Inhibition of PLK4 might enhance the anti‐tumour effect of bortezomib on glioblastoma via PTEN/PI3K/AKT/mTOR signalling pathway

Abstract Glioblastoma (GBM) is one of the most common aggressive cancers of the central nervous system in adults with a high mortality rate. Bortezomib is a boronic acid–based potent proteasome inhibitor that has been actively studied for its anti‐tumour effects through inhibition of the proteasome. The proteasome is a key component of the ubiquitin‐proteasome pathway that is critical for protein homeostasis, regulation of cellular growth, and apoptosis. Overexpression of polo‐like kinase 4 (PLK4) is commonly reported in tumour cells and increases their invasive and metastatic abilities. In this study, we established a cell model of PLK4 knockdown and overexpression in LN‐18, A172 and LN‐229 cells and found that knockdown of PLK4 expression enhanced the anti‐tumour effect of bortezomib. We further found that this effect may be mediated by the PTEN/PI3K/AKT/mTOR signalling pathway and that the apoptotic and oxidative stress processes were activated, while the expression of matrix metalloproteinases (MMPs) was down‐regulated. Similar phenomenon was observed using in vitro experiments. Thus, we speculate that PLK4 inhibition may be a new therapeutic strategy for GBM.

Drug Administration (FDA) for treatment of refractory multiple myeloma (MM). Previous studies have shown that bortezomib disrupts multiple downstream signalling pathways, such as the NF-κB signalling pathway and ubiquitin-proteasome pathway, and has an important role in the regulation of cell cycle, mitosis, cell viability, proliferation and apoptosis in glioblastoma cells. [6][7][8] It also blocks autophagy flux mediated by the 26S proteasome, thereby inhibiting the elimination of damaged organelles, 9 and inhibits the proliferation of glioblastoma. Polo-like kinase 4 (PLK4) is critical for embryonic development and cell cycle control. PLK4 localizes to the nucleolus during the G 2 phase, moves into the centrosomes during early M phase, and then into the cleavage furrow during cytokinesis. 10 The expression of PLK4 remains low in G 0 , and in early-to-mid-G 1 phase, while it increases in late G 1 , S and G 2 phases and is at the highest level during mitosis. 11 A previous study reported that PLK4 knockdown reduces invasion, induces epithelial phenotype in breast cancer cells and plays an important role in cancer metastasis. 12 Another study found that increase in PLK4 expression enhances the resistance of GBM cells to radiotherapy, and inhibition of PLK4 enhances the effect of temozolomide via reduction of IKBKE phosphorylation; besides, depletion of PLK4 in glioma cells using CRISPR/CAS9 significantly decreased the proliferation, viability and survival of glioma cells, increasing the susceptibility of cancer cells to DNA damage agents. 13,14 Another study reported that a combination of bortezomib and PLK inhibitor showed a significantly higher anti-tumour effect compared with bortezomib treatment alone in multiple types of head and neck cancers. 15 Therefore, we hypothesized that inhibition of PLK4 might enhance the therapeutic effect of bortezomib in the treatment of GBM. In this study, we found that bortezomib could significantly inhibit the proliferation of glioma cells using MTT assay and further found that this effect was mediated by the PTEN/ PI3K/AKT/mTOR signalling pathway. We also found that the apoptotic pathway and oxidative stress response were activated following bortezomib treatment, while the degradation of extracellular matrix (ECM) was decreased. Furthermore, inhibition of PLK4 enhanced the effect of bortezomib in glioma cells, while overexpression of PLK4 reduced this effect. Using in vitro experiments, we found similar phenomena as in the xenograft experiments. Thus, we speculate that inhibition of PLK4 expression may be a new therapeutic strategy for GBM. were purchased from Yeasen. Anti-PLK4 (ab137398), PTEN (ab32199), ATM (ab78), p-ATM (ab81292), ATR (ab2905), p-ATR (ab227851), AKT (ab179463), p-AKT (ab38449), p-mTOR (ab109268), mTOR (ab2732), HIF-1α (ab1), MMP2 (ab37150), MMP9 (ab73734), Bcl-2 (ab692), Bax (ab32503), Superoxide Dismutase 1 (ab13498), Thioredoxin (ab26320), caspase-3 (ab13847) and cleaved caspase-3 (ab2302) antibodies were purchased from Abcam. FGF (ab99979), EGF (ab217772), TGF-β (ab100647) and VEGF (ab222510). ELISA kits were purchased from Abcam. ROS detection kit (E004-1-1) was purchased from Nanjing Jiancheng Bioengineering Institute. DNA Damage Quantification Colorimetric Kit (K253-25) was purchased from Biovision. 293T (CRL-11268), LN-18 (CRL-2610), A172 (CRL-1620) and LN-229(CRL-2611) cells were purchased from American Type Culture Collection (ATCC).

