Quinacrine upregulates p21/p27 independent of p53 through autophagy-mediated downregulation of p62-Skp2 axis in ovarian cancer

We have previously shown that the anti-malarial compound Quinacrine (QC) inhibits ovarian cancer (OC) growth by modulating autophagy. In the present study we extended these studies to identify the molecular pathways regulated by QC to promote apoptosis independent of p53 status in OC. QC exhibited strong anti-cancer properties in OC cell lines in contrast to other anti-malarial autophagy inhibiting drugs. QC treatment selectively upregulated cell cycle inhibitor p21, and downregulated F box protein Skp2 and p62/SQSTM1 expression independent of p53 status. Genetic downregulation of key autophagy protein ATG5 abolished QC-mediated effects on both cell cycle protein p21/Skp2 as well as autophagic cargo protein p62. Furthermore, genetic silencing of p62/SQSTM1 resulted in increased sensitivity to QC-mediated apoptosis, downregulated Skp2 mRNA and increased accumulation of p21 expression. Likewise, genetic knockdown of Skp2 resulted in the upregulation of p21 and p27 and increased sensitivity of OC cells to QC treatment. In contrast, transient overexpression of exogenous p62-HA plasmid rescued the QC-mediated Skp2 downregulation indicating the positive regulation of Skp2 by p62. Collectively, these data indicate that QC-mediated effects on cell cycle proteins p21/Skp2is autophagy-dependent and p53-independent in high grade serious OC cells.


Results
Autophagy modulators affect cell cycle differentially. Our previous finding showed that QC induces autophagy in OC cells. To determine if other autophagy modulators have similar effects, we first analyzed autophagy and cell cycle related proteins by Western blot analysis (Fig. 1A). We observed that when C13 cells are treated with QC, Bafilomycin A (BafA) and 3-methyladenine (3-MA) at indicated doses, QC treatment resulted in degradation or downregulation of p62 expression in C13 cells whereas BafA and 3-MA in stabilization of p62 expression (Fig. 1A). LC3B was also stabilized by treatment with both QC and BafA. Importantly, QC but not BafA or 3-MA treatment stabilized the cell cycle inhibitor p21 expression in these cells. Furthermore, QC treatment downregulated Skp2, an F-box protein known to target p21/p27 for degradation, while BafA and 3-MA showed no effect on Skp2/p21 expression. These data indicate that while autophagy inhibitors BafA and 3MA stabilized p62 expression, QC treatment downregulated p62 expression, which is an indication of autophagy induction 9 . Since p21 is a direct target of p53 and QC has been reported previously to stabilize p53 11 , we next examined whether QC is able to promote these observed changes in a p53-dependent manner. QC treatment of p53 null SKOV3ip1 and p53 mutant OV90 (S215R) cell lines resulted in the downregulation of p62, Skp2 and upregulation of p21 expression ( Fig. 1B and C). These data indicate that QC-mediated effects on p62, Skp2 and p21 expression are independent of p53 expression. To further verify these findings, we generated stable clones expressing wild type p53, R249S, R273H and R280H mutant forms of p53 in SKOV3ip1 cells that are void of p53 expression as described in the materials and methods. Western blot analysis showed that QC treatment downregulated p62 and Skp2 expression and upregulated p21 and LC3B expression in these cells irrespective of p53 status (Fig. 2). Of note, QC treatment also stabilized p53 expression in wild type and mutant p53 expressing SKOV3ip1 stable clones further confirming that QC-mediated effects are independent of p53 status. This is consistent with previous reports that QC stabilization of p53 results in p53-dependent and p53-independent tumor cell death 12 .

QC-mediated p62 downregulation is independent of proteasomal degradation. It has been
shown that p62 is degraded via proteasomal-as well as autophagic-dependent manner 30 . Therefore, to determine whether QC-mediated p62 degradation is proteasome-dependent, we co-treated cells with QC and proteasomal inhibitors such as lactacystin and velcade for 24 hours. Western blot analysis revealed that QC-mediated p62 degradation was not altered upon inhibition of proteasomes with either 10 µM lactacystin or 2 µg/ml velcade treatment. Importantly, inhibition of proteasomal activity further stabilized p21 and p27 expression (Fig. 3A). We suspected that the induction of p21 and p27 expression in response to QC treatment might also inhibit CDK activities. Therefore, we determined the phosphorylation status of CDK substrates by Western blot analysis using anti-phospho CDK substrate antibody. Treatment with QC attenuated phosphorylation of several CDK substrates indicating that QC blocked CDK activities irrespective of proteasomal inhibition (Fig. 3A). Similarly, Western blot analysis using anti-ubiquitin antibody revealed that proteasomal inhibitors arrested proteins modified with poly-ubiquitinated chains demonstrating the efficacy of proteasomal inhibitors. These data indicate that QC-mediated p62 downregulation is independent of proteasomal degradation.
