K. Berniak and P. Rybak contributed equally to this work.
Relationship between DNA damage response, initiated by camptothecin or oxidative stress, and DNA replication, analyzed by quantitative 3D image analysis
Article first published online: 11 JUL 2013
Copyright © 2013 International Society for Advancement of Cytometry
Cytometry Part A
Volume 83, Issue 10, pages 913–924, October 2013
How to Cite
Berniak, K., Rybak, P., Bernas, T., Zarębski, M., Biela, E., Zhao, H., Darzynkiewicz, Z. and Dobrucki, J. W. (2013), Relationship between DNA damage response, initiated by camptothecin or oxidative stress, and DNA replication, analyzed by quantitative 3D image analysis. Cytometry, 83: 913–924. doi: 10.1002/cyto.a.22327
This article is an expanded version of the communication, which was delivered during 2012 ISAC Congress in Leipzig.
- Issue published online: 20 SEP 2013
- Article first published online: 11 JUL 2013
- Manuscript Accepted: 6 JUN 2013
- Manuscript Revised: 29 MAY 2013
- Manuscript Received: 21 JAN 2013
- NCN. Grant Numbers: UMO-2011/01/B/NZ3/00609, 2012/05/E/ST2/02180
- EU Structural Funds. Grant Number: BMZ No. POIG.02.01.00–12-064/08
- NCI. Grant Number: CA RO1 28 704
- R.A. Welke Cancer Research Foundation
- DNA damage;
- DNA damage response;
- histone H2AX phosphorylation;
- DNA replication;
- oxidative stress;
- hydrogen peroxide;
- cell cycle;
- base excision repair
A method of quantitative analysis of spatial (3D) relationship between discrete nuclear events detected by confocal microscopy is described and applied in analysis of a dependence between sites of DNA damage signaling (γH2AX foci) and DNA replication (EdU incorporation) in cells subjected to treatments with camptothecin (Cpt) or hydrogen peroxide (H2O2). Cpt induces γH2AX foci, likely reporting formation of DNA double-strand breaks (DSBs), almost exclusively at sites of DNA replication. This finding is consistent with the known mechanism of induction of DSBs by DNA topoisomerase I (topo1) inhibitors at the sites of collisions of the moving replication forks with topo1-DNA “cleavable complexes” stabilized by Cpt. Whereas an increased level of H2AX histone phosphorylation is seen in S-phase of cells subjected to H2O2, only a minor proportion of γH2AX foci coincide with DNA replication sites. Thus, the increased level of H2AX phosphorylation induced by H2O2 is not a direct consequence of formation of DNA lesions at the sites of moving DNA replication forks. These data suggest that oxidative stress induced by H2O2 and formation of the primary H2O2-induced lesions (8-oxo-7,8-dihydroguanosine) inhibits replication globally and triggers formation of γH2AX at various distances from replication forks. Quantitative analysis of a frequency of DNA replication sites and γH2AX foci suggests also that stalling of replicating forks by Cpt leads to activation of new DNA replication origins. © 2013 International Society for Advancement of Cytometry
It has been demonstrated that a relationship between DNA damage response (DDR, also defined as DNA damage signaling) and DNA replication can be studied and quantitatively analyzed within a population of adherent cells in vitro, using laser scanning cytometry (LSC). Cell measurements by LSC, combined with multivariate data analysis, made it possible to correlate the drug-induced DNA damage response in individual cells with a specific phase of the cell cycle, inhibition of replication, an increased level of a specific protein, or induction of apoptosis ([1-9]). However, LSC analysis can neither provide information on the relationship between these processes in time and in space within the same nucleus, nor can it reveal whether DNA damage response occurs within or outside the sites of DNA replication (“replication factories”). When DNA damage signaling occurs at the sites of DNA replication, it is likely that the damage is causally related to movement of DNA replication forks. It is also possible to envisage a different scenario, where these two events coincide in time, but not in space, i.e., the DNA damage signaling occurs during S-phase, but is induced away from replication forks because the two events are not causally related. Thus, in order to discern between these two scenarios, in addition to LSC analysis, spatial patterns of DDR and replication foci need to be investigated in 3D by microscopy.
