miR-4310 promotes glioma cell proliferation, migration, and invasion in vitro
To explore the roles of miR-4310 in glioma progression, we first compared the expression of miR-4310 between 8 NB tissues and 26 glioma tissues. We found that glioma tissues showed higher levels of miR-4310 than NB tissues (Fig.1 A).
To further verify the biological function of miR-4310 in glioma cells, we transfected the U87 and LN229 glioma cell lines with the miR-4310 lentivirus and its negative control (NC). We thus obtained two groups of glioma cell lines that could stably express miR-4310 and their respective negative control cell lines. We named them U87-LV-NC and U87-LV-miR-4310, and LN229-LV-NC and LN229-LV-miR-4310 and confirmed high miR-4310 overexpression efficiency in these lines using qPCR (Supplementary Fig.1 A-B). This allowed us to use them as tool cell lines in our subsequent studies.
First, we studied the in vitro effects of miR-4310 expression on cell proliferation. For this purpose, we applied Edu incorporation and MTT assays on our U87 and LN229 glioma cell lines. Our results revealed that the overexpression of miR-4310 promoted cell proliferation significantly, whereas the suppression of miR-4310 expression restored the proliferation rate (Fig.1B-C).
Next, we performed the Transwell chamber, Boyden chamber, and wound healing assays, which showed that overexpressed miR-4310 could promote glioma cell migration and invasion, and this function could be restored by miR-4310 inhibition (Fig.1D-F).
In conclusion, our in vitro results showed that miR-4310 could indeed promote the proliferation, migration, and invasion of glioma cell lines.
miR-4310 promotes tumorigenesis in vivo.
The above results prompted us to perform an in vivo tumor formation experiment by subcutaneously injecting U87-miR-4310 or control cells into nude mice. For this purpose, subcutaneous tumor formation was attempted in 10 nude mice, of which tumorigenesis was successfully achieved in 9, and 1 died halfway. After 30 days of implantation, 8 out of the 9 mice injected with U87-miR-4310 cells had larger tumor burdens (Fig. 2A) and displayed higher expression of Ki67 and proliferating cell nuclear antigen (PCNA) in tumor tissues relative to controls (Fig. 2B). These results suggested that miR-4310 significantly promoted tumorigenesis in vivo.
Biological function of miR-4310 is achieved by activating the PI3K/AKT pathway and EMT-associated genes
The epithelial-mesenchymal transition (EMT) process and the PI3K/AKT signaling pathway are known to have an inseparable relationship with tumor cell proliferation, migration, and invasion. Thus, we sought to investigate the expression levels of proteins related to EMT and the PI3K/AKT pathway by western blot. Our results showed that overexpression of miR-4310 upregulated the zinc finger E-box binding homeobox 1 (ZEB1) and N-cadherin, and downregulated E-cadherin (Fig. 1G). We then examined the effect of miR-4310 on the PI3K/AKT pathway. We found that overexpression of miR-4310 significantly increased the phosphorylation of PI3K and AKT, but not their total protein levels and downregulated the downstream molecules of PI3K/AKT, p21, and p27 (Fig. 1H). All the above effects were restored in the presence of the miR-4310 inhibitor.
The above experimental results show that miR-4310 achieves its biological functions by activating the EMT and the PI3K/AKT signaling pathway.
miR-4310 directly targets PTEN
The biological function of miRNAs is achieved by binding to their target gene mRNA. PTEN is an important factor that antagonizes the PI3K/AKT signaling pathway. Therefore, we supposed that PTEN might be the target gene of miR-4310. miR-4310 could release the antagonistic effect of PTEN on the PI3K/AKT pathway by targeting PTEN, thereby activating this pathway.
