https://orcid.org/0000-0003-2468-901X
Abstract
YWHAE gene product belongs to the 14-3-3 protein family that mediates signal transduction in plants and mammals. Protein-coding and non-coding RNA (lncRNA) transcripts have been reported for this gene in human. Here, we aimed to functionally characterize YWHAE-encoded lncRNA in colorectal cancer-originated cells. RNA-seq analysis showed that YWHAE gene is upregulated in colorectal cancer specimens. Additionally, bioinformatics analysis suggested that YWHAE lncRNA sponges miR-323a-3p and miR-532-5p that were predicted to target K-Ras 3′UTR sequence. Overexpression of YWHAE lncRNA resulted in upregulation of K-Ras gene expression, while overexpression of both miR-323a-3p and miR-532-5p had an inverse effect, detected by RT-qPCR. Consistently, western blot analysis confirmed that YWHAE lncRNA overexpression upregulated K-Ras/Erk1/2 and PI3K/Akt signaling pathways, while miR-323a-3p and miR-532-5p overexpression suppressed both pathways in HCT116 cells. Furthermore, dual luciferase assay validated the direct interaction of miR-323a-3p and miR-532-5p with K-Ras 3′UTR sequence and supported the sponging effect of YWHAE lncRNA over both miRNAs. These results suggested YWHAE lncRNA as an oncogene that exerts its effect through sponging miR-323a-3p and miR-532-5p and in turn, upregulates K-Ras/Erk1/2 and PI3K/Akt signaling pathways. Consistently, flow cytometry analysis, MTT assay and measuring cyclin D1 gene expression, confirmed the cell cycle stimulatory effect of YWHAE lncRNA, while miR-323a-3p and miR-532-5p showed an inhibitory effect on cell cycle progression. Finally, wound-healing assay supported the cell migratory effect of YWHAE lncRNA in HCT116 cells. This study identified a novel mechanism involving YWHAE-encoded lncRNA, miR-323a-3p and miR-532-5p in regulating HCT116 cell survival and suggested a potential therapeutic avenue for colorectal cancer.
Introduction
In the recent years, long non-coding RNAs have attracted an extensive assessment as potential key players in diverse biological processes (1). LncRNAs have been defined as transcripts longer than 200 nucleotides (2) that are either untranslated into proteins or translated into truncated unfunctional ones (3). These transcripts have been implicated in a wide range of developmental processes and diseases including different types of cancers (4); however, the mechanism by which they act is still largely debatable (1). Many lncRNAs have been identified and distinct mode of actions have been attributed to them, such as functioning as signal, scaffold, guide, enhancer and decoy RNAs (5). Decoy lncRNAs, present decoy binding sites and sequester their corresponding regulatory factors, thereby limiting their availability. These lncRNAs interact with proteins including transcription factors, catalytic proteins and subunits of chromatin modifying complexes or hybridize with other RNA molecules such as miRNAs, thereby modulating gene transcription (6). Decoy lncRNAs that function as miRNA sequesters are known as competing endogenous RNAs (ceRNAs) (7). Such lncRNAs can act either as oncogenic factors through sponging tumor suppressor microRNAs or inversely in several cancer types (8, 9). This mechanism may lead to translational activation of miRNA target genes, affecting cell growth and migration, cell cycle status and cell viability (10, 11).
miRNAs are single-stranded RNA molecules of 18–22 nucleotides in length. These small non-coding RNAs are evolutionary conserved and play an indispensable role in maintaining cellular identity and homeostasis (12). miRNAs regulate gene expression through direct interaction with their specific target transcripts. This process can either completely degrade the target mRNAs or block their translation depending on the degree of interaction between the miRNA seed region (6–8 nucleotide long) and the microRNA recognition site (MRS) found in the target mRNA (13). miRNAs were identified as an essential player in various disease conditions including cancers, where they function either as oncogenes or as tumor suppressors (14, 15).
Recent studies have identified new functions of previously discovered miRNAs that are regulating signaling pathways involved in cancer cell proliferation and migration (16). miR-532-5p have been reported to target the K-Ras/MEK/Erk1/2 signaling pathway in lung adenocarcinoma cells (17) and inhibited colorectal cancer progression through the suppression of PI3K/Akt signaling pathway (18). Furthermore, miR-323a-3p has been shown to exert a tumor-suppressive activity in bladder cancer partly through downregulation of the PI3K/Akt signaling pathway (19). Historically, K-Ras/MEK/Erk1/2 and PI3K/Akt signaling pathways have been shown to maintain cell viability (20, 21) and were involved in tumor cell survival (22). Additionally, several studies have confirmed the cross-talk mechanism between these two pathways where a constitutively activated mutant K-Ras can simultaneously promote both Erk1/2 and Akt phosphorylation (23–26). Both pathways are known to act synergistically to promote cell survival, proliferation and cell motility through activation of the cell cycle regulator cyclin D (27).
In the current study, we discovered a novel mechanism involving three key players: YWHAE-encoded lncRNA, miR-323a-3p and miR-532-5p that act together to control HCT116 colorectal cancer cell survival through targeting both K-Ras/Erk1/2 and PI3K/Akt signaling pathways (Fig. 6).
Results
YWHAE gene is upregulated in colon cancer tissues, inversely to miR-323a-3p and miR-532-5p
In order to gain an overview about the correlation between YWHAE gene, miR-323a-3p and miR-532-5p in colorectal cancer, we investigated their expression pattern between colon cancer and colon normal tissues. The data collected from RNA-seq atlas (Fig. 1A) (28) and TCGA (COAD/READ) (Fig. 1B) (29) databases confirmed that YWHAE gene is upregulated in colon cancer tissues compared to that of the normal ones (P < 0.0001).YWHAE gene encodes two transcript variants, one is protein coding (14-3-3E), however, the second is non-coding due to the presence of a pre-mature stop codon in the second exon (E1’) (Fig. 1C). In our study, we were interested in the elaboration of YWHAE ncRNA transcript and its relation with colon cancer. For this reason, we amplified both transcript variants (Fig. 1D), and the non-coding RNA (1753 bp) was extracted from the gel and cloned in an expression vector for ectopic expression in HCT116 cell line. Consequently, the expression pattern of both YWHAE transcript variants was investigated between colorectal normal tissues and colorectal cancer tissues. Results showed that both transcript variants have higher expression level in colorectal cancer tissues compared to the normal ones (Fig. 1E). Additionally, data analysis collected from starbase (30) and Gene Expression Omnibus (GEO) (31) showed that miR-323a-3p (Fig. 1F) and miR-532-5p (Fig. 1G) expression levels were higher in colon normal tissues compared to colon adenocarcinoma tissues (P < 0.0001 and P < 0.001, respectively). Furthermore, in order to investigate the correlation between YWHAE-encoded lncRNA, YWHAE mRNA, K-Ras mRNA and the two studied miRNAs, RT-qPCR was performed against these five transcripts in three different colorectal cancer cell lines (SW480, HT29 and HCT116) (Fig. 1H/I). Results indicated that both YWHAE-encoded lncRNA and K-Ras transcripts proportionally increased along the different stages of colorectal cancer cells where they showed the highest expression level in HCT116 cell line; however, this proportionality was absent between YWHAE mRNA and K-Ras transcripts where YWHAE mRNA showed the highest expression level in SW480 cell line (Fig. 1H). On the other hand, miR-323a-3p and miR-532-5p showed the lowest expression level in HCT116 cell line (Fig. 1I). To further investigate the correlation between these two miRNAs and YWHAE gene or K-Ras gene, analyzed data were collected from starbase database, which show a negative correlation between YWHAE gene and both miRNAs (miR-323a-3p and miR-532-5p) (P < 0.05) (Supplementary Material, Fig. A/E). Furthermore, additional data collected from the same database supported the negative correlation between K-Ras gene and the two mentioned miRNAs (P < 0.05) (Supplementary Material, Fig. B/E) (30). Additionally, starbase data showed positive correlation between YWHAE gene and K-Ras gene in colorectal cancer tissues (P < 0.001) (Supplementary Material, Fig. C/E) (30).
