MLK3 Is Associated With Poor Prognosis in Patients With Glioblastomas and Actin Cytoskeleton Remodeling in Glioblastoma Cells

Mixed lineage kinase 3 (MLK3) has been implicated in human melanoma and breast cancers. However, the clinical significance of MLK3 in human gliomas and the underlying cellular and molecular mechanisms remain unclear. We found that MLK3 proteins were highly expressed in high-grade human glioma specimens and especially prevalent in primary and recurrent glioblastoma multiforme (GBM). High levels of MLK3 mRNA were correlated with poor prognosis in patients with isocitrate dehydrogenase (IDH)-wild-type (wt) gliomas. Furthermore, genetic ablation of MLK3 significantly suppressed the migration and invasion abilities of GBM cells and disrupted actin cytoskeleton organization. Importantly, MLK3 directly bound to epidermal growth factor receptor kinase substrate 8 (EPS8) and regulated the cellular location of EPS8, which is essential for actin cytoskeleton rearrangement. Overall, these findings provide evidence that MLK3 upregulation predicts progression and poor prognosis in human IDH-wt gliomas and suggest that MLK3 promotes the migration and invasion of GBM cells by remodeling the actin cytoskeleton via MLK3-EPS8 signaling.


INTRODUCTION
Gliomas, one of the most prevalent forms of primary brain tumors, are classified into four grades (grade I-IV) in line with the World Health Organization (WHO) 2016 classification criteria (1). Glioblastoma multiforme (GBM) is a malignant grade IV tumor with poor prognosis. According to the statistics, the 3-year overall survival of patients with GBM is approximately 12% (2). Tumor invasion and immune evasion account for the main causes of recurrence and death in patients with GBM. Therefore, understanding the underlying cellular and molecular events may provide promising molecular markers for GBM diagnosis, prognosis, and targeted therapy.
Mixed lineage kinase 3 (MLK3), also known as mitogenactivated protein kinase kinase kinase 11 (MAP3K11) and Srchomology 3 (SH3) domain-containing proline-rich kinase (SPRK), is a member of the serine/threonine protein kinase family encoded by the MAP3K11 gene in humans (3). MLK3 is involved in human melanoma and breast cancers (4)(5)(6). High levels of MLK3 mRNA are found in metastatic primary malignant melanoma tissues (4). Breast tumors express a comparable level of MLK3 proteins, whereas MLK3 kinase activity is profoundly reduced and inversely correlated with tumor grades in human epidermal growth factor receptor (EGFR) 2-positive breast cancer tissues (5,6). In diverse human cancer cell lines, MLK3 is involved in multiple cellular processes, including proliferation, proapoptosis, migration, and invasion (4)(5)(6)(7)(8)(9)(10)(11). MLK3-JNK signaling has been reported to be related to EGFR activation-driven migration and invasion of GBM cell line (10). However, the pathophysiological function of MLK3 in the progression and prognosis of human gliomas remains unknown, and how MLK3 promotes the development of gliomas has not been well understood.
Cancer cell migration and invasion involves integrated complexes and is a dynamic process that requires actin cytoskeletal rearrangement to change the cell shape and generate the driving force for cell movement. A group of regulatory molecules are involved in cytoskeletal remodeling, including EGFR kinase substrate 8 (EPS8) (12)(13)(14)(15). EPS8 is responsible for actin cytoskeleton formation and facilitates the migratory and invasive capacities of GBM cells (16). Furthermore, recent findings implicate that actin cytoskeleton remodeling drives cancer cell resistance to antitumor immunity (17). Altogether, elucidating the roles of MLK3 in actin cytoskeleton regulation is essential for understanding glioma progression and invasion.
In this study, we examined the expression of MLK3 in human glioma tissue specimens. Additionally, we determined the correlation between MLK3 protein and mRNA levels and glioma progression and poorer prognosis in patients with GBM. Furthermore, we investigated whether and how MLK3 is involved in GBM cell migration, invasion, and actin cytoskeletal remodeling. Our data provide evidence that MLK3 is a valuable biomarker for predicting the prognosis and towards targeted therapy of GBM.

Human Tissue Analysis
Glioma tissues (WHO grade I, n = 6; grade II, n = 23; grade III, n = 20; and grade IV, n = 48) were obtained from the Department of Pathology of the Affiliated Hospital of Xuzhou Medical University between 2016 and 2017. All samples were identified by pathologists according to the 2016 WHO classification criteria.

