lncRNA GAS5 Inhibits Cell Migration and Invasion and Promotes Autophagy by Targeting miR-222-3p via the GAS5/PTEN-Signaling Pathway in CRC.

Colorectal cancer (CRC) is a frequently occurring lethal disorder with heterogeneous outcomes and drug responses. Recent studies have demonstrated that long non-coding RNAs (lncRNAs) play a critical role in carcinogenesis. Hence, the aim of this study was to investigate the role of lncRNA growth arrest-specific 5 (GAS5) in CRC cells via mediation of the microRNA-222-3p (miR-222-3p)/GAS5/phosphatase and tensin homolog (PTEN)-signaling pathway. HCT116 and SW480 cells were collected and treated with small interfering (si)-lncRNA GAS5, overexpressing (oe)-lncRNA GAS5, miR-222-3p mimic, miR-222-3p inhibitor, or si-lncRNA GAS5 + miR-222-3p mimic. The miR-222-3p level and mRNA and protein levels of GAS5, Beclin1, light-chain 3B (LC3B), PTEN, and Akt were detected. Besides, cell migration, invasion, and apoptosis as well as acidic vesicular organelles (AVOs) were examined respectively. Xenografts in nude mice were also performed to detect tumorigenesis in vivo. Results suggested that the downregulation of lncRNA GAS5 decreased the expressions of Beclin1, LC3B, and PTEN. When treated with oe-lncRNA GAS5 or miR-222-3p inhibitor, HCT116 and SW480 cells exhibited suppressed invasion and migration abilities and increased apoptotic cells and autophagosome and AVO activities. Moreover, overexpression of GAS5 inhibited the tumorigenesis of CRC cells in vivo. Taken together, lncRNA GAS5 upregulated the expression of PTEN by functioning as a competing endogenous RNA (ceRNA) of miR-222-3p, thus inhibiting CRC cell migration and invasion and promoting cell autophagy.

Introduction

Colorectal cancer (CRC) is a fatal disease and is the third most common cancer globally. It is estimated that each year over 500,000 patients worldwide die from CRC. In China, CRC ranks as the fourth most common malignancy, with increasingly high incidence and mortality.2, 3 It is acknowledged that CRC is caused by uncontrollable replication of epithelial cells attached on the inner side of the colon. The symptoms of CRC patients are affected by multiple factors, including different stages, sites, and sizes of tumors, and they might differ for each individual, varying from abnormal weight loss to changes in defecation and pain. Despite improved diagnosis and therapy together with less-exposed social risks, CRC still has a poor 5-year survival rate of only 64.9%. Despite current treatments, such as chemotherapy, efficient regimens designed to prevent recurrence are still absent. Due to the poor response of many patients to current treatment regimens, it is of paramount importance to develop novel therapies for CRC.

In recent years, long non-coding RNAs (lncRNAs) have emerged as new hotspots for disease treatment, especially in malignant tumors.5, 6 Particularly, lncRNAs have been implicated in cell fate and gene expression during the occurrence and progression of certain cancers, including CRC. Moreover, lncRNA involvement in selected biological processes includes competitive binding to microRNA (miRNAs). For example, growth arrest-specific transcript 5 (GAS5), a member of the lncRNA family as a tumor suppressor, has been reported to induce cell apoptosis in endometrial cancer by downregulating miR-103, although GAS5 small nucleolar RNAs (snoRNAs) were found to be correlated with p53 expression and DNA damage in CRC. The accumulating literature has demonstrated that GAS5 is downregulated in CRC tissues and CRC cell lines.9, 10

At the same time, a prior study mentioned that miRNAs play critical roles in CRC via the mediation of gene expression. For instance, miR-222, which is highly expressed in CRC, can regulate disintegrin and metallopeptidase domain 17 (ADAM-17), a protein involved in drug resistance.1, 11 Additionally, certain miRNAs are associated with tumorigenesis by the inhibition of phosphatase and tensin homolog (PTEN), an anti-oncogene gene situated on 10p23, which was reported to be related to cancer cell proliferation and migration.

