HDAC-inhibition counteracts everolimus resistance in renal cell carcinoma in vitro by diminishing cdk2 and cyclin A

Background Targeted therapies have improved therapeutic options of treating renal cell carcinoma (RCC). However, drug response is temporary due to resistance development. Methods Functional and molecular changes in RCC Caki-1 cells, after acquired resistance to the mammalian target of rapamycin (mTOR)-inhibitor everolimus (Cakires), were investigated with and without additional application of the histone deacetylase (HDAC)-inhibitor valproic acid (VPA). Cell growth was evaluated by MTT assay, cell cycle progression and apoptosis by flow cytometry. Target molecules of everolimus and VPA, apoptotic and cell cycle regulating proteins were investigated by western blotting. siRNA blockade was performed to evaluate the functional relevance of the proteins. Results Everolimus resistance was accompanied by significant increases in the percentage of G2/M-phase cells and in the IC50. Akt and p70S6K, targets of everolimus, were activated in Cakires compared to drug sensitive cells. The most prominent change in Cakires cells was an increase in the cell cycle activating proteins cdk2 and cyclin A. Knock-down of cdk2 and cyclin A caused significant growth inhibition in the Cakires cells. The HDAC-inhibitor, VPA, counteracted everolimus resistance in Cakires, evidenced by a significant decrease in tumor growth and cdk2/cyclin A. Conclusion It is concluded that non-response to everolimus is characterized by increased cdk2/cyclin A, driving RCC cells into the G2/M-phase. VPA hinders everolimus non-response by diminishing cdk2/cyclin A. Therefore, treatment with HDAC-inhibitors might be an option for patients with advanced renal cell carcinoma and acquired everolimus resistance.


Background
Over the past years the therapy for renal cell carcinoma (RCC) has undergone change, with better understanding of the molecular biology of RCC leading to the development of several targeted agents. The phosphatidylinositol-3 kinase (PI3K)/Akt/mammalian target of rapamycin (mTOR) pathway has been identified as a pivotal key regulator. mTOR has an impact on various cellular functions, including cell growth, proliferation and cell survival [1]. Two mTOR-inhibitors, temsirolimus (Torisel®) and everolimus (Afinitor®), have been approved by the FDA to treat advanced RCC due to prolonged progression free survival [2][3][4]. However, targeted therapy is not curative in metastatic RCC and drug response is limited. Recently, it has been shown that chronic mTOR-inhibition evokes undesired feedback mechanisms in RCC cells, which may lead to resistance development [5,6]. Undesirable feedback has also been demonstrated in prostate cancer cells after chronic exposure to everolimus, indicating molecular alterations tied to acquired resistance [7]. Agents targeting such feedback loops and cross talk with other pathways involved in acquired resistance to mTORinhibition are, therefore, urgently required [8]. Studies and clinical trials have demonstrated that histone deacetylase (HDAC)-inhibitors in combination with other therapies, including targeted therapies, induce synergistic or additive anti-tumor effects [9][10][11]. It has also been reported that during chronic application of everolimus, combination with the HDAC-inhibitor valproic acid (VPA) contributes to sustained anti-tumor activity [6]. Additionally, HDAC-inhibitors have been shown to re-sensitize tumor cells to cytotoxic drug treatment [12,13]. Hence, HDAC-inhibition might prove promising in reversing everolimus resistance in RCC. To follow up on a pilot study employing everolimus resistant RCC Caki-1 cells (Caki res ) [6], resistance dependent functional and molecular aberrations were investigated in the same cell line. Further investigation was designed to determine whether Caki res cell growth could be influenced by the HDAC-inhibitor VPA, whereby the growth behavior of Caki res compared to VPA treated Caki res cells was evaluated. It is shown that everolimus resistance contributes to a significant increase in the IC 50 , an elevated percentage of G2/M-phase cells and distinct up-regulation of the cell cycle activating proteins cdk2 and cyclin A. VPA counteracted everolimus resistance by significantly inhibiting tumor growth and reducing cdk2 and cyclin A. Thus, VPA might represent a new promising treatment option for RCC patients with acquired everolimus resistance.

