1. Localization of gDNA and gDNA^OX in MCF-7 cells

To find out the intracellular locations of gDNA and gDNA^OX, a number of DNA probes were synthesized and differentially labeled. gDNA^red and pBR322^green probes were labeled using nick-translation with SpectrumRed and SpectrumGreen, respectively. In MCF-7 cells, gDNA^red and pBR322^green demonstrate similar granulated, clumped staining patterns in the periphery of the cytoplasm, visible in approximately 70% of cells (Figure 1А). More detailed analysis showed that intracellular distribution of labeled DNA fragments is sample specific (Figure 1В). In cells treated with both gDNA^red and pBR322^green, some areas of the cytoplasm are stained with one, but not the other type of labeled DNA. Areas stained with more sequence-diverse gDNA^red are present in larger numbers and occupy a larger volume of the cell. In gDNA^red stained cells there was also a diffuse staining near the nuclear envelope that was visible at a higher magnification (x 200). Our observations indicate that at least some exogenous gDNA fragments are imported into the cell.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Staining of MCF7 cells with various types of labeled DNA.

A - gDNA^red, nuclei are stained with DAPI (x40); B – merged staining patterns of gDNA^red and pBR322^green (x200); С – merged staining patterns of gDNA^red-ox and FITC-conjugated antibodies to 8-oxodG (x200); D – FACS analysis of early endosomal marker EEA1; the distribution of the cells with varying EEA1 contents.

Final concentrations of added DNA in the media were at 50 ng/mL; cells were incubated with DNA for 30 min before fixation in 3% formaldehyde. In case of staining with FITC-conjugated antibodies to 8-oxodG, fixed cells were pretreated with 0.1% Triton Х100 for permeation.

https://doi.org/10.1371/journal.pone.0077469.g001

To determine the intracellular locations for gDNA^OX, a composite probe was produced by slow renaturation of nick-translation labeled gDNA^red and gDNA^OX (gDNA^red-OX). Similar to gDNA^red, this composite labeled probe was also located at the periphery of the cytoplasm (Figure 1С), however, in case of the composite probe gDNA^red-OX, a substantial portion of the labeled fragments were found inside of the cytoplasm near the nucleus. To confirm that this diffuse staining corresponded to oxidized DNA, we stained the cells with FITC-conjugated antibodies to 8-oxodG (Figure 1C). Our data indicate that gDNA^OX is imported into the cell at a substantially larger degree than gDNA. After entering the cell, gDNA^OX locates in the cytoplasm, forming foci around the nucleus.

Endocytosis is one of the common ways of delivery of exogenous compounds into the cell. The formation of novel endosomes is accompanied by an increase in expression of early endosome antigen 1 protein (EEA1), known as an early endosomal biomarker [26]. Using FACS, we demonstrated that exposure to DNA^OX leads to an increase of the proportion of cells that express high levels of EEA1 (Figure 1D). These observations are in concert with visual patterns of intracellular staining for gDNA^OX.

It is known that intracellular sensors are capable of binding to DNA fragments either inside the cytoplasm (AIM2, RIG1, STING) [27] or within the endosomes (TLR9) [28]. Interestingly, 2-hours exposure to gDNA^OX stimulates the expression of mRNAs encoding AIM2, TLR9 and RIG1 (Table 1). Two DNA sensors, AIM2 and TLR9, were studied in greater details (Figure 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. The exposure to gDNA^OX (50 ng/mL) leads to a transient increase in expression cytoplasmic DNA sensor AIM2, while not changing expression levels of TLR9.

A - intracellular localization of AIM2 (FITC-conjugated antibodies) and labeled probe gDNA^red-ox (x40). B – the ratio of the levels of AIM1 [1] and TLR9 [2] – encoding RNAs to the levels TBP-encoding reference mRNA in cells exposed to gDNA or gDNA^OX for 2 hrs (grey columns) and 48 hrs (black columns).

C and D – Flow cytometry detection of AIM2 (C) and TLR9 (D) expression in MCF-7. Cells were stained with AIM2 (C) or TLR9 (D) antibody (secondary PE-conjugated antibodies). Panels D [1] and E [1] – control cells plots: FL2 versus SSC. R: gated area. Panels C [2] and D [2]: median signal intensity of FL2 (R) in MCF-7 cells (mean value for three independent experiments). Panels C [3] and D [3]: relative proportions of AIM2- or TLR9-positive cells in R gates [1]. Background fluorescence was quantified using PE-conjugated secondary antibodies.

*p < 0.05 against control group of cells, non-parametric U-test.

https://doi.org/10.1371/journal.pone.0077469.g002

2. Exposure to gDNA^OX induces short-term oxidative stress

To study possible influence of gDNA and gDNA^OX on the intracellular levels of reactive oxygen species (ROS), the ROS were measured using dichlorodihydrofluorescindiacetate (H2DCFH-DA) dye that rapidly penetrates cell membranes, and gets trapped in the cytosol in its deacetylated form. Nonfluorescent DCFH transforms to fluorescent DCF by a variety of ROS radicals and, therefore, serves as a sensitive intracellular marker for oxidative stress [29]. Figure 3A depicts the results of the ROS levels analysis in living cells. In untreated control cells, DCF dye diffusely associates with the surface of the cell, and may be removed from the membrane by PBS washing. Most common sources of ROS at cellular membrane are enzymes of NOX family [30]. In cells treated with gDNA (50 ng/mL), H2DCFH-DA stain visualizes both the membrane and some amount of intracellular granules. The PBS wash does not influence cytoplasmic granule staining. Patterns of DCF granules and labeled gDNA^red probe stains approximately overlap (Figure 3C), possibly indicating that an interaction of gDNA with some cellular constituents stimulates ROS biosynthesis at the place of contact. This observation aligns well with previously stated hypothesis that ecDNA may somehow directly stimulate enzymatic activity of NOX proteins [5].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. The exposure to gDNA^OX leads to an increase in the production of ROS.

