ANTXR1 Facilitates Tumor Growth in Glioma via Deactivating MAPK Signaling Pathway

Background: Gliomas are commonly known as primary brain tumors and associated with frequent recurrence and an unsatisfactory prognosis in despite of extensive research in the underlying molecular mechanisms. In this study, we aimed to examine the role of ANTXR1 in glioma tumorigenesis and explore its downstream regulatory mechanism. Methods: We detected overexpression of ANTXR1 in glioma cell lines (SHG-44 and U251) in comparison with those in normal brain tissues. Result: Glioma cell growth and migratory ability were dramatically impaired as a result of silencing ANTXR1 by shANTXR1 lentivirus. ANTXR1 blockade also accelerated cell apoptosis since ANTXR1 held back apoptosis via targeting G2 phrase during cell mitosis. In vivo xenograft models veried in vitro ndings above that the solid tumors stripped from mice were much lighter and smaller after depletion of ANTXR1 than controls. We also mechanically probed the downstream pathways and disclosed that overexpression of ANTXR1 abrogated the levels of MAKP9 and apoptisis-related protein HTRA2, but augmenting CCND1 and CDK6 levels in glioma cells. Our ndings allow us to demonstrate that ANTXR1 acts as a tumor promoter in glioma induction through attenuating MAPK9-mediated gene transcription and HTRA2-induced apoptisis but intensifying CCND1-mediated proliferation. Conclusion: Together, we declare that ANTXR1 plays an indispensable role in glioma tumorigenesis via deactivating MAPK signaling and apoptosis pathway but activating PI3K/AKT-mediated cell growth. Our study provide a valuable clue to targeting ANTXR1 as a molecular target and a promising anticancer agent in glioma clinical therapeutics.


Background
Brain tumors are associated with signi cant morbidity and mortality due to aggressiveness of local growth and metastasis outside the central nervous system [1]. Glioma was estimated to comprise 28% of brain and other central nervous system tumors, and 80% of malignant brain tumors [2], which has been considered as a primary malignant brain tumor. It is thought to originate from neuroglial stem cells and categorized as astrocytic, oligodendroglial or ependymal tumors based on the histopathologic features [3]. Typically, glioma is characterized by high degree of in ltrative growth which contributes to frequent recurrence [4]. Despite the role of molecular markers in determination of low-grade gliomas has been generally con rmed, i.e., IDH1and IDH2mutations, it is not necessarily true in the high-grade glioma and the speci c molecular markers still remain further investigations [5,6]. To date, the therapeutic approach of glioma is usually surgical resection in conjunction with chemotherapy, radiotherapy, immunotherapy or gene therapy. However, the combined treatments were not always effective and glioma patients in many cases suffered from high risk of recurrence and a very poor prognosis [7]. Therefore, it still matters to seek for advances in therapeutic strategies of gliomas based on its molecular pathogenesis.
Page 3/21 ANTXR1 (Anthrax toxin receptor 1), also known as Tumor endothelial marker 8 (TEM8), is encoded on chromosome 2 in human as a type I transmembrane protein and is a highly-conserved cellular receptor for anthrax toxin secreted by Bacillus anthracis [8,9]. The tripartite anthrax toxin is composed of protective antigen (PA), lethal factor and edema factor [10], among which the PA component binds to ANTXR1 and then the complex mediates delivery of toxin to cytoplasm. Through binding collagen types I and VI, ANTXR1 also acts a pivotal role in endothelial cell attachment and migration [11,12], and servers as a promoter in the process of pathological angiogenesis. ANTXR1 was originally found upregulated in the endothelial cells of colorectal cancer during tumor-related vasculature [13], and subsequently was identi ed overexpressed in various cancers, such as gallbladder carcinomas, breast cancer [14] and prostate cancer [15], in comparison with that in normal tissues [16]. A previous study suggested that the mutation in ANTXR1 gene would lead to disorganized angiogenesis in response to the inhibitory effect on downstream signaling pathways in infantile hemangioma [17]. Besides, downregulation of ANTXR1 slowed down tumor growth in a few cancer models, including melanoma, breast cancer, colorectal cancer and lung cancer [18]. However, little is known about the role of ANTXR1 in the development of glioma.
In this study, we thus exploited a series of in vitro and in vivo experiments to disclose and identify biological functions of ANTXR1 in glioma progression, through which we aimed to evaluate the possibility of ANTXR1 to serve as a novel molecular therapy target. Our ndings indicate the ANTXR1mediated apoptosis via deactivating MAPK signaling pathway and underscore the vital role of ANTXR1 in glioma progression, as well as its further potential application as a targeted molecular agent.

