A positive feedback loop between ZNF205‐AS1 and EGR4 promotes non‐small cell lung cancer growth

Abstract Accumulating evidences revealed that long noncoding RNAs (lncRNAs) are frequently implicated in non‐small cell lung cancer (NSCLC). Herein, we reported the identification of a novel NSCLC‐associated functional lncRNA ZNF205 antisense RNA 1 (ZNF205‐AS1). ZNF205‐AS1 was increased in NSCLC tissues and cell lines, and associated with poor prognosis of NSCLC patients. Bioinformatics prediction, combined with experimental verification revealed that early growth response 4 (EGR4) directly bound to ZNF205‐AS1 promoter, increased the promoter activity of ZNF205‐AS1, and activated ZNF205‐AS1 transcription. Intriguingly, ZNF205‐AS1 transcript directly interacted with EGR4 mRNA, increased EGR4 mRNA stability, and up‐regulated EGR4 expression via RNA‐RNA interaction. Thus, ZNF205‐AS1 and EGR4 formed a positive feedback loop. Through regulating EGR4, ZNF205‐AS1 activated its own promoter activity. EGR4 was also increased in NSCLC and the expression of ZNF205‐AS1 was significantly positively correlated with EGR4 in NSCLC tissues. Gain‐of‐function and loss‐of‐function assays demonstrated that both ZNF205‐AS1 and EGR4 promoted NSCLC cell growth in vitro and NSCLC tumour growth in vivo. Concurrently depleting ZNF205‐AS1 and EGR4 more significantly repressed NSCLC tumour growth in vivo. Collectively, our study demonstrated that the positive feedback loop between ZNF205‐AS1 and EGR4 promotes NSCLC growth, and implied that targeting this feedback loop may be promising therapeutic strategy for NSCLC.

RNAs (lncRNAs) have attracted increasing attentions. 6 LncRNA is a class of transcript with more than 200 nucleotides in length, but do not encode proteins. [7][8][9] Aberrant expressions of lncRNAs have been revealed in various pathological status, particular in cancers. [10][11][12][13] Furthermore, lncRNAs also play important roles in various pathophysiological processes, including cancers. [14][15][16][17][18] In NSCLC, several lncRNAs are reported to participate in tumourigenesis and/or development of NSCLC. LncRNA MALAT-1 is frequently revelated to regulate NSCLC metastasis. 19 LncRNA MEG3 is reported to regulate cisplatin resistance of NSCLC. 20 MetaLnc9 promotes NSCLC metastasis via activation of AKT/mTOR pathway. 21 LncRNA VELUCT regulates NSCLC cell viability. 22 Knockdown of LINC01614 inhibits NSCLC progression. 23 Compared with these limited number of lncRNAs reported to play roles in NSCLC, more lncRNAs are revealed to be aberrantly expressed in NSCLC. 24 Among these thousands of lncRNAs aberrantly expressed in NSCLC, many lncRNAs may also have critical roles, which need further investigation. Seiler et al performed a functional siRNA screen to search the lncRNAs regulating NSCLC cell viability. 22 Among the list of candidate targets, we noted ZNF205 antisense RNA 1 (ZNF205-AS1) (NCBI Reference Sequence: NR_024167.1), which has a relative high score in the screen, but whose roles in cancers are unknown.
In this study, we further investigated the expression pattern and biological roles of ZNF205-AS1 in NSCLC. We also explored the reason contributing to the aberrant expression of ZNF205-AS1 in NSCLC. Intriguingly, we identified a positive feedback loop between ZNF205-AS1 and transcription factor Early Growth Response 4 (EGR4) in NSCLC, which significantly promoted NSCLC growth.  Where indicated, NSCLC cells were treated with 50 µM α-amanitin (Sigma-Aldrich, Saint Louis, MO, USA) for 0-24 hours.

| Vectors construction and transfection
The complementary DNA (cDNA) encoding ZNF205-AS1 was PCR- The transfection and cotransfection of vectors were performed using Lipofectamine 3000 (Invitrogen) according to the protocol.

| Stable cell lines construction
To obtain ZNF205-AS1 or EGR4 stably overexpressed cells,

| Chromatin immunoprecipitation (ChIP) assay
ChIP assay was carried out in indicated NSCLC cells with the EZ- according to the protocols. Then, 3 µg of purified biotin-labelled RNAs were incubated with 1 mg of SPC-A1 whole-cell lysates at 25°C for 1 hour. Next, the complexes were isolated using streptavidin agarose beads (Invitrogen). The RNAs enriched in the pulldown material were detected by qPCR as above described.  Zeiss Photomicroscope (Carl Zeiss, Oberkochen, Germany) and quantified via counting at least ten random fields.

