Background
As the most frequently seen primary intracranial tumour, glioma has the highest fatality rate of tumours in the central nervous system of adults, and half of cases are glioblastomas [
1]. It is difficult to remove the tumour completely. To make things worse, unclear boundaries with normal tissue easily lead to tumour recurrences [
2]. According to gene expression profiling studies, GBM can be categorized into three transcriptionally defined and clinically related subtypes: proneural (PN), mesenchymal (MES) and classical (CL) [
3]. Despite the fact that standard therapies including surgery, chemotherapy with TMZ and radiotherapy have developed rapidly, the one-year survival rate remains only 40.6%, and the five-year survival rate is only 5.6% [
1]. The resistance to cell death is one of the hallmarks of GBM cells [
4], therefore, it is crucial to clarify the specific molecular mechanisms of this phenomenon to identify a novel therapeutic target for GBM.
p38 mitogen-activated protein kinases are a class of mitogen-activated protein kinases (MAPKs) that are responsive to various stresses and are involved in different cell processes, including cell death [
5,
6]. p38 MAPK signalling plays an important role in GBM; p38 MAPK signalling pathway activation is achieved via a phosphorylation cascade [
7], and phosphorylated p38 can regulate key proteins in autophagy and apoptosis, leading to cell death. As a result, p38 is considered an inducer of apoptosis in GBM [
8]. However, the regulation of p38 MAPK signalling in GBM still remains unclear. Hence, understanding the regulation of p38 MAPK may provide a new insight into the mechanisms of cell death and potential strategies for GBM therapy.
As an atypical member of the Rho GTPase family, RND2 does not have detectable GTPase activity. The most distinctive function of RND2 is its inhibitory effect on Rho kinase-mediated biological functions, including actin cytoskeleton formation and phosphorylation of myosin light chain phosphatase [
9]. Recent studies have also pointed out that RND2 is a key regulator of neuronal movement in the development of the brain and is also essential in regulating the actin cytoskeleton of cells [
10,
11]. Relatedly, as a novel and specific effector of Rnd2 GTPase, Rapostlin induces neurite branching [
12]. However, the pathological role of RND2 in human GBM progression has not been investigated and associated animal studies need to be explored.
Here, we have provided evidence showing that RND2 is an oncogene in human glioblastomas. RND2 expression levels were significantly upregulated in human glioblastomas and suggest a poor prognosis in patients. Additionally, RND2 physically interacted with p38 and decreased p38 phosphorylation, which reduces expression of its downstream substrates, LC3B, Beclin-1, cleaved-caspase3 [
13] and induces p62 [
14,
15]. In conclusion, our findings revealed a new function for RND2 in GBM cell death and provide new insight into the inhibitory effect of RND2 on regulatory mechanisms of p38/MAPK activation.
Methods and materials
The expression profiles of RND2 in different human cancers were downloaded from the TCGA database, while the profiles in normal human tissues were based on information from the HPA database (
https://www.proteinatlas.org). RND2 expression profiles in different glioma subtypes were analysed based on the GlioVis portal (
https://gliovis.bioinfo.cnio.es) [
16].
Human GBM and control brain tissues
Human control brain tissues and GBM tissues were acquired from the Department of Neurosurgery, Renmin Hospital of Wuhan University. GBM tissues were sampled during surgeries and stored at − 80 °C. Control brain tissues were collected from patients during emergency surgeries for traumatic brain injury. The procurement and use of tissue in this study was approved by the Renmin Hospital of Wuhan University’s Institutional Ethics Committee of the Faculty of Medicine (approval number: 2012LKSZ (010) H). The histological diagnosis of glioma was confirmed by the pathologists of the Department of Pathology at the Renmin Hospital of Wuhan University. All tumour samples were subjected to pathological examination and related molecular testing (MGMT, 1p19q, and IDH1/IDH2), and all were defined according to the 2016 WHO classification [
17]. All clinical information for the patients is listed and presented in Supplemental Table S
1.
Cell culture
The human renal epithelial cell line (293 T) and human GBM cell lines (U87 and U251) were from the Cell Bank of the Shanghai Institute of Biochemistry and Cell Biology (Shanghai, China). Information detailing the U251 and U87 cell lines, the generation of the U87 stable cell line, and cell culture methods was described in our previous study [
18]. Cells were cultivated in DMEM along with 1% penicillin/streptomycin and 10% foetal bovine serum, and the incubating temperature was 37 °C, with 5% CO2. The STR Authentication is listed in the supplemental materials.