Based on the abovementioned facts, our study boldly investigated the role of lncRNA GAS5 in CRC by mediating the miR-222-3p/GAS5/PTEN-signaling pathway, on which little attention has been focused in previous studies. As a result, this paper represents a concise version for clinicians by incorporation of the possible functions of lncRNA GAS5 in the treatment of CRC via the miR-222-3p/GAS5/PTEN-signaling pathway.

Results

lncRNA GAS5 Regulates PTEN by Competitive Binding to miR-222-3p

lncRNA GAS5 and miR-222 were reported to be involved in the occurrence of CRC cells.9, 12 In this study, we further verified the expression of lncRNA GAS5 and miR-222-3p in HCT116 and SW480 CRC cells. The results showed (Figure 1A) that the expression of lncRNA GAS5 was significantly lower in the HCT116 and SW480 cell lines compared with that in the normal NCM460 cell line, whereas the expression of miR-222-3p was significantly higher (p < 0.05). Dual-luciferase reporter gene assay, RNA immunoprecipitation (IP), and RNA pull-down assays were conducted to verify the relationship among lncRNA GAS5, miR-222-3p, and PTEN. Online bioinformatics analysis software showed that there were specific binding sites between the 3′ UTR of the PTEN gene and sequences of miR-222-3p and between lncRNA GAS5 and sequences of miR-222-3p.

Figure 1

lncRNA GAS5 Regulates PTEN by Competitive Binding to miR-222-3p Based on the Target Prediction Program and Determination of Luciferase Activity

(A) The expressions of miR-222-3p and GAS5 determined by qRT-PCR. (B) Verification of the target relationship between miR-222-3p and PTEN. (C) Verification of the target relationship between miR-222-3p and GAS5. (D) The binding between lncRNA GAS5 and miR-222-3p examined by RNA pull-down assay. (E) The binding between lncRNA GAS5 and Ago2 examined by RNA IP. miR-222-3p, microRNA-222-3p; PTEN, phosphatase and tensin homolog; GAS5, growth arrest-specific 5; IP, immunoprecipitation; lncRNA, long non-coding RNA. *p < 0.05 compared with empty vector group; #p < 0.05 compared with NCM460 cell line; Δp < 0.05 compared with Ago2. The data were measurement data and presented as the mean ± SD and analyzed by the independent sample t test. The experiment was independently repeated three times.

The dual-luciferase reporter assay revealed that, compared with the empty vector group, luciferase activity of the PTEN wild-type (WT) 3′ UTR was markedly inhibited by miR-222-3p (p < 0.05), whereas that of PTEN mutant (MUT) 3′ UTR was not notably affected (p > 0.05), indicating that miR-222-3p specifically bound to PTEN-3′ UTR and reduced the expression of PTEN after transcription (Figure 1B). At the same time, compared with the empty vector group, luciferase activity of the lncRNA GAS5 WT-binding site was markedly inhibited by miR-222-3p (p < 0.05), whereas that of the lncRNA GAS5 MUT-binding site was not notably affected (p > 0.05). It was demonstrated that miR-222-3p specifically bound to lncRNA GAS5 (Figure 1C). The RNA pull-down assay showed that, compared with the MUT-miR-222-3p, the WT-miR-222-3p that bound to lncRNA GAS5 was increased remarkably (p < 0.05), suggesting that miR-222-3p could directly bind to lncRNA GAS5 (Figure 1D). The Ago2 co-immunoprecipitation assay demonstrated that lncRNA GAS5 binding to Ago2 was increased compared with that binding to immunoglobulin G (IgG) (p < 0.05), suggesting that lncRNA GAS5 can bind to Ago2 protein (Figure 1E). Taken together, these results suggest that lncRNA GAS5 could regulate PTEN through competitively binding to miR-222-3p.