Clinical samples and immunohistochemical staining
Tissue microarrays of glioma samples from 134 patients (15-80 years old, glioma histopathological grade , and ) and normal brain samples from 24 people were from Jiangxi Provincial People's Hospital A liated to Nanchang University. The usage of all samples were approved with permission from all patients and followed the International Ethical Guidelines for Biomedical Research Involving Human Subjects issued by the Council for International Organization of Medical Sciences (CIOMS). The histopathological grading of the glioma tissue microarrays here was based on the 2016 World Health Organization classi cation of tumors of the central nervous system [19]. In terms of immunohistochemical staining, we rstly baked the slides of glioma tissues and matched normal brain tissues in oven at 65 °C for 30 min followed by dewaxing and rinsing them for several times. After retrieving the antigen by disposing with citric acid antigen at 100°C for 10 min, we blocked the slides using 3% H 2 O 2 for 5 min. ANTXR1 antibody (1:250, Cat. #ab21270, Abcam) was next added on the slides at 37°C for 1 h and then the second antibody. The staining was nished after treatment of DAB and hematoxylin. Then we photographed the slides and evaluated the staining percentage and the intensity, according to which we assigned values to each slide and determined the outcomes of immunohistochemistry. Speci cally, the assignment score of a slide was the integer from 1 to 4 when the staining percentage was less than 25%, 50%, 75% and 100%, respectively. Similarly, the staining intensity was also quanti ed and scored from 0 to 3 based on the color depth with 0 signifying no dyeing and 1-3 light yellow, pale brown and dark brown.

Western blotting
Glioma cells were prepared and lysed by 1 × cold lysis buffer. Total proteins from cells were then extracted applying BCA protein detection kit (Cat. #23225, HyClone-Pierce). After separated by 10% SDS-PAGE, protein extraction was shifted to a polyvinylidene di uoride membrane and incubated using blocking liquid for 1 h. We next probed the blots using primary antibodies of ANTXR1 (1:1000, Cat. (1:3000, Cat. #A0208, Beyotime) as secondary antibody. The signal of membranes was nally detected utilizing immobilon Western Chemiluminescent HRP Substrote (Cat. #RPN2232, Millipore).

MTT assay
Cell proliferation of SHG-44 and U251 cell lines was detected via MTT assay. To put it simply, cell suspension (2000 cell/well) was prepared after treatment of trypsin and then cultured in 96-well plates. Cell plates were treated and stained by MTT solution (20 μL 5 mg/mL) 4 h before culture termination. At the end of cultivation, we discarded culture medium, disposed with 100 μL DMSO and shook for 5 min. For the next 5 consecutive days, the absorption values of cells were obtained at OD490 nm with a Microplate Reader and thus viability of lentivirus-transfected cells was nally acquired.

Flow cytometry
We employed ow cytometry to examined cell apoptosis and cell cycle of both glioma cell lines. At the very beginning, cells were trypsinized and rinsed in cold PBS (pH=7.2~7.4). As to the detection of apoptosis, cells were then re-rinsed with 1 × binding buffer and stained by Annexin V-APC for 15 min from light. The percentage of cell apoptotic rate was thus determined by FACSCalibur (BD Biosciences). In terms of cell cycle detection, we xed cells suspension with cold ethanol (70%) at 4°C for at least 1 h and then removed ethanol. Following centrifugation and re-rinse of PBS, cells were stained using PI solution (40×, 2 mg/mL: 100×RNase, 10 mg/mL: 1×PBS =25:10:1000) for 30 min. Lastly, FACSCalibur (BD Biosciences) was employed to detect cell cycle distribution (G1, S and G2).

