miR-4293 Upregulates lncRNA WFDC21P by Directly Suppressing mRNA-Decapping Enzyme 2 to Promote Lung Carcinoma Proliferation

Background: Emerging evidence shows that lncRNA WFDC21P could promote STAT3 phosphorylation and microRNA 4293 SNP is associated with the risk of carcinomas, but the oncogenic functions of WFDC21P and miR-4293 in lung carcinoma are unclear. Methods: mRNA sequencing of lung carcinoma and control para-carcinoma tissues was performed to screen the potential targets. WFDC21P and miR-4293 levels were evaluated in lung carcinoma cells and tissues by qRT-PCR. The function of WFDC21P and miR-4293 on proliferation, apoptosis and metastasis were assessed by MTT, FACS, western blot, transwell assays, colony formation assays and xenografts experiment. RNA immunoprecipitation assays were implemented to verify the relationship between WFDC21P and mRNA-decapping enzyme 2 (DCP2). Furthermore, gain/loss of miR-4293 functions were used to determine its targeting relationship of DCP2. Results: WFDC21P expression is markedly enhanced in lung carcinoma tissue and cells. Moreover, WFDC21P promotes tumor cell proliferation and metastasis but suppresses apoptosis. Mechanistic investigations identied DCP2 can directly bind to WFDC21P and down-regulates its expression. DCP2 as a direct target of miR-4293 and its expression is suppressed by miR-4293. Consequently, miR-4293 can further promote WFDC21P expression by regulating DCP2. With positive correlation to WFDC21P expression, miR-4293 also plays oncogenic role in lung carcinoma. Furthermore, knockdown of WFDC21P results in functional attenuation of miR-4293 on tumor promotion. In vivo xenograft growth is also promoted by both WFDC21P and miR-4293. Conclusion: Our results establish oncogenic roles for both WFDC21P and miR-4293, and demonstrate that interactions between miRNAs and lncRNAs through DCP2 are important in lung carcinoma pathogenesis.


Background
Lung carcinoma remains a leading cause of cancer death globally. In spite of advances in early detection and standard treatment, lung cancer is often diagnosed at an advanced stage and has a poor prognosis. [1,2] The rst step in developing effective diagnosis and therapy is to elucidate the molecular mechanism operative in lung carcinoma. Recent large-scale genomic analyses of lung carcinoma identi ed many potential targets that can be exploited in diagnosis and therapies. [3,4] Non-coding RNAs (ncRNAs), such as long non-coding RNAs (lncRNAs) and microRNAs (miRNAs), are regulators of intracellular and intercellular signaling in lung cancer and modulate cell signaling to control diverse cellular processes, including proliferation, migration and apoptosis. Compared to miRNAs, lncRNAs exhibit more wide range of size and functions. [5] In tumorigenesis of lung carcinoma, some lncRNAs serve as tumor suppressive roles such as MEG3 and GAS5, while some are oncogenes like HOTAIR and MALAT1. [6][7][8][9] As important regulators in gene expression network, lncRNA can modulate mRNA stability, translation and post-translation in cytoplasm. In nucleus, lncRNAs are capable of controlling nuclear architecture and transcription. [10] Among the versatile functions of lncRNA, the crosstalk between lncRNA and miRNA is probably the most reported. As the competing endogenous RNA, lncRNAs can act as miRNA decoys to modulate gene expression. [11][12][13] Indeed, miRNAs can also directly or indirectly affect the expression of lncRNAs. [14] In the process of miRNA-mediated gene silencing, miRNA-induced silencing complexes can recruit the cellular mRNA decay machinery to degrade mRNAs and lncRNAs in which decapping complexes are important factors. [15,16] Components of decapping complexes, including DCP1 and DCP2, can remove 7-methylguanosine (m7G) cap to initiate mRNAs decay. [17][18][19] Similarly, decapping complexes also modulate the expression of lncRNAs through RNA degradation, which serves a vital role in transcriptional regulation speci cally at inducible genes. [20] The role of lncRNA degradation in the pathogenesis of carcinoma has not been clari ed.
WFDC21P is a pseudogene, belonging to WAP four-disul de core domain family which has been reported playing important roles in pathogenesis of lung carcinoma. [21,22] Moreover, WFDC21P was identi ed as a positive regulator of dendritic cell differentiation by promoting STAT3 phosphorylation.
[23] However, the function of WFDC21P in lung carcinoma is still unknown. In this study, we presented evidence that WFDC21P could be degraded by directly binding to DCP2, and that WFDC21P is highly expressed in lung carcinoma tissue and promotes cell proliferation and migration, but inhibits apoptosis. More importantly, we revealed that miR-4293 can elevate the expression of WFDC21P by regulating DCP2. Here, we Apoptosis Assays Apoptotic cells were detected using an Annexin V-FITC/PI kit (KeyGEN Biotech. Co. Ltd., Nanjing, China) according to the manufacturer′s instructions. A total of 1 × 10 4 cells was collected and analyzed by ow cytometry as previous described [24,25].

