Histone deacetylase inhibitors (HDACi) are a class of putative chemotherapeutic agents for which the mechanism of toxicity has not been fully identified. To explore the possibility that HDACi are genotoxic, human TK6 lymphoblastoid cells were exposed to trichostatin A (TSA) and genetic damage was measured. TSA caused a dose-dependent increase of G1-arrested cells at 24 h that correlated with increasing levels of p21 and apoptosis. Significantly elevated frequencies of structural chromosomal aberrations in cells exposed to TSA were observed using both the kinetochore-antibody micronucleus assay and nonbanding metaphase chromosome analysis. Increased tail intensities, indicative of elevated levels of DNA damage, were observed using the alkaline comet assay. Elevated levels of phosphorylated histone γH2AX protein were observed as early as 3 h following TSA exposure and peaked at 12 h for 200nM TSA. Significant levels of aneuploidy at the 200nM TSA dose were observed using metaphase analysis, but interestingly, kinetochore-positive micronuclei were not detected at any dose using the kinetochore micronucleus assay, suggesting that TSA induces aneuploidy via a nondisjunction event rather than chromosome lagging. Increases in chromosomal loss and breakage were observed using simultaneous FISH metaphase analysis of chromosomes 5, 7, 8, and 21, consistent with data obtained from the micronucleus and metaphase chromosome analyses. We conclude that TSA is both a clastogen and aneugen in the TK6 cell line and propose that the observed cytostatic and apoptotic properties of TSA may partially be due to this genotoxicity.
Histone deacetylase inhibitors (HDACi) are a class of putative chemotherapeutic agents that induce cell cycle arrest, apoptosis, and differentiation in a variety of cancer cell lines (Minucci and Pelicci, 2006). These characteristics, along with the fact that many HDACi synergistically enhance cell death to DNA-damaging agents, have pushed them forward as lead cancer treatment compounds with many entering, or currently undergoing, clinical trials (Minucci and Pelicci, 2006). The mechanisms by which HDACi elicit this cytotoxicity, however, is not completely understood (Drummond et al., 2005) and continues to be a topic of investigation. One common explanation is that HDACi alter transcription and that cellular cytotoxicity is just a by-product of these transcriptional changes. While this hypothesis remains to be verified (Johnstone and Licht, 2003), it is supported by the observation that tumor suppressor proteins, like p21, increase following chemical exposure to HDACi (Blagosklonny et al., 2002). Another hypothesis is that HDACi-induced cytotoxicity is a consequence of heterochromatic hyperacetylation and genotoxicity (Johnstone and Licht, 2003). It is theorized that the proteins normally associated with the heterochromatin are released upon decondensation, giving rise to chromosome missegregation and subsequent aberrant mitosis (Johnstone and Licht, 2003; Warrener et al., 2003). The finding that the mutation of HDACs in fission yeast reduces the fidelity of chromosome segregation supports this hypothesis (Grewal et al., 1998). Interestingly, this phenotype in fission yeast is mimicked by exposure to the HDACi trichostatin A (TSA) (Ekwall et al., 1997; Grewal et al., 1998).
Despite the possibility that HDACi may be genotoxic, a thorough investigation of their ability to cause chromosomal damage has not been conducted to our knowledge. To address this need, human TK6 lymphoblastoid cells were exposed to the classic HDACi TSA, which shares structural similarity to the clinically relevant HDACi suberoylanilide hydroxamic acid (SAHA). The studies presented here demonstrate that TSA has both clastogenic and aneugenic properties. Using multiple approaches, including the kinetochore-micronucleus assay, metaphase chromosome analysis (both with and without fluorescence in situ hybridization [FISH]), single-cell gel electrophoresis, and phosphorylated histone γH2AX protein analysis, we demonstrate that exposure to TSA causes chromosome breakage in this cell line. Increases in aneuploidy are observed using metaphase and FISH chromosome analysis, but not with the kinetochore-micronucleus assay, suggesting that chromosome missegregation occurs via a nondisjunction event. In addition, TSA causes dose-dependent increases in G1 cell cycle arrest and apoptosis with accompanying increases of p21 but not caspase 3 protein levels. We conclude that TSA is both a clastogen and aneugen in the TK6 cell line and posit that the observed cytostatic and apoptotic characteristics of TSA may partially be due to this genotoxicity.