А – Microscopy-based evaluation of MCF-7 cells sequentially treated with DNA (50 ng/mL) and H2DCFH-DA (control, gDNA, gDNA^ox [1]) and incubated for 30 minutes (x100). Alternatively, MCF-7 cells were incubated with DNA (50 ng/mL) for 1 hour followed by addition of H2DCFH-DA and photography 30 minutes later (gDNA^ox [2]). B - MCF-7 cells exposed to gDNA^ox (0.5h; 50ng/mL), were sequentially treated with Mito-tracker TMRM (15 min) and H2DCFH-DA (15 min) (x200). C - Co-detection of labeled probe gDNA^red (50 ng/mL) and DCF after 30 minutes of incubation. D - The results of the quantification of fluorescence using plate reader [1]. The time kinetics of fluorescence outputs in cells sequentially treated with H2DCFH-DA and, three minutes later, a DNA sample at final concentration of 5 or 50 ng/mL [2]. The same for cells pretreated with DNA (final concentration 5 ng/mL) for one hour, with subsequent addition of H2DCFH-DA. *) p < 0.05 against control group of cells, non-parametric U-test.

https://doi.org/10.1371/journal.pone.0077469.g003

In cells treated with gDNA^OX (50 ng/ml), intracellular ROS-producing granules arise fast, and their numbers are substantially larger than in cells treated with gDNA (Figure 3A, inset gDNA^OX[1]). These events are accompanied by changes in the morphology of MCF-7 cells, including an increase in size of nuclei and cytoplasmic swell. It is important to note that observed cellular responses are rapid and short-living. Described changes in staining patterns and cell morphology are seen only in case of sequential additions of H2DCFH-DA and gDNA^OX to MCF-7 media. When cells were pre-treated with gDNA^OX for 1 hour, then studied using а H2DCFH-DA dye, the number ROS-synthesizing granules seen in cells was lower and their intensities were less bright than in case of no-pretreatment protocol (Figure 3А inset gDNA^OX [2]). Even more interesting, in pre-treatment protocol, some cells stopped ROS biosynthesis at all, and became even less bright then non-treated control cells (darker cells that are less fluorescent than the background (Figure 3А inset gDNA^OX (b)).

The observed phenomena were independently confirmed in a study of DCF generation kinetics using quantification with a fluorescent reader (Figure 3D). When MCF-7 cells were treated with DNA immediately after addition of H2DCFH-DA to the media, a dramatic increase in the intensity of DCF fluorescence was observed. These increases were at the highest rates of increase during first 20 minutes after the addition of DNA to the media (coefficient k1), then, with time, these rates drop (coefficient k2) (Figure 3D[1], Table inset). k1 and k2 coefficients were dependent on type and concentrations of DNA treatment: gDNA^OX (5ng/mL) > gDNA (5ng/mL) > gDNA^OX (50 ng/mL) ≥ gDNA (50 ng/mL) > control. These effects were not seen when cells were pretreated with DNA for 1 hour before the addition of H2DCFH-DA (Figure 3D[2]).

Taken together, the results of these experiments indicate that treatment with gDNA^OX rapidly induces ROS biosynthesis in MCF-7 cells. In parallel, the opposite process of the suppression of ROS generation, or ROS quenching, is initiated. As larger the amounts of gDNA^OX were added to the media, the more rapid was the development of ROS quenching.

A bulk of the intracellular ROS is generated by mitochondria. An increase in oxidative metabolism in mitochondria may lead to the diffusion of ROS into cytoplasm and subsequent increase in perimitochondrial detection of ROS by DCF. To test this hypothesis, we sequentially stained the cells exposed to 50 ng/mL of gDNA^OX for 30 minutes with Mito-tracker (TMRM red) and DCF (Figure 3B). A majority of Mito-tracker and DCF signal were located close to each other, with partially overlaps (yellow signal, Figure 3B). In intact cells, H2DCFH-DA does not stain mitochondria (Figure 3А, control). Our observations point that in the cells exposed to oxidized DNA, a majority of endogenous ROS is generated by mitochondria.

3. Exposure togDNA^OX stimulates an increase in the levels of oxidative modification of cell’ own DNA

It is likely that intensive production of ROS observed immediately after exposure of cells to gDNA^OX may result in the damage to cellular DNA. To visualize this damage, fixed MCF-7 cells were stained with PE-labeled anti-8-oxodG antibodies (Figure 4). As compared to non-treated control cells, in MCF-7 cultures treated with either gDNA or gDNA^OX, the amounts of stained cells were increased (Figure 4A (x20). At larger magnifications, three types of staining patterns may be detected (Figure 4B): (1) – nuclear staining; (2) – cytoplasmic staining; (3) – staining for micronuclei. In non-treated control populations of MCF-7 cells, PE-labeled anti-8-oxodG antibodies predominantly stain micronuclei. In populations treated with gDNA^OX, there was an increase in the amounts of cells with nuclear staining (Figure 4E). As our previous experiments showed that gDNA^red-OX is located in cytoplasm and does not penetrate the nucleus, observed staining of nuclei shall be attributed to the damage of cell’ own DNA.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. The analysis of 8-oxodG content in cells exposed to either gDNA or gDNA^OX (50 ng/mL).

