miR-182-5p Induced by STAT3 Activation Promotes Glioma Tumorigenesis

Glioma is one of the most common types of primary brain tumor in humans. Glioblastoma is the highest grade and most aggressive malignant glioma (1). A major feature of malignant glioma is its diffuse infiltrative growth into surrounding brain tissue (2). The oncogenic transcription factor STAT3, a component of the Janus-activated kinase (JAK)/STAT3 signaling pathway, plays a key role in many physiologic and pathophysiologic processes including cancer growth, invasion, and metastasis (3–5). STAT3 is mainly activated via tyrosine phosphorylation (Tyr705) by growth factor/cytokine receptors and non–receptor tyrosine kinases, which results in STAT3 nuclear translocation and thus subsequent activation of various STAT3 target genes that regulate cell proliferation, migration, and invasion (6, 7). In glioma, activation of STAT3 often positively correlates with tumor malignancy and indicates poor prognosis (8–10).

miRNAs are a class of evolutionally conserved small, noncoding RNA molecules (∼19–23 nucleotides) that are involved in posttranscriptional regulation of gene expression in animals, plants, and viruses (11, 12). The alteration in miRNA expression profiles represents a common feature in all human cancers, indicating the importance of such alterations in cancer progression and prognosis (13). miR-182-5p is a well-described oncomiR that is overexpressed in a number of malignancies including glioma (14), lung (15), breast (16), prostate (17), liver (18), and colorectal (19) cancers. Given the importance of both miR-182-5p and STAT3 in the processes linked to cancer, we sought to test whether both may be part of a common pathway. Moreover, STAT3-induced transcription of protein-coding genes has been extensively studied; however, little is known about STAT3-regulated miRNA gene transcription.

In this study, we present evidence that STAT3-induced miR-182-5p promotes glioma cell growth, migration, and invasion by suppressing PCDH8 (protocadherin 8). In addition, the STAT3–miR-182-5p–PCDH8 axis is critical for glioma tumor growth in vivo. We found that expression of both p-STAT3 and miR-182-5p is upregulated in glioblastoma compared with their expression in normal brain tissues; however, expression of PCDH8 is higher in normal brain tissues than in tumor tissues. The expression levels of PCDH8 are inversely correlated with those of p-STAT3 or miR-182-5p in glioblastoma tissues. Therefore, our data shed light on the molecular mechanisms of the growth- and invasion-promoting function of the axis and suggested that the axis may include potential prognostic markers and therapeutic targets.

The expression vector encoding STAT3 CA was obtained from Addgene (Plasmid #24983). SMARTpool siRNA duplexes specific for STAT3, PCDH8, and a nontargeting siRNA (siControl) were purchased from Dharmacon. Recombinant human EGF was purchased from R&D Systems. miR-182-5p mimics and inhibitor were obtained from Sigma-Aldrich. WP1066 was synthesized at The University of Texas MD Anderson Cancer Center (Houston, TX). STAT3 (#4904) and p-STAT3 (#9131) were purchased from Cell Signaling Technology. PCDH8 (#ab55507) was purchased from Abcam. β-Actin (#sc-47778) was purchased from Santa Cruz Biotechnology.

HEK 293T cells and the human glioma cell lines SW1783, Hs683, LN229, and U87 were purchased from ATCC between years 2009 and 2015. The HFU251 human glioma cell line was obtained from our brain tumor center core and was described previously (20). The cell lines were subjected to cell line characterization and authentication using short tandem repeat profiling in every 6 months, and the cells were passaged continuously for fewer than 6 months after receipt in our laboratory for relevant studies reported here. All of the cell lines were grown in DMEM containing 10% FBS (HyClone).

HEK 293T cells were transfected with FuGENE HD (Roche). SW1783, Hs683, HFU251, and U87 cells were transfected with X-tremeGENE HP (Roche). SW1783 and U87 cells were transfected with siRNA using Lipofectamine RNAiMAX (Invitrogen).

