1.

Introduction

Fast and reliable cell viability assays are highly desired for any experiments with cells in vitro. For example, cell viability evaluation is important when physical methods based on temporary increase in membrane permeability, such as sonoporation, are used to deliver exogenous molecules to the cells. Cell sonoporation can be induced by exposing cells to high intensity ultrasound (US) in the presence of microbubbles (MB). This leads to facilitated delivery of drug, gene, or other bioactive molecules across sonoporated cell membrane.¹ However, sonoporation can also induce the death of the treated cells.^2–4 Therefore, it is crucial to evaluate cell viability when performing sonotransfer experiments. A few membrane integrity-based cell viability assays are compatible with sonoporation experiments; these use propidium iodide,^1,5,6 ethidium bromide (EB),^7,8 or trypan blue (TB)^7–9 fluorescent dyes. It should be noted that these fluorescent dyes have to be added to the cell suspension 10 to 20 min after the sonoporation treatment for estimation of cell viability.

Cell viability assays based on cell metabolic activity evaluation, e.g., MTT or XTT assays,^10,11 can also be employed for cell viability evaluation after sonoporation. However, as previously indicated, metabolic activity-based assays may lead to imprecise evaluation of cell viability.¹² Another conventional cell viability test is clonogenic assay, which is accurate but time and labor consuming.^13,14

It should be noted that none of the presented viability assays enable simultaneous evaluation of cell viability and efficiency of exogenous molecule transfer. Here, we present a simple method for the evaluation of sonoporation efficiency and cell viability in a single experiment that utilizes commonly used fluorescent dyes: EB and TB. We also show that interaction of both fluorescent dyes used in this method is at least partially related to Förster resonance energy transfer (FRET). EB emission energy is absorbed by TB through radiationless energy transfer in a protein-rich environment, such as the cytoplasm of the cell. Understanding the mechanisms of the method can help to optimize the conditions to get more reliable results not only after cell sonoporation but also after other physical cell treatment such as cell electroporation. At this point, we show that the assay is compatible not only with fluorescence microscopy but also with flow cytometry (FC) that eventually reduce assay cost and time for assay performance.

2.

Methods

3.

Results

The aim of this study was to develop a simple, fast, and cheap method that would allow distinguishing between unaffected, reversibly sonoporated (sonotransfered with EB), and irreversibly sonoporated (dead) cells.

First, we examined the coincidence between the EB fluorescence spectrum and the TB absorption spectrum [Fig. 1(a)]. Presented overlapping spectra indicate the possibility of TB to quench EB fluorescence via one of two mechanisms: (i) radiative energy transfer when EB fluorescence is absorbed by TB and (ii) radiationless energy transfer from EB to TB (FRET). We investigated EB, the absorption and emission of both fluorophores, as well as quenching by TB in the presence and in the absence of FBS. FBS contains proteins, thus imitating protein-rich environment, similar to cell cytoplasm. In FBS environment, broad emission spectrum of EB centered at 620 nm [Fig. 1(a), dashed line]. TB absorption spectrum in FBS had maximum at 595 nm [Fig. 1(a), dashed line], whereas in water environment the maximum of TB absorption spectrum was shifted to 583 nm (data not shown). After mixing both dyes in FBS, EB emission was effectively quenched; it diminished from around 25,000 a.u. (no TB present) to around 4000 a.u. [Fig. 1(b), dotted line].

Fig. 1

(a) Spectral overlap between EB emission and TB absorption (filled gray area). Normalized spectra of EB absorption (470 nm peak, solid line), TB absorption (595 nm peak, solid line), EB emission (620 nm peak, dashed line), and TB emission (710 nm peak, dotted line) in FBS. (b) Emission spectra of EB (dashed line), TB (dotted line), and EB in the presence of TB (solid line) in FBS. Insert: absorption difference of TB between FBS and water.

To understand the influence of the observed TB absorption spectrum shift in FBS environment on the quenching of EB, additional data analysis was performed. Figure 1(b) (insert) shows the increase of differential TB absorption in 500 to 700 nm spectral region when TB was diluted in FBS. Notably, EB emission maximum is positioned between these values. Therefore, the slight increase in EB emission when in FBS environment (25,000 a.u.) as compared to pure water environment (18,000 a.u.) (data not shown) was in part compensated with the increase of TB absorption. Moreover, the significant absorption increase of TB diluted in FBS precisely overlaps with the EB peak in FC filter transmission zone (545 to 625 nm), leading to the decrease of donor fluorophore fluorescence.

