Combined experimental and computational analysis of DNA damage signaling reveals context‐dependent roles for Erk in apoptosis and G1/S arrest after genotoxic stress
2012; Springer Nature; Volume: 8; Issue: 1 Linguagem: Inglês
10.1038/msb.2012.1
ISSN1744-4292
AutoresAndrea R. Tentner, Michael J. Lee, Gerry J Ostheimer, Leona D. Samson, Douglas A. Lauffenburger, Michael B. Yaffe,
Tópico(s)Genomics, phytochemicals, and oxidative stress
ResumoArticle31 January 2012Open Access Source Data Combined experimental and computational analysis of DNA damage signaling reveals context-dependent roles for Erk in apoptosis and G1/S arrest after genotoxic stress Andrea R Tentner Andrea R Tentner Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Michael J Lee Michael J Lee Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Gerry J Ostheimer Gerry J Ostheimer Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Leona D Samson Leona D Samson Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Douglas A Lauffenburger Douglas A Lauffenburger Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Michael B Yaffe Corresponding Author Michael B Yaffe Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Andrea R Tentner Andrea R Tentner Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Michael J Lee Michael J Lee Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Gerry J Ostheimer Gerry J Ostheimer Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Leona D Samson Leona D Samson Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Douglas A Lauffenburger Douglas A Lauffenburger Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Michael B Yaffe Corresponding Author Michael B Yaffe Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA Search for more papers by this author Author Information Andrea R Tentner1, Michael J Lee1, Gerry J Ostheimer1, Leona D Samson1, Douglas A Lauffenburger1 and Michael B Yaffe 1 1Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, Cambridge, MA, USA *Corresponding author. Departments of Biology and Biological Engineering, David H Koch Institute for Integrative Cancer Research at MIT, Massachusetts Institute of Technology, 77 Massachusetts Avenue, 76-353, Cambridge, MA 02139, USA. Tel.: +1 617 452 2103; Fax: +1 617 452 4978; E-mail: [email protected] Molecular Systems Biology (2012)8:568https://doi.org/10.1038/msb.2012.1 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info Following DNA damage, cells display complex multi-pathway signaling dynamics that connect cell-cycle arrest and DNA repair in G1, S, or G2/M phase with phenotypic fate decisions made between survival, cell-cycle re-entry and proliferation, permanent cell-cycle arrest, or cell death. How these phenotypic fate decisions are determined remains poorly understood, but must derive from integrating genotoxic stress signals together with inputs from the local microenvironment. To investigate this in a systematic manner, we undertook a quantitative time-resolved cell signaling and phenotypic response study in U2OS cells receiving doxorubicin-induced DNA damage in the presence or absence of TNFα co-treatment; we measured key nodes in a broad set of DNA damage signal transduction pathways along with apoptotic death and cell-cycle regulatory responses. Two relational modeling approaches were then used to identify network-level relationships between signals and cell phenotypic events: a partial least squares regression approach and a complementary new technique which we term 'time-interval stepwise regression.' Taken together, the results from these analysis methods revealed complex, cytokine-modulated inter-relationships among multiple signaling pathways following DNA damage, and identified an unexpected context-dependent role for Erk in both G1/S arrest and apoptotic cell death following treatment with this commonly used clinical chemotherapeutic drug. Synopsis Data-driven modeling was used to analyze the complex signaling dynamics that connect DNA repair with cell survival, cell-cycle arrest, or apoptosis. This analysis revealed an unexpected role for Erk in G1/S arrest and apoptotic cell death following The cell-fate choice between apoptosis and cell-cycle arrest after doxorubicin treatment is dose dependent and modulated by TNF. An early, transient G1/S arrest following doxorubicin treatment acts as a cell-fate decision point between re-entry into S-phase and apoptotic cell death. Two complementary relational modeling methods identify the proliferation and survival signal, ERK, as potentially causal for the G1/S arrest and death phenotypes. Inhibitor experiments validate a role for ERK