To extend our prior findings linking induction of DDR and DNA replication ([5, 9]), in this study, we used 3D fluorescence confocal microscopy and image analysis to reveal spatial relationship between DDR and DNA replication. The DNA damage signaling inflicted by the DNA topoisomerase I inhibitor camptothecin (Cpt) or hydrogen peroxide (H2O2) was used as a test case to validate and optimize a method of quantitative image analysis dedicated to investigating spatial relationships between DNA replication sites and DDR foci.
It is known that Cpt induces DDR, revealed as phosphorylation of histone H2AX on Ser139 (induction of γH2AX foci) and activation of ATM (Ataxia talangiectasia mutated kinase), exclusively in cells replicating DNA ([8-10]). In the case of oxidative DNA damage inflicted by H2O2, the induction of γH2AX foci was apparent in both, DNA replicating as well as not replicating cells. The level of γH2AX expression, however, was significantly higher in S-phase cells (Fig. 1). On the other hand, it was uncertain to what extent the sites representing DNA damage signaling, namely γH2AX foci, known to mark primarily the presence of DNA double-strand breaks [(DSBs; ([12, 13])], were spatially associated with the DNA replication sites in S-phase cells exposed to H2O2.
In this report, we describe a method of quantitative analysis of a spatial relationship between γH2AX foci and DNA replication sites, the latter detected by labeling newly synthesized (nascent) DNA with EdU (5-ethynyl-2′-deoxyuridine) and “click chemistry” ([7, 14]), in cells exposed to Cpt or H2O2. Our analysis of a spatial association between these foci in 3D images of A549 human lung adenocarcinoma cell nuclei confirms that Cpt induces γH2AX foci almost exclusively in association with DNA replication sites. In contrast, although the frequency of H2O2-induced γH2AX foci is distinctly elevated in S-phase cells, they are located predominantly outside of sites of DNA replication.
MATERIALS AND Methods
In vitro cultures of human lung adenocarcinoma A549 intended for imaging experiments were grown on round coverslips submerged in Petri dishes, as described previously ([6, 15]).
Inducing DNA Damage by Cpt or H2O2 and Labeling Nascent DNA
DNA precursor analog, EdU (10 μM, Click-iT® EdU AlexaFluor® 488 Imaging Kit cat. # C10337; Invitrogen/Molecular Probes) was added to cell cultures 30 min before adding Cpt or H2O2. Cpt or H2O2 was added to culture medium at a final concentration of 0.2 μM or 0.2 mM, respectively. After 30 min of exposure to a drug (i.e., 60 min of exposure to EdU), cells were fixed. Untreated cells were incubated with EdU for 60 min. Detection of EdU was performed according to manufacturer's instructions. In control experiments, we demonstrated that EdU, at a concentration and exposure time we used, induced no detectable increase of the level of phosphorylation of histone H2AX, indicating that no DDR was evoked by the precursor (data not shown).
Detection of γH2AX
Cells were fixed with methanol-free formaldehyde (4%, EMS, Hatfield, PA) and treated with Triton X-100 (0.1%, Sigma, Poznań, Poland). Blocking was done overnight in 3% (w/v) BSA (Sigma, Poznań). Phospho-specific (Ser139) γH2AX mAb (Upstate Biotechnology, Lake Placid, NY) was used, followed by a secondary antibody AlexaFluor® 568 goat anti-mouse IgG (H+L), cat. # A11004 (Invitrogen/Molecular Probes).
Laser Scanning Cytometry
LSC experiments were performed as described previously ([5-7]).
