Abstract
Background: Overexpression of cyclin D1 is associated with resistance to chemotherapy and radiotherapy in several types of cancer including head and neck squamous cell carcinoma (HNSCC). Materials and Methods: Cyclin D1 was silenced in an HNSCC cell line and its effect tested in sensitizing the cells to cisplatin, in vitro as well as in vivo. The HNSCC cell line NT8e, which is a chemoresistant, cyclin D1 over-expressing cell line, was used in the study. RNAi (shRNA) against cyclin D1 was designed and cloned in a vector. Results: Stable silencing of cyclin D1 resulted in delayed cell cycle progression and significantly sensitized the cells to cisplatin. Effective cell kill was achieved at a much lower therapeutic dose in vivo. Conclusion: Suboptimal concentrations of cisplatin could be used in vivo to eradicate xenograft tumors indicating the promise of combining vector-based cyclin D1 silencing with chemotherapy to achieve maximum tumor regression.
Head and neck cancer is the most common type of cancer reported in the Indian population (1) ranking first in males and third in females (2). Despite clinical advances, the 5-year survival of such patients remains poor. Out of the various factors associated with the poor survival rate of patients with head and neck squamous cell carcinoma (HNSCC), drug resistance is of prime importance. Resistance to therapy in HNSCC has been shown to correlate with deregulated p53 and cyclin D1 levels (3).
Cyclin D1 gene (CCND1), an oncogene located at 11q13, encodes a regulatory subunit of the CDK4 and CDK6 holoenzyme complex, which phosphorylates and deactivates the tumor suppressor protein retinoblastoma (RB) (4). Phosphorylation of RB releases E2F family of transcription factors which then proceed to activate genes that are essential for advancing the cell cycle into the late G1 and S phases (5). Cyclin D1 has been shown to be overexpressed in majority of cancer types including HNSCC (4, 6, 7), and is correlated with poor drug/radiation response of various cell lines (8-12). Antisense molecules to inhibit the expression of cyclin D1 have been reported to reduce the tumorigenicity of xenografts or increase the sensitivity to drugs such as cisplatin (13-15). Recently an RNAi approach for gene silencing has gained attention owing to its specificity and utility at low doses to bring about the same silencing effect as antisense molecules (16). siRNA duplexes have been tried (17), but the stability of RNA duplexes, once delivered in vivo remains a challenge.
Radiation resistance is well correlated with overexpression of cyclin D1, epidermal growth factor – receptor (EGF-R) (10) and p53 dysfunctions in several types of cancers. It has been shown that cells having deregulated p53 fail to undergo apoptosis following various damage stimuli to DNA (18), and this explains why p53-defective cells are resistant to radiation induced cell death. On the other hand cyclin D1 overexpression in the breast cancer cell line MCF7 is shown to increase the radiation sensitivity (19). Taken together, these observations suggest that sensitivity to radiation is tissue/cell type-specific, and thus assessment of sensitization approaches should be carried out in a tissue/cell specific manner.
Materials and Methods
Cell lines and culture conditions. HNSCC cell line NT8e, established in our laboratory (20), was used for in vitro as well as in vivo experiments. Human embryonic kidney cell line HEK293 was used for transient transfection studies to validate shRNA constructs. Fetal buccal mucosa cells FBM (normal immortalized epithelial cells) were obtained from Dr. A.N. Bhisey, Cancer Research Institute, Mumbai, India. All the cells were maintained in Dulbecco's modified Eagle's medium (Life Technologies, Grand Island, NY, USA) with 10% fetal calf serum (Life Technologies).