To determine whether QC-mediated p62 degradation is autophagy dependent, we utilized mouse embryonic fibroblasts (MEFs) lacking ATG5 expression. ATG5 plays a critical role in promoting autophagy and it has been shown that depletion of ATG5 inhibits autophagy 31,32 . Therefore, we next examined the effect of QC treatment on wild type and ATG5 null MEFs. Western blot analysis revealed that QC treatment promoted p62 downregulation, induced LC3B expression and increased apoptosis as indicated by increased PARP cleavage in WT MEFs but not in ATG5 null MEFs (Fig. 3B). Skp2 levels were too low to be detected in wild type MEFs whereas they were elevated in p62 null MEFs. Collectively, these data indicate that ATG5 is essential for QC-mediated effects. To further verify these effects, we next investigated the effect of QC on shRNA-mediated ATG5 depleted C13 cells. C13 cells were stably selected with shRNA against ATG5 and NTC respectively as described in materials and methods. Western blot analysis confirmed complete knockdown of ATG5 in C13HBF2 stable clone (Fig. 3C). While QC treatment downregulated p62 and Skp2 expression in NTC-C13 cells, it did not affect p62 and Skp2 Similarly, significant upregulation of p21 expression was observed in QC-treated NTC-C13 cells but not in ATG5 depleted C13 cells. Consistent with these results, QC treatment in ATG5 depleted cells resulted in diminished apoptotic response as reflected by lower level of cleaved PARP and cleaved caspase 3 expression (Fig. 3C). Importantly, no change in LC3B induction was observed in NTC and ATG5 depleted clones. Taken together, these data indicate that QC-mediated apoptosis and autophagic response is dependent on ATG5 levels.
We next wanted to examine the temporal regulation of autophagy and apoptosis upon treatment with QC. Western blot analysis of QC treated C13 and Hey8MDR cells show that p62 expression was downregulated as early as 4 hours following treatment in C13 cell and 8 hours in HeyA8MDR cells, whereas appearance of cleaved PARP was detected in a 16 to 24 hours window (Fig. 3D). These data suggest that QC treatment triggers autophagic clearance of p62 leading to apoptosis in both C13 and HeyA8MDR cancer cells.
To determine whether PARP cleavage was due to the activation of initiator caspases as well as effector caspases, we did immunoblotting in C13 and HeyA8MDR cells treated with 5.0 and 10 µM QC for 24hrs. We observed activation of Caspase 8 and Caspase 9, and increased the activity of Caspase 3, indicating that both extrinsic and intrinsic apoptotic pathways are activated by QC (Fig. S1).
p62 knockdown promotes QC-mediated p21 upregulation via Skp2. We have previously reported that QC downregulated p62 expression and that p62 expression plays a critical role in promoting cell survival 13 . Since we observed parallel changes in p62, Skp2 and p21 expression upon QC treatment, we next wanted to determine the relationship between p62 and cell cycle inhibitor p21. To further determine the role of p62 in QC-mediated cell death, we generated p62 knockdown cells via lentiviral mediated shRNA in OV2008 cells as described in materials and methods. We treated NTC and p62 knockdown OV2008 cells with QC and evaluated the effect on apoptosis and cell cycle proteins. Western blot analysis showed that treatment with QC resulted in extensive apoptosis in p62 depleted cells when compared to NTC OV2008 cells as reflected in the levels of cleaved PARP (Fig. 4A). Importantly, p62 knockdown also resulted in diminished basal levels of Skp2 which was further downregulated upon QC treatment in OV2008 cells. The decrease in Skp2 expression was also found to be associated with an increase in p21 expression which was further increased upon QC treatment in p62 depleted OV2008 cells. In order to confirm that the effects are not specific to only OV2008 cancer cells, we generated NTC and two separate p62 shRNA expressing stable clones in C13 cells. Consistent with our previous findings, QC treatment downregulated p62 and promoted increased apoptosis in p62 knockdown clones as indicated by cleaved PARP expression without causing any significant change in autophagy in NTC and p62 shRNA clones. Interestingly, knockdown of p62 in C13 cells also resulted in p21 upregulation and Skp2 downregulation, changes that were further enhanced upon QC treatment (Fig. 4B). Together, these data indicate that p62 positively regulated Skp2 levels in these cells.
