Cytometry 13:31-38 (1992)
8 1992 Wiley-Liss, Inc.
Assessment of Radiation-Induced DNA Strand Breaks and Their Repair in Unlabeled Cells Maja Item Affentranger and Werner Burkart Paul Schemer Institute, Radiation Hygiene, CH-5232 Villigen PSI, Switzerland Received for publication May 10, 1991; accepted July 30, 1991
Radiation induced damage, i.e., the induction of DNA strand breaks, was studied on the level of single, unlabeled cells. DNA strand breaks were determined by direct partial alkaline unwinding in intact cell nuclei followed by staining with acridine orange, a development of a proposal f i s t described by B. Rydberg (Int J Radiat Biol46:521-527, 1984). The ratio of green fluorescence (doublestranded DNA) to red fluorescence (single-stranded DNA) in single cells was taken as a measure of DNA strand breaks. CHO-K1 and M3-1 cells irradiated with X-rays show a dose dependent induction of DNA strand breaks. Incuba-
The quantitiative assessment of DNA strand breaks induced by ionizing radiation (8) and of their subsequent repair represents an important field in radiobiological investigation. Well established biochemical techniques are available to quantify DNA breaks in large cell populations (1,13,3). In many cases, the possibility to study the behavior of individual cells after irradiation might be of importance, for example in microdosimetry and radiotherapy, or in proliferating cells. There are methods available to determine DNA strand breaks in single cells, like using the dye acridine orange in microfluorimetry (16) or microelectrophoretic studies (15). However, these assays are very time-consuming and severely limited in the number of cells which can be processed. The flow cytometric technique, first described by Rydberg (17) and further developed in our laboratory, allows the analysis of a statistically relevant number of cells and yields the distribution of DNA strand breaks in cell populations. Dose-effect curves can be determined for these populations. The method has the potential to correlate parameters like cell cycle stages with primary DNA damage and repair kinetics.
tion at 37°C after irradiation leads to repair of breaks. A repair halflife of about 10-11 min can be determined. Cell cycle specific differences in the induction of DNA strand breaks or repair behavior are not detectable at the resolution achieved so far. This new method offers two major advantages: the resolution of DNA damage and repair on the level of single cells and no need for labeling, thereby allowing for DNA damage and repair to be assessed in biopsy material from tumor patients. Key terms: DNA repair, acridine orange, X-ray
MATERIALS AND METHODS Cells and Growth Conditions Chinese hamster ovary cells (CHO-K1; No. CCL-61, American Type Culture Collection, Rockville, MA), Chinese hamster bone marrow cells (M3-1; kindly supplied by Dr. M. R. Raju, Los Alamos, NM), and freshly isolated human peripheral lymphocytes were used. CHO-K1 cells were cultivated in HAMS F-10 (Flow Laboratories) and M3-1 cells in MEM-a (Gibcol-both supplemented with 10% fetal calf serum (Seromed), 25 Eiml penicillin, and 25 pg/ml streptomycin. Cells were cultivated as monolayer in 75 cm2 culture flasks and kept in a n incubator (Cytoperm, Hereus) perfused with a water-saturated 95% air-5% C 0 2 mixture. Peripheral human lymphocytes were separated from heparinized blood samples (30 year old donor) in a density-gradient (Lymphoprepm, Nycomed AS, Norway) and suspended in RPMI 1640 medium for experimental use. Irradiation and Repair Cells were irradiated on ice in culture medium a t a dose rate of 1.3 Gyimin (Philips X-ray machine, model
32
AFFENTRANGER AND BURKART
2.25%ULGT-agarose
Table 1 Influence of RNAse Treatment on the Ratio of GreenlRed Fluorescence Intensity of CHO-Kl Cells
cell suspension
Untreated cells RNAse treated cells 1 ml alkaline solution for 45min, room temp.
Angle (") between abscissa and cells 14 14
MG 323, 240 kV, 10 mA, filtered with 1 mm A h ; dose rate was checked with a Farmer dosemeter 2570). Irradiated cells were incubated at growth conditions to allow repair of DNA damages.
-for another 10rnin at 65'C
9 analysis by flow cytornetry
FIG1. Preparation of cells for flow cytometric analysis of singleand double-stranded DNA. Cells embedded in a n agarose layer were treated with a n alkaline solution to initiate DNA unwinding. The effect is stopped by removing the alkaline solution and washing with PBS buffer. To allow flow cytometic analysis, cells are kept in solution by melting and mixing the cell-agar layer with PBS, followed by a staining procedure with acridine orange.
Single Cell Assay Induction of DNA strand breaks and their repair were determined by alkaline unwinding of DNA in intact cell nuclei combined with acridine orange staining, based on an idea first described by Rydberg (17) (see Fig. 1). Fifty microliters of 2.25% ultra low gelling temperature agarose (SeaPrepa 15/45, FMC; melted in PBS and held at 37°C) was quickly transferred into a flat glass tube and mixed with 25 ~1 ice-coldcell suspension (5 x lo6 cellsiml). The bottom surfaces of the tubes were previously precoated with thin layers of 0.1% agarose (Sigma, LGT). The layers were always freshly prepared, spreading 10 pl of 0.1% agarose (melted in wa-
Green fluorescence FIG.2. Scattergrams showing distribution of CHO-K1 cells irradiated with different doses (X-ray) and stained with acridine orange after alkaline unwinding.