Immunohistochemical Staining
Paraffin-embedded glioma tissue sections were deparaffinized and hydrated using xylene and graded alcohols. Antigen retrieval was performed with high pressure for 3 min. The tissues were blocked with 3% bovine serum albumin for 20 min at room temperature. Primary antibodies were incubated overnight at 4°C. Then, biotinylated secondary antibodies were incubated for 30 min at room temperature. The staining was carried out using an ABC reagent kit (VECTASTAIN) and a 3,3-diaminobenzidine peroxidase substrate reagent kit (Vector, #SK-4100). The nuclei were treated with hematoxylin staining, and the sections were mounted on glass slides. Images were acquired with Nikon microscopy. The evaluation of MLK3 staining was as previously described (18). According to staining intensity and area, MLK3 staining was categorized into scores 0-12. The median level of MLK3 (score 6) as the cut-off point, the samples were divided into low/high expression of MLK3 groups.

Cell Culture and Transfection
The human GBM cell lines U87, U118, U251, U343, and T98G were maintained in Dulbecco's modification of Eagle's medium (DMEM, Gibco, #12000-014) supplemented with 10% fetal bovine serum (FBS). Cells were grown at 37°C and 5% CO 2 . All cell lines were authenticated through short tandem repeat DNA fingerprinting from Cell Bank/Stem Cell Bank, The Committee of Type Culture Collection of Chinese Academy of Sciences (Shanghai, China) in August 2018. Plasmid transfections were carried out with Lipofectamine 3000 (Invitrogen, #L3000-015).

Transwell Migration and Invasion Assays
Cells were treated with serum-free DMEM for 14 h. Then, cells (5×10 4 ) were suspended in serum-free DMEM and added to the upper chamber of a 24-well transwell plate (Corning, #3422). For the invasion assay, the upper chambers were precoated with a Matrigel Basement Membrane Matrix (BD Biosciences, #356234). DMEM supplemented with 10% FBS was added to the lower chambers. The cells were cultured at 37°C for 10 h2 4 h. The chambers were washed with phosphate-buffered saline (PBS), fixed with 4% paraformaldehyde for 20 min and washed with PBS. The cells on the upper surface of the membrane were removed with a cotton swab. The cells on the bottom surface were stained with Giemsa staining.

Immunofluorescence Analysis
Cells were plated on glass slides, washed with PBS, fixed with 4% paraformaldehyde for 10 min at 4°C, permeabilized with Triton X-100 (0.2%) for 15 min at room temperature, and blocked with 10% normal goat serum for 2 h at room temperature. Primary antibodies were incubated overnight at 4°C. The fluorescent secondary antibodies were incubated for 2 h at room temperature. The nuclei were stained with 4-6diamidino-2-phenylindole (DAPI) (Sigma, #D8417). Coverslips were mounted on the glass slides, and images were taken with confocal microscopy.

Immunoblot Analysis
Cell total protein was separated by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes. After blocking with 3% bovine serum albumin for 2 h, the membranes were incubated with the primary antibody overnight, followed by horseradish peroxidaseconjugated secondary antibody for 1 h at 4°C. The specific proteins were detected by enhanced chemiluminescence reagent. The band intensity was quantified with ImageJ software.

GST Pull-Down Assay
GST-fused MLK3 (1-104 aa) and MLK3 (632-847 aa) proteins were expressed in BL21 cells and purified with the Pierce ™ GST Protein Interaction Pull-Down Kit (Thermo, #21516) according to the manufacturer's instructions. The purified proteins were incubated with the protein lysates from pcDNA3.1-EPS8-transfected HEK293 cells for 2 h at 4°C. The beads were washed and boiled in loading buffer for the immunoblot assay.

Statistical Analysis
All data are presented as the mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism 7.0 software. Pearson correlation analysis and chi-square (X 2 ) tests were conducted to detect correlations. The Kaplan-Meier method was applied to assess overall survival. For parametric data, two-tailed Student's t-tests and one-way ANOVA were used to examine differences. For non-parametric data, a twosided Mann-Whitney test was used. P < 0.05 was considered statistically significant.

