CircZNF609 and circNFIX as possible regulators of glioblastoma pathogenesis via miR-145-5p/EGFR axis

Glioblastoma is a rare and deadly malignancy with a low survival rate. Emerging evidence has shown that aberrantly expressed circular RNAs (circRNAs) play a critical role in the initiation and progression of GBM tumorigenesis. The oncogenic function of circZNF609 and circNFIX is involved in several types of cancer, but the role and underlying mechanism of these circRNAs in glioblastoma remain unclear. In this study, we hypothesized that circZNF609 and circNFIX may regulate EGFR through sponging miR-145-5p. Herein, we assessed the expression levels of circZNF609, circNFIX, miR-145-5p, and EGFR using quantitative polymerase chain reaction in glioblastoma patients and normal brain samples. The results showed that circZNF609, circNFIX, and EGFR expression levels were upregulated and miR145-5p was downregulated (p = 0.001, 0.06, 0.002, and 0.0065, respectively), while there was no significant association between clinicopathological features of the patients and the level of these genes expression. We also found a significant inverse correlation between miR145-5p and the expression of cZNF609, cNFIX and EGFR (p = 0.0003, 0.0006, and 0.009, respectively). These findings may open a new window for researchers to better understand the potential pathways involved in GBM pathogenesis. In conclusion, it may provide a new potential pathway for the development of effective drugs for the treatment of GBM patients.

regulating miRNAs. 6,7.MicroRNAs (miRNAs), a subset of non-coding RNAs conserved across species, regulate gene expression by binding to the 3′-untranslated region of target mRNAs 8 .In recent years, many studies focused on the regulatory role of miRNAs as the core of the competitive endogenous RNA (ceRNA) hypothesis that explains the relationship between different classes of RNAs including circRNA, long noncoding RNA (lncRNA), miRNAs and mRNAs 9 .Hsa-miR-145-5p, one of the targets of several dysregulated circRNAs identified in studies, has been widely confirmed to be associated with cancer 8,10 .Evidence has shown that miR-145-5p can be regulated by different mechanisms and can involve in tumorigenesis and progression through binding to 3′-UTRs of target mRNAs 10,11 .Several studies have reported that miR-145-5p can significantly suppress tumor progression, including in colorectal cancer (CRC) 12 , lung cancer 13 , and glioma 14 via targeting epidermal growth factor receptor (EGFR) gene.EGFR is a trans-membrane receptor tyrosine kinase of the ERBB family and has a role in cell proliferation, differentiation and cell survival.An overexpression of either EGFR or EGFR-activating mutations is associated with poor clinical prognosis in various human tumors 15,16 .Despite many reports on the molecular mechanism of EGFR, circRNA-induced EGFR regulation in glioblastoma is still poorly understood.
In the current study, using literature review and bioinformatics analysis from different web tools and databases including NCBI Gene Expression Omnibus (GEO) (https:// www.ncbi.nlm.nih.gov/ geo/), Circular RNA Interactome database (https:// circi ntera ctome.irp.nia.nih.gov/), miRNet (http:// www.mirnet.ca) and IntaRNA web servers (https:// rna.infor matik.uni-freib urg.de/ IntaR NA/ Input.jsp), and miRTarBase (http:// miRTa rBase.cuhk.edu.cn/), we predicted the presence of a regulatory axis, namely, circZNF609, circNFIX/miR-145-5p/EGFR axis.CircZNF609 was the alias of hsa_circ_0000615 which derived from the host gene ZNF609.Dysregulation of CircZNF609 has been reported in a few cancers such as lung adenocarcinoma 17 , nasopharyngeal carcinoma 18 , and hepatocellular carcinoma 19 .Another circRNA, hsa_circ_0049658, originated from host gene nuclear factor IX (NFIX) mRNA.Previous studies have shown that circNFIX is upregulated in liver cancer, non-small cell lung cancer and glioma, and functions as an oncogene involved in cancer metabolism, chemotherapy resistance, migration and invasion 20 .However, the influence and mechanism of these circRNAs in glioblastoma needs further research, and whether circZNF609 and circNFIX may play a role as ceRNAs in GBM pathogenesis remains to be investigated.So, we hypothesized that circZNF609 (circ0000615) and circNFIX (circ0049658) regulate EGFR expression through sponging miR-145-5p.The relative expressions of circZNF609, circNFIX, miR-145-5p and EGFR were then evaluated in glioblastoma compared to normal brain tissues.Also, the relationship between the expression of these genes and the clinical features of patients such as age, gender and tumor diameter was assessed.

