2.1.

Gene Construction

For this study, we used a genetically encoded FRET sensor CD3 (CFP-DEVD-DsRed2). The details of construction protocols were described previously,⁷ in which the enhanced mutant ECFP was referred to using the more concise name CFP throughout, and DsRed was generated from pDsRed2 (Clontech).

2.2.

Cell Culture, Transfection, and Screening

A human adenoid cystic carcinoma (ACC-M) cell line was obtained from the China Center for Type Culture Collections (Wuhan, China).⁸ The TK-GFP vector was kindly provided by Professor Ariane Söling (Germany).⁹ ACC-M cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, Sigma-Aldrich) supplemented with penicillin G sodium $(100\mathrm{units}∕\mathrm{mL})$ , streptomycin sulfate $(100\mu \mathrm{g}∕\mathrm{mL})$ and 10% fetal bovine serum at $37°\mathrm{C}$ in a humidified atmosphere of 5% ${\mathrm{CO}}_2∕95\%$ air. ACC-M cells were transfected with the plasmid TK-GFP, which expresses the herpes simplex virus thymidine kinase gene (HSV- $tk$ ) and EGFP (enhanced mutant green fluorescence protein) in the same bicistronic message, by FuGENE6 Transfection Reagent (Roche, Switzerland). Stable TK-GFP-expressing ACC-M cell clones (ACC-M-TK-GFP) were screened with $800\mu \mathrm{g}∕\mathrm{mL}$ G418 (Promega Corp., Madison, Wisconsin). ACC-M-TK-GFP cells and ACC-M cells were transfected with CD3 using the same methods. Stable CD3-expressing ACC-M (ACC-M-CD3) cell clones were screened with $800\mu \mathrm{g}∕\mathrm{mL}$ G418.

2.3.

Imaging system for FRET Measurement and Multicolor Fluorescent Imaging

FRET measurements were performed using an FV1000 laser confocal scanning microscope (Olympus, Japan) with a Plan Apo $60\times$ oil immersion objective having 1.42 numerical aperture (NA). The fluorescence images of CD3-expressing ACC-M cells were detected by excitation at $458\mathrm{nm}$ (Ar laser), and the emitted fluorescence bands were separated by a grating and detected by photomultiplier tubes (PMT) in a cyan fluorescent protein (CFP) channel (465 to $485\mathrm{nm}$ ) and FRET channel (580 to $650\mathrm{nm}$ ). For reflecting FRET efficiency of CD3 in living cells, the ratio of FRET $\left(R_{\mathrm{FRET}}\right)$ was calculated as follows:^{10, 11}

For fluorescent imaging of multicolor labeled cells, the fluorescence images of ACC-M-TK-GFP-CD3 cells were obtained using laser confocal scanning microscopy, in which the CFP channel used excitation at $458\mathrm{nm}$ and emission at 465 to $485\mathrm{nm}$ , the DsRed channel used excitation at $514\mathrm{nm}$ and emission at 580 to $650\mathrm{nm}$ , and the GFP channel used excitation at $488\mathrm{nm}$ and emission at 500 to $550\mathrm{nm}$ . The dead cells induced by GCV were identified by morphological changes in transmission images, such as membrane blebbing and cell shrinkage.

3.1.

Real-Time Imaging of Caspase-3 Activity During HSV- $tk$ /GCV–Induced Apoptosis in Living ACC-M Cells

The genetically encoded probe CD3 was transfected into ACC-M-TK-GFP cells, giving TK-GFP and CD3 co-expressed ACC-M cells, named ACC-M-TK-GFP-CD3 cells. The fluorescence images of ACC-M-TK-GFP-CD3 cells showed CD3 localized in the cytosol and TK-GFP localized in the nucleus (Fig. 1 ). To monitor in real time GCV-induced apoptosis in HSV- $t{k}^{+}$ cells, TK-GFP and CD3 co-expressed ACC-M cells were treated with $1.5\mathrm{mg}∕\mathrm{mL}$ GCV, and then images of the ACC-M-TK-GFP-CD3 cells were obtained at about 15-min intervals. A time series of $R_{\mathrm{FRET}}$ images, transmission images, and multichannel fluorescence images of ACC-M-TK-GFP-CD3 cells is shown in Fig. 2a . The FRET ratio of CD3 in ACC-M-TK-GFP-CD3 cells obviously decreased as the caspase-3 activity triggered the cleavage of the CD3 probe.^{7, 12} In the plot of $R_{\mathrm{FRET}}$ versus time [Fig. 2b] and the plot of diameter of cells versus time [Fig. 2c], it is apparent that caspase-3 was activated before cell morphological changes began. Furthermore, once the activation of caspase-3 had been initiated in the cytosol, the activation process was completed within several minutes.⁷ Using ACC-M-CD3 cells expressing only CD3, and without TK-GFP, as a negative control, the FRET ratio of CD3 in ACC-M-CD3 cells was stable and the cells still survived $8\mathrm{h}$ and $27\min$ after treatment with $1.5\mathrm{mg}∕\mathrm{mL}$ GCV. It was especially notable that mitosis of ACC-M-CD3 cells was not affected by GCV (Fig. 3 , arrow). This indicates that GCV did not induce the activation of caspase-3 in ACC-M-CD3 cells without TK expression. The same experiments were repeated at least three times.

