circHUWE1 Exerts an Oncogenic Role in Inducing DDP-Resistant NSCLC Progression Depending on the Regulation of miR-34a-5p/TNFAIP8

Background Circular RNAs (circRNAs) are reported as competing endogenous RNAs (ceRNAs) and play key roles in non-small-cell lung cancer (NSCLC) progression. Thus, this study was aimed at clarifying underlying molecular mechanisms of circHUWE1 in NSCLC. Methods The quantitative real-time polymerase chain reaction (RT-qPCR) and western blot analyses were used for examining circHUWE1, microRNA-34a-5p (miR-34a-5p), and tumor necrosis factor alpha-induced protein 8 (TNFAIP8). IC50 of cisplatin (DDP) in A549/DDP and H1299/DDP cells and cell viability were analyzed by the Cell Counting Kit 8 (CCK-8) assay. Colony forming assay was performed to assess colony forming ability. Cell apoptosis and cell cycle distribution were determined by flow cytometry. Migrated and invaded cell numbers were examined by transwell assay. The association among circHUWE1, miR-34a-5p, and TNFAIP8 was analyzed by dual-luciferase reporter and RNA immunoprecipitation assays. A xenograft experiment was applied to clarify the functional role of circHUWE1 in vivo. Results circHUWE1 was upregulated in NSCLC tissues and cells, especially in DDP-resistant groups. circHUWE1 downregulation inhibited DDP resistance, proliferation, migration, and invasion while it induced apoptosis and cell cycle arrest of DDP-resistant NSCLC cells, which was overturned by silencing of miR-34a-5p. TNFAIP8 was a functional gene of miR-34a-5p, and the suppressive effects of miR-34a-5p overexpression on DDP-resistant NSCLC progression were dependent on the suppression of TNFAIP8. circHUWE1 inhibition also delayed tumor growth of DDP-resistant NSCLC cells. Conclusion circHUWE1 functioned as a promoter in DDP-resistant NSCLC by interaction with miR-34a-5p-TNFAIP8 networks, providing novel insight into DDP-resistant NSCLC diagnosis and treatment.


Introduction
Non-small-cell lung cancer (NSCLC) is a most common pulmonary subtype and is a major reason of tumor-caused death all over the world [1,2], accounting 80% proportion of primary lung cancer [3]. Although a considerable amount of research has been done to investigate lung cancer treatment, the mortality rate of NSCLC patients is increasing [4]. Except for surgery, chemotherapy, targeted therapy, and immunotherapy have been developed for cancer pharmacological treatment [5,6]; furthermore, novel promising anticancer approaches including cell therapy have been evolved in preclinical studies [7]. However, innate or acquired pharmacological resistance represents important obstacles to overcome. As well-documented, noncoding RNAs including circular RNAs (circRNAs) and microRNAs (miRNAs) are one of the pharmacological resistance mechanisms in tumor cells [8,9]. Cisplatin (DDP) is the backbone of first-line chemotherapy for NSCLC [10]. Therefore, exploring the pathogenesis mechanism of DDP resistance is important for NSCLC physiotherapy.
circRNAs are the endogenous transcripts with a covalently closed continuous loops that form via linking the end of a 3′ exon with an upstream 5′ exon [11,12]. Indeed, recent evidence supported that circRNAs acted as a critical regulator in the development of tumorigenesis [13]. For instance, circ_0000003 could regulate NSCLC cell malignant behaviors through mediating miR-338-3p [14]. circHUWE1 (hsa_circ_0004396) is coded by HUWE1 gene and located  International Journal of Genomics on chrX (53641494-53641706). A previous microarray profiles suggested a series of highly expressed circRNAs, including circHUWE1 [15]. We hypothesized that circHUWE1 played a carcinogenic role in NSCLC tumorigenesis. miRNAs, single strand and noncoding RNAs, could modulate almost all biological processes [16]. Mechanically, miRNAs could target the 3 ′ untranslated region (3 ′ UTR) of target mRNAs, thereby participating regulation of corresponding mRNAs expression [17]. miR-34a-5p dysfunction has been revealed in different cancers, and miR-34a-5p could be targeted by circRNAs to act as a tumor suppressor role [18][19][20]. Nevertheless, little is known on the relevance of circHUWE1 and miR-34a-5p in NSCLC.
Tumor necrosis factor alpha-induced protein 8 (TNFAIP8) is a member of a founding member of the TIPE family [21]. The upregulation of TNFAIP8 was found in a variety of tumors, which was highly linked to antiapoptotic and oncogenic gene expression [22,23]. In addition, TNFAIP8 also played a significant role in drug resistance of tumor cells [24]. Therefore, the underlying regulatory function of TNFAIP8 involved NSCLC was investigated.

