Background
The cell cycle is comprised of a series of highly coordinated events culminating in cell growth and division. Cyclin-dependent kinases (CDK) and their cyclin counterparts strictly regulate and drive cell cycle progression and different CDK/cyclin complexes are responsible for the timely occurrence of each phase transition in order to maintain genetic integrity throughout generations. Cancer cells have been frequently found to have a de-regulated CDK activity allowing them to escape the normal cell cycle and proliferate uncontrollably. For these reasons CDKs have been considered attractive targets for cancer therapy and several CDK-inhibitors have been developed and are under intense investigation[
1].
R-Roscovitine (Seliciclib, CYC202; herein referred to as Roscovitine), one of the most promising members of the CDK-inhibitor family, is an orally available adenosine analogue prominently targeting CDK2 (also affecting CDKs 1, 7 and 9 at a much lower rate)[
2] with a low off-target effect on other members of the human kinome[
3], and a nice toxicity profile[
4]. In preclinical studies Roscovitine has shown significant
in vitro and
in vivo antitumor activity on a wide panel of human cancers and is currently in phase II clinical trials[
5]. Since preclinical experimentation, it has become evident that, CDK-inhibitors, such as Roscovitine, may actually curb the activity of DNA repair machinery[
6,
7], hence becoming an attractive candidate for therapeutic association with either radiation therapy[
8,
9] or genotoxic agent-based chemotherapy[
10]. However, the mechanism of this inhibition is still elusive.
One of the proposed means for CDK-inhibitors to affect DNA repair is through checkpoint deregulation[
11‐
13], but increasing evidence supports a complex network of direct interactions between individual CDKs and proteins that play a key role in DNA damage repair (DDR). It is known that different DNA repair pathways are preferentially activated at specific stages of the cell cycle possibly suggesting a functional crosstalk between CDK/cyclin complexes and DNA repair mechanisms[
14]. In particular, CDK2 has been shown to interact with p53[
15], BRCA1[
16], BRCA2[
17], Ku70[
18] and both, CDK1 and CDK2, can modulate BRCA1-BARD1 activity[
13,
19]. Moreover, CDK2 knock-down cells have an attenuated capacity to repair DNA damage suggesting a pivotal role for CDK2[
7] in DDR. Given the ability of CDKs to compensate for each other
in vivo, overall CDK activity has been proposed to be influential in DDR regulation[
20] however CDK2 function seems to have a specific role in some survival pathways[
21].
Cyclins, similarly to CDKs, have been correlated to DDR. Cyclin E levels are upregulated under genotoxic stress conditions[
22] and a post-translational cleavage generates an 18-amino acid peptide, which has been shown to interact with Ku70[
18] promoting the release of the pro-apoptotic factor Bax from the inactivating complex Bax/Ku70. Moreover, an increasing amount of data suggests an important role in DDR for the A-type cyclins, and in particular for cyclin A1. Differing from cyclin A2, ubiquitously expressed during the S and G2/M phases of the cell cycle, cyclin A1 is a testis-specific cyclin, which interacts with CDK2 and is involved in germ cell meiosis and spermatogenesis[
23]. Cyclin A1 may have a role in carcinogenesis, as it has been found to be over-expressed in acute myeloid leukemia and various other tumour types[
23‐
25], however, its role in cancer is still particularly obscure. In somatic non-testicular tissues, cyclin A1 is not expressed or is expressed at very low basal levels. After genotoxic insult, cyclin A1 mRNA is upregulated
in vitro[
26] and
in vivo[
27]. At a molecular level, human CDK2/cyclin A1 complexes interact with members of the Ku family and phosphorylate Ku70[
27,
28], a pivotal player in the non-homologous end-joining (NHEJ) double strand break (DSB) repair pathway. Furthermore, under genotoxic conditions the kinase activity of CDK2/cyclin A1 complex increases, while the relative kinase activity of CDK2/cyclin A2 decreases and the CDK2/cyclin A1 complex out-competes with CDK2/cyclin A2 for Ku70 binding[
28]. Moreover, it has recently been found that CDK2 phosphorylation status and structure changes upon the cyclin A family member with which it is bound [
29] suggesting a non-redundant function between CDK2/cyclin A1 and CDK2/cyclin A2 complexes. Finally cyclin A1 knockout mice and
Xenopus embryos exhibited a clear defect in DNA repair[
27,
30] and are more prone to undergo apoptosis[
31].