Cellular senescence of indicated NSCLC cells was evaluated using
Senescence-associated β-galactosidase (SA-β-gal) staining with the Senescence β-Galactosidase Staining Kit (Beyotime) in accordance with the protocol. The results were collected using the Zeiss Photomicroscope and quantified via counting at least 10 random fields.

| ZNF205-AS1 was increased in NSCLC and correlated with poor prognosis of NSCLC patients
To investigate the expression pattern of ZNF205-AS1 in NSCLC, we collected 90 pairs of NSCLC tissues and matched adjacent noncancerous lung tissues, and measured the expression of ZNF205-AS1 in these tissues using qPCR. As presented in Figure 1A, ZNF205-AS1 was significantly increased in NSCLC tissues compared with adjacent noncancerous lung tissues. Correlation regression analyses of the association between ZNF205-AS1 expression levels and clinicopathological characteristics of NSCLC patients displayed that increased ZNF205-AS1 expression levels were associated with poor pathological differentiation (P = 0.035), great tumour diameter (P = 0.049), lymph nodes metastasis (P = 0.033), and advanced TNM stage (P = 0.020) (Table 1). Furthermore, Kaplan-Meier survival analyses in these 90 NSCLC patients displayed that increased ZNF205-AS1 expression levels were associated with poor overall survival (P = 0.0008) ( Figure 1B). In addition, ZNF205-AS1 expression levels in normal bronchial epithelial cell line 16HBE and NSCLC cell lines PC-9, NCI-H1299, NCI-H23, SK-MES-1, and SPC-A1 was measured using qPCR. As presented in Figure 1C, ZNF205-AS1 was consistently increased in NSCLC cell lines compared with normal bronchial epithelial cell line. Collectively, these results demonstrated the increased expression of ZNF205-AS1 in NSCLC and the association between ZNF205-AS1 and poor prognosis of NSCLC patients.

ZNF205-AS1
To investigate the reasons contributing to the elevation of ZNF205-AS1 in NSCLC, we screened the promoter region of ZNF205-AS1 using JAS-PAR (http://jaspar.genereg.net/), 26 and predicted two EGR4 binding sites, locating at −15 and −218 upstream of the transcription start site of ZNF205-AS1 (Figure 2A). To investigate whether EGR4 binds to the predicted sites at ZNF205-AS1 promoter, ChIP assays were carried out with EGR4 specific antibody. As presented in Figure 2B, the ZNF205-AS1 promoter region containing the predicted binding sites was specific enriched by EGR4 specific antibody, whereas a distal region of ZNF205-AS1 promoter without the EGR4 binding sites was not enriched. To further investigate whether EGR4 modulates the transcriptional activity of ZNF205-AS1 via binding to ZNF205-AS1 promoter, dual luciferase reporter assays were carried out. The promoter of ZNF205-AS1 from −1025 to +47 bp upstream of the transcription start site was cloned into the pGL3-basic firefly luciferase reporter. The constructed or empty pGL3basic firefly luciferase reporter was cotransfected with EGR4 overexpression or empty plasmids into PC-9 cells. As presented in Figure 2C, ectopic expression of EGR4 increased luciferase activity of the constructed reporter. In addition, the constructed luciferase reporter was cotransfected with two independent EGR4 specific shRNAs into SPC-A1 cells. The results displayed that knockdown of EGR4 by two independent shRNAs both decreased luciferase activity of the constructed reporter ( Figure 2D). These results suggested that EGR4 activated the promoter activity of ZNF205-AS1. Next, the effects of EGR4 on the expression of ZNF205-AS1 was detected. We constructed EGR4 that stably overexpressed and control PC-9 cells via transfecting EGR4 overexpression and empty plasmids, respectively. The overexpression efficiency was verified using western blot ( Figure 2E). The expression of ZNF205-AS1 in EGR4 stably overexpressed and control PC-9 cells was measured using qPCR. As displayed in Figure 2F, ectopic expression of EGR4 up-regulated ZNF205-AS1. We also stably depleted EGR4 in SPC-A1 cells via transfecting two independent ZNF205-AS1 specific shRNAs.
The knockdown efficiency was verified using western blot ( Figure 2G).
The expression of ZNF205-AS1 in EGR4 stably depleted and control SPC-A1 cells was measured using qPCR. As displayed in Figure 2H, EGR4 knockdown down-regulated ZNF205-AS1. Collectively, these results suggested that EGR4 directly bound to the promoter of ZNF205-AS1 and activated the transcription of ZNF205-AS1.