Reagents and antibodies
Antibodies used in these experiments included the following: anti-RND2 (13844–1-AP, Proteintech, USA), anti-Rho7/Rnd2 (GXT56070, GeneTex, USA), anti-p-p38 (#4511, Cell Signaling Technology, USA), anti-p38 (#9212, Cell Signaling Technology, USA), anti-cleaved-caspase3 (ab32042, Abcam, UK), anti-caspase3 (NB100-56708SS, Novus, USA), anti-BAX (50599–2-Ig, Proteintech, USA), anti-GAPDH (#5174, Cell Signaling Technology), anti-P62 (M162–3, Medical Biological Laboratories, Japan), anti-Beclin1 (11306–1-AP, Proteintech, USA), anti-LC3B (GB11124, Servicebio, China), anti-DYKDDDK/Flag-tag (ANT102, Antgene, China), and anti-His-tag (D291–3, Medical Biological Laboratories, Japan). The autophagy inhibitor wortmannin (3-MA) and the autophagy activator rapamycin (Sirolimus) (S1039, USA) were purchased from Selleck (S2758, USA).
Quantitative real-time PCR (qPCR) and RNA extraction
The extraction of total RNA from tissues and cells was carried out using the Trizol reagent (Invitrogen, USA). For the reverse transcription of RNA, the PrimeScript RT Reagent Kit (RR047A, Takara, Japan) was used to synthesize cDNA. Using SYBR Premix Ex Taq II (RR820A, Takara), we performed qPCR to detect mRNA levels following the specifications provided by the manufacturers. qPCR was performed on a 2.1 Real-Time PCR System using Bio-Rad CFX Manager (Bio-Rad, USA). The relative Ct method was adopted to compare the data of the experimental group and the control group, and GAPDH was set as internal control. The primer sequences are listed in Supplemental Table S
3. The clinical information of the GBM patients who provided samples is listed in Supplemental Table S
2.
shRNA transfection
Specific shRNA targeting of human RND2 (shRND2) and negative control shRNA (shNC) were purchased from RiboBio Corporation (Guangzhou, China). Referring to the specifications, Lipofectamine 3000 transfection reagent (L3000015, Thermo Fisher Scientific) was used in the transfection. The sequences of the different shRND2 constructs are provided in Supplemental Table S
3.
DNA construction and transfection
RND2 cDNA was subcloned with a Flag tag (Flag-RND2) into the pcDNA3.1 vector. Full-length p38 cDNA was subcloned with a 6x His tag (His-p38) into the pcDNA3.1 vector. Transfections were carried out with the help of the transfection reagent Lipofectamine 3000 (L3000015, Thermo Fisher Scientific) in accordance with the specifications.
Flow Cytometric analysis
An Annexin V-PE/7-AAD kit (Becton Dickinson, New Jersey, USA) was used to measure the apoptosis rate of GBM cells. GBM cells were collected and then washed with PBS three times. Then, samples were stained with Annexin V-PE/7-AAD for 15 min in the dark. One hour after staining, the specific apoptosis of GBM cells was analysed using a FACSCalibur flow cytometer (Becton Dickinson). Negative staining for both 7-AAD and Annexin V-PE suggested that the cells were still viable with no apoptosis. Cells in the early stage of apoptosis were positive for Annexin V-PE and negative for 7-AAD. Positive staining for both 7-AAD and Annexin V-PE meant that the cells were in the late stage of apoptosis, or were already dead. To calculate the total apoptosis rate and carry out statistical analysis, the sum of the upper and lower right quadrants was calculated.
Mitochondrial membrane potential (ΔΨm) assay
For early stage apoptosis, the collapse of Δψm function is a hallmark event. Δψm variations were detected by capturing the images of cells after JC-1 staining (Yeasen, Shanghai, China) using an Olympus BX51 microscope (Olympus, Japan) operated followed the manufacturer’s specifications. We also recorded the ratio of aggregated JC-1 (red fluorescence) to monomeric JC-1 (green fluorescence). ImageJ software was used to detect the fluorescence intensity. A drop in the red/green fluorescence intensity ratio indicated the loss of ΔΨm.