Cyclin D1 down regulation using vector driven RNAi. A: Validation of shRNA constructs for cyclin D1 was carried out by co-transfecting hemagglutinin-cyclin D1 (HA-CCND1) and shRNA constructs in HEK293 cells. The upper panel shows Western blot for HA-cyclin D1 and the lower panel shows Western blot of β-actin, used as internal loading control. pshCyD1-1st set showed potent silencing of HA-cyclin D1 and was used for transfecting NT8e cells. B: Stably transfected NT8e cell clones were characterized further and probed with antibodies to cyclin D1, cyclin D2, phosphorylated retinoblastoma (pRB), and RB as described in the Materials and Methods. Unaltered levels of cyclin D2 showed the specificity of shRNA construct used to silence cyclin D1. Cyclin D1 down-regulation resulted in reduced phosphorylation of RB protein as compared to NT-VC cells although total RB levels remained unchanged. EV, Empty vector.
Selection of stable clone carrying CCND1 shRNA. To achieve sequence specific long term silencing of cyclin D1 in NT8e cells, we first designed three different shRNA sequences for CCND1 based on its mRNA sequence, and cloned them in psiSTRIKE™ (Promega, Madison, CA, USA). These constructs were referred as pshCyD1-1st set, pshCyD1-2nd set, and pshCyD1-3rd set. Each of these shRNA constructs was co-transfected along with hemagglutinin (HA) CCND1 in HEK293 cells. Western blot analysis using HA-specific antibodies (sc-805, Santa Cruz Inc. Santa Cruz, CA, USA) 72 h post transfection revealed that pshCyD1-1st set was able to bring down HA-CCND1 levels potently (Figure 1A). This construct pshCyD1-1st set (sequence of insert as shown in Table I) was then transfected into NT8e cells, and after selection with 400 μg/ml G418 (Life Technologies), stable cell clones were picked up and expanded. Cyclin D1 expression in these clones was examined by Western blot. NT-VC (NT8e cells stably transfected with vector alone as control) and NT-shCyD1 (NT8e cells stably transfected with shRNA against CCND1) were used in all experiments.
Western blot analysis. Cells were lysed in Laemmeli buffer and placed in boiling water for 10 min. Lysates were spun down, supernatant collected and western blotting carried out using standard protocol. Antibodies for cyclin D1 (clone DCS-6, sc-20044, dilution 1:1000), cyclin D2 (sc-452, dilution 1:500), pRB (sc-12901-R dilution 1:1000) and total RB (clone IF8, sc-102, dilution 1:1000) were purchased from Santa Cruz Inc. Santa Cruz, CA, USA, while antibody for β-actin was procured from Sigma (clone ac-74, dilution 1:1000, Sigma-Aldrich Corporation, Saint Louis, MO, USA). All the secondary antibodies were procured from Santa Cruz Inc. USA, and used at a dilution of 1:5000.
In vitro sensitivity to cisplatin (cis-diamminedichloroplatinum). A total of 5×103 NT-VC and NT-shCyD1 cells were plated per well in 96-well plates and treated with increasing doses of cisplatin ranging from 1-6 μM. One day post cisplatin treatment, cells were washed and fresh medium was added to the wells. Cells were further incubated for 48 h, fixed with 50% Trichloro Acetic Acid, stained with sulphorhodamine B (SRB) and the O.D. readings taken at 540 nm.
Cell cycle analysis. Effect of CCND1 silencing on cell cycle progression was checked by synchronizing the cells at G1 using mimosine (Sigma-Aldrich Corporation) for 24 h. After 24 h, mimosine block was released and fresh medium was replenished. Cells were harvested at different time points, stained with propidium iodide and analyzed on flow cytometer (FACSCalibur, Becton Dickinson, San Jose, CA, USA) using Modfit software.
Cell size determination. Effect of CCND1 silencing on cell size of NT-shCyD1 was assessed by laser confocal microscopy (LCM). Cells were grown on coverslips, and cell size was determined by LCM.
Growth analysis. To assess the effect of cyclin D1 on cellular growth, 20,000 NT-shCyD1, and NT-VC cells were plated on 35 mm plates, and harvested every 24 h. Viable cell counts were taken using trypan blue dye exclusion method.
Colony formation assay. NT-shCyD1 and NT-VC cells were plated in 35 mm culture plates at 30-40% confluence and treated with escalating doses of cisplatin ranging from 1-6 μM for 24 h. Cells were then trypsinized and 500 cells from each plate were then plated in 60 mm culture dishes. The cells were incubated for two weeks, resulting colonies were fixed with normal buffered formalin (NBF) (10% formaldehyde in 1×PBS), stained with crystal violet, and colony counts were taken.