RNA-Seq data showing the differential expression of YWHAE gene, miR-323a-3p and miR-532-5p between colon cancer tissues and colon normal tissues. (A) RNA-Seq Atlas data showing the differential expression of YWHAE gene between colon cancer tissues (n = 1839 samples) and colon normal tissues (n = 293 samples). Unpaired t-test was used for statistical analysis. (B) TCGA data showing the differential expression of YWHAE gene between colorectal (COAD-READ) cancer tissues (n = 222 samples) and colorectal normal tissues (n = 22 samples). (C) Schematic representation showing YWHAE transcript variant 1 (mRNA) and transcript variant 2 (ncRNA). Gray boxes represent the exonic sequences, and the straight lines represent the intronic sequences. Transcript variant 2 (ncRNA) contains an additional exon (the second one from the left) that is 33 bp in length and holds an immature stop codon. The start and mature stop codons are indicated by black arrows. The real-time PCR primer reverse 1 (R1) was used for specific detection of YWHAE lncRNA; however, R2 primer was designed at the exon-exon junction of E1 and E2 for specific detection of YWHAE mRNA. (D) Amplified PCR products of both YWHAE transcript variants using specific primers along the first and the last exonic sequences (Table 1). The ncRNA and the mRNA are 1753 bp and 1720 bp in length, respectively. (E) Experimental tissue analysis showing the differential expression of both YWHAE transcript variants between colorectal cancer tissues (n = 10) and colorectal normal tissues (n = 6). (F) Figure adapted from starbase data showing the relative expression of miR-323a-3p in colon adenocarcinoma (COAD) tissues (n = 450) compared to colon normal tissues (n = 8). (G) GEO data showing the relative expression of miR-532-5p in colorectal adenocarcinoma tissues (n = 33) compared to colorectal normal tissues (n = 15). (H) Relative expression of YWHAE lncRNA, YWHAE mRNA and K-Ras gene in three different stages of colorectal cancer cell lines, detected by RT-qPCR. (I) Relative expression of miR-323a-3p and miR-532-5p in three different stages of colorectal cancer cell lines normalized against the expression level of U48, detected by RT-qPCR. *P < 0.05. Error bars represent SEM.
YWHAE-encoded lncRNA promotes activation of K-Ras/Erk1/2 and PI3K/Akt signaling pathways in HCT116 cells, inversely to miR-323a-3p and miR-532-5p
Subsequently, we tested whether YWHAE-encoded lncRNA plays an important role in K-Ras/Erk1/2 and PI3K/Akt signaling pathways. Firstly, YWHAE-encoded lncRNA was significantly downregulated in HCT116 cells (Fig. 2B) using shRNA against its specific exon E1’ (Fig. 2A) then, RT-qPCR was performed against K-Ras gene (Fig. 2C). Results showed that YWHAE lncRNA downregulation induced a significant downregulation of K-Ras gene at the mRNA level (Fig. 2C). Additionally, YWHAE-encoded lncRNA overexpression (Fig. 2D) significantly upregulated K-Ras gene at the mRNA level (Fig. 2E). Erk1/2 protein is well known to act downstream of the K-Ras protein and that K-Ras/B-Raf/MEK/Erk1/2 is one of the widely known MAPK signaling pathways (32). For this reason, western blot analysis was performed against phospho-Erk1/2 to confirm the effect of YWHAE-encoded lncRNA on K-Ras/MEK/Erk1/2 signaling pathway. Our results confirmed that phosho-Erk1/2 has been significantly upregulated after YWHAE-encoded lncRNA overexpression in HCT116 cells compared to cells transfected with an empty expression vector (mock) (Fig. 2F/G). Additionally, several studies have shown that there is a cross-talk between K-Ras/Erk1/2 and PI3K/Akt signaling pathways and that a constitutively activated mutant K-Ras can bypass the activation of Erk1/2 in order to activate PI3K/Akt signaling pathway (25, 26). To confirm this phenomenon, western blot analysis was performed, and the results confirmed that phospho-Akt has been significantly upregulated after YWHAE-encoded lncRNA overexpression compared to cells transfected with the mock vector (Fig. 2F/G). Thereafter, we investigated whether miR-323a-3p and miR-532-5p exert an effect on K-Ras/Erk1/2 and PI3K/Akt signaling pathways in HCT116 cells. Firstly, miR-323a-3p and miR-532-5p were significantly overexpressed in HCT116 cell line (Fig. 2H) then, RT-qPCR was performed against K-Ras gene. Results indicated that overexpression of both miRNAs significantly downregulated K-Ras gene at the mRNA level (Fig. 2I). Bioinformatics analysis predicted three MRSs for miR-323a-3p and one MRS corresponding to miR-532-5p (Fig. 2J). For this purpose, we used several bioinformatics miRNA site prediction tools including MiRDB (microRNA Database) (33) and Targetscan (34) in order to precisely determine the nucleotide interactions between the miRNAs’ seed region and K-Ras 3′UTR (Fig. 2K). Later on, luciferase assay was performed to investigate the direct interaction between miR-323a-3p and miR-532-5p with K-Ras 3′UTR. Co-transfection of HEK293-T cells with expression vectors containing miR-323a or miR-532 precursor along with psi-check expression vector holding K-Ras 3′UTR cloned downstream of the luciferase gene, showed a significant decrease in the luciferase activity compared to mock transfected cells (Fig. 2L). In order to confirm the specificity of the miRNAs-K-Ras 3′UTR interaction, AKT1 and AKT2 3′UTRs were cloned downstream of the luciferase gene and co-transfected with either miR-323a or miR-532. AKT1 and AKT2 3′UTRs were used as non-targets since both miRNAs have no predicted MRSs in their 3′UTR sequences based on MiRDB and Targetscan softwares. Co-transfection of both miRNAs with either AKT1 or AKT2 3′UTR showed no significant change in the luciferase activity compared to mock transfected cells (Fig. 2L). Furthermore, western blot analysis was performed in order to confirm the effect of both miRNAs on K-Ras/Erk1/2 and PI3K/Akt signaling pathways. miR-323a-3p and miR-532-5p overexpression in HCT116 cells significantly downregulated phospho-Akt and phospho-Erk1/2 in comparison to mock transfected cells (Fig. 2M/N).