Specimens collection and ethics
A total of 51 samples, including 30 tumor tissues from glioblastoma patients and 21 normal brain tissues were collected from the Department of Neurosurgery of Sina Hospital and Shafa Neuroscience Research Center at Khatam Al-Anbia Hospital.All tissue samples were diagnosed by two independent pathologists.All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.The study was approved by the Ethical Committee of Tehran University of Medical Sciences (Ethical code: IR.TUMS.MEDICINE.REC.1401.487),and written informed consent was obtained from all subjects.None of the patients underwent radiotherapy or chemotherapy before surgery.The clinicopathological features including age, gender, and tumor size of patients were attained from their medical records.Control brain samples were taken from patients with a kind of cerebral hemorrhage that have needed surgery and cerebral lobectomy at the same time.The characteristics of the study population are shown in Table 1.

Bioinformatics analysis and primer design
We screened differential miRNAs in glioblastoma and normal tissues from the GEO dataset (GSE165937, NCBI) using the edgeR package in R software with |logFC|≥ 1, adjust P value < 0.05.Volcano diagram of differentially expressed miRNAs was plotted using ggplot2.
The suitable primers were designed with Primer Blast and Gene Runner software.Moreover, the specificity of divergent primers for hsa_circ_0000615 and hsa_circ_0049658 was checked by circPrimer 1.2 software.The sequences of primers in this study are shown in Table 2.

RNA extraction and quantitative real-time PCR
Approximately 15 mg of frozen tissue samples using liquid nitrogen, were homogenized.Total RNA was then extracted using Kiazol reagent (Kiazist, Iran) according to the manufacturer's instructions, and then quantity and purity of RNAs were measured by Nanodrop 2000C (Thermo Scientific, USA) and electrophoresis on 1.8% (wt/vol) agarose gel.RNA samples were converted into cDNA using the cDNA synthesis kit (Yekta Tajhiz Azma, Iran).Reverse transcription of mRNA and miRNA was performed using random primers and specific stem-loop primers, respectively.The reaction condition was set at 70 °C for 5 min, 42 °C for 60 min, and 70 °C for 5 min.
The relative expression levels of all genes were evaluated using RealQ Plus 2 × Master Mix Green without Rox (Ampliqon, Denmark) using the LightCycler 96 (Roche, Germany).Thermal cycling conditions for circZNF609, circNFIX and EGFR include preincubation at 95 °C for 900 s followed by amplification in 40 cycles at 95 °C for 15 s and 60C for 50 s.The suitable annealing temperature for miR-145-5p was 58 °C.SnRNA U6 as a housekeeping gene for miR-145-5p; and GAPDH as a reference gene for circRNAs and EGFR.PCR products were analyzed by melting curve analysis and agarose gel electrophoresis.The back-splice junction sequences of circZNF609 and circNFIX were confirmed by Sanger sequencing.All of the qPCR reactions were performed in duplicate and data were calculated by comparing Ct values.

Statistical analysis
All statistical analysis was performed using GraphPad Prism 8.0 (GraphPad Software, Inc., San Diego, CA) software.Normality was checked by the Shapiro-Wilk test.The relative expression of circZNF609, circNFIX, miR-145-5p, and EGFR was evaluated in glioblastoma tumors and normal brain tissues using the unpaired sample t test.The association between the expression of circRNAs, miR-145-5p and EGFR with clinicopathological features of patients was calculated using Mann-Whitney and one-way analysis of variance (ANOVA) tests.Pearson correlation coefficient was used for the correlation of expression of genes analyses.Also, the receiver operating characteristic (ROC) curve was employed to investigate the diagnostic potential of circRNA biomarker(s) by the GraphPad Prism v.8 software.A value of p < 0.05 was considered statistically significant.