Mixed Lineage Kinase 3 Is Highly Expressed in High-Grade Human Glioma Specimens
First, 97 clinical glioma tissue specimens were analyzed with immunohistochemistry (IHC) to determine the protein expression and subcellular distribution of MLK3. As shown in Figures 1A, B, the expression levels of MLK3 protein were higher in glioma tissues than in para-tumor tissues. Clearly, MLK3 proteins were mainly localized in the cytoplasm of glioma cells. Next, we assessed the correlation between MLK3 levels and clinicopathological characteristics (tumor size and WHO grade). IHC analysis showed that the levels of MLK3 protein were closely correlated with age and tumor grade ( Table 1). MLK3 staining was weaker in low-grade glioma tissues (grade I and II) than in high-grade glioma tissues (grade III and IV) ( Figures 1A, C). In addition, there was significant variation of MLK3 levels among patients within the same grade gliomas ( Figure 1D). The percentage of high MLK3 level cases increased with the grade of glioma.
Next, we explored MLK3 mRNA levels in glioma tissues by using RNA sequencing (RNA-seq) data from the Chinese Glioma Genome Atlas (CGGA) database and the Freije dataset from the Oncomine database. The results showed that the levels of MLK3 mRNA were significantly increased in high-grade gliomas (grade IV) and closely associated with the progression of gliomas (Figures 1E, F). MLK3 mRNA levels showed significant variation within the same grade ( Figure 1G), which is consistent with the results of IHC analysis.
These findings suggest that MLK3 is significantly upregulated in high-grade human glioma tissues and positively associated with a malignant phenotype. The detection of several molecular markers, including IDH1 mutation, 1p/19q codeletion, MGMT promoter methylation status, and EGFR amplification has been applied with clinical diagnoses of gliomas (1). We analyzed the correlation between MLK3 levels and the status of these biomarkers by using CGGA data and IHC analysis. The results revealed that the expression of MLK3 mRNA in human gliomas was correlated with IDH status but not 1p/19q codeletion and MGMT promoter methylation status. While it was negatively related to EGFR mRNA levels ( Table 2). We also analyzed the correlation between MLK3 levels and downstream regulators. The results showed that MLK3 mRNA in human gliomas was negatively related to MAPK8, MAPK9, and MAPK10 ( Table 2). Additionally, MLK3 was highly expressed in IDH-wt gliomas and especially prevalent in IDH-wt GBM (59/95) (Figure 2A).
In IDH-wt GBM, high levels of MLK3 frequently occurred in primary GBM (43/71), recurrent GBM (9/15) and secondary GBM (7/9) ( Figure 2B). The results of the IHC analysis confirmed that the MLK3 protein was abundantly expressed in IDH-wt GBM (24/33) ( Figure 2C) and are in accordance with the results of the bioinformatics analysis. We further studied the correlation between MLK3 expression and the overall survival of patients with gliomas by bioinformatics analysis. The Kaplan-Meier analysis showed that high levels of MLK3 were associated with the poor prognosis of patients ( Figure  2D). Glioma patients with the IDH-mutant (mut) phenotype had a better prognosis than patients with the IDH-wt phenotype. Remarkably, in IDH-wt gliomas, patients with comparatively lower levels of MLK3 had a higher overall survival rate than patients with high levels of MLK3 ( Figure 2E). By contrast, there was no difference in the overall survival of IDH-mut patients with either high or low levels of MLK3 ( Figure 2E).
These data suggest that MLK3 upregulation predicts poorer prognosis in IDH-wt gliomas.

Mixed Lineage Kinase 3 Promotes Glioblastoma Multiforme Cell Migration and Invasion and Is Required for Actin Cytoskeleton Rearrangement
To assess the roles of MLK3 in the biological behaviors of glioma cells, we first detected the protein levels of MLK3 in high-migration glioma cell lines (U87, U118, and U251) and low-migration glioma cell lines (U343 and T98G) 48 h after serum starvation ( Figure 3A).   Figure S1, Figures 3B, C). The loss of MLK3 robustly restrained cell migration in U251 and U118 cells compared to wt cells ( Figures 3D, E). The Matrigel assay showed that MLK3 ablation evidently suppressed U251 and U118 cell invasion (Figures 3F, G). The above results suggest that MLK3 is required for the migration and invasion of GBM cells in vitro. Next, the MLK3 kinase inhibitors CEP-701 was used to test the roles of MLK3 activity in the regulation of GBM cell migration and invasion. Immunoblot analysis showed that CEP-701 effectively inhibited the activity of MLK3 (Supplementary Figure S2A). A wound healing assay showed that CEP-701 (at the appropriate concentration) significantly inhibited the wound closure of U87, U251, and T98G cells in comparison to vehicle-treated cells Figures S2B-D). Transwell assays showed that CEP-701 (400 nM) markedly suppressed the migration of U251 and U87 cells (Supplementary Figures S2E, F). The Matrigel assay revealed that the invasion ability of U251 and U87 cells was remarkably reduced after treatment with CEP-701 (400 nM) (Supplementary Figures S2G, H).