Wound-healing assay
Glioma SHG-44 and U251 cells were seeded into 6-well dishes. After cell growing for 72 h, we scratched a wound across the cell layer using a 96-wounding replicator (Cat. #VP408FH, VP scienti c) and then rinsed slightly for 2-3 times. SHG-44 cells were cultured for 8 h and 24 h while U251 cells for 24 h and 48 h, following which we took uorescence micrographs and accordingly calculated the cell migration rate.

Apoptosis pathway assay
Apoptosis-related proteins and signaling pathways involved in this research was determined utilizing Human apoptosis antibody array (Cat. #ab134001, Abcam). Simply, after U251 cell samples were washed and lysed, we measured the concentration of the whole proteins via BCA protein assay kit (HyClone-Pierce). The proteins were thus incubated with blocked antibody array membrane throughout the night at 4°C, followed by continue incubation with Streptavidin-HRP for 1 h. Subsequently, we employed enhanced chemiluminescence (ECL) (Amersham) for visualizing proteins and got the gray results by ImageJ software.

In vivo tumorigenesis
Mice used for constructing xenograft models were 28-day-old and BALB/c nude females (Beijing Vitalriver Experimental Animal Technology Co., Ltd), which were randomly separated (shANTXR1 vs. shCtrl groups) and subcutaneously injected with 0.2 mL (2 × 10 7 cells/mL) lentivirus-transfected SHG-44 cell suspensions. The body weight, as well as tumor length and width (for tumor volume) of each mouse model was rst measured after 35 days of the construction of xenograft models, since which the measurement was executed every seven days for another four weeks. On the day of the last measurement, we disposed the experimental mice using 0.7% sodium pentobarbital (10 μL/g, SIGMA) and determined the uorescence expression in vivo through a Perkin Elmer IVIS Spectrum (Waltham). In the end, we executed all mice and removed the solid tumors for Ki-67 staining (primary antibody: Ki-67, 1:200, Cat. #ab16667; secondary antibody: HRP, 1:400, Cat. #ab6721, Abcam). Animal experiments in this research were carried out on a protocol approved by Ethics committee of Jiangxi Provincial People's Hospital A liated to Nanchang University.

Statistical analysis
We performed cell experiments in triplicate and expressed the data as mean ± SD. Differences between two groups were examined by t-test and Mann-Whitney U test, with sign test used for the difference of ANTXR1 expression. Spearman correlation test was exploited to estimate the relationship between ANTXR1 level and clinical characteristics, while Kaplan-Meier survival analysis was detecting how high/low level of ANTXR1 affected overall survival of glioma patients. We run all analysis and plotting in SPSS 20.0 and GraphPad Prism software 7.0 with P < 0.05 as statistical signi cant.

Overexpression of ANTXR1 gene in glioma tissues
The expression patterns of ANTXR1 in glioma tissues and normal brain tissues were explored through immunohistochemistry analysis. Sign test was subsequently employed to detect the difference between glioma tissues and normal brain tissues. The results showed that expression of ANTXR1 was upregulated in glioma tissues in comparison with that in normal tissues (P < 0.001) ( Figure 1A, Table 1). In regard to clinicopathologic characteristics of the patients with glioma, we also found a signi cantly positive correlation between the malignant grade of glioma and the expression level of ANTXR1 (P < 0.01) (Tables 2-3). ANTXR1 expression was also more likely to upregulate in the recurrent glioma tissues than in the non-recurrent ones (P < 0.001) (Tables 2-3). However, we failed to determine an age-or sexspeci c difference in the expression levels of ANTXR1 in patients with glioma (P > 0.181) ( Table 2). We next utilized Kaplan-Meier survival analysis and found that a higher expression of ANTXR1 predicated a lower survival probability of the glioma patients (P < 0.05) ( Figure 1B).