The effect of YWHAE lncRNA, miR-323a-3p and miR-532-5p overexpression on K-Ras/Erk1/2 and PI3K/Akt signaling pathways. (A) Schematic representation showing the shRNA’s secondary structure. The shRNA’s loop will be processed by dicer enzyme at the DCS leading to the production of the sense strand that will hybridize with E1’ of the lncRNA. (B) RT-qPCR result showing successful downregulation of YWHAE lncRNA after transfection of HCT116 cells with shRNA. (C) RT-qPCR result showing significant downregulation of K-Ras mRNA after transfection of HCT116 cells with shRNA. (D) RT-qPCR result showing successful overexpression of YWHAE lncRNA in HCT116 cells. (E) RT-qPCR result showing K-Ras upregulation at the mRNA level following YWHAE lncRNA overexpression. (F) YWHAE lncRNA overexpression significantly upregulated phospo-Erk1/2 and phospho-Akt proteins, detected by western blot analysis. (G) Protein band quantification was performed using Image-J software in order to ensure the result’s significance. (H) Successful miR-323a-3p and miR-532-5p overexpression in HCT116 cells shown by RT-qPCR. (I) RT-qPCR results showed that miR-323a-3p and miR-532-5p overexpression was followed by significant downregulation of K-Ras gene expression. (J) Schematic representation showing the MRSs of miR-323a-3p (black boxes) and miR-532-5p (white box) in K-Ras 3′UTR. The dashed blue arrows mark the length of K-Ras 3′UTR. (K) The numbers on top of each sequence represent the distance from the first nucleotide of K-Ras 3′UTR sequence. (L) Summarized luciferase assay result supporting the direct interaction between miR-323a-3p or miR-532-5p with K-Ras 3′UTR sequence. Co-transfection of HEK293-T cells with either miR-323a or miR-532 cassettes along with the luciferase K-Ras 3′UTR sequence construct, significantly downregulated the luciferase activity. However, co-transfection of both miRNA cassettes with the non-targets’ 3′UTR, made no significant change in the luciferase activity. (M) Overexpression of miR-323a-3p or miR-532-5p significantly downregulated the phosphorylation level of both Akt and Erk1/2 proteins, detected by western blot analysis. (N) Protein band quantification was performed using ImageJ software in order to ensure the result’s significance. ns means non-significant. *P < 0.05. Error bars represent SEM.
YWHAE-encoded lncRNA probably interacts and sequesters miR-323a-3p and miR-532-5p
According to the bioinformatics analysis, one miRNA sponging site (MSS) was predicted for each of the studied miRNAs in YWHAE-encoded lncRNA sequence (Fig. 3A). For this purpose, we used several bioinformatics MSS prediction tools including LncTar (35), IntaRNA (36) and most importantly RNA Hybrid (37) in order to predict the extension of the nucleotide interactions between the miRNAs and YWHAE-encoded lncRNA (Fig. 3B). Furthermore, we performed luciferase assay in order to investigate the direct interaction between miR-323a-3p or miR-532-5p with YWHAE lncRNA.
Sponging effect of YWHAE lncRNA over miR-323a-3p and miR-532-5p. (A) Schematic representation showing miR-323a-3p (black) and miR-532-5p (white) sponging sites (MSSs) within the 6th exon of YWHAE lncRNA sequence. (B) Bioinformatics prediction of the base pairing between the two studied miRNAs and YWHAE lncRNA sequences, using RNA hybrid software. (C) Co-transfection of HEK293-T cells with either miR-323a or miR-532 precursor and the luciferase construct fused with YWHAE lncRNA, resulted in significant downregulation of the luciferase activity in comparison to mock transfected cells. However, co-transfection of the miRNAs’ precursor with the non-targets’ 3′UTR (AKT1 and AKT2) made no significant change in the luciferase activity in comparison to mock transfected cells. (D) Co-transfection of HEK293-T cells with either YWHAE lncRNA or YWHAE mRNA and the luciferase construct fused with K-Ras 3′UTR, resulted in significant upregulation of the luciferase activity in comparison to mock transfected cells. However, co-transfection of both YWHAE transcript variants with the non-target 3′UTR (APPL1), made no significant change in the luciferase activity in comparison to mock-transfected cells. ns means non-significant. *P < 0.05. Error bars represent SEM.
Overexpression of miR-532-5p or miR-323a-3p in HEK293-T cells, along with a construct holding the lncRNA-cloned downstream of the luciferase gene, showed a significant decrease in the luciferase activity compared to mock transfected cells (Fig. 3C). In order to confirm the specificity of the miRNAs–LncRNA interaction, AKT1 and AKT2 3′UTRs were also used as non-target controls. Co-transfection of both miRNAs with either AKT1 or AKT2 3′UTR showed no significant change in the luciferase activity compared to mock transfected cells (Fig. 3C). Additionally, further luciferase assay was performed in order to confirm the positive correlation between YWHAE gene and K-Ras gene. Co-transfection of either YWHAE lncRNA or YWHAE mRNA with K-Ras 3′UTR significantly upregulated the luciferase activity compared to mock transfected cells. Although both YWHAE transcript variants showed significant effect on the K-Ras luciferase activity, however, YWHAE lncRNA exerted higher effect than that of YWHAE mRNA (Fig. 3D). Furthermore, data adapted from starbase database showed no correlation between YWHAE gene and APPL1 gene (Supplementary Material, Fig. D). For this purpose, APPL1 3′UTR was used as non-target control to confirm the specific effect of YWHAE gene on K-Ras gene. Therefore, co-transfection of both YWHAE transcript variants with the non-target 3′UTR (APPL1) showed no significant change in the luciferase activity compared to mock transfected cells (Fig. 3D).
YWHAE-encoded lncRNA overexpression enhances cell cycle progression in HCT116 cells, inversely to miR-323a-3p and miR-532-5p
YWHAE-encoded lncRNA overexpression in HCT116 cells resulted in a significant reduction in the sub-G1 (Freq. sub-G1 = 4.12%) and G1 (Freq. G1 = 43.3%) cell populations compared to mock-transfected cells (Freq. sub-G1 = 11.69, Freq. G1 = 47.44). Furthermore, lncRNA overexpression resulted in a significant increase in the S (Freq. S = 34.2%) and G2 (Freq.G2 = 18.38%) cell populations compared to mock transfected cells (Freq. S = 27.8%, Freq. G2 = 13.07) (Fig. 4A/B). MTT assay was also performed in order to strengthen our results. LncRNA overexpression significantly upregulated HCT116 cell viability and proliferation in comparison to mock-transfected cells (Fig. 4C). The effect of YWHAE-encoded lncRNA on cell cycle was also investigated at the molecular level. Cyclin D1 is a cell cycle regulatory molecule responsible for cell cycle progression from G1 to S phase (38). YWHAE lncRNA overexpression significantly upregulated cyclin D1 expression at the mRNA level (Fig. 4D). Additionally, western blot analysis was performed against cyclin D1, and the results confirmed that cyclin D1 was significantly upregulated at the protein level following YWHAE-encoded lncRNA overexpression in comparison to mock-transfected cells (Fig. 4E/F). Furthermore, our results showed that both miR-323a-3p and miR-532-5p exerted an inhibitory effect on K-Ras/Erk1/2 and PI3K/Akt signaling pathways. For this purpose, we investigated the effect of both miRNAs overexpression on HCT116 cell cycle status. miR-323a-3p or miR-532-5p overexpression significantly upregulated the frequency of sub-G1 (Freq. sub-G1 = 15.68 and 14.92, respectively) and G1 (Freq. G1 = 50.7 and 49.32, respectively) cell populations compared to the mock control (Freq. sub-G1 = 3.16, Freq. G1 = 43.24). Additionally, overexpression of both miRNAs significantly downregulated the frequency of S (Freq. S = 25.85 and 28.08, respectively) and G2 (Freq. G2 = 7.77 and 7.68, respectively) cell populations compared to the mock control (Freq. S = 37.38, Freq. G2 = 16.22) (Fig. 4G/H). Consistently, MTT assay showed that HCT116 cell viability and proliferation were significantly reduced after overexpression of both miRNAs compared to mock transfected cells (Fig. 4I). For investigating the effect of both miRNAs at the molecular level of the cell cycle, RT-qPCR and western blot analysis were also performed against cyclin D1. Results showed that overexpression of both miRNAs significantly downregulated cyclin D1 at the mRNA (Fig. 4J) and protein (Fig. 4K/L) levels.
