We describe a rapid and quantitative flow cytometric method for determining the apoptotic or anti-apoptotic potential of a gene in various cell types. A plasmid carrying green fluorescent protein (GFP) is co-transfected with an expression vector encoding the gene of interest. Subsequently cells are stained with propidium iodide and, utilising flow cytometry, transfected, GFP-expressing single cells are detected and apoptotic cells in this population are identified by their DNA content of <2 N. The method detects apoptosis as reliably as established methods using in situ nick-end labelling but is faster, easier and less expensive.
Apoptosis or programmed cell death (PCD) is a genetically controlled cellular suicide mechanism for selectively eliminating unwanted cells (for reviews see 1–4 ). PCD is an obligatory process in a variety of biological processes including embryonic and neuronal development, immune system regulation, organogenesis, tissue homeostasis and in the prevention of malignancies such as tumor growth and viral infection. Apoptosis is characterised by plasma membrane blebbing, cell shrinkage, nuclear condensation, endonucleolytic cleavage of genomic DNA into internucleosomal length fragments and the formation of apoptotic bodies.
Current methods for studying apoptosis include the assessment of morphological changes at the cellular level by light-, electron- or time-lapse microscopy in combination with vital fluorescent dyes (reviewed in 5 , 6 ), the use of Annexin V to monitor the loss of membrane phospholipid asymmetry during apoptosis ( 7 ) or assays to detect DNA fragmentation by agarose gel electrophoresis ( 8–10 ), by in situ nick-end labelling (TUNEL) ( 11 ), or by using an enzyme-linked immunoabsorbent assay ( 12 ). Most of these methods, however, are cumbersome, inappropriate or, in the case of the TUNEL method, expensive for studying the effect of putative cell death genes in transient transfection assays.
We have developed a method suitable for the flow cytometric determination of apoptosis in a large number of samples by propidium iodide (PI) in a population of cells transiently co-transfected with a plasmid encoding enhanced green fluorescent protein (eGFP) ( 13 ) and a plasmid carrying a gene of interest. This assay combines the FACS-optimised GFP mutant (eGFP) ( 13 ) as a marker to detect transiently transfected cells and the reduced fluorescence of the DNA-binding dye PI in the apoptotic subpopulation as a marker for apoptosis ( 14–17 ). The reduced fluorescence of PI in apoptosing cells results in the appearance of a characteristic sub-2 N fluorescence peak with respect to the G 0 /G 1 cell cycle region.
In order to validate this method, we initially used the established murine tumor cell lines (βTC and βHC) which originate from β cell tumors ( 18 ) of transgenic mice in which the insulin gene regulatory region (Rip) targets the expression of the simian virus 40 large T-antigen (Tag) to the β cells of the pancreatic islets (Rip1Tag2) ( 19 ).
Approximately 80 000 cells were seeded in a 3 cm diameter well of a six-well dish and grown in DMEM supplemented with 10% FCS (v/v), 2 mM glutamine, 100 IU penicillin and 100 µg/ml streptomycin to 70% confluency. Cells were transfected with 1 µg of a plasmid encoding enhanced GFP (pEGFP-C1; Clontech) together with 1 µg of either a control plasmid (pMEX), or a pCMV plasmid carrying either the pro-apoptotic adenovirus E1A or the anti-apoptotic adenovirus E1B-19K gene ( 20 , 21 ) using 10 µl LipofectAMINE Reagent (GIBCO-BRL) according to the manufacturer's recommendations. Following transfection, the cells were allowed to recover in complete medium for 16 h and were then either left untreated or an apoptotic stimulus (800 ng/ml staurosporine; Sigma) ( 22 , 23 ) was added for an additional 16 h. At 32 h after transfection, floating cells were collected and combined with trypsinised, adherent cells, washed twice with 4 ml PBS and fixed [2% paraformaldehyde (PFA), 100 mM NaCl, 300 mM sucrose, 3 mM MgCl 2 , 1 mM EGTA, 10 mM PIPES pH 6.8] at room temperature for 30 min, washed twice with 4 ml PBS and post-fixed in ice-cold 70% EtOH for 10–14 h.
