Long noncoding RNA ARHGAP5-AS1 inhibits migration of breast cancer cell via stabilizes SMAD7 protein

Background: Tumor metastasis is the main cause of death from breast cancer patients and cell migration plays a critical role in metastasis. Recent studies have shown long non-coding RNAs (lncRNAs) play an essential role in the initiation and progression of cancer. In the present study, the role of a LncRNA, ARHGAP5-AS1 in breast cancer was investigated. Methods: Bioinformation was analyzed for the expression of ARHGAP5-AS1. qRT-PCR was conducted to verify the expression of ARHGAP5-AS1 in breast cancer specimens. Transwell migration assays and F-actin staining were utilized to estimate cell migration ability. RNA pulldown assays and RNA immunoprecipitation were used to prove the interaction between ARHGAP5-AS1 and SMAD7. Western blot and immunouorescence imaging were used to examine the protein levels. Dual luciferase reporter assays were performed to evaluate the activation of TGF-β signaling. Results: Compared to MDA-MB-231 cells, the expression of LncRNA ARHGAP5-AS1(NR_027263) was signicantly suppressed in its highly metastatic subtype MDA-MB-231-LM2 cells. Functional study showed ARHGAP5-AS1 could inhibit cell migration via suppression of stress bers in breast cancer cell lines. Afterwards, SMAD7 was further identied to interact with ARHGAP5-AS1 by its PY motif and thus its ubiquitination and degradation was blocked due to reduced interaction with E3 ligase SMURF1 and SMURF2. Moreover, ARHGAP5-AS1 could inhibit TGF-β signaling pathway due to its inhibitory role on SMAD7. Conclusions: Overall, these ndings demonstrate that ARHGAP5-AS1 inhibits breast cancer cell migration and could server as a novel biomarker for breast cancer metastasis and a potent target for the treatment in the future. novel activating protein antisense cell line


Background
Breast cancer is one of the most common cancer and accounts for 29% (246,660) of all new cancers among the worldwide women [1]. According to the latest statistics, released by the American Cancer Society, about 276,480 new cases of invasive breast cancer are expected to be diagnosed in the United States in 2020, which means one in eight women will get breast cancer during their lifetime. In addition, 42,170 deaths per year makes breast cancer the second leading cause of cancer-related mortality among women in the USA [2]. The main cause of death from breast cancer is due to metastasis to other organs (e.g. bone, lung, brain and liver) [3]. Nevertheless, the mechanisms underlying the metastatic dissemination remain poorly understood, which causes a critical barrier for breast cancer therapy. Cell migration is regarded to be one of the essential step involved in metastasis [4]. Reduced migration suppressed the dissemination of cancer cells in a breast cancer mouse model [5]. It is well characterized that cancer cell could achieve migration via rearrangement of cytoskeleton [6]. Thus, expanding the understanding of mechanisms underlying cytoskeleton reorganization may promote the understanding of cell migration and hence cancer metastasis.
Recent studies have shown that long non-coding RNAs (lncRNAs) play an essential role in the initiation and progression of cancer. LncRNAs are gene transcripts which have more than 200 nucleotides in length and no potential of translation [7]. LncRNAs have been proposed to carry out diverse functions, including transcriptional regulation in cis or trans, organization of nuclear domains, and regulation of proteins or RNA molecules [8]. It is now widely understood that lncRNAs could identify cellular pathologies such as cancer, provide prognostic value, or even serve as therapeutic options for cancer patients [9]. For example, over-expression of the lncRNA HOTAIR in early-stage, surgically resected breast cancer is highly predictive of progression to metastatic disease and overall survival [10]. Silencing the expression of the lncRNA GUARDIN triggered apoptosis and senescence, enhanced cytotoxicity of additional genotoxic stress and inhibited cancer xenograft growth. Thus, GUARDIN may constitute a target for cancer treatment [11].
The transforming growth factor b (TGFb) signaling pathway is a key player in metazoan biology, and its misregulation can result in tumor development. Pathological forms of TGFb signaling promote tumor growth and invasion, evasion of immune surveillance, as well as cancer cell dissemination and metastasis [12]. Smad6 and Smad7 are inhibitory Smads that negatively control TGFβ pathway activity in response to feedback loops and antagonistic signals [13]. SMAD7, as one of the key inhibitor of TGFb signaling, negatively regulates the whole pathway via multiple mechanisms [14]. In the cytoplasm, SMAD7 can compete the binding site of type-1 TGFb receptor (TGFb R1) with SMAD2/3 and hence inhibit signal transduction by blocking phosphorylation of SMAD2/3 [15]. Besides, SMURF1 and SMURF2 are recruited to TGFb R1 by SMAD7 and induce the degradation of receptors [16]. Meanwhile in the nucleus, SMAD7 disrupts the binding of SMAD2/3/4 complex with DNA [17].
In our present study, we identi ed a novel lncRNA, Rho GTPase activating protein 5 antisense RNA 1 (ARHGAP5-AS1), which expression is downregulated in highly metastatic breast cancer cell line MDA-MB-231-LM2. Functional study showed that ARHGAP5-AS1 could inhibit migration of breast cancer cells through the inhibition of stress bers. Moreover, we found that ARHGAP5-AS1 could interact with SMAD7 and stabilized SMAD7 protein via blocking the interation between SMAD7 and its E3 ligase (SMURF1 & SMURF2), and hence inhibited its ubiquitination and degradation. Further more, ARHGAP5-AS1 could inhibit TGFβ signaling pathway by upregulation of TGFb R1. Taken together, these ndings demonstrate that ARHGAP5-AS1 could server as a novel biomarker for breast cancer metastasis and a potent target for the treatment in the future. (Gibco) supplemented with 10% FBS. SKBR3 and BT549 were cultured in RPMI 1640 (Hyclone) medium supplemented with 10% FBS. All cell lines were incubated at 37 ℃ in a humidi ed atmosphere of 5% CO 2 and 95% air except for MDA-MB-231 which were raised in a humidi ed atmosphere containing 100% air.