(Supplementary
Actin cytoskeleton remodeling is a vital event in cell migration and invasion. To clarify the role of MLK3 in actin cytoskeleton organization, F-actin and vinculin (for focal adhesion) staining were carried out to identify changes in the actin cytoskeleton. The results showed that MLK3 ablation in U251 and U118 cells resulted in marked changes of the actin cytoskeleton, and actin based "mass-like structures" appeared on the edge of cells (Figures 3H, I). MLK3 ablation increased stress fibers in U251 cells ( Figure 3H), while the distribution of vinculin obviously decreased on the edge of U251 ko and U118 ko cells (Figures 3H, I). More interestingly, the altered cell morphology and more filamentous protuberances were observed by ectopic expression of MLK3 in T98G cells (Supplementary Log-rank test. NS, not significant. *P < 0.05, **P < 0.01. Figure S3). These findings indicate that MLK3 contributes to actin cytoskeleton remodeling.

Mixed Lineage Kinase 3 Directly Binds to Epidermal Growth Factor Receptor Kinase Substrate 8
Previous studies have indicated that EPS8 controls actin cytoskeleton reorganization (20,21). To elucidate the molecular mechanism underlying MLK3-mediated cell migration and invasion, we identified the molecular interactions of MLK3 and EPS8. Co-IP analysis showed that overexpressed ( Figure 4A) and endogenous MLK3 ( Figure 4B) interacted with EPS8 in U251 cells. Furthermore, in vitro GST pull-down analysis showed that the c-terminal domain of MLK3 (1-104 aa) and the proline-rich domain (PRD) of MLK3 (632-847 aa) directly bound to EPS8 ( Figure 4D). In addition, MLK3 cooperates with EPS8 and affects the overall survival of patients with gliomas. Bioinformatics analysis showed that high levels of MLK3 and EPS8 in gliomas are correlated with a significantly worse overall survival in patients ( Figure 4E). These results suggest that MLK3-EPS8 signaling may be involved in the development of gliomas.

Mixed Lineage Kinase 3 Regulates the Localization of Epidermal Growth Factor Receptor Kinase Substrate 8
To further explore the mechanisms of MLK3-mediated actin cytoskeleton remodeling, we examined the localization and expression of EPS8 in MLK3-depleted U118 and U251 cells. Immunoblot assays showed that EPS8 was upregulated in MLK3-depleted U118 and U251 cells (Figures 5A, B). Additionally, the immunofluorescence assay results showed that the subcellular localization of EPS8 was disrupted in MLK3-ablated cells, and mass-like structures were assembled on the cell edges ( Figures 5C, D). These data suggest that MLK3 is critical for actin cytoskeleton rearrangement by regulating EPS8 localization in GBM cells.