The effect of YWHAE lncRNA, miR-323a-3p and miR-532-5p overexpression on cell cycle progression and cell viability. (A) Cell cycle graphs after transfection of HCT116 cells with mock or YWHAE-encoded lncRNA. (B) Transfection of HCT116 cells with YWHAE-encoded lncRNA significantly decreased the frequency of Sub-G1 and G1 phases, and significantly increased the frequency of S and G2 phases. (C) MTT assay showed that YWHAE lncRNA overexpression increased the cell viability and proliferation of HCT116 cells compared to the mock control. (D) YWHAE lncRNA overexpression significantly induced upregulation of cyclin D1 at the mRNA level in comparison to mock transfected cells. (E) Western blot analysis showed that YWHAE lncRNA overexpression significantly upregulated cyclin D1 at the protein level in comparison to mock transfected cells. (F) Protein band quantification was performed using ImageJ software in order to confirm the result’s significance. (G) Cell cycle graphs of mock, miR-323a-3p and miR-532-5p transfected HCT116 cells. (H) Overexpression of miR-323a-3p or miR-532-5p in HCT116 cells significantly increased the frequency of cells in Sub-G1 and G1 phases and significantly decreased the frequency of cells in S and G2 phases. (I) MTT assay showed that miR-323a-3p or miR-532-5p overexpression significantly decreased the cell viability and proliferation of HCT116 cells compared to the mock control. (J) miR-323a-3p or miR-532-5p overexpression significantly induced downregulation of cyclin D1 at the mRNA level in comparison to mock-transfected cells. (K) Western blot analysis showed that miR-323a-3p or miR-532-5p overexpression significantly downregulated cyclin D1 at the protein level compared to mock transfected cells. (L) Protein band quantification was performed using ImageJ software in order to confirm the result’s significance. *P < 0.05. Error bars represent SEM.
YWHAE-encoded lncRNA enhances HCT116 cell migration
YWHAE lncRNA was shown to have the highest expression level in HCT116 cells compared to SW480 and HT29 cell lines (Fig. 1G). HCT116 cells transfected with an empty vector (mock) showed significant migration only after 72 h of transfection; however, transfection of HCT116 cells with YWHAE lncRNA showed significant migration after 24 h and progressively enhanced until 72 h after transfection (Fig. 5A/B). Furthermore, YWHAE lncRNA overexpression exerted a significant migratory effect on HCT116 cells in comparison to mock transfected cells after 24, 48 and 72 h of transfection (Fig. 5C).
The effect of YWHAE lncRNA overexpression on HCT116 cell migration. (A) Wound healing assay showing the cell migration difference between mock and YWHAE lncRNA transfected cells after 24, 48 and 72 h of transfection. (B) Mock-transfected cells showed significant cell migration after 72 h of transfection; however, YWHAE lncRNA transfected cells showed significant cell migration after 24 h and started to progressively increase until 72 h after transfection. (C) YWHAE lncRNA exerted a significant migratory effect on HCT116 in comparison to mock-transfected cells after 24, 48 and 72 h of transfection. The percentage of the wound area was measured using ImageJ software. ns means non-significant. *P < 0.05. Error bars represent SEM.
Discussion
Recent studies have suggested that lncRNAs may act as competing endogenous RNAs through direct interaction with miRNAs. This RNA–RNA hybridization process reduces the availability of the small non-coding RNAs contributing to the modulation of miRNA target gene expression (7). A huge number of miRNAs and lncRNAs have been shown to be upregulated or downregulated in different cancer types (39, 40) participating in the formation of diverse disease conditions and functioning either as oncogenes or as tumor suppressors (10, 11). YWHAE gene encodes two validated transcript variants, one is protein coding (14-3-3E) and the second is a non-coding transcript (Fig. 1C/D). A recent study has shown that silencing YWHAE gene expression suppresses the proliferation of colon cancer cells through induction of differential gene expression. These genes were mainly involved in cell proliferation, apoptosis and cell cycle progression (41). Our RNA-seq data analysis showed that YWHAE gene has higher expression level in colon cancer tissues in comparison to colon normal tissues (Fig. 1A/B). Additionally, our experimental study confirmed that both YWHAE transcript variants have higher expression level in colorectal cancer tissues compared to the normal ones (Fig. 1E). However, miR-323a-3p and miR-532-5p showed an inverse expression pattern (Fig. 1F/G). Additional RNA-seq data confirmed the negative correlation of YWHAE gene and K-Ras gene with respect to miR-323a-3p and miR-532-5p and supported the positive correlation between YWHAE gene and K-Ras gene (Supplementary Material, Fig. A/B/C/E). Furthermore, we have confirmed the positive proportionality between YWHAE-encoded lncRNA and K-Ras gene where they gradually increase during the different stages of colorectal cancer cell lines (Fig. 1H); however, this proportionality was not observed between YWHAE mRNA and K-Ras gene supporting the specific correlation between YWHAE-encoded lncRNA and K-Ras gene. YWHAE-encoded lncRNA and K-Ras gene showed the highest expression level in HCT116 cell line inversely to that of miR-323a-3p and miR-532-5p (Fig. 1H/I).
HCT116 is classified as a stage 4 or Dukes’ D colorectal cancer cell line according to the TNM and Dukes’ staging systems, respectively, and it is known to express a constitutively activated mutant (G13D) K-Ras (42). Previous studies confirmed that K-Ras/B-Raf/MEK/Erk1/2 signaling pathway (shortly Erk pathway) has been implicated in the formation of different cancer types due to its oncogenic features (43, 44). Additionally, a wide range of studies confirmed the cross-talk mechanism between K-Ras/Erk1/2 and PI3K/Akt signaling pathways where a constitutively activated mutant K-Ras can bypass the Erk signal to the PI3K/Akt signal (25, 26, 45). Both pathways functionally co-regulate the same transcription factors and promote cell cycle progression and cell survival through the activity of cyclin D (27).