RNA extraction and quantitative real-time PCR
Total RNA of cancer tissues and cell line were extracted from Trizol reagents (Invitrogen#15596018) according to the manufacturer's instructions. cDNA was obtained by AMV reverse transcription system (TAKARA#2621) followed the manufacturer's protocol. Quantitative real-time PCR was performed with SYBR Green PCR master mix reagent (ABI#4472908) and speci c primers for SMAD7-F/R (ATGCTGTGCCTTCCTCCGCTG/CCACGCACCAGTGTGACCGA) and ARHGAP5-AS1-F/R (GGCCCCTGATTCAGTACGTT/GCGTGAACAGGGGTCTTTTG). Data analysis of ARHGAP5-AS1 expression in breast cancer cell lines was normalized by internal control 28S RNA and evaluated using 2 ∆∆C t method, while data analysis of breast cancer tissues was normalized by 18S RNA and presented by -∆Ct.

RNA Fluorescent in situ hybridization
This experiment was performed with Fluorescent In Situ Hybridization Kit (Ribo#R11060.1) according to the manufacturer's protocol. Brie y, 1.5 × 10 5 MDA-MB-231 cells or LM2 cells was seeded onto cover slides in 24-well culture plate. LncRNA ARHGAP5-AS1 was detected by speci c probe with Cy3 labeling.
Fluorescence pictures were taken by Nikon Eclipse Ti microscope (Nikon, Kanagawa, Japan).

Protein microarray screening
Sense and antisense lncRNA were transcripted in vitro with Cy5 labeling. Synthesized lncRNA were incubated with HuProt TM microarray (CDI Laboratories, Inc.) which contained more than 20000 full length human proteins with GST tag. This procedure was completed by Wayen biotechnologies (Shanghai), Inc. according to the manufacturer's instructions. In brief, proteome microarrays were blocked with blocking buffer for 1 h at room temperature before incubation of transcripted lncRNA in blocking buffer with microarray at room temperature for 1 h. TBST followed by MilliQ water were used to wash the microarrays three times for 5 min. Dried slides were scanned using GenePix 4000B. Images were analyzed using GenePix Pro 6.0.

Statistical analyses
All experiments were repeated at least three times or otherwise mentioned. The p-values for comparison between two groups were obtained by Student's t-test (two-tailed). All values were presented as Mean ± STD and the p-value 0.05 was considered to be statistically signi cant.  Figure 1C). According to the location of NR_027263, it was named by Rho GTPase activating protein 5 antisense RNA 1 (ARHGAP5-AS1). To determine the identity of ARHGAP5-AS1, the translational potential of ARHGAP5-AS1 was detected by in vitro translational system. The results showed ARHGAP5-AS1 cannot encode any protein ( Figure 1D Figure 2E) as well as SKBR3 cells ( Figure   2F). Conversely, over-expression of ARHGAP5-AS1 reduce the formation of stress bers in LM2 cells ( Figure 2G) as well as BT549 cells ( Figure 2H). Taken together, ARHGAP5-AS1 inhibited migration of breast cancer cells through the inhibition of stress bers formation.