Transwell Migration Assay
Transwell migration assays were performed using Corning Costar Transwell chambers with 8 µm pore size membranes (Sigma, St Louis, MO, USA). After transfection, cancer cells incubated in 500 µL serum free medium were seeded into the upper chamber. The lower chamber was lled with 600 µL 1640 medium supplemented with 10% calf serum. After 24 hours, cells on the upper surface of the membrane were removed with cotton swabs. Lower chamber cells were stained with 1% crystal violet (Sigma) in 2% ethanol for 20 min. Excess crystal violet was removed by quickly immersing the insert in ddH 2 O for 3-4 s.
Stained cells were counted under a microscope (DM6000B, Leica). The assay was repeated three times for each group.

Colony Formation
After transfection in 6 wells plates, cancer cells were incubated for 48 hours. Then, transfected cells were harvested, and 1.5 × 10 3 cells per treatment group were seeded into 6 cm plates for colony formation. On the 12th day, colonies were stained with 0.25% crystal violet. Colonies were countered and reported as the number ± SEM. RNA immunoprecipitation (RIP) experiments were performed using a Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Merk millipore, Darmstadt, Germany) according to the manufacturer′s instructions. Antibodies for RIP assays against DCP2 were purchased from Abcam-Trading Company (Shanghai, China).

Continuously Monitor Cell Migration
To continuously monitor cancer cell migration, 1 × 10 4 cells per well were seeded into the top chamber of a CIM plate, which was then incubated on the RTCA station (xCELLigence System, Roche, Mannheim, Germany). Changes in cell migration were monitored in real-time.

Xenografts in Mice
A549 cells stably expressed WFDC21P, or untreated A549 cells (5 × 10 6 ) in 0.1 mL PBS, transfected with miR-4293, WFDC21P-small interfering RNAs (siRNAs) (Supplemental Table 2), WFDC21P were transplanted subcutaneously into the right or left anks of 5-6-week-old male BALB/c-nu/nu mice weighing 18-20 g (Charles River. Beijing, China). The maximum and minimum tumor diameters were measured every 3 d. The second time of injection the xenografts with siRNA or miRNA (Supplemental Table 2) was performed at the 14 day after rst time treatment. The mice were house in speci c-pathogen free (SPF) environment. Tumor volume (mm 3 ) = A x B 2 /2, where A and B are the maximum and minimum tumor diameters, respectively. After 4 weeks, nu mice were given euthanasia by amobarbital injection of 3 times standard does. The tumors were anatomized and weighed. All animal experiments were approved by the Committee on the Ethics of Animal Experiments of Binzhou Medical University and performed in accordance with the National Institutes of Health guide for the care and use of Laboratory animals. SPSS 22.0 software (IBM Corp., Armonk, NY, USA) was used to analyze statistical signi cance. Normally distributed data were expressed as mean ± SD, and Student's t test was used to compare two averages and ANOVA was used for mean comparison of multiple groups. If the assumption of homogeneity of variance is accepted or rejected, LSD test or Games-Howell test is used to compare the mean of different samples. Abnormally distributed data were expressed as median (interquartile range), the Mann-Whitney U test was used to compare two groups, and Kruskal-Wallis H test were used to compare multiple groups. Pearson's correlation was used to analyze the association between different variables. p < 0.05 is considered statistically signi cant differences.