A recent study showed that miR-532-5p exerts a tumor suppressive activity in lung adenocarcinoma cells through targeting K-Ras oncogene (17). Additionally, this miRNA inhibited colorectal cancer cell progression through suppression of the PI3K/Akt signaling pathway (18). Another new study performed on miR-323a-3p confirming its inhibitory effect on bladder cancer cell migration and invasion partly through suppression of the PI3K/Akt signaling pathway (19). In our study, we have shown that YWHAE-encoded lncRNA induced upregulation of both K-Ras/Erk1/2 and PI3K/Akt signaling pathways (Fig. 2D/E/F/G) and silencing its activity promoted downregulation of K-Ras gene at the mRNA level (Fig. 2A/B/C). On the contrary, miR-323a-3p and miR-532-5p overexpression exerted an inverse activity where they promoted downregulation of both signaling pathways (Fig. 2I/M/N). In order to ensure the direct effect of both miRNAs on K-Ras gene, we firstly used bioinformatics prediction tools to check for the existence of miRNA recognition sites in K-Ras 3′UTR sequence where we got fruitful predictive results (Fig. 2J/K). Then, luciferase assay confirmed the direct interaction between both miRNAs and K-Ras 3′UTR sequence (Fig. 2K). Since our study showed that YWHAE-encoded lncRNA promoted a completely inverse mode of action in comparison to miR-323a-3p and miR-532-5p on Erk and Akt signaling pathways, we were interested to investigate if our lncRNA is regulating the two signaling pathways through targeting both miRNAs. For this purpose, luciferase assay was performed to ensure the direct interaction between YWHAE lncRNA and the studied miRNAs, and the results were extremely consistent with our predictions (Fig. 3A/B/C). Additionally, another luciferase assay simultaneously confirmed the positive correlation between YWHAE-encoded lncRNA and K-Ras gene and supported the sponging effect of YWHAE-encoded lncRNA over both miRNAs. Although both YWHAE transcript variants share common MSSs in their 3′UTR (Fig. 1C) and both of them enhanced the K-Ras luciferase activity; however, YWHAE-encoded lncRNA showed higher significant effect than that exerted by YWHAE mRNA (Fig. 3D). This indicates that the major sponging activity is mainly mediated by YWHAE-encoded lncRNA probably due to the difference in the overall secondary structure of both transcript variants.
Erk1/2 and Akt signaling pathways are known to promote cell cycle progression (20, 21). Accordingly, flow cytometry analysis and MTT assay were performed. Our results showed that overexpression of YWHAE-encoded lncRNA significantly enhanced HCT116 cell proliferation and viability (Fig. 4A/B/C). However, miR-323a-3p and miR-532-5p overexpression significantly inhibited cell proliferation and viability (Fig. 4G/H/I). Cyclin D1 is a cell cycle regulatory molecule required for the progression from G1 to S phase (38) and is widely known to act downstream of both K-Ras/Erk and PI3K/Akt signaling pathways (27). Since a significant modulation was observed at the G1 and S phases of the cell cycle after YWHAE-encoded lncRNA and both miRNAs overexpression, we were interested to study the cell cycle status at the molecular level. Overexpression of YWHAE-encoded lncRNA significantly upregulated cyclin D1 at the mRNA and protein levels (Fig. 4D/E/F). However, overexpression of miR-323a-3p and miR-532-5p significantly downregulated cyclin D1 at both levels (Fig. 4J/K/L).
Additionally, Erk1/2 and Akt signaling pathways were shown to promote cell migration in a cancer-specific type, and the inhibition of these two pathways will almost abolish this mechanism (46–49). Correspondingly, we were interested to investigate the effect of YWHAE-encoded lncRNA on HCT116 cell migration. Results confirmed that YWHAE lncRNA overexpression significantly and progressively promoted HCT116 cell migration after 24, 48 and 72 h of transfection (Fig. 5A/B/C). This study provides a novel mechanism consisting of YWHAE lncRNA, miR-323a-3p and miR-532-5p, K-Ras/Erk1/2, PI3K/Akt and cyclin D1, which is responsible for the regulation of colorectal cancer cell survival and reveals a new pathway by which colorectal cancer cells can nourish themselves (Fig. 6). Our discovery introduces YWHAE-encoded lncRNA as a new key player that could be manipulated to control colorectal cancer progression.
Schematic representation summarizing the role of YWHAE-encoded lncRNA, miR-323a-3p and miR-532-5p in colorectal cancer cell proliferation and migration. YWHAE lncRNA directly interacts and sequesters miR-323a-3p and miR-532-5p. Both miRNAs will not have the accessibility to hybridize with K-Ras 3′UTR. K-Ras mRNA will be freely available to be translated to its corresponding protein. Constitutively activated mutant K-Ras protein will activate PI3K/Akt and Erk1/2 signaling pathways through a phosphorylation process. Activated phosphorylated Akt and Erk1/2 enhance cyclin D1 expression. Cyclin D1 expression promotes cell proliferation and migration.
Materials and Methods
Bioinformatics analysis
YWHAE mRNA (NM_006761.4) and YWHAE ncRNA (NR_024058.1) sequences were collected and analyzed from the NCBI database. RNA-seq data were collected and analyzed from RNA-seq atlas (28), GEO (31, 41), TCGA (29) and starbase (30) databases in order to investigate the relative expression of YWHAE gene, miR-323a-3p and miR-532-5p between colon cancer and colon normal tissues. The data was analyzed as a mean value of several samples and presented as log2 relative expression for each tissue type (Fig. 1A/B/F/G). starbase database was also used to study the correlation between YWHAE, KRas, APPL1, miR-323a-3p and miR-532-5p (Supplementary Material, Fig. A/B/C/D) (30). MiRDB (33) and Targetscan (34) softwares were used to predict the MRSs in K-Ras 3′UTR sequence (Fig. 2J/K). LncTar (35), IntaRNA (36) and most importantly RNA hybrid (37) were used to predict the MSSs in YWHAE-encoded lncRNA sequence (Fig. 3A/B).
Tissue samples
A total of 6 colorectal normal tissues and 10 colorectal cancer tissues were obtained from Talighani’s Hospital, Tehran, Iran, and were stored at −80°C. Total RNA was extracted using RiboEx reagent (Krishgen Biosystems, India) then cDNA was synthesized according to the manufacturer’s protocol (Invitrogen).