Images of cell nuclei with fluorescently labeled EdU and γH2AX were recorded using Leica LSC SP5 confocal microscope (Leica Microsystems GmbH, Wetzlar, Germany). The following instrumental parameters were used: 63× HCX PL APO CS NA 1.4 oil immersion lens, excitation 488 (Ar) and 561 nm (HeNe), emission detection bands 500–550 nm for AlexaFluor488 (Click-iT EdU), and 600–660 nm for AlexaFluor568 (immunofluorescence, γH2AX), registration in sequential mode, scanning 8000 Hz (resonant scanner), with 8–16 frames averaged.
Image Processing and Analysis
3D images of DNA replication and histone phosphorylation were deconvolved using AutoDeblur software v.X2.2 (Media Cybernetics, Rockville, MD). The area occupied by the nucleus was selected and the histogram of fluorescence intensities was edited to remove noise and background. The images of DNA replication (EdU) sites had negligible background, whereas the immunofluorescence images of γH2AX foci always resulted in a low but detectable level of nonspecific signal, which had to be carefully estimated and removed by histogram editing. The 3D object counter plugin () for Image J was used to determine the position and the number of 3D objects in a stack. The coordinates of the centers of mass (barycenters) of each focus were calculated. This procedure was repeated for both fluorescence channels (replication and histone phosphorylation). A script written in Python was used to measure the distances from a selected γH2AX focus to all EdU sites. The set of all distances from all γH2AX foci to their nearest EdU site were used to construct a histogram of the distances to the nearest-neighbor. This procedure was repeated for each nucleus.
Quantitative Analysis of Distances Between γH2AX Foci and DNA Replication (EdU) Sites
Our previous attempt to quantify the spatial relationship between DNA replication sites and γH2AX foci by means of colocalization analysis, based on 2D images, yielded results that gave only an approximate indication of the existence of spatial dependence between the two events ([6, 8]). In these prior estimates, correlation coefficients depended on the areas (or volumes) of the differentially labeled replication versus phosphorylation regions. However, it is not the area of the overlap of these sites, but the spatial proximity between them in 3D space that has to be used as a criterion to distinguish whether the damage signaling is induced within the replicating or nonreplicating DNA regions. In this study, therefore, for the purpose of analysis of interdependencies between γH2AX foci and replication sites, we measured the distances between the centers of mass (barycenters) of γH2AX foci and the nearest DNA replication regions. Representative images (Fig. 2) and analyses of distances between barycenters of EdU sites and γH2AX foci in untreated cells as well as cells exposed to Cpt or H2O2 (over 350 nuclei) are shown in Figures 3-5; the numerical data are summarized in Tables 1 and 2.
|non. repl.||Early S||Middle S||Late S|
|Total number/cell||Control||107 ±56||175 ±36||79 ±43|
|Cpt||227 ± 143*||276 ± 89*||99 ± 64|
|H2O2||155 ± 66*||285 ±166*||93 ± 45|
|Foci volume (median) [μm3]||Control||0.0365||0.0515||0.0877|
|Repl. to repl. distance (median; max) [μm]||Control||0.9; 3.76||0.95; 2.82||1.22; 5.24|
|Cpt||0.74; 2.85||0.77; 2.92||1.01; 5.9|
|H2O2||0.83; 2.46||0.77; 2.88||1.10; 5.05|
|Percent of all replication foci, accompanied by γH2AX signal||Control||9.87%||6.47%||7.33%|
|Histone PHOSPHORYLATION foci (γH2AX)|
|non. repl.||Early S||Middle S||Late S|
|Total number /cell||Control||11 ± 18||26 ± 20||30 ± 16||25 ± 15|
|Cpt||8 ± 7||94 ± 68*||166 ± 66*||60 ± 27*|
|H2O2||83 ± 31||70 ± 17*||129 ± 33*||112 ± 26*|
|Foci volume (median) [μm3]||Control||0.0505||0.0393||0.0503||0.0449|
|γH2AX to γH2AX distance (median, max) [µm]||Control||1.79; 15.97||1.35; 10.62||1.57; 6.43||1.65; 7.44|
|Cpt||2.33; 14.72||1.03; 6.54||0.99; 4.4||1.39; 5.41|
|H2O2||1.07; 3.14||1.18; 4.34||1.17; 3.78||1.24; 5.16|