A - Cells stained with PE-labeled anti-8-oxodG antibodies and DAPI (x20). B - Three types of anti-8-oxodG stain distribution observed in cells treated with gDNA^OX (x100). Cell were incubated with DNA samples for 1 hour, fixed with 3% formaldehyde, permeated with 0,1 % triton X100 and stained with anti-8-oxodG (PE-conjugated secondary antibodies). C – colocalization of 8-oxodG with mitochondria. Cells were incubated with gDNA^OX for 0.5 hour, обработаны Mito-tracker (30 nM, 15 min), photographed, then fixed with 3% formaldehyde, permeated with 0,1 % triton X100, stained with anti-8-oxodG antibodies (FITC-conjugated secondary antibodies) and photographed again. D - 8-oxodG content in DNA exposed cells pre-treated with NAC (FACS analysis). Cells were incubated with NAC (0.15 mM) for 30 min, then exposed to gDNA^OX for 1 hour and analyzed using anti-8-oxodG antibodies (PE-conjugated secondary antibodies). Background fluorescence was quantified using PE-conjugated secondary antibodies. E - Relative proportions of nuclei stained for 8-oxodG in non-treated control cells, cells exposed to gDNA, cells exposed to gDNA^OX (grey columns). Light grey column reflects cells pre-treated with NAC and exposed to gDNA^OX. *p < 0.05 against control group of cells, non-parametric U-test.

https://doi.org/10.1371/journal.pone.0077469.g004

An increase of mitochondrial biosynthesis of the ROS in gDNA^OX exposed cells demonstrated above (Figure 3В) may lead to an increase in the level of oxidation in mitochondrial DNA that, in turn, may explain observed cytoplasmic staining for gDNA^red-OX shown at Figure 1C. On Figure 4C, one may see that some 8-oxodG signals do not merge with gDNA^red-OX. In cells pretreated with antioxidant N-acetyl-cysteine (NAC) (0.15 mM) for 30 minutes before exposure to gDNA^OX, the levels of oxidation in cellular DNA were substantially lower than in cells not treated with NAC (Figure 4D and 4E).

4. Exposure to gDNA^OX stimulates an increase in strand breaks in cell’ own DNA

One of well-known feature of DNA oxidation is an accumulation of single- and double strand DNA breaks (SSBs and DSBs). To quantify SSBs and DSBs in MCF-7 cells exposed to either gDNA or gDNA^OX, we employed comet electrophoresis in alkaline conditions (Figure 5А). Three types of nuclei were enumerated: nuclei with intact DNA (Figure 5А [1], Type I); nuclei with some degree of chromatin fragmentation (Type II); nuclei with substantial fragmentation of DNA (Type III). In majority of cases, the nuclei of non-treated control are classified as either Type I or Type II, while Type III nuclei are seen predominantly in cells treated with gDNA^OX. Depending on how long the cells were exposed to gDNA^OX, the proportions of Type III nuclei may differ. Figure 5A also presents the comet tail moments [2] and % tail DNA [3]. After 30 minutes of incubation of MCF-7 cells with gDNA^OX, the amounts of DNA breaks drastically increase, while similar treatment with gDNA leads to moderate elevation of chromatin fragmentation levels. After 2 hours of incubation either with gDNA or gDNA^OX, the amounts of DNA breaks decrease, and their number falls to below of that found in respective gate-specific populations in non-treated control cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. DNA damage in cells exposed to either gDNA or gDNA^OX at final concentration 50 ng/mL for 30 min and 2 hours.

А – comet assay in alkaline conditions [1]. - Digital photography of the nuclei with varying degree of DNA damage [2,3]; - cumulative histograms for tail moment and percentage of DNA within tails. The reliability of differences with the control in the obtained distributions was analyzed by means of Kolmogorov–Smirnov statistics (the table shows the values of D and α).

B - dsDNA breaks in cells exposed to gDNA^OX (50ng/mL, 1 hour).Cells were processed for immunofluorescence staining with anti γH2AX antibody (x40) [1].- Three detected types of nuclei are denoted by numbers: 1- nucleus with multiple dsDNA breaks, 2- nucleus with a few dsDNA breaks, 3- nucleus with intact DNA [2]. - Example of a micronucleus with dsDNA breaks.

С – FACS analysis of γ-foci A: there main fractions of the cells as evident in gating areas R1, R2, R3 [1], the distribution of γH2AX fluorescence intensities [2], relative proportions of cells within gating areas R1-R3 [3]. *p < 0.05 against control group of cells, non-parametric U-test.

https://doi.org/10.1371/journal.pone.0077469.g005

Observations described above were independently confirmed using another common technique for visualization of DSBs, an immunostaining with antibodies against the histone γH2AX, phosphorylated by serine-139. This form of H2AX is known to rapidly accumulate at DNA loci flanking the DSB site [31]. MCF-7 cells stained with FITC-conjugated antibodies to Ser-139 phosphorylated histone γН2АХ are shown at Figure 5B [1]. Stained slides also included three different cell populations of γН2АХ positive cells. In this experiment, cells were classified as Type 1 cells when they had multiple phospho-γН2АХ foci. Most of the γН2АХ positive cells were classified as Type 2 cells (between 2 and 10 distinct γН2АХ foci per cell), and Type 3 cells with no signs of the focal phospho- γН2АХ staining.

In anti-γН2АХ staining, overall fluorescence intensity of the cell is proportional to the number of γН2АХ foci per cell, and, therefore, to amount of DSBs. Using FACS, three gated areas, R1 to R3, were studied (Figure 5C[1,2]). Cells within gate R1 have largest FL1 (γH2AX); this is interpreted as multiple DSBs (Type 1 cells, Figure 5B). Gate R2 contains cells with not numerous γH2AX (Type 2 cells). Gate R3 contains the largest number of cells; most of these cells are intact with no DSBs (Type 3 cells). In MCF-7 cultures, an exposure to gDNA^OX (1h) leads to a 1.5-folds increase in the number of cells within gate R1 that is paralleled by a decrease in the number of cells within R2. After 24 hours of exposure to gDNA^OX, the amounts of cells with multiple DSBs decrease to the levels below that that in non-treated control cells (Figure 5C[3]). A treatment with gDNA evokes similar, but less pronounced type of cellular response that in its magnitude does not reach significance when compared to non-treated control cells (p>0.05).

These observations indicate that, in MCF-7 cells, short-term exposure to gDNA^OX results in both single- and double strand DNA breaks. Longer durations of the treatment (between 2 and 24 hours) evoke some type of compensatory response that leads to a decrease in the levels of chromatin fragmentations across cell populations.