WFDC21P Increases in Lung Carcinoma Cells and Tissues
Carcinoma tissue lysates prepared from 3 NSCLC patients and controls from their corresponding paracarcinoma tissues were subjected to RNA-seq analysis to detect mRNA expression. Mining of the RNAseq dataset identi ed that the expression of 1392 mRNAs were elevated, and 1303 were down-regulated (fold change > 2.0, p < 0.05) (Fig. 1A). Accumulated evidence indicated that WFDC family played great roles in the progression of carcinoma.[28-31] Additionally, some of them has been identi ed as biomarkers of diagnosis of lung cancer. [32][33][34] Here, the members of WFDC family were selected to do further comprehensive analysis of their roles in lung cancer. Our results demonstrated that WFDC1 was attenuated while WFDC2 was signi cantly elevated in carcinoma tissue. As a member of WFDC family, we also found that lncRNA-WFDC21P may have enhanced expression in carcinoma tissue (Fig. 1B). Moreover, the role of WFDC21P in tumorigenesis of lung cancer is unclear till now. Therefore, we collected more NSCLC sample tissues and corresponding non-tumor tissues to further investigate the WFDC21P expression in lung cancer. Compared with lung para-carcinoma samples, WFDC21P levels were elevated in lung carcinoma samples (Fig. 1C). Additionally, we also found that the expression of WFDC21P in cancer cell lines (A549, H1299 and H1975) was markedly elevated compared with human bronchial epithelial cells (HBE) (Fig. 1D). These results indicate that WFDC21P may play signi cant roles in lung carcinoma pathogenesis.

WFDC21P Enhances Cell Proliferation and Migration
In order to examine the effects of WFDC21P on cell proliferation, migration and apoptosis, overexpression or interference of WFDC21P was performed in A549 or H1975 cells ( Fig. 2A and C). It was found that WFDC21P overexpression markedly enhanced A549 or H1975 cell growth ( Fig. 2B and S1A), while transfection with an effective si-WFDC21P inhibited cell proliferation ( Fig. 2D and S1B). The role of WFDC21P in apoptosis was also examined, and WFDC21P knockdown increased apoptosis ( Fig. 2E and S1C). WFDC21P can promote STAT3 phosphorylation in dendritic cells.
[23] To further investigate whether the role WFDC21P in cell apoptosis is related to STAT3 signals, the expression of apoptotic factors related to STAT3 was examined. The expression of STAT3 and anti-apoptosis factors (Bcl2 and BIRC3) was signi cantly attenuated in si-WFDC21P-treated cells, but was elevated by WFDC21P overexpression. WFDC21P knockdown resulted in Bax elevation, while WFDC21P overexpression decreased Bax levels. Additionally, WFDC21P knockdown resulted in higher levels of cleaved caspases (caspase-3, 8 and 9). As expected, WFDC21P overexpression suppressed cleavage of caspases. The effects of WFDC21P on the examined apoptosis factors in A549 or H1975 cells may be owned to the enhanced phosphorylation of STAT3 resulted from WFDC21P elevation ( Fig. 2F and S1D). Therefore, we next examined the effects of WFDC21P on tumor cell migration. WFDC21P overexpression increased the number of cells penetrating the Transwell membrane. Conversely, WFDC21P knockdown attenuated the cell penetration ( Fig. 2G-J and S1E-H). In the colony formation assay, overexpression of WFDC21P resulted in an elevation of colony number. As expected, WFDC21P knockdown decreased the number of colonies (Fig. 2K-N and S1I-L).
These results indicate that WFDC21P is proto-oncogene in lung carcinoma.