Clinicopathological features
The median age of participants was 49 years (range, 12-75 years).The gender ratio of male to female patients was 5:1.The average ages at the time of diagnosis were 51.88 ± 15.87 and 55.8 ± 12.25 years for male and female patients, respectively.Most tumors were found in the frontal and temporal lobes of the brain, in two cases the tumor was located in the brain stem, which were diagnosed at a young age.No family history of brain cancer or other types of cancer has been found in patients.Using bioinformatic analysis of raw data from the GEO dataset, we investigated the expression levels of miRNAs in GBM tissues compared to normal brain tissues (Supplementary Table 1).According to hypothesis of this study, we focused on miR-145 analysis.The results indicated that the expression of miR-145-5p was significantly lower in glioblastoma tissues than in normal tissues (log FC: − 1.25, adj p value = 0.0342) (Fig. 1).

miRNA-target interactions
The interaction between hsa-miR-145-5p and its target genes, which was performed using miRNet, represented as a network.The resulting miRNA-target interactions showed that 1602 circRNAs and 238 genes interact with hsa-miR-145-5p.The genes investigated in this study were indicated as bigger nodes (Fig. 2).We hypothesized that circZNF609 and circNFIX regulates EGFR through sponging miR-145-5p.The interaction sites between circ-ZNF609 (circ0000615) and circNFIX (circ0049658) with miR-145-5p and also binding sites between miR-145-5p and EGFR are illustrated in Fig. 3A-C), respectively.

Evaluating the expression differences of cZNF609, cNFIX, miR-145 and EGFR between glioblastomas and normal tissues
CircZNF609 (hsa_circ_0000615) is derived from the circulation of the second exon of ZNF609.This back splice junction site was verified by the Sanger sequencing (Fig. 4A).CircNFIX (hsa_circ_0049658) is reverse spliced from the mRNA exons 3-8 of the host gene nuclear factor IX (NFIX) (Fig. 4B).Expression analysis showed a significant upregulation of circZNF609 in GBM samples compared to normal brain tissue (p = 0.001) (Fig. 5A), whereas circNFIX gene expression was increased in tumor samples compared to normal samples, but not significantly (p = 0.06) (Fig. 5B).We also found downregulation of miR-145-5p (p = 0.0065) (Fig. 5C) and upregulation of EGFR (p = 0.002) in GBM samples compared to normal tissue (Fig. 5D).

Receiver operating characteristic (ROC) curve analysis
The possible clinical utility of cNFIX and cZNF609 as a distinct GBM biomarker was evaluated by ROC curve analysis (Fig. 8A, B).The ROC curve analysis for cZNF609 indicates that this biomarker was well able to discriminate between tumoral and non-tumoral tissues due to its high level of AUC, sensitivity, and specificity (p = 0.001) (Fig. 8B).

Association between expression of circZNF609, circNFIX, miR-145-5p and EGFR and clinicopathological parameters
We assessed the association of circZNF609, circNFIX, miR-145-5p and EGFR expression levels in tumoral vs normal brain tissues with clinicopathological parameters.There was no significant difference in age, gender and tumor size among the groups (p > 0.05).