| Ethical statement
Animal studies were performed according to the principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Animals, and the experiments in this study were approved by the Health Animal Care and Use Committee of the Shanxi Dayi hospital.

| MTT assay
Cells were seeded into a 96-well plate and cultured for 24 hours, followed by further incubation with 500 nmol/L bortezomib for 24 hours. MTT assay was performed accordingly as described in a previous study. 18 Cells were incubated with 5 mg/mL MTT reagent for 5 hours, and the optical density at 560 nm was measured using a CMax Plus microplate reader (MD). The number of cells was measured using a cell counter (Countess, Thermo) each day. A proliferation curve was constructed based on the cell counts.

| RNA extraction
RNA extraction was performed according to the following protocol.
Briefly, cells were lysed with lysis buffer and incubated at room temperature for 5 minutes. After incubation, the cell lysate was centrifuged at 12 000 g for 10 minutes (4°C). Then, chloroform was added into the cell lysate and oscillated for 15 seconds. The cell lysate was further centrifuged at 12 000 g for 10 minutes (4°C) and washed with 75% ethanol. The RNA was collected by centrifugation at 7500 g for 10 minutes (4°C) and stored at −80°C until it was used in experiments.

| Reverse transcription and real-time quantitative polymerase chain reaction (qPCR)
Reverse transcription was performed according to the following protocol. Briefly, the reaction mixture was made up as recommended and the reaction steps were set up as follows: reverse transcription:  with 5% skim milk at room temperature for 1 hours, followed by incubation with primary antibodies overnight at 4°C. Then, the membranes were incubated with the secondary antibody at room temperature for 1 hours. The grey value of each group was detected using chemiluminescence and analysed using the IPP 6.0 software.

| Flow cytometry
Cells were first grouped and treated as previously described. The cells in each group were fixed in 70% ethanol overnight followed by incubation with 20 μg/mL RNase A for 30 minutes at 37°C. After washing with PBS, the cells were incubated with 50 μg/mL at room temperature for 30 minutes away from light. The apoptosis cells were detected using FC500 flow cytometer (Beckman).

| Detection of cellular ROS concentration
Detection of cellular ROS concentration was performed according to the following protocol. Briefly, the cells and xenograft tissues were digested with trypsin and then washed with PBS three times. The cells were then incubated with the detection probe at a concentration of 10 μmol/L for 30 minutes. After resuspending in PBS, the fluorescence intensity was detected using a Multiskan FC microplate reader.

| Measurement of DNA damage
Cellular DNA was extracted using a DNA extraction kit (D1700, Solarbio

| Statistical analysis
The data are presented as the mean ± SEM Each experiment was repeated three times independently. One-way ANOVA was used to analyse the differences between groups using GraphPad 7.0 software. P value <.05 was set as a statistically significant difference.

| Effect of bortezomib on proliferation of glioma cells
As shown in Figure 1A, the viability of LN-229 cells decreased significantly in the PI group (P < .05) compared to the NC and TG groups; however, the viability in PO group did not change significantly (P > .05). The expression of PLK4 was significantly higher in the PO