WFDC21P Interacts with, and is Regulated by DCP2
Although WFDC21P has the function of promoting carcinoma, the detail mechanisms were still unknown. DCP2, a key component for degrading RNA, is involved in the pathogenesis of some carcinoma. [35] We assumed that WFDC21P is regulated by DCP2, and is involved in the pathogenesis of lung cancer. To investigate this hypothesis, we rst screened for an effective siRNA against DCP2 (Fig. 3A). Then, we detected the role of DCP2 in regulating WFDC21P expression. As expected, pcDNA-WFDC21P transfection elevated WFDC21P levels compared with the scrambled control. After DCP2 knockdown, irrespective of whether WFDC21P was overexpressed or not, WFDC21P was signi cantly elevated compared with the scrambled control. More importantly, DCP2 knockdown combined with pcDNA-WFDC21P transfection resulted in further increased WFDC21P expression compared with control treatment (pcDNA-WFDC21P + Scram) (Fig. 3B). As a regulated target of WFDC21P, STAT3 phosphorylation was also used to evaluate the effects of DCP2 on WFDC21P. As expected, WFDC21P overexpression after attenuating DCP2 by siRNA could further enhance STAT3 phosphorylation compared with scrambled control treatment. STAT3 phosphorylation was also higher in cells co-transfected with si-DCP2 and pcDNA-WFDC21P compared with control treatment (Fig. 3C). DCP2 can bind to RNA and promote its degradation by decapping.
Therefore, RNA immunoprecipitation (RIP) was performed to check the direct interaction between WFDC21P and DCP2. As a result, WFDC21P was signi cantly enriched by DCP2 immunoprecipitation compared with control IgG precipitation, indicating that WFDC21P directly interacts with DCP2 (Fig. 3D).
These results indicate that DCP2 may promote WFDC21P degradation by directly interacting with WFDC21P.

DCP2 is a Direct Target of miR-4293
Furthermore, we predicted that DCP2 mRNA might be a target of miR-4293 by using miRNA TargetScan (http://www.targetscan.org/vert_72/) (Fig. 3E). A GFP reporter plasmid containing DCP2-3′UTR was then constructed and used to co-transfect cells with miR-4293. miR-4293 transfection resulted in considerable attenuation of both GFP intensity and the percentage of GFP-positive cells. Cells were then co-transfected with mu-4293 (containing a mutated site of DCP2 mRNA binding) and the DCP2-3′-UTR GFP reporter plasmid. mu-4293 could not reduce the percentage of GFP positive cells compared with the wide type miR-4293. To further con rm that miR-4293 binds to the DCP2-3′UTR sequence, we constructed a mu-DCP2-3′UTR GFP reporter plasmid (with a mutant binding site for miR-4293) in which we deleted the "CTG" sequence from the DCP2-3′UTR GFP reporter plasmid. As a result, cells, co-transfected with mu-DCP2-3′UTR and miR-4293, had no signi cant attenuation in the percentage of GFP-positive cells ( Figure   S2). Next, we transfected cancer cells (A549 and H1975) with miR-4293 and its inhibitor (ASO-4293). As expected, miR-4293 transfection signi cantly attenuated DCP2 expression compared to transfection of both ASO-4293 and scrambled control both in A549 and H1299 cells (Fig. 3F). These results indicate that miR-4293 can directly bind to the DCP2-3′UTR to regulate the expression of DCP2. miR-4293 Increases WFDC21P and STAT3 Levels through Regulating DCP2 The above results indicate that WFDC21P is degraded by interaction with DCP2 and DCP2 expression is suppressed by miR-4293. However, the effects of miR-4293 on WFDC21P were still to be elucidated. Firstly, we prepared miR-4293 (mimic), ASO-4293 (inhibitor) and mu-4293 (mutant). A549 cells were then transfected with miR-4293, ASO-4293 or mu-4293 respectively. After transfection, miR-4293 mimic could signi cantly elevate miR-4293 level, which could be markedly attenuated by ASO-4293 (Fig. 3G). By detecting WFDC21P level, we found that miR-4293 treatment could markedly enhance WFDC21P expression compared with the scrambled control. In contrast, ASO-4293 treatment suppressed WFDC21P expression compared with the scrambled control (Fig. 3H). In addition, miR-4293 treatment further increased WFDC21P levels (Fig. 3I). These results indicate that miR-4293 elevation can promote the expression of WFDC21P.
Next, STAT3 activation was used to evaluate the relationship between WFDC21P and miR-4293. As expected, overexpression of both WFDC21P and miR-4293 resulted in more STAT3 phosphorylation compared with only WFDC21P or miR-4293 treatment (Fig. 3J). To further investigate the role of WFDC21P in stimulating STAT3 pathway, we treated WFDC21P-overexpressed cells with AG490 (a speci c inhibitor to STAT3 signaling, 50 µmol/L), and found that AG490 treatment reduced WFDC21Pinduced STAT3 activation (Fig. 3K).
We also found that WFDC21P, by regulating STAT3 pathway, may have a positive feedback effect on the expression of miR-4293. STAT3 pathway was deactivated by AG490 treatment, which also could result in the decreased levels of miR-4293. Moreover, miR-4293 elevation resulting from WFDC21P overexpression was further alleviated by using AG490 to deactivate STAT3 pathway (Fig. 3L). These results indicate that there might be a feedback loop between miR-4293 and WFDC21P in which miR-4293 increases WFDC21P levels and WFDC21P might upregulate miR-4293 expression by activation of STAT3 pathways.