Discussion
Glioblastoma is the most aggressive diffuse glioma of astrocytic lineage and is considered a grade IV glioma based on 2021 WHO classification of CNS tumors 22 .The current standard treatment for glioblastoma involves surgery, chemotherapy and, radiotherapy 23 .However, as the average overall survival of patients is low, new diagnostic and treatment approaches are needed.Due to the rapid progression, even with aggressive multimodal therapy, glioblastoma remains almost incurable 24 .It is known that one of the most important cell signaling pathways involved in the development of glioblastoma is epidermal growth factor receptor.The EGFR signaling cascade activates several intracellular signaling pathways, including the mitogen-activated protein kinase (MAPK), phosphoinositide 3-kinase (PI3K), signal transducer and activator of transcription 3 (STAT3) pathways, and thus can cause proliferation and tumor growth, angiogenesis, and alteration of the tumor microenvironment 25 .Existing evidence has demonstrated that EGFR overexpressed in ~ 60% of primary GBM patients.Therefore, a better understanding of the EGFR signaling network and its interactions with other pathways is essential to elucidate the mechanisms of resistance and to develop better therapeutic agents 24,25 .Today, cancer treatment has moved towards pathway-based biomarkers therapies 26 .The research has provided new insight to explore the molecular mechanisms involved in glioblastoma 3 .Many studies revealed that noncoding RNAs such as circRNAs, lncR-NAs and microRNAs could regulate many important pathological processes including initiation, progression, metastasis, and drug resistance.Several lines of evidence indicate the role of circular RNA molecules in different types of cancers, including glioma 27 .However, the molecular mechanism of circRNAs remains largely unknown.circRNAs are endogenous single-stranded RNAs with covalently closed-loop structures and are highly stable, conserved, and abundantly expressed in various tissues 28 .Circ-ZNF609 (circBase ID: hsa_circ_0000615), with a full length of 874 bp, located at chr15: 64791491-64792365, is derived from the circulation of the second exon of ZNF609 29 .Circ-NFIX (hsa_circ_0049658) located in chr 19 (13183860-13192669), backspliced by exons 3-8 and 695 bp in length 30 .These circRNAs have highly conserved and fundamental biological functions among different species, suggesting that they play a critical role in cellular processes or organismal development 31,32 .Circ-ZNF609 and circ-NFIX have been identified as circular RNAs expressed in neural tissue and may therefore play an important role in neural functions or processes 28 .Recently, several studies have reported dysregulation of cZNF609 and cNFIX in various tumors and diseases 7,18,30,33,34 .In breast cancer, circZNF609 plays an oncogenic role by elevating p70S6K1 expression via sponging miR-145-5p 35 .According to the study by Pan et al., circNFIX attenuated pressure overload-induced cardiac hypertrophy by regulating miR-145-5p/ATF3 axis 36 .Thus, MiR-145-5p might be sponged by cZNF609 and cNFIX, and its expression was regulated by these circRNAs.Previous studies have shown that the expression of miR-145-5p is decreased in GBM patients, suggesting it can be a promising diagnostic and prognostic biomarker for this condition 37 .Another study reported that circZNF609 promotes glioma cell migration and proliferation via the miR-134-5p/BTG-2 axis 38 .In laryngeal squamous cell carcinoma (LSCC), elevated level of circZNF609 has been associated with poor patient survival.Moreover, the functional investigation indicated that circZNF609 regulated LSCC cell proliferation and invasion by sponging miR-342-3p and upregulating EGFR 8 .Xiao et al. 20 demonstrated that circNFIX regulates the HCC progression and glutaminolysis by targeting miR-3064-5p/HMGA2 axis.Xu et al., discovered through bioinformatics and experimental analysis that circNFIX promotes glioma progression by regulating miR-34a-5p via Notch signaling pathway.Several miRNAs, including miR-145-5p, were also reported to have binding sites with circNFIX in this study 7 .Therefore, aberrant expression of these circRNAs plays an important role in tumor proliferation, cell cycle regulation, invasion, and metastasis by modulating complementary miRNAs or target mRNAs linked to cancer-related signaling pathways [39][40][41] .These findings suggest that circZNF609 and circNFIX may act as predictive biomarkers for cancer prognosis and as promising targets for cancer therapy.However, the specific role of circ-ZNF609 in glioblastoma remains largely unexplored.
In this study, we first predicted the circZNF609, circNFIX/miR-145-5p/EGFR axis by extracting data from the literature and bioinformatics analysis using databases and web tools.Then, we evaluated the expression of each gene of this axis in 30 glioblastoma tissues and 21 normal brain tissues.
The results demonstrated that circZNF609 expression was significantly up-regulated in tumor tissues (p = 0.001).CircNFIX gene expression was not significant in tumor versus normal samples (p = 0.06).In comparison, a study by Ding et al., performed on 65 glioma samples and 15 normal samples, showed that the expression level of circNFIX was significantly upregulated in glioma tissues compared to the normal control group, and there was a significant difference between the low and high grade groups.Furthermore, knockdown of circNFIX suppressed glioma progression in vitro and in vivo, possibly by regulating the miR-378e/RPN2 axis as a ceRNA 42 .This finding supports the previous studies indicating an oncogenic mechanism for cZNF609 and cNFIX in different types of tumors 6,18,20,30 .Also, miR-145-5p and EGFR expression levels in glioblastoma tumors compared with normal brain tissues were decreased and increased, respectively (p = 0.0065, p = 0.002), while there was no statistically significant difference between expression levels of circZNF609, circNFIX, miR-145-5p, and EGFR genes with clinicopathological characteristics of patients, including age, gender and tumor size (p > 0.05).To use the relation of expression of these genes in clinic, further studies with more samples of different types of glioma are necessary.According to the ROC curve analysis, we found that circZNF609 transcript levels had more than 70% specificity in this regard, indicating that this circRNA can serve as good predictive biomarker to distinguish malignant from non-malignant tissues.Therefore, in samples suspicious for this disease, assessing the expression level of cZNF609 can be useful in evaluating the malignancy of the tissue.In contrast, the AUC of cNFIX was 0.636 with 53.3% sensitivity and 61.9% specificity in GBM patients (p = 0.1038) (Fig. 8A).Thus, it is not a good biomarker to discriminate tumor from normal tissues.
We also found an inverse correlation between EGFR and miR-145-5p expression levels (r = − 0.359 and p = 0.009).By analyzing miR-145 expression levels in GBM patients in this present study, 16 out of 30 patients showed ≥ two-fold reduction of miR-145 expression, which was associated with upregulation of the EGFR expression.
Based on bioinformatics analysis, EGFR was identified as a putative target of miR-145-5p (Fig. 2).Three binding sites for this microRNA were predicted in the EGFR gene using miRTarBase (Fig. 3C).In addition, there is strong evidence of interaction between the EGFR and miR-145 in glioma.Lu et al. showed that restoration of miR-145 in glioma cells significantly reduced in vitro proliferation, migration and invasion.Overexpression of miR-145 also reduced the expression of ADAM17 and EGFR 14 .Based on these findings, we hypothesize that miR-145 may modulate EGFR expression.However, we speculate that there may be other target genes that competitively interact with miR-145.These need further investigation.
The present study is limited by the relatively small number of patients.Further studies with a larger group of patients and an approximately equal number of male and female participants are needed to determine reliable prognostic value.

