Sample collection and analysis using flow cytometry

Samples were collected using a FACScan flow cytometer (Becton Dickinson) and analysed using the Lysis II software package (Becton Dickinson). BrdUrd incorporation was detected using a monoclonal primary antibody to BrdUrd, and was revealed using a fluorescent (FITC) secondary antibody, which has a maximum emission spectrum of 515 <^< 545 nm. The green FITC fluorescence emission signal was directly related to BrdUrd incorporation, and was collected on a log scale. Parallel

incorporation of the DNA stain propidium iodide (PI) (maximum emission 590 <X< 640 nm) allowed the simultaneous analysis of BrdUrd incorporation and DNA content. This type of staining protocol (Wilson et al. 1994; Nagasawa et al. 1994; Higashikubo et al.

1996) allowed the identification of the three main populations of interest: GO/Gl, S and G2/M, whose movement could then be followed using a series of analysis windows generated in the software package (LYSIS II). Debris or clusters of cells were eliminated from analysis by the use of a doublet discrimination module on the PI signal.

The populations of interest used when analysing the cell-cycle data are shown in Figure 4.2. This shows the histogram profile of the total cell population, which contains both the BrdUrd-labelled and unlabelled cells. During analysis, these windows were overlaid onto plots of the individual populations, so the fraction of labelled and unlabelled cells can be calculated for each phase and analysed over time.

Figure 4.3 a-c shows examples of the plots generated during the dose response analysis of individual cell-cycle phase delays. A G2/M delay was observed in all cell lines. Estimation of the length of delay in the exit of cells from G2/M was performed by plotting the entry of BrdUrd-labelled cells, corrected for cell division (see Chapter 3), into GO/Gl. The entry of cells into G l with time followed a sigmoidal trend and was fitted using a non-linear regression procedure (Johns and Joiner, 1991) with a generalised logit equation as described in Chapter 3 (equation 3.3). This equation was fitted to the individual data points for each experiment, and the upper and lower limits were optimally defined for each cell line. The extent of G2/M delay of BrdUrd-labelled cells was determined from the shift in the inflection points of the best fit curves of unirradiated controls compared to and irradiated samples (Figure 4.3 a).

The G2/M delay of G2 phase cells was observed by following the movement of BrdUrd unlabelled G2 cells through this phase with time after radiation treatment. A delay in radiation treated cells was indicated by an increase in the time required for the fraction of cells in this phase to decrease with time, relative to unirradiated controls.

The extent of G l/S transition delay was analysed by plotting the entry of G l cells into early S with time, see Figure 4.3 b. Differences between the half maximum peak height times were used to calculate the extent of delay at G l/S border. A similar analysis was performed for the determination of a delay in the progression of BrdUrd- labelled cells through S-phase, see Figure 4.3c.

F igure 4.1. C ell irradiation system for cell cycle perturbation experim ents. T he Stuart incubator in w hich the cells w ere irradiated w as set at 37°C. C ells w ere transported to and from the X -ray m achine to the long-term incubators using pre-w arm ed trays w ith polystyrene lids to m inim ise heat loss.

o 00 M2 _{M3 M4} o 200 400 600 FL3-Area 800 1000

Figure 4.2. Cell cycle delay analysis o f D N A histogram s. R egions o f interest w ere isolated using com puter generated regions (LY SIS II softw are, B ecton D ickinson). Each population is identified by differences in DNA content (x-axis). G 2/M cells have tw ice the D N A content o f GO/Gl cells, as they are preparing to divide. S-phase cells have a variable DNA content, as D N A synthesis is in various stages o f com pletion. M 1 show s cells w hich are in GO/Gl and this w indow w as used to follow the entry of B rdU rd-labelled cells into this phase after progression through G2/M . T his facilitates the analysis o f a G 2/M delay o f S phase cells at the tim e o f irradiation. M 2 show s cells in early S and facilitates the analysis o f G l/S delay. M3 show s cells in m id-S and facilitates the analysis o f S-phase delay. M 4 is placed around G 2/M cells and allow s the analysis o f the progression o f G2 cells at the tim e o f irradiation through G 2 .

80 A. G2/M delay 5 60 _c ^ 4 0 0 Ü 3 "2 20 CÛ 0 10 20 3 0 4 0 C/D E _c (/) 0 Ü ■D 0 "0 0 Time (h) C/D _{B . G1/S delay} 0 0 C 0 5 10 15 20 2 5 Time (h) 100 _{I I I I I ] I I I I I I I I I I} C. S delay 8 0 - Time (h) O O G y 1 Gy A 2 Gy

Figure 4.3. R adiation-induced cell-cycle delay. A: G2 delay dose response. A lterations in the rate o f entry o f SW 48 cells into G l w ith increasing dose indicate a delay in G2,

and is quantified by the difference betw een curve inflection points. B: G l/S delay dose

response. A delay in progression o f B rdU rd-unlabelled cells through G l/S is estim ated from d ifferences in the tim e taken to reach h a lf m axim um peak height. C : S -phase delay. A delay in the m ovem ent o f B rdU rd-labelled cells in m id-S through S phase is estim ated from the time taken to reach half m inim um peak height.