Drury and Kmiec [36] demonstrated the importance of target association in the gene repair reaction and suggested that the rate-limiting step is the pairing of the ssODN to the genomic site. Thus, we assume that strategies aimed at increasing correction levels might be facilitated by events that enable more efficient oligonucleotide binding. Several agents have been employed to achieve this goal including trichostatin A (TSA), which is a potent histone deacetylase inhibitor [37]. When added to cell cultures following ssODN delivery, TSA was able to increase gene repair up to 10-fold, most likely by altering the chromosome structure and enabling oligonucleotide binding [38]. Slowing down the process of DNA replication by extending S phase or synchronizing cells in S is another way to enhance the frequency of gene repair [14,18,19,39].
Our current work focused on examining this second manipulation in greater detail by studying more carefully the function of S phase in gene repair. During this phase of the cell cycle, chromatin is in a more open conformation, accessible to proteins/DNA complexes that undergo a search for homology and interact at specific sites within the genome. In this study, we confirm that incubating cells with thymidine leads to elevated gene repair levels by synchronizing the cells along the G1/S border and enhancing the total number of amenable targets when, after release, the cells cycle into S phase in the presence of the oligonucleotide [10,22]. We also show that the inherent loss of gene correction with time [12] can be counterbalanced by inhibiting the replication of the entire cell population, and preventing the "dilution" of the repaired population of cells by the unimpeded non-corrected cell population. Lastly, we present a novel finding that shows that the long term effect of thymidine treatment can result in the induction of cellular senescence.
Treatment with thymidine leads to an increased level of cellular dTTP, ultimately resulting in a feedback inhibition of ribonucleotide reductase. Subsequently, cells are starved of dCTP, due to the allosteric inhibition of the reduction of CDP [40], resulting in a slowing of DNA replication and an accumulation of cells that traverse S phase (referred to as a thymidine block [24]. By incubating DLD-1 cells in 1 mM thymidine for 24 hours, we were able to stall approximately 90% of the cell population at the G1/S border. Following the introduction of the oligonucleotide, the block was removed and the cells were allowed to enter S and G2 in the presence of the ssODN. The high repair efficiency seen with thymidine pre-treatment was most likely due to the increased number of accessible target sites that in S (3 times more cells than in the non-synchronized cells) or cycling through S within the first 8 hours following electroporation. Upon removal of thymidine, cells were able to once again replicate at high levels further indicating the increased number of cells in S phase. Thus, our data indicate that elevated levels of correction are mostly due to an increase of cellular targets in S phase, whose open chromatin structures facilitate ssODN association and ultimately gene repair events.
These results on the importance of S phase in the ssODN-coupled repair differs from the conclusions drawn in Wu et al. [14] that increasing the proportion of cells in S phase does not enhance repair levels. Our data contrast these observations by identifying S phase as the critical cell cycle stage in which gene repair occurs. We find that the highest overall gene repair was evident at 8 hours following the ssODN introduction, for both 1 mM thymidine pre-treated and non-treated cells (0.66% and 0.12%, respectively; see Additional file 1). In addition, the maximal value occurred following the 1 mM thymidine pre-treatment, coinciding with the highest percent of cells in S phase at 8 hours. When cells were pre-treated with thymidine and then sustained in thymidine for 8 hrs following ssODN electroporation, gene repair decreased to only 0.21% by this time point, concurrent with an altered cell cycle profile in which a lower percent of cells were in S phase and a higher percent remained at G1 or in early S (see Additional file 2). However, when thymidine was simply added after ssODN introduction (without a pre-treatment), gene repair levels and cell cycle profiles remained remarkably unchanged, as compared to the non-treated non-synchronized cells. Thus, our results support the aforementioned observation that highlights the importance of S phase in the gene repair reaction. What is interesting to note, however, is the fact that although the gene repair is lower when thymidine is used in both the pre- and post-treatment, its repair level remains constant from 8 to 48 hours while a decrease is observed when thymidine is removed following the addition of the ssODN. Ultimately, this change in repair frequency results in the appearance of higher levels of gene repair following a 1–2 mM thymidine treatment than a 1-0 mM at 48 hours following ssODN introduction (see Figure 4, bar a and c).