ARHGAP5-AS1 interacts with the PY motif of SMAD7
LncRNAs can function as protein scaffold or microRNA sponge to regulate cellular processes in cytoplasm, while regulating gene transcription in cis or trans in the nucleus [9]. ARHGAP5-AS1 mainly existed in cytoplasm rather than in nucleus ( Figure 3A), so we focused on the proteins interacting with ARHGAP5-AS1. In order to explore the mechanism of ARHGAP5-AS1 inhibiting migration of breast cancer cells, protein microarray screening was utilized to nd the interacting protein with ARHGAP5-AS1. Firstly, sense and antisense ARHGAP5-AS1 was transcripted in vitro with Cy5 labeling ( Figure S1A). The screening showed that there were 26 proteins interacting with ARHGAP5-AS1 sense excluding with antisense ( Figure S1B). There were ve proteins correlated to breast cancer progression among the interacting proteins with ARHGAP5-AS1 ( Figure S1C). SMAD7 was proved to interact with ARHGAP5-AS1 by RNA pull down assay ( Figure 3B). More importantly, the enrichment of SMAD7 and ARHGAP5-AS1 complex could be detected by RNA immunoprecipitation ( Figure 3C). Furthermore, truncations with Flag tag were constructed to determine which domain of SMAD7 interacting with ARHGAP5-AS1. Precisely, truncation 1 lost C-terminal MH2 domain while truncation 2 remained N-terminal MH1 domain only, compared to full-length SMAD7 ( Figure 3D). The interactions of different truncations with ARHGAP5-AS1 were detected by RNA pull down assay. The results showed that both full-length and truncation 1 SMAD7 were able to interact with ARHGAP5-AS1, but truncation 2 lost the interaction with ARHGAP5-AS1 ( Figure   3E). It suggested that PY motif was essential for the interaction between ARHGAP5-AS1 and SMAD7. Taken together, ARHGAP5-AS1 interacted with SMAD7 through its PY motif.