In vivo studies. Animal experiments were carried out after obtaining approval from Institutional Animal Ethics Committee. Four to six week old, female NMRi nude mice (~22 gm in weight), bred in the ACTREC Animal House, Mumbai were taken for the study. A total of 107 NT-shCyD1 and NT-VC cells were injected in NMRi nude mice (n=5 for each group). After the tumors attained a size of ~5 mm, they were treated with suboptimal dose of cisplatin (2 mg/kg body weight) once a week, for four weeks. After the treatment was over, tumor sizes were recorded and mice were sacrificed. The tumor volumes were calculated as described earlier (21) using the formula (a×b2)/2, where a and b are two longest perpendicular diameters. Tumor volume ratios were calculated as tumor volume after treatment/ tumor volume before treatment.
Statistical analysis. Statistical analysis for SRB and colony formation assay was carried out using one-way ANOVA. For in vivo experiments statistical analysis was by independent t-test.
Results
Stable silencing of CCND1 in NT8e cells is sequence specific. Three different shRNA sequences for CCND1 were cloned into psiSTRIKE™ (Promega) referred to as pshCyD1-1st set, pshCyD1-2nd set, and pshCyD1-3rd set. Validation of these shRNA constructs was done by co-transfection of HA-CCND1 with each one of these constructs in HEK293 cells. Western blot analysis using HA specific antibodies revealed that pshCyD1-1st set was able to bring down HA-cyclin D1 levels potently (Figure 1A). NT8e cell line was then used for transfection with the selected plasmid clone pshCyD1-1st set. Three stably transfected NT8e clones showing potent down regulation in CCND1 were obtained (Figure 1B). Comparison of cyclin D1 levels was made against vector transfected NT8e cells referred to as NT-VC. Cells showing stable silencing of CCND1 were referred to as NT-shCyD1A, NT-shCyD1B, and NT-shCyD1C cells. To test if the shRNA constructs had no off target effects, cyclin D2 protein levels were checked in NT-shCyD1A, NT-shCyD1B, and NT-shCyD1C cells. Cyclin D2 levels remained unaltered in all three clones (Figure 1B).
Cyclin D1 knock-down results in reduced pRB phosphorylation. Cyclin D1 catalyses the phosphorylation of RB (pRB) (5) to de-repress a subset of proliferation-associated E2F target genes (22). Thus it was expected that silencing of CCND1 would result in reduced phosphorylation of RB protein. Clones showing stable silencing of CCND1 expressed reduced level of pRB compared to vector control, while the total RB protein level remained unaltered indicating the effect of down-regulation of CCND1 in these clones (Figure 1B).
Progression of NT-shCyD1 cells through the cell cycle is delayed, resulting in retarded cellular growth. G1 Cyclins including types D and E, play an important role in the cell cycle pathway (23). Since cyclin D1 is known to catalyze the inactivation of RB thereby aiding the transcription of E2F target genes necessary for cells to enter S phase of the cell cycle, we reasoned that down-regulation of CCND1 may result in delayed progression of NT-shCyD1 cells through the cell cycle. NT-shCyD1 and NT-VC cells were synchronized in G1 phase by mimosine and harvested at different time points after releasing mimosine block. Cell cycle analysis revealed a greater tendency of NT-shCyD1 cells to stay arrested at the G1 phase of the cell cycle compared to NT-VC cells (Figure 2A). We also observed a high G1 population even in the unsynchronized condition (60.08% NT-shCyD1 cells in G1 vs. 44.91% NT-VC), suggesting that CCND1 down-regulation delays the progression of cells through the cell cycle. This delay was also reflected in reduced cellular proliferation as indicated by reduced cell numbers (Figure 2B).