Downregulation of lncRNA GAS5 Promotes miR-222-3p Level but Inhibits Levels of PTEN, LC3B, and Beclin1

The relative expression of miR-222-3p and the mRNA expression of the GAS5/miR-222-3p/PTEN pathway-related genes GAS5, PTEN, light-chain 3B (LC3B), and Beclin1 in the CRC HCT116 cell line were detected by qRT-PCR, as shown in Figure 2A. Compared with the control group, miR-222-3p expression and mRNA expressions of GAS5, PTEN, LC3B, and Beclin1 did not differ significantly in the empty vector group (all p > 0.05). Compared with the control group, higher mRNA expressions of Beclin1, LC3B, PTEN, and GAS5 were observed in the overexpressing (oe)-lncRNA GAS5 and miR-222-3p inhibitor groups in which miR-222-3p expression was downregulated, whereas lower mRNA expressions of Beclin1, LC3B, PTEN, and GAS5 were obtained in the miR-222-3p mimic, small interfering (si)-lncRNA GAS5, and si-lncRNA GAS5 + miR-222-3p mimic groups in which miR-222-3p expression was upregulated. Compared with the si-lncRNA GAS5 and miR-222-3p mimic groups, the mRNA expressions of Beclin1, LC3B, PTEN, and GAS5 were reduced, but the miR-222-3p expression was elevated in the si-lncRNA GAS5 + miR-222-3p mimic group (all p < 0.05). However, no obvious changes were noted in all the above indicators between the si-lncRNA GAS5 and miR-222-3p mimic groups (p > 0.05). Similar results were found in the CRC SW480 cell line (Figure 2B). It was concluded that the downregulation of lncRNA GAS5 could promote the miR-222-3p level but inhibit the mRNA levels of PTEN, LC3B, and Beclin1.

Figure 2

lncRNA GAS5 Inhibits miR-222-3p Level but Enhances mRNA levels of PTEN, LC3B, and Beclin1

(A) The expression of miR-222-3p and mRNA expressions of genes related to the GAS5/PTEN-signaling pathway, including GAS5, Beclin1, LC3B, and PTEN in the HCT116 cell line, as determined by qRT-PCR. (B) The expression of miR-222-3p and mRNA expressions of genes related to the GAS5/PTEN-signaling pathway, including GAS5, Beclin1, LC3B, and PTEN in the SW480 cell line, as determined by qRT-PCR. miR-222-3p, microRNA-222-3p; GAS5, growth arrest-specific 5; LC3B, light-chain 3B; PTEN, phosphatase and tensin homolog; lncRNA, long non-coding RNA. *p < 0.05 compared with control group; Δp < 0.05 compared with si-lncRNA GAS5 and miR-222-3p mimic groups. The data were measurement data, presented as the mean ± SD and analyzed by one-way ANOVA. The experiment was independently repeated three times.

Downregulation of lncRNA GAS5 Inhibits Protein Levels of PTEN, LC3B, Cleaved Caspase-3, and Beclin1 but Increases the Phosphorylation of Akt

Western blot assay was performed to determine the protein expression of signaling pathway-related factors in the CRC HCT116 and SW480 cell lines. Compared with the control group, protein expressions of PTEN, LC3BII/I, Beclin1, and cleaved caspase-3 did not show significant differences in the empty vector group (all p > 0.05). In contrast to the control group, the oe-lncRNA GAS5 and miR-222-3p inhibitor groups revealed increased protein expressions of PTEN, LC3BII/I, Beclin1, and cleaved caspase-3 but decreased phosphorylation of Akt, while the miR-222-3p mimic, si-lncRNA GAS5, and si-lncRNA GAS5 + miR-222-3p mimic groups exhibited reduced protein expressions of PTEN, LC3BII/I, Beclin1, and cleaved caspase-3 but elevated the phosphorylation of Akt (all p < 0.05). Protein expressions of Beclin1, LC3B II/I, cleaved caspase-3, and PTEN were lower while the phosphorylation of Akt was higher in the si-lncRNA GAS5 + miR-222-3p mimic group than those in the si-lncRNA GAS5 and miR-222-3p mimic groups (all p < 0.05). However, protein expressions of Beclin1, LC3B II/I, cleaved caspase-3, and PTEN did not differ significantly between the si-lncRNA GAS5 and miR-222-3p mimic groups (p > 0.05). The protein expression of Akt remained nearly the same among all groups (Figures 3A and 3B). Besides, the results were similar in the CRC SW480 cell line (Figures 3C and 3D). It was suggested that the downregulation of lncRNA GAS5 could inhibit the protein levels of PTEN, LC3B, cleaved caspase-3, and Beclin1 while increasing the phosphorylation of Akt.