The drop in the proportion of DSB-containing cells after short-term exposure to oxidized or control DNA may be explained either by the repair of the breaks, or by apoptosis/detachment of damaged cells, or both. To evaluate these possibilities, we enumerated cells that remain in the media after its removal from cell layer, and cells removed from the layer after PBS wash. In cultures exposed to oxidized DNA for 2 hours, the proportion of detached cells remained similar to that in cultures exposed to genomic DNA and non-treated control cultures (approximately 2% of total amount of cells in given culture). Similar results were obtained in experiments aimed at direct evaluation of apoptosis (see below). Therefore, it is likely that the decrease in the proportion of cells with DSBs observed after exposure to gDNA or gDNA^OX is due to an increase in DNA repair.

5. Exposure to gDNA^OX leads to an increase in genome instability

Single- and double strand DNA breaks are known to result in the loss of chromosome stability that is especially prominent in actively proliferating cells [32]. A thorough study of the nuclei of the cells incubated with gDNA^OX revealed pronounced chromosome instability (Figure 6). At concentrations of 50 ng/mL, an exposure of actively proliferating, low confluency MCF-7 cells to gDNA^OX results in the formation of multiple micronuclei (Figure 6A[1]) and other nuclear anomalies such as nucleoplasmic bridges and nuclear buds (Figure 6A[2]), as well as in decondensation of mitotic chromosomes (Figure 6A[3]). All of these events are signs of profound replication stress that is known to develop in actively proliferation cell cultures undergoing various stress treatments [32]. Similarly treated cell cultures with lower proportions of proliferating cells, for example, confluent or serum starved cultures show substantially lesser the amounts of chromatin changes. Proportions of micronuclei-containing cells in cultures grown in varying conditions are show at Figure 6В. In non-treated control MCF-7 cells, the frequency of cells with micronuclei was around 7%, a number that is similar to that reported in other studies [33]. In actively proliferating cultures exposed to gDNA^OX, the micronuclei were detected in about 40% of cells. Exposure to gDNA also leads to increase in the amounts of cells with micronuclei, but in this case an increase is not significant. Many micronuclei formed after the treatment with gDNA^OX were positively stained for both PE-labeled anti-8-oxodG (Figure 4B and Figure 6C) and anti-phospho-γН2АХ antibodies that highlight DSBs (Figure 5B [2]).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. Genome instability in MCF-7 cells exposed to gDNA^OX at final concentration 50 ng/mL for 24 hours.

A – multiple micronuclei [1], chromatin bridges [2], M-phase chromatin decondensation [3], non-treated control cells [4] (x100).

B – proportions of cells with micronuclei in non-treated control cells, cells exposed to gDNA, cells exposed to gDNA^OX. Grey columns: non-confluent, actively proliferating MCF-7 culture. Black columns: MCF-7 cells at high confluency. *p < 0.05 against control group of cells, non-parametric U-test.

С - Exposure to gDNA^OX (50 ng/mL, 2 hours) induces formation of 8-oxodG-containing micronuclei (x100).

https://doi.org/10.1371/journal.pone.0077469.g006

These observations indicate that, in MCF-7 cells, an exposure to gDNA^OX induces genome instability that is, most likely, secondary to accumulation of large the amounts of SSBs and DSBs.

6. Exposure to gDNA^OX arrests cell cycle

One of the most important consequences of genome instability is the block of cell proliferation due to activation of the DNA damage checkpoints. Cell cycle-related consequences of exposure to or gDNA were studied in MCF-7 cells that were harvested 48 hours after addition of DNA (50 ng/mL) to the media (Figure 7).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 7. Proliferation and cell cycle of MCF-7 cells exposed to gDNA or gDNA^OX at final concentration 50 ng/mL for 48 hours (FACS).

A: (1) - fixed cells stained with anti-Ki-67 antibodies (green color). Background fluorescence was quantified using FITC-conjugated secondary antibodies (grey color) [2]. - proportion of Ki-67-positive cells in total cell population [3]. - the average signal intensity of FL1 (Ki-67+). Cells were cultivated either in absence (dark grey columns) or in presence of 0.15 mM NAC (light grey columns). B: (1) - fixed cells stained with anti-PCNA antibodies (green color). Background fluorescence was quantified using FITC-conjugated secondary antibodies (grey color) [2]. - proportion of PCNA-positive cells in total cell population [3]. - the average of the median signal intensities of FL1 (PCNA+).

C: (1) - distribution of fluorescence intensities of the cells stained with пропидий йодидом. (2) - содержание в популяции клеток с количеством ДНК, соответствующим G1-, S and G2/M –фазам клеточного цикла.

*p < 0.05 against control group of cells, non-parametric U-test.

https://doi.org/10.1371/journal.pone.0077469.g007

To investigate these cultures, cells were stained with antibodies to the proliferation markers Ki-67 and PCNA [34,35] and enumerated by FACS. Additionally, cell counts were also performed after DNA-specific propidium iodide (PI) treatment. Figure 7A shows the distribution of the cells with various Ki-67 contents. In control MCF-7 cultures, Ki-67 stains approximately 45% of cells. After exposure to gDNA^OX, the proportion of Ki-67-positive cells decreased to 30% (Figure 7A[2]). These decreases were paralleled by the decrease in mean fluorescence intensity per each Ki-67-positive cell by 40% that is indicative of the decrease in amounts of Ki-67 in individual cells. Similar results were obtained using another well-known marker of proliferation, PCNA (Figure 7B[1-3]). It seems that observed block of proliferation is ROS-dependent, as the changes in KI-67 staining of the cells pre-treated with antioxidant NAC (0.15mM) and exposed to same amounts of oxidized DNA were not significant (Figure 7С[2,3]).

The data collected after the staining with propidium iodide (PI) point to similar direction (Figure 7С[1]). After exposure to gDNA^OX, the proportion of G0/G1 cells increased, while proportions of the cells in S- and G2/M phases decreased (Figure 7С[2]). These observations indicate that, in a substantial proportion of previously proliferating MCF-7 cells, the exposure to gDNA^OX and, to a lesser degree, to gDNA, blocks the cell cycle in G0/G1.