TUNEL assay
A feature of apoptotic cells is the fragmentation of DNA, which can be measured using a TUNEL kit. We followed the protocol offered by the manufacturer of the TUNEL kit (Roche Diagnostics, Mannheim, Germany). Images of stained cells were collected using the Olympus BX51 microscope (Olympus, Japan). ImageJ software was used to count TUNEL positive cells.
Western blotting
Cells or tissues were lysed in RIPA buffer with protease and phosphatase inhibitors (cocktails from Roche and PMSF from Beyotime) for 30 min at 4 °C. The protein concentration was detected using a BCA kit (Biosharp, China). Proteins were separated by SDS-PAGE and then transferred onto a PVDF membrane, which was incubated with primary antibodies (including anti-GAPDH, anti-Flag, anti-β-actin, anti-p-p38, anti-Rnd2, anti-p38, anti-His, anti- p62, anti-caspase3, anti-LC3B, anti-Beclin1, anti-BAX, and anti-cleaved-caspase3) overnight and secondary antibodies for an hour. The proteins were delineated with a LI-COR Odyssey Infrared imaging system (LI-COR Bioscience, USA). ImageJ software was used to detect the grey value of the blots. The relative protein quantity was normalized to GAPDH.
Immunoprecipitation assays
Cell lines U87 and U251 were co-transfected with the Flag-Rnd2 (Miaoling Biology, China) and His-p38 (Miaoling Biology, China) plasmids. Cells were lysed in IP buffer containing 1% NP-40, 50 mM NaF, 2 mM Na3VO4, 4 mM Na pyrophosphate and protease inhibitors 48 h after transfection. A total of 3 μg of antibodies (anti-Flag, anti-His or IgG; Beyotime) were added to the cell lysates, and the samples were incubated with 30 μL Protein A/G (Santa Cruz Biotechnology) at 4 °C overnight. The precipitates were washed 5 times or more with IP buffer and were boiled for 5 min in 40 μL 1.5x loading buffer (Beyotime), followed by western blot.
Immunofluorescence assays
Cells were fixed in 4% paraformaldehyde for 15 min, treated with 0.1% Triton-X for 10 min and blocked with 1% BSA for 1 h. The samples were then incubated with primary antibodies (anti-p38, anti-Rnd2, anti-pp38, anti-Lc3B, anti-Cleaved-Caspase3) overnight, followed by Alexa Fluor-labelled secondary antibodies (Antgene, China). The nuclei were stained by DAPI (ANT046, Antgene, China). The Olympus BX51 microscope (Olympus) and a FV1200 confocal microscope (Olympus) were used to take pictures. ImageJ software was used to count positive cells.
Immunohistochemistry
The tissues were embedded in paraffin after being fixed in 4% paraformaldehyde and cut into slices. After being hydrated, the slices were treated with 3% H2O2 for 10 min and blocked with 1% BSA for 1 h. The samples were then incubated with primary antibodies (anti-LC3B, anti-cleaved-Caspase, anti-p-p38, anti-BAX, anti-Rnd2) overnight, followed by HRP-labelled secondary antibodies (Servicebio, China). DAB (Servicebio, China) was used for dyeing and haematoxylin was used to stain the nuclei. Pictures were taken with the Olympus BX51 microscope (Olympus). A semiquantitative score was applied to describe the distribution and intensity of RND2 staining (0 = negative, 1 = weak, 2 = moderate, 3 = strong, and 4 = strong and widely distributed).
Assay of green fluorescent protein-LC3 Puncta
RND2 plasmids were transfected into GBM cells that stably expressed green fluorescent protein (GFP)-LC3. After 2 days, transfected cells were fixed with 4% paraformaldehyde, and then a confocal laser scanning microscope (Olympus, Japan) was used to visualize GFP-LC3 puncta in the cells. The number of GFP-LC3 puncta was calculated from at least 100 cells.
Transmission Electron microscopy (TEM)
Cells transfected with CTRL or RND2 plasmids were fixed with an electron fixation solution containing 2.5% glutaraldehyde. The cells were then post-fixed in 1% osmic acid. Next, a graded series of ethanol was used to dehydrate the specimens. They were then placed in capsules contained embedding medium and heated at 70 °C for approximately 9 h. The specimen sections were stained by uranyl acetate and alkaline lead citrate. Finally, the stained sections were observed using a TEM (Hitachi HT7700, Tokyo, Japan).