Establishment of ANTXR1-knockdown model in vitro
To construct ANTXR1-knockdown model in glioma cell lines, we rst packaged shRNA targeting ANTXR1 in lentivirus vector which was labeled by green uorescent protein (GFP) and established ANTXR1-silenced lentivirus. Then we transfected ANTXR1-shRNA (shANTXR1) or shCtrl lentivirus into SHG-44 and U251 cell lines. The two transfected cell lines were observed 72h after lentiviral transfection under uorescence microscope and the photomicrographs demonstrated a >80% uorescence e ciency with cells in good conditions (Figure 2A). Subsequently, the inhibitory effects of lentiviral transfection on the expression of ANTXR1 mRNA level were detected by means of qRT-PCR, with the knockdown e ciency of ANTXR1 in SHG-44 and U251 cells as high as 62.3% (P < 0.001) and 60.8% (P < 0.001), respectively ( Figure 2B). The following western blotting veri ed this nding and demonstrated that ANTXR1 protein was notably downregulated in both glioma cell lines in comparison with controls ( Figure 2C). Collectively, the ANTXR1-knockdown models were successfully established and thus prepared for the following experiments.
Silencing of ANTXR1 hampered in vitro growth and migration of glioma cells We next monitored and compared the capacity of cell growth and migration between ANTXR1-silenced groups and controls in two cell lines. The Celigo cell counting assay unraveled that both cell lines exhibited signi cantly slower cell proliferation rate in the rst 5 consecutive days (SHG-44 cells, P < 0.001, fold change = -2.0; U251 cells, P < 0.001, fold change = -7.0) in comparison with those in shCtrl group ( Figure 3A), indicating a remarkably restriction in cell growth resulted by ANTXR1 knockdown. Thus, we applied ow cytometry to detect cell apoptosis and cell cycle in SHG-44 and U251 cell lines. The results proposed that cell apoptosis was signi cantly enhanced in shANTXR1 groups (SHG-44 cells, P < 0.001, fold change = 9.1; U251 cells, P < 0.001, fold change = 3.1) compared with controls ( Figure 3B). With regard to cell cycle, we disclosed that knockdown of ANTXR1 signi cantly increased the percentage of cells in G1 phrase in SHG-44 cell line (P < 0.01) through ow cytometry but it was not true in U251 cell line. However, silencing of ANTXR1 signi cantly reduced the proportion of cells in S phrase (SHG-44 cells, P < 0.001; U251 cells, P < 0.01) and increased these in G2 phrase in shANTXR1 groups of both cell lines (P < 0.001) ( Figure 3C), which denoted that ANTXR1 held back apoptosis via targeting G2 phrase during cell mitosis.
The ability of tumor cell migration was then detected and quanti ed in both cell lines. The results of wound-healing assay displayed that the cell migration rate (24 h) was lessened by 57% (P < 0.01) after harboring shANTXR1 in SHG-44 cells, while it also dropped by the same count in shANTXR1 group (48 h) compared with controls in U251 cells (P < 0.001) ( Figure 4A). The following transwell assay indicated similar outcomes that knocking down ANTXR1 was associated with 63% (P < 0.001) and 54% (P < 0.001) reduction of migration ability in SHG-44 and U251 cells, respectively ( Figure 4B). On the whole, our data pointed out ANTXR1 downexpression substantially suppressed glioma cell growth and migration in experimental cell lines in vitro.
Underlying molecular mechanism of ANTXR1 modulating apoptosis in glioma cells The increment of cell apoptosis rate allowed us for a further human apoptosis antibody array. We perceived that among the human apoptosis-related proteins, HTRA was upregulated with other proteins downregulated (i.e., Bcl-2, Bcl-w, clAP-2, HSP60, HSP70, IGF-I, IGF-II, Livin, sTNF-R2, TNF-α, TNF-β, TRAILR-3, TRAILR-4 and XIAP) after knockdown of ANTXR1 in the U251 cell line ( Figure 5A-C). Western blotting was thereafter employed in U251 tumor cells and the outcomes exhibited an involvement of the upstream MAPK9 signaling pathway. Speci cally, the transfection of shANTXR1 had a bearing on overexpression of MAPK9 but downregulation of P-AKT, CCND1 and CDK6, while the expression level of AKT barely changed in the U251 cell line ( Figure 5D).