Vectors construction
For cloning YWHAE lncRNA, total RNA was extracted from colorectal cancer tissues using RiboEx reagent (Krishgen Biosystems, India) then cDNA was synthesized according to the manufacturer’s protocol (Invitrogen). Specific YWHAE gene cloning primers were used to amplify both YWHAE transcript variants (Fig. 1C/D) (Table 1). After that, ~ 2 Kbp of YWHAE lncRNA was extracted and purified from the gel and cloned into pTG19-T vector using T4 Ligase (Fermentas, USA) later on, YWHAE lncRNA was sub-cloned into pcDNA3.1(+) expression vector using EcoRI and HindIII restriction enzymes (Takara, Japan). Concerning miR-323a-3p and miR-532-5p cloning, genomic DNA was extracted from the blood of a healthy individual according to the manufacturer’s protocol (molecular cloning, Green and Sambrook) and used as a template for PCR amplification. Specific cloning primers were used in order to amplify both miRNA hairpin precursors (pre-miR-323a and pre-miR-532) (Table 1). The amplified miRNA precursors were extracted and purified from the gel and cloned into pTG19-T vector using T4 Ligase (Fermentas, USA), and after that miR-323a-3p and miR-532-5p precursors were sub-cloned into pEGFP-C1 expression vector using Pst1/Kpn1 and Sal1/EcoR1 restriction enzymes, respectively. For performing luciferase assay, YWHAE lncRNA was sub-cloned from pcDNA3.1(+) vector into psi-check2 vector using Xho1 and Not1 enzymes. K-Ras, AKT1, AKT2 and APPL1 3′UTRs were amplified using specific cloning primers (Table 1). Due to the large size of K-Ras 3′UTR (~5Kbp), two separate fragments (1st fragment, 340 bp; 2nd fragment, 1686 bp) specifically holding MRSs for both miRNAs were amplified, and later on were sewed to get ~2Kbp 3′UTR. The cloning primers carry restriction enzyme sites for Xho1 and Not1 at their both ends. After PCR amplification of the mentioned 3′UTRs, the PCR products were digested with Xho1 and Not1 restriction enzymes and directly cloned into psi-check2 vector. The vector constructs were sequenced to confirm their correct orientation and the absence of mutations.
List of the cloning primers
| Cloning primer’s name | Primer sequence, 5′ to 3′ |
|---|---|
| YWHAE-encoded LncRNA | Forward: GGAAGTCCCGGATTGAGGCG Reverse: GATTCTGTATCTGTGGCTCCAGTCAG |
| K-Ras 3′UTR | Forward (1): ACTCGAGGGTGAGGGAGATCCGACAATACAGA Reverse (1): CTCATAAGCACTGCAGTTCCGCTAACAGTCTGCATGGAGCAGG Forward (2): GGAACTGCAGTGCTTATGAGGGG Reverse (2): AGCGGCCGCGCTGCTGCTACCTTTGGGC |
| AKT1 3′UTR | Forward: ACTCGAGGGACTGCGCTGGACGATAGC Reverse: AGCGGCCGCTGAAGAATTTGGAGGGAAGGTTCC |
| AKT2 3′UTR | Forward: ACTCGAGCAGCCTCCAGCCTCACCTTTG Reverse: AGCGGCCGCGCTGTGCCCACACTACGAGACC |
| APPL1 3′UTR | Forward: GCCAGCATTTAAGTAGTCATGGC Reverse: CTAGCTCTTAGGATCTCTCAG |
| Pre-miR-323a-3p | Forward: GTGGCTGGTAGTTCATGATGCGC Reverse: CAAGCCAGAGCAGTGTGCTGAG |
| Pre-miR-532-5p | Forward: GACCTGCATGCCAGATTCTAAGC Reverse: GGGATGGCTGCTAACAGTGTACC |
| shRNA | Forward: CGCGGATCCCTGCGGGTTCAGTTCCAGAGCACAACTACTACCAACAGTGC Reverse: CCCAAGCTTTTCCAAAAAACTGCGGGTTCCTTTCCAGAGCACTGTTGGTAGTAG |
| Cloning primer’s name | Primer sequence, 5′ to 3′ |
|---|---|
| YWHAE-encoded LncRNA | Forward: GGAAGTCCCGGATTGAGGCG Reverse: GATTCTGTATCTGTGGCTCCAGTCAG |
| K-Ras 3′UTR | Forward (1): ACTCGAGGGTGAGGGAGATCCGACAATACAGA Reverse (1): CTCATAAGCACTGCAGTTCCGCTAACAGTCTGCATGGAGCAGG Forward (2): GGAACTGCAGTGCTTATGAGGGG Reverse (2): AGCGGCCGCGCTGCTGCTACCTTTGGGC |
| AKT1 3′UTR | Forward: ACTCGAGGGACTGCGCTGGACGATAGC Reverse: AGCGGCCGCTGAAGAATTTGGAGGGAAGGTTCC |
| AKT2 3′UTR | Forward: ACTCGAGCAGCCTCCAGCCTCACCTTTG Reverse: AGCGGCCGCGCTGTGCCCACACTACGAGACC |
| APPL1 3′UTR | Forward: GCCAGCATTTAAGTAGTCATGGC Reverse: CTAGCTCTTAGGATCTCTCAG |
| Pre-miR-323a-3p | Forward: GTGGCTGGTAGTTCATGATGCGC Reverse: CAAGCCAGAGCAGTGTGCTGAG |
| Pre-miR-532-5p | Forward: GACCTGCATGCCAGATTCTAAGC Reverse: GGGATGGCTGCTAACAGTGTACC |
| shRNA | Forward: CGCGGATCCCTGCGGGTTCAGTTCCAGAGCACAACTACTACCAACAGTGC Reverse: CCCAAGCTTTTCCAAAAAACTGCGGGTTCCTTTCCAGAGCACTGTTGGTAGTAG |
List of the cloning primers
| Cloning primer’s name | Primer sequence, 5′ to 3′ |
|---|---|
| YWHAE-encoded LncRNA | Forward: GGAAGTCCCGGATTGAGGCG Reverse: GATTCTGTATCTGTGGCTCCAGTCAG |
| K-Ras 3′UTR | Forward (1): ACTCGAGGGTGAGGGAGATCCGACAATACAGA Reverse (1): CTCATAAGCACTGCAGTTCCGCTAACAGTCTGCATGGAGCAGG Forward (2): GGAACTGCAGTGCTTATGAGGGG Reverse (2): AGCGGCCGCGCTGCTGCTACCTTTGGGC |
| AKT1 3′UTR | Forward: ACTCGAGGGACTGCGCTGGACGATAGC Reverse: AGCGGCCGCTGAAGAATTTGGAGGGAAGGTTCC |
| AKT2 3′UTR | Forward: ACTCGAGCAGCCTCCAGCCTCACCTTTG Reverse: AGCGGCCGCGCTGTGCCCACACTACGAGACC |
| APPL1 3′UTR | Forward: GCCAGCATTTAAGTAGTCATGGC Reverse: CTAGCTCTTAGGATCTCTCAG |
| Pre-miR-323a-3p | Forward: GTGGCTGGTAGTTCATGATGCGC Reverse: CAAGCCAGAGCAGTGTGCTGAG |
| Pre-miR-532-5p | Forward: GACCTGCATGCCAGATTCTAAGC Reverse: GGGATGGCTGCTAACAGTGTACC |
| shRNA | Forward: CGCGGATCCCTGCGGGTTCAGTTCCAGAGCACAACTACTACCAACAGTGC Reverse: CCCAAGCTTTTCCAAAAAACTGCGGGTTCCTTTCCAGAGCACTGTTGGTAGTAG |
| Cloning primer’s name | Primer sequence, 5′ to 3′ |
|---|---|
| YWHAE-encoded LncRNA | Forward: GGAAGTCCCGGATTGAGGCG Reverse: GATTCTGTATCTGTGGCTCCAGTCAG |
| K-Ras 3′UTR | Forward (1): ACTCGAGGGTGAGGGAGATCCGACAATACAGA Reverse (1): CTCATAAGCACTGCAGTTCCGCTAACAGTCTGCATGGAGCAGG Forward (2): GGAACTGCAGTGCTTATGAGGGG Reverse (2): AGCGGCCGCGCTGCTGCTACCTTTGGGC |
| AKT1 3′UTR | Forward: ACTCGAGGGACTGCGCTGGACGATAGC Reverse: AGCGGCCGCTGAAGAATTTGGAGGGAAGGTTCC |
| AKT2 3′UTR | Forward: ACTCGAGCAGCCTCCAGCCTCACCTTTG Reverse: AGCGGCCGCGCTGTGCCCACACTACGAGACC |
| APPL1 3′UTR | Forward: GCCAGCATTTAAGTAGTCATGGC Reverse: CTAGCTCTTAGGATCTCTCAG |
| Pre-miR-323a-3p | Forward: GTGGCTGGTAGTTCATGATGCGC Reverse: CAAGCCAGAGCAGTGTGCTGAG |
| Pre-miR-532-5p | Forward: GACCTGCATGCCAGATTCTAAGC Reverse: GGGATGGCTGCTAACAGTGTACC |
| shRNA | Forward: CGCGGATCCCTGCGGGTTCAGTTCCAGAGCACAACTACTACCAACAGTGC Reverse: CCCAAGCTTTTCCAAAAAACTGCGGGTTCCTTTCCAGAGCACTGTTGGTAGTAG |
Short hairpin RNA designing
Short hairpin RNA (shRNA) was designed against exon E1’ of YWHAE lncRNA (Fig. 2A) according to the reference instructions (50). ShRNA was amplified using specific primers (Table 1) and cloned in pRNA H1.1/Neo expression vector (Genscript, USA). RNA fold Webserver (51) was used to predict the shRNA’s secondary structure (Figure.2A). PHDcleav software (52) was used for prediction of human dicer cleavage site (DCS) and processing of the shRNA.