For STAT3 or PCDH8 silencing of expression, lentiviral particles were produced with use of a lentivirus packaging mix (ViraPower, Invitrogen). For viral transduction, the cells were seeded at 50%–60% confluence. The cells were treated overnight, with the medium containing lentivirus harvested 48 or 72 hours after transfection. Virus particles were concentrated and purified by ultra-high-speed centrifugation (25,000 × g for 2 hours at 4°C). Puromycin or hygromycin (Sigma-Aldrich) was used to select stable clones.

Cells were lysed with RIPA lysis buffer; cell lysates were then subjected to an 8%–12% SDS-PAGE, transferred onto polyvinylidene difluoride membrane, and probed with antibodies, as previously described (21).

The miR-182-5p promoter was produced by PCR by using primers 5′-TTTACGCGTGTTGTTGTTGAGACAGAATCTCGCT-3′ (forward) and 5′- TTTAAGCTTCCTGCCGACCCTGCGGAGAGA-3′ (reverse). In the promoter assay, SW1783 cells were transfected with pGL3-miR-182-5p, STAT3 CA plasmid, and Renilla. The wild-type (WT) and mutant 3′-UTRs luciferase reporter of the pcdh8 gene were generated by annealing the forward and reverse oligonucleotides and were then cloned into psi-Check2 vector. Oligonucleotides for the pcdh8 3′-UTR were as follows: WT, 5′-TCGAGATAATAAAAGCTAAAGTGGAAGTTATTGCCAAAGGAACTGTCTGTAGC-3′ (forward) and 5′-GGCCGCTACAGACAGTTCCTTTGGCAATAACTTCCACTTTAGCTTTTATTATC-3′ (reverse); mutant, 5′- TCGAGATAAT AAAAGCTAAAGTGGAAGTTATGGACCAAGGAACTGTCTGTAGCC-3′ (forward) and 5′-GGCCGCTACAGACAGTTCCTTGGTCCATAACTTCCACTTTAGCTTTTATTATC-3′ (reverse). In the 3′-UTR reporter assay, HEK 293T cells were transfected with the psi-Check2-PCDH8 3′-UTR (WT or mutant), Renilla, together with miR-182-5p mimics/inhibitor or with control mimics/inhibitor. Luciferase activity was measured in a 1.5-mL Eppendorf tube with the Promega Dual-Luciferases Reporter Assay Kit (Promega E1980) according to the manufacturer's protocols after transfection. Relative Renilla luciferase activity was normalized to firefly luciferase activity.

Total RNA was extracted with use of TRIzol reagent (Invitrogen) for miR-182-5p or PCDH8 mRNA analyses. For detection of miR-182-5p expression, stem-loop RT-PCR was performed. Quantitative real-time RT-PCR was performed by using the SYBR green reagent with an ABI Prism 7000HT sequence detection system. Relative expression was evaluated by a comparative C_T method and normalized to the expression of U6 small RNA or GAPDH. The stem loop primer for miR-182-5p is 5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACTGGATACGACAGTGTG-3′. The qPCR primer for miR-182-5p was purchased from Sigma.

U87 or LN229 cells (2 × 10⁶) were prepared for the chromatin immunoprecipitation (ChIP) assay by using a ChIP Assay Kit (Cell Signaling Technology) according to the manufacturer's protocol. The resulting precipitated DNA specimens were analyzed by using PCR to amplify a 298-bp region (site 1) of the miR-182-5p promoter with the primers 5′-GGTTTCACCATGCTAACCAGGCT-3′ (forward) and 5′- TAGCCATAAGCATCACCCAAGGAG-3′ (reverse) and a 300-bp region (site 2) of the miR-182-5p promoter with the primers 5′-GGCTCAAGAATCTACATTTCAACAG-3′ (forward) and 5′-ACACACCCCCACTACAGGGCTCT-3′ (reverse). The negative control is an encoding region of mir-182-5p, which was amplified by PCR with the primers 5′-ATAGTTGGCAAGTCTAGAAC-3′ (forward) and 5′-TCTGTCTCTTCCTCAGCACA-3′ (reverse). The PCR products were resolved electrophoretically on a 2% agarose gel and visualized with use of ethidium bromide staining.