MiR-4293 and STAT3 Levels also Increased in NSCLC Cells and Tissues
The above results indicate that WFDC21P expression and STAT3 activation can be upregulated by miR-4293. However, the role of miR-4293 and STAT3 in the pathogenesis of NSCLC requires clari cation.
Firstly, we examined the expression of miR-4293 in carcinoma cell lines (A549, H1299 and H1975) and found that miR-4293 expression was signi cantly elevated compared to HBE (Fig. 4A). As expected, miR-4293 elevation was also found in NSCLC tissues compared with para-carcinoma tissues (Fig. 4B). As a direct target of WFDC21P, STAT3 phosphorylation was increased in lung carcinoma cells (Fig. 4C). Similarly, STAT3 expression was also increased in lung carcinoma tissue (Fig. 4D). Based on the above results, we further analyzed possible correlations among WFDC21P, STAT3 and miR-4293 in the NSCLC samples. We found that WFDC21P expression was positively correlated with STAT3 expression (p < 0.01, Fig. 4E). miR-4293 and WFDC21P expression was also in positive correlation (p = 0.015, Fig. 4F), as was the relationship between miR-4293 and STAT3 expression (p < 0.01, Fig. 4G).

MiR-4293 Promotes Cell Proliferation and Migration
In order to further study the role of miR-4293 in lung carcinoma, MTT assays were performed and showed that miR-4293 transfection could signi cantly promote A549 or H1975 cell proliferation compared with cells transfected with scrambled control or ASO-4293 inhibitor. ASO-4293 transfection suppressed cancer cell proliferation compared with the scrambled control ( Fig. 5A and S3A). The apoptosis assay showed that miR-4293 treatment decreased lung cancer cell (A549 or H1975) apoptosis compared with the scrambled control. More importantly, miR-4293 treatment signi cantly increased survival capacity compared with ASO-4293 treatment ( Fig. 5B and S3B).
Next, we investigated the effects of miR-4293 on apoptosis related factors. We found that the expression of anti-apoptosis factors (Bcl2 and BIRC3) could be elevated by miR-4293 transfection, but was attenuated by ASO-4293 treatment compared with scrambled control. However, the expression of Bax, a pro-apoptosis factor was suppressed by miR-4293 treatment and enhanced by ASO-4293 transfection. We also examined the expression of caspases in uenced by miR-4293 and ASO-4293. miR-4293 treatment resulted in less cleavage of caspase-3, caspase-8 and caspase-9, which play essential roles in apoptosis. Whereas, ASO-4293 treatment induced the cleavage of caspase-3, caspase-8 and caspase-9 compared with miR-4293. Importantly, the apoptotic factors examined are closely related to STAT3 signaling pathways; therefore, we detected STAT3 expression and found that miR-4293 treatment could signi cantly enhance STAT3 phosphorylation compared with scrambled control, in contrast, ASO-4293 treatment suppressed STAT3 phosphorylation compared with miR-4293 transfection ( Fig. 5C and S3C). To investigate the possible role of miR-4293 in tumor cell migration, we conducted Transwell migration assays, which demonstrated that A549 or H1975 cells transfected with miR-4293 have enhanced migration capacity compared with cells transfected with scrambled control ( Fig. 5D and E, S3D and E). Additionally, miR-4293 treatment could signi cantly increase survival of A549 or H1975 cells seeded sparsely in a clonogenic assay compared with the scrambled control. Conversely, cells transfected with ASO-4293 formed fewer colonies compared with scrambled control ( Fig. 5F and G, S3F and G). miR-4293 promoted cell proliferation and migration but inhibited apoptosis. These results indicate that miR-4293 has oncogenic potential in lung carcinoma.