| EGR4 was increased and positively correlated with ZNF205-AS1 in NSCLC tissues
To evaluate whether the regulation of ZNF205-AS1 by EGR4 exist in vivo, the expression of EGR4 in the same 90 pairs of NSCLC tissues and adjacent noncancerous lung tissues used in Figure 1A was measured using qPCR. Consistent with ZNF205-AS1, EGR4 was also increased in NSCLC tissues compared with adjacent noncancerous lung tissues ( Figure 2I). Next, the expression correlation between ZNF205-AS1 and EGR4 in these 90 NSCLC tissues was calculated.
As presented in Figure 2J, the expression of ZNF205-AS1 was significantly positively correlated with that of EGR4 in NSCLC tissues (r = 0.7193, P < 0.0001), supporting the positive modulation of ZNF205-AS1 by EGR4 in vivo.

RNA interaction
Due to the significant correlation between ZNF205-AS1 and EGR4 in NSCLC, we next investigated whether ZNF205-AS1 also modulate and control PC-9 cells via transfecting ZNF205-AS1 overexpression and empty plasmids, respectively. The overexpression efficiency was verified using qPCR ( Figure 3A). We also stably depleted ZNF205-AS1 in SPC-A1 cells via transfecting two independent ZNF205-AS1 specific shRNAs. The knockdown efficiency was verified using qPCR ( Figure 3B). The mRNA levels of EGR4 in ZNF205-AS1 stably overexpressed and control PC-9 cells, and ZNF205-AS1 stably depleted and control SPC-A1 cells was measured using qPCR. As displayed in Figure 3C-D, ectopic expression of ZNF205-AS1 up-regulated EGR4 mRNA levels, and while knockdown of ZNF205-AS1 by two independent shRNAs both down-regulated EGR4 mRNA levels. Furthermore, the protein levels of EGR4 in ZNF205-AS1 stably overexpressed and control PC-9 cells, and ZNF205-AS1 stably depleted and control SPC-A1 cells was measured using western blot.
As displayed in Figure 3E-F, ectopic expression of ZNF205-AS1 upregulated EGR4 protein levels, and while knockdown of ZNF205-AS1 by two independent shRNAs both down-regulated EGR4 protein levels. Thus, these results suggested that ZNF205-AS1 up-regulated EGR4 expression transcriptionally or post-transcriptionally.
To explore the detailed mechanisms underlying the positive regulation of EGR4 by ZNF205-AS1, we first detected the subcellular distribution of ZNF205-AS1 using purification of cytoplasmic and nuclear RNA, followed by qPCR. As displayed in Figure 3G

| The autoregulatory loop of ZNF205-AS1
The above results demonstrated that EGR4 activated ZNF205-AS1 transcription, and while ZNF205-AS1 also up-regulated EGR4,  Figure 4D). Collectively, these results suggested that ZNF205-AS1 activated its own promoter activity via promoting the occupation of EGR4 on its own promoter.  Figure 6D). Collectively, these results demonstrated that