33
DNA STRAND BREAKS AND THEIR REPAIR
FIG.3. Three-dimensional views of the distribution of CHO-K1 cells irradiated with different doses (X-ray). Same histograms as Figure 2 (intensity of green (x) vs. red (y) fluorescence; z-axis gives the number of cells).
Flow Cytometry
["I 50 I Angle
0
5
10
15
20
25
30
35
Dose [Gy] FIG.4. Dose response curves given as angle between the greeniged ratio of untreated and irradiated CHO-K1 cell populations (three experiments).
ter) over the bottom surface, kept on ice for 10 min, and dried at room temperature for another 10 min. The tubes with the cell-agarose mixture were placed on ice for gelling. One milliliter of alkaline solution (0.05 N NaOH, 0.95 N NaCl) was added on top of the agarose layer and kept in the dark a t room temperature. After 45 min, the alkaline solution was replaced by 2 ml PBS for 20 min. PBS was replaced a second time and the tubes were now placed in a 40°C waterbath. After 10 min the tubes were transferred into a 65°C waterbath for another 10 min and slowly inverted several times to mix the melted agarose with the buffer. Tubes were stored a t 4°C ready to be stained before flow cytometric measurements.
Flow cytometric analyses were carried out with a PARTEC PAS I1 Cytofluorograph. Treated cells, which were held at 4°C in a agaroseiPBS suspension (see above), were centrifuged for 5 min a t 1,300 rpm and resuspended in 0.1 ml ice-cold PBS.The samples were mixed on ice with 0.2 ml of solution A (0.1%TritonX-100, 0.08 N HC1, 0.15 N NaC1; pH 1.3). After 45 s, cells were stained by adding 0.6 ml of solution B (10 pgiml acridine orange in lop3M EDTA, 0.15 N NaC1, 0.125 M Na,PO,, 0.04 M citric acid; pH 6.0) and analysed within 10 min after staining. Cells passing the object field (objective Nikon, 40 times) were illuminated with a 100 W high pressure mercury light source with the excitation filter 450-490 nm. The fluorescence light was detected by two bialkali-photomultipliers. Green fluorescence was measured in a wavelength interval from 590 to 515 nm (TK 590, OG 515), and red fluorescence above 630 nm (TK 590, RG 630). The data are shown as dot- and contour plots, three dimensional graphs, and one parameter histograms.
RESULTS Quantification of radiation induced DNA strand breaks in mammalian cells is done indirectly by flow cytometric analysis (for details see Methods and Fig. 1). Controlled alkaline treatment of irradiated cells, where every breakage site serves as a starting point for DNA unwinding, results in a damage dependent amount of single- and double-stranded regions in the cellular DNA. The fluorescent dye acridine orange intercalates into double-stranded DNA, emitting green fluorescence with maximum emission at 530 nm. The
34
AFFENTRANGER AND BURKART
Green fluorescence FIG.5. Scattergrams showing distribution of CHO-K1 cells irradiated with 15 Gy (X-ray) and then incubated for different time-periods at 37°C to allow DNA strand-break repair. The relation of doublestranded (green fluorescence) to single-stranded DNA (red fluorescence) is shown.
FIG.6. Three-dimensional views of the histograms from Figure 5 (intensity of green fluorescence; z-axis gives the number of cells).
Angle
["I
~-
(XI
vs. red (y)
-
30
20 1 10
O L 0
10
20
30
40
50
Repair Incubation [min]
60
70
FIG. 7. Decrease of the angle between the greedred ratio of untreated and irradlated CHO-Kl cell populatlons (15 Gy) wlth length of repair incubation.
35
DNA STRAND BREAKS AND THEIR REPAIR
Control
Control 15Gy 0 min repair
*GY 15Gy 5min repair
3GY
f3 F
15Gy 75min repair
0, 0
6GY
E !
w
15 Gy 10min repair
9GY
15 Gy
l7min repair
15Gy
15Gy
30min repair
15Gy
30 Gy
Green fluorescence
6Omin repair Green fluorescence
Fm. 8. One parameter histograms showing number of CHO-K1 cells vs. green fluorescence. a: Shift of GO/Gl-peak vs. smaller channel numbers as a response from different doses (X-rays). b: Irradiated
cell populations (15Gy) demonstrate a shift of the GO/Gl-peak to higher channel numbers with time-incubation at 37°C for DNA strand break repair.