For the cell proliferation assay, 5 × 10³ cells were plated in 96-well plates. Cell growth was determined by using a standard tetrazolium bromide (MTT) assay and verified by counting with trypan blue and a hemocytometer. Data represent the means ± SD of 3 independent experiments.

After transfection, SW1783 or U87 cells (5 × 10⁴) in a 500-μL volume of serum-free medium were placed in the upper chambers of a Transwell and incubated at 37°C for 16 hours for the migration or invasion assay. The cells that penetrated the uncoated (migration) or Matrigel-coated (invasion) filters were counted at a magnification of ×200 in 15 randomly selected fields, and the mean number of cells per field was recorded.

All mouse experiments were approved by the Institutional Animal Care and Use Committee of The University of Texas MD Anderson Cancer Center. We intracranially injected glioblastoma stable cells into 4-week-old female athymic nude mice. Five mice were injected for each group. Mice were euthanized 4 weeks after the glioblastoma cell injection. The brain of each mouse was harvested, fixed in 4% formaldehyde, and embedded in paraffin. Tumor formation and phenotype were determined by histologic analysis of hematoxylin and eosin (H&E)-stained sections.

Human or mouse normal brain and brain tumor tissues were stored frozen at −75°C until use for qPCR analysis. For immunohistochemical (IHC) analysis, mouse tumor tissues were fixed and prepared for staining. The specimens were stained with Mayer's hematoxylin and subsequently with eosin (Biogenex Laboratories) or with antibodies against p-STAT3 and PCDH8. The tissue sections from paraffin-embedded human glioblastoma specimens were stained with antibodies against p-STAT3 and PCDH8. IHC analyses were performed on the tissue arrays by a standard immunostaining protocol (22). The use of human brain tumor specimens was approved by the Institutional Review Board at The University of Texas MD Anderson Cancer Center and written informed consent was obtained from the patients for the use of clinical specimens for research.

The significance of the data from patient specimens was determined by the Pearson correlation coefficient test. The significance of the in vitro data and in vivo data between experimental groups was determined by Student t test or Mann–Whitney U-test. P < 0.05 was considered to be statistically significant.

To ascertain the role of miR-182-5p in glioma tumorigenesis, we first determined that miR-182-5p expression by quantitative real-time RT-PCR analysis was significantly increased in human glioblastoma tissues compared with its expression in normal human brain tissues (Fig. 1A). Quantitative real-time RT-PCR analysis revealed that miR-182-5p expression was markedly increased in a panel of glioma cell lines tested (Hs683, SW1783, HFU251, and U87), compared with its expression in normal human astrocytic cells. Among the glioma cell lines, miR-182-5p showed lower expression in the grade III–derived glioma cell lines (Hs683 and SW1783) relative to the glioblastoma grade IV cell lines (HFU251 and U87; Fig. 1B). Furthermore, we found that the glioblastoma cell lines expressed higher levels of activated STAT3 (p-STAT3) than the grade III glioma cell lines (Fig. 1C), suggesting the levels of miR-182-5p directly correlated with the expression of activated STAT3. When the SW1783 cells were transfected with a constitutively activated mutant of STAT3 (STAT3 CA), miR-182-5p was increased (Fig. 1D). In addition, when the SW1783 cells were treated with a ligand for STAT3 activation, EGF, miR-182-5p was upregulated (Fig. 1E), suggesting that STAT3 activation positively regulates miR-182-5p expression. In contrast, inhibition of STAT3 by siRNA or WP1066 (23–25), a STAT3 inhibitor, strongly reduced miR-182-5p expression levels in U87 cells (Fig. 1F and G).