ARHGAP5-AS1 stabilizes SMAD7 via inhibition of ubiquitination of SMAD7
PY motif of SMAD7 is responsible for the interaction between SMAD7 and its E3 ligase thus is important for the degradation of SMAD7 [20]. Due to the interaction of ARHGAP5-AS1 with PY motif of SMAD7, whether the stability of SMAD7 was in uenced by ARHGAP5-AS1 was investigated. As expected, knockdown of ARHGAP5-AS1 reduced the protein level of SMAD7 in MDA-MB-231 cells and SKBR3 cells, while over-expression of ARHGAP5-AS1 increased the protein level of SMAD7 in LM2 cells and BT549 cells ( Figure 4A-4B). The degradation of SMAD7 was controlled by ubiquitin and proteosome [21]. The proteosome inhibitor MG132 was applied to block the effects of ARHGAP5-AS1 on SMAD7. The results showed that MG132 blocked the degradation of SMAD7 caused by knockdown of ARHGAP5-AS1 ( Figure   4C). In order to certify that the reduction of SMAD7 by knockdown of ARHGAP5-AS1 is due to the increase in degradation rather than the decrease in its mRNA level, the relative mRNA level of SMAD7 was detected by qRT-PCR. The results showed that whether knockdown or over-expression of ARHGAP5-AS1 had no in uence on the mRNA level of SMAD7 ( Figure S2A-S2B). Meanwhile, knockdown of ARHGAP5-AS1 accelerated the degradation of SMAD7 when the protein synthesis inhibitor cycloheximide (CHX) stimulation applied ( Figure 4D). Degradation of proteins in proteosome system results from polyubiquitination of proteins, so that the ubiquitination of SMAD7 was detected by immunoprecipitation. The results showed that over-expression of ARHGAP5-AS1 reduced ubiquitination of SMAD7 in 293T cells ( Figure 4E). Moreover, ARHGAP5-AS1 had a negative effect on the interaction between SMAD7 and its E3 ligases SMURF1 and SMURF2 ( Figure 4F). On the contrast, ARHGAP5-AS1 had no effect on the interaction between SMAD7 and its deubiquitinase OTUD1 ( Figure S2C). Taken together, ARHGAP5-AS1 stabilized SMAD7 via reduction of ubiquitination as well as the degradation of SMAD7.
SMAD7 mediates the negative effect of ARHGAP5-AS1 on cell migration As SMAD7 antagonizes TGFβ signaling, it has critical role in TGF-β induced cytoskeleton rearrangement [22]. In order to determine whether SMAD7 mediated the suppression of cell migration by ARHGAP5-AS1, antisense DNA oligos was utilized to knockdown SMAD7 which mimicked knockdown of ARHGAP5-AS1.
The knockdown e ciency was detected by western blot, and the results showed e cient knockdown of SMAD7 ( Figure 5A). Immuno uorescence by TRITC-phalloidin showed increase in stress ber formation in MDA-MB-231 cells with knockdown of SMAD7 ( Figure 5B). As a result, knockdown of SMAD7 enhanced migration of MDA-MB-231 cells ( Figure 5C). In sum, knockdown of SMAD7 had the similar effect on cell migration to knockdown of ARHGAP5-AS1. To investigate whether SMAD7 was able to rescue the promotion of migration by reduction of ARHGAP5-AS1, the stable over-expression of SMAD7 MDA-MB-231 cells was constructed. The expression of SMAD7 was detected by western blot. The results showed that knockdown of ARHGAP5-AS1 reduced the protein level of SMAD7 in plvx-EV cells, while ectopic expression of SMAD7 blocked the reduction of SMAD7 in plvx-SMAD7-OE cells ( Figure 5D). Immuno uroscence of F-actin showed that knockdown of ARHGAP5-AS1 promoted the formation of stress ber in plvx-EV cells, but had no effects in plvx-SMAD7-OE cells ( Figure 5E). The similar results were achieved in migration assay ( Figure 5F). Taken together, knockdown of SMAD7 exerted similar function to ARHGAP5-AS1. Moreover, ectopic expression of SMAD7 blocked the enhancement of migration by reduction of ARHGAP5-AS1 in MDA-MB-231 cells.
Knockdown of ARHGAP5-AS1 promotes TGF-β signaling via SMAD7 Due to the different inhibitory mechanisms of SMAD7 in nucleus and cytoplasm [14], the subcellular location of SMAD7 was detected with knockdown of ARHGAP5-AS1. The results of immuno uorescence showed no change of SMAD7 location by knockdown of ARHGAP5-AS1 in MDA-MB-231 cells ( Figure 6A). Besides, extraction of nuclear and cytoplasmic SMAD7 was consistent with observation in immuno uorescence ( Figure 6B). Notably, expression of SMAD7 in cytoplasm was reduced by knockdown of ARHGAP5-AS1, which was similar to previous results ( Figure 4A). SMAD7 turned off TGFβ signaling by inducing the degradation of TGFb R1 [16]. In order to investigate the effects of ARHGAP5-AS1 on TGFβ signaling, the components of TGFβ pathway were detected by western blot. The results showed increased expression of TGFb R1 and prolonged activation of TGFβ signaling, which in turn presented by increased phosphorylation of SMAD2 in response to TGFβ1 stimulation ( Figure 6C). The similar pattern was showed by knockdown of SMAD7 ( Figure 6D). Moreover, we constructed four SMAD-binding elements (SBE) into pGL4.27 vector and performed dual luciferase reporter assays in 293T cells. The results showed that the activity of re y luciferase was enhanced by the treatment of TGF-β and over-expression of ARHGAP5-AS1 inhibited the enhancement of luciferase activity ( Figure 6E). Taken together, knockdown of ARHGAP5-AS1 promoted TGFβ signaling by upregulation of TGFb R1.