Taken together these data support that during genotoxic stress differential transcriptional levels and activity of cyclin A family members may redirect CDK2 toward DNA repair resulting in a modulation of NHEJ. Since one of the most relevant effects of CDK inhibitors is the downregulation of cell cycle related cyclins, we investigated if the inhibition of DNA repair mechanisms by Roscovitine may also occur through the modulation of the expression levels of cyclin A family members. Physiological CDK-inhibition, in fact, results in cyclin downregulation through the inhibition of E2F-family transcription factors, which drive and regulate cell cycle-related cyclin transcription. Given that the promoter of the cyclin A1 gene,
CCNA1, is different from the other cell cycle-related cyclins, not being under the regulation of E2Fs[
32], here we analyzed the effects of Roscovitine on cyclin A1 expression and modulation of DNA repair mechanisms. We demonstrated that under DNA damaging conditions cyclin A1 is strongly upregulated and localizes to the nucleus. Although Roscovitine alone was not sufficient to reduce the basal levels of cyclin A1, in contrast to cell cycle related cyclins, Roscovitine treatment could abolish the DNA damage-induced cyclin A1 upregulation, reducing NHEJ and significantly hindering DNA repair over time.
Discussion
Under genotoxic conditions the CDK2/cyclin A1 complex increases its functional kinase activity and the ability to phosphorylate Ku70. In addition, here we demonstrated upon treatment with different DNA damaging agents (doxorubicin or γ-irradiation) a marked dose dependent increase in the RNA and protein levels of cyclin A1, which is independent of cell cycle phase redistribution. Conversely cyclin A2 (whose expression is tightly related to the S and G2-M phases of the cell cycle) is downregulated under genotoxic stress conditions as a result of the check-point activation and consequent decrease of the S phase fraction. This switch in the respective levels of the A-family cyclins may be functionally relevant to redirect CDK2 activity toward DNA repair, especially given the findings that the ectopic overexpression of cyclin A1 increased
in-vitro NHEJ activity and that cyclin A1 depletion, as demonstrated by others[
27], results in an impaired DNA DSB repair ability.
DNA DSBs are considered the most lethal form of DNA damage and CDK inhibition has been shown to potentially affect the two major DSB repair pathways (HR and NHEJ)[
7]. Various mechanisms have been proposed to explain this effect such as the deregulation of the DNA damage-induced checkpoint signalling cascade[
13] or the downregulation of specific genes involved[
33,
34]. Roscovitine is an oral 2,6,9 trisubstituted purine analog currently under phase II investigation, which competes with ATP for the catalytic binding site on CDK2 (but also CDKs 1, 7 and 9 with a much lower affinity) with a demonstrated antitumor activity in many human cancer models and a nice toxicity profile.
One of the most prominent effects of the drug is the inhibition of CDK2/cyclin E complexes, which causes a decrease in Rb phosphorylation and a consequent inactivation of E2F family members, thus leading to cyclin transcriptional downregulation and ultimately to cell cycle arrest. This strong transcriptional depression of most of the cell cycle related cyclins further enforces the drug's inhibitory effect on CDK/cyclin complexes. Furthermore, Roscovitine has been shown to downregulate several other genes involved in a wide spectrum of cellular functions[
35,
36], probably as a result of partial CDK7/cyclin H and CDK9/cyclin T inhibition[
37]. In addition, whole genome ChIP-on-chip analysis recently mapped E2F transcription factor family members to the promoters of many more genes than were traditionally associated with the cell cycle[
38], suggesting an alternative mechanism to explain these transcriptional effects.
We investigated the effects that Roscovitine may have on cyclin A1 transcription as one of the possible mechanisms through which CDK2 inhibition may curb DNA DSB repair activity. The promoter of the cyclin A1 gene, CCNA1 is not E2F-dependent and, consistently, increasing doses of Roscovitine did not repress cyclin A1 basal transcription levels in contrast to cyclins A2, B, D and E. However, we demonstrated that Roscovitine at doses preferentially inhibiting CDK2 but not CDK7 and 9 completely abolished cyclin A1 DNA damage-induced upregulation, thus suggesting that residual CDK2 activity is required for cyclin A1 upregulation. In addition Roscovitine co-administered with doxorubicin was able to largely modify the patterns of cell cycle phase distribution in comparison to doxorubicin only treatment. This resulted in an augmented S phase and consequently in an increased expression of cyclin A2. The combined treatment thus resulted in the complete inversion of the doxorubicin-induced switch between cyclin A1 and cyclin A2.