|γH2AX to repl. distance (median, max) [μm]||Control||0.86; 3.57||0.87; 3.98||1.28; 4.76|
|Cpt||0.53; 2.91||0.55; 3.43||0.37; 4.51|
|H2O2||0.77; 3.32||0.74; 3.89||1.23; 6.07|
|Percent of all γH2AX foci associated with replication, minus control (number/cell)||Control||35.7% (26)||35.3% (11)||31.3% (8)|
|Cpt||79.3% (54)*||68.4% (93)*||100% (36)*|
|H2O2||49.9% (22)*||50.2% (49)*||33.6% (29)*|
In untreated cells, there was a very low level of histone H2AX phosphorylation in nonreplicating cells (Figs. 1 and 2) (∼11 phosphorylation foci/nucleus; Table 2), and a higher number in EdU-labeled cells (∼25–30/nucleus). This is consistent with the previous data reporting lower level of constitutive DNA damage signaling in G1- compared with S-phase cells ([1, 4, 5]). Because the pattern of DNA replication is distinctly different in early-, mid-, and late S-phase ([8, 18]) (Fig. 2), these stages were analyzed separately.
The histograms of distances in 3D space between the γH2AX foci and the nearest DNA replication regions in Cpt-treated cells demonstrate that a large proportion of γH2AX foci are located close (less than 0.5 μm) to the sites of replicating DNA (Figs. 4A, 4D, and 4G). It should be noted that these histograms embrace Cpt-induced as well as endogenous phosphorylation foci, the latter reporting constitutive DNA damage signaling induced primarily by endogenous oxidants ([1-3]). However, in untreated cells, very few phosphorylation foci were found within 0.5 μm of the nearest replication site (Figs. 3A, 3C, and 3E).
The shape of the histograms representing untreated cells differs significantly between the subsequent S-phase sub-stages. Clearly, the three stages of S-phase pose different problems for analysis of distances between γH2AX and EdU sites. It is noticeable that the number of replication foci (Table 1) and the ranges of the measured distances are different in the three morphologically recognizable sub-stages of S-phase in untreated and Cpt-treated cells. The left side of each histogram (short distances) represents γH2AX foci located within DNA replicating areas (Figs. 4A, 4D, and 4G). Likely, they represent DNA damage (DSBs) inflicted by Cpt at the site of DNA replication forks. The right side of the histogram represents γH2AX foci located at a relatively large distance from the replication sites. The width of the histograms (which is defined by replication-independent endogenous γH2AX foci located far from replication) is larger in mid- and late- (Figs. 4D and 4G) than in early S-phase (Fig. 4A; Table 2).
In early S-phase of Cpt-treated cells, the DNA replication sites are very numerous (227 ± 143/nucleus; Table 1); the median distance between the sites in 3D space is ∼0.74 μm and the median distance between γH2AX foci and the nearest DNA replication region is 0.53 μm. Consequently, all γH2AX foci, even the ones that may be unrelated to replication, are found quite near to some EdU incorporation sites. Mid-S sub-phase is characterized by an even larger number of DNA replication sites (276 ± 89/nucleus). However, they are clustered primarily around nucleoli and under the nuclear envelope. They are separated by a median distance of 0.77 μm. In late S-phase, in Cpt-treated cells, DNA replication is confined to relatively few (99 ± 64/nucleus), well spaced areas (the median EdU-to-EdU sites distance is 1.01 μm); some γH2AX foci are found even as far as 4.51 μm from the nearest EdU-region. In other words, the histograms extending further into long distances for mid and late S-phase reflect the increasing average distances between DNA replicating regions in the course of progression through S-phase. Considering the above, the late S-phase cells are expected to have the lowest number of γH2AX foci unrelated to replication that are, however, in proximity to the replication regions. Consequently, this period appears to be the most suitable for assessing the spatial relationship between DNA damage signaling (histone H2AX phosphorylation) and DNA replication.