mechanism of acridine orange binding to singlestranded DNA and RNA still remains unclear. Electrostatic and dye-base stacking interactions are involved between the nucleic acids and the dye, resulting in a maximum emission at 640 nm (5,10,11). The emission ratio of the two characteristic wavelength regions of
acridine orange, when bound to single- or doublestranded DNA, is taken t o determine induced DNA strand breakages. In this system, the RNA present in the cell does not contribute to the orange signal, because the mild alkaline treatment used for DNA-unwinding is sufficient to
36
AFFENTRANGERANDBURKART
Angle ["I
50 I
l0
0 0
5
10
L 15
20
1 25
30
35
Dose [Gy] Angle
40 1
["I
-1
bi
30k 20
I
I\
0
10
20
30
40
50
60
70
Repair incubation [mi n] FIG.9. Radiation induced (X-ray) DNA strand breaks and their repair behavior in M3-1 cells. a: Dose response curve given as angle between the greedred ratio of untreated and irradiated M3-1 cell populations. b Time-dependent decrease of the angle between the greedred ratio of untreated and irradiated M3-1 cells.
break down cellular RNA. Cells incubated with RNAse and untreated cells show no difference in the greedred fluorescence ratio (see Table 1). CHO-K1 cells irradiated with various X-ray doses show a dose dependent increase in red as well as a decrease in green fluorescence intensity, depending on the applied doses (Figs. 2, 3). In the dot plot presentation (Fig. 2 ) , the total cell population distributed in the different cell cycle stages and cell debris forms a stripe that moves uniformly toward higher red fluorescence values depending on the radiation induced DNA strand breaks. A quantitative evaluation of this shift is achieved by setting a line along the mean fluorescence ratios (Figs. 2, 3). The angle between untreated and irradiated cells shows a dose dependent increase (Fig. 4). The repair behavior of cells treated with 15 Gy and subsequently incubated at 37°C to allow DNA strand break repair results in an increase in the intensity of green fluorescence and a concomitant decrease of the red fluorescence signal, depending on the length of repair incubation (Figs. 5 , 6). After 60 min the fluorescence signal ratio reaches the level of the control values. The quantitative evaluation of the graphs using the angle between untreated and irradiated cells with
repair incubation shows a repair halflife of about 1011 min (Fig. 7). Considering the green fluorescence intensity only, a uniform movement of the GOiG1-peak to smaller channel numbers is observed when cell populations are irradiated with different doses (Fig. 8). The contrary, a shift of the GOiG1-peak to higher channel values, occurs with time-dependent repair incubation of the cells. Such one-parameter signals can be helpful in the delineation of the cell cycle stages. The same experiments were carried out with M3-1 cells-Chinese hamster bone marrow cells. An evaluation of the results as angle between untreated and treated cell populations is demonstrated in Figure 9. A dose-dependent increase of the cellular damages follows (Fig. 9a). Irradiated M3-1 cells which were allowed t o repair the induced DNA strand breaks show a time dependent decrease of the damages (Fig. 9b). Again, a repair halflife of about 11min is determined. Both for the induction of primary damage and for repair kinetics, no significant differences between CHO-K1 and M3-1 were seen. Isolated peripheral human lymphocytes were irradiated with 3 and 6 Gy (X-ray) in another experiment (Fig. 10). A dose-dependent shift in the ratio of the fluorescence signals is again seen. As a non-proliferating cell population, cells of the S- and G2iM-phase are absent. Therefore, a determination of the angle between untreated and treated cells is more difficult.
DISCUSSION An optimization of the flow cytometric determination of DNA strand breaks in single cells showed that several experimental details are extremely important and must be adhered to properly. The concentration of the cells embedded in agar has to be at or below a concentration of lo7 cells per ml. At higher cell concentrations, clumping between cells, debris, and free DNA-threads is observed after alkaline treatment when the agar is melted. It is no longer possible to produce suspension of single cells. Close attention is also necessary when acridine orange from different origins and lots is used. It is recommended to compare only experiments stained with the same product in order to minimize dye artifacts. In addition to the concentration of the f luorochrome, the salt concentrations are also in need of careful attention. Intercalation and electrostatic binding show a different sensitivity to ionic strength. Electrostatic binding to the phosphates of the single-stranded DNA is especially sensitive to ionic strength. In our setting the staining solution is composed of 2.28 x lop5 M acridine orange (final concentration 1.53 x 10 M acridine orange) in a phosphate-citrate buffer containing lop3 M EDTA and 0.15 N NaCl at pH 6.0 (for details see Methods) to obtain dye binding to singleand double-stranded DNA with similar efficiency (5). Cells stained with acridine orange were microscopically examined to give an impression of the state of the
'
DNA STRAND BREAKS AND THEIR REPAIR
37
Green fluorescence FIG.10. Contour plots of human peripherial lymphocytes irradiated with 3 and 6 Gy of X-rays, respectively.