Following fixation, cells were washed twice with 4 ml PBS and divided to allow comparison of our method to the commonly used modified TUNEL method using fluorescent Cy5-CTP. One half of the sample was treated with RNase A (50 µg/ml) in PBS for 30 min, washed twice with 4 ml PBS and stained with PI (50 µg/ml) in PBS for 30 min before FACS analysis. The other half of the sample was incubated in 50 µl of terminal deoxynucleotidyltransferase (TdT) reaction mixture (Boehringer Mannheim; 200 mM potassium cacodylate, 25 mM Tris-HCl pH 6.6, 0.25 mg/ml bovine serum albumin, 1 mM CoCl 2 , 0.25 nmol of FluoroLink Cy5AP3-dCTP [Amersham], 12.5 U TdT) for 1 h at 37°C. The sample was then washed twice with 4 ml PBS, treated with RNase A (50 µg/ml) in PBS for 30 min, washed twice with 4 ml HBS (Note: HBS is used from this step onwards since DAPI tends to cause micro-precipitates in PBS), stained with DAPI (10 µg/ml; Boehringer Mannheim) in HBS for 20 min and analysed on a Becton Dickinson FACS Vantage machine. Flow cytometric analysis of the PI-stained cells was carried out on a Becton Dickinson FACScan equipped with a doublet discrimination module.
As shown in Figure 1A , our method reliably detects the expected apoptotic or anti-apoptotic effect of the adenovirus E1A or E1B-19K gene products, respectively ( 20 , 21 ). The addition of staurosporine following transfection further exacerbates the observed effects. Similar results are obtained using fluorescent Cy5-CTP and the TdT reaction to measure apoptosis. Although the sensitivity of the GFP/PI method is below that of the Cy5-CTP/TdT method ( Fig. 1A ), the normalised amount of apoptosis, expressed as a ratio of the apoptosis maxima obtained with the E1A gene product within each detection method, is nearly identical ( Fig. 1B ).
To demonstrate the general usefulness of our assay, the GFP/PI method was tested on a non-transformed rat fibroblast cell line (Rat1A) which is less sensitive to apoptotic stimuli. Rat1A cells were transiently transfected using either LipofectAMINE as described for the β tumor cells above, or polyethylenimine-(PEI 2000)-DNA-adenovirus complexes as described ( 24 ). A comparison of the different transfection and apoptotic detection methods using Rat1A cells is presented in Table 1 . As observed with the β tumor cell lines, the GFP/PI method is a reliable apoptotic assay for transiently transfected Rat1A cells ( Table 1 ) as well as for a human lung carcinoma cell line A549 (ATCC CCL-185) ( 24 ).
The PFA/EtOH fixation procedure used in our protocol is essential. Omission of the mild PFA fixation step leads to the rapid leakage of eGFP from the cell and impairs detection of transiently transfected cells. The EtOH fixation/permeabilisation step allows diffusion of low molecular weight DNA products from apoptotic cells and is required to detect the characteristic sub-2 N DNA peak upon PI staining ( 6 , 17 ).
In conclusion, we present a rapid, efficient and reproducible method for determining the apoptotic potential of a given gene of interest in a variety of cell lines. Our method is applicable to and compatible with a variety of cell types and transfection methods. It combines the ease of using GFP as a marker for transfection and sub-2 N PI staining as a simple, inexpensive assay for apoptosis. Furthermore, this assay can be carried out on a relatively simple FACScan analyser and does not require elaborate FACS sorter technology or expertise.
The authors wish to thank Johannes Hoffman for providing the TdT/Cy5-CTP protocol for FACS analyses, Eileen White for pCMV-E1A and pCMV-E1B-19K and Thomas Bader and Kanaga Sabapathy for critical reading of the manuscript. This work was supported in part by the Austrian Industrial Promotion Fund.