Correcting for Intrinsic (endogenous) γH2AX Foci Unrelated to Cpt and H2O2 Treatment
It is important to note that the histograms shown in Figures 4A, 4D, and 4G as well as 5A, 5D, and 5G embrace not only the lesions induced by Cpt or H2O2, but the intrinsic γH2AX foci as well. The intrinsic foci are present in untreated cells [Fig. 2 (top row) and Fig. 3] where, as mentioned, they report constitutive DNA damage signaling induced most likely by endogenous oxidants generated in mitochondria as by-products of aerobic respiration ([1-3, 6, 19]). Thus, the specific analysis of DNA damage signaling induced by Cpt or H2O2 requires separating Cpt- and H2O2-induced foci from the intrinsic γH2AX foci. This was achieved by subtracting the histograms representing control, untreated cells (i.e., the histograms shown in Fig. 3) from the histograms representing all γH2AX foci in Cpt- or H2O2-treated cells (Fig. 4B, E, H and Fig. 5B, E, H). All the three corrected histograms representing Cpt-treated cells (Fig. 4B, E, H) are strongly asymmetric and show a majority of the γH2AX foci located within no more than 1 μm of the nearest replication site. This result was expected because Cpt is known to cause DSBs at the sites of moving replication forks (). The histograms representing the H2O2 treatment do not show such asymmetry, revealing that histone γH2AX foci do not cluster in close proximity of DNA replication sites (Fig. 5B, E, H). This issue is discussed further in the text. The precise assessment of the numbers of DSBs (or γH2AX foci) occurring in and outside of replication areas as a result of both treatments requires analysis of the distances separating γH2AX foci from the nearest replication sites, as described below.
Counting DNA Replication-Dependent γH2AX Foci
The goal of this work is to quantitatively analyze a relationship between the induction of γH2AX foci and DNA replication (EdU sites). Toward this end, it is necessary to measure frequency (the actual number) of replication-dependent and independent γH2AX foci. This quantitative analysis hinges upon a correct criterion for distinguishing the dependent (neighboring) from independent (located afar) events. Such distinction has to rely on the knowledge of distances between the barycenters of the γH2AX and EdU labeled sites. A question arises as to what distance between the barycenters is to be expected when H2AX phosphorylation occurs within a replicating area—in its center or at the periphery. The barycenters of the dependent events (i.e., damage signaling induced in a replicating area) are not expected to coincide ideally in 3D space for the following reasons: (i) in the protocol, the EdU labeling pulse of the replication sites was of 30 min duration. A large area of nascent DNA was, therefore, labeled exclusively on one side of the DSB created by Cpt, before replication forks encountered Cpt-stabilized DNA-topoisomerase I (topo1) complexes terminating their progression (Fig. 6). H2AX phosphorylation occurred during the span of 30 min after the initial exposure to Cpt or H2O2 along megabase domain of DNA on both flanks of the DSB (Fig. 6) along hundreds of nucleosomes ([13, 20]); (ii) some contribution to a separation between the barycenters of the labeled nascent DNA and histone phosphorylation regions may also come from chromatin unfolding, relaxation, and movements that occur following the damage ().
In addition to the above, the nascent DNA (EdU labeled) and chromatin with phosphorylated H2AX regions are, to some degree, misrepresented by optical microscopy (Fig. 6). Specifically, the size of images of very small foci exceeds their actual physical size in as much as even the smallest labeled regions are represented by an ellipsoid of ∼250–500 nm in diameter and at least 700 nm in length. Thus, the H2AX phosphorylation that occurred in the replication focus, and the replication area itself, are represented by relatively large, possibly overdrawn, regions of fluorescence. In summary, even in the case, where the damage is inflicted at a particular replication fork, the barycenters of both labeled regions (γH2AX and replication) can eventually be separated by a measurable distance at a time when the cell is fixed. The knowledge about this particular distance is of importance for distinguishing replication-dependent and independent damage events.
