Expression and correlation of EGFR-AS1, miR-149-5p and ELP5mRNA in BC patients
We first investigated several differentially expressed lncRNAs in BC and found that EGFR-AS1 was markedly upregulated. Moreover, TCGA and Starbase database analysis revealed that the level of EGFR-AS1 was upregulated in BC tissues compared with normal tissues (Fig. 1A–C). Similarly, the level of ELP5 was upregulated in BC tissues (Fig. 1 D–E). We then assessed the levels of EGFR-AS1, miR-149-5p, and ELP5 mRNA in 33 pairs of matched BC tissues by qRT-PCR, and found that the expression of EGFR-AS1 and ELP5 mRNAs in the BC tissues was upregulated, whereas that of miR-149-5p was downregulated (Figure 1 F–H). Furthermore, we investigated the relationship between EGFR-AS1, miR-149-5p, and ELP5 mRNA in BC tissues. Pearson’s correlation analysis indicated negative correlations between miR-149-5p and EGFR-AS1(R2=0.060 P<0.05) and miR-149-5p and ELP5mRNA(R2=0.090 P<0.05) in the BC tissues and a positive correlation between EGFR-AS1 and ELP5 mRNA(R2=0.204 P<0.01) (Fig. 1I–K).
Effect of EGFR-AS1 knockdown on proliferation, migration, and invasion of BC cells in vitro
The results of qPCR analysis indicated higher levels of EGFR-AS1 in MCF-7, MDA-MB-231, MDA-MB-453, BT-549, and SK-BR-3 BC cell lines than in MCF-10A cells (Fig. 2A). Moreover, MCF-7 and MDA-MB-231 cells had the highest expression of EGFR-AS1 among the BC cell lines. Therefore, we selected MCF-7 and MDA-MB-231 cells for further analysis. To evaluate the biological functions of EGFR-AS1 in BC cell proliferation and migration, transient transfection of EGFR-AS1-expressing cell lines was performed, wherein si-EGFR-AS1-siRNA was used for knockdown functional analysis. The level of EGFR-AS1 were detected by qPCR to demonstrate the effectiveness of transfection (Fig. 2B). The growth curve of the MTT proliferation assay showed that the restraint of EGFR-AS1 significantly inhibited the growth of MCF-7 and MDA-MB-231 cells (Fig. 2C). Moreover, after transfection with si-EGFR-AS1, cell viability was more efficiently inhibited. Colony formation assay revealed that EGFR-AS1 decreased the number and size of colonies in the MCF-7 and MDA-MB-231 cells (Fig. 2D), with the same results obtained using EdU assays (Fig. 2E). Therefore, we determined that EGFR-AS1 knockdown markedly decreased cell proliferation. These results demonstrate that EGFR-AS1 promotes BC cell proliferation.
To further confirm that EGFR-AS1 facilitated the migration and invasion of BC cells, we assessed whether EGFR-AS1 knockdown affected the expression of proteins associated with cell migration and invasion, namely, ROCK1, RhoA, RhoB, and matrix metalloproteinase (MMP)2. Result show that the expression of all of these proteins was dysregulated in EGFR-AS1 knockdown cultures (Fig. 3A).
In addition, local tumor invasion is the first step in tumor metastasis and requires significant changes in the adhesion and migration of tumor cells, which is reminiscent of the development of EMT. Therefore, we determined the levels of EMT-related proteins and found that EGFR-AS1 knockdown notably decreased the levels of EMT-related proteins, such as N-cadherin, SLUG, SNAIL, and Vimentin (Fig. 3B). Moreover, transwell assays showed that EGFR-AS1 knockdown significantly inhibited the migration and invasion of BC cells compared with those of control cells (Fig. 3 C–D). According to the literature, EGFR-AS1 knockdown also suppresses the expression of PI3K/AKT pathway markers [20], which agrees with our results (Fig. 3 E).
Collectively, these findings demonstrated that EGFR-AS1 promoted the proliferation, migration, and invasion of BC cells, consistent with our biofunction and western blot results. In addition, we evaluated patient prognosis using the TCGA database, which suggested that EGFR-AS1 expression is associated with poor prognosis in BC (Fig. 3 F–G).