Cyclin D1 down-regulation results in delayed progression through G1 phase of the cell cycle. A: Cell cycle analysis of NT-shCyD1 and NT-VC cells. Cell cycle analysis showed that NT-shCyD1 had an increased tendency of remaining in G 1 arrest compared to NT-VC cells. B: Growth analysis showed that NT-shCyD1 cells grew slower compared to NT-VC cells. Data are mean±S.D.
Determination of cell size. A: NT-shCyD1 cells (right) were larger in size than NT-VC cells (left) as seen in Laser Confocal Microscopy (LCM) transmission images. B: Graphical representation of average cell sizes. Data are mean±S.D. *p≤0.01.
NT-shCyD1 cells have larger cell size. Overexpression of cyclin D1 is shown to result in reduced cell size (24). On the contrary it has been shown that mice lacking cyclin D1 are reduced in size and have severe defects in neurological development (25). Thus we determined the effect of CCND1 down-regulation on cellular size. The size of NT-shCyD1 cells was found to be higher (mean area=5049.3±1419.8 μm2) compared to NT-VC cells (mean area=1749.6±342.8 μm2), suggesting that CCND1 down-regulation resulted in increased cell size in these cells (Figure 3).
Cyclin D1 silencing sensitizes NT8e cells to cisplatin induced cell death in vitro. NT8e cells do not express p53 despite DNA damage induced by irradiation (Figure 4A). In the recent past, mutant p53 or loss of p53 functions has consistently been shown to be a cause for chemoresistance of HNSCC cells. NT8e cells survived physiological levels of cisplatin when tested in vitro (data not shown). We tested if silencing of CCND1 in NT-shCyD1 cells reduces drug doses required for cell death and result in a better cell kill by cisplatin. A drastic reduction in cellular survival of NT-shCyD1 cells compared to NT-VC cells was observed when they were treated with increasing doses of cisplatin. CCND1 silencing significantly sensitized NT-shCyD1 cells to cisplatin (mean IC50=0.9 μM, p≤0.001), compared to NT-VC cells (mean IC50=2.45 μM) (Figure 4A and B). An increase in cell kill may be explained by the observation that CCND1 down-regulation delays G1-S transition, but does not stop the cell cycle completely (Figure 2A). Thus at a given time, more cells simultaneously enter the S-phase, incorporating a greater number of cisplatin adducts, resulting in efficient cell kill, compared to a situation where cells are in different phases of the cell cycle and hence may show different sensitivity to a DNA-damaging agent. A more accurate way of determining the cytotoxicity of a drug is through the colony formation assay as it gives sufficient time to the cells to recover from cytotoxic shock. A significant drop (p≤0.01) in the colony forming ability of NT-shCyD1 cells was seen when they were treated with increasing doses of cisplatin as compared to NT-VC, cells especially at lower doses (Figure 4D and E). This difference in the chemosensitivity was not apparent at higher doses, probably because the high drug doses are cytotoxic even to NT-VC cells.
A suboptimal dose of cisplatin is sufficient for NT-shCyD1 tumor regression in vivo. To test the effect of CCND1 silencing on chemosensitivity achieved in vivo, 107 NT-shCyD1 or NT-VC cells were injected into NMRi nude mice (n=5 for each group) and allowed to form tumors. CCND1 silencing results in reduced tumorigenicity of cells (13). Consistent with this, NT-shCyD1 cells took longer (~one week delay) to form tumors (~5 mm) compared to NT-VC cells. When tumors were treated with a suboptimal dose of cisplatin (2 mg/kg body weight), once a week, for four weeks, NT-shCyD1 tumors shrunk significantly (p≤0.01), while NT-VC tumors either did not regress or grew slowly. The mean tumor volume ratio at the end of treatment for NT-shCyD1 was 0.26 compared to 2.4 for NT-VC after the treatment (Figure 5). Untreated tumors grew unchecked.