| ZNF205-AS1 promoted NSCLC cell growth
The autoregulatory loop of ZNF205-AS1. A, After transiently transfecting ZNF205-AS1 overexpression plasmids into PC-9 cells, ChIP assay was carried out using EGR4 specific antibody. The retrieved DNA was quantified by qPCR to detect the occupation of EGR4 on the promoter of ZNF205-AS1. Results are displayed as percentage of input DNA. B, After transiently transfecting ZNF205-AS1 specific shRNAs into SPC-A1 cells, ChIP assay was carried out using EGR4 specific antibody. The retrieved DNA was quantified by qPCR to detect the occupation of EGR4 on the promoter of ZNF205-AS1. Results are displayed as percentage of input DNA. C, Luciferase reporter assay in PC-9 cells cotransfected with the ZNF205-AS1 promoter reporter construct and ZNF205-AS1 overexpression plasmid. The ratios of firefly luciferase (FLU) activity to Renilla luciferase (RLU) activity are displayed. D, Luciferase reporter assay in SPC-A1 cells cotransfected with the ZNF205-AS1 promoter reporter construct and ZNF205-AS1 specific shRNAs. The ratios of firefly luciferase (FLU) activity to Renilla luciferase (RLU) activity are displayed. Results are displayed as mean ± SD of three independent experiments. *P < 0.05, **P < 0.01, ns, not significant, by Student's t test (A, C) or one-way ANOVA followed by Dunnett's multiple comparison test (B, D) F I G U R E 3 ZNF205-AS1 stabilized EGR4 mRNA via RNA-RNA interaction. A, The expression of ZNF205-AS1 in ZNF205-AS1 stably overexpressed and control PC-9 cells was detected by qPCR. B, The expression of ZNF205-AS1 in ZNF205-AS1 stably depleted and control SPC-A1 cells was detected by qPCR. C, The mRNA levels of EGR4 in ZNF205-AS1 stably overexpressed and control PC-9 cells were detected by qPCR. D, The mRNA levels of EGR4 in ZNF205-AS1 stably depleted and control SPC-A1 cells was detected by qPCR. E, The protein levels of EGR4 in ZNF205-AS1 stably overexpressed and control PC-9 cells were detected by western blot. F, The protein levels of EGR4 in ZNF205-AS1 stably depleted and control SPC-A1 cells was detected by western blot. G, The subcellular distribution of ZNF205-AS1 in SPC-A1 cells was detected by qPCR. U6 and β-actin were used as nuclear and cytoplasmic control, respectively. H, Schematic diagram of the predicted RNA-RNA interaction between EGR4 mRNA and ZNF205-AS1 transcript. I, The RNA-RNA interaction between EGR4 mRNA and ZNF205-AS1 transcript was detected by RNA pulldown assay using in vitro transcribed biotin-labelled ZNF205-AS1. The retrieved RNA was quantified by qPCR and displayed as percentage of input RNA. J, After transiently transfecting ZNF205-AS1 overexpression plasmids into PC-9 cells, the stability of EGR4 mRNA over time was detected by qPCR relative to time 0 after blocking new RNA synthesis with α-amanitin (50 µM) and normalized to 18S rRNA (transcribed by RNA polymerase I and not influenced by α-amanitin). K, After transiently transfecting ZNF205-AS1 specific shRNAs into SPC-A1 cells, the stability of EGR4 mRNA over time was detected by qPCR relative to time 0 after blocking new RNA synthesis with α-amanitin (50 µM) and normalized to 18S rRNA. L, After transiently transfecting EGR4 overexpression plasmids into PC-9 cells, the stability of ZNF205-AS1 transcript over time was detected by qPCR relative to time 0 after blocking new RNA synthesis with α-amanitin (50 µM) and normalized to 18S rRNA. M, After transiently transfecting EGR4 specific shRNAs into SPC-A1 cells, the stability of ZNF205-AS1 transcript over time was detected by qPCR relative to time 0 after blocking new RNA synthesis with α-amanitin (50 µM) and normalized to 18S rRNA. Results are displayed as mean ± SD of three independent experiments. **P < 0.01, ***P < 0.001, ns, not significant, by Student's t test (A, C, J, L) or one-way ANOVA followed by Dunnett's multiple comparison test (B, D, I, K, M) consistent with ZNF205-AS1, EGR4 also promoted NSCLC cell growth in vitro.