QC-mediated p21 stabilization is post-transcriptional. We next investigated whether the increase in p21 expression was due to its increased transcription or post-translational stabilization. To this end, we transiently transfected pcDNA3.1-p21 Flag expression vector in OV2008 NTC and OV2008 p62 shRNA cells followed by QC treatment. Western blot analysis showed that QC treatment, upregulated p21-Flag expression in NTC as well as p62shRNA cells (Fig. 5A). Interestingly, p21 expression was high at basal levels in p62 knockdown cells and was significantly more upregulated upon treatment with QC when compared with NTC cells. These data indicate that p21 was stabilized at the post-transcriptional level upon QC treatment.
We further tested the effect of p62 expression on p21 stability. For this purpose, we utilized two contrasting cellular models 1) ATG5 depleted C13 cells to achieve high levels of cellular p62 and 2) p62 knockdown c13 cells showing depleted cellular p62 levels. The cells were treated with 20 µg/ml cyclohexamide (CHX) for indicated time intervals. Western blot analysis showed that ATG5 knockdown C13 cells had increased levels of p62 and Skp2 expression which remained stable up to 6 hours of cyclohexamide treatment, whereas NTC cells exhibited increased degradation of p62 (Fig. 5B). Interestingly, CHX treatment had more pronounced effect on stability of p21 levels in ATG5 depleted cells harboring high levels of both p62 and Skp2. Half-life of p21 was markedly lower in ATG5 deficient cells. In contrast, p62 knockdown C13 cells showed high levels of p21 expression owing to lower levels of Skp2 thereby enhancing the half-life of p21 (Fig. 5C). These experiments further confirm the mechanistic basis of QC-mediated cellular changes involving the p62-Skp2 axis.
QC promotes Skp2 downregulation in an autophagy dependent manner. In order to determine if Skp2/p21 regulation by QC and p62 is autophagy dependent given that p62 is a bona fide autophagy cargo protein, we co-treated C13 cells with QC and autophagy inhibitor BafA and then checked Skp2 expression. Western blot analysis revealed that QC downregulated p62, Skp2 and upregulated p21 and p27 respectively (Fig. 6A). Co-treatment with BafA rescued QC mediated downregulation of p62 and Skp2 expression. Consistent with these data, BafA also blocked QC mediated upregulation of p21 and p27. To determine whether BafA also rescued Skp2 levels at transcriptional level, the above treated samples were subjected to qPCR as described in materials and methods. Real time PCR analysis revealed that co-treatment with BafA rescued QC mediated Skp2 mRNA downregulation (Fig. 6B). These data support the hypothesis that autophagy inhibition with BafA prevents the downregulation of Skp2 thereby resulting in downregulation of the cell cycle inhibitors p21 and p27. Furthermore, our data show that QC treatment downregulated Skp2 at the transcriptional level. We next determined whether p62 overexpression in p62 depleted cells will rescue its effects. For this purpose, we generated C13 cells stably expressing shRNA targeting 3′UTR region of p62 which were then transfected with pcDNA-p62-HA plasmid. Western blot analysis revealed efficient p62 knockdown in these cells (Fig. 7A). Transfection of full length p62 plasmid restored expression of Skp2 and downregulated p21 expression in p62 depleted cells. These data indicate that p62 positively regulates Skp2 expression. We further evaluated whether p62 knockdown affects Skp2 mRNA. To this end, we utilized NTC and p62 knockdown C13 cells treated with QC. Real time analysis showed that QC treatment as well as p62 knockdown downregulated Skp2 mRNA in C13 cells (Fig. 7B), whereas there was no effect of QC on p62 mRNA (Fig. 7C).
