However, no obvious localized intensification was found inside cells even by imaging at the shortest interval $\left(15\mathrm{ms}\right)$ .

Fig. 3

Calcium bursts after four $30\text{‐}\mathrm{ns}$ , $2.5\text{‐}\mathrm{MV}∕\mathrm{m}$ pulses: (a) Calcium Green fluorescence images of Jurkat cells at 0, 0.6, 1.2, and $2.5\mathrm{s}$ after shocking, showing strong intensification, and (b) quantitative intensity changes in different cells induced by four pulses. Gray lines (Cell) represent individual cells; dark line (Avg) is the average value.

4.2.

PI Penetration Control

To examine cell membrane integrity after nanoelectropulse treatment, PI influx experiments were conducted with both nanosecond, high-field and millisecond, low-field pulses. As a conventional indicator of cell membrane permeability, PI has been widely used to study the reversible membrane breakdown induced by electrical pulses.^{4, 5, 6, 22} PI is normally excluded from viable cells with intact cell membrane. In the cytoplasm, PI molecules bind with nucleic acid, with an associated 20- to 30-fold increase in fluorescence, providing a convenient and reliable method for checking membrane integrity.²³ No PI fluorescence intensification occurred in cells after four $30\text{‐}\mathrm{ns}$ , $3.5\text{‐}\mathrm{MV}∕\mathrm{m}$ pulses, but a single low-voltage, millisecond pulse ( $5\mathrm{ms}$ , $100\mathrm{kV}∕\mathrm{m}$ ) caused immediate PI entry into cells (within hundreds of milliseconds). Images of PI diffusion after portion (Fig. 4 ) demonstrate that PI molecules cross the cell membrane at the anode pole of the cell, then diffuse gradually throughout the cell. This asymmetric uptake of PI is in agreement with classical observations of electroporation.^{22, 24, 25} These control experiments demonstrate that the cell membrane is not permeant to small dye molecules and that it remains intact after nanoelectropulses, although it can be porated by long electrical pulses.

Fig. 4

PI asymmetric penetration induced by electroporation: $5\mathrm{ms}$ , $100\mathrm{kV}∕\mathrm{m}$ , 1 pulse: (a) image before pulse (0s), the cell in PI solution is dark; (b) image at $5\mathrm{s}$ after pulse exposure, a bright new-moon edge emerges on the bottom of cell, close to the anode electrode. It indicates PI enters cells at the anode first; (c) image at $10\mathrm{s}$ , the bright area expands at the anode and a weak fluorescence appears at the cathode; and (d) image at $5\min$ , the whole cell becomes extremely bright. (e) to (h) Cells before and after four $30\text{‐}\mathrm{ns}$ , $2.5\text{‐}\mathrm{MV}∕\mathrm{m}$ pulses. The arrow in images shows the direction of electric fields.

4.3.

Externalization of Phosphatidylserine (PS)

To study the transient responses of the cell membrane after nanoelectropulse exposure, FM 1-43 was used to track the PS translocation during temporal observations. FM 1-43 is a membrane probe widely utilized for vesicle cycling in living cells, but an indication of the phospholipid scrambling of plasma membrane was also reported.²⁶ On the other hand, as an early indicator of apoptosis, PS externalization is commonly detected^{14, 16} with a fluorescent conjugate of annexin V. However, the annexin V assay requires incubation and washing, and is not readily adapted to real-time observations. In contrast, FM 1-43 can be observed directly without washing because its quantum yield of fluorescence is thousands of times greater in the lipid environment of the cell membrane than it is in aqueous solvent. Translocation of PS to the external leaflet of the membrane lipid bilayer attracts more FM 1-43 from the medium into the membrane, and the fluorescence further increases (Fig. 5 ). Before pulsing [Fig. 5a] the cell membrane displays a uniform weak emission, whereas strong fluorescence appears at the region closer to the anode electrode after nanoelectropulse application [Figs. 5b and 5c], indicating a localized, pulse-driven translocation of PS in this area of the cell membrane. FM 1-43 fluorescence intensification occurs at the anode pole of the cell within milliseconds after pulse exposure, then gradually increases by 10 to 30% over the next few seconds.

Fig. 5

FM 1-43 fluorescence images of Jurkat cells: (a) images at 0, 2, and $4\mathrm{s}$ after pulsing ( $30\mathrm{ns}$ , $2.5\mathrm{MV}∕\mathrm{m}$ , four pulses), intensification occurred at anode areas but not cathodes; (b) amplified images of anode areas; and (c) profile of the left cell in image (a), where an arrow shows the location and direction of the profile. The solid line is the intensity distribution of the cell at $4\mathrm{s}$ after pulses and the dotted line represents that before pulses. There are some bright spots inside cells because actively growing cells cycle stained membrane into intracellular vesicles.