Figure 3

lncRNA GAS5 Increases Protein Levels of PTEN, LC3B, Cleaved Caspase-3, and Beclin1 but Decreases the Phosphorylation of Akt

(A) The protein bands of Beclin1, LC3B II/I, PTEN, Akt, p-Akt, cleaved caspase-3, and GAPDH in HCT116 cells detected by western blot assay. (B) The quantitative analysis of relative protein expression levels of Beclin1, LC3B II/I, PTEN, Akt, and cleaved caspase-3 and the phosphorylation of Akt in HCT116 cells. (C) The protein bands of Beclin1, LC3B II/I, PTEN, Akt, p-Akt, cleaved caspase-3, and GAPDH in SW480 cells detected by western blot assay. (D) The quantitative analysis of relative protein expression levels of Beclin1, LC3B II/I, PTEN, Akt, and cleaved caspase-3 and the phosphorylation of Akt in SW480 cells. miR-222-3p, microRNA-222-3p; GAS5, growth arrest-specific 5; LC3B, light-chain 3B; PTEN, phosphatase and tensin homolog; lncRNA, long non-coding RNA; p-Akt, phosphorylated Akt. *p < 0.05 compared with control group; Δp < 0.05 compared with si-lncRNA GAS5 and miR-222-3p mimic groups. The data were measurement data, presented as the mean ± SD and analyzed by one-way ANOVA. The experiment was independently repeated three times.

lncRNA GAS5 Suppresses HCT116 and SW480 Cell Migration Ability

To observe whether miR-222-3p or lncRNA GAS5 could affect CRC cell migration, we performed the scratch test. As shown in Figures 4A and 4B, compared with the control group, the migration ability of HCT116 cells in the empty vector group did not show obvious changes (p > 0.05), but it was enhanced in the si-lncRNA GAS5, miR-222-3p mimic, and si-lncRNA GAS5 + miR-222-3p mimic groups (p < 0.05) and inhibited in the oe-lncRNA GAS5 and miR-222-3p inhibitor groups (all p < 0.05). Compared with the si-lncRNA GAS5 and miR-222-3p mimic groups, cells in the si-lncRNA GAS5 + miR-222-3p mimic group exhibited enhanced migration (p < 0.05). Nevertheless, no significant difference in cell migration ability was observed between the si-lncRNA GAS5 and miR-222-3p mimic groups (p > 0.05). Similar results were found in the CRC SW480 cell line (Figures 4C and 4D). According to the above results, upregulation of lncRNA GAS5 could inhibit CRC cell migration ability.

Figure 4

lncRNA GAS5 Suppresses HCT116 and SW480 Cell Migration Ability

(A) A scratch test was conducted to examine the migration of HCT116 cells in the seven groups at time points of transfection for 0 and 48 h. (B) The wound-healing rate in HCT116 cells in the seven groups. (C) The scratch test was used to detect the migration of SW480 cells in the seven groups at time points of transfection for 0 and 48 h. (D) The wound-healing rate in SW480 cells in the seven groups. miR-222-3p, microRNA-222-3p; GAS5, growth arrest-specific 5; lncRNA, long non-coding RNA. *p < 0.05 compared with control group; Δp < 0.05 compared with si-lncRNA GAS5 and miR-222-3p mimic groups. The data were measurement data, presented as the mean ± SD and analyzed by one-way ANOVA. The experiment was independently repeated three times.