RNase R and Actinomycin D Treatment.
For RNase experiments, 10 μg of total RNA was treated with 3 U/mg RNase R (Invitrogen) to degrade linear RNAs for 30 min at 37°C. In addition, 2 mg/mL of actinomycin D was purchased from Sigma-Aldrich (Merck KGaA, Darmstadt, Germany) and then was used to assess the stability of RNAs.
2.6. Cell Proliferation. A total of 3000 NSCLC cells were added to each well of 96-well plates with 0.2 mL of culture media and then cultured at 37°C in a humidified atmosphere with 5% CO 2 . After that, cells were interacted with Cell Counting Kit 8 (10 μL of CCK-8; Invitrogen) for 2 h. The optical density was assessed on a multiwell scanning spectrophotometer (BioTek, Winooski, VT, USA). IC 50 values of the DDP treatment were determined using SPSS software (version 19.0, IBM, Chicago, IL, USA). Moreover, colony forming was conducted as previous description [25].

Flow Cytometry
Assay. The single-cell suspension of NSCLC cells was prepared and fixed by 75% ethanol at 4°C for 12 h. The binding buffer containing Annexin Vfluorescein isothiocyanate and propidium iodide (BD Biosciences, San Diego, CA, USA) was used to stain cells in the dark at room temperature for 15 min, following by apoptosis assay under the flow cytometry (Becton Dickinson, San Jose, CA, USA). For cell cycle measurement, transfected NSCLC cells or control was collected and then fixed by 75% ethanol at 4°C for 12 h. After that, cells were incubated with 1 mL of PI/TritonX-100 staining solution containing RNase A for 30 min. Flow cytometry (Becton Dickinson) was used examine cell cycle, and ModFit software (Becton Dickinson) was used for data analysis.
2.10. RNA Immunoprecipitation (RIP) Assay. Approximately, 1 × 10 7 cells NSCLC cells were resuspended in RIP lysis buffer from Imprint® RNA immunoprecipitation kit (Sigma-Aldrich) at 4°C. Subsequently, the cell extract was interacted with magnetic beads embracing IgG or AGO2 antibody at 4°C for 4 h. After purifying by proteinase K treat-ment, then isolated immunoprecipitated RNA was subjected to RT-qPCR.

10
International Journal of Genomics were obtained from Vital River Laboratory (Beijing, China) and then assigned to four groups (6 mice per group). The A549/DDP cells (2 × 10 7 ) transfected with sh-circHUWE1 or control were implanted into the right back near the forelimb of BALB/c nude mice. Furthermore, BALB/c nude mice in the DDP groups were treated with 5 mg/kg of DDP every week by intraperitoneal injection after 7 d injection, with phosphate-buffered saline as control. The volume of xenograft tumor was calculated according to the equation: volume = 1/2 ðlength × width 2 Þ. The xenograft was resected for weight measurement and then stored for further research. For immunochemistry for Ki-67 and TNFAIP8, a study was conducted as previous description [26] using Ki-67 (ab231172; 1 : 100 dilution; Abcam) and TNFAIP8 immunochemistry kit (#Yaji Biological, Shanghai, China).
2.13. Statistical Analysis. RNA or protein expressed values of each sample were assessed by SPSS software and presented as the mean ± standard deviation. The differential significance was performed by Student's t-test or analysis of variance followed by Bonferroni's test, and P value less than 0.05 was considered statistically significant. Pearson's analysis was used to analyze the correlations.