prepared cells. Human lymphocytes as compared against CHO-K1 and M3-1 cells show a higher cell loss of about 20%. To better distinguish between cell-derived signals and those from debris and cell clumps, an additional parameter, such as the cell size, should be included in the system. Variation of greedred fluorescence intensity ratios in an uncontrolled way can be avoided by standardizing the measurements and setting for example the non irradiated cells at a constant angle. Unstained cell samples with partially unwound DNA can be stored a t 4°C without showing alterations in the ratio of green (double-) to red (single-stranded DNA) fluorescence intensity. During the examined time-period of 7 d, DNA renaturation was not detected. The quantification of radiation-induced DNA strand breaks measured as a fluorescence ratio of stained single- and double-stranded DNA shows good reproduction between independent experiments (Figs. 4,7). The transformation of the dot plots into angles between untreated and treated cells leads to very similar dose response curves for CHO-K1 agd M3-1 (Figs. 4,9). From the time-course of repair, identical repair half lives of about 11 min were determined. These values are in good agreement with results given by other authors (2,4,6,12).The breaks repaired by such kinetics belong to the fast repaired DNA lesions and are mainly singlestrand breaks (7,121. At present, the sensitivity of this single cell method allows the detection of X-ray induced damage in the range of 1 Gy (Figs. 4,9). This means that fluorescence shifts are detectable that result from about 1,000 single strand breaks (14) and 40 double strand breaks per cell (7). This translates to about 1break per lo6 base pairs. By our measurements it is not possible to distinguish between single- and double-strand breaks.
One possibility to increase the sensitivity is the selection of photomultipliers of higher sensitivity in the red region. The upper dose limit was determined to be about 50 Gy for the standard protocol (data not shown). With shorter unwinding time or with a different alkaline unwinding solution, it is possible to work even with higher doses. It is theoretically possible to determine DNA-content and to distinguish between cells in the GOIGl-, S-, and G2iM-phasesof the mitotic cell cycle from the length of the vector (Figs. 3, 6 ) . Cell cycle specific differences in radiation-induced damage or repair behavior cannot be distinguished at the resolution achieved so far. With the resolution obtained a t present in routine operation, first applications of the single cell method include clinical work. Especially with tumor biopsy material, where in-vivo labeling is not possible, such investigation can readily be carried out and provide insights on DNA damage and repair. Dosimetric studies are also conceivable. Another application is the investigation of cellular effects from clastogenic agents like bleomycin (9).
LITERATURE CITED AhnstrBm G, Erixon K: Fhdiation-induced strand breal-age in DNA from mammalian cells; strand separation in alkaline solution. Int J Radiat Biol 23:285-289, 1973, Ahnstrom G, Edvardsson KA: Radiation-induced single-strand breaks in DNA determined by rate of alkaline strand separation and hydroxylapatite chromatography: an alternative to velocity sedimentation. Int J Radiat Biol 26493-497, 1974. Blocher D: DNA double strand breaks in Ehrlich ascites tumour cells at low doses of X-rays. I. Determination of induced breaks by centrifugation a t reduced speed. Int J h d i a t Biol 42:317-328, 1982. Bryant PE, Blocher D: Measurement of the kinetics ofDNA double strand break repair in Ehrlich ascites tumour cells using the unwinding method. Int J Radiat Biol 38:335-347, 1980.
38
AFFENTKANGER AND BURKART
5. Darzynkiewicz Z, Kapuscinski J Acridine orange: A versatile probe of nucleic acids and other cell constituents. In: Flow Cytmnetry and Sorting, Melamed MR, Lindmo T, Mendelsohn ML (eds). J. Wiley-Liss, Inc., New York, 1990, pp 291-314. 6 . Dikomey E, Franzke J: DNA denaturation kinetics in CHO cells exposed t o different X-ray doses and after different repair intervals using the alkaline unwinding technique. Radiat Environ Biophys 27:29-37, 1988. 7. Elkind MM: DNA repair and cell repair: are they related? Int J Radiat Oncol Biol Phys 5:1089-1094, 1979. 8. Frankenberg-Schwager M: Review of repair kinetics for DNA damage induced in eukaryotic cells in vitro by ionizing radiation. Radiother Oncol 14:307-320, 1989. 9. Item Affentranger M, Burkart W: Differences in the distribution of DNA strand breaks in single cells induced by X-rays or bleomycin. In preparation. 10. Kapuscinski J , Darzynkiewicz Z: Interactions of acridine orange with double-stranded nucleic acids. Spectral and afinity studies. J Biomol Struct Dyn 5:127-143, 1987. 11. Kapuscinski J, Darzynkiewicz Z, Melamed M R Interactions of
acridine orange with nucleic acids. Properties of complexes of acridine orange with single-stranded ribonucleic acid. Biochem Pharmacol 323679-3694, 1983. 12. Kelland LR, Steel G G Inhibition of recovery from damage induced by ionizing radiation in mammalian cells. Radiother Oncol 13:285-299, 1988. 13. Kohn KW, Grirnek-Ewig RA: Alkaline elution analysis, a new approach to the study of DNA single-strand interruptions in cells. Cancer Res 33.1849-1853, 1973. 14. Lett JT, Sun C: The production of strand breaks in mammalian DNA by X-rays: At different stages in the cell cycle. Radiat Res 44:771-787, 1970. 15. Oestling 0, Johanson KJ: Microelectrophoretic study of radiation-induced DNA damages in individual mammalian cells. Biochem Biophys Res Commun 123:291-298, 1984. 16. Rydberg B, Johanson KJ: Estimation of DNA strand breaks in single mammalian cells. In: DNA Repair Mechanisms, Friedberg EC, Fox CF (eds). Academic Press, New York, 1978, pp 465-468. 17. Rydberg B: Detection of DNA strand breaks in single cells using flow cytometry. Int J Radiat Biol 46521-527, 1984.