Next, we analyzed the sequence of the miR-182-5p promoter by using the STAT3 consensus sequences AA (N4) TT and AA (N5) TT. We identified 2 putative STAT3-binding sites in the miR-182-5p promoter (sites 1 and 2, Fig. 1H). To provide direct evidence that miR-182-5p is a direct transcriptional target of STAT3, we performed standard ChIP assays in glioma cells using primers that flank a 298-bp region (containing STAT3 binding site 1) or a 300-bp region (containing STAT3 binding site 2) of the miR-182-5p promoter. We found that in U87 cells, endogenous STAT3 protein bound to both of the STAT3-binding regions of the miR-182-5p promoter, but not to an miRNA-encoding region (MER) of miR-182-5p (Fig. 1H). Similar results was detected in LN229 cells (Fig. 1H). Moreover, knocking down STAT3 by siRNA in U87 cells reduced the binding of miR-182-5p promoter to STAT3 protein (Supplementary Fig. S1). These findings indicated that STAT3 protein directly interacts with the miR-182-5p promoter. To provide further evidence that miR-182-5p is a direct transcriptional target of STAT3, we examined whether STAT3 activation transactivates the miR-182-5p promoter by using a miR-182-5p promoter luciferase reporter with the STAT3 CA or the control vector. Luciferase activity driven by the miR-182-5p promoter increased in SW1783 cells overexpressing STAT3 CA (Fig. 1I). These results suggested that STAT3 is involved in miR-182-5p transcriptional activation.

To determine whether miR-182-5p promotes an oncogenic phenotype of glioma, we performed overexpression or knockdown function assays in glioma cells by using an miR-182-5p mimics or inhibitor. Overexpression of miR-182-5p resulted in a significant increase in viability in Hs683 and SW1783 cells in an MTT assay (Fig. 2A). In contrast, as shown in Fig. 2B, miR-182-5p inhibition significantly reduced glioma cell growth compared with the scramble control cells in HFU251 and U87 cells, as measured by an MTT assay.

Furthermore, we performed Transwell migration and invasion assays to determine whether miR-182-5p regulates glioma cell migration and invasion abilities. We found that miR-182-5p increased the ability of SW1783 cells to migrate and invade (Fig. 2C). In contrast, miR-182-5p inhibition significantly reduced U87 cells migration and invasion abilities (Fig. 2D).

To identify the miR-182-5p–mediated downstream regulator, 3 target prediction algorithms (PicTar, TargetScan, and miRanda) were used. We analyzed first 100 picks from each of the three algorithms and found that 5 genes are overlapped, namely, ARF4, CTTN, EDNRB, PCDH8, and RASA1 (Fig. 3A). To determine whether miR-182-5p targets these genes, we generated luciferase reporter plasmids that harbor miR-182-5p target sequences in 3′-UTR (untranslated repeat) of these genes. By using these luciferase reporter plasmids, we found that miR-182-5p mimics did not significantly inhibit the activities of 3′-UTR luciferase reporters of ARF4, CTTN, EDNRB, or RASA1 gene (Supplementary Fig. S2). However, expression of miR-182-5p mimics caused a strong inhibition in the activities of 3′-UTR luciferase reporter of PCDH8 gene (Supplementary Fig. S2). Indeed, PCDH8 contains miR-182-5p target and seed motifs, located in the 3′-UTR (Fig. 3B). It has been shown that PCDH8 is a tumor suppressor in a variety of cancers, and its overexpression is associated with a better prognosis of those cancers (26). Therefore, we decided to further evaluate PCDH8 in this study. Expression of miR-182-5p in Hs683 and SW1783 cells resulted in a significant decrease in PCDH8 protein levels (Fig. 3C), whereas inhibition of miR-182-5p in HFU251 and U87 cells resulted in increased PCDH8 protein levels (Fig. 3D). Analyses with use of 3′-UTR of luciferase reporter plasmids containing the miR-182-5p target sequences (WT or mutant) on PCDH8 were performed to determine whether PCDH8 was a direct target of miR-182-5p (Fig. 3B). Expression of miR-182-5p caused a significant decrease in luciferase activity of the WT PCDH8 3′-UTR but not of the mutant form (Fig. 3E). In contrast, inhibition of miR-182-5p increased the luciferase activity in the WT PCDH8 3′-UTR but not in the mutant form (Fig. 3F). Together, these results suggest that miR-182-5p downregulates PCDH8 expression by directly targeting its 3′-UTR.