Roscovitine, alone or under DNA damaging conditions, was able to diminish cyclin A1 protein levels as well. Such transcriptional and post-transcriptional repression was observed in different NSCLC, prostate and breast cancer cell lines and we propose that this potentiates and synergizes the Roscovitine-mediated CDK2 inhibition thus resulting in a significant decrease of cellular NHEJ ability. In fact, we observed that combination treatment led to an increase in DNA DSBs and overall DNA damage over-time, further substantiating, not only the importance of CDK-inhibitors in combination therapy but also the role of CDKs in DNA repair mechanisms. While these findings were supported by genome-wide mircroarray analysis, we also observed a significant effect on key genes involved in other DNA repair pathways.
Methods
Cell Culture and Serum Starvation
The following solid cancer human cell lines were purchased from and authenticated by American Type Culture Collection (ATCC; Manassas, VA) and cultured at 37°C in a humidified atmosphere of 5% CO2 in air, within the appropriate medium according to supplier recommendations supplemented with 10% (v/v) heat-inactivated fetal bovine serum (Atlanta Biologicals; Lawrenceville, GA) and 100U of Penicillin and 100 μg/ml of Streptomycin (Sigma-Aldrich; St. Louis, MO): NSCLC cell lines A549 and H23, breast cancer cell lines MCF-7 and MDA-MB-231, prostate cancer cell lines LNCAP and DU145, and the adenovirus transformed human embryonic kidney epithelial cells HEK293FT. Cells were regularly sub-cultured according to ATCC recommendations with a 0.25% trypsin-EDTA solution (Sigma). To obtain synchronous populations of cells, confluent plates of A549 cells were incubated in media supplemented with 0.1% (v/v) heat-inactivated fetal bovine serum for 96 hours. Cells were then sub-cultured into serum-containing medium and time points were taken every four hours.
Drugs, irradiations and treatments
Doxorubicin was obtained from BioMol International (Plymouth Meeting, PA). Lyopholized drug was re-suspended into a 1:1 mixture of dimethyl sulfoxide (DMSO; Fisher Scientific; Pittsburgh, PA) and MilliQ filtered H2O (Millipore; Bellerica, MA) to a concentration of 4.31 mM, aliquoted for use and stored at -20°C. Roscovitine was obtained from Signa Gen Laboratories (Gaithersburg, MD). Lyophilized drug was re-suspended into DMSO to a concentration of 14.1 mM, aliquoted and stored at -20°C until use. Fresh dilutions from the stock solutions were prepared for each treatment. Taxol was obtained from USB Corporation (Cleveland, OH). Lyophilized drug was re-suspended into DMSO to a concentration of 5.86 mM, aliquoted and stored at -20°C until use. MG-132 (Z-Leu-Leu-Leu-al) was obtained from Sigma. Lyophilized drug was re-suspended into DMSO to a concentration of 10 mg/ml, aliquoted and stored at -20°C until use. Irradiations were performed in an AECL Gamma Cell 40, Cs-137 irradiator at a dose rate of 1 Gy/minute for respective doses. In treatments throughout this article the control samples refer to cells treated with an equal concentration (v/v) of DMSO as in the highest drug concentration used per experiment.
Western Blot Analysis and SDS-PAGE
Equal amounts (50-100 μg) of whole cell lysates were resolved by SDS-PAGE and transferred to a nitrocellulose membrane (Whatman Inc., Piscataway, NJ) by wet electrophoretic transfer. Non-specific binding sites were blocked for 1 hour at room temperature with 3% non fat dry milk (NFM) in tris-buffered saline containing 0.01% Tween-20 (TBS-T) and probed with the following primary antibodies in 3% NFM in TBS-T overnight at 4°C; rabbit anti-cyclin A1 (sc-15383; Santa Cruz Biotechnology Inc.; Santa Cruz, CA), mouse anti-cyclin A2 (CY-A1; Sigma), mouse anti-cdc2 (A17; Abcam, Cambridge, MA), rabbit anti-CDK2 (sc-163; Santa Cruz), rabbit anti-p53 (sc-6243; Santa Cruz), mouse anti-Hsp70 (sc-24; Santa Cruz), mouse anti-p130/Rb2 full length (610262; BD Biosciences, San Jose, CA), rabbit anti-serine 952 phosphorylated p130/Rb2 (sc-16298; Santa Cruz), rabbit anti-serine-2 phosphorylated RNA polymerase II (A300-654A; Bethyl Laboratories Inc., Montgomery, TX), rabbit anti-serine-5 phosphorylated RNA polymerase II (A300-655A; Bethyl), mouse anti-α-tubulin (sc-58666; Santa Cruz), and mouse anti-ser139 phosphorylated histone γH2AX (Millipore cat. #05636; lot# DAM1567248). Membranes were washed for 15 minutes in TBS-T and then incubated for 1 hour with either goat anti-mouse (31432; Pierce; Rockford, IL) or mouse anti-rabbit (31464; Pierce) horseradish peroxidase conjugated IgG at a dilution of 1:10,000 in 3% NFM in TBS-T. This was followed by 15 minutes of wash in TBS-T and enhanced chemiluminescence (ECL; Amersham, Buckinghamshire, UK) according to the manufacturer's instructions. All western blot images included in article are representative of at least three consecutive independent experiments.