| Effect of bortezomib on expression of apoptosis-related genes in glioma cells
As shown in Figure 2, the expression of apoptosis-related genes in the glioma cells was measured using qPCR method. In the LN-229 cells, compared to the NC group, the expression of TNF-α was significantly higher in the PI and PO groups (P < .05), and compared to the TG group, the expression was significantly higher in the PI group (P < .05). Compared to the NC group, the expression of caspase-9 was significantly higher in all treatment groups (P < .05), and compared to the TG group, the expression was significantly higher in the PI group (P < .05). The expression of caspase-3 was significantly increased in the PI and PO groups (P < .05), and compared to the TG group, the expression was significantly increased in the PI group (P < .05). The expression of Bax was significantly increased in all treatment groups compared to the NC group (P < .05), and compared to the TG group, the expression was significantly increased in the PI group (P < .05).
Similar trends were observed in the LN-18 and A172 cells. In serum samples, the expression of TNF-α was significantly increased in all treatment groups compared to the NC group (P < .05), and compared to the TG group, the expression of TNF-α was significantly increased in the PI group (P < .05). The change in expression of caspase-9 presented a similar trend to that of TNF-α. The expression of caspase-3 was significantly increased in the TG and PI group compared to the NC group (P < .05), and compared to the TG group, the expression of caspase-3 was significantly increased in the PI group (P < .05).
The expression of Bax was significantly increased in all treatment groups compared to the NC group (P < .05), and compared to the TG group, the expression of Bax was significantly increased in the PI group (P < .05).

| Detection of apoptotic cells by flow cytometry
As shown in Figure 3, apoptotic cells were detected using flow cytometry. In LN-299, LN-18 and A172 cells, the proportion of apoptotic cells were significantly higher in the TG and PI groups (P < .05) compared to the NC group, and significantly higher in the PI group compared to the TG group (P < .05). However, these changes were not significant following overexpression of PLK4.

| Effect of bortezomib on expression of caspase-3 and caspase-9
As shown in Figure 4, the ratio of cleaved caspase-3/caspase-3 in LN-299, LN-18, A172 cells and xenografts was significantly increased in TG and PI groups (P < .05) compared with NC group, and the ratio was decreased in PO group (P < .05) compared with NC group. And compared with TG group, the ratio of cleaved caspase-3/caspase-3 was significantly increased in PI group (P < .05) and decreased in PO group. The changing in expression of caspase-9 presented a similar trend with the ratio of cleaved caspase-3/caspase-3.

| Effect of bortezomib on activation of PTEN/ PI3K/AKT/mTOR signalling pathway in glioma cells
As shown in Figure 5, the expression of PTEN was significantly higher in the TG and PI groups (P < .05), and significantly lower in the PO group (P < .05) compared to the NC group. The expression of PTEN decreased significantly in the PO group (P < .05) compared to the TG group. The ratio of p-AKT/AKT decreased significantly in the PI and increased significantly in PO groups (P < .05) compared to the NC and TG groups. The ratio of p-mTOR/mTOR was significantly lower in the PI group (P < .05) and higher in PO group (P < .05) compared to the NC group. The ratio of p-mTOR/mTOR was significantly lower in PI group (P < .05) compared to the TG group. The ratio of p-ATR/ATR was significantly higher in all the treatment groups (P < .05) compared to the NC group and was higher in the PI group (P < .05) and was significantly lower in the PO group (P < .05) compared to the TG group. Changes in the ratio of p-ATM/ATM presented a similar trend as that of p-ATR/ATR. Similar trends were also observed in A172 and LN-18 cells (Figures 6, 7). In the xenografts (Figure 8)

| Effect of bortezomib on the expression of target proteins in glioma cells
As shown in Figure 9, the expression of HIF-1α was significantly lower in the PI group (P < .05) and significantly higher in the PO  (Figures 10, 11). In the xenografts (Figure 12)

| Effect of bortezomib on secretion of angiogenesis-related molecules in glioma cells
As shown in Figure 13 The concentration of TGF-β was significantly lower in the PI group compared to the NC and TG groups (P < .05). The concentration of EGF was significantly lower in the PI group compared to the NC and TG groups (P < .05). The concentration of FGF was significantly lower in all treatment groups compared to the NC group (P < .05) and was significantly lower in the PI group compared to the TG group (P < .05).

| Relative concentration of cellular ROS
As shown in Figure 14A, relative concentration of cellular ROS in LN-299 cells was significantly higher in the TG and PI groups (P < .05) compared to the NC group, and significantly higher in the PI group (P < .05) compared to the TG group. The concentration of cellular ROS was significantly lower in the PO group (P < .05) compared to the TG group, although it did not change significantly in the PO group compared to the NC group. Similar trends were observed in the LN-18 and A172 cells, and in the xenografts.