This maximum distance between the centers of images representing regions of H2AX phosphorylation and damaged DNA replication area is a sum of four values: (i) the average radius of a replication region (in early and mid-S phase, approximately 400 nm and 600 nm in late S), (ii) the distance between the barycenters of replication and histone phosphorylation which arises from the fact that phosphorylation occurs on both sides of replication, whereas nascent DNA is located only on one side of the DNA break, we assume this to reach ∼100 nm, (iii) the error in measuring the position of each barycenter which may reach ∼100 nm in our system, and (iv) moving the centers of fluorescently labeled regions of replication and histone phosphorylation apart during a chromatin relaxation () which might reach 100 nm in the case of the experimental protocol we used. Thus, we estimate “the critical distance” to be ∼700 nm for the early- and mid-S phase, and ∼900 nm for the late S-phase cells.
The accuracy of the estimate of the size of the “critical distance” is fairly limited. Thus, we investigated the influence of setting the value of the “critical distance” on the number of events that are consequently qualified as dependent or independent. The number of H2AX phosphorylation foci recognized as related to replication, plotted as a function of the selected size of the “critical distance,” is shown in Figure 7. As expected, this dependence is very weak in control cells, where intrinsic H2AX phosphorylation appears to be unrelated to replication. The dependence is very strong for Cpt-treated cells, because in this case a large proportion of γH2AX foci are located close to replication. In cells treated with H2O2, the dependence is less strong, because the majority of γH2AX foci are seen at a large distance from replication. Conspicuously, the curves representing late S-phase in Cpt- and in H2O2-treated cells exhibit a change in slope at ∼400 and ∼1000 nm, respectively. These slope changes correspond to distances delineating the areas characterized by abundant and infrequent H2AX phosphorylation foci. As most Cpt-induced H2AX phosphorylation foci are close to replication, the curve bends near the value corresponding to a short distance, whereas in H2O2-treated cells this distance is larger.
Relationship Between γH2AX Foci Induced by Cpt or H2O2 and Replication
Based on the available knowledge, it was reasonable to expect that Cpt can damage DNA in replicating regions only, whereas H2O2 might damage replicating as well as nonreplicating sections ([5, 6, 8]). Analysis of the 3D images of replication and histone phosphorylation should resolve the question whether the previously reported higher susceptibility of replicating versus nonreplicating cells to H2O2-induced damage arises from specific lesions induced in replicating DNA.
The early-, mid-, and late-S histograms representing distances from Cpt-induced phosphorylation foci to the nearest replication region show a similar tendency—a vast majority of phosphorylation sites are found close to replication (Fig. 4). As described above, late S-phase reveals the spatial relationship between the two events most clearly. According to the definition of the “critical distance” described above, in late S-phase all Cpt-induced phosphorylation foci are found to be adjacent to the replication regions present when Cpt exposure commenced (Fig. 4I), and all such replication regions are associated with the γH2AX foci. Unexpectedly, in early and mid S-phase, some phosphorylation sites are detected at a distance greater than 700 nm from replication sites. This seems to contradict the notion that all Cpt-induced damage occurs only in replication forks, or suggests that the critical distance was set incorrectly, at too low a value. However, we note that the replication forks that terminated their movement shortly after addition of EdU or commenced replication just before cell fixation are likely to have incorporated only an undetectable amount of the precursor, as explained in the scheme in Figure 8. In such cases, the replication that was a target of Cpt is not detected and the γH2AX focus is incorrectly counted as replication-independent. There are also some replication foci detected in early S with no apparent association to γH2AX (Fig. 4C and a scheme in Fig. 8). However, this sub-stage of S-phase is characterized by a large number of small replication foci. It is, therefore, possible that the period of 30 min is not sufficient for Cpt (at the concentration we used) to bind (and induce damage) in all replication forks active in early S, or some replication foci are newly formed, as discussed below.