To determine whether miR-182-5p can functionally target PCDH8, we performed rescue experiments to further demonstrate the importance of PCDH8 in glioma cell growth, migration, and invasion by miR-182-5p. As shown in Fig. 4A, PCDH8 expression was significantly inhibited by two different siRNAs specifically targeting PCDH8 compared with its expression in the control in SW1783 cells. We then tested the effect of PCDH8 knockdown in the cells. Knockdown of PCDH8 enhanced the proliferation, migration, and invasion of SW1783 cells (Fig. 4B–D, respectively). Furthermore, we found that the downregulated miR-182-5p expression by its inhibitor significantly suppressed the proliferative, migration, and invasive capabilities of U87 cells, whereas siRNA knockdown of PCDH8 compromised the inhibitory effect mediated by miR-182-5p (Fig. 4E–G). Thus, our data confirmed that PCDH8 is a functional target by miR-182-5p.

To evaluate the importance of the STAT3–miR-182-5p–PCDH8 pathway in glioma progression, we first evaluated the consequences of targeting of STAT3 in cell growth. As anticipated, cell growth in U87 cells was significantly decreased with STAT3 knockdown or WP1066 treatment. When miR-182-5p mimics or PCDH8 siRNA was transfected into the STAT3-inhibited U87 cells, the proliferative potential of these cells was restored (Fig. 5A and B). In addition, apoptosis in U87 and HFU251 cells significantly induced with STAT3 knockdown or WP1066 treatment, whereas the proapoptotic effect of siSTAT3 or WP1066 in these cells was inhibited by miR-182-5p mimics or PCDH8 siRNA (Supplementary Fig. S3A and S3B). However, we found that miR-182-5p mimics or PCDH8 siRNA alone did not affect apoptosis of U87 cells (Supplementary Fig. S3C). These results, in combination with the results in Figs. 2 and 4 showing that miR-182-5p as well as PCDH8 play roles in cell proliferation, suggest that suppression of STAT3 inhibition–induced apoptosis by miR-182-5p mimics or PCDH8 siRNA is likely caused by increased cell proliferation.

Next, we examined the role of STAT3–miR-182-5p–PCDH8 pathway in cell migration and invasion. We found that migration and invasion of U87 cells were notably blocked after STAT3 knockdown (Fig. 5C) and after WP1066 treatment (Fig. 5D). However, miR-182-5p mimics or PCDH8 siRNA transfection, respectively, rescued these inhibition effects on migration and invasion (Fig. 5C and D). Moreover, cell–cell adhesion in HFU251 cells was significantly increased by STAT3 knockdown (Supplementary Fig. S3D). When miR-182-5p mimics or PCDH8 siRNA was transfected into the STAT3-inhibited HFU251 cells, the cell–cell adhesion was inhibited (Supplementary Fig. S3D). However, cell–matrix adhesion in U87 and HFU251 cells was not significantly affected by STAT3 knockdown, and miR-182-5p mimics or PCDH8 siRNA also did not change cell–matrix adhesion in the STAT3-inhibited U87 and HFU251 cells (Supplementary Fig. S3E). Thus, the effect of STAT3–miR-182-5p–PCDH8 pathway in cell migration and invasion is likely due, at least in part, to the role of the pathway in cell–cell adhesion. Collectively, the above results suggest that the miR-182-5p–PCDH8 axis is critical for STAT3-induced glioma cell growth, migration, and invasion.