in maintenance of the early, transient arrest and in promotion of apoptosis from this arrested state following doxorubicin-induced DNA damage. Introduction The DNA damage response (DDR) constitutes an evolutionarily conserved set of signaling pathways that are essential for maintaining genomic integrity in all eukaryotic cells (Hoeijmakers, 2001; Jackson and Bartek, 2009). Mutations and/or acquired defects that compromise the function of these DDR pathways results in enhanced mutagenesis, and are thought to underlie the development and progression of cancer. (Kastan and Bartek, 2004; Halazonetis et al, 2008). Paradoxically, the defective DDR of tumor cells facilitates their killing by genotoxins, rendering these types of DNA damaging agents useful for cancer chemotherapy (Kennedy and D'Andrea, 2006; Helleday et al, 2008; Litman et al, 2008; Powell and Kachnic, 2008). A variety of chemotherapeutic agents in common clinical use, including camptothecin, etoposide, and doxorubicin, inhibit Topoisimerases I and II (Topo I/II), respectively, creating large numbers of DNA double-strand breaks (DSBs). These lesions are thought to be particularly cytotoxic, and initiate a canonical DNA damage signaling pathway through activating the upstream PI3-kinase-like kinase (PIKK) ATM (Hoeijmakers, 2001; Jackson and Bartek, 2009). Recruitment of Mre11, Rad50, and Nbs1 to the DSB results in an ATM autoactivation loop that increases and sustains ATM activity, and together with the PIKKs ATR and DNA-PK, drives the accumulation of the phosphorylated form of the histone variant H2AX (so-called γH2AX) in the area surrounding the break (Furuta et al, 2003; Pommier, 2006; Harper and Elledge, 2007; Jackson and Bartek, 2009). γH2AX is required for the efficient retention of additional signaling molecules including MDC1, RNF8, RNF168, BRCA1, along with repair factors into foci near the break, and facilitates the activation of downstream signaling pathways that mediate phenotypic responses to damage—that is, cell-cycle arrest, apoptosis, or repair and cell-cycle re-entry (Figure 1A). ATM effectors involved in DNA damage-induced cell-cycle arrest and induction of programmed cell death include the checkpoint kinase Chk2 and the multi-functional transcription factor p53. Chk2, together with the kinases Chk1 and MK2, phosphorylates and functionally inactivates Cdc25 family members, thereby blocking activation of CyclinE/A-Cdk2 and CyclinA/B-Cdk1 complexes and establishing G1/S and G2/M checkpoints (Bartek and Lukas, 2003; Donzelli and Draetta, 2003; Reinhardt et al, 2007). p53 assists in maintenance of DNA damage-mediated cell-cycle arrest through upregulation of the cyclin-dependent kinase inhibitor p21 and/or initiates programmed cell death through transactivation of pro-apoptotic Bcl-2 protein family members (Levine and Oren, 2009). Figure 1.A strategy for systematic analysis of signal transduction pathways involved in the DDR. (A) An expanded DNA damage signaling network that reports on internal cellular states and external microenvironment cues indicating the intersection and integration of multiple distinct signaling pathways that respond to or modulate chromatin integrity, cellular stress, survival, apoptosis, and cell-cycle progression or arrest. Indicated nodes and responses in this expanded DNA damage-response network were interrogated by quantitative western blotting for the total (W) or phospho-forms (pW) of signaling molecules, ELISA-based activity assay (E), or flow cytometry (F). (B) Experimental approach: cells were treated +/− doxorubicin (Dox; 0, 2, or 10 μM) and +/− TNFα (0 or 100 ng/ml) for 4 h. Four hours following treatment, the media was replaced with fresh media containing 1% FBS. Cell extracts for signal measurement were collected at 0, 0.25, 0.5, 1, 1.5, 2, 4, 8, 12, 16, and 24 h following treatment. Whole cell samples were collected and fixed for cellular-response measurements at 6, 12, 24, and 48 h following treatment. Download figure Download PowerPoint While cell fate after DNA damage is known to be regulated by these core DDR pathways, additional signaling pathways governing general stress and survival responses such as the PI3-kinase pathway, the NF-κB pathway, and the three major MAPK pathways are also likely to contribute (Figure 1A). Since extracellular signals are transduced through these pathways, they may serve as information processing junctions whereby signals from the microenvironment are integrated with the DDR. A common component of the tumor microenvironment is the inflammatory cytokine TNFα, which is present within many tumor types and appear