Depletion of ANTXR1 refrained glioma progression in vivo
To verify the effect ANTXR1 exerted on glioma progression in vivo, we subcutaneously injected SHG-44 cells that were transfected with shCtrl or shANTXR1 into nude mice to construct xenograft models and monitored the development of potential tumors. The tumor in mice models of the negative controls grew much faster and the volumes were much larger than those in the shANTXR1 group (P < 0.05) ( Figure 6A). The outcomes of bioluminescence photographing were shown in Figure 6B and total bioluminescent intensity of controls was signi cantly higher than that in shANTXR1 group (P < 0.01). Strikingly, tumor weight of controls also outweighed that after treatment with ANTXR1 lentivirus (P < 0.01) ( Figure 6C). Ki-67 staining proved that this proliferation protein marker was notably downexpressed followed by the depletion of ANTXR1 in comparison with that in control group ( Figure 6D). As a consequence, disruption of ANTXR1 evoked considerable suppression on in vivo tumorigenesis of glioma.

Discussion
Unlike other cancers locating noncerebral tissues, glioma is derived from lesions in the primary central nervous system and is seriously threatening mental health. Patients with glioma were also more likely to be cognitively impaired and suffer from more psychological symptoms [20], highlighting the urgent need for more effective treatments of this intractable disorder. This research displayed that in human clinical glioma tissues, ANTXR1 expression was correlated with the severity of the disorder indicated by the clinicopathologic features. Followed by ANTXR1-lentivirus treatment in glioma cell lines, cell apoptosis were dramatically promoted in vitro, while cell growth and proliferation were limited with similar circumstances encountered in the ability of cell movement and migration. Further western blotting on the regulatory signaling and the downstream apoptosis pathway unveiled the involvement of MAPK signaling pathway and the alterations of apoptosis-related proteins owing to the silencing of ANTXR1. In vivo trials veri ed the aforementioned ndings and ANTXR1 has shown its effectiveness in advancing glioma development. In light of these discoveries, we proposed that ANTXR1 participated in and modulated the process of tumor growth through deactivating MAPK pathway and activating downstream apoptosis pathway in glioma patients.
Consistently, researchers had announced the increased expression level of ANTXR1 and declared that miR-26b-3p acted as its critical upstream modulator in glioma progression [21]. The present study tried to unfold and trace its downstream regulatory mechanisms. The ndings here implied the pivotal role of ANTXR1 in MAPK signaling pathway as its upstream deactivator since downregulation of MAPK9 was greatly attributable to the presence of ANTXR1. MAPK (mitogen-activated protein kinase) signaling cascades include several signaling components (RAS-MAPK) and have been suggested to participate in mediating cellular signal transmissions, thus regulating cell growth and proliferation [22]. This pathway is involved in numerous biological processes and its dysregulation was commonly found in many syndromes and disorders, including a number of human tumors [23,24]. MAPK9, also known as JNK2, was a member of MAPKs targeting speci c transcription factors and thus modulating gene expression during the cell-cycle process [25]. In colitis-induced colorectal carcinoma, MAPK9 functioned as a tumor suppressor and in ammation-triggered inactivation of MAPK9 would result in abnormalities of MAPK9dependent cell cycle [26]. MAPK9 de ciency was also found to support the development of breast cancer [27]. Nevertheless, the carcinostasis of MAPK9 was not always true. In melanoma, MAPK9 played a positively role in cell proliferation and its activity was of the essence in tumorigenesis [28].