MATERIALS AND METHODS
Cell culture and chemical treatments.
The human TK6 lymphoblastoid cell line (ATCC, Manassas, VA) was chosen because HDACi have been proposed for the treatment of lymphomas and leukemias. The cell line was maintained in RPMI1640 medium (GIBCO, San Diego, CA) containing 10% fetal bovine serum (Omega Scientific, Tarzana, CA) and 1% penicillin and streptomycin (Omega Scientific) under standard conditions. TSA (CAS# 58880-19-6, MW 302.37, > 99% purity; Sigma Aldrich, St Louis, MO) was dissolved in dimethylsulfoxide for all experiments. Cells were dosed with 0.1% DMSO and 25, 50, 100, 200, or 400nM TSA at a concentration of 3.5–4 × 105 cells/ml.
Cell cycle analysis and apoptosis assay.
For cell cycle analysis, asynchronous TK6 cells were exposed to TSA for 24 h and 1 × 105 cells were isolated from culture and washed twice with ice-cold phosphate-buffered saline (PBS) and subsequently fixed for 15 min in 1% paraformaldehyde/PBS on ice, washed twice with PBS, and resuspended in 1 ml of propidium iodide (PI) solution containing 2.5 μg/ml PI in PBS and 25 μl RNAse A (1 mg/ml) and stained overnight at 4°C. For analysis of apoptosis, TK6 cells were exposed to TSA for 3, 6, and 24 h and subsequently resuspended in fresh medium lacking TSA for another 21 h. This protocol was chosen specifically to mimic the exposures used for the various genotoxicity assays (described below). TK6 cells were prepared for flow cytometric annexin V analysis following the manufacturer's protocol (BD Biosciences Pharmingen, San Diego, CA), except the cells were not washed several times with ice-cold PBS as it increased background necrosis and apoptosis levels. At least 1 × 104 TK6 cells were analyzed on a Beckman Coulter EPICS XL-MCL flow cytometer using System II software for each experiment and repeated three times.
Doses above 200nM TSA were not used for any of the genotoxicity studies due to excessive apoptosis observed at the 24- or 45-h time points in the Annexin V assay. The procedure for the cytokinesis blocked in vitro micronucleus assay was performed as previously described (Eastmond and Tucker, 1989) with minor modifications. Cells were exposed for 20 h and were prepared for the analysis of micronuclei following a 24-h recovery period. A total of 2000 binucleated cells were analyzed per experiment for the presence of micronuclei containing the calcinosis, Raynaud phenomenon, esophageal dysmotility, sclerodactyly, and telangiectasia (CREST) syndrome, kinetochore antibody CREST-kinetochore antibody. Metaphase chromosome preparation and multiple probe FISH was performed as previously described (Smith et al., 1998; Zhang et al., 1998) with minor modifications enabling for the simultaneous analysis of chromosomes 5, 7, 8, and 21. These specific chromosomes were chosen because previous studies had demonstrated that they are selectively lost and gained upon chemical exposure to benzene and its metabolites benzoquinone and hydroquinone (Smith et al., 1998; Zhang et al., 1998). Cells were exposed to TSA for 20 h and prepared for metaphase analysis following a 24-h recovery period. A total of 500 metaphases were analyzed for the presence of structural and numerical chromosomal aberrations for each experiment (total of two experiments) whereas 1000 metaphases were analyzed using multiple probe FISH for only one experiment. The alkaline comet assay was performed as previously described (Singh et al., 1988) with some modifications. Cells were exposed to TSA for 3, 6, and 24 h prior to preparation for analysis of DNA damage in the comet assay. Five hundred randomly chosen cells per slide were scanned and analyzed with CometScan imaging software (MetaSystems, Germany). Cells were subsequently screened manually to exclude cells that did not meet stringent requirements (i.e. poor staining, out of focus, and oddly shaped). The average ± standard error of the mean for tail intensity, a measure of total DNA damage (Collins, 2002), from two independent experiments is shown for the 3- and 24-h time periods but for only one experiment at the 6-h time point. All slides for all assays were coded to prevent observer bias.