to contribute to tumor progression (Ben-Baruch, 2003). In addition, exogenous administration of TNFα has been used to induce tumor cell death, particularly in combination with genotoxic agents (Mocellin and Nitti, 2008). The relative importance of these additional signal transduction pathways, and the manner in which their signals are integrated together with those from the canonical DDR pathways to control the fate of DNA-damaged cells, is poorly understood. Therapeutic manipulation of these pathways, however, could significantly alter and potentially improve the clinical response of tumors to DNA damaging therapies (Samuels and Ericson, 2006; Ahmed and Li, 2007; Reinhardt et al, 2007; Roberts and Der, 2007). To pursue this integrative understanding in a systematic manner, we collected quantitative time-resolved signaling data on the activation status of key nodes in several molecular signaling pathways that could potentially impact cell-cycle control and induction of cell death, together with similar data on selected components of the core DDR pathway, in an osteosarcoma cell line treated with varying doses of doxorubicin in the presence or absence of the inflammatory tumor cytokine TNFα (Szlosarek et al, 2006). In parallel, we collected quantitative time-resolved phenotypic data on the apoptotic and cell-cycle regulatory response of the cell population following DNA damage. We used computational modeling approaches to identify critical context-dependent molecular signals that modulate damage-induced cell-cycle arrest and the initiation of programmed cell death. Results An experimental treatment-response protocol for quantifying phenotypic changes and information flow within an expanded DNA damage signaling network Following DNA damage, cells display complex dynamic phenotypes that connect cell-cycle arrest in G1, S, or G2/M, and DNA repair with decisions related to survival, cell-cycle re-entry, permanent cell-cycle arrest, or cell death. How these cellular decisions are made remains poorly understood, but must involve the integration of large amounts of molecular signaling information distributed across pathways that reflect both the current internal state of the cell and the state of the surrounding microenvironment (Figure 1A). To explore this molecular signaling–phenotypic response relationship, we treated U2OS osteosarcoma cells with varying concentrations of doxorubicin, a clinically relevant anti-cancer agent that generates DNA DSBs by inhibiting Topo II, in the presence or absence of TNFα, a cytokine commonly found in tumor microenvironments (Szlosarek et al, 2006). Six specific treatment conditions were examined encompassing exposure to no (0 μM), low (2 μM), or high (10 μM) doxorubicin, in the presence or absence of 100 ng/ml of TNFα, for 4 h followed by transfer to drug-free medium without doxorubicin or TNFα. Cell lysates were prepared at 11 time points between 0 and 24 h after exposure to doxorubicin, and used in quantitative western blotting or ELISA assays to measure 17 molecular signals (Figure 1B; Supplementary Table S1). Cells were fixed at four time points between 0 and 48 h after damage for flow cytometry-based assays of cell-cycle progression and cell death. Together, these measurements provide a high-dimensional quantitative time-dependent data set that captures dynamic changes in signaling pathways and cellular responses. The cell-fate choice between apoptosis and cell-cycle arrest after doxorubicin treatment is dose dependent and modulated by TNFα In the absence of DNA damage, minimal levels of apoptosis were observed over the 48h course of the experiment, as evidenced by <5% of the cells staining positively for the apoptotic markers cleaved caspase-3 and cleaved PARP (Figure 2A, upper panels, and 2B). The addition of TNFα to undamaged cells resulted in only a minimal increase in basal levels of apoptosis. Treatment with either 2 or 10 μM doxorubicin caused a marked dose- and time-dependent increase in apoptotic cell death. The presence of TNFα dramatically increased the fraction of cells undergoing DNA damage-induced cell death, and shortened the time to apoptosis, particularly within 12 h after exposure to doxorubicin. For example, 6 h after treatment, the amount of apoptotic cell death observed in the TNFα+2 μM doxorubicin-treated samples was greater than threefold that observed in samples treated with 2 μM doxorubicin alone (Figure 2B, gray and black squares; 14 versus 4%, respectively), and instead was comparable to the amount of apoptosis observed in the doxorubicin alone-treated samples after an additional 6 h of incubation (12 h time point, 17% death). These effects were even more pronounced following treatment with a higher concentration of doxorubicin. The amount of apoptotic cell death observed 6 h following treatment with TNFα+10 μM doxorubicin was nearly sixfold greater than that observed with 10 μM doxorubicin alone (Figure 2B, gray and black diamonds, 7 versus 40%, respectively) and was comparable to the amount of apoptosis observed in the 10-μM doxorubicin-treated samples an additional 18 h later (24 h time point, 36%). Figure 2.TNF enhances dose-dependent cell death following doxorubicin-induced DNA damage with minimal affect on dose-dependent cell-cycle arrest. (A) Representative scatter plots for cleaved caspase-3 and cleaved PARP as measured simultaneously by flow cytometry in single cells, for all six treatment conditions, at 12 h following treatment. (B) Apoptosis in treated cell populations (percentage of the cell-population staining positively for both apoptotic markers cleaved caspase-3 and cleaved PARP), for all six treatment conditions investigated at 6, 12, 24, and 48 h following treatment. (C) Representative cell-cycle profiles based on PI staining of DNA content for cells exposed to 0, 2, or 10 μM doxorubicin +/− TNFα at 6, 12, 24, and 48 h following treatment. (D) Quantification of the percentage of the treated cell populations in the G1 (a), S (b), G2/M (c), and pHH3+/M (d) phase of the cell-cycle, and the sub-G1 population (e) at 6, 12, 24, and 48 h following each treatment. (B, D) Mean values from n⩾4 independent experiments; error bars denote standard error of the mean. All measurements of cell-cycle stage and apoptosis represent 'window' measurements that do not reflect the cumulative fraction of the initial population, but rather the fraction of the population that is still measurable at the indicated time point. For example, apoptosis measurements at later time points will not include cells that have died and disintegrated between the time of treatment and the time of measurement. Source data is available for this figure in the Supplementary Information. Source Data for Figure 2B [msb20121-sup-0001-SourceData-S1.xls] Source Data for Figure 2C [msb20121-sup-0002-SourceData-S2.xls] Source Data for Figure 2D [msb20121-sup-0003-SourceData-S3.xls] Download figure Download PowerPoint In contrast to the ability of TNFα to enhance DNA damage-induced cell death, this cytokine had little effect on the cell-cycle arrest observed in response to DNA damage by either low- or high-dose doxorubicin (Figure 2C, compare left and right columns, and 2D). In untreated control cells, with or without TNFα, we observed an initial slight rise in the S and G2 population, followed by a slow progressive increase in the percentage of G1 cells, and a corresponding drop in S and G2, as the cells approached confluence. After DNA damage, dramatic cell-cycle changes were observed in the surviving cell populations treated with low-dose doxorubicin, whereas more subtle cell-cycle effects were seen in the surviving cells after high-dose treatment. Both 2 and 10 μM doxorubicin caused a very slight increase in the percentage of G1 cells 6 h after treatment (Figure 2D). In cells exposed to low-dose doxorubicin, this was followed by a significant accumulation of the S-phase population at 24 h, along with a significant depletion of the G1 population (Figure 2Da and Db, black squares). By 48 h, both the G1 and S-phase populations of these 2 μM doxorubicin-treated cells were reduced from their levels at 24 h, and a significant increase in the population of G2-arrested cells was observed (Figure 2C and Dc). The addition of TNFα reduced the S-phase accumulation seen following 2 μM doxorubicin at 24 h, and reduced the percentage of G2-arrested cells at 48 h (Figure 2D, gray squares). Importantly, this was accompanied by an increase in the sub-G1 population at both of these time points. In cells treated with high-dose doxorubicin, we also observed a steady and significant depletion in the G1 population from 6 to 48 h. However, this decline was muted compared with the decline observed after low-dose doxorubicin treatment (Figure 2D, black diamonds). Similarly, the time-dependent rise in the percentage of S-phase cells at 24 h was also greatly reduced after 10 μM treatment, and no meaningful increase in the percentage of G2-arrested cells was observed over the time course of the experiment. Taken together, these observations reflect alternative cell-fate choices in response to low versus high-dose doxorubicin treatment. Less damaged cells preferentially trigger cell-cycle arrest, while