Immunostaining
Following respective drug treatments, cells grown directly on sterilized glass coverslips were fixed and permeabilized for 10 minutes in 70% cold methanol (MeOH), immunostained (for cyclin A1 and γH2AX) and analyzed as previously described[
40].
Flow cytometry
Cells (1 × 106) were collected, after respective drug treatments, washed, resuspended in 1 ml of PBS and fixed and permeabilized for at least 10 minutes in 70% cold ethanol. After fixation cells were pelleted, washed 3 times with PBS, re-suspended into a primary antibody solution (10 μg/ml antibody diluted in PBS) and incubated on ice for 15 minutes. Cells were then pelleted, washed 3 times with PBS, re-suspended into FITC-conjugated secondary antibody solution (10 μg/ml) and incubated for 15 minutes on ice protected from the light. Cells were washed 3 times in PBS and re-suspended in propidium iodide staining solution, 10 μg/ml propidium iodide (from stock of 0.5 mg/ml in 0.38 mM sodium citrate pH 7.0) and 25 μg/ml DNase-free RNase A (from stock of 10 mg/ml RNase A in 10 mM Tris pH 7.5 and 15 mM NaCl) diluted in PBS. Cells were incubated at 37°C for a minimum of 30 minutes protected from light and analyzed immediately by flow cytometry utilizing an Epics XL-MCL BeckmanCoulter (The Wistar Institute, Philadelphia, PA). Graphs represent average fluorescence intensity or average percentage of cells found in cell cycle phase over three consecutive independent experiments.
Reverse Transcriptase-PCR and Real time (RT-PCR)
Total RNA from cell lines was extracted using the High Pure RNA Isolation Kit (Roche) following the manufacturer's instruction. cDNA was synthesized from 1 μg of total RNA by using random hexamers as primers and moloney murine leukemia virus reverse transcriptase (Invitrogen, Carlsbad, CA) according the manufacturer's protocol in a final volume of 20 μl. As a control for genomic contamination a reverse transcription (RT) reaction was carried out without the addition of the reverse transcriptase (RT-). After cDNA synthesis, samples were diluted 1:10 and 4 μl was used in each real time polymerase chain reaction (real time PCR). cDNA was amplified using species specific intragenic primers for
CCNA1[
23],
CCNA2,
CCNB1,
CCND3,
CCNE1,
TP53 and
GAPDH genes (Additional File
4). Real time PCR was carried out utilizing SybrGreen Master Mix (Roche, Basel, Switzerland) following the manufacturer's instructions in a final reaction volume of 10 μl. Reactions were performed on a LightCycler 480 II (Roche Diagnostics, Indianapolis, IN) with an initial denaturation of 5 minutes at 95°C; 45 cycles of 10 seconds at 95°C, 20 seconds at 60°C, and 10 seconds at 72°C where fluorescence was acquired. Each sample was run in triplicate and data was analyzed using the comparative Ct method with GAPDH as the endogenous control and control cells as the reference sample in each experiment. Final data points represent the average fold change respect to control (2
^-ΔΔCt) or expression levels respect to GAPDH (2
^-ΔCt) of at least three consecutive independent experiments.
Alkaline Comet Assay
After appropriate drug treatments, cells were harvested and analyzed utilizing the alkaline comet assay as previously described[
41,
42]. Briefly, cells were mixed in a suspension of low melting point agarose and spread on agarose-coated slides. Once the agarose solidified, slides were incubated in lysis buffer followed by electrophoresis to allow migration of DNA and detection of DNA damage. Cells were then stained with 1 μg/mL ethidium bromide and analyzed using the fluorescence microscope Olympus BX40 (Melville, NY) with a Spot-RT digital camera and software (Webster, NY). At least 200 cells were evaluated per experimental point. Visual scoring of comet images using fluorescence microscopy was performed according to Norbury[
43]. Briefly, each nucleus is assigned a score from 0-4 depending on the relative intensity of DNA fluorescence in the tail (0 = no damage, 4 = >80% of DNA found in the tail) and the final score is calculated as the average DNA damage found in all cells counted from three consecutive independent experiments. Statistical analysis was carried out using a standard student's t test.