Skp2 knockdown promotes QC mediated p21 upregulation. Our data shows that p62, by positively regulating cell cycle protein Skp2 leads to the degradation of p21 and p27. To gain further insight into whether Skp2 directly regulated p21 and p27 in OC cells, we generated stable knockdown of Skp2 in C13 (p53 wild type), OV90 (p53 mutant) and SKOV3ip1 (p53 null) cells. Two separate shRNA against the coding region of Skp2 was utilized and stable knockdown was determined by western blot analysis. Knockdown of Skp2 in C13 and OV90 cells resulted in upregulation p21 expression. In contrast, Skp2 knockdown in SKOV3ip1 cells did not show any change in p21 expression (Fig. 8A). To further understand the effect of QC in this context, we treated NTC and Skp2 shRNA OV90 and SKOV3ip1 cells with increasing concentrations of QC. Western blot analysis revealed that while Skp2 knockdown stabilized p21 in OV90 cells, QC treatment upregulated p21 in Skp2 knockdown cells both in OV90 (Fig. 8B) and Skov3ip1 cells (Fig. 8C). While these data demonstrate that QC mediated upregulation of p27 and p21 is Skp2-dependent, it appears that there may be additional mechanisms that could contribute towards upregulation of p21 in SKOV3ip1 cells. Equally important is the observation that there was no effect on QC mediated p62 degradation in these Skp2 deficient cells which also exhibited increased PARP cleavage indicating that Skp2 is downstream of p62. We observed activation of Caspase 8 and Caspase 9, and increased the activity of Caspase 3, indicating that both extrinsic and intrinsic apoptotic pathways are activated by QC (Fig. S1).

Discussion
Autophagy modulators have been used against several diseases including cancer. In this study, we have characterized the cellular effects mediated by anti-malarial drug Quinacrine (QC) in OC cell lines. Previously, we have shown that QC effectively attenuated growth of OC cells both in vitro and in vivo by promoting autophagic mediated cell death 13 . Although QC has been shown to modulate autophagy, the cellular mechanisms responsible for mediating its effects are not clearly defined. We have now uncovered a unique mechanism of action for QC that affects two major players of two distinct pathways (namely autophagy and cell cycle) critical for supporting proliferation and survival of cancer cells. Mechanistically, we show that QC has two important cellular targets p62/SQSTM1 and F-box protein Skp2. Although QC has been shown to exert antitumor activity in several solid tumor cell line models, this is the first paper to show that QC-induced autophagic degradation of p62 leads to cell cycle inhibition by upregulating CDK inhibitors p21 and p27 expression in OC cells. In addition, it is a common perception that autophagy and cell cycle arrest are a result of stress-induced nutrient deprivation and/or a   15 . Our data thus challenges the notion that autophagy is essentially a pro-survival pathway supporting cancer growth.
P62 is a known bona fide substrate for autophagy. Induction of autophagy triggers breakdown of several cellular proteins leading to either cell survival or cell death in a context dependent manner 6,10,[33][34][35] . In this study, we show that QC-mediated downregulation of p62 and Skp2 expression promoted apoptosis. Autophagy inhibition, but not proteasomal inhibition, rescued QC-mediated p62 degradation. Similarly, ATG5-null MEFs exhibited increased levels of p62, Skp2 and no autophagic accumulation of LC3B in response to QC treatment. More importantly, we show that QC-mediated effects on p62, Skp2 and p21 are independent of p53 status. Data also indicated that QC-mediated autophagic degradation of p62 was intact in cells lacking functional p53. Although QC has been shown to stabilize p53 and promote apoptosis, our data shows that QC-mediated effects on autophagy and cell cycle inhibitors are independent of p53. More specifically, we show that QC treatment upregulated p21 (a transcriptional target of p53) at the protein level. These data emphasize that downregulation of both critical cell survival proteins p62 and Skp2 is required for QC-mediated autophagy. Additionally we also show that QCmediated cell death involves both the intrinsic and extrinsic pathways to induce apoptosis.