The relevance of our in vitro findings was next confirmed in vivo by intracranial injection of U87 cells into athymic nude mice. U87 shControl cells did show detectable tumor formation within 4 weeks after injection. In contrast, depleting U87 cells of STAT3 significantly suppressed tumor formation, which was abrogated by miR-182-5p overexpression or knockdown of PCDH8 (Fig. 6A). Similar effects were also observed in HFU251 stable cells (Fig. 6B). We then used IHC staining to evaluate the expression levels of p-STAT3 and PCDH8 in normal brain tissues and in brain tumors produced by U87 cells in nude mice. The expression of PCDH8 was higher in normal brain tissues than in tumor tissues, whereas p-STAT3 expression in the nucleus was higher in tumor tissues (Fig. 6C). Moreover, the expression level of miR-182-5p was significantly higher in tumor tissues than in normal brain tissues by quantitative real-time RT-PCR analysis (Fig. 6D). Together, these findings suggest that the STAT3-miR-182-5p-PCDH8 pathway promotes tumor progression in vivo.

To determine whether our findings have clinical relevance, we examined p-STAT3 and PCDH8 expression levels in serial sections of 20 human primary glioblastoma specimens by IHC analyses. The expression level of PCDH8 in the tumors was significantly lower than that in adjacent normal brain tissues; however, the level of p-STAT3 in tumors was high (Fig. 7A). We further found in tumors that levels of p-STAT3 inversely correlated with levels of PCDH8 expression (Fig. 7B). Quantification of staining showed that this correlation was statistically significant among the 20 specimens (r = −0.607; P < 0.01; Fig. 7B). Moreover, there was a statistically significant correlation between miR-182-5p levels and PCDH8 transcript levels by quantitative real-time RT-PCR analysis (r = −0.614; P < 0.01; Fig. 7C). These results further support a critical role for the STAT3–miR-182-5p–PCDH8 pathway promoting tumor progression in human glioblastoma.

It is well-known that miRNAs are critical for the development of the malignant phenotype of glioma cells (27, 28).miR-182-5p is emerging as an important regulator of various physiologic and pathologic processes. miR-182-5p has also been reported to be an oncogene in many types of cancers (14–19, 29–31). In this study, we showed that miR-182-5p promoted glioma cell growth, migration, and invasion. More importantly, the clinical relevance of miR-182-5p was confirmed in human glioblastoma samples. Therefore, these results firmly established miR-182-5p as a functional mediator of glioma tumorigenesis.

The molecular mechanisms of miR-182-5p regulation in cancer are unclear. STAT3-induced transcription of protein-coding genes has been extensively studied in human cancers; however, the role of STAT3 in the transcription of non–protein-coding genes, such as miRNAs, is less explored. Recent results indicated that the STAT3 transcription factor plays an important role in miRNA expression. For example, STAT3 activated by IL6 directly represses the miR-34a gene, which was shown to be necessary for invasion and metastasis of human colorectal cancer cells and was related to nodal and distant metastasis in patients (32).Furthermore, STAT3-mediated activation of miRNA-23a has been shown to decrease glucose production by directly targeting PGC-1α and G6PC in hepatocellular carcinoma, thereby facilitating liver tumorigenesis (33). STAT3-activated miR-21 and miR-181b-1 are also required for sphingosine-1-phosphate lyase downregulation that promotes colon carcinogenesis (34).These data cumulatively suggest that STAT3 and miRNAs play a critical role in human cancers. In this study, we provided strong evidence that the activation of miR-182-5p induced by STAT3 activation is important for glioma cell growth, migration, and invasion. We found that STAT3 directly binds two sites in miR-182-5p promoter regions and is required for transcriptional induction of miR-182-5p. In contrast, STAT3 activation by EGF treatment resulted in the upregulation of miR-182-5p. Conversely, inhibition of STAT3 by siRNA or WP1066 strongly reduced miR-182-5p expression levels, suggesting that STAT3 and miR-182-5p are required for glioma cell growth, migration, and invasion. This study further confirmed that STAT3 acts as a key molecule in the regulation of miRNA in human glioma. Moreover, the results of our study are consistent with the finding in a previous reporter that miR-182 could be a valuable predictor of poor prognosis for patients with glioma (14). Interestingly, another previous study reported that elevated miR-182 levels in proneural subtype glioblastoma tumors negatively correlate with Bcl2L12 levels, and miR-182 increased the apoptosis induced by pan-kinase inhibitor staurosporine in glioblastoma cells through downregulation of Bcl2L12 (35). It has been shown that activated STAT3 protect tumor cells from apoptosis by upregulating the transcription of Bcl2 family genes such as Bcl-2, Bcl-xL and Mcl-1 (36). Thus, it is possible that in STAT3-activated tumor cells, the effect of miR-182–Bcl2L12 axis on apoptosis is counteracted by the effects of Bcl-2, Bcl-xL, and Mcl-1, whereas miR-182-PCDH8 axis–induced cell growth could further enhance the antiapoptotic effects of Bcl-2, Bcl-xL, and Mcl-1.