Exposure to everolimus induced resistance in RCC cells
24 h exposure to ascending concentrations of everolimus [0.1-1 μM] induced a dose dependent significant reduction in the number of Caki par cells compared to the untreated control (set to 100%, Figure 1A) with an IC 50 of 0.78 ± 0.23 nM. Everolimus resistance (Caki res ) was evidenced by a significant shift of the IC 50 to 10.47 ± 3.14 nM.
Resistance towards everolimus significantly enhanced the G2/M-phase Evaluation of cell cycle progression revealed significant alterations after acquired everolimus resistance. The G2/ M-phase percentage was increased in unsynchronized Caki res cells, compared to Caki par , and was accompanied by a decrease in the S-phase ( Figure 1B). Synchronization of the cells led to a similar shift, additionally reducing the percentage of G0/G1-phase cells in Caki res ( Figure 1B).
Re-treatment of Caki res with therapeutic everolimus concentrations caused an increase in the G2/M-phase Treatment of Caki par for 24 h with 1, 5 or 50 nM everolimus dose-dependently reduced S-and G2/M-phase cells, while the percentage of G0/G1-phase cells increased ( Figure 2A). Re-treatment with everolimus had no significant effect on any cell phase in Caki res , regardless of the concentration ( Figure 2B). Therefore, all further re-treatment investigation was performed with 1 nM everolimus.
Resistance-dependent alteration in tumor growth was associated with modulated protein expression After 24 h exposure to 1 nM everolimus, Caki par revealed a decrease in phosphorylated Akt (pAkt) and p70S6 kinase (pp70S6K) compared to untreated Caki par ( Figure 3). Concomitantly, Akt's negative regulator PTEN (pPTEN) was activated by 1 nM everolimus. The 24 h application of 1 nM everolimus to Caki par caused a distinct decrease in the cell cycle activating proteins cdk1 and cdk2 as well as in cyclin A and cyclin B, whereas the negative cell cycle regulator p27 was elevated. Compared to Caki par , Caki res displayed an activation of pAkt and considerable elevation of cdk1, cdk2, cdk4 and cyclin E, whereas p27, p53 and p73 were diminished. Re-treating Caki res with 1 nM everolimus evoked additional activation of pAkt and pp70S6K, a further augmentation of cdk2 and cyclin A, along with deactivation of pPTEN. However, the expression of p27, p53 and p73 was elevated in Caki res after re-treatment.

The HDAC-inhibitor VPA inhibited tumor growth in Caki par and Caki res
Application of the HDAC-inhibitor VPA [1 mM] to Caki par cells for 1 or 2 weeks contributed to a significant reduction in cell growth ( Figure 4A,B), although to a lesser extent than that from 1 nM everolimus exposure. Exposing Caki res to VPA [1 mM] also led to significantly diminished tumor growth ( Figure 4A,B). The VPA induced growth inhibition in Caki res was significantly higher than that in Caki par .
Resistance-dependent alterations in protein expression were counteracted by HDAC-inhibition through VPA VPA caused deactivation of pAkt and pp70S6k in Caki res after 1 or 2 weeks ( Figure 5A). In addition, the expression of Akt and p70S6k was diminished with VPA after 1 week. In contrast, the activity of pPTEN was enhanced after 1 or 2 weeks of VPA treatment, compared to untreated Caki res . Applying VPA for 1 or 2 weeks to Caki res caused a considerable decrease in cdk2 and cyclin A and an elevation in p27 ( Figure 5A). VPA treatment resulted in increased acetylation (1 or 2 weeks) and increased total content of histone H3 and H4 (2 weeks) in Caki res ( Figure 5B).

Resistance towards everolimus did not affect apoptosis in RCC
Apoptosis was not influenced by treatment with everolimus in either Caki par or Caki res (data not shown). In good accordance, examination of the apoptosis proteins caspase 3 and PARP by western blot showed no differences between Caki par and Caki res and no changes were apparent after treatment with VPA (data not shown).

siRNA knock-down
Since cdk2 and cyclin A were distinctly increased in RCC res and were mainly affected by VPA treatment, their functional relevance during resistance dependent tumor growth was evaluated by siRNA knock-down. Cdk2 and cyclin A siRNA blockade, verified by western blot analysis ( Figure 6A), resulted in significant growth inhibition in both Caki par and Caki res , compared to untreated and siRNA controls ( Figure 7A + B left). The impact of HDAC1 and HDAC2 as targets of VPA was also determined by siRNA blockade. HDAC1 and HDAC2 siRNA knock-down contributed to an increase in histone H3 and H4 acetylation in Caki par and Caki res ( Figure 6B). The observed elevation of histone H3 and H4 acetylation was accompanied by significantly reduced tumor growth in Caki par and Caki res , compared to untreated and siRNA controls ( Figure 7A + B right).