Intracranial Xenograft model
PBS was used to suspend U87-MG cells, which stably expressed lentivirus RND2 or CTRL plasmids, at a concentration of 1 × 105 cells/μL, and the cells were then injected into the right striatum of 6 week old Balb/c nude mice by stereotactic implantation; a blank control group that only received PBS was included. For the analysis of survival, the mice were under periodic monitoring and they were sacrificed when serious neurological symptoms appeared and/or an evident loss of weight (more than 20% of their body weight) occurred. We removed and weighed the whole mouse brains. The values (weight of control group/RND2 overexpression group - blank control group) were calculated and statistically analysed. All samples were then fixed in 4% paraformaldehyde. The brains were kept for further analysis and were embedded in paraffin. The Institutional Animal Care and Use Committee of the Renmin Hospital of Wuhan University approved all animal experiments mentioned above.
Statistical analysis
All experiments were done in triplicate and were replicated at least once. All data are expressed as the means ± standard deviations. Statistical analyses were carried out with GraphPad Prism 7 and SPSS version 19.0. Unpaired Student t tests were used in the comparison of means between two groups. p values less than 0.05 were considered to be statistically significant. One-way analysis of variance (ANOVA) was performed to determine the differences between groups. When the analysis showed significance, post hoc testing that targeted the differences between groups was carried out using the Student-Newman-Keuls test. The Pearson correlation coefficient was used to analyse the correlation between RND2 and other genes. *P < 0.05, **P < 0.01, ***P < 0.001 was considered significant statistics.
Discussion
Our results showed for the first time that the RND2/p38/MAPK signalling axis regulates cell death including autophagic activities and apoptosis in GBM. Upregulated RND2 expression in GBM was defined as a negative predictor in patients. Constitutively expressed or induced RND2 decreased the phosphorylation of p38 through physical interaction, thereby inhibiting GBM cell autophagy and apoptosis (Fig.
7h).
GBM is the most frequently seen malignant primary tumour in the central nervous system of adults. Even with ever-accelerating treatments, including radical surgery, radiotherapy and chemotherapy, the overall survival time of patients suffering from GBM only remains at approximately 18 months [
1]. The evasion of and resistance to apoptosis are hallmarks of malignant tumours [
25], which suggests that apoptosis may be a therapeutic strategy for anti-tumour drugs. Additionally, GBM cells lack the intrinsic apoptosis pathway, which leads to chemo-resistance and treatment failure [
26,
27]. Furthermore, autophagy is identified as type II programmed cell death, especially in cells with apoptosis deficiencies [
28]. However, other studies have shown that autophagy can inhibit the development of GBM in the early stage but promote the survival of GBM cells in the late stage. In recent years, clinical studies of autophagy inhibitors in glioblastoma have not yielded the expected results, which indicates that the role autophagy plays in cell death is complicated [
29]. Consequently, it is necessary to explore more specific targets that regulate autophagy to inhibit the development of GBM. Hence, how to induce glioblastoma cell death by autophagy and apoptosis is an urgent problem to solve and is significant for clinical treatment.
RND2 is a member of the RND subfamily, which is a subfamily of the Rho GTPases. The main feature of RND2 is its lack of intrinsic and GAP-stimulated GTPase activities [
30]. In addition, its function does not rely on GDP/GTP exchange but rather on transcriptional, post-translational, and post-transcriptional mechanisms [
31]. The activities of RND2 have been explored not only in normal tissue development but also in disease states [
10]. However, the activities of RND2 in cancer have not yet been demonstrated. Additionally, the mechanistic and direct role of RND2 in GBM tumour genesis is totally unexplored. To date, only a few proteins including Rapostlin, MgcRacGAP and Vps4-A have been identified as RND2 binding partners [
12,
32,
33]. Our study is the first to advance the knowledge of RND2 in GBM and cell death. Our data showed that the level of RND2 expression was dramatically decreased in GBM compared with normal brain tissues. Besides, the expression level of RND2 was higher in PN and CL compared with MES. As we all know, MES subtype is generally considered to hint worse prognosis, however, it’s showed that the favourable outcome of the proneural GBM subtype was because patients were IDH mutant and when those patients are excluded from analysis, the proneural subtype has a worse prognosis than other subtypes [
34]. The PN group expresses genes associated with the process of neurogenesis [
35] and RND2 plays an important role in the development of brain [
10]. This may explain that the expression of RND2 is higher in PN compared with MES. Furthermore, RND2 was negatively correlated with patient prognosis, while it was positively correlated with tumour size, suggesting that RND2 is a potential target for treating GBM.