These findings were utilized to measure EB sonotransfer and cell membrane integrity. Obtained results allowed discrimination among unaffected cells, EB positive cells (reversible sonoporation), and dead cells (irreversible sonoporation). Indeed, experiments with CHO cell culture showed the capability to discriminate among all three cell groups (see Fig. 2). Images were acquired using fluorescence microscope and processed with the open source imaging program, ImageJ. The bright-field illumination images and the fluorescence images were captured separately and merged together. In Fig. 2(a), red fluorescent cells depict both reversibly sonoporated and dead cell populations, since EB enters both of the cell groups. To discriminate among these groups, TB (which only enters dead cells) was added 15 min after US exposure [Fig. 2(b)]. Cell fluorescence images show that EB emission was quenched by TB only in some of the cells, which allowed the assumption that the cells that were still fluorescent after addition of TB did not have compromised membrane, thus denying TB entry. Moreover, when observed under bright-field illumination, TB intensively stained the dead cells. Therefore, the analysis of merged fluorescence and bright-field illumination images allowed clear separation of the populations of the dead and the viable cells. We tested this new method for evaluation of sonotransfer and cell viability after cell sonoporation experiments, using US conditions described above (1-MHz central frequency, 500-kPa PNP, 10% duty cycle with pulse repetition frequency of 1000 Hz, and total exposure duration 2 s). Cell exposure to US under these conditions resulted in $9.6\pm 1\%$ EB sonotransfer and $17.3\pm 0.2\%$ of dead cells. The rest of the cells were unaffected. In comparison, $6.4\pm 1.7\%$ dead and $1.4\pm 1.4\%$ EB positive cells were observed in control.

Fig. 2

Microscopy images after sonoporation, (a) in the presence of EB ($600\mu \mathrm{M}$) alone or (b) EB ($600\mu \mathrm{M}$) $+\mathrm{TB}$ ($229\mu \mathrm{M}$ or 0.02% w/v). The images were acquired under bright-field illumination or under fluorescence excitation ($560\pm 40\mathrm{nm}$) with emission filter ($630\pm 60\mathrm{nm}$), then merged together after background subtraction using open source program Image J. Cells were insonicated at 1 MHz, 500 kPa PNP acoustic pressure, 10% duty cycle, and 1 kHz repetition frequency for 2 s in the presence of Sonovue MBs.

The same experiment was performed to test the compatibility of the new sonotransfer efficiency and cell viability evaluation method with FC. Figure 3 shows FC results in control group [Fig. 3(a)] and after US application [Fig. 3(b)]. In the control group, only one cell population peak (C1) with low autofluorescence intensity was detected [Fig. 3(a)]. Therefore, C1 region corresponded to viable and unaffected cell population. When cells were treated with US, three populations were distinguishable [Fig. 3(b)]. The low fluorescent cell population in C1 region denoted unaffected cells with no EB inside. The population with the highest fluorescence (marked as C3 region) showed the cells with EB fluorescence that was not quenched by TB, thus indicating that no TB entered the cells. The last cell population, which is more fluorescent than C1, but less fluorescent than C3, was marked within C2 region. This indicated that C2 population had accumulated EB, but its fluorescence was quenched by the presence of TB inside the cells. Therefore, C2 population represented dead cells. Both the sonotransfer and the decrease in cell viability shown in Fig. 3 go in agreement with our previous study.³

Fig. 3

FC results in the presence of EB ($600\mu \mathrm{M}$) and TB (0.02% w/v) at (a) control conditions and (b) after sonoporation. 488 nm excitation laser and u emission filter transmitting from 545 to 625 nm were used in FC. Cells were insonicated at 1 MHz, 500 kPa PNP acoustic pressure, 10% duty cycle, and 1 kHz repetition frequency for 2 s in the presence of Sonovue MBs. C1 represents viable nonaffected cells, C2 represents irreversibly sonoporated (dead) cells, and C3 represents reversibly sonoporated cells.

Fluorescence lifetime measurements were employed to investigate the energy transfer pathways between EB and TB. As previously indicated, the experiments in vitro were performed in the absence or in the presence of FBS. The estimation shows that without FBS EB relaxes monoexponentially with the time constant $\tau =1.75\pm 0.02\mathrm{ns}$ (data not shown). The time dependence of EB fluorescence intensity decay measurements is shown in Fig. 4. The bimodal behavior of EB fluorescence decay was observed in the presence of FBS, with the time constants ${\tau }_{1\mathrm{d}}=1.79\pm 0.1\mathrm{ns}$ (short component) and ${\tau }_{2\mathrm{d}}=5.2\pm 0.14\mathrm{ns}$ (long component), Fig. 4(a). A different time dependence of EB relaxation was detected in the presence of FBS and TB. The two-exponent fitting gives slightly different time constants of ${\tau }_{1\mathrm{da}}=1.55\pm 0.08\mathrm{ns}$ (short component) and ${\tau }_{2\mathrm{da}}=4.67\pm 0.21\mathrm{ns}$ (long component) [Fig. 4(b)].

Fig. 4

Kinetic measurements of EB fluorescence in FBS, (a) in the presence and (b) in the absence of TB. EB and FBS final concentration was kept at 6:1 molar ratio. TB:FBS molar ratio was 2.29:1.