Discussion
Chronic everolimus treatment led to drug resistant RCC cells. It was possible to hinder resistance by applying the HDAC-inhibitor VPA.
Caki res revealed a 13-fold higher IC 50 than Caki par . This IC 50 change is within the range of the 4 to 22 foldchange used to define drug-resistance [14,15], indicating clear-cut everolimus resistance. The IC 50 shift was associated with a considerable increase in the G2/M-phase, whereby S-phase cells were shifted into the G2/M-phase and the G0/G1-phase fraction was reduced. Such a shift has also been observed during lung cancer drug resistance with an accelerated G2/M-phase transition [16]. In prostate cancer cells everolimus resistance has also revealed a higher G2/M-phase cell cycle fraction [7]. Based on a recent study, chronic everolimus application to RCC cells resulted in an accumulation of G2/M-phase cells [6]. The G2/M-shift may, therefore, be characteristic of chronic everolimus exposure and be associated with resistance development. Everolimus resistance contributed to characteristic molecular changes, including activation of the everolimus target molecules Akt and p70S6K. Re-treatment of Caki res with 1 nM everolimus evoked additional activation of Akt and p70S6k, whereas the activity of the Akt negative regulator, PTEN, was diminished. Akt is a key molecule with multiple functions, including cell growth and survival. Tumor progression and resistance development in RCC in vitro and in vivo towards different agents has been associated with increased activity of the PI3K/Akt/mTOR signaling pathway [5,8,17,18]. Enhanced activity of Akt has also been shown to be involved in bone metastasis, larger tumor size, grades III/IV tumors and shorter disease-free survival in RCC [19][20][21]. Moreover, elevated Akt phosphorylation has been associated with hyperproliferation and overexpression of cell cycle proteins [22]. Indeed, the present study shows that the cell cycle activating proteins cdk2 and cyclin A were both over-expressed in Caki res compared to Caki par , and further increased after re-treatment with everolimus. The finding that proteins involved in mitotic control were further up-regulated after applying a therapeutic everolimus concentration is clinically relevant, since mitotic activity of tumor cells is often accelerated, once resistance has developed. In the present investigation the number of mitotic cells significantly increased when Caki res cells were exposed to low-dosed everolimus. This finding might, therefore, explain why RCC patients, treated with temsirolimus or everolimus often develop progressive disease [23]. The same progression has been observed in different gynecologic cancers as well as estrogen receptor-positive breast cancer and colorectal carcinoma, indicating a correlation between treatment resistance and enhanced aggressiveness characterized by accelerated tumor growth [24][25][26].
The functional relevance of cdk2 and cyclin A in tumor growth was verified by siRNA knock down, revealing significant growth inhibition after cdk2 and cyclin A loss. Cdk2 and cyclin A establish complexes in the S-phase and are required for entrance into the G2/Mphase. Indeed, low expression of cdk2 and cyclin A has been shown to be associated with cell cycle arrest and accumulation of cells in the S-phase [27]. Everolimus resistance has also been associated with a considerable increase in cdk2 in prostate cancer [7] and in RCC cells [6]. Thus, augmented cdk2 seems closely related to nonresponsiveness towards everolimus and deserves attention in overcoming resistance development. High levels of cyclin A have been associated with a worse outcome and have been proposed as a prognostic factor in breast cancer, as well [28][29][30]. Similarly, a cyclin A increase in RCC has been associated with elevated tumor size and poor survival [31]. In good accordance with the present findings regarding Caki res , cyclin A expression has been shown to be inversely correlated with the expression of the cell cycle negative regulator p27 in RCC [31]. It may, therefore, be concluded that resistance development towards everolimus is accompanied by elevated cdk2/cyclin A, driving tumor cells from the S-into the G2/M-   phase, leading to a more aggressive tumor phenotype with enhanced growth capacity.
The HDAC-inhibitor VPA caused a significant decrease in RCC tumor growth. Since VPA's growth inhibitory effect on Caki-1 was even more pronounced in Caki res than in Caki par , VPA seems to re-sensitize the tumor cells to everolimus. However, it may also be concluded that chronic treatment with everolimus sensitizes the cells to VPA treatment. Although this is speculative, various studies have shown that HDAC-inhibitors in combination with everolimus induce synergistic anti-tumor effects [10]. HDAC-inhibitors have been implicated in the re-sensitization of tumor cells to cytotoxic drug treatment [12,13] and concomitant application of VPA with chemo-or targeted therapies has shown that VPA prevents tumor cells from becoming resistant [6,32]. VPA may, therefore, counteract resistance dependent feedback loops and reverse molecular alterations in everolimus resistant cells. VPA treatment did deactivate proteins associated with mitosis in the Caki res cells, particularly Akt and p70S6k. Both cdk2 and cyclin A were strongly enhanced in Caki res and were considerably reduced in the presence of VPA. Thus, cdk2 and cyclin A could serve as predictive indicators for a response to VPA. In several tumor entities application of VPA for up to 2 weeks has resulted in counter-regulation of the cdk-cyclin axis, contributing to significant growth inhibition [6,33]. Since cdk2/cyclin A reduction and growth inhibition in Caki res after application with VPA was accompanied by acetylation of histone H3 and H4, epigenetic modification might be involved in the anti-tumor effect. Other investigators have also reported an association between histone H3 and H4 acetylation, cdk2 reduction and diminished growth in bladder and prostate cancer cells [33,34]. Knock-down of HDAC1 and HDAC2, responsible for deacetylation of histone H3 and H4 [35,36], has induced considerable acetylation of histone H3 and H4 in Caki res , correlating with significant growth inhibition. The present investigation also indicates that histone H3 and H4 acetylation diminishes tumor growth, probably by influencing the cell cycle activating proteins cdk2/ cyclin A.