We used TargetScan (http://www.targetscan.org/) and predicted two PTEN sites that could be direct targets of miR-4310 (Fig. 3A). Overexpression of miR-4310 downregulated the PTEN protein levels but did not affect the mRNA level (Fig. 3B-C). This indicates that miR-4310 can regulate the translation of the PTEN mRNA. This result was confirmed by immunohistochemistry staining in xenografts derived from U87-LV-NC and U87-LV-miR-4310 cells (Fig. 3D).
To further prove that miR-4310 directly targets PTEN, we performed the dual-luciferase reporter gene assay. Our results showed that the luciferase activity was abrogated when co-transfected with wild-type PTEN reporter and miR-4310 mimics (Fig. 3E lane 5). This result could be reversed by co-transfection with wild-type PTEN reporter and miR-4310 inhibitor (Fig. 3E lane 6) but was not affected when co-transfected with mutant PTEN reporter and miR-4310 mimics or inhibitor (Fig. 3E lane 14, 15). Moreover, when we singly co-transfected with mutant site 1 or site 2 of PTEN reporter and miR-4310 mimics, we could still observe that luciferase activity was abrogated (Fig. 3E lane 8, 11), but the degree of abrogation was less than that of lane 5 (Fig. 3E). Thus, the results of the dual-luciferase reporter gene assay illustrate that PTEN is the target gene of miR-4310, and both binding sites are active.
To understand whether the biological function of miR-4310 is achieved through PTEN, we transiently transfected a PTEN plasmid into the U87-LV-miR-4310 and LN229-LV-miR-4310 cell lines to test whether overexpression of PTEN could inhibit the proliferation, migration, and invasion of these cells. The obtained results showed that the overexpression of PTEN could inhibit U87-LV-miR-4310 and LN229-LV-miR-4310 proliferation, migration, and invasion (Fig. 4A-D). The immunoblotting assay data also confirmed that the overexpression of PTEN could affect the EMT and PI3K/AKT associated proteins. (Fig. 4E-F)
Collectively, these results support our initial hypothesis that miR-4310 promotes the activation of the PI3K/AKT pathway by targeting PTEN, thus causing the release of the antagonistic effect of PTEN on the signaling pathway.
SP1 induces the expression of miR-4310 by binding to its promoter region.
SP1 is a zinc finger transcription factor that binds to GC-rich motifs of many promoters. The encoded protein is involved in many cellular processes, including cell differentiation, cell growth, apoptosis, immune responses, response to DNA damage, and chromatin remodeling. According to reports, SP1 plays an important role in the tumorigenesis, progression, and drug resistance of gliomas [24-27]. We found that the promoter region of miR-4310 also contains GC-rich fragments. Thus, we supposed that SP1 may be an upstream signaling molecule of miR-4310 and may regulate miR-4310 expression by binding to the promoter region of miR-4310.
We used UCSC (http://genome.ucsc.edu), PROMO (http://alggen.lsi.upc.es/), and JASPA (http://jaspar.genereg.net/) to predict whether SP1 could bind to certain sequences in the promoter region of miR-4310. Two sites were identified that SP1 could bind to: site A (-1508~-1499) and site B (-1939~-1930) (Fig. 5A). We overexpressed SP1 in U87 and LN229 glioma cell lines and used PCR to detect the level of miR-4310. The results showed that the expression level of miR-4310 increased (Fig. 5B). Similarly, the PTEN/PI3K/AKT pathway and EMT related proteins showed the same regulation. (Fig.5C-D) This proves that SP1 may act as an upstream factor of miR-4310 to regulate the expression of miR-4310.
We then verified our hypothesis using chromatin immunoprecipitation (ChIP) and agarose gel electrophoresis assay on U87 and LN229 cells. Site A and site B sequences were enriched in the anti-SP1 group, indicating that SP1 can bind to site A and site B (Fig. 5E). Subsequently, we used electrophoresis mobility shift assay (EMSA) and dual-luciferase reporter gene assay to further confirm this conclusion (Fig.5 F-G). These results also supported that SP1 could induce the expression of miR-4310 by binding to its promoter region.