Cell culture and transient transfection
HT29 and HCT116 cell lines were cultured in RPMI 1640; however, SW480 and HEK293-T cells were cultured in HG-DMEM and DMEM-F12 media, respectively (Invitrogen). The culture media was supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin (Sigma), 10% fetal bovine serum (Invitrogen) and incubated at 37°C with 5% CO2. These cell lines were obtained from Pasteur Institute/Iran. HCT116 cell transfection with YWHAE lncRNA, pre-miR-323a-3p, pre-miR-532-5p or mock (control) was done using TurboFect reagent (Fermentas, USA) according to the manufacturer’s instructions. HEK293-T cell co-transfection with psi-check2 vector carrying YWHAE lncRNA, K-Ras, AKT1, AKT2 or APPL1 3′UTRs and pre-miR-323a-3p or pre-miR-532-5p was also performed using TurboFect reagent (Fermentas, USA) according to the manufacturer’s instructions. Transfection rate was determined by visualizing GFP (Green Fluorescent Protein) expression after 24 h of transfection using inverted fluorescence microscopy (Nikon eclipse Te2000-s).
RNA extraction and cDNA synthesis
Total cellular RNA was extracted from the cells using RiboEx reagent (Krishgen Biosystems, India) according to the manufacturer’s protocol. The quality and yield of the extracted RNA was analyzed using agarose gel electrophoresis and spectrophotometry, respectively. Later on, cDNA synthesis was performed according to the following protocol: (1) DNase 1 treatment at 37°C for 30 min to avoid genomic DNA contamination then followed by inactivation at 72°C with EDTA. (2) For miRNA detection, Poly-A tailing was performed for 1 μg of RNA using 2.5 U Poly A polymerase (NEB, UK), 2 μl of 10 mM ATP and incubated at 37°C for 40 min. (3) For mRNA reverse transcription enhancement, random hexamer and oligo dT mix primers were used; however, for miRNA reverse transcription enhancement, anchored or universal oligo-dT primers were added to the previous mixture and incubated at 65°C for 5 min. (4) cDNA synthesis was performed using the Prime Script II reverse transcriptase (RT) (Takara, Japan) at 42°C for 70 min followed by RT inactivation at 72°C for 12 min.
Reverse transcription PCR
For the cloning purpose, cDNA was amplified using specific cloning primers according to the following protocol: initial denaturation at 95°C for 10 min, and the second step includes denaturation at 95°C for 15 s, annealing at 60°C for 25 s and extension at 72°C for 1 min. This step was repeated for 35 cycles. The last step includes final extension at 72°C for 5 min. The cloning PCR primers used in the following experiment are presented in Table 1.
Real-time quantitative PCR
RT-qPCR was performed using 50 ng of cDNA template and SYBR Premix Kit (Takara Biotechnology, Japan) in StepOne™ system (Applied Biosystems, USA). The running method was as the following: initial denaturation at 95°C for 10 min, and the second step includes denaturation at 95°C for 10 s, annealing at 60°C for 25 s and extension at 72°C for 30 s. This step was repeated for 40 cycles. The primers used in the following experiment are presented in Table 2.
List of the real-time PCR primers
| Real-time PCR primer’s name | Primer sequence, 5′ to 3′ |
|---|---|
| YWHAE-encoded LncRNA | Forward: GGAAGTCCCGGATTGAGGCG Reverse: GGTTCAGTTCCAGAGCACGTC |
| YWHAE mRNA | Forward: CATTTTTGCTGCCCGGACG Reverse: CATTGACTCCACCATTTCGTCG |
| K-Ras gene | Forward: GGAGAGAGGCCTGCTGAAAATGAC Reverse: CTCCTCTTGACCTGCTGTGTCGAG |
| miR-323a-3p | Forward: CACATTACACGGTCGACCT Reverse: AACTCAAGGTTCTTCCAGTCACG |
| miR-532-5p | Forward: CATGCCTTGAGTGTAGGACC Reverse: AACTCAAGGTTCTTCCAGTCACG |
| Cyclin D1 | Forward: CAGAGTGATCAAGTGTGACCC Reverse: CGTCGGTGGGTGTGCAAGC |
| GAPDH | Forward: GCCACATCGCTCAGACAC Reverse: GGCAACAATATCCACTTTACCAG |
| U48 | Forward: TGACCCCAGGTAACTCTGAGTGTGT Reverse: AACTCAAGGTTCTTCCAGTCACG |
| Real-time PCR primer’s name | Primer sequence, 5′ to 3′ |
|---|---|
| YWHAE-encoded LncRNA | Forward: GGAAGTCCCGGATTGAGGCG Reverse: GGTTCAGTTCCAGAGCACGTC |
| YWHAE mRNA | Forward: CATTTTTGCTGCCCGGACG Reverse: CATTGACTCCACCATTTCGTCG |
| K-Ras gene | Forward: GGAGAGAGGCCTGCTGAAAATGAC Reverse: CTCCTCTTGACCTGCTGTGTCGAG |
| miR-323a-3p | Forward: CACATTACACGGTCGACCT Reverse: AACTCAAGGTTCTTCCAGTCACG |
| miR-532-5p | Forward: CATGCCTTGAGTGTAGGACC Reverse: AACTCAAGGTTCTTCCAGTCACG |
| Cyclin D1 | Forward: CAGAGTGATCAAGTGTGACCC Reverse: CGTCGGTGGGTGTGCAAGC |
| GAPDH | Forward: GCCACATCGCTCAGACAC Reverse: GGCAACAATATCCACTTTACCAG |
| U48 | Forward: TGACCCCAGGTAACTCTGAGTGTGT Reverse: AACTCAAGGTTCTTCCAGTCACG |
List of the real-time PCR primers
| Real-time PCR primer’s name | Primer sequence, 5′ to 3′ |
|---|---|
| YWHAE-encoded LncRNA | Forward: GGAAGTCCCGGATTGAGGCG Reverse: GGTTCAGTTCCAGAGCACGTC |
| YWHAE mRNA | Forward: CATTTTTGCTGCCCGGACG Reverse: CATTGACTCCACCATTTCGTCG |
| K-Ras gene | Forward: GGAGAGAGGCCTGCTGAAAATGAC Reverse: CTCCTCTTGACCTGCTGTGTCGAG |
| miR-323a-3p | Forward: CACATTACACGGTCGACCT Reverse: AACTCAAGGTTCTTCCAGTCACG |
| miR-532-5p | Forward: CATGCCTTGAGTGTAGGACC Reverse: AACTCAAGGTTCTTCCAGTCACG |
| Cyclin D1 | Forward: CAGAGTGATCAAGTGTGACCC Reverse: CGTCGGTGGGTGTGCAAGC |
| GAPDH | Forward: GCCACATCGCTCAGACAC Reverse: GGCAACAATATCCACTTTACCAG |