Upregulation of circHUWE1 Was Related to DDP
Resistance. In this study, we will investigate the role of cir-cHUWE1 in NSCLC. As shown in Figure 1(a), circHUWE1 was a transcript of the HUWE1 gene exon 24. Convergent and divergent primer amplification showed that convergent primers amplified circHUWE1 products from both cDNA and gDNA, while divergent primers amplified circHUWE1 from cDNA only (Figure 1(b)). We also found that cir-cHUWE1 was upregulated in NSCLC tissues and cells relative to controls; importantly, circHUWE1 level was higher in DDP-resistant tissues and cells than DDP-sensitive tissues and cells, suggesting that circHUWE1 was related to DDP resistance (Figures 1(c) and 1(d)). Additionally, RT-qPCR revealed that circHUWE1 can resist RNase R degradation (Figures 1(e) and 1(f)). We also noticed that the half-life of cir-cHUWE1 was longer than the half-life of HUWE1 mRNA, and circHUWE1 was mainly expressed in the cytoplasm of

12
International Journal of Genomics A549/DDP and H1299/DDP cells but not the nucleus (Figures 1(g)-1(j)). The association of circHUWE1 and DDP resistance was investigated in NSCLC.

Depletion of circHUWE1 Could Suppress the Progression of DDP-Resistant NSCLC.
To systematically illustrate the role which circHUWE1 played in NSCLC, A549/DDP and H1299/ DDP cells were transfected with sh-circHUWE1. circHUWE1 was greatly decreased in A549/DDP and H1299/DDP cells after transfection with sh-circHUWE1 (Figure 2(a)). Moreover, the elevations of IC 50 in A549/DDP and H1299/DDP cells compared with A549 and H1299 cells were blocked by circHUWE1 inhibition (Figures 2(b)-2(e)). The results of colony forming and CCK-8 assays suggested that depletion of circHUWE1 could inhibit proliferation ability of A549/DDP and H1299/ DDP cells (Figures 2(f)-2(h)). The silencing of circHUWE1 also induced cell cycle progress arrest at S phase and cell apoptosis in A549/DDP and H1299/DDP cells, as confirmed by flow cytometry assay (Figures 2(i) and 2(j)). When sh-circHUWE1 was transfected in, migration and invasion inhibition were observed in A549/DDP and H1299/DDP cells (Figures 2(k)-2(n)). The protein levels of PCNA and MMP 13 were decreased while c-caspase 3 was increased in circHUWE1-silencing A549/DDP and H1299/DDP cells (Figures 2(o)-2(q)). Therefore, knockdown of circHUWE1 has the potency to suppress the progression of DDP-resistant NSCLC.
3.6. miR-34a-5p Could Bind to TNFAIP8 and Acted as a Tumor Suppresser in DDP-Resistant NSCLC Cells. To understand the mechanism by which miR-34a-5p suppressed NSCLC progression, we hypothesized that TNFAIP8 was a target of miR-34a-5p. The association between TNFAIP8 and miR-34a-5p was explored. TNFAIP8 was upregulated in A549/DDP and H1299/DDP cells after transfection with TNFAIP8 ( Figure 6(a)). Transfection with TNFAIP8 into 549/DDP and H1299/DDP cells could rescue the downregulation on of TNFAIP8 induced by miR-34a-5p (Figure 6(b)). IC 50 of DDP in A549/DDP and H1299/DDP cells was reduced by miR-34a-5p mimic but elevated by cotransfecting with TNFAIP8 (Figures 6(c)-6(f)). Not surprising, cell proliferation was inhibited by miR-34a-5p overexpression and rescued by transfecting with TNFAIP8 into A549/DDP and H1299/DDP cells (Figures 6(g)-6(i)). When transfecting with miR-34a-5p mimic, cell cycle was arrested while apoptosis was enhanced in A549/DDP and H1299/DDP cells, which was abrogated by overexpression of TNFAIP8 (Figures 6(j)-6(l)). Migration and invasion were inhibited in miR-34a-5p-induced A549/DDP and H1299/DDP cells while TNFAIP8 upregulation effectively reversed the suppressive effects on migration and invasion (Figures 6(m) and 6(n)). The miR-34a-5p upregulation decreased PCNA and MMP 13 while increased c-caspase 3 in A549/DDP and H1299/DDP cells, which was overturned by TNFAIP8 upregulation (Figures 6(o)-6(q)). These results suggested    , tumor growth and tumor weight were significantly inhibited in the sh-circHUWE1 group compared with the sh-NC group, and the inhibitory effects of circHUWE1 silencing on tumor growth were enhanced by treatment with DDP treatment. Besides, the expression of circHUWE1 and TNFAIP8 was decreased whereas that of miR-34a-5p was increased in removed tissues collected from the sh-circHUWE1 and sh-circHUWE1+DDP groups compared with matched controls (Figures 7(d)-7(f)). In addition, the immunochemistry suggested a significant decrease in both Ki-67 and TNFAIP8 expression in the circHUWE1 knockdown group compared to controls (Figure 7(g)). Taken together, circHUWE1 silencing suppressed tumor growth of DDP-resistant NSCLC in vivo.