8 1992 Wiley-Liss, Inc.
Assessment of Radiation-Induced DNA Strand Breaks and Their Repair in Unlabeled Cells Maja Item Affentranger and Werner Burkart Paul Schemer Institute, Radiation Hygiene, CH-5232 Villigen PSI, Switzerland Received for publication May 10, 1991; accepted July 30, 1991
Radiation induced damage, i.e., the induction of DNA strand breaks, was studied on the level of single, unlabeled cells. DNA strand breaks were determined by direct partial alkaline unwinding in intact cell nuclei followed by staining with acridine orange, a development of a proposal f i s t described by B. Rydberg (Int J Radiat Biol46:521-527, 1984). The ratio of green fluorescence (doublestranded DNA) to red fluorescence (single-stranded DNA) in single cells was taken as a measure of DNA strand breaks. CHO-K1 and M3-1 cells irradiated with X-rays show a dose dependent induction of DNA strand breaks. Incuba-
The quantitiative assessment of DNA strand breaks induced by ionizing radiation (8) and of their subsequent repair represents an important field in radiobiological investigation. Well established biochemical techniques are available to quantify DNA breaks in large cell populations (1,13,3). In many cases, the possibility to study the behavior of individual cells after irradiation might be of importance, for example in microdosimetry and radiotherapy, or in proliferating cells. There are methods available to determine DNA strand breaks in single cells, like using the dye acridine orange in microfluorimetry (16) or microelectrophoretic studies (15). However, these assays are very time-consuming and severely limited in the number of cells which can be processed. The flow cytometric technique, first described by Rydberg (17) and further developed in our laboratory, allows the analysis of a statistically relevant number of cells and yields the distribution of DNA strand breaks in cell populations. Dose-effect curves can be determined for these populations. The method has the potential to correlate parameters like cell cycle stages with primary DNA damage and repair kinetics.
tion at 37°C after irradiation leads to repair of breaks. A repair halflife of about 10-11 min can be determined. Cell cycle specific differences in the induction of DNA strand breaks or repair behavior are not detectable at the resolution achieved so far. This new method offers two major advantages: the resolution of DNA damage and repair on the level of single cells and no need for labeling, thereby allowing for DNA damage and repair to be assessed in biopsy material from tumor patients. Key terms: DNA repair, acridine orange, X-ray
MATERIALS AND METHODS Cells and Growth Conditions Chinese hamster ovary cells (CHO-K1; No. CCL-61, American Type Culture Collection, Rockville, MA), Chinese hamster bone marrow cells (M3-1; kindly supplied by Dr. M. R. Raju, Los Alamos, NM), and freshly isolated human peripheral lymphocytes were used. CHO-K1 cells were cultivated in HAMS F-10 (Flow Laboratories) and M3-1 cells in MEM-a (Gibcol-both supplemented with 10% fetal calf serum (Seromed), 25 Eiml penicillin, and 25 pg/ml streptomycin. Cells were cultivated as monolayer in 75 cm2 culture flasks and kept in a n incubator (Cytoperm, Hereus) perfused with a water-saturated 95% air-5% C 0 2 mixture. Peripheral human lymphocytes were separated from heparinized blood samples (30 year old donor) in a density-gradient (Lymphoprepm, Nycomed AS, Norway) and suspended in RPMI 1640 medium for experimental use. Irradiation and Repair Cells were irradiated on ice in culture medium a t a dose rate of 1.3 Gyimin (Philips X-ray machine, model
32
AFFENTRANGER AND BURKART
2.25%ULGT-agarose
Table 1 Influence of RNAse Treatment on the Ratio of GreenlRed Fluorescence Intensity of CHO-Kl Cells
cell suspension
Untreated cells RNAse treated cells 1 ml alkaline solution for 45min, room temp.
Angle (") between abscissa and cells 14 14
MG 323, 240 kV, 10 mA, filtered with 1 mm A h ; dose rate was checked with a Farmer dosemeter 2570). Irradiated cells were incubated at growth conditions to allow repair of DNA damages.
-for another 10rnin at 65'C
9 analysis by flow cytornetry
FIG1. Preparation of cells for flow cytometric analysis of singleand double-stranded DNA. Cells embedded in a n agarose layer were treated with a n alkaline solution to initiate DNA unwinding. The effect is stopped by removing the alkaline solution and washing with PBS buffer. To allow flow cytometic analysis, cells are kept in solution by melting and mixing the cell-agar layer with PBS, followed by a staining procedure with acridine orange.
Single Cell Assay Induction of DNA strand breaks and their repair were determined by alkaline unwinding of DNA in intact cell nuclei combined with acridine orange staining, based on an idea first described by Rydberg (17) (see Fig. 1). Fifty microliters of 2.25% ultra low gelling temperature agarose (SeaPrepa 15/45, FMC; melted in PBS and held at 37°C) was quickly transferred into a flat glass tube and mixed with 25 ~1 ice-coldcell suspension (5 x lo6 cellsiml). The bottom surfaces of the tubes were previously precoated with thin layers of 0.1% agarose (Sigma, LGT). The layers were always freshly prepared, spreading 10 pl of 0.1% agarose (melted in wa-
Green fluorescence FIG.2. Scattergrams showing distribution of CHO-K1 cells irradiated with different doses (X-ray) and stained with acridine orange after alkaline unwinding.