This line of evidence was also supported by qRT-PCR analysis at the level of mRNA encoding inducible cell cycle arrest proteins, including CDKN2A (p16INK4), CDKN1A (p21CIP1/WAF1) and TP53 (Table 1). Cell cycle changes evoked by treatment with gDNA were similar to those of gDNA^OX, but substantially less pronounced.

7. Exposure to either gDNA^OX or gDNA supports cell survival

It was noted that the total amount of cells harvested 48 hours after exposure to gDNA^OX or gDNA were similar to those of non-treated control populations (Figure 8А). As the proliferation activities of cells treated with either gDNA^OX or gDNA were, at least in part, blocked (Figure 7), it was important to evaluate overall levels of cell death in all studied populations.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 8. Cell death in MCF-7 cultures exposed to either gDNA or gDNA^OX at final concentration 50 ng/mL for 48 hours.

A. Total number of cells in studied cell population.

B. (FACS) – enumeration of cells with sings of early apoptosis [1]. - the distribution of fluorescence intensities of the cells stained with Annexin V-FITC (green color) или FITC-conjugated secondary antibodies (grey color) [2]. - control cells plots: FL1 versus SSC. R: gated area [3]. - the proportion of Annexin V -positive cells in total cell population.

C. Evaluation of modified nuclei in three studies typed of MCF-7 cultures. (1) -Example of Hoechst33342 staining; (2) - Graph of the proportion of cells with modified nuclei in three studied types of MCF-7 cultures.

D. Electrophoresis [1] and evaluation of ecDNA concentrations [2] in the media of non-treated control cells and cells exposed to either gDNA or gDNA^OX. Dashed line indicates amounts of ecDNA that should be present in the media when exogenous DNA is taken into account. *p < 0.05 against control group of cells, non-parametric U-test.

https://doi.org/10.1371/journal.pone.0077469.g008

To quantify cells in early apoptosis, we used FITC-conjugated Annexin V (Figure 8B[1-3]). After two hours of exposure either gDNA^OX or gDNA, the proportion of the apoptotic cells went down approximately by 25%, but observed changes had not reached significance (p>0.05)). However, after 48 hours of exposure to either gDNA^OX or gDNA, the proportion of apoptotic cells in treated cultures decreased to the levels twice less than in control MCF-7 cultures.

To evaluate overall levels of cell death in all studied populations, nuclear morphology was evaluated in all populations after staining with Hoechst33342 (Figure 8C [1,2]). If condensed and fragmented chromatin was detected, the cell was marked as apoptotic. After exposure to gDNA^OX (48 hours, 50 ng/mL), the amount of cells with apoptotic nuclei decreased three folds.

To further assess various aspects of cell death, we extracted ecDNA from cell-free media conditioned by non-treated control cells and cells treated either with gDNA or gDNA^OX for 48 hours (50 ng/mL). Extracted DNA fragments were analyzed by gel electrophoresis to assess their size distribution (Figure 8D[1]). The length of DNA fragments extracted from cell-free media conditioned by non-treated control cells, varied between 15 kb and 0.1 kb, and included visible mono- and dinucleosome bands that are contributed to the ecDNA pool by dying apoptotic cells [36]. In cells treated either with gDNA or gDNA^OX, these bands were less prominent. The decrease in relative abundance of mono- and dinucleosome bands was in concert with the overall decrease in total amounts of ecDNA extracted from cell-free media and quantified using RiboGreen stain (Figure 8D[2]). In media of MCF-7 cells exposed to exogenous DNA, the final concentrations of ecDNA should be around 190 ng/mL (a sum of concentrations of endogenously produced DNA at 140 ng/mL and added DNA at 50 ng/mL); However, cell-free media of cells treated with exogenous DNA had substantially lower concentrations of DNA, in fact, after treatment with gDNA, these concentrations were 1.7 times lower than expected. After treatment with gDNA^OX, these concentrations were 6 times lower than expected. These drastic drops in DNA concentrations may be explained by the decrease of overall levels of apoptosis and DNA release in gDNA or gDNA^OX treated cultures.

Figure 8 presents evidence that in gDNA^OX treated MCF-7 cultures and, to lesser degree, in gDNA treated cells, the levels of cell death substantially decrease as compared to non-treated controls. Additional supportive evidence for this statement is presented in Table 1 that summarizes the changes in expression levels for mRNAs encoding cell survival and DNA repair related proteins. In two hours after adding gDNA^OX to MCF-7 culture, levels of mRNA for BCL2, BCL2A1 (Bfl-1/A1), BCL2L1 (BCL-X), BIRC3 (c-IAP1) and BRCA1 increase 1.2 to 6.4 folds, and stay elevated for at least 48 hours. In case of treatment with gDNA, these genes also tend to increase their mRNA biosynthesis, up to 1.9 - 3.5 times, but these changes in expression levels are delayed as compared to the treatment with gDNA^OX and reach significance only after 48 hours. Interestingly, in case of treatment with gDNA, the expression levels of mRNA encoding for key component of DSB repair machinery BRCA1, were not altered.

8.  Exposure to either gDNA^OX or gDNA leads to a decrease in activity of NRF2 and an increase in activity of NF-kB and STAT3

NF-E2-related factor 2 (NRF2) is known to participate in the development of adaptive response in fibroblasts and mesenchymal stem cells cultivated in the presence of gDNA^OX [5,7]. After 2 hours of exposure of MCF-7 cells to gDNA^OX, the levels of NRF2 mRNA increase (Table 1). At the same time point, there is an increase in the expression of the gene KEAP1 that encodes for a cytoplasmic protein partner of NRF2, capable of blocking its transcription factor activity [37]. As evident from FACS data, protein levels of NRF2 after treatment with gDNA do not change (Figure 9A). An exposure to gDNA^OX for 2 hours leads to a decrease of NRF2 levels. Fluorescent microscopy studies showed that exposure to gDNA^OX leads to a change in the NRF2 staining pattern. In non-treated control MCF-7 cells, NRF2 is located both in the nucleus (^~50% of cells) and in the cytoplasm (most of the cells), while in cells exposed to gDNA^OX NFR2 is found exclusively in the cytoplasm (Figure 9B), thus, indicating suggesting that its transcriptional activator function is blocked.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 9. Decrease in activity of transcriptional factor NRF2 in MCF-7 cells exposed to gDNA^OX at final concentrations of 50 ng/mL for 2 hours.