lncRNA GAS5 Represses HCT116 and SW480 Cell Invasion Ability

To observe whether miR-222-3p or lncRNA GAS5 could affect CRC cell invasion, we conducted the Transwell assay. As shown in Figures 5A and 5B, compared with the control group, the invasion ability of the HCT116 cells in the empty vector group did not show any significance (p > 0.05), but it was promoted in the si-lncRNA GAS5, miR-222-3p mimic, and si-lncRNA GAS5 + miR-222-3p mimic groups (p < 0.05) and suppressed in the oe-lncRNA GAS5 and miR-222-3p inhibitor groups (all p < 0.05). Compared with the si-lncRNA GAS5 and miR-222-3p mimic groups, the invasion ability was enhanced in the si-lncRNA GAS5 + miR-222-3p mimic group (all p < 0.05). However, no significant difference was found between the si-lncRNA GAS5 and miR-222-3p mimic groups (p > 0.05). Similar results were found in the CRC SW480 cell line (Figures 5C and 5D). According to the aforementioned results, upregulation of lncRNA GAS5 could restrict CRC cell invasion ability.

Figure 5

lncRNA GAS5 Enhances HCT116 and SW480 Cell Invasion Ability

(A) The cell invasion of HCT116 cells after transfection in the seven groups (×200). (B) The number of invasive HCT116 cells after transfection in the seven groups. (C) The cell invasion of SW480 cells after transfection in the seven groups (×200). (D) The number of invasive SW480 cells after transfection in the seven groups. miR-222-3p, microRNA-222-3p; GAS5, growth arrest-specific 5; lncRNA, long non-coding RNA. *p < 0.05 compared with control group; Δp < 0.05 compared with si-lncRNA GAS5 and miR-222-3p mimic groups. The data were measurement data, presented as the mean ± SD and analyzed by one-way ANOVA. The experiment was independently repeated three times.

lncRNA GAS5 Induces HCT116 and SW480 Cell Autophagy

In addition, to investigate the role of miR-222-3p and lncRNA GAS5 in CRC cell autophagy, we observed the appearance of the autophagosome or other autophagic structures in HCT116 and SW480 cells under transmission electron microscopy, which is the most direct way to prove the existence of autophagy. The results in the HCT116 cells after treatment for 24 and 48 h were consistent with those in the SW480 cells after treatment for 48 h. A certain amount of vesicle-like structures and autolysosome structures wrapped up with kytoplasm was observed in the control group. The number of vesicle-like structures and autolysosome structures did not show obvious changes in the empty vector group compared with that in the control group, but it was larger in the oe-lncRNA GAS5 and miR-222-3p inhibitor groups and smaller in the si-lncRNA GAS5 and miR-222-3p mimic groups. However, few such structures were found in the si-lncRNA GAS5 + miR-222-3p mimic group (Figures 6A–6C). Furthermore, acridine orange staining further confirmed the results above (Figures 6D and 6E). It was suggested that upregulation of lncRNA GAS5 could promote CRC cell autophagy.

Figure 6

lncRNA GAS5 Induces Cell Autophagy Based on Observation under Electron Microscopy

(A and B) HCT116 cell after treatment for (A) 24 and (B) 48 h under electron microscopy, respectively (scale bar, 500 nm). (C) SW480 cell after treatment for 48 h under electron microscopy (scale bar, 500 nm). (D) Acidic vesicular organelle (AVO) activities of HCT116 cells by acridine orange staining (×400). (E) AVO activities of SW480 cells by acridine orange staining (×400). Red arrows referred to the autophagic vacuoles. miR-222-3p, microRNA-222-3p; GAS5, growth arrest-specific 5; lncRNA, long non-coding RNA. *p < 0.05 compared with control group; Δp < 0.05 compared with si-lncRNA GAS5 and miR-222-3p mimic groups. The data were measurement data, presented as the mean ± SD and analyzed by one-way ANOVA. The experiment was independently repeated three times.

lncRNA GAS5 Promotes HCT116 and SW480 Cell Apoptosis

Flow cytometry was performed to explore the effects of lncRNA GAS5 and miR-222-3p on the apoptosis of HCT116 and SW480 cells after transfection for 24 or 48 h. As shown in Figures 7A–7D, compared with the control group, the HCT116 cell apoptosis rate did not differ significantly in the empty vector group (p > 0.05), but it decreased in the si-lncRNA GAS5, miR-222-3p mimic, and si-lncRNA GAS5 + miR-222-3p mimic groups (p < 0.05), and it increased in the oe-lncRNA GAS5 and miR-222-3p inhibitor groups (all p < 0.05). HCT116 cells in the si-lncRNA GAS5 + miR-222-3p mimic group exhibited a lower apoptosis rate compared with the si-lncRNA GAS5 and miR-222-3p mimic groups (p < 0.05). No significant difference was found between the si-lncRNA GAS5 and miR-222-3p mimic groups (p > 0.05). Similar trends were found in the CRC SW480 cell line after transfection for 48 h (Figures 7E and 7F). The above findings suggested that upregulation of lncRNA GAS5 could induce CRC cell apoptosis.