33
DNA STRAND BREAKS AND THEIR REPAIR
FIG.3. Three-dimensional views of the distribution of CHO-K1 cells irradiated with different doses (X-ray). Same histograms as Figure 2 (intensity of green (x) vs. red (y) fluorescence; z-axis gives the number of cells).
Flow Cytometry
["I 50 I Angle
0
5
10
15
20
25
30
35
Dose [Gy] FIG.4. Dose response curves given as angle between the greeniged ratio of untreated and irradiated CHO-K1 cell populations (three experiments).
ter) over the bottom surface, kept on ice for 10 min, and dried at room temperature for another 10 min. The tubes with the cell-agarose mixture were placed on ice for gelling. One milliliter of alkaline solution (0.05 N NaOH, 0.95 N NaCl) was added on top of the agarose layer and kept in the dark a t room temperature. After 45 min, the alkaline solution was replaced by 2 ml PBS for 20 min. PBS was replaced a second time and the tubes were now placed in a 40°C waterbath. After 10 min the tubes were transferred into a 65°C waterbath for another 10 min and slowly inverted several times to mix the melted agarose with the buffer. Tubes were stored a t 4°C ready to be stained before flow cytometric measurements.
Flow cytometric analyses were carried out with a PARTEC PAS I1 Cytofluorograph. Treated cells, which were held at 4°C in a agaroseiPBS suspension (see above), were centrifuged for 5 min a t 1,300 rpm and resuspended in 0.1 ml ice-cold PBS.The samples were mixed on ice with 0.2 ml of solution A (0.1%TritonX-100, 0.08 N HC1, 0.15 N NaC1; pH 1.3). After 45 s, cells were stained by adding 0.6 ml of solution B (10 pgiml acridine orange in lop3M EDTA, 0.15 N NaC1, 0.125 M Na,PO,, 0.04 M citric acid; pH 6.0) and analysed within 10 min after staining. Cells passing the object field (objective Nikon, 40 times) were illuminated with a 100 W high pressure mercury light source with the excitation filter 450-490 nm. The fluorescence light was detected by two bialkali-photomultipliers. Green fluorescence was measured in a wavelength interval from 590 to 515 nm (TK 590, OG 515), and red fluorescence above 630 nm (TK 590, RG 630). The data are shown as dot- and contour plots, three dimensional graphs, and one parameter histograms.
RESULTS Quantification of radiation induced DNA strand breaks in mammalian cells is done indirectly by flow cytometric analysis (for details see Methods and Fig. 1). Controlled alkaline treatment of irradiated cells, where every breakage site serves as a starting point for DNA unwinding, results in a damage dependent amount of single- and double-stranded regions in the cellular DNA. The fluorescent dye acridine orange intercalates into double-stranded DNA, emitting green fluorescence with maximum emission at 530 nm. The
34
AFFENTRANGER AND BURKART
Green fluorescence FIG.5. Scattergrams showing distribution of CHO-K1 cells irradiated with 15 Gy (X-ray) and then incubated for different time-periods at 37°C to allow DNA strand-break repair. The relation of doublestranded (green fluorescence) to single-stranded DNA (red fluorescence) is shown.
FIG.6. Three-dimensional views of the histograms from Figure 5 (intensity of green fluorescence; z-axis gives the number of cells).
Angle
["I
~-
(XI
vs. red (y)
-
30
20 1 10
O L 0
10
20
30
40
50
Repair Incubation [min]
60
70
FIG. 7. Decrease of the angle between the greedred ratio of untreated and irradlated CHO-Kl cell populatlons (15 Gy) wlth length of repair incubation.
35
DNA STRAND BREAKS AND THEIR REPAIR
Control
Control 15Gy 0 min repair
*GY 15Gy 5min repair
3GY
f3 F
15Gy 75min repair
0, 0
6GY
E !
w
15 Gy 10min repair
9GY
15 Gy
l7min repair
15Gy
15Gy
30min repair
15Gy
30 Gy
Green fluorescence
6Omin repair Green fluorescence
Fm. 8. One parameter histograms showing number of CHO-K1 cells vs. green fluorescence. a: Shift of GO/Gl-peak vs. smaller channel numbers as a response from different doses (X-rays). b: Irradiated
cell populations (15Gy) demonstrate a shift of the GO/Gl-peak to higher channel numbers with time-incubation at 37°C for DNA strand break repair.