| U48 | Forward: TGACCCCAGGTAACTCTGAGTGTGT Reverse: AACTCAAGGTTCTTCCAGTCACG |
| Real-time PCR primer’s name | Primer sequence, 5′ to 3′ |
|---|---|
| YWHAE-encoded LncRNA | Forward: GGAAGTCCCGGATTGAGGCG Reverse: GGTTCAGTTCCAGAGCACGTC |
| YWHAE mRNA | Forward: CATTTTTGCTGCCCGGACG Reverse: CATTGACTCCACCATTTCGTCG |
| K-Ras gene | Forward: GGAGAGAGGCCTGCTGAAAATGAC Reverse: CTCCTCTTGACCTGCTGTGTCGAG |
| miR-323a-3p | Forward: CACATTACACGGTCGACCT Reverse: AACTCAAGGTTCTTCCAGTCACG |
| miR-532-5p | Forward: CATGCCTTGAGTGTAGGACC Reverse: AACTCAAGGTTCTTCCAGTCACG |
| Cyclin D1 | Forward: CAGAGTGATCAAGTGTGACCC Reverse: CGTCGGTGGGTGTGCAAGC |
| GAPDH | Forward: GCCACATCGCTCAGACAC Reverse: GGCAACAATATCCACTTTACCAG |
| U48 | Forward: TGACCCCAGGTAACTCTGAGTGTGT Reverse: AACTCAAGGTTCTTCCAGTCACG |
Dual-luciferase reporter assay
For this purpose, YWHAE lncRNA, ~ 2 kbp sized sequence of K-Ras 3′UTR, 964 bp of AKT1 3’UTR, ~ 2.5 kbp of AKT2 3′UTR and 1091 bp of APPL1 3′UTR were PCR amplified and cloned downstream of the luciferase gene in psiCHECK-2 vector (Table 1). AKT1, AKT2 and APPL1 3’UTRs were used as non-target controls in this experiment. HEK293-T cells were co-transfected with constructs carrying YWHAE lncRNA, K-RAS, AKT1, AKT2 or APPL1 3′UTRs along with pre-miR-323a-3p, pre-miR-532-5p or mock expression vectors in a 48-well plate. Luciferase activity was measured 48 h after transfection using the Dual-Stop and Glo Luciferase System (Promega, USA) according to the manufacturer’s instructions. Firefly luciferase activity was normalized against the Renilla luciferase signal.
Protein extraction and western blotting
HCT116 transfected cells were lysed using Ripa buffer according to the manufacturer’s protocol (cell signaling) followed by Bradford assay for measuring protein concentration. After that, the extracted proteins were subjected to 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (Sigma, USA) and then transferred to polyvinylidene fluoride (PVDF, Thermo Scientific, USA) membrane. Later on, membrane blocking was performed at room temperature for 2 h using 5% BSA (Sigma, USA) diluted in Tris-buffered saline containing 0.1% Tween (TBST, Bio basic, Canada). Then, the membrane was incubated with the primary antibodies, phospho-Akt (1:500, Santa Cruz Biotechnology), phospho-Erk1/2 (1/500, Santa Cruz Biotechnology), cyclin D1 (1/500, Santa Cruz Biotechnology) and β-actin (1:1000; Santa Cruz Biotechnology) overnight at 4°C. Sheep anti-rabbit IgG-Horse Radish Peroxidase (HRP) (1/5000, Santa Cruz Biotechnology) (against p-Erk1/2 and cyclin D1) and goat anti-mouse IgG-HRP (1:3000; Santa Cruz Biotechnology) (against p-Akt and β-actin) secondary antibodies were incubated for 2 h at room temperature. Actin protein was used as a loading control. The bands were visualized using ECL reaction kit (Beyotime, China), recorded in Canon EOS 60D and quantified using ImageJ software.
Flow cytometry analysis
Cells were trypsinized and harvested, washed with PBS and stained with Propidium Iodide (Sigma, USA) after 48 h of transfection. The test was performed in duplicates. Analysis of the cell cycle was done using the automated multicolor flow cytometry system: BD FACS Calibur Flow Cytometer (BD Biosciences). Flow cytometry results were analyzed using FlowJo-V10-CL Software.
MTT assay
Approximately 10 000 cells were seeded per well of the 96-well plates. Later on, the complete medium was removed and MTT mixture was added to the cells. The plate was completely covered with an aluminum foil and incubated for 4 h at 37°C. After that, the medium was removed using a syringe and DMSO was added to the cells and the plate was incubated for 15 mins at 37°C. DMSO enhances the dissolving of the MTT precipitate. Finally, the optical density of each well was measured using Elisa reader at 570 nm.
Wound-healing assay
HCT116 cells were seeded in 24-well plates and transfected with YWHAE lncRNA. Cells were immediately scratched after changing the transfection medium, and photos were taken after 24, 48 and 72 h of transfection using light microscopy. The percentage of the wound area was measured using ImageJ software.
Statistical analysis
Real-time PCR results were normalized against the endogenous expression of the controls (U48 or GAPDH genes) and analyzed according to the 2−ΔΔCt (Delta-Delta Ct) and 2−ΔCt (Delta Ct) algorithms using Microsoft Excel. GraphPad Prism 6 (San Diego, CA) was used to perform statistical paired or unpaired t-test and graph construction. RNA-Seq Data in Figure 1 was analyzed using unpaired t-test. Results with P-value less than 0.05 (*) were considered as statistically significant.
Acknowledgements
Precious thanks to Dr Abdullah Medlej for his technical advice. We also thank Talighani’s Hospital for supporting us with the tissue samples.
Conflict of Interest statement. None declared.
Funding
Tarbiat Modares University financial aids.
Author contributions
H.B. contributed in the conception and design of the study, performed the experiments, analyzed the data and wrote the manuscript. B.S. participated in the development of methodology. BS and MB reviewed and/or revised the manuscript. All authors read and approved the final manuscript.