A FACS: the average of the median signal intensities in cells stained with anti-NRF2 antibodies after various exposures. B - Fluorescent microscopy of cells stained to NRF2 (x40). C - Graph of the proportion of cells with nuclear staining for NRF2 in three studied types of MCF-7 cultures. *p < 0.05 against control group of cells, non-parametric U-test.

https://doi.org/10.1371/journal.pone.0077469.g009

NF-κB and STAT3 control the expression of anti-apoptotic and cell cycle control and proliferation genes. Both of these transcriptional factors are activated in response to various kinds of stress. In particular, NF-κB and STAT3 were found to play pivotal roles in various aspects of tumorigenesis [38,39]. Here we present an analysis of activity of these two transcription factors in cells exposed to either gDNA or gDNA^OX.

NF-κB.

The exposure to gDNA^OX leads to a rapid, 1.8-3.6 fold increase in the levels of mRNAs encoding components of the NF-κB pathway, including MAP4K4, MYD88, NFKB1 and TIRAP (Table 1).The effects of exposure to gDNA are seen substantially later, at 48 hours post exposure (MAP4K4, MYD88 and TIRAP). After 2 hours of exposure to either gDNA or gDNA^OX, the amount of NF-κB (p65) proteins increase 1.5 fold (FACS, Figure 10C), and decrease 48 hours later. Fluorescent microscopy evaluation of gDNA^OX-treated MCF-7 cells confirms activation of NF-κB as evident from the translocation of this factor into the nucleus (Figure 10A). After 2 hours of exposure, the fraction of MCF-7 cells with nuclear staining for NF-κB increases from 12% to 56% (Figure 10B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 10. Increase in activity of transcriptional factor Nf-kB in MCF-7 cells exposed to gDNA^OX at final concentrations of 50 ng/mL for 2 hours.

A Fluorescent microscopy of cells stained with anti-p65 (FITC) antibodies (x40). B Graph of the proportion of cells with nuclear staining for Nf-kB in three studied types of MCF-7 cultures.

C, D (FACS) - the average signal intensity of FL1 (p65) in cells stained with anti-p65 (C) and Ser529-phosphorylated р65 (D) antibodies [1]. - distribution of fluorescence intensities of the cells stained with Ser529-phosphorylated р65 antibodies (FITC) (green color) или FITC-conjugated secondary antibodies (grey color) [2]. - proportion of Ser529-phosphorylated р65 -positive cells in total cell population [3]. - the average of the median signal intensities of FL1 (Ser529-phosphorylated р65 +). Cells were cultivated either in absence (dark grey columns) or in presence of 0.15 mM NAC (light grey columns).

https://doi.org/10.1371/journal.pone.0077469.g010

It is known that NF-κB (р65) is activated by phosphorylation, which plays a key role in the regulation of its transcriptional activity and is associated with nuclear translocation. For instance, upon treatment with TNFa, Ser529 of р65 is phosphorylated by casein kinase II [40]. Flow cytometry quantification (Figure 10D) demonstrates that exposure to gDNA^OX leads to an increase of the proportion of cells that contain Ser529-phosphorylated р65, thus, confirming that NF-κB in these cells is transcriptionally active [40]. The exposure to gDNA does not increase the proportion of cells with Ser529-phosphorylated р65. The pre-treatment with antioxidant NAC at 0.15mM for 30 minutes before addition of same amount of oxidized DNA prevented an increase in the levels of Ser529-phosphorylated р65 that remained similar to that in control cells (Figure 10D [2,3]). Therefore, we may conclude that oxidized DNA dependent activation of NF-κB is mediated by an increase in local production of ROS.

STAT3.

Two hours exposure to gDNA^OX also leads to an increase in the expression of mRNA for STAT3 and STAT6 (3 and 1.6 fold, respectively) (Table 1), while exposure to gDNA results in significant activation of STAT3 and STAT6 only at the 48 hour time point. Both FACS and fluorescent microscopy show that non-treated control MCF-7 cells express substantial amounts of STAT3 (Figure 11A[1,2], 11B[1] и 11С). Importantly, in these cells STAT3 is located exclusively in the nuclei. These observations indicate that STAT3 in active in control MCF-7 cultures. Published studies describing activity of Stat3 in MCF-7 contradict each other. Some authors showed that in MCF-7 Stat3 is phosphorylated and located in the nuclei [41]. Other studies failed to detect activity of Stat3 in MCF-7 [42]. Stat3 activity may change in response to growth factors and cytokines [38,39]. Therefore, observed disagreements may be explained by differing cultivation conditions, in particular, by type of the serum supplementation. Interestingly, supplementation of the media with antioxidant NAC leads to decrease in activity of Stat3 (Figure 11B[2]).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 11. Activity of STAT3 is stimulated in MCF-7 cells exposed to either gDNA or gDNA^OX at final concentrations of 50 ng/mL.

A FACS: Frequency plot for fluorescence intensities in cells stained with anti-STAT3 antibodies [1] and the average of the median signal intensities of FL1 (STAT3) in these cells [2].

B Fluorescent microscopy of cells stained with STAT3 antibodies (x20) [1]. - non-treated control cells and cells exposed to either gDNA or gDNA^OX for 2 hours [2]. - cells pre-treated for 30 min by 0.15mM NAC, then exposed to either gDNA or gDNA^OX for 2 hours.