Conclusion
Our study demonstrated that circZNF609 and circNFIX may play a role in the pathogenesis of glioblastoma through the miR-145/EGFR axis, which requires functional studies.Further in vitro and in vivo model studies are needed to better understand the mechanism of this axis.Identifying molecular dysregulations in the EGFR signaling pathway may pave the way for potential therapeutic interventions and improve medical diagnosis.

Figure 1 .
Figure 1.Volcano diagram of differentially expressed miRNAs in glioblastoma.|log2FC|> 1 and adj p-value < 0.05 were considered as cut-off values for the identification of DE-miRNAs.Blue and red dots denote the downregulated and upregulated miRNAs in glioblastoma samples; grey dots denote miRNAs without differential expression between GBM and normal samples.The arrow represents the miR-145-5p.

Figure 2 .
Figure 2. Predicted target genes of miR-145-5p.miRNet was used for establishing a network of miR-145-5p and their target genes.Blue rectangle represents has-mir-145-5p; purple and orange dots represent target genes and circRNAs; the dots shown with two colors represent circRNA-gene.

Figure 7 .
Figure 7. Grouped graph of expression correlation of cZNF609, cNFIX, miR-145-5p and EGFR in GBM and normal tissues as calculated by Pearson correlation.

Figure 8 .
Figure 8. ROC curve analysis of (A) circNFIX and (B) circZNF609 expression for discrimination of GBM tumors from normal tissues.

Table 1 .
Clinicopathological features of cases.