NT-shCyD1 cells are sensitized for cisplatin-induced cell kill in vitro. A: NT8e cells do not express detectable levels of p53 despite irradiation. B and C: NT-shCyD1 cells are sensitive to cisplatin induced cell kill. NT-shCyD1 cells grew in significantly fewer number of colonies compared to NT-VC cells. D: Colony formation assay. E: Graphical representation of the average number of colonies formed after the respective treatments of NT-shCyD1 cells and NT-VC cells with cisplatin. Data are mean±S.D. *p≤0.01, **p≤0.01, ***p≤0.01.
Discussion
Cyclin D1 overexpression not only is a poor prognostic marker in various tumor types (26, 27) but also results in drug resistance in tumor cells (11, 12, 28), which means that higher drug doses need to be used for efficient killing of tumor cells. This results in severe side-effects due to toxicity to normal tissue. Sensitization of tumor cells to conventional therapies has been explored by various researchers in the recent past and has shown some promise. Ablation of cyclin D1 expression using antisense molecules has been shown to confer chemosensitivity of tumor cells towards drugs such as cisplatin (14). However, RNA interference has been shown to be much more potent than antisense molecules. siRNA silencing of either eIF4E or CCND1 combined with cisplatin treatment was shown to bring about increased cell kill suggesting that these siRNAs alone or in combination with conventional cytotoxic agents such as cisplatin may be useful for therapy of HNSCC (17).
Cyclin D1 silencing sensitizes NT-shCyD1 tumors to cisplatin in vivo. NT-VC and NTshCyD1 cells were grown as xenografts in nude mice and divided into four groups with n=5 mice per group. Cisplatin was injected i.p. at 2 mg/kg body weight, once a week for four weeks. Mean tumor volume ratio: tumor volume after the treatment/tumor volume before the treatment. NT-shCyD1 tumors showed significant (*p≤0.01) reduction in size when compared to NT-VC tumors treated with cisplatin. The bar shows the unaltered tumor volume ratio (ratio=1).
The use of RNAi has become an important tool in current laboratory research, and there are increasing efforts to bring RNAi into clinics, given its specificity and ability to potently silence the gene of interest. A combination of such a targeted approach for gene silencing with conventional treatments such as chemotherapy and radiation therapy to cure cancer becomes important because conventional treatments are known to cause normal tissue toxicity. Combination therapy may enable doses of chemotherapeutic drugs or radiation to be reduced.
In the present study we used CCND1 shRNA to achieve its efficient and stable silencing in HNSCC cells overexpressing cyclin D1. CCND1 silencing was sequence specific as was evident by unaltered levels of cyclin D2. Silencing of Cyclin D1 brought about reduced phosphorylation of RB. Silencing CCND1 a major regulator of G1 phase progression, resulted in slower progression of NT-shCyD1 cells compared to NT-VC cells through G1 phase of the cell cycle as evidenced by flow cytometric analysis. This slower progression can be attributed not only to the silencing of CCND1, but also to the resulting hypophosphorylation of pRB. This slower progression through the G1 phase was concurrent with the slower growth rate of NT-shCyD1 when compared to NT-VC cells. Thus our data are in line with the earlier observations that CCND1 silencing has a negative effect on proliferative properties of cancer cells (17).
Overexpression of cyclin D1 results in reduced cell size (24). To the best of our knowledge, no study has described the effect of CCND1 down-regulation on cell size. In the present study we observed an increase in cell size of NTshCyD1 cells compared to NT-VC cells. Recently, the AKT/mTOR pathway has been implicated in the control of cellular size and proliferation (29, 30). Both of these molecules fall under EGF-R signaling pathways, which are again implicated in cellular proliferation (31). It is important to note that the activation of EGF-R pathway culminates in transcription of genes associated with increased cell proliferation and angiogenesis, by exerting its effect on genes such as CCND1 and Vascular Endothelial Growth Factor (VEGF) (31). There are reports that the mTOR inhibitor rapamycin diminished AKT-mediated increases in cell size, mitochondrial membrane potential, and cell survival (29). Furthermore, cyclin D1 regulates mitochondrial function in vivo by inhibiting mitochondrial activity and aerobic glycolysis (32). In the light of this knowledge, it is intriguing to investigate if cyclin D1 is involved in the regulation of cellular size.