F I G U R E 11
Effect of bortezomib on activation of target proteins in A172 cells. A, Expression of HIF-1α, Bcl-2, Bax, SOD1, TRX, MMP-2 and MMP-9 using Western blotting. B, C, D, E and F, Quantitative analysis of each protein. Data were presented as mean ± SD (n = 3). NC, control group, TG, bortezomib treatment group, PI, bortezomib treatment combined with PLK4 inhibition group, PO, bortezomib treatment combined with PLK4 overexpression group. *P < .05 vs NC group; #P < .05 vs TG group

| Quantification of DNA damage
As shown in Figure 14B, DNA damage in LN-299 cells was significantly higher in all treatment groups (P < .05) compared to the NC group, and significantly higher in the PI group (P < .05) compared to the TG group. DNA damage was significantly lower in the PO group (P < .05) compared to the TG group. Similar trends were observed in the LN-18 and A172 cells, and in the xenograft tissues.

| Measurement of the diameter of xenograft tissues
As shown in Figure 15, the diameter of xenograft tissues was significantly decreased in TG group and PI group (P < .05) compared with NC group and was not significantly changed in PO group. And compared with TG group, the diameter of xenograft tissues was significantly decreased in PI group (P < .05) and significantly increased in PO group (P < .05).

| D ISCUSS I ON
Gliomas are primary tumours that originate from neuroglial stem or progenitor cells. They constitute nearly 30% of brain and central nervous system tumours, and 80% of all malignant brain tumours. 20 Gliomas are the most common primary intracranial tumours in adults, and half of all newly diagnosed gliomas are classified as glioblastomas, which are more common than low-grade gliomas. 21 The occurrence of glioblastoma is associated with a short survival and fatal outcome with limited treatment options. The PLK family was first discovered over 30 years ago 22  . 29 A previous study found that reducing mTOR activity increases HIF-1α activity and the production of reactive oxygen species (ROS), leading to the oxidative stress in cells. 30 Reduction in mTOR activity leads to the phosphorylation of 4E-BP1, resulting in dissociation with the cap-binding protein eukaryotic initiation factor 4E (eIF4E), and reduction of cap-dependent mRNA translation. 31 In this study, we found that following bortezomib treatment, the activation of PI3K/mTOR signalling pathway was inhibited, with increased expression of PTEN.
We also noticed that the activation of ATM/ATR was increased, and inhibition of PLK4 enhanced these effects. Hence, we speculate that HIF-α remains stable under normal oxygen conditions, while under hypoxia, HIF-α is hydroxylated by specific prolyl hydroxylase domain-containing proteins (PHDs) at two critical prolyl residues that results in its rapid degradation by the proteasome via ubiquitination. 33 However, the relationship between ROS and HIF-α remains obscure. A previous study found that hypoxia increases the production of ROS in the electron transport chain, resulting in increased stability and activity of HIF-1α via inhibition of PHD activity, 34 while another study found that increased production of ROS could promote the degradation of HIF-1α via the ubiquitin-proteasome system. 35 Here, we found that the expression of HIF-1α was decreased in the bortezomib treatment group, and therefore, we speculate that although hypoxia induces the produc-   is also a critical protease that has a vital role in multiple biological processes, especially in degradation of ECM through proteolytic cleavage to regulate ECM remodelling. 42 MMP-9 also plays an important role in regulation of tumour microenvironment via tumour invasion, angiogenesis and metastasis. 43 In this study, we found that bortezomib treatment decreased the expression of MMPs, contributing to the maintenance of ECM and basement membrane, which inhibits the invasive and metastatic ability of cancer cells, and results in an anti-tumour effect. Inhibition of PLK4 enhanced the effect of bortezomib in glioma cells. The MMPs are also regulated at the transcriptional level by multiple molecules, such as EGF, FGF, VEGF and TGF-β. 44 In this study, we found that the concentration of these cytokines decreased following bortezomib treatment, and this effect was further enhanced by inhibition of PLK4, resulting in the down-regulation of the MMPs.
In this study, we established a cell model for PLK4 overexpres-

ACK N OWLED G EM ENTS
None.

CO N FLI C T O F I NTE R E S T
None.