DISCUSSION
GBMs are devastating malignancies with highly migratory and invasive behavior. Despite advancements in clinical diagnosis and treatment, recurrence remains universal, and available salvage strategies are rare. Therefore, identifying special molecular markers for the early diagnosis, prognosis evaluation, and targeted therapy of GBM is of great significance. Our findings provide the first evidence that MLK3 upregulation plays an essential role in tumor progression and correlates with an unfavorable overall survival in patients with IDH-wt GBM. MLK3 may serve as a valuable diagnostic and prognostic marker and may be a promising therapeutic target for GBM therapy. To date, patients with IDH-wt gliomas exhibit the poorest outcomes, and few targeted agents are therapeutically effective for IDH-wt gliomas. The present study provides evidence that MLK3 might be required for the evaluation of prognosis and targeted therapy of IDH-wt gliomas. Previous studies indicated that the IDH and IDH2 mutations are more frequent in grade II-III gliomas and secondary GBMs (22)(23)(24). Approximately 80% of secondary GBMs have somatic mutations in IDH, which are absent in primary GBM (25). Patients with IDH mutations have a better prognosis, which is consistent with our findings. An increase in R (-)-2-hydroxyglutarate in gliomas harboring IDH mutations has been found to prevent the association of MLK3 with Cdc42 and further inhibit the activity of MLK3, which may be one of the reasons for the better prognosis in patients with IDH-mut gliomas (26,27). More notably, our findings showed that IDH-mut gliomas present with low levels of MLK3, suggesting that mutant IDH not only blocks MLK3 activity but also reduces MLK3 expression. Although there is no correlation between the MLK3 levels and overall survival of patients with IDH-mut gliomas, high levels of MLK3 are positively correlated with poor prognosis in IDH-wt gliomas. Therefore, the prognosis of patients with IDH-wt gliomas can be reasonably evaluated by MLK3. In addition, this study demonstrated that MLK3 cooperates with EPS8, which might aid in prognosis prediction for patients with gliomas.
MLK3 appears to function as an oncogene to promote cell migration and invasion through the MLK3/JNK pathway in human breast and gastric cancer cells (7,8,10), the MLK3/ FRA-1/MMP-1 axis in human triple-negative breast cancer cells (28), or MLK3/ERK signaling in ovarian and colorectal cancer cells (9,11). Here, our data showed that MLK3 is responsible for the migration and invasion abilities of GBM cells via MLK3/EPS8 signaling. MLK3 has been reported to contribute to EGF-induced GBM cell migration and invasion (10). Here, we found that MLK3 mRNA levels in human gliomas was negatively related to EGFR and MAPK8, MAPK9, MAPK10 mRNAs levels, and MLK3 downregulation reduces GBM cell migration and invasion without EGF induction, suggesting that the MLK3 regulation of cell migration and invasion is EGF signaling-independent. Cell migration is a multipurpose process (29,30) and drives the progression of cancer (31,32). The alteration of the actin cytoskeleton is critical in the process of cell migration and generates the forces that push cell migration. MLK3 silencing enhances stress fibers in breast cancer cells (33). In this study, we confirmed that genetic depletion of MLK3 disturbs actin cytoskeleton organization. However, the molecular mechanisms underlying actin cytoskeleton remodeling remain unclear. Previous studies showed that EPS8 directly binds to actin and blocks actin prolongation by its effector domain (648-821 aa) via its capping activity to activate Rac signaling (34,35). Therefore, the accurate localization of EPS8 is essential and contributes to actin rearrangement. In this study, we found that MLK3 interacts with EPS8 and controls the localization of EPS8. Following the loss of MLK3 in U118 and U251 cells, EPS8 expression was apparently increased, and its location was disarranged, which resulted in a number of masslike structures on the cell edge. We verified that the regions of MLK3 (1-104 aa) with an SH3 domain and MLK3 (632-847 aa) containing PRD directly bind with EPS8. Targeting the two domains of MLK3 could be used as a valuable strategy for developing drugs for glioma therapy. Taken together, our results confirmed a novel mechanism by which MLK3 promotes cell migration and invasion via MLK3/EPS8 signaling.
The actin cytoskeleton is a dynamic cell structure that is attributed to diverse cellular processes, including cell morphogenesis, membrane trafficking, cell division, and immune response (17,(36)(37)(38). In addition, actin cytoskeleton dynamic remodeling has emerged as a critical event for glioma invasion. Therefore, controlling the change in actin is a feasible approach for glioma therapy. Currently, immunological therapies are crucial strategies for glioma therapy. Thus, tumor cells that escape from immunological surveillance are a barrier to effective immunotherapies. A study confirmed that the alteration of actin dynamics in tumor cells facilitates immune evasion (17). It would be interesting to ascertain whether MLK3 is involved in immune evasion of gliomas by regulating actin dynamics.
In summary, our findings have illustrated that as an oncogene, MLK3 may be a crucial regulator of the progression of gliomas and is associated with poor prognosis. As a consequence, the deletion of MLK3 in GBM cells decrease the capacity for cell migration and invasion and disrupt actin cytoskeleton remodeling. Furthermore, we provided a novel mechanism by which MLK3 facilitates glioma cell migration by regulating actin skeleton remodeling via MLK3/EPS8 signaling. Therefore, MLK3 is a potential target for therapy.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
In this study, human glioma tissue samples were approved for use by the Ethics Committee of the Affiliated Hospital of Xuzhou Medical University (No. XYFY2018-KL056-01). All samples were obtained and analyzed previously by pathologists according to the 2016 WHO classification criteria. Written informed consent has been provided by the participants' legal guardian/next of kin prior to the surgery. For re-use of these samples in this study, written informed consent was not required in accordance with the national legislation and the institutional requirements.

AUTHOR CONTRIBUTIONS
YZ and X-YH designed the study. J-MS and Y-PW designed and performed the staining and scoring protocols for the human glioma tissues. YZ and Z-CS performed the cellular and molecular experiments and analyzed the data. F-JC assisted with the gene knockout by CRISPR/Cas9, and YZ and X-YH wrote the manuscript. All authors contributed to the article and approved the submitted version.