Knockdown of WFDC21P Attenuates the Oncogenic Role of miR-4293 The above results show that miR-4293 exert an oncogenic role and increases WFDC21P expression, which indicates that miR-4293 may exert its oncogenic role by promoting WFDC21P expression. To better understand whether the role of miR-4293 can be attenuated by WFDC21P knockdown, we rst evaluated cell proliferation co-in uenced by miR-4293 and si-WFDC21P. WFDC21P knockdown suppressed cell proliferation and miR-4293 transfection promoted cell proliferation. More importantly, the increase in cell proliferation resulting from miR-4293 transfection was markedly alleviated by si-WFDC21P (Fig. 6A). The apoptosis assay showed that WFDC21P knockdown promoted apoptosis and also signi cantly abolished the increased cell viability resulting from miR-4293 transfection (Fig. 6B).
We then examined the expression of Bcl2, anit-apoptosis factors, which was enhanced by miR-4293 transfection and suppressed by WFDC21P knockdown compared with scrambled control. Moreover, Bcl2 elevation caused by miR-4293 treatment was attenuated by WFDC21P knockdown. Besides, miR-4293 transfection with WFDC21P knockdown resulted in lower expression of Bcl2 compared with miR-4293 transfection with scrambled control. The expression of Bax was elevated by WFDC21P knockdown but was suppressed by miR-4293 transfection compared with scrambled control. si-WFDC21P alleviated the downregulation of Bax expression affected by miR-4293. Accordingly, WFDC21P knockdown induced more caspase-3 cleaved than miR-4293 treatment. miR-4293 transfection attenuated the elevation of cleavage caspase-3 induced by WFDC21P knockdown. More importantly, we found that miR-4293 transfection enhanced STAT3 phosphorylation, which was alleviated by si-WFDC21P (Fig. 6C). RTCA station analysis was performed to continuously observe cells migration, which gradually increased in miR-4293-treated cells but was blocked by si-WFDC21P treatment ( Fig. 6D and E).
To further study whether the oncogenic role of miR-4293 can be attenuated by blocking WFDC21P, we investigated whether ASO-4293 inhibition of lung cancer cell proliferation can be restored by WFDC21P overexpression. We found that the proliferative capacity was suppressed by ASO-4293 inhibitor and enhanced by WFDC21P overexpression compared with control. More interestingly, the inhibition of cell proliferation by ASO-4293 was restored to a certain extent after WFDC21P overexpression (Fig. 6F).
Compared with control treatment, ASO-4293 transfection increased apoptosis. Even though WFDC21P overexpression could promote cell viability, the percentage of apoptotic cells was still elevated by ASO-4293 transfection (Fig. 6G). By examination of apoptosis related factors, we found that the expression of Bcl2 was decreased in ASO-4293-transfected cells, but increased in WFDC21P overexpression cells. ASO-4293 transfection with WFDC21P overexpressing resulted in attenuation of Bcl2 expression compared with control treatment. In contrast, Bax expression was signi cantly enhanced by ASO-4293 transfection. WFDC21P overexpression partially attenuated Bax expression induced by ASO-4293. As expected, WFDC21P overexpression decreased the cleavage of caspase-3 compared with scramble control. And attenuation of caspase-3 cleavage caused by WFDC21P overexpression was alleviated by ASO-4293 treatment. We also found that ASO-4293 resulted in less activated STAT3, while WFDC21P overexpression increased STAT3 activation. WFDC21P overexpression also partially increased STAT3 levels induced by ASO-4293 (Fig. 6H). As expected, ASO-4293 treatment could also alleviate migration capacity by WFDC21P overexpression (Fig. 6I-L). Together, these results indicate that oncogenic activity of miR-4293 can be attenuated by blocking WFDC21P expression.