The same reasoning, which was described in the case of Cpt, can be applied to the analysis of images of cells exposed to H2O2 (Fig. 5). In the case of H2O2-induced damage, the late S-phase is also the period that most accurately can reveal the spatial relationships between replication and damage. We have analyzed the 3D images and subtracted the histograms representing endogenous (intrinsic) phosphorylation in the same way as in the case of Cpt (Fig. 5B, E, H). The histograms representing distances between H2O2-induced γH2AX foci and replication indicate that there are more H2AX phosphorylation sites in S-phase than in nonreplicating cells, consistent with the previous data obtained by LSC ([5, 6]). However, only some γH2AX foci were actually being induced in the replication areas. Thus, no clear preference for the induction of γH2AX within the replicating regions was apparent (Fig. 5C, F, I). This observation is also consistent with the fact that γH2AX foci are more frequent in nonreplicating H2O2-treated cells (∼83 ± 31/nucleus) than in nonreplicating untreated control (∼11 ± 18/nucleus), i.e., DNA damage signaling is strongly elevated following exposure to H2O2 regardless of the phase of the cell cycle. These observations reinforce the conclusion that, although H2O2-induced γH2AX foci are more frequent in S-phase than in G1, DNA damage signaling is generated principally outside of the replicating regions.
Activation of New DNA Replication Origins (ORI) in Cpt- and H2O2-Treated Cells
It is conspicuous that in all stages of S-phase the total number of replication sites is greater in Cpt- and H2O2-treated cells than in untreated cells (Table 1). This observation appears to be in conflict with the fact that Cpt induces DSBs and halts replication, since a lower, rather than a higher number of detectable replication foci should be expected in cells challenged with this drug. The increased number of replication foci that we observed may actually reflect activation of new replication forks in response to stalling the already existing ones by Cpt. It is possible that the newly activated replication origins are active sufficiently long so as to incorporate a detectable amount of the precursor before being also stalled by Cpt. One can hypothesize that, despite the presence of Cpt, some new replication origins (ORI) are activated (as discussed in () and the references cited therein) at distances large enough to be detected by confocal microscopy. Notably, the increase of the number of replication sites is higher in the case of Cpt, which specifically attacks replication forks, but not as high in the case of H2O2, which exhibits only a limited tendency to damage replicating regions.
The Method—Advantages and Limitations
The method of quantitative analysis of spatial relationship between H2AX phosphorylation and DNA replication, which is introduced in this report, yields the expected results in the case of Cpt-induced damage. The most suitable period for analysis is late S-phase. The nearest neighbor analysis of γH2AX and replication sites indicate that a great majority or even all of the damage events induced by Cpt occur in the regions of replicating DNA, as would be expected ([12, 13, 23]). This result validates the method as a tool in quantitative assessment of spatial relationships between discrete cellular events.
The analysis of H2O2-induced DNA damage leads to a conclusion that, although the DDR is elevated in S-phase, the damage is not inflicted preferentially in replicating DNA. Combined with the Cpt data, this observation reinforces the notion that the method we describe can be applied to studies of spatial relationship between discrete cellular events. Moreover, the analysis of H2O2-induced DNA damage opens an interesting question about a possible mechanism that confers higher vulnerability of DNA to oxidative damage in S-phase. As mentioned, the data presented here and previously ([5, 6]) indicate that whereas the incidence of DNA damage signaling induced by H2O2, as detected by γH2AX foci, is higher in S-phase, it is not preferential to the DNA replicating sites revealed by the incorporated EdU. It is known that the primary product of oxidation of DNA bases is 8-oxo-7,8-dihydro-2′-deoxyguanosine (abbreviated 8-oxo-dG) () and it may be expected to be generated randomly in nuclear DNA. There is also evidence demonstrating that oxidative stress can induce signaling pathways and lead to induction of DNA replication fork stalling and general replication stress (). Our data indeed show that only some γH2AX sites induced in cells exposed to H2O2 are associated with replication, while most of them are located at some distance from the replication sites that have incorporated EdU. This can occur if the DNA oxidation lesions, and resulting local changes in chromatin structure, are marked by histone H2AX phosphorylation, while replication forks are stalled at the same time by the signaling pathways. These two phenomena could result in the forks failing to reach the original 8-oxoG lesions during the course of the experiment. This is in contrast to DNA damage induced by Cpt, where DSBs are formed directly at the sites of the replication forks. However, the issue of higher incidence of H2AX foci in cells exposed to H2O2 warrants further studies.