mechanism of acridine orange binding to singlestranded DNA and RNA still remains unclear. Electrostatic and dye-base stacking interactions are involved between the nucleic acids and the dye, resulting in a maximum emission at 640 nm (5,10,11). The emission ratio of the two characteristic wavelength regions of
acridine orange, when bound to single- or doublestranded DNA, is taken t o determine induced DNA strand breakages. In this system, the RNA present in the cell does not contribute to the orange signal, because the mild alkaline treatment used for DNA-unwinding is sufficient to
36
AFFENTRANGERANDBURKART
Angle ["I
50 I
l0
0 0
5
10
L 15
20
1 25
30
35
Dose [Gy] Angle
40 1
["I
-1
bi
30k 20
I
I\
0
10
20
30
40
50
60
70
Repair incubation [mi n] FIG.9. Radiation induced (X-ray) DNA strand breaks and their repair behavior in M3-1 cells. a: Dose response curve given as angle between the greedred ratio of untreated and irradiated M3-1 cell populations. b Time-dependent decrease of the angle between the greedred ratio of untreated and irradiated M3-1 cells.
break down cellular RNA. Cells incubated with RNAse and untreated cells show no difference in the greedred fluorescence ratio (see Table 1). CHO-K1 cells irradiated with various X-ray doses show a dose dependent increase in red as well as a decrease in green fluorescence intensity, depending on the applied doses (Figs. 2, 3). In the dot plot presentation (Fig. 2 ) , the total cell population distributed in the different cell cycle stages and cell debris forms a stripe that moves uniformly toward higher red fluorescence values depending on the radiation induced DNA strand breaks. A quantitative evaluation of this shift is achieved by setting a line along the mean fluorescence ratios (Figs. 2, 3). The angle between untreated and irradiated cells shows a dose dependent increase (Fig. 4). The repair behavior of cells treated with 15 Gy and subsequently incubated at 37°C to allow DNA strand break repair results in an increase in the intensity of green fluorescence and a concomitant decrease of the red fluorescence signal, depending on the length of repair incubation (Figs. 5 , 6). After 60 min the fluorescence signal ratio reaches the level of the control values. The quantitative evaluation of the graphs using the angle between untreated and irradiated cells with
repair incubation shows a repair halflife of about 1011 min (Fig. 7). Considering the green fluorescence intensity only, a uniform movement of the GOiG1-peak to smaller channel numbers is observed when cell populations are irradiated with different doses (Fig. 8). The contrary, a shift of the GOiG1-peak to higher channel values, occurs with time-dependent repair incubation of the cells. Such one-parameter signals can be helpful in the delineation of the cell cycle stages. The same experiments were carried out with M3-1 cells-Chinese hamster bone marrow cells. An evaluation of the results as angle between untreated and treated cell populations is demonstrated in Figure 9. A dose-dependent increase of the cellular damages follows (Fig. 9a). Irradiated M3-1 cells which were allowed t o repair the induced DNA strand breaks show a time dependent decrease of the damages (Fig. 9b). Again, a repair halflife of about 11min is determined. Both for the induction of primary damage and for repair kinetics, no significant differences between CHO-K1 and M3-1 were seen. Isolated peripheral human lymphocytes were irradiated with 3 and 6 Gy (X-ray) in another experiment (Fig. 10). A dose-dependent shift in the ratio of the fluorescence signals is again seen. As a non-proliferating cell population, cells of the S- and G2iM-phase are absent. Therefore, a determination of the angle between untreated and treated cells is more difficult.
DISCUSSION An optimization of the flow cytometric determination of DNA strand breaks in single cells showed that several experimental details are extremely important and must be adhered to properly. The concentration of the cells embedded in agar has to be at or below a concentration of lo7 cells per ml. At higher cell concentrations, clumping between cells, debris, and free DNA-threads is observed after alkaline treatment when the agar is melted. It is no longer possible to produce suspension of single cells. Close attention is also necessary when acridine orange from different origins and lots is used. It is recommended to compare only experiments stained with the same product in order to minimize dye artifacts. In addition to the concentration of the f luorochrome, the salt concentrations are also in need of careful attention. Intercalation and electrostatic binding show a different sensitivity to ionic strength. Electrostatic binding to the phosphates of the single-stranded DNA is especially sensitive to ionic strength. In our setting the staining solution is composed of 2.28 x lop5 M acridine orange (final concentration 1.53 x 10 M acridine orange) in a phosphate-citrate buffer containing lop3 M EDTA and 0.15 N NaCl at pH 6.0 (for details see Methods) to obtain dye binding to singleand double-stranded DNA with similar efficiency (5). Cells stained with acridine orange were microscopically examined to give an impression of the state of the
'
DNA STRAND BREAKS AND THEIR REPAIR
37
Green fluorescence FIG.10. Contour plots of human peripherial lymphocytes irradiated with 3 and 6 Gy of X-rays, respectively.