Transient transfections
The human cyclin A1 IMAGE clone 5172478 (GenBank:BC036346.1) was purchased from ATCC (MGC-34627) transformed into DH5α heat-shock competent E. coli cells and grown on LB agar plates or in broth with 100 μg/ml Ampicillin (Fisher) at 37°C. Plasmid DNA was extracted using the Genopure Plasmid Midi Kit (Roche) following manufacturer's instructions then verified by restriction enzyme digestion and gel electrophoresis. HEK293FT cells were transiently transfected using a 6:2 ratio of Fugene HD (Roche) and plasmid DNA (2 μg) following manufacturer's protocol. Enhanced yellow fluorescent protein (pEYFP) plasmid DNA was utilized as a control for transfection efficiency at the same concentration. Cells were analyzed after 36 hours of transfection by western blot and fluorescence microscopy to confirm expression of transfected protein and then utilized in experiments as described.
In vitro NHEJ assay
The
in vitro NHEJ assay was performed on respectively treated cell lysates as previously described[
44] utilizing 120 μg of protein extract and 60 μg of purified BamHI (Roche) digested pCI-neo plasmid DNA (Promega). A reaction including the incubation of 20 μM Wortmannin with whole cellular lysate for 15 minutes on ice before the addition of digested plasmid DNA was included as a negative control for NHEJ activity in each experiment. After incubation samples were diluted 1:10, phenol chloroform 25:24:1 (Fisher) extracted, and ethanol precipitated overnight at 4°C. DNA was resuspended into 20 μl of Tris-EDTA buffer and 1 μl was utilized in each real time PCR reaction. To detect plasmid re-ligation one set of primers amplified an intact region of the plasmid to act as the endogenous control, while a second set of primers bound both up-stream and down-stream of the enzymatic cut site. Samples were run in triplicate with each primer pair following the real-time PCR protocol described above. Final results represent the average fold change (2^
-ΔΔCt) in re-ligation respect to control, over three consecutive independent experiments.
Microarray Analysis
Total RNA was isolated by Trizol (Invitrogen). Fifteen μg of total RNA was converted to cDNA by using Superscripts reverse transcriptase (Invitrogen), and T7-oligo-d(T)24 (Geneset) as a primer. Second-strand synthesis was performed using T4 DNA polymerase and E.Coli DNA ligase and them blunt ended by T4 polynucleotide kinase. cDNA was purified by phenol-chloroform extraction using phase lock gels (Brinkmann). Then cDNAs were
in vitro transcribed for 16 hours at 37°C by using the IVT Labelling Kit (Affymetrix) to produce biotinylated cRNA. Labelled cRNA was isolated by using the RNeasy Mini Kit column (QIAGEN). Purified cRNA was fragmented to 200-300 mer using a fragmentation buffer. The quality of total RNA, cDNA synthesis, cRNA amplification and cRNA fragmentation was monitored by capillary electrophoresis (Bioanalizer 2100, Agilent Technologies). Fifteen micrograms of fragmented cRNA was hybridised for 16 hours at 45°C with constant rotation, using a human oligonucleotide array U133 Plus 2.0 (Genechip, Affymetrix, Santa Clara, CA). After hybridisation, chips were processed by using the Affymetrix GeneChip Fluidic Station 450 (protocol EukGE-WS2v5_450). Staining was made with streptavidin-conjugated phycoerythrin (SAPE)(Molecular Probes), followed by amplification with a biotinylated anti-streptavidin antibody (Vector Laboratories), and by a second round of SAPE. Chips were scanned using a GeneChip Scanner 3000 G7 (Affymetrix) enabled for High-Resolution Scanning. Images were extracted with the GeneChip Operating Software (Affymetrix GCOS v1.4). Quality control of microarray chips was performed using the AffyQCReport software[
45]. A comparable quality between microarrays was demanded for all microarrays within each experiment.
Microarray Statistical Analysis
The background subtraction and normalization of probe set intensities was performed using the method of Robust Multiarray Analysis (RMA) described by Irizarry et al.[
46]. To identify differentially expressed genes, gene expression intensity was compared using a moderated t test and a Bayes smoothing approach developed for a low number of replicates[
47]. To correct for the effect of multiple testing, the false discovery rate, was estimated from p-values derived from the moderated t test statistics[
48]. The analysis was performed using the affylmGUI Graphical User Interface for the limma microarray package[
49].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
MF and CES designed experiments, performed the research, analyzed the data and wrote the paper. DF performed microarray experiments and analysis. FR performed experiments and analyzed the data. LB, AR and AG designed experiments and wrote the paper. All authors critically reviewed and edited the paper.