Clinical relationship among miR-4310, SP1, and PTEN.
To further clarify the relationship among miR-4310, PTEN, and SP1, we used in situ hybridization (ISH) and immunohistochemistry (IHC) to semi-quantitatively analyze their expression levels in 86 paraffin-embedded glioma tissue samples. Among the 52 samples with low expression levels of miR-4310, 59.6% (31/ 52) also exhibited high expression of PTEN and 40.4% (21/52) exhibited low expression of PTEN, whereas 26.9% (14/52) exhibited high expression of SP1 and 79.1% (38/52) exhibited low expression of SP1. Similarly, of the 34 samples with high expression levels of miR-4310, 40.5% (15/34) presented with high expression of PTEN and 59.5% (19/34) with low expression of PTEN, whereas 58.8% (20/34) presented with high expression of SP1 and 41.2% (14/34) with low expression of SP1 (Fig. 6A-B). Our results showed that SP1 positively correlated with miR-4310 (Chi-square test, p=0.003; Spearman correlation coefficient ρ=0.319). Maybe due to the small size of our sample, we were not able to find a significant negative correlation between miR-4310 and PTEN, but they still had a negative tendency to some extent (Chi-square test, p=0.188; Spearman correlation coefficient ρ=-0.152). The other clinical characteristics of glioma patients are summarized in Table 1.
miR-4310, SP1, and PTEN are independent prognostic factors for glioma.
According to CGGA (http://www.cgga.org.cn), we analyzed the expression levels of SP1 and PTEN in each glioma grade and performed Kaplan-Meier survival analysis based on SP1 and PTEN expressions. The results of this analysis show that the expression level of SP1 increases with the WHO grade of glioma (Supplementary Fig. 1 C, E). Moreover, the Kaplan-Meier survival analysis showed that high expression of SP1 is associated with poor prognosis in glioma patients (Supplementary Fig. 1 D, F). Similarly, high expression of PTEN is associated with a good prognosis in glioma patients (Supplementary Fig. 1 G-H). We still performed a Kaplan–Meier survival analysis on 86 glioma patients. The result showed that patients with high SP1 expression had poorer OS rates than those exhibiting low SP1 expression (Supplementary Fig. 1 I).
Although neither p-value reached significance, in these tissue samples, patients with high miR-4310 expression or PTEN expression tended to have longer or shorter survival, respectively (Supplementary Fig. 1 J-K). We suspect that the non-significance is due to the small sample size and mutual interference among SP1, PTEN, and miR-4310.
Furthermore, we used univariate and multivariate COX regression to analyze whether miR-4310, SP1, and PTEN are independent prognostic factors for glioma. We first used data from these 86 cases to analyze the relationship between various factors and the overall survival (OS) of glioma patients (Table 2). Age, gender, WHO grade, SP1, PTEN, and miR-4310 were included in this study as possible prognostic factors.
The multivariate COX regression results revealed that miR-4310 (p= 0.002) and WHO grade (p<0.0001) are prognostic factors for glioma. We next performed a survival analysis based on the miR-4310 expression (Fig. 7D), and these results suggested that high expression of miR-4310 is associated with poor prognosis in glioma patients.
In addition, considering that SP1 and PTEN are also common risk factors clinically, we also used the CGGA-693 database for a further COX regression analysis. Age, gender, WHO grade, SP1, PTEN, IDH mutation, chemotherapy, and radiotherapy were included in the study as possible prognostic factors. The results based on CGGA-693 revealed that WHO grade (p<0.0001), SP1 (p=0.006), PTEN (p<0.0001), IDH mutation(p<0.0001), and chemotherapy (p=0.004) are independent prognostic factors for glioma (Table 3). Survival analysis results suggest that high expression of SP1 is associated with poor prognosis in glioma patients. Similarly, high expression of PTEN correlates with benign prognosis in glioma patients. (Fig. 6 D-E)