Conclusion
The present findings reveal that resistance development towards the mTOR-inhibitor, everolimus, is associated with undesired feedback loops, including activation of the Akt/mTOR signaling pathway and increased cdk2/ cyclin A expression, and is associated with cell cycle progression and tumor growth. Evidence is provided that re-treatment with everolimus not only fails to inhibit tumor progression in Caki res , but activates progression. Since resistance is a serious problem in treating RCC the HDAC-inhibitor VPA could be employed to impair cdk2/cyclin A expression. Acetylation of histone H3 and H4 plays a pivotal role in this process, significantly inhibiting tumor cell growth. Patients with renal cell carcinoma and acquired everolimus resistance might, therefore, benefit from treatment with VPA. In vivo investigation and clinical trials are necessary to verify tumor growth inhibition by VPA in everolimus resistant renal cell carcinoma.

Drugs
Everolimus (RAD001, Afinitor™, Novartis Pharma AG, Basel, Switzerland) was dissolved in DMSO as a 10 mM stock solution and stored as aliquots at -20°C. Prior to experiments, everolimus was diluted in cell culture medium. Resistance towards everolimus was induced by treating Caki-1 cells with stepwise ascending concentrations from 1 nM up to 1 μM. The tumor cells were further exposed to 1 μM everolimus twice a week for over one year. Tumor cells, resistant to everolimus, were designated Caki res and control cells, sensitive to everolimus, were designated Caki par .
Besides comparing characteristics of Caki res to Caki par , the response to therapeutic everolimus concentrations (drug re-treatment) was also investigated. Preparation for everolimus re-treatment was carried out by incubating Caki res cells for three days with everolimus-free medium. Subsequently, 1, 5 or 50 nM everolimus was applied to the Caki res and Caki par cells.
Valproic acid (VPA) (G. L. Pharma GmbH, Lannach, Austria) was applied at a final concentration of 1 mM to Caki res and Caki par cells twice a week over a total of either 1 or 2 weeks. Control cell cultures remained untreated.
To evaluate toxic effects of applied drugs, cell viability was determined by trypan blue (Gibco/Invitrogen, Darmstadt, Germany).

Apoptosis
To detect apoptosis the expression of Annexin V/propidium iodide (PI) was evaluated using the Annexin V-FITC Apoptosis Detection kit (BD Pharmingen, Heidelberg, Germany). Tumor cells were washed twice with PBSbuffer, and then incubated with 5 μl of Annexin V-FITC and 5 μl of PI in the dark for 15 min at room temperature. Cells were analyzed on a FACScalibur (BD Biosciences, Heidelberg, Germany). The percentage of vital, necrotic and apoptotic cells (early and late) in each quadrant was calculated using Cell-Quest software (BD Biosciences).

Measurement of tumor cell growth and proliferation
Cell growth was assessed using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dye reduction assay (Roche Diagnostics, Penzberg, Germany). RCC cells (50 μl, 1×10 5 cells/ml) were seeded onto 96-well tissue culture plates. After 24, 48 and 72 h MTT (0.5 mg/ml) was added for an additional 4 h. Thereafter, cells were lysed in a buffer containing 10% SDS in 0.01 M HCl. The plates were incubated overnight at 37°C, 5% CO 2 . Absorbance at 550 nm was determined for each well using a microplate ELISA reader. Each experiment was done in triplicate. After subtracting background absorbance, results were expressed as mean cell number. IC 50 values were calculated on the basis of the dose response analysis of Caki par and Caki res using GraphPad Prism 5.0.

Cell cycle analysis
Cell cycle analysis was performed with RCC cultures grown to subconfluency and carried out after 24 h using both asynchronous and synchronous cell populations. Caki-1 cells were synchronized at the G1-S boundary with aphidicolin (1 μg/ml) 24 h before starting cell cycle analysis and subsequently resuspended in fresh (aphidicolin free) medium for 2 h. Asynchronous or synchronous tumor cell populations were stained with propidium iodide using a Cycle TEST PLUS DNA Reagent Kit (Becton Dickinson) and then subjected to flow cytometry with a FACScan flow cytometer (Becton Dickinson). 10,000 events were collected from each sample. Data acquisition was carried out using Cell-Quest software and cell cycle distribution, calculated with ModFit software (Becton Dickinson). The number of gated cells in the G1-, S-or G2/M-phases were expressed in%.

Statistics
All experiments were performed 3-6 times. Statistical significance was investigated by the Wilcoxon-Mann-Whitney-U-test. Differences were considered statistically significant at a p-value less than 0.05.