ZNF205-AS1 and EGR4 inhibited NSCLC tumour growth in vivo
ZNF205-AS1 and EGR4 formed a positive feedback loop in NSCLC, and both ZNF205-AS1 and EGR4 promoted NSCLC cell growth in vitro. We next investigated the significances of targeting the positive feedback loop between ZNF205-AS1 and EGR4 for NSCLC. We constructed ZNF205-AS1 and EGR4 concurrently stably depleted SPC-A1 cells via transfecting ZNF205-AS1 and EGR4 specific shRNAs. The knockdown efficiency of EGR4 was confirmed using western blot ( Figure 7A). The knockdown efficiency of ZNF205-AS1 was confirmed using qPCR ( Figure 7B). ZNF205-AS1 and EGR4 concurrently depleted and control SPC-A1 cells were subcutaneously implanted into nude mice. Subcutaneous tumour growth was detected every three days ( Figure 7C). Subcutaneous tumours were excised and weighed at the 21th day after injection ( Figure 7D). As presented in Figure 7C-D, ZNF205-AS1 or EGR4 knockdown both repressed subcutaneous tumour growth. The concurrent knockdown of ZNF205-AS1 and EGR4 more significantly repressed tumour growth. Proliferation marker Ki67 IHC staining of subcutaneous xenografts further supported the growth repressive roles of ZNF205-AS1 or EGR4 knockdown, and more significant growth repressive roles of concurrent ZNF205-AS1 and EGR4 knockdown ( Figure 7E). Apoptosis marker TUNEL staining of subcutaneous xenografts displayed that ZNF205-AS1 or EGR4 knockdown both promoted cell apoptosis of subcutaneous tumours, and concurrent knockdown of ZNF205-AS1 and EGR4 more significantly promoted cell apoptosis of subcutaneous tumours ( Figure 7F). Furthermore, we F I G U R E 5 ZNF205-AS1 promoted NSCLC cell growth. A, Cell viabilities of ZNF205-AS1 stably overexpressed and control PC-9 cells were detected using Glo cell viability assay. The relative cell viability to 0 h is presented. B, Cell growth of ZNF205-AS1 stably overexpressed and control PC-9 cells were detected using EdU immunofluorescence staining; scale bars = 100 µm. C, Cell viabilities of ZNF205-AS1 stably depleted and control SPC-A1 cells were detected using Glo cell viability assay. The relative cell viability to 0 hour is presented. D, Cell growth of ZNF205-AS1 stably depleted and control SPC-A1 cells were detected using EdU immunofluorescence staining; scale bars = 100 µm. Results are displayed as mean ± SD of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, by Student's t test (A, B) or one-way ANOVA followed by Dunnett's multiple comparison test (C, D) and control SPC-A1 cells was determined using SA-β-gal staining. As presented in Figure 7I, neither ZNF205-AS1 nor EGR4 knockdown changed senescence of SPC-A1 cells. Collectively, these results suggested that targeting the positive feedback loop between ZNF205-AS1 and EGR4 inhibited NSCLC tumour growth in vivo, but did not regulate senescence.