Discussion
circRNAs had been declared to participate in drug resistance in NSCLC. For example, NSCLC cell-derived exosomes containing high circ-CPA4 level induced stemness and increased the innate DDP resistance in both A549 and H1299 cells through the let-7 miRNA/PD-L1 axis [27]. circSMARCA5 was originally identified in glioblastoma multiforme [28], and its overexpression was inclined to increase sensitivity to cisplatin in breast cancer MCF7 cells by interacting with host gene [29]. In terms of our research, circHUWE1 was found to be highly expressed in NSCLC and was related to DDP resistance. Function analysis experiments demonstrated that circHUWE1 downregulation inhibited DDP resistance and malignant phenotypes of DDP resistance-acquired NSCLC A549 and H1299 cells, which was dependent on regulating miR-34a-5p and TNFAIP8.
DDP treatment is a common way for tumor chemotherapy, such as lung cancer, but successful chemotherapy is still not satisfied following the development of multidrug resistance [30,31]. Thus, exploring the pathogenesis mechanism and DDP resistance was important to identify possible ther-apeutic strategy for NSCLC. According to a previous report, circHUWE1 was regarded as an oncogene in regulating the invasion and proliferation of colorectal cancer cells [32]. By the way, little literatures had been reported to describe a cir-cHUWE1 role in human cancers, except for colorectal cancer [32] and NSCLC (this study). Mechanically, the putative function of circRNA includes serving as miRNA sponges, and the cross-talk between circRNA and miRNAs is of great importance in the progression of tumorigenesis [33]; however, whether circHUWE1 functioned in NSCLC through sponging miRNAs remained unclear before this study. Here, we identified circHUWE1 as a key regulator of the progression of DDPresistant NSCLC by sponging miR-34a-5p. miR-34a-5p belongs to the miR-34 miRNA family, and most of them promote tumor development by exerting an antioncogenic role in tumorigenesis [34][35][36]. Supporting a tumor-inhibitory role for miR-34a-5p in NSCLC, luteolin exerted significant cancer inhibitory activity through upregulation of miR-34a-5p level in NSCLC [37]. Importantly, the antitumor roles of miR-34a-5p were implicated in DDP resistance [38,39]. Consistent with previous conclusions, we verified that miR-34a-5p acted as an inhibitor for DDP resistance in NSCLC by negatively regulating the expression of TNFAIP8.
Together, these findings identified a previously unknown regulatory axis circHUWE1/miR-34a-5p/TNFAIP8 in DDPresistant NSCLC progression, representing the first functional characterization of circHUWE1 and confirming it as a key driver for DDP resistance.
By the way, emerging role of circRNAs [40], long noncoding RNAs [41,42], and miRNAs [43] had been welldocumented in cancer development and progression. Moreover, serum extracellular vesicle-(EV-) derived circRNAs could be highly stable minimally invasive diagnostic biomarkers [44]. Nevertheless, this study did not further explore a circHUWE1 role in NSCLC A549 and H1299 cells and circHUWE1 expression in EV from patients' blood and cell culture medium, at least temporarily.

Conclusion
In summary, circHUWE1 downregulation inhibited DDP resistance, proliferation, migration, and invasion while it

18
International Journal of Genomics induced apoptosis and cell cycle arrest of DDP resistanceacquired NSCLC cells depending on the regulation of the miR-34a-5p/TNFAIP8 axis, supporting an important role of circHUWE1 in DDP-resistant NSCLC progression. This work provided a novel promising target for DDP-resistant NSCLC diagnosis.

Data Availability
No data were used to support this study.

Conflicts of Interest
The authors declare that they have no conflicts of interest.