cells with more extensive DNA damage activate programmed cell death. This cell-fate choice is further modulated away from cell-cycle arrest and toward cell death by the presence of TNFα. An early G1/S checkpoint constitutes a cell-fate decision point The pattern of cell-cycle progression observed in response to DNA damage induced by doxorubicin, particularly under lower-dose conditions, suggests that the damaged cells may undergo an early transient arrest at the G1/S boundary between 0–12 h, followed by a more or less synchronous cell fate transition of either release into S-phase or initiation of apoptotic cell death by 24 h. To probe this process in more detail, asynchronous U2OS cells were exposed to 2 μM doxorubicin, and cell progression analyzed by flow cytometry at more frequent intervals between 0 and 42 h. As shown in Figure 3Aa–Ae, an accumulation of G1-arrested cells was evident as early as 3 h after damage, followed by a drop in, and broadening of, the G1/S peak at 15 h and the appearance of a clear shoulder off of the G1/S peak by 18 h, indicating S-phase entry. Under these DNA damage conditions, the greatest increase in cell death occurred between 6 and 12 h (Figure 2B), when the cells appeared to be arrested in G1 (Figure 3A). Direct flow cytometry analysis of cells stained for both DNA content and cleaved caspase-3 (cC3) further confirmed that the bulk of cC3-positive cells had 2 N DNA content, suggesting an apoptotic G1 population (Figure 3B). This finding was unexpected, since it is generally believed that DSB generation, as well as sensing and processing of the protein-linked DSBs induced by Topo II poisons such as doxorubicin is strongest in S-phase, when Topo II levels are high (Goswami et al, 1996) and replication forks are likely to encounter these impassable DNA lesions (Howard et al, 1994; Hong and Kreuzer, 2000; Bartek et al, 2004; Nitiss, 2009). In agreement with this, there is also evidence that doxorubicin-induced cell death may preferentially occur during S-phase (Grdina et al, 1980; Meyn et al, 1980) Figure 3.Cell death following DNA damage occurs from cells arrested in G1 or at the G1/S boundary. (A) Representative flow cytometry profiles showing DNA content for cells exposed to 0 (black) or 2 μM (gray) doxorubicin, at time points between 3 and 42 h following treatment. Treatment with low-dose doxorubicin (gray traces) results in an accumulation in the G1 population (2 N DNA content) as compared with untreated cells (black traces) by as early as 3 h following treatment. This accumulation persists through at least 15 h. By 24 h following treatment, the accumulation in G1 is completely abrogated and an accumulation in the S-phase population is apparent, suggesting synchronous release into S-phase. (B) Representative scatter plots showing DNA content and cleaved caspase-3 in cells treated with 2 μM Dox (left) or 2 μM Dox+TNFα (right) at 24 h following treatment. The majority of cells staining positively for cleaved caspase-3 have 2 N DNA content (red box), consistent with apoptosis in G1 or at the G1/S boundary. (C) Cells were synchronized in late G1 and at the G1/S boundary via a double thymidine block and released directly into aphidicolin-containing media to block progression through S-phase. Cells were treated with or without 2 μM Dox 1 h later and analyzed 8 h later for apoptosis by flow cytometry (top). The percentage of cells staining positively for both apoptotic markers is indicated from two biological replicates in each of two independent experiments (bottom). (D) Cells treated with either 0 or 2 μM Dox were pulsed with BrdU for 4 h beginning 15 h following treatment. At 48 h following treatment, cells were analyzed by flow cytometry for DNA content and BrdU incorporation (top). Overlays of the cell cycle/BrdU distributions from cells treated with either 0 (black) or 2 μM (gray) Dox reveal an accumulation of BrdU-positive cells with 4 N DNA content in populations treated with 2 μM doxorubicin, indicating that the population of cells synchronously released from a G1 or G1/S arrest into S between 12 and 24 h following treatment is the same population that accumulates in G2/M by 48 h following treatment. No BrdU-positive cells re-entered the G1 population by 48 h after 2 μM Dox, consistent with a persistent G2/M arrest. Left and right panels are identical, with the Dox-treated cells shown either behind (left) or in front of (right) the untreated controls. (E) Cells treated with 0 or 2 μM Dox as in (D) were collected at 48 h following treatment and analyzed for DNA content and the mitotic marker, pHH3. Mitotic cells are