WFDC21P and MiR-4293 Promote Tumorigenesis in vivo.
Both miR-4293 and WFDC21P have oncogenic activities in lung carcinoma cells. To further investigate the roles of miR-4293 and WFDC21P in vivo, we hypodermically injected A549 cells stably treated with plasmid, or miRNA, or siRNA, or controls into nude mice to produce xenografts. Every 3 days, we measured the volumes of the xenografts. One month after injection, the mice were sacri ced. Stably expression of WFDC21P resulted in enhanced tumor growth both in volumes and weights compared with pcDNA control (Fig. 7A-C). We examined expression of WFDC21P and STAT3 phosphorylation in the xenografts. As expected, WFDC21P overexpression in injected xenografts resulted in elevation of WFDC21P and STAT3 phosphorylation ( Figure S 4A and B).
We then evaluated the effects of miR-4293 or WFDC21P knockdown on xenograft growth. miR-4293 treatment enhanced growth in both volumes and weights of xenografts compared with scrambled control. In contrast, WFDC21P knockdown resulted in xenografts growth that was signi cantly lower than that produced by miR-4293 overexpression and also than scrambled control (Fig. 7D-F). Compared with scrambled control, miR-4293 treatment resulted in elevation of miR-4293, WFDC21P and STAT3 phosphorylation in xenografts. Conversely, these measures were attenuated by WFDC21P knockdown ( Figure S4C-E).
To better understand the cooperative action of miR-4293 and WFDC21P in tumor growth, we established xenografts with A549 cells co-treated with miR-4293 and scrambled control, miR-4293 and ASO-4293, as well as miR-4293 and siRNA-WFDC21P. As expected, xenografts with miR-4293 and scrambled cotreatment had the highest growth rate of these three groups. Interestingly, ASO-4293 and si-WFDC21P could abolish the tumor promoting effect of miR-4293 almost equally (Fig. 7G-I). Moreover, ASO-4293 attenuated the decrease in WFDC21P or STAT3 levels in xenografts by miR-4293 and si-WFDC21P ( Figure   S4F-H). Similarly, tumor promotion resulting from WFDC21P stable overexpression was also alleviated by ASO-4293 treatment (Fig. 7J-L). ASO-4293 treatment could also alleviate elevation of WFDC21P and STAT3 phosphorylation in xenografts resulting from WFDC21P overexpression in injected cells ( Figure   S4I and J). The above results support the ability of miR-4293 and WFDC21P to promote tumor growth in vivo.