Additionally, ANTXR1 shortage also led to reduction of AKT phosphorylation, implying that silencing of ANTXR1 blocked tumorigenesis through inactivating PI3K/AKT signaling pathway and further misregulating CDKs. In our experiments, CDK6, which was particularly needed for G1/S transition, were found overexpressed in glioma cell lines while lack of ANTXR1 lessened CDK6 expression. This result proposed an ANTXR1-mediated association between MAPK9 inactivation and the expression of cell cycle-related proteins. CCND1, also named cyclin D1, functioned as a regulatory subunit of CDK4 or CDK6, both of which were vital regulators in the stepwise progression of cell cycle [29]. CCND1 was considered to be a pan-caner actor and dysregulation of CCND1 had linkage with pathogenesis of human tumors, e.g., prostate cancer [30], retinoblastoma [31] and head and neck squamous cell carcinoma [32]. Ampli cation of CCND1 also conferred unfavorable prognosis in malignant tumors [33]. In agreement with this, we also demonstrated a ANTXR1-mediated cell cycle disruption and growth promotion via activating PI3K/AKT signaling in glioma patients.
Regarding the downstream apoptosis pathway, we declared a series of decreases in apoptosis-related proteins but an increase in HTRA2 following depletion of ANTXR1. HTRA2 (high temperature requirement A2), locating in the mitochondrial inter membrane, has been reported to bind apoptosis inhibitory factors and evoke cell apoptosis to maintain mitochondrial homeostasis, which acknowledgedly served as a favorable target for anticancer agents [34,35]. In prostate cancer, a reduction in HTRA2 abrogated cell apoptotic activity through interacting with C-terminus of ITGA7, a primary laminin receptor [36]. Intracerebral hemorrhage in rat models gave rise to upregulation of HTRA2 and initiated HTRA2-triggered neuronal apoptosis [37]. Researchers previously announced that HTRA2 performed its pro-apoptotic function by interacting with XIAP (X-linked inhibitor of apoptosis protein) in a caspase-dependent manner [38]. Consistent with this notion, our study also detected overexpression of XIAP but downexpression of HTRA2, whereas the case was exactly the opposite after silencing ANTXR1. We hence drew the conclusion that ANTXR1-mediated dysregulation of MAPK signaling in glioma cells contributed to abnormalities in apoptotic pathway by downregulating HTRA2.

Conclusions
In summary, we investigated the ANTXR1 role and its downstream molecular mechanisms in regulating malignant biological behaviors of glioma. Our results sounded ANTXR1 acted as a tumor promoter in glioma induction via deactivating MAPK signaling and apoptosis pathway but activating PI3K/AKTmediated cell growth, speci cally, attenuating MAPK9-mediated gene transcription and HTRA2-induced apoptisis but intensifying CCND1-mediated proliferation. Findings from this study, in conjunction with previous work on ANTXR1 can provide a promising option as anticancer agents for the clinical therapeutics of glioma.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
None.
Authors' contributions Dengfeng Wan, Aijun Liang and Chaoyang Zhou have made substantial contributions to the concept and design of this study. Jianzhong Zhang, Jingxing Leng, Bin Xi and Bin Zhou, conducted the experiments. Yu Yang, Ronglan Zhu, Liangchen Zhong and Xingxing Jiang conducted data analysis. Aijun Liang and Chaoyang Zhou produced the manuscript. All authors have approved the nal draft. Tables   Table 1 The difference in expression patterns of ANTXR1 between glioma tissues and normal brain tissues suggested by immunohistochemistry analysis.  Note, the cut-off age point (43 years old) was the median age of the patients whose ages ranged from 15-80 years old.     shRNA. *P < 0.05, **P < 0.01.