Total cell lysates were collected from 1.5 to 2 × 106 cells using 100 μl of radioimmunoprecipitation assay lysis buffer. Protein concentrations were determined by the Bradford assay (Bio-Rad, Hercules, CA). Equal protein amounts (40 μg) were resolved by polyacrylamide gel electrophoresis, transferred onto nitrocellulose membranes, and immunoblotted for p21, caspase 3, and γH2AX (Upstate, Charlottesville, VA). Proteins were visualized using the enhanced chemiluminescence (ECL) method per manufacturer's protocol (Amersham Biosciences, United Kingdom). merFilm (Kodak BioMax XAR; Eastman Kodak, Rochester, NY) was exposed and developed using the Konica SRX-101 developer (Wayne, NJ). Caspase 3 protein levels did not change following TSA exposure and was used as the loading control for all experiments. Scion Image software (http://www.scioncorp.com/Frederick, MD) was used to quantify phosphorylated histone γH2AX and caspase 3 protein levels.
Ordinary least squares was used to estimate the effect for all analyses—for those analyses with different numbers of cells evaluated for each observation, we used simply the number of cells as an analytic weight. For each regression we also included categorical “experiment” as a variable to adjust for experimental variability. For trend test, dose was entered as a continuous variable; we also did an analysis of categorical dose, entering the appropriate indicator variables for each dose, except 0 (baseline). For all analyses, the p values reported are from robust standard errors (MacKinnon and White, 1985), which relaxes the constant residual variance assumptions of the standard inference returned by ordinary least squares.
TSA Induces Apoptosis and G1 Cell Cycle Arrest in TK6 Cells
Flow cytometric analysis measuring annexin V binding to phosphatidylserine demonstrated that apoptosis levels significantly increased (p values < 0.01) at the 24- and 45-h time periods for cells exposed to 400nM TSA (Fig. 1A). Increases in apoptosis were also observed with the 200nM TSA dose at the 45-h time point but were not statistically significant (p value = 0.37). Flow cytometric analysis measuring PI staining of DNA demonstrated that 24 h TSA exposures induced a significant (trend test p value < 0.001) dose-dependent G1 cell cycle arrest that coincided with decreasing levels of G2/M cells (Fig. 1B). TSA-induced apoptosis and G1 cell cycle arrest coincided with increasing levels of p21 protein levels (Fig. 1C). Doses of 200 and 400nM TSA caused noticeable increases in p21 protein levels as early as 3 h following exposure, peaking at 12 h, and decreasing at the 24-h time point. Caspase 3 protein levels for the 200 and 400nM TSA doses did not change throughout the 24-h time period analyzed (Fig. 1C).
TSA Causes Chromosomal Breakage in the TK6 Cell Line
The kinetochore antibody micronucleus assay can be used to discriminate between micronuclei that have been formed through either chromosomal loss or breakage. Micronuclei staining negative will have been formed through a chromosome breakage (clastogenic) event whereas micronuclei staining positive are formed through chromosomal loss, although they may also contain chromosome fragments. Kinetochore-antibody micronucleus assay analysis demonstrated that TSA caused significant increases in chromosomal breakage (micronuclei that do not stain positive for the CREST antibody; trend test p value < 0.001) in the TK6 cell line (Fig. 2A). Significant increases of kinetochore-negative micronuclei were observed for both the 50 and 200nM doses (p values = 0.01 and < 0.001, respectively). Further studies utilizing a second method based upon direct metaphase chromosome analysis were conducted in order to confirm these initial findings. Metaphase analysis revealed that significant increases in structural chromosomal aberrations (breaks, gaps, and rearrangements; trend test p value = 0.016) occurred following treatment with TSA (Fig. 2A), mimicking what was observed using the micronucleus assay. The trend test for SCA was also similarly significant when gaps were excluded (p = 0.015). Simultaneous analysis of chromosomes 5, 7, 8, and 21 using multiple probe FISH also demonstrated that TSA caused increases in chromosomal breakage of these chromosomes (Fig. 2B), providing additional evidence that TSA is a clastogen in the TK6 cell line. Elevated levels of phosphorylated histone γH2AX, a protein that localizes to the site of a DNA double-strand break, regardless of the origin (Pilch et al., 2003), were observed following exposure to both 200 and 400nM TSA (Fig. 2C). While both doses induced noticeable increases in phosphorylated γH2AX as early as 3 h following exposure, the 200nM dose appeared to peak at 12 h whereas the 400nM dose appeared to cause phosphorylated γH2AX to peak at or after 24 h (Fig. 2C).