Various groups have shown that attenuation of cyclin D1 expression can have profound antitumor effects. Antisense molecules have been reported to inhibit the expression of cyclin D1 in tumor cells, retard their proliferation (33) and to reduce the tumorigenicity of xenografts or increase their sensitivity to drugs such as cisplatin (13-15). Owing to its specificity and usage at low doses to bring about the same silencing effect as antisense molecules, RNAi seems to be a better suited molecule. siRNA duplexes have been demonstrated to lower drug resistance (17), but the stability of RNA duplexes, once delivered in vivo, remains a challenge. Vector driven RNAi serves as an effective alternate to siRNA duplexes. Not only is vector-driven RNAi as effective as siRNA duplexes in silencing the desired gene, but it is also cost effective and easy to handle (34).
Recently CCND1 shRNAs expressed from lentiviral vectors have been shown to increase sensitivity of HNSCC to cisplatin (35), however, several ethical issues remain to be addressed before lentiviruses can be routinely used in clinics. Furthermore, in the same study, cell lines which were made resistant to cisplatin by exposure to cisplatin were used. In such cases, it is difficult to predict whether the observed resistance to cisplatin is actually a result of cyclin D1 overexpression. Thus it is important to assess the effect of silencing druggable targets such as cyclin D1 in cells/tumors which mimic as closely as possible clinical manifestations.
In our studies, silencing of cyclin D1 in NT8e cells using vector-driven RNAi resulted in enhanced chemosensitivity to cisplatin. NTshCyD1 cells also took longer (≈delay of one week) to produce these tumors compared to NT-VC cells. When treated with 2 mg/kg of body weight of cisplatin, NTshCyD1 tumors regressed dramatically compared to NT-VC tumors which either grew slowly upon treatment or remained unaffected. Untreated tumors arising from NTshCyD1 cells grew slowly, while untreated tumors arising from NT-VC cells grew fast. To experimentally reduce the tumor burden in nude mice, a suboptimal concentration of cisplatin (2 mg/kg body weight) was used while higher cisplatin concentrations (up to 5 mg/kg body weight) have been used in other studies (35).
It is surprising why cyclin D1 ablation in cells overexpressing or ‘addicted’ to cyclin D1 should be deleterious to them. Classified as a classical cell cycle regulatory gene, mechanistically, it seems unlikely that such a molecule should increase the chemosensitivity. Recently, a well-orchestrated series of experiments by Jirawatnotai et al. (36) show that cyclin D1 has an unexpected and previously undescribed function in DNA repair. Their studies show that cyclin D1 is brought to damaged chromosomal sites by Breast Cancer 2 protein (BRCA2), and facilitates the recruitment of Homologous Recombination repair protein RAD51. They further show that this function is independent of its CDK activity, which further gives superiority to cyclin D1 as a target over CDKs. Thus disruption of cyclin D1 function explains why it, not only, slows down tumor growth, but also sensitizes tumors to DNA-damaging agents. With the myriad of functions cyclin D1 possesses inside the cells, it is of great importance to consider cyclin D1 as a therapeutic target to achieve efficient tumor regression especially in HNSCC, where an elevated levels of cyclin D1 is a well-known cause of poor treatment outcome. Furthermore, the promise of combined conventional therapy and targeted therapies based on shRNAs has a great relevance to achieving such a regression at lower drug concentrations. Once optimized in clinics, it holds the potential of benefiting patients by reducing the toxic side effects of conventional treatments.
Acknowledgements
This work was supported by a grant from the Indo-French Centre for Promotion of Advanced Research (IFCPAR), New Delhi. VK was recipient of a fellowship from the Council for Scientific and Industrial Research, GOI, India. The Authors would like to thank Mr. Ganesh V. Joshi for his help with the animal experiments.
Footnotes
-
Disclosure Statement
None.
- Received October 5, 2011.
- Revision received November 28, 2011.
- Accepted November 30, 2011.
- Copyright© 2012 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