Maintenance of EGFR mRNA stability by EGFR-AS1 via complementary binding
Previous studies have demonstrated that lncRNAs can be broadly divided into cis-acting lncRNAs, affecting the expression or chromatin state of nearby genes, and trans-acting lncRNAs that perform various functions in the cell [20, 21, 31, 32]. Considering that EGFR-AS1 is transcribed on the opposite strand of EGFR, we analyzed the RNA sequence base pairing between EGFR-AS1 and EGFR mRNA. A long base-pairing with a length of 194 bp was predicted by NCBI BLAST (http://blast.ncbi.nlm.nih.gov/) (Fig. 4A). Next, we detected the expression of EGFR mRNA in the same paired 33 BC tissues by qRT-PCR, and observed an upregulation in the expression of these mRNAs (Fig. 4B). Moreover, we observed that EGFR expression was positively correlated with EGFR-AS1 in BC tissues (Fig. 4C). Previous studies have demonstrated the ability of EGFR-AS1 to cis-regulate the neighboring gene EGFR in hepatocellular carcinoma (HCC) tissues [33]. Thus, we investigated whether EGFR-AS1 affects EGFR expression in BC tissues and cells. EGFR protein abundance was significantly higher in EGFR-high samples compared with EGFR-low BC tissue samples (Fig. 4D). Furthermore, the expression of EGFR mRNA and abundance of EGFR protein were detected. EGFR was markedly decreased following EGFR-AS1 knockdown in BC cells (Fig. 4E and F). Thereafter, we investigated whether EGFR-AS1 knockdown suppressed EGFR mRNA expression following treatment with actinomycin D (ActD), a transcriptional inhibitor. The mRNA stability of EGFR were tested by qRT-PCR in 0h,2h,4h,8h,12h in BC cells. The greatest effect was observed after 4 h, indicating that EGFR mRNA stability decreased following EGFR-AS1 silencing (Fig. 4G). These results demonstrated that EGFR-AS1 binds to EGFR mRNA and affects its stability in BC cells. Finally, nucleoplasmic separation experiments revealed that EGFR-AS1 localizes in both the cytosol and nucleus of BC cells (Fig. 4H) and, thus, participates in target gene regulation.
Effects of ELP5 on proliferation, metastasis, and invasion of BC cells
ELP5 is a component of the RNA polymerase II elongator complex, a multiprotein complex associated with the RNA polymerase II holoenzyme, and is involved in transcriptional elongation [26]. ELP5 participates in controlling cell motility and the tumorigenicity of melanoma cells [25]. To examine the effects of ELP5 expression on tumorigenesis-related processes in BC cells, we performed several experiments. First, we analyzed the level of ELP5 in 33 paired BC and adjacent tissue samples using western blotting and determined that the expression of ELP5 was higher in BC tissues than in adjacent normal tissues (Fig. 5A and S1a). Moreover, western blotting and qPCR results indicated higher levels of ELP5 in MCF-7, MDA-MB-231, MDA-MB-453, BT-549, and SK-BR-3 BC cell lines compared to MCF-10A cells (Fig. 5B and C). Next, immunochemistry analysis of 15 paired BC and paired normal tissue samples showed higher levels of ELP5 in BC tissues than in adjacent normal tissues (Fig. 5D).
We then constructed ELP5 siRNA and overexpression plasmid constructs to explore the function of ELP5 in BC (Fig. S1b). Changes in ELP5 protein level affected the expression of proteins involved in cell migration and invasion as well as that of EMT-related markers, with ELP5 depletion resulting in decreased expression of RhoA, ROCK1, MMP2, E-cadherin, Vimentin, Snail, and Zeb and an increase in RhoB expression. However, ELP5 overexpression caused the opposite effect, suggesting that ELP5 promotes the proliferation, migration, and invasion of BC cells (Fig. 5E-Fand S1c-d). ELP5 depletion inhibited the expression of PI3K/AKT pathway markers, whereas ELP5 overexpression promoted the expression of these markers (Fig. 5G and S1e). We then evaluated the effects of this pathway using LY294002, which inhibits PI3K expression and reduces phosphorylation in the MCF-7 and MDA-MB-231 cells transfected with P-ELP5. LY294002 administration reduced the induction of the various cellular functions, and the negative correlation between ELP5 expression and that of RhoA, ROCK1, P21, and cyclinD1, was reversed (Fig. 5H-Iand S1f-h).
EGFR-AS1 as a miR-149-5p molecular sponge
It is widely acknowledged that lncRNAs can transcriptionally, or post-transcriptionally, activate downstream genes [33, 34]. As mentioned before, EGFR-AS1 is principally located in the cytoplasm and plays a primary role in post-transcriptional regulation [20, 21, 33]. The ceRNA network is a common post-transcriptional regulatory pattern of lncRNAs. To verify the hypothesis that EGFR-AS1 acts as a ceRNA to regulate mRNA expression, we predicted the target miRNAs of EGFR-AS1 by StarBase, miRBase, and RegRNA. We then selected the top three common miRNAs, namely hsa-miR-149-5p, hsa-miR-195-3p, and hsa-miR-661, which can be absorbed by EGFR-AS1. Moreover, qRT-PCR analysis was performed to analyze the expression of these three miRNAs. In this case, we confirmed the most evident regulatory effect of EGFR-AS1 and miR-149-5p in BC cell lines (Figure. 6A and S2a). Reports have shown that miR-149-5p plays a vital role in the suppression of cancer by targeting various oncogenes that regulate tumor-related processes [14, 16, 22–24]. Our analysis suggested that the level of miR-149-5p was markedly lowered in BC cell lines compared to MCF-10A cells (Figure. 6B).