Figure 6 demonstrates the representative intensity changes, which were observed in hundreds of cells, at different membrane areas after pulsing. This immediate pulse-induced fluorescence increase always occurs at the anode pole, never at the cathode end of the cell or at the equatorial regions. Cathodic pole fluorescence in fact often exhibits a pulse-induced fluorescence decrease. However, the polarized fluorescence pattern dissipates after about 1 min as a result of lateral (circumferential) diffusion of PS within the membrane outer leaflet, which corresponds with fluorescent annexin V observations.⁹ This predominantly anodic polarization pattern is consistent with published images of electroporated cells.^{24, 25} Based on both previous reports^{9, 14} of pulse-induced PS externalization visualized with conventional annexin V assay and the association of the phospholipid scrambling with FM 1-43 membrane staining,²⁶ we assumed that the polarized transient intensification in pulsed cells are indicative and associated with PS externalization. The images can be interpreted in two possible ways: (1) high electric fields cause PS translocation directly by reorienting the head group dipole and (2) transient nanopores formed by nanoelectropulses provide a pathway for PS molecules to migrate from the inner membrane leaflet to the external face of the cell.¹⁶ Because FM 1-43 is a doubly charged cation, one might also assume that a pulse-driven electrophoretic migration of the dye into the membrane could produce a fluorescence increase at the anodic pole of the cell. To test this idea, we conducted similar experiments with DiI ( $1,1^{\prime }$ -dioctadecyl- $3,3,3^{\prime },3^{\prime }$ -tetramethylindocarbocyanine perchlorate, Molecule Probes Inc., Eugene, Oregon), a membrane-labeling dye that is also a divalent cation and that has been used in the studies of electrophoresis of cell membrane components.²⁷ No fluorescence intensification of DiI in the membrane was observed even with much higher electric fields and pulse counts than we used for the FM 1-43 experiments. The fluorescence weak decay at cathodes may be due to the PS internalization in high electric fields. Because the population of PS on the outer leaflet is much lower than the inner membrane leaflet, the change is relatively small. The polarized intensification was recorded in more than hundreds of cells in later experiments. Here, we show only a group of representative images. These results provide important information for studies of electroperturbation mechanisms and membrane response models.

Fig. 6

Plot of fluorescence intensity of FM 1-43 at different areas of the cell membrane. The fluorescence kept rising at anodes after pulsing, but not at equatorial and cathodic regions, where the intensity even slightly decayed. C1 and C2 indicate two cells. Anod, Equat, and Cath represent the areas on the cell membrane close to the anode electrode, between anode and ground, and the ground electrode.

4.4.

Nucleus Perturbation

Although evidence supports the hypothesis that nanoelectropulses mainly disrupt intracellular compartments, the effects on intracellular organelles and processes are not fully understood, especially for the nucleus and its contents—only a few reports in this specific area have appeared.^{13, 15, 28, 29} DNA damages and fragmentation induced by electric pulses were found based on electrophoresis results,^{13, 15, 28} where DNA molecules were analyzed after extracted from lysed cells. Chen observed the nuclear response in living cells—a slight pulse-induced fluorescence quenching of acridine-orange (AO)-stained nuclei, which they suggested resulted from the DNA∕AO complex moving out of the pulse-damaged nuclear envelope.²⁹ Understanding the responses of chromatin and nuclei to pulsed high electric fields, especially in living cells, may benefit the potential applications of nanoelectropulse to cellular manipulation and cancer therapy.

In this paper, we also describe observations of nuclear morphological and chromatin∕DNA structural changes after nanoelectropulse treatment of cells stained with the nucleic acid probe Hoechst 33342 (H342). Figure 7 shows Jurkat cells stained with H342, a common dye for double-stranded DNA and widely utilized for nuclear staining of living cells. After multiple nanoelectropulses, nuclear enlargement and fluorescence quenching were observed (Fig. 7B). In the image taken at $10\mathrm{s}$ , the intensity changes were still small, but after $5\min$ , they became significant. These observations are in agreement with the quantitative intensity records from sequential images of pulsed and control cells, seen in Fig. 7B(f).

Fig. 7

Nuclear enlargement and H342 fluorescence quenching after 200 pulses ( $30\mathrm{ns}$ , $3.5\mathrm{MV}∕\mathrm{m}$ ). A: a cell stained with H342, a differential interference contrast (DIC) image of the same cell, and a composite. B: (a), (b), and (c) fluorescence images of a nucleus at $0\mathrm{s}$ , $10\mathrm{s}$ , and $5\min$ after pulsing, respectively, showing quenching of the H342 fluorescence; (d) and (e) demonstrate the postpulse enlargement of the nucleus and the reticular pattern associated with the pulse-induced reduction in fluorescence intensity; and (f) fluorescence intensity changes of H342 in pulsed cells ( $30\mathrm{ns}$ , $3.5\mathrm{MV}∕\mathrm{m}$ , 200 pulses) and control cells.

The pulse-induced decrease in H342 fluorescence may indicate DNA damage or chromatin structural changes that modify the number and type of sites available for specific and nonspecific binding^{30, 31, 32} by H342. We suggest that DNA fragmentation or modifications of chromatin configuration may be caused by either directly pulsed megavolt-per-meter electric fields or as a result of intracellular or specifically intranuclear fluctuations in pH or ionic homeostasis.