| DISCUSSION
Genome and transcriptome sequencings have found more than 58 000 lncRNAs in human cells, but the number of protein-coding genes is only about 21 000, implying that the lncRNA landscape may be more complex and various. 30 Indeed, many aberrantly expressed lncRNAs have been identified in cancers. 31 Most of lncRNAs are temporal-spacial specifically and disease specifically expressed. 32 Although several lncRNAs have been found to play roles in cancers and regarded as cancer-associated lncRNAs, 33  LncRNA ZNF205-AS1 was identified as a candidate. In this study, we further investigate the expression and biological roles of ZNF205-AS1 in NSCLC. We found that ZNF205-AS1 was increased in NSCLC tissues and cell lines compared with adjacent noncancerous lung tissues and normal bronchial epithelial cell line, respectively.
Furthermore, increased expression of ZNF205-AS1 was positively associated with poor pathological differentiation, great tumour diameter, lymph nodes metastasis, advanced TNM stage, and poor overall survival of NSCLC patients. These data implied that ZNF205-AS1 F I G U R E 6 EGR4 promoted NSCLC cell growth. A, Cell viabilities of EGR4 stably overexpressed and control PC-9 cells were detected using Glo cell viability assay. The relative cell viability to 0 h is presented. B, Cell growth of EGR4 stably overexpressed and control PC-9 cells were detected using EdU immunofluorescence staining; scale bars = 100 µm. C, Cell viabilities of EGR4 stably depleted and control SPC-A1 cells were detected using Glo cell viability assay. The relative cell viability to 0 h is presented. D, Cell growth of EGR4 stably depleted and control SPC-A1 cells were detected using EdU immunofluorescence staining; scale bars = 100 µm. Results are displayed as mean ± SD of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, by Student's t test (A, B) or one-way ANOVA followed by Dunnett's multiple comparison test (C, D) HE ET AL. These results demonstrated that ZNF205-AS1 functioned as an oncogene in NSCLC and ZNF205-AS1 may also be regarded as a cancer-associated lncRNA in NSCLC.
Although many aberrantly expressed lncRNAs have been identified in cancers, the reasons contributing to the dysregulation of lncRNAs are relatively less studied. In this study, using computational screen, we identified EGR4 binding sites on the promoter of ZNF205-AS1. EGR4 is a transcription factor, which belongs to the early growth response (EGR) family of immediate early genes. 34 Previous study mainly focused on the roles of EGR4 in male infertility and neural development. 35 42 Feedback loops could amplify the effects of interaction molecules in caners and more significantly promote the aberrant expression of these molecules. 43 In this study, we further found that targeting the feedback loop between EGR4 and ZNF205-AS1 via concurrently depleting EGR4 and ZNF205-AS1 significantly repressed NSCLC tumour growth in vivo. Although concurrent depletion of EGR4 and ZNF205-AS1 for clinical application is difficult until now, but the combination of EGR4 inhibitor and ZNF205-AS1 siRNA therapeutics in the future may be promising therapeutic strategy for NSCLC.
In conclusion, this study demonstrated that lncRNA ZNF205-AS1 formed a positive feedback loop with EGR4, which contributed to the up-regulations and the oncogenic roles of ZNF205-AS1 and EGR4 in NSCLC. Our data suggested that the positive feedback loop between ZNF205-AS1 and EGR4 may be promising therapeutic target for NSCLC.

ACKNOWLEDG EMENTS
This work was supported by the Medical and Health Research Fund of Zhejiang Province (2017KY165).

CONFLI CT OF INTEREST
The authors declare that they have no conflict of interest. F I G U R E 7 Inhibition of the positive feedback loop between ZNF205-AS1 and EGR4 significantly repressed NSCLC tumour growth in vivo. A, The expression of EGR4 in ZNF205-AS1 and EGR4 concurrently depleted and control SPC-A1 cells was detected by western blot. B, The expression of ZNF205-AS1 in ZNF205-AS1 and EGR4 concurrently depleted and control SPC-A1 cells was detected by qPCR. Results are displayed as mean ± SD of three independent experiments. ***P < 0.001 by one-way ANOVA followed by Dunnett's multiple comparison test. C, ZNF205-AS1 and EGR4 concurrently depleted and control SPC-A1 cells were subcutaneously injected into nude mice. Tumour volumes were detected every three days. D, Subcutaneous tumour weights were detected at the 21th day after injection. E, Ki67 immunohistochemistry staining of tumours derived from ZNF205-AS1 and EGR4 concurrently depleted or control SPC-A1 cells; scale bars = 50 μm. F, TUNEL staining of tumours derived from ZNF205-AS1 and EGR4 concurrently depleted or control SPC-A1 cells; scale bars = 50 μm. For C-F, results are displayed as mean ± SD of n = 5 mice in each group. *P < 0.05, **P < 0.01 by Mann-Whitney test. G,H, ZNF205-AS1 (G) and EGR4 (H) expression levels in NSCLC tissues with strong or weak Ki67 staining intensity. The median Ki67 staining intensity was used as the cutoff (n = 90 NSCLC tissues). Results are displayed as median with interquartile range. P < 0.0001 by Mann-Whitney test. I, Cellular senescence of ZNF205-AS1 and EGR4 concurrently depleted and control SPC-A1 cells was detected by SA-β-gal activity. ns, not significant, by one-way ANOVA followed by Dunnett's multiple comparison test