The main difference between EtBr fluorescence lifetime in $+\mathrm{TB}$ and $-\mathrm{TB}$ solutions is the statistically significant reduction of both short ${\tau }_1$ and long ${\tau }_2$ components. The mean amplitude weighted lifetime $⟨\tau ⟩=\sum a_i{\tau }_i$ of EtBr was calculated both in the presence ($⟨{\tau }_{\mathrm{da}}⟩=2.29\mathrm{ns}$) and absence ($⟨{\tau }_{\mathrm{d}}⟩=2.61\mathrm{ns}$) of the acceptor (TB). FRET efficiency ($E$) was calculated by $E=1-⟨{\tau }_{\mathrm{da}}⟩/⟨{\tau }_{\mathrm{d}}⟩=12\%$. These findings allow the conclusion that the loss of EB fluorescence in FBS solution in the presence of TB is mainly governed by FRET.

FRET is expected in the regions where EtBr and TB are colocalized on the protein.

4.

Discussion

In this work, we present an FRET-based method to evaluate efficiency of reversible and irreversible sonoporation in a single experiment. To this end, we use two common fluorescent dyes: EB and TB, which are already used for cell staining in sonoporation setting.^7,8 TB is also used for animal cell^9,17 and plant cell staining.¹⁸ The presented method is based on the EB fluorescence quenching by TB. EB is a DNA-intercalating agent, commonly used as a fluorescent dye. After binding to DNA, the fluorescence of this dye is intensified 10-fold.¹⁹ It has also been shown that EB binds to proteins.²⁰ EB has been already shown to be quenched by a number of compounds, including porphyrin cations and quaternary ammonium compounds,²¹ antitumor drug amsacrine,²² natural compound emodin,²³ or by newly synthetized dibipyridyl ruthenium(II) complex of p-tert-butyltetrathiacalix4arene,²⁴ and bromodeoxyuridine.^25,26 On the other hand, it has also been shown that TB fluorophore could be used as a quenching agent. With its maximal absorption spectrum at pure water being 590 nm,²⁷ TB has been used to quench cell autofluorescence in retinal pigment epithelia cells,²⁸ alveolar macrophages from smokers,²⁹ human mesenteric blood vessels,³⁰ and extracellular fluorescence.³¹

Our experiments indicate that EB emission can be efficiently quenched by TB absorption in a cellular environment. After sonoporation, a donor fluorophore (in our case EB) has been able to localize in the close proximity to TB, leading to the energy transfer to TB through nonradiative dipole–dipole coupling.³² The increase of EB emission spectrum peak in the FBS environment could be explained by the fact that EB is able to interact with proteins and DNA.²⁰ TB is also able to interact with proteins, which could explain the shift in absorption peak.³³ We propose that FRET between EB and TB might be enabled by protein interaction with both of the fluorescent molecules, thus explaining the occurrence of the phenomenon in protein-rich (e.g., intracellular) but not aqueous environment.

EB emission and TB absorption spectra in protein environment strongly coincide with the transmittance characteristics of the standard FC filter. Therefore, the use of both fluorophores can be easily adapted to FC measurements for the detection of sonotransfer efficiency and cell viability. Consequently, this allows discrimination of three distinct cell populations: (i) nonaffected cells, (ii) reversibly sonoporated cells, and (iii) irreversibly sonoporated (dead) cells.

The method can be also adapted for fluorescence microscopy. Indeed, TB both stains the dead cells under visible light and quenches EB fluorescence. For better discrimination of distinct cell populations, microscope images have to be processed with imaging processing software, such as open source ImageJ. As we show in this study, image processing allows the discrimination among nonaffected cells, reversibly sonoporated cells, and irreversibly sonoporated (dead) cells. Taking all of this into account, our approach represents a simple, reliable, and cheap method to simultaneously evaluate exogenous molecule delivery and cell viability.

The method we present is partially based on a FRET mechanism as proven by fluorescence lifetime measurements. The alternative energy transfer pathway would require photon energy to be released by the donor molecule and reabsorbed by the acceptor, either TB or other intracellular chromophores capable of absorption in donor fluorescence wavelength. Therefore, a proximity-based FRET mechanism makes the combination of EB and TB appropriate for desired quenching and allows the separation among three distinct cell groups. Moreover, EB–TB dye pair could be used as an FRET-based method for the general research in vitro, including multicolor FC applications with tandem fluorophores acting as FRET pairs.

The existence of EB and TB as a specific donor acceptor pair in protein-rich environment (as compared to aqueous environment) clearly identifies the excitation energy loss pathway. Fluorescence quenching effects such as FRET are especially important when the model system (BSA-EtBr-TB) environment changes to the cellular environment. The disadvantage of FRET estimation obtained from the dye–protein interaction in cuvette was that only the average of FRET efficiency over the volume of the sample has been obtained. Usually not all EtBr and TB molecules bind to the protein, and therefore, the calculated FRET efficiency is lowered. Consequently, in cellular environment, the result can be shifted toward higher FRET efficiency due to efficient/preferential EtBr and TB binding to cell DNA and plasma proteins. The investigation of EtBr quenching mechanism in cells would require further investigation including combination of fluorescence lifetime imaging and Förster resonance energy transfer (FLIM-FRET) experimentation.