TSA Causes DNA Breaks
Comet assay analysis of TK6 cells exposed to TSA for 3, 6, and 24 h demonstrated that TSA increased DNA damage in a dose- and time-specific manner (Fig. 3A). A significant increase in tail intensity was observed at the 6-h time point (trend test p value = 0.02) but was not observed at any other time points. Tail intensities for the 200nM TSA dose increased at the 6-h time point but decreased at 24 h whereas the 25 and 50nM doses had increased tail intensities at the 24-h time point but not at the 3 or 6 h (Fig. 3A). Differences between studies were observed for the same doses using comet analysis and, as such, are cautiously interpreted. Representative images of damaged and nondamaged TK6 cells are presented in Figure 3B. Quantitative analysis of phosphorylated histone γH2AX levels of TK6 cells exposed to 200nM TSA (Fig. 3C) demonstrated that protein levels fluctuated in a time-dependent manner closely mimicking the levels of DNA damage identified with comet analysis (Fig. 3A).
TSA Causes Aneuploidy
Analysis of TSA exposed TK6 cells with the kinetochore-antibody micronucleus assay failed to identify increases in chromosomal loss (Fig. 4A). These data contradict the findings obtained through direct metaphase chromosome analysis (Fig. 4A), which identified significant increases of chromosomal loss (trend test p value = 0.02 and a p value of 0.03 for the 200nM TSA dose). Further analysis using multiprobe FISH to simultaneously evaluate chromosomes 5, 7, 8, and 21 also demonstrated that aneuploidy of these chromosomes increases following 200nM TSA exposure (Fig. 4B).
Histone deacetylase inhibitors induce cytotoxicity in a variety of cancer cell lines through a mechanism that has yet to be elucidated. The possibility that these cytostatic and apoptotic effects may be due to their ability to induce genotoxicity has not yet been fully addressed. We studied the effects of the hydroxamic acid TSA on the human TK6 lymphoblastoid cell line to explore the possibility that HDACi-induced cytotoxicity may partially be explained by DNA breakage and chromosome damage. Using multiple approaches, including protein analysis and several standard genetic toxicology assays, we demonstrate that TSA has both clastogenic and aneugenic properties in the TK6 cell line. While the primary focus of the current study was to identify whether TSA is genotoxic, a secondary aim was to identify if the TK6 cell line exhibited cytostatic effects similar to those observed in other cancer cell lines following TSA exposure. We identified that TSA elevated p21 protein levels, caused G1 cell cycle arrest, and induced apoptosis in a time- and dose-dependent manner, findings consistent with previous studies in other cell lines (Blagosklonny et al., 2002; Fandy et al., 2005). Caspase 3 protein levels did not change following either the 200 or 400nM TSA 24-h exposures, but were not investigated at longer time points or using higher doses. Therefore, it would be premature to declare that this result is indicative of a caspase 3–independent apoptotic mechanism in the TK6 cell line. Another, more likely scenario, is that the apoptosis, while sufficient to be observed by flow cytometric analysis of annexin V, may not have been high enough to be observed by caspase 3 protein analysis. Thus, we cannot currently rule out the possibility that apoptosis in the TK6 cell line occurs through a caspase 3–mediated mechanism following TSA exposure.
TSA has previously been identified to cause chromosome missegregation in fission yeast (Ekwall et al., 1997; Grewal et al., 1998) through a mechanism that likely involves disruption of protein localization at centromeric heterochromatin (Ekwall et al., 1997). The current model suggests that histone deacetylation is a necessary step preceding RNAi-directed heterochromatic methylation and subsequent cohesin recruitment and binding (Ekwall, 2004). Inhibition of the initial deacetylation step will thus block a series of ensuing events necessary for proper chromosome segregation. Because many basic biological processes are highly conserved between yeast and mammals, it is possible that HDACs also play a critical role in chromosome segregation in human cells. Data presented in this study support this hypothesis as we observed increased levels of aneuploidy in human TK6 cells exposed to TSA. Using direct metaphase chromosome analysis with and without multiprobe FISH, we identified elevated levels of aneuploidy with a threshold dose of 200nM following 24 h of exposure. Interestingly, we failed to identify significant increases in aneuploidy at any dose using the kinetochore-micronucleus assay. This discrepancy can likely be attributed to the fact that the micronucleus assay cannot detect nondisjunction events as the aberrant chromosome will segregate to one of the two daughter nuclei during mitosis and will not form a micronucleus. Direct metaphase chromosome analysis circumvents this problem because total numbers of chromosomes are enumerated. Our findings in the TK6 cell line, supported by several other reports in other mammalian cell lines (Cimini et al., 2003; Taddei et al., 2001), indicate that TSA induces chromosomal missegregation. While the understanding of the mechanism by which this HDACi-induced chromosomal loss occurs in mammalian cells lags behind that for the fission yeast, there is nonetheless accumulating evidence to indicate that the HDACi TSA is an aneugen.