Both p62 and Skp2 are well known for their proliferative and pro-tumorigenic roles in multiple cancers. It is important to note that they are both elevated in OC and have been shown to serve as poor prognostic factor in overall survival in this patient population 36,37 . Importantly, elevated levels of both p62 and Skp2 have been associated with therapeutic resistance [38][39][40] and as such. efforts are currently underway to identify small molecule inhibitors 37,39,[41][42][43] to block their function and/or activity. While several reports indicate that autophagy plays a pro-survival role in a variety of cancers 44 , other studies indicate that autophagy induced by anti-neoplastic agents results in cell death 44 . In our experimental conditions, QC treatment induced autophagy and resulted in cell death in multiple OC cells. By using ATG5 deficient MEFs and specific shRNA targeting ATG5 in C13 OC cells, we showed that induction of apoptosis by QC was completely blocked in MEFs and to a significant degree in C13 cells depleted of ATG5 expression. The observed modest change in C13 can be attributed to the presence of other ATG related genes that could play role in autophagy. Nevertheless, ATG5 depletion had effect on QC mediated apoptosis and on its cellular targets. Consistent with these findings, QC treatment in ATG5 deficient cells or co-treatment with autophagy blocker Bafilomycin A in C13 cells also attenuated degradation of p62 as well as downregulation of Skp2 expression.
It is not surprising that downregulation of Skp2 expression by QC was associated with upregulation of cell cycle inhibitors such as p21 and p27 given that Skp2 has been shown to promote proteasomal degradation of p21 and p27. It has been also shown that autophagy is associated with several phases of the cell cycle 45 . However, the molecular mechanism linking both of these processes is not clearly defined. Our data provided explanation of p53 independent p21 upregulation observed in p53 null/mutated OC upon QC treatment. Other investigators have also reported QC mediated cell cycle inhibition in other cancer types 46 . Our data expands this understanding and , shRNA2 (HAX50) were exposed with QC at 5 and 10 µM for 24 hours. Cells were harvested and subjected to Western blot analysis using anti-p27, anti-p21, anti-LC3B, anti-cleaved PARP, and anti-GAPDH antibodies. (C) Skovip3 NTC and Skp2 shRNA1 (HAX49), shRNA2 (HAX 50) were exposed with QC at 5 and 10 µM for 24 hours. Cells were harvested and subjected to Western blot analysis using anti-p62/SQSTM1, anti-Skp2, anti-p21, anti-p27, anti-LC3B, anti-cleaved PARP, anti-PDI and anti-GAPDH antibodies.
SCiENtifiC RepoRts | (2018) 8:2487 | DOI:10.1038/s41598-018-20531-w highlights the presence of a relationship between cell cycle and autophagy. By blocking autophagy by chemical inhibitor Bafilomycin A, we demonstrate that QC's effects on p21 and Skp2 were reversed. This prompted us to ask whether p62 downregulation has any role in Skp2/p21 regulation. Genetic silencing of p62 downregulated Skp2 expression and caused upregulation of cell cycle inhibitors p21 and p27. Further treatment with QC resulted in extended half-life of p21 in p62 and Skp2 deficient cells thereby resulting in significant upregulation of these cell cycle inhibitors. Further, we confirmed that p62 positively regulated Skp2 mRNA. These findings reveal the connection between autophagic substrate p62 and Skp2 and hence cell cycle inhibitor p21. Ultimately, our work indicates a link between autophagy and cell cycle. Additionally, we showed temporal regulation of apoptosis and autophagy upon QC treatment. Onset of autophagy in the form of p62 degradation was present as early as 4 hours of QC treatment in C13 cells whereas apoptotic marker appeared at later time points. These findings suggest that autophagy is an early indicator leading to apoptosis upon QC treatment.
Upon genetic silencing of either p62 and/or Skp2 expression, OC cells became sensitive to QC treatment. This was associated with pronounced upregulation of cell cycle inhibitors p21 and p27 (Model in Fig. 9) in three different OC cell lines carrying wild type p53, mutant p53 and p53 null. These findings along with other findings using SKOV3 cell lines overexpressing wild type and mutant p53 indicated that QC mediated effects on autophagy and cell cycle were largely independent of p53. However, it is plausible that presence of p53 and its resulting stabilization by QC might further promote apoptotic effects. Indeed, previous reports indicated that QC stabilized p53 without causing any genotoxicity and that it promoted apoptosis in colon, breast and cervical cancers [47][48][49] .
In conclusion, in this study, we showed that QC is able to induce apoptosis by modulating autophagy and cell cycle independent of p53 status. This is particularly clinically relevant given that the majority of high grade serous OC harbors p53 mutations. Importantly, our work provides additional insight into the mechanisms by which QC arrests cell growth in OC. Finally, these findings suggest that Skp2, an important oncogene that is overexpressed in OC and is associated with chemo resistance, may be a novel target for QC treatment.