С [1] - evidence for nuclear localization of STAT3 (x100), the nuclei were stained with DAPI [2]. - to evaluate the background, the cells were treated with normal rabbit IgG and FITC-conjugated secondary antibodies.

https://doi.org/10.1371/journal.pone.0077469.g011

After 2 hours of exposure to either gDNA^OX or gDNA, the amounts of STAT3 increase, with no changes in its localization. In 24 hours, the amounts of STAT3 protein start to decrease and in 48 hours after the addition of DNA, samples reach their initial levels (Figure 11A[2]). In the case of exposure to gDNA^OX, these effects are more pronounced than in the case of gDNA. The pre-treatment with antioxidant NAC at 0.15mM for 30 minutes before addition of same amount of oxidized DNA prevented activation of STAT3.

Both gDNA^OX- and gDNA-induced activation of NF-κB and STAT3 leads to an increase in the expression levels of genes encoding components of MAPK and JNK/p38 pathway: FOS, JUN and MAPK8 (JNK1). In parallel, we observed an increase in the expression of genes that encode soluble cytokines (Table 1). For IL10, IL6, IL8 and TNFa, the levels of mRNA increase 1.8-5.3 folds; two hours after adding DNA sample to the media, in gDNA^OX- treated MCF-7 cells, the levels of these mRNAs are 2-3 times higher than those in cells treated with gDNA. Additionally, we observed expression stimulating effects of gDNA^OX on cell adhesion and migration molecules ICAM1, PECAM1, SELE, SELP, VCAM1,and RHOA, growth factor encoding genes VEGFA, BMP4 and BMP2 and pluripotent stem cell-related genes NANOG, OCT4 and GATA-4 (Table 1).

Cell culture

ER/PR-positive MCF-7 breast cancer cells were purchased at ATCC, Manassas, USA (Cat: HTB-22). Human embryonic lung fibroblasts were retrieved from the biospecimen collection maintained by the Research Centre for Medical Genetics, Russian Academy of Medical Sciences collection and grown as described in [7]. Ethical approval for the use of primary human cells was obtained from the Committee for Medical and Health Research Ethics of Research Centre for Medical Genetics, Russian Academy of Medical Sciences (2012, approval number 5).

MCF-7 cells were cultured in DMEM medium supplemented with 10% (v/v) fetal calf serum, 2 mM L-glutamine, 100 units/mL penicillin, and 100 μg/mL of streptomycin. Cells were grown in a humidified atmosphere with 5% CO2 in air at 37°C. Before treatment with DNA probes, cells were grown for 24 h or 72 h in slide flasks.

Flow cytomery

Before flow cytometry, cells were washed in Versene solution, than treated with 0.25 % trypsin under control of light microscopic observation. Cells were transferred to the Eppendorf tubes, washed with culture media, then centrifuged and resuspended in PBS. Staining of the cells with various antibodies was performed as described below. Briefly, to fix the cells, the paraformaldehyde (Sigma) was added at a final concentration of 2 % at 37°C for 10 min. Cells were washed three times with 0.5% BSA-PBS and permeabilized with 0.1% Triton X-100 (Sigma) in PBS for 15 min or with 70% ethanol at 4°C. Cells (~ 50 x 10³) were washed three times with 0.5% BSA-PBS and stained with 1 - 2 μg/mL FITC-γH2AX (Ser139) antibody (Temecula California), FITC-Ki-67 antibody, PCNA, 8-oxodG, EEA1, AIM2, TLR9, NRF2, NF-κB (p65) , S529 NF-κB (p65) and STAT3 antibodies (Abcam) for 3 h at 4°C, then again washed thrice with 0.5% BSA-PBS and stained with 1 μg/mL secondary FITC-conjugated or PE-conjugated antibodies (Abcam) for 1 h at 4°C. To quantify intracellular DNA, cells were treated with propidium iodide and RNAase A. To quantify the background fluorescence, we stained a portion of the cells with secondary FITC(PE)-conjugated antibodies only. Cells were analyzed at CyFlow Space (Partec, Germany).

Fluorescent microscopy

Cell images were obtained using the AxioScope A1 microscope (Carl Zeiss).

Extraction of the DNA fragments from the cells or the cell-free media

To extract extracellular DNA, cells were removed from the media by centrifugation at 460 x g, followed by mixing of 3 mL of the media with 0.3 mL of the solution containing 1% sodium lauryl sarcosylate, 0.02 M EDTA, and 75 μg/mL RNAse A (Sigma, USA), incubation for 45 min, then the 24-h treatment with proteinase K (200 μg/mL, Promega, USA) at 37°C. Intact gDNA was extracted from primary human embryonic fibroblasts (HEFs) [7]. To extract genomic DNA, cells separated, and the DNA was extracted form lysed cells. After two cycles of the purification with saturated phenolic solution, DNA fragments were precipitated by adding two volumes of ethanol in the presence of 2M ammonium acetate. The precipitate was then washed with 75% ethanol twice, then dried and dissolved in water. The concentration of DNA was determined by measuring fluorescence intensity after DNA staining with the RiboGreen (Molecular Probes/Invitrogen, CA, USA). Mean size of untreated gDNA fragments was 30 kb. To match gDNA and gDNA^OX samples in its mean size, gDNA was hydrolyzed by DNAse I until size distribution of its fragments became from 0.2 to 15 kb.