Discussion
This study provides the rst evidence that LncRNA ARHGAP5-AS1 was downregulated in human breast cancer specimens. Furthermore, ARHGAP5-AS1 could inhibit cell migration via suppression of stress bers in breast cancer cells. More importantly, to our knowledge it is the rst lncRNA could interact with SMAD7, which is an important inhibitory Smads that negatively control TGFb signaling pathway. The present study demonstrated that ARHGAP5-AS1 impaired the degradation of SMAD7 via interaction with its PY motif, blocking the interation with its E3 ligase (SMURF1 & SMURF2) and thus decreased its ubiquitination. Knockdown of SMAD7 mimicked the promotion of cell migration by knockdown of ARHGAP5-AS1. Meanwhile, ectopic expression of SMAD7 blocked the increase in migration and stress ber formation by knockdown of ARHGAP5-AS1. Moreover, ARHGAP5-AS1 inhibited the formation of stress ber and hence migration via suppression of TGFβ signaling by stabilized SMAD7.
ARHGAP5-AS1 located on the human chromosome 14 and was rst identi ed as a transcriptional isoform of ARHGAP5 in gastric cancer by large-scale sequencing [23]. Non-coding RNA was considered as transcription "noise" in the past. However, more and more studies showed an emerging and diverse role in cancer pathways [9]. For example, lncRNA NKILA suppresses breast cancer metastasis by interaction with p65 and blocking IκB phosphorylation [24]. Here we illustrated that ARHGAP5-AS1 was a long non-coding RNA and its expression was downregulated in breast cancer tissues, which data were achieved from TCGA database. Meanwhile, the downregulation of ARHGAP5-AS1 was further evidenced in the collected human breast cancer tissues (n = 22), compared to the corresponding adjacent normal tissues, suggesting a tumor-suppressor role in breast caner. Thus, we investigated the role of ARHGAP5-AS1 in initiation and progression of breast cancer. As a result, ARHGAP5-AS1 inhibited migration of breast cancer cells. Moreover, stress ber formation was reduced by ARHGAP5-AS1. It's a pity that we have not studied the regulatory mechanism of ARHGAP5-AS1 expression in breast cancer cells. We thought there could be a possibility that the expression of ARHGAP5-AS1 is regulated by some transcription factors which expression are quite different in breast cancer tissues.
LncRNA functions as protein scaffold, which are already well established [25]. We demonstrated that ARHGAP5-AS1 interacted with PY-motif of SMAD7. PY motif is responsible for the interaction of SMAD7 with WW domain of E3 ligase [20]. Indeed, ARHGAP5-AS1 occupied PY motif and blocked the ubiquitination of SMAD7 by E3 ligase (SMRUF1 & SMURF2), so that reduced the degradation of SMAD7 protein. SMAD7 is a powerful inhibitor of TGFβ signaling, antagonizing TGFβ signaling via multiple mechanisms [14]. TGFβ signaling pathway exerts tumor-suppressive effects in cancer initiation, yet it promotes epithelial to mesenchymal transition (EMT), cell invasion as well as migration in cancer progression [26,27]. The dual role of TGFb in cancer has long been noted, but its mechanistic basis, operating logic, and clinical relevance have remained elusive. TGFβ induces cytoskeleton reorganization by upregulation of RhoGEFs, such as NET1 [28] and GEFH1 [29]. While SMAD7 inhibits TGFβ induced actin reorganization and RhoA activation [22]. It is expectable that TGFβ induced stress bers and cell migration could be blocked by ARHGAP5-AS1 due to its interaction with SMAD7 and increased the stability of the latter. However, the function of ARHGAP5-AS1 in breast cancer cells was detected without exogenous TGFβ stimulation. Tumor cells themselves are able to produce TGF-β ligands [30], [31], which might be a reasonable explanation. Indeed, knockdown of SMAD7 mimicked the promotion effect of cell migration by knockdown of ARHGAP5-AS1. Furthermore, ectopic expression of SMAD7 was able to block the promotion of cell migration by knockdown of ARHGAP5-AS1. Above results were also achieved without TGFβ stimulation. To verify the effects of ARHGAP5-AS1 on TGFβ stimulated signaling, we further knockdown the expression of ARHGAP5-AS1 in MDA-MB-231 cells and these cells were starved for 12 h and treated with 0.3 ng/mL TGFβ1. Prolonged activation of TGFβ signaling was demonstrated by upregulation of TGFb R1 and increased phosphorylation of SMAD2 compared to the control cells.
Besides, four fragments of SMAD-binding element (SBE) were cloned into pGL4.27 vector. ARHGAP5-AS1 overexpression showed reduced luciferase activity with stimulation of TGFβ, proving the inhibitory role on TGFβ signaling.
In this study, we give a clear interpretation that ARHGAP5-AS1, which is downregulated in breast cancer tissue, suppresses the cell migration and stress ber formation in breast cancer cells. Furthermore, SMAD7, which is an important inhibitory Smads that negatively control TGFb signaling pathway, is demonstrated to be interacted with ARHGAP5-AS1 and its protein level is regulated by ARHGAP5-AS1. Thus, our ndings provide valuable clues toward understanding the mechanisms of human breast cancer progression and present an opportunity to develop more effective clinical therapies in the future.

Conclusions
Our experiments, results, and analysis indicated that ARHGAP5-AS1 inhibits cell migration through its inhibitory role on SMAD7 in breast cancer. ARHGAP5-AS1 could server as a novel biomarker for breast cancer metastasis and a potent target for the treatment in the future.

Declarations
Authors' contributions