boxed, indicating stable G2 arrest after doxorubicin treatment. (F) A model for the early G1/S decision point. Source data is available for this figure in the Supplementary Information. Source Data for Figure 3A [msb20121-sup-0004-SourceData-S4.xls] Source Data for Figure 3B [msb20121-sup-0005-SourceData-S5.xls] Source Data for Figure 3C [msb20121-sup-0006-SourceData-S6.xls] Download figure Download PowerPoint There is strong evidence that the primary mechanism of doxorubicin action is via Topo II inhibition (Nitiss, 2009). Levels of Topo II in tumors in vivo are correlated with the apoptotic response to doxorubicin, while downregulation of Topo IIa is associated with acquired resistance of tumors to this drug (Ogiso et al, 2000; Tanner et al, 2006). Importantly, however, doxorubicin is well known to cause additional types of DNA damage through Topo II-independent effects, including free radical-induced single-strand breaks, direct Doxorubicin-DNA adducts, and inter-strand crosslinks (Swift et al, 2006 and references therein). To directly address whether doxorubicin-induced DSBs were present in the G1 population, we treated cells with 2 μM doxorubicin and used flow cytometry to measure the percentage of cells staining positively for γH2AX as a function of cell-cycle stage. As shown in Supplementary Figure S1, elevated levels of γH2AX could be detected in roughly 50% of the G1 cells, a value that while large, was somewhat lower than that observed in S and G2 cells where the response rate was ∼90%. Thus, while other types of doxorubicin-induced DNA damage almost certainly exist in G1 cells, our data does indicate the presence of drug-induced DSBs during this phase of the cell cycle. Interestingly, Ju et al (2006) have reported topoisomerase-mediated generation of transient DSBs during normal gene transcription, providing a possible mechanism for enhanced DSB formation by Topo II inhibition even in the absence of DNA replication. To further explore whether cell death could occur directly from G1 or at the G1/S boundary without a requirement for replicative DNA synthesis, cells were synchronized at the G1/S border using a double thymidine block, and released into media containing the DNA polymerase inhibitor aphidicolin to block progression into S-phase. The cells were then treated with 2 μM doxorubicin 1 h later. Under these conditions, the extent of apoptosis was comparable to, or higher than that observed in the DNA-damaged asynchronous population (Figure 3C), consistent with doxorubicin-induced cell death occurring in G1 or at the G1/S border in the absence of continued DNA replication. As shown in Figure 3Ag–Aj, the subpopulation of cells in the asynchronous population treated with 2 μM doxorubicin that had not undergone apoptosis had largely progressed into mid-S-phase by 24 h, ultimately accumulating in G2 between 30 and 42 h after damage. To ensure that the increase in the G2 population at late times observed by flow cytometry represented the same population of cells that had transited through S-phase, rather than the appearance of an artifactual G2 peak from selective death and drop-out of the S-phase cells at these late time points, we pulsed the asynchronous damaged cells with BrdU at 15 h, when a portion of this population is synchronously entering S-phase, and followed the subsequent cell-cycle progression of these BrdU-labeled cells. As shown in Figure 3D, at 48 h following treatment, the 4 N DNA-containing cells are strongly BrdU+, indicating that the cell population that entered S between 15 and 19 h following treatment transited through S and into G2 by 48 h. These cells stain negatively for phospho-Histone H3 (pHH3; Figure 3E), indicative of G2 arrest. Furthermore, we found that these cells were unable to form proliferative colonies in culture, suggesting that the G2 arrest is persistent (data not shown, manuscript in preparation). Together, these data indicate that DNA-damaged cells undergo a transient arrest in G1 or at the G1/S border. This arrest serves as a decision point from which cells either die or re-enter the cell cycle by releasing synchronously into S and subsequently arresting in G2 (c.f. Figure 3F), perhaps as a consequence of defective G1 and/or S-phase checkpoints in this cancer cell line. Cells show complex patterns of signaling network activation in response to low and high amounts of DNA damage in the presence or absence of cytokine cues To explore specific signaling pathways and molecules responsible for mediating the different phenotypic responses observed in response to low- and high-dose doxorubicin, in the
Referência(s)