Generation of the DNA samples

Quantification of mRNA levels

Total mRNA was isolated from cells using RNeasy Mini kit (Qiagen, Germany). After the treatment with DNAse I, RNA samples were reverse transcribed by Reverse Transcriptase kit (Sileks, Russia). The expression profiles were obtained using quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) with SYBRgreen PCR MasterMix (Applied Biosystems). Three housekeeping genes, ACTB, GADPH and TBP, were evaluated as possible reference genes in MCF-7 exposed to oxidized DNA. An expression of TBP was found the most stable and the employed as reference standard in further experiments. The mRNA levels were analyzed in several independent experiments using the StepOne Plus (Applied Biosystems); the technical error (%CV) was approximately 2%. All PCR products were run in the polyacrylamide gel (PAGE) to confirm their size. The following primers were used (Sintol, Russia):

AIM2 (F: CAGAAATGATGTCGCAAAGCAA, R: TCAGTACCATAACTGGCAAACAG)

BCL2 (F:GCCTTCTTTGAGTTCGGTGG, R: ATCTCCCGGTTGACGCTCT)

BCL2A1 (Bfl-1/A1) (F:TACAGGCTGGCTCAGGACTAT, R: CGCAACATTTTGTAGCACTCTG)

BCL2L1 (BCL-X) (F:CGACGAGTTTGAACTGCGGTA, R: GGGATGTCAGGTCACTGAATG)

BIRC3 (c-IAP1) (F:AAGCTACCTCTCAGCCTACTTT, R: CCACTGTTTTCTGTACCCGGA)

BMP2 (F:ACTACCAGAAACGAGTGGGAA, R: CATCTGTTCTCGGAAAACCTGAA)

BMP4 (F:AAAGTCGCCGAGATTCAGGG, R: GACGGCACTCTTGCTAGGC)

BRCA1 (F:TGTGAGGCACCTGTGGTGA, R: CAGCTCCTGGCACTGGTAGAG)

CDKN2A (p16INK4) (F:ATGGAGCCTTCGGCTGACT, R: TAACTATTCGGTGCGTTGGG)

CDKN1A (p21CIP1/WAF1) (F:GGAAGACCATGTGGACCTGT, R: ATGCCCAGCACTCTTAGGAA)

FOS (F:GGGGCAAGGTGGAACAGTTAT, R: CCGCTTGGAGTGTATCAGTCA)

GATA-4 (F:GCCCAAGAACCTGAATAAATCTAAG, R: AGACATCGCACTGACTGAGAACGTC)

ICAM1 (F:CGTGCCGCACTGAACTGGAC, R: CCTCACACTTCACTGTCACCT)

IL10 (F:AAGGCGCATGTGAACTCCC, R: ACGGCCTTGCTCTTGTTTTC)

IL6 (F:AAATTCGGTACATCCTCGACGGCA, R: AGTGCCTCTTTGCTGCTTTCACAC)

IL8 (F:ACTGAGAGTGATTGAGAGTGGAC, R: AACCCTCTGCACCCAGTTTTC)

JUN (F:TCCAAGTGCCGAAAAAGGAAG, R: CGAGTTCTGAGCTTTCAAGGT)

KEAP1 (F:GTGGTGTCCATTGAGGGTATCC, R : GCTCAGCGAAGTTGGCGAT)

MAP4K4 (F:GAGCCACAGGTACAGTGGTC, R: AAGCCTTTTGGGTAGGGTCAG)

MAPK8 (JNK1) (F:AGAAGCTAAGCCGACCATTTC, R: TCTAGGGATTTCTGTGGTGTGA)

MYD88 (F: GGCTGCTCTCAACATGCGA, R: TGTCCGCACGTTCAAGAACA);

NANOG (F:GCTGAGATGCCTCACACGGAG, R: TCTGTTTCTTGACTGGGACCTTGTC);

NFKB1(F:CAGATGGCCCATACCTTCAAAT, R: CGGAAACGAAATCCTCTCTGTT);

NRF2 (NFE2L2) (F:TCCAGTCAGAAACCAGTGGAT, R: GAATGTCTGCGCCAAAAGCTG);

OCT4 (F:TGGAGAAGGAGAAGCTGGAGCAAAA, R: GGCAGATGGTCGTTTGGCTGAATA);

PECAM1 (F:CCAAGGTGGGATCGTGAGG, R: TCGGAAGGATAAAACGCGGTC);

RHOA (F:TGGAAAGACATGCTTGCTCAT, R: GCCTCAGGCGATCATAATCTTC);

RIG1 (F:GAGATTTTCCGCCTTGGCTAT, R: CCGTTTCACCTCTGCACTGTT);

SELE (F:CAGCAAAGGTACACACACCTG, R: CAGACCCACACATTGTTGACTT);

SELP (F:CAGACCACTCAACCAGCAG, R: GGCCGTCAGTCGAGTTGTC);

STAT3 (F:GGGTGGAGAAGGACATCAGCGGTAA, R: GCCGACAATACTTTCCGAATGC);

STAT6 (F:GTTCCGCCACTTGCCAATG, R: TGGATCTCCCCTACTCGGTG);

STING (F: CCAGAGCACACTCTCCGGTA, R: CGCATTTGGGAGGGAGTAGTA);

TIRAP (F:ATGGTGGCTTTCGTCAAGTCA, R: TCAGATACTGTAGCTGAATCCCG);

TLR9 (F: CCCACCTGTCACTCAAGTACA, R: GTGGCTGAAGGTATCGGGATG);

TP53 (F:TTTGGGTCTTTGAACCCTTG, R: CCACAACAAAACACCAGTGC);

TNFa (F: CAGCCTCTTCTCCTTCCTGAT, R: GCCAGAGGGCTGATTAGAGA);

VCAM1 (F:GGGAAGCCGATCACAGTCAAG, R: AAATTCGGTACATCCTCGACGGCA);

VEGFA (F:AGGCCAGCACATAGGAGAGA, R: TTTCTTGCGCTTTCGTTTTT);

TBP (reference gene) (F: GCCCGAAACGCCGAATAT, R: CCGTGGTTCGTGGCTCTCT).

Blocking ROS

Some experiments were supplemented with controls exposed to both oxidized DNA and antioxidant N-acetyl-cysteine (NAC) at 0.15 mM. In these cases, NAC was added to the media 30 minutes before exposure to DNA.

Statistics

All reported results were reproduced at least three times as independent biological replicates. In FACS, the mean values of signal intensities were analyzed. The Figures show the average data and the standard deviation (SD). The significance of the observed differences was analyzed using non-parametric Mann-Whitney U-tests. P-values < 0.05 were considered statistically significant и marked at Figures with (*). Data were analyzed with StatPlus2007 Professional software (http://www.analystsoft.com).