prepared cells. Human lymphocytes as compared against CHO-K1 and M3-1 cells show a higher cell loss of about 20%. To better distinguish between cell-derived signals and those from debris and cell clumps, an additional parameter, such as the cell size, should be included in the system. Variation of greedred fluorescence intensity ratios in an uncontrolled way can be avoided by standardizing the measurements and setting for example the non irradiated cells at a constant angle. Unstained cell samples with partially unwound DNA can be stored a t 4°C without showing alterations in the ratio of green (double-) to red (single-stranded DNA) fluorescence intensity. During the examined time-period of 7 d, DNA renaturation was not detected. The quantification of radiation-induced DNA strand breaks measured as a fluorescence ratio of stained single- and double-stranded DNA shows good reproduction between independent experiments (Figs. 4,7). The transformation of the dot plots into angles between untreated and treated cells leads to very similar dose response curves for CHO-K1 agd M3-1 (Figs. 4,9). From the time-course of repair, identical repair half lives of about 11 min were determined. These values are in good agreement with results given by other authors (2,4,6,12).The breaks repaired by such kinetics belong to the fast repaired DNA lesions and are mainly singlestrand breaks (7,121. At present, the sensitivity of this single cell method allows the detection of X-ray induced damage in the range of 1 Gy (Figs. 4,9). This means that fluorescence shifts are detectable that result from about 1,000 single strand breaks (14) and 40 double strand breaks per cell (7). This translates to about 1break per lo6 base pairs. By our measurements it is not possible to distinguish between single- and double-strand breaks.
One possibility to increase the sensitivity is the selection of photomultipliers of higher sensitivity in the red region. The upper dose limit was determined to be about 50 Gy for the standard protocol (data not shown). With shorter unwinding time or with a different alkaline unwinding solution, it is possible to work even with higher doses. It is theoretically possible to determine DNA-content and to distinguish between cells in the GOIGl-, S-, and G2iM-phasesof the mitotic cell cycle from the length of the vector (Figs. 3, 6 ) . Cell cycle specific differences in radiation-induced damage or repair behavior cannot be distinguished at the resolution achieved so far. With the resolution obtained a t present in routine operation, first applications of the single cell method include clinical work. Especially with tumor biopsy material, where in-vivo labeling is not possible, such investigation can readily be carried out and provide insights on DNA damage and repair. Dosimetric studies are also conceivable. Another application is the investigation of cellular effects from clastogenic agents like bleomycin (9).
LITERATURE CITED AhnstrBm G, Erixon K: Fhdiation-induced strand breal-age in DNA from mammalian cells; strand separation in alkaline solution. Int J Radiat Biol 23:285-289, 1973, Ahnstrom G, Edvardsson KA: Radiation-induced single-strand breaks in DNA determined by rate of alkaline strand separation and hydroxylapatite chromatography: an alternative to velocity sedimentation. Int J Radiat Biol 26493-497, 1974. Blocher D: DNA double strand breaks in Ehrlich ascites tumour cells at low doses of X-rays. I. Determination of induced breaks by centrifugation a t reduced speed. Int J h d i a t Biol 42:317-328, 1982. Bryant PE, Blocher D: Measurement of the kinetics ofDNA double strand break repair in Ehrlich ascites tumour cells using the unwinding method. Int J Radiat Biol 38:335-347, 1980.
38
AFFENTKANGER AND BURKART
5. Darzynkiewicz Z, Kapuscinski J Acridine orange: A versatile probe of nucleic acids and other cell constituents. In: Flow Cytmnetry and Sorting, Melamed MR, Lindmo T, Mendelsohn ML (eds). J. Wiley-Liss, Inc., New York, 1990, pp 291-314. 6 . Dikomey E, Franzke J: DNA denaturation kinetics in CHO cells exposed t o different X-ray doses and after different repair intervals using the alkaline unwinding technique. Radiat Environ Biophys 27:29-37, 1988. 7. Elkind MM: DNA repair and cell repair: are they related? Int J Radiat Oncol Biol Phys 5:1089-1094, 1979. 8. Frankenberg-Schwager M: Review of repair kinetics for DNA damage induced in eukaryotic cells in vitro by ionizing radiation. Radiother Oncol 14:307-320, 1989. 9. Item Affentranger M, Burkart W: Differences in the distribution of DNA strand breaks in single cells induced by X-rays or bleomycin. In preparation. 10. Kapuscinski J , Darzynkiewicz Z: Interactions of acridine orange with double-stranded nucleic acids. Spectral and afinity studies. J Biomol Struct Dyn 5:127-143, 1987. 11. Kapuscinski J, Darzynkiewicz Z, Melamed M R Interactions of
acridine orange with nucleic acids. Properties of complexes of acridine orange with single-stranded ribonucleic acid. Biochem Pharmacol 323679-3694, 1983. 12. Kelland LR, Steel G G Inhibition of recovery from damage induced by ionizing radiation in mammalian cells. Radiother Oncol 13:285-299, 1988. 13. Kohn KW, Grirnek-Ewig RA: Alkaline elution analysis, a new approach to the study of DNA single-strand interruptions in cells. Cancer Res 33.1849-1853, 1973. 14. Lett JT, Sun C: The production of strand breaks in mammalian DNA by X-rays: At different stages in the cell cycle. Radiat Res 44:771-787, 1970. 15. Oestling 0, Johanson KJ: Microelectrophoretic study of radiation-induced DNA damages in individual mammalian cells. Biochem Biophys Res Commun 123:291-298, 1984. 16. Rydberg B, Johanson KJ: Estimation of DNA strand breaks in single mammalian cells. In: DNA Repair Mechanisms, Friedberg EC, Fox CF (eds). Academic Press, New York, 1978, pp 465-468. 17. Rydberg B: Detection of DNA strand breaks in single cells using flow cytometry. Int J Radiat Biol 46521-527, 1984.