The currently described method of measuring spatial (3D) immediacy between the DNA replication sites and DNA damage sites (γH2AX) induced by H2O2 or by Cpt are also relevant to analysis of damage sites induced by endogenous oxidants related to constitutive DNA damage response ([1-5]) as well as to other DNA damaging agents that induce localized DNA damage response in forms of foci. These may be foci of ATM, p53 binding protein 1 (53BP1), Rad51 or Rad50.
The method of quantitative analysis of spatial relationship between discrete subnuclear events can be refined further. A better spatial resolution of foci in 3D images is required and may be possible with super-resolution 3D optical microscopy, like SIM, PALM, STORM, STED, etc. The algorithm for quantifying the probability of phosphorylation occurring accidently within a replication region, as a function of the density of foci is reported in this article which accompanies this report () (same issue). This improvement should assist in achieving a better assessment of a correlation between two types of events characterized by a high density, which is the case of replication and damage in early S-phase. It is worth noting that the approach we are describing here bridges the gap between LSC analysis of large numbers of cells and analysis of high spatial resolution deconvolved 3D confocal images of individual cells. Another example of an approach which makes it possible to exploit compatibility between a wide spectrum of cytometric methods, including flow cytometry, LSC, wide field, and confocal microscopy, has been described recently ([27, 28]). This report also describes analysis of proximity between DNA replication and DNA foci, albeit using a different detection scheme (in situ proximity ligation assay).
The accuracy of detection of DNA damage signaling has limitations because of a relatively high and uneven background which seems always present in immunofluorescence detection of γH2AX. This background may be a consequence of (i) the low level constitutive DNA damage induced by endogenous oxidants, (ii) autofluorescence, and (iii) nonspecific binding of the fluorochrome-tagged antibodies. The analysis would certainly benefit from better understanding or minimizing this background. It should also be noted that the intensity of γH2AX fluorescence diminishes with time when its Ser139 becomes dephosphorylated in the course of DSBs repair ().
Obviously, the method can only be applied to studies of nuclei with limited density of independent foci. Because at high density, foci cannot be resolved and counted only a certain range of damage levels can be analyzed accurately. A high-resolution, high-sensitivity microscope capable of detecting and imaging smaller foci in 3D space will expand the capabilities of this method to studies of a wider range of damage levels.
- The method of quantitative analysis of spatial relationship between discrete nuclear events described here is suitable for analysis of interdependencies between H2AX histone phosphorylation and DNA replication foci in cells subjected to treatments resulting in DSBs. This approach has a potential to be used for analysis of any other discrete cellular events detectable as a pattern of foci by 3D optical microscopy.
- In cells subjected to H2O2, only a minor proportion of H2AX histone phosphorylation sites coincide with replicating areas. Thus, an increased level of histone phosphorylation which is seen in S-phase of cells subjected to oxidative stress induced by H2O2 is not a direct consequence of DSBs inflicted on replicating DNA.
- DNA damage and stalling of replicating forks caused by Cpt (or, to a lesser degree, H2O2) does not globally inhibit replication in the nucleus but appears to induce activation of new replication origins.
P.R. is a recipient of SET doctoral studentship from Jagiellonian University. A travel grant awarded by ISAC to P.R. to attend cytoa2012 is gratefully acknowledged.
- 26Analysis of spatial correlations between patterns of DNA damage response and DNA replication in nuclei of cells subjected to replication stress or oxidative damage. Cytometry Part A 2013; 83A:925–932., , , , , , .
- 28A computational platform for robotized fluorescence microscopy (II): DNA damage, replication, checkpoint activation, and cell cycle progression by high-content high-resolution multiparameter image-cytometry. Cytometry Part A 2013;83A:344–355., , .