Fig. 1

Fluorescence and transmission images of TK-GFP and CD3 co-expressed ACC-M cells. The fluorescent images of ACC-M-TK-GFP-CD3 cells were detected using a laser confocal scanning microscope in (a) the CFP channel with excitation at $458\mathrm{nm}$ and emission at 465 to $485\mathrm{nm}$ , (b) the DsRed channel with excitation at $514\mathrm{nm}$ and emission at 580 to $650\mathrm{nm}$ , and (c) the GFP channel with excitation at $488\mathrm{nm}$ and emission at 500 to $550\mathrm{nm}$ , respectively. A transmission image of ACC-M-TK-GFP-CD3 cells is also shown (d). The scale bar is $20\mathrm{\mu }\mathrm{m}$ .

Fig. 2

Single-cell imaging FRET changes of CD3 in response to caspase-3 activity during GCV-induced apoptosis in ACC-M-TK-GFP-CD3 cells. (a) A time series of $R_{\mathrm{FRET}}$ images of the ACC-M-TK-GFP-CD3 cells treated with $1.5\mathrm{mg}∕\mathrm{mL}$ GCV are shown in the top. The color bar represents the FRET ratio of the CD3 probe. The transmission image (first column) and multichannel fluorescent images (second column: DsRed channel; third column: CFP channel; fourth column: GFP channel) of the ACC-M-TK-GFP-CD3 cells at $6\mathrm{h}$ , $23\mathrm{mins}$ and $9\mathrm{h}$ , $21\mathrm{mins}$ after GCV treatment are shown in the middle row and bottom row, respectively. (b) The plot of $R_{\mathrm{FRET}}$ versus time. (c) The plot of the diameter of cells versus time. The diameter of cells is calculated using the dimension of the longest axis from the transmission images.

Fig. 3

Single-cell imaging FRET ratio of CD3 in living ACC-M-CD3 cells treated with GCV. Bright field images and a time series of $R_{\mathrm{FRET}}$ images of ACC-M-CD3 cells treated with $1.5\mathrm{mg}∕\mathrm{mL}$ GCV showed that the FRET ratio of the CD3 probe was stable and the cells still survived. The mitosis of ACC-M-CD3 cells was not affect by GCV (arrow). The scale bar is $20\mathrm{\mu }\mathrm{m}$ .

3.2.

Real-Time Imaging of the Bystander Effect of the HSV- $tk$ /GCV System on Living ACC-M Cells

To confirm the relationship of caspase-3 activation and the bystander effect of the HSV- $tk$ /GCV system in living cells, we mixed ACC-M-TK-GFP cells and ACC-M-CD3 cells in a 1:3 ratio and seeded them onto a 35-mm “coverslip-bottom culture dish.” After the cells were cultured for $24\mathrm{h}$ , $1.5\mathrm{mg}∕\mathrm{mL}$ GCV was added to induce apoptosis. Mesnil ^{4, 13} reported that the bystander effect involved a transfer of phosphorylated ganciclovir molecules from HSV- $t{k}^{+}$ cells to HSV- $t{k}^{-}$ cells through gap junctions. Therefore, we chose to image the ACC-M-CD3 cells growing adjacent to ACC-M-TK-GFP cells. The FRET images of the ACC-M-CD3 cells adjacent to ACC-M-TK-GFP cells showed that the FRET ratio of the CD3 probe remained stable even as the cells died [Fig. 4a ]. The same experiment was repeated at least three times. In Fig. 4b, the plots of FRET ratio $(I_{\mathrm{FRET}}∕I_{\mathrm{CFP}})$ versus time were calculated from four individual ACC-M-CD3 cells in each repeat experiment. Error bars indicate SD, and the values of $I_{\mathrm{FRET}}∕I_{\mathrm{CFP}}$ did not normalize. Three approximately parallel lines showed that the FRET ratio of CD3 in ACC-M-CD3 cells was stable. This indicated that the bystander effect of the HSV- $tk$ /GCV system can kill the adjacent HSV- $t{k}^{-}$ cells via a caspase-3-independent pathway.

Fig. 4

Single-cell imaging FRET ratio of CD3 in living ACC-M-CD3 cells mixed with ACC-M-TK-GFP cells after GCV treatment. (a) Bright field images and a time series of $R_{\mathrm{FRET}}$ images of ACC-M-CD3 cells adjacent with ACC-M-TK-GFP cells treated with $1.5\mathrm{mg}∕\mathrm{mL}$ GCV showed that the FRET ratio of the CD3 probe in ACC-M-CD3 cells remained stable until the cells’ death. The scale bar is $20\mathrm{\mu }\mathrm{m}$ . (b) The plots of the FRET ratio $(I_{\mathrm{FRET}}∕I_{\mathrm{CFP}})$ versus time were calculated from four individual ACC-M-CD3 cells in three repeated experiments.