PCDH8, a member of the cadherin family, has been shown to be a tumor suppressor involved in tumorigenesis. PCDH8 inhibits oncogenesis in epithelial human cancers by suppressing tumor cell proliferation and inhibiting cell migration (26). The PCDH8 gene promoter is frequently methylated in renal cell carcinoma (∼58%), which suggests that it could act as a tumor suppressor or that its tumor suppressor activity is downmodulated in the methylated state (37, 38).Moreover, several studies report that the epigenetic inactivation of PCDH8 by promoter methylation is associated with the loss of its tumor-suppressive function in human nasopharyngeal carcinoma (39), gastric cancer (40), bladder cancer (41),and prostate cancer (42). However, the status of PCDH8 expression in gliomas is unknown. In this study, we found that PCDH8 expression was significantly downregulated in mouse or human brain tumor tissue as compared with normal brain tissue. We also observed a significant negative correlation between p-STAT3 and the PCDH8 protein in human glioblastoma. Moreover, PCDH8 expression was significantly increased by miR-182-5p inhibitor. These findings indicate that the main mechanism in the reduction of PCDH8 protein is miRNA regulation at the posttranscriptional level. Certainly other miRNAs can target PCDH8 such as miR-429, which has been shown to be a regulator of the epithelial-to-mesenchymal transitional during embryo implantation (43).miR-429 has been shown to inhibit glioma invasion through suppression of a big MAPK (44); however, the link between miR-429 and STAT3 is unknown.

In summary, STAT3 activation induces miR-182-5p to repress PCDH8 expression, leading to glioma cell growth, migration, and invasion. Our results reveal the STAT3–miR-182-5p–PCDH8 axis as potential targets for the development of new therapeutic agents for human glioma.

No potential conflicts of interest were disclosed.

Conception and design: J. Xue, A. Zhou, S. Huang

Development of methodology: J. Xue, A. Zhou, Y. Wu, K. Lin, S. Huang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Xue, Y. Wu, S.-A. Morris, R. Verhaak, G.N. Fuller, K. Xie, S. Huang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Xue, Y. Wu, S.-A. Morris, K. Lin, S.B. Amin, K. Xie, A.B. Heimberger, S. Huang

Writing, review, and/or revision of the manuscript: J. Xue, A. Zhou, S.-A. Morris, G.N. Fuller, A.B. Heimberger, S. Huang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Lin, S. Huang

Study supervision: S. Huang

The authors thank Tamara Locke in The University of Texas MD Anderson Cancer Center's Department of Scientific Publications for editing this article.

This work was supported in part by U.S. National Cancer Institute grants R01CA157933, R01CA182684, P50CA127001, and CA16672 (Cancer Center Support Grant) and the National Natural Science Foundation of China (81502184 to J. Xue).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.