We identified that TSA caused DNA damage using the kinetochore micronucleus assay, metaphase chromosome analysis (with and without FISH), alkaline comet assay, and phosphorylated histone γH2AX protein quantification by western blotting. Significant increases in kinetochore-negative micronuclei were observed at both the 50 and 200nM doses using the micronucleus assay with very similar results being observed with metaphase chromosome analysis (Fig. 2A). Additional cytogenetic studies using multiple probe FISH supported these initial findings and indicated that TK6 cells exposed to TSA had elevated levels of damage in chromosomes 5, 7, 8, and 21 (Fig. 2B). No threshold effect was observed using the micronucleus assay, but appears to be present above the 100nM TSA dose using metaphase analysis both with and without FISH analysis. Western blot analysis from three separate experiments demonstrated that increases in phosphorylated histone γH2AX protein levels occurred following 200 and 400nM TSA exposures (Fig. 2C). Analysis of the TSA exposed TK6 cells was also performed using the alkaline comet assay to ensure further the validity of these results. Although the results from the comet assay are not as definitive as those obtained from the other assays, increased levels of DNA damage are observed and correlate tightly with the γH2AX protein levels (Fig. 3A and 3C). The comet assay results were initially puzzling as the 25- and 50nM doses only have elevated levels of damage at the 24-h time point and far exceed the amount of damage detected at the same time in the 200nM or 15μM etoposide-positive control (data not shown). The observation that the γH2AX protein levels change in a temporal manner very similar to the amount of DNA damage observed with the comet assay indicates that the time- and dose-specific results observed are not likely due to an artifact with the assay. We consider that variations may be due to a combination of cell cycle kinetics and chromosome decondensation that is dose specific.
Previous studies in Tetrahymena thermophila and mouse embryonic fibroblasts have failed to identify that TSA causes chromosome breakage (Duharcourt and Yao, 2002; Yaneva et al., 2005), results that contrast with the findings presented here using the human TK6 cell line. These data raise the question of whether the clastogenic effect of TSA is specific to TK6 cells. As previously described, the TK6 cells responded in a manner similar to other cell lines exposed to TSA, exhibiting G1 cell cycle arrest, increased p21 protein levels, apoptosis, and chromosome missegregation. Additionally, TSA has been reported to alter chromosome and interphase chromatin condensation in HeLa and PtK1 cells (Cimini et al., 2003; Toth et al., 2004), results that provide a possible mechanism for our findings that TSA causes chromosome breakage. We thus theorize that it is unlikely that the clastogenic effects of TSA will be limited only to the TK6 cell line. Furthermore, these results highlight the possibility that other HDACi may also cause DNA and chromosome damage as TSA is often considered the model HDACi.
Using a variety of approaches, we present data demonstrating that the HDACi TSA is both a clastogen and an aneugen. Because these effects are observed to occur simultaneously with p21 induction and precede the induction of apoptosis, we propose that the cytotoxicity of TSA may partially be due to its genotoxicity. Further work is required to elucidate the mechanism of TSA-induced genotoxicity and identify whether this, and/or its effects on transcription, plays more of a role in its cytotoxicity.
Supplementary data are available online at http://toxsci.oxfordjournals.org/.
The authors would like to thank Cliona McHale, Randi Paynter, and Noe Galvan for their invaluable discussions and technical expertise.
- chromosome abnormality
- antineoplastic agents
- cell line
- comet assay
- dna damage
- mitotic metaphase
- micronucleus tests
- nondisjunction, genetic
- chromosome analysis
- chromosome loss
- toxic actions
- histone deacetylase inhibitor