Recently, it was reported that the repair frequency decreases as a function of recovery time following ssODN introduction, and this lack of maintenance of the repair frequency is due to the selective recovery and proliferation of non-corrected cells [12,22]. Although it is unclear whether the corrected cells are arrested in late S phase [22] or at the G2/M border [12], it is apparent that the single-stranded DNA oligonucleotide activates a damage response pathway through the upregulation and autophosphorylation of ATM [11,22,41]. This activation, in turn, leads to a phosphorylation of the downstream checkpoint kinases, Chk1 and Chk2 at Ser317 and/or Ser 345 and at Thr68, respectively [22,42-45]. Due to the induction of the damage response and the specificity of the Chk1 and Chk2 activation within the corrected cells, these cells are preferentially stalled in the cell cycle and do not divide. In contrast, the non-corrected cells do not uniformly contain activated Chk1/Chk2 and thus, no halt of cellular replication is observed [22]. Because of this lingering cell cycle block, the overall correction frequency appears to decrease as the non-corrected cells resume normal replication while the vast majority of corrected cells become arrested. In contrast to our system in which correction efficiency is measured among a background of non-corrected cells, protocols that utilize selection to isolate corrected cells, in effect, exclude all other cells from the population, only allowing corrected non-arrested cells to continue to grow. Because of this, non-corrected cells are not able to out compete and over grow the corrected cells, and thus, a decrease in gene repair is not evident [46]. In a more biologically relevant target, in which there is no such selective pressure, however, it is imperative to accommodate and respond to the varying replicative abilities of both the non-corrected and corrected cell populations.
In this report, we establish a protocol that utilizes a thymidine treatment following the introduction of the oligonucleotide, allowing for a population wide reduction in DNA replication activity, regardless of correction state of the individual cells. The post-treatment led to a decrease in cell replication by an average of nearly 8 fold at each time point and as a result, correction levels remained constant. This observation was also true for cells not pre-treated with 1 mM thymidine prior to oligonucleotide electroporation (data not shown). In addition, removing the thymidine following ssODN electroporation at 48 hours and 72 hours allowed for the maintenance of substantial levels of correction up to 144 hours. The wash-out of the thymidine treatment was accompanied by the resumption of cellular replication and a normal division rate. This implies that thymidine can be used to slow the replication of the entire cell culture temporarily in order to preserve a high level of gene correction.
In contrast, treating the cells with 2 mM thymidine for 144 hours resulted in the population entering into cellular senescence, as determined by a physiological change in cell morphology and a positive staining for SA β-galactosidase. Interestingly, the development of senescence was time-dependent as removal of thymidine at 48 hours, but not 72 hours, in an assay lasting 144 hours showed a minimal number of senescent cells, indicating that there is a window of time in which the transition between survival and senescence is determined.
Normal cells can revert to a state of replicative senescence within several days, when exposed to certain physiological stress factors [see [47,48] and references therein]. In addition, hydroxyurea (HU), which depletes cells of dNTPs and rapidly shuts down cell replication [49,50], has been shown to induce senescence following a long treatment [51,52]. Not only could the entry into senescence after a prolonged treatment with thymidine be due to the cellular stress following an imbalance of ribonucleotide reductase, but also due to activation of other pathways that lead to cell senescence. Excess thymidine in the cell has been shown to activate an ATM-mediated protein kinase cascade which is followed by an ATR-mediated response; both of which are required for cell survival following the treatment [53]. The response is believed to emanate from an alteration in the chromatin state [54] or the formation of an altered DNA structure, known as a "chicken foot," following genome-wide stalling of replication forks [53,55-57], which p53 is able to recognize and bind to [58]. Both structures can activate ATM and along with a prolonged p53 induction, has been shown to mediate the onset of a senescent phenotype [59-61].
Answers to all your Biology Questions