Next, the dual-luciferase reporter assay was used to detect direct binding between EGFR-AS1 and miR-149-5p based on their complementary sequences. An EGFR-AS1 fragment was constructed and inserted downstream of the luciferase reporter gene. We then co-transfected a miR-149-5p mimic with the reporter gene into MCF-7 and MDA-MB-231 cells. A significant reduction in luciferase reporter activity was observed relative to co-transfection with control RNA, while the co-transfection of miR-149-5p and mut-EGFR-AS1 did not induce a significant reduction in the luciferase signal (Fig. 6C–D and S2b). Therefore, a direct interaction between EGFR-AS1 and miR-149-5p was confirmed.
To investigate whether miR-149-5p was regulated by EGFR-AS1 in BC cells, we assessed the levels of ELP5, ROCK1, RhoA, and P21 via the co-transfection of cells with si-EGFR-AS1 and miR-149-5p inhibitors. Interestingly, the inhibition of ELP5, ROCK1, RhoA, and P21 protein expression induced by si-EGFR-AS1 was reversed by miR-149-5p inhibition (Fig. 6E and S2c). Thus, we sought to determine whether the biological function of EGFR-AS1 in BC cells could also be reversed by miR-149-5p inhibitors. Our analysis revealed reduced proliferation, as indicated by the MTT proliferation assay, colony formation experiments, and EdU assay. Likewise, reduced metastasis and invasion revealed by the transwell assay in BC cells and mediated by EGFR-AS1 knockdown was successfully restored by miR-149-5p inhibitors (Fig. 6F-I and S2d-g).
ELP5 is the functional target of miR-149-5p in BC
To identify miR-149-5p target genes, we screened the StarBase and TargetScan databases for the predicted targets of miR-149-5p and combined them with the genes that were significantly upregulated in the TCGA BC database. Using this strategy, we identified 12 genes (Fig. S3a).
Experiments were then performed to identify the interaction between miR-149-5p and ELP5. First, we detected the levels of miR-149-5p mimics and inhibitor in the transient transfection of cell lines by qPCR to confirm successful transfection (Fig. S3b-e). Next, the binding sequence of miR-149-5p to ELP5 3′UTR was predicted and obtained (Fig. 7A). We then examined the luciferase activity of the reporter containing wild-type ELP5 (ELP5-WT) or mutant-type ELP5 (ELP5-MUT) in cells transfected with miR-149-5p mimics. The results suggested that the luciferase activity of the ELP5-WT vector was only restrained by miR-149-5p mimics (Fig. 7B and S4a). qPCR and western blotting were then used to verify the expression of ELP5 at the mRNA and protein levels. The results showed that the mRNA expression and protein levels of ELP5 were negatively regulated by miR-149-5p (Fig. 7C-D and S4b-c).
To further investigate whether ELP5 was regulated by miR-149-5p in BC cells via the co-transfection of cells with the miR-149-5p inhibitor and si-ELP5, we measured the protein levels of ELP5, ROCK1, RhoA, and P21 by western blotting. Interestingly, inhibition of ELP5, ROCK1, RhoA, and P21 protein expression induced by the miR-149-5p inhibitor was effectively reversed by ELP5 knockdown (Fig. 7E and S4d). We then sought to determine whether the biological function of miR-149-5p inhibitors in BC cells could also be reversed by ELP5 knockdown, and observed increased proliferation, as indicated by the MTT proliferation assay, colony formation experiments, and EdU assay (Fig. 7F-H and S4e-h). Moreover, increased metastasis and invasion ability indicated by the transwell assay in BC cells mediated by miR-149-5p inhibition were successfully restored by ELP5 knockdown in BC cells (Fig. 7I).