Conclusions
We demonstrate that QC treatment promotes anti-tumorigenic effects in OC cells by downregulating the protein expression of p62 and Skp2 and by triggering the protein expression of p21 and p27. We propose a novel molecular mechanism of action for QC in OC where the cell cycle related protein p21 and the oncogene Skp2 are both altered upon QC treatment in an autophagy dependent but p53 independent manner.

Materials and Methods
Cell culture. SKOV3, C13, OV2008, OVCAR 3 and 293 T cells were grown in recommended growth media according to American Type Culture Collection (ATCC) (VA, USA) as described previously 13 . OV90 cells were grown in OSE media consisting of 50:50 medium 199:105 (Sigma-Aldrich, St. Louis, MO) supplemented with 15% fetal bovine serum, and sodium bi-carbonate (2.2 gm/L). Antibiotic penicillin and streptomycin was added in all the growth media (Thermo Fisher Scientific, Waltham, MA, USA). All cell lines were kept in humidified incubator at 37 C with 5% CO 2 . The authenticity of the cell lines used in this report was confirmed by STR genotyping at the Genome Analysis Core, Rochester, MN. ATG5 wild type and null mouse embryonic fibroblasts were gift from Dr. Dan Billadeau, Mayo Clinic, Rochester, MN.  Table 1. Densitometry analyses were performed using ImageJ, and graphs were plotted using GraphPad Prism.

Transient Transfections. OC cells were transiently transfected with Lipofectamine 2000 (Thermo Fisher
Scientific, MA, USA) according to manufacturer's protocol. Briefly, plasmids were transfected in cancer cells in serum free medium followed by addition of serum containing medium. After 48 hours of transfections, cells Western Blot analysis. Protein lysates were prepared by lysing the OC cell lines in lysis buffer (Cell Signaling Inc, MA, USA) containing 50 mM Tris-HCl (pH 7.4), 350 mM NaCl, 0.25% Nonidet P-40, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM glycerol phosphate, 1 mM sodium orthovanadate, and 30% glycerol with protease inhibitors. Protein was estimated by BCA method, separated on SDS-PAGE and transferred onto PVDF membrane using Trans-Blot Turbo (Bio-Rad, CA, USA). The membranes were blocked with 5% BSA in TBST for 1 hour (50 mM Tris-HCL, pH 8.0, 10 mM NaCl and 0.1% Tween 20) at room temperature followed by incubation with the indicated antibodies at 4 °C overnight. Membranes were washed with TBST and incubated with either mouse or rabbit-800 IR dye and finally scanned under Odyssey Fc Imaging system (Bio-Rad, CA, USA). Experiments were conducted three times.
RNA isolation and cDNA synthesis. RNA extraction was performed using Qiagen RNA isolation kits following manufacturer's instruction. 1 µg of RNA was reverse transcribed using the quantitect reverse transcription cDNA synthesis kit following manufacturer's instruction (Qiagen, MD, USA).
Quantitative real time PCR. Quantitative real-time PCR (qRT-PCR) was carried out using SYBR-Green PCR Master Mix (Applied Biosystems, Foster City, CA, USA), using the CFX96 TouchTM Real-Time PCR detection System (Bio-Rad, CA, USA). Oligonucleotides were synthesized by Integrated DNA Technologies (IDT). Primer sequences for the genes analyzed are p62FP:5′-TGAAACACCGGACATTCGG-3′, p62RP: 5′-TCAGGAAATTCACACTCCGGATC, Skp2FP:5′CTGGGTGTTCGTGATTCTCTG-3′, Skp2RP: 5′-GCTGGGTGATGGTCTCTG-3′ and GAPDH FP:5′-ACATCGCTCAGACACCATG-3′ and GAPDH RP:5′-TGTAGTTGAGGTCAATGAAGGG-After an initial step of 3 min denaturation at 95 °C, the amplification conditions were 45 cycles of 95 °C 10 sec, for denaturation and 55 °C 30 sec for annealing and elongation, with a final extension at 72 °C for 10 min. Normalization across samples was performed using the average of the constitutive gene human GAPDH primers and calculated by 2 −ΔΔCt method as previously described 50 . Binding efficiencies of primers sets for both target and reference genes were similar. All samples were run in triplicates and repeated twice. Expression levels of the genes in the two designated groups were analyzed by an unpaired t test using GraphPad PRISM (version 6.0; GraphPad Software).