Discussion
lncRNAs are involved in cell proliferation, metastasis and apoptosis, and can also play oncogenic or tumor suppressive roles. Here, we show that WFDC21P is signi cantly elevated in NSCLC tissues and promotes tumor cell proliferation and metastasis but suppresses apoptosis. DCP2 can directly bind to WFDC21P and promote its downregulation. More interestingly, DCP2 is found to be a target of miR-4293 and down-regulated by miR-4293. It has been reported STAT3 is a target gene of WFDC21P.
[23] Importantly, our results con rm that miR-4293 enhances the expression of WFDC21P and promotes STAT3 phosphorylation by a novel mechanism in directly regulating DCP2. In addition, WFDC21P elevates miR-4293 expression by promoting STAT3 activation. As a result of elevated STAT3 activity, apoptosis is suppressed and proliferation is enhanced through Bcl2 elevation, Bax2 attenuation and reduced caspase cleavage (Fig. 8). Together, these ndings provide evidence that miR-4293 plays an oncogenic role in NSCLC by suppressing DCP2-mediated WFDC21P degradation.
Accumulated studies have indicated that the prominent class of lncRNA share a common function in negatively impacting on miRNA-mediated gene regulation.[36, 37] Interestingly, we found that, in contrast to ceRNA, not only WFDC21P could be up-regulated by miR-4293, but miR-4293 expression could be enhanced by WFDC21P through positive feedback. Both WFDC21P and miR-4293 play oncogenic roles in NSCLC. To better understand the mechanism, a key step is to determine miR-4293 targets and we revealed DCP2 to be such a target. As a critical component of the decapping complex, DCP2 is responsible for RNA degradation. [38,39] By decapping a subset of mRNAs, DCP2 plays an important role in cancer pathogenesis by affecting processes such as cell migration and apoptosis. [40][41][42][43][44] More interestingly, DCP2 is regulated at the post-transcriptional level rather than being transcriptionally silenced, and miRNAs are an effective way to down-regulate DCP2. DCP2 mRNA has a fairly long 3′-UTR of 7.2 kb that may contain many potential miRNA binding sites. miR-141-3p/200a-3p may be involved in repression of DCP2 expression during renal development; therefore, miRNAs can not only directly suppress the expression of target genes by binding the 3′-UTR but can also regulate gene expression by involving DCP2 activity. As a novel nding, we observed that miR-4293 can bind to DCP2 transcripts and down-regulate their expression. Several studies indicate that DCP2 appears to target distinct transcripts and the reported targets are almost all mRNAs. [40,45,46] However, we observed that DCP2 can bind to WFDC21P, a long non-coding RNA, to participate in cancer pathogenesis. We further clari ed that WFDC21P degradation played important roles in the tumorigenesis of lung carcinoma, and showed that DCP2 directly regulated by miR-4293, down-regulated WFDC21P expression. The downregulation of lncRNAs mediated by DCP2 may become a novel underlying mechanism of lncRNAs in different kinds of carcinoma. However, DCP2 is not the only one component of the decapping enzyme [47], and the speci c underlying mechanism for DCP2 promoting WFDC21P degradation needs further study.
Although we did not con rm a direct interaction between WFDC21P and miR-4293, WFDC21P expression can still be elevated by miR-4293. More importantly, miR-4293 upregulated WFDC21P in both NSCLC cells and xenografts. In addition, the expression of miR-4293 was positively correlated with WFDC21P levels in lung carcinoma. Hence, the elevation of miR-4293 in NSCLC tissues might account for the increase in WFDC21P levels. Recent studies reveal that lncRNAs, as competing endogenous RNA, can bind to related miRNAs to further suppress the function of the miRNAs. [14,48] However, we found WFDC21P could elevate miR-4293 levels via a positive feedback loop. Furthermore, the oncogenic role of miR-4293 was blocked by suppressing WFDC21P. miRNAs have the potential to target hundreds of mRNAs because of imperfect complementarity when binding. Different targets result in different roles of miRNAs. [49][50][51] Although WFDC21P did not directly interact with miR-4293, we found that WFDC21P may play a major role in the function of miR-4293, and WFDC21P can be considered as a downstream regulator of the miR-4293 signaling pathway. In the present study, we observed that WFDC21P knockdown almost completely abolished the effects of miR-4293 on proliferation, apoptosis and metastasis. WFDC21P overexpression could only restore the effects of ASO-4293 suppression to a certain extent, possibly because of DCP2 activation. Transfection of ASO-4293 almost abolished suppression of DCP2 by miR-4293, and DCP2 activation may be greater with ASO-4293 treatment. Even though WFDC21P was overexpressed, its degradation was upregulated by activated DCP2.
STAT3 plays a critical role in many pathophysiological processes, including cancer cell proliferation, antiapoptosis and metastasis. [52,53] We found that STAT3 expression was not only elevated in lung carcinoma, but also positively correlated with the expression of both WFDC21P and miR-4293. In addition to enhancing STAT3 activation, WFDC21P possibly participates in the regulation of STAT3 signaling.[54, 55] As a regulator of WFDC21P, miR-4293 could further elevate STAT3 phosphorylation. Moreover, by regulating STAT3 activation, WFDC21P could elevate the expression of miR-4293. Hence, STAT3 may play a critical role in the proto-oncogenic capacity of WFDC21P and miR-4293. STAT3 deactivation may suppress the function of WFDC21P and miR-4293.

Conclusions
Many lncRNAs, as competing endogenous factors, have been reported to participate into cancer pathogenesis by suppressing the function of miRNAs. Differently, this study demonstrates a novel interaction pattern between lncRNA and miRNA, and we found that miR-4293 up-regulates the expression of WFDC21P through its degradation by directly mediating DCP2, which plays important roles in the tumorigenesis of lung carcinoma. Moreover, WFDC21P interference almost completely abolished the effects of miR-4293 on the proliferation of lung carcinoma cells, which indicates that WFDC21P knockdown might be an effective strategy for lung cancer therapy. The present study provided a valuable theoretical basis for the discovery of lung carcinoma therapeutic targets and diagnosis markers based on WFDC21P and miR-4293. the Medical Ethics Committee of Binzhou Medical University. Patients provided written informed consent and the study procedures were fully explained before study inclusion. The authors con rmed that we have obtained written consent from the patients to publish this manuscript.

Consent for publication
The authors con rmed that we have obtained written consent from the patients to publish this manuscript.

Availability of data and materials
All data generated during this study are included either in article or in the additional les.

Declaration of Interests
The authors declare that they have no con icts of interest concerning this article.

Funding
The present study was supported by the National Natural Science Foundation of China (No.81772281