Role of the EGFR-AS1-miR-149-5p-ELP5/PI3K/AKT axis in the metabolic alteration of BC cells and promotion of BC development
Based on these results, we hypothesized that EGFR-AS1 promotes the proliferation, migration, and invasion of BC cells by sponging miR-149-5p; thus, EGFR-AS1 regulates ELP5 expression. To test this hypothesis, we performed an immunohistochemical analysis that showed that the ELP5 protein expression in BC tissues was positively correlated with EGFR-AS1 expression (Fig. 8A). As mentioned above, qRT-PCR further confirmed the expression level of ELP5 and indicated that it was upregulated in 33 BC tissues relative to adjacent normal tissues (Fig. 8B). We then determined whether EGFR-AS1 was co-expressed with ELP5. According to the median expression of ELP5, we sorted the 33 BC tissue samples into a low ELP5 (n = 16) and high ELP5 (n = 16) expression group. We then detected the EGFR-AS1 level in the two groups and found that the level of EGFR-AS1 in the high ELP5 group was significantly higher than that in the low ELP5 group (Fig. 8C). To further analyze the regulatory role of EGFR-AS1 on ELP5, we knocked down EGFR-AS1 expression in MCF-7 and MDA-MB-231 cells using EGFR-AS1 siRNA. The results showed that EGFR-AS1 knockdown decreased ELP5 levels, as indicated by qRT-PCR and western blotting (Fig. 8D and 8E). Thus, we demonstrated the positive regulation of ceRNA ELP5 by EGFR-AS1.
According to previous studies, the EMT-related markers N-CA and Vimentin are both regulated by EGFR-AS1, and the ELP5.PI3K/AKT pathway can also be activated by EGFR-AS1 and ELP5. Through rescue assays and based on the results of further western blotting analysis of ELP5, N-CA, Vimentin, PI3K, P-AKT, and AKT protein abundance, we determined that ectopic ELP5 expression reversed the inhibitory effects of EGFR-AS1 suppression (Fig. 8F).
Role of EGFR-AS1 silencing in the promotion of docetaxel sensitivity in BC cells
Chemotherapy is crucial to the effective treatment of BC; however, it is limited by the development of chemoresistance [7, 8, 35]. Increasing evidence supports the idea that lncRNAs are key regulators of drug resistance [16, 17]. Therefore, we examined the ability of EGFR-AS1 to act as a mechanism of resistance to the targeted drug docetaxel. First, to test the IC50 of docetaxel in MCF-7 and MDA-MB-231 cells, they were treated with a concentration gradient of docetaxel. After 48 h, the IC50 values of docetaxel in MCF-7 and MDA-MB-231 cells were 9.59 nM and 4.771 nM, respectively (Fig. 9A). Next, the concentrations of docetaxel corresponding to the individual IC50s and half the IC50 concentration were added, and the EGFR-AS1 and miR-149-5p levels were assessed by qRT-PCR. In the two cell lines, the level of EGFR-AS1 increased with increasing drug concentration, indicating that the increase in EGFR-AS1 was docetaxel-dependent (Fig. 9B). Moreover, the level of miR-149-5p decreased with increasing drug concentration, thus, demonstrating that the decrease in miR-149-5p was also docetaxel-dependent (Fig. 9C).
Similar results were obtained via western blotting. In the two cell lines, ELP5 was increased in a docetaxel-dependent manner (Fig. 9D and 9E). These results indicated that EGFR-AS1, miR-149-5p, and ELP5 are involved in the resistance to docetaxel. After silencing EGFR-AS1, the IC50s of MCF-7 and MDA-MB-231 cells decreased to 5.06 nM and 2.36 nM, respectively (Fig. 9F). Colony formation and MTT proliferation assays were then conducted following addition of the IC50 concentration of docetaxel to MCF-7 and MDA-MB-231 cells. Cells with lower EGFR-AS1 expression had lower cell viability (Fig. 9G and 9H). These experiments confirmed our hypothesis that silencing EGFR-AS1 can promote sensitivity to docetaxel in BC cells, which provides new insights into its clinical application.
EGFR-AS1 knockdown as an in vivo inhibitor of BC cell tumorigenesis
We then performed an in vivo murine study to identify the role of EGFR-AS1 in BC growth regulation. MCF-7 cells stably transfected with sh-EGFR-AS1, or sh-NC were subcutaneously injected into BALB/c nude mice. After 21 days, the tumors were removed, measured and weighed. The results indicated that the mice injected with EGFR-AS1-silenced MCF-7 cells developed markedly smaller tumors than the control cells (Fig. 10A). Moreover, as can be seen from the tumor growth curve, EGFR-AS1 knockdown significantly suppressed tumor growth compared with the negative control group (Fig. 10B). Similar results were observed for tumor volume and tumor weight (Fig. 10C and 10D). Moreover, following harvesting of tumors qRT-PCR and western blot studies were performed, revealing that the mRNA expression of EGFR-AS1 and ELP5 was decreased (Fig. 10E and 10F), and the expression of ELP5, EMT-related proteins (N-CA, Vimentin, Slug, Snail), and PI3K/AKT pathway-related markers was similarly reduced in the tumors of mice injected with EGFR-AS1-silenced MCF-7 cells (Fig. 10G-I). These results suggested that EGFR-AS1 promotes in vivo BC tumor growth and metastasis.