Our study firstly identified and validated that p38 was the substrate for RND2. Besides, we found that RND2 decreased the phosphorylation of p38 by directly binding to p38 in the cytoplasm of cancer cells and we observed that RND2-mediated p38 MAPK signalling was critical for autophagy activities. The activity of p38 MAPK was related to the enhanced expression of autophagic markers (such as ATG5/ATG12 and LC3B) and apoptotic markers (such as PARP and caspase-3) [
36,
37]; thus, p38 MAPK is quite critical in regulating the balance between survival and cell death [
5]. As we know, there are many downstream substrates that could be regulated by p38 mitogen-activated protein kinase, which regulates various cellular processes through a cascade of complex phosphorylation [
38]. p38 signalling suppresses tumourigenesis and promotes apoptosis in various types of cancers. Consequently, it is important to understand the mechanisms of how specific substrates are recognized and regulated by p38. There may be some factors, such as availability, concentration, and subcellular localization of upstream proteins [
39], that influence substrate selection. It has been found that p38 is expressed in both the cytoplasm and the nucleus (but mainly in the cytoplasm), which demonstrates that the distribution of p38 might be critical for regulating substrates. Hence, the potential binding partners of p38 like MK2, MKK3, TAB1 may regulate p38 in different ways such as through phosphorylation sites or subcellular localization [
40‐
42]. It is possible RND2 associates with these proteins complex and mediates p38 nucleus-cytoplasm transportation directly or indirectly.
Our study further found that RND2/p38/MAPK signalling axis downregulates cell death, owing to the fact that the inhibition of autophagy coincided with the inhibition of apoptosis- and autophagy-mediated cell death, with the inhibition of autophagy further attenuating apoptosis. However, p38 played a dual role in autophagy. For example, the phosphorylation of p38 could promote the expression of the key autophagy protein Beclin-1, leading to cell death [
14]. Inversely, p38 MAPK inhibited autophagy by phosphorylating ULK1 [
24].
Our data provided significant evidence that RND2 inhibited autophagy through the de-phosphorylation of p38 specifically. Autophagy has significant functions under different pathological conditions but crosstalk with apoptosis is still controversial [
43]. It has been pointed out that autophagy occurs before apoptosis and that it is a necessary condition. Autophagy suppresses the development of tumours by inhibiting the expression of oncogenes and promoting pro-apoptotic factors to induce the cell death [
44]. However, chemo-resistance is one of the most common problems during anti-cancer therapy and autophagy is closely related to this process through its involvement in the avoidance of chemotherapy-induced apoptosis in different cancers. Moreover, autophagy provides nutrients for cells in a hypoxic or starvation environment to protect tumour cells from apoptosis in GBM [
45]. In addition, autophagy and apoptosis were found to be antagonistic or synergistic under certain conditions. Apoptosis and autophagy could occur simultaneously to trigger cell death, while apoptosis could also accelerate the transformation of cells to autophagic cell death [
46,
47]. Above all, our study revealed that the RND2/p38/MAPK signalling axis downregulated autophagy and apoptosis at the same time and that apoptosis could be influenced by RND2-mediated autophagy to improve the survival of GBM cells. Our results enrich our knowledge of the mechanism by which autophagy is precisely regulated in glioblastomas, which may provide a potential solution for TMZ chemo-resistance to TMZ in GBM cells which have emerged as a challenging problem in clinical practice [
48]. Lack of programmed cell death is an important reason which caused chemo-resistance. Recently, it has been demonstrated that autophagy plays a prodeath role in GBM cells treated with chemotherapeutic agents, by enhancing autophagy-mediated apoptosis [
49]. Similarly, our results show that RND2 promote the survival of cells by reducing autophagy and autophagy-mediated apoptosis in GBM. In a summary, our data not only suggests RND2 as an alternative therapeutic target for malignant human cancers such as GBM but also provides a solid foundation for the development of a compound targeting RND2, which could be transformed into clinical applications and combined with chemotherapies.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.