Figure 7

lncRNA GAS5 Promotes HCT116 and SW480 Cell Apoptosis

(A) HCT116 cell apoptosis after treatment for 24 h measured by flow cytometry. (B) HCT116 cell apoptosis rate in the six groups. (C and D) HCT116 cell apoptosis after treatment for 48 h measured by flow cytometry. (E and F) SW480 cell apoptosis after treatment for 48 h measured by flow cytometry. (D) SW480 cell apoptosis rate in the six groups. miR-222-3p, microRNA-222-3p; GAS5, growth arrest-specific 5; lncRNA, long non-coding RNA. *p < 0.05 compared with control group; Δp < 0.05 compared with si-lncRNA GAS5 and miR-222-3p mimic groups. The data were measurement data, presented as the mean ± SD and analyzed by one-way ANOVA. The experiment was independently repeated three times.

Overexpression of GAS5 Inhibits Tumorigenesis of CRC Cells In Vivo

To study the effect of miR-222-3p and GAS5 on tumor growth in vivo, a xenograft experiment in nude mice was performed. The results showed that the influence of injection of HCT116 cells on tumorigenesis in the mice did not significantly differ from that of injection of SW480 cells (p > 0.05). Compared with the control group, no significant change was observed in the size and weight of the tumors in the empty vector group (p > 0.05), whereas the size and weight of the tumors notably increased in the si-lncRNA GAS5 group, miR-222-3p mimic group, and si-lncRNA GAS5 + miR-222-3p mimic group (p < 0.05), but they decreased in the oe-lncRNA GAS5 and miR-222-3p inhibitor groups (p < 0.05). Compared with the si-lncRNA GAS5 group and the miR-222-3p mimic group, the size and weight of the tumors in the si-lncRNA GAS5 + miR-222-3p mimic group increased significantly (p < 0.05), and no significant difference was noted between the si-lncRNA GAS5 group and the miR-222-3p mimic group (p > 0.05) (Figures 8A–8C). We further analyzed the expression of PTEN in tumors (Figures 8G and 8H), and the results were consistent with the western blot analysis. The results showed that the overexpression of GAS5 and silencing of miR-222-3p both inhibited the tumorigenesis of CRC cells in vivo.

Figure 8

Overexpression of GAS5 Inhibits Tumorigenesis of CRC Cells In Vivo

(A) The xenograft tumor in mice after the injection of HCT116 cells. (B) The size of the tumor induced by the injection of HCT116 cells. (C) The weight of the tumor induced by the injection of HCT116 cells. (D) The xenograft tumor in mice after the injection of SW480 cells. (E) The size of the tumor induced by the injection of SW480 cells. (F) The weight of the tumor induced by the injection of SW480 cells. (G and H) The expression of PTEN in transplanted tumor tissues (×400). miR-222-3p, microRNA-222-3p; GAS5, growth arrest-specific 5; lncRNA, long non-coding RNA. *p < 0.05 compared with control group; Δp < 0.05 compared with si-lncRNA GAS5 and miR-222-3p mimic groups. The data were measurement data, presented as the mean ± SD and analyzed by one-way ANOVA. N = 5.

Discussion

The current study was intended to uncover the effects of lncRNA GAS5 and its relative genes, miR-222-3p, and PTEN on cell migration, invasion, and autophagy in CRC. The results suggested that lncRNA GAS5 inhibits cell migration and invasion and promotes autophagy in CRC by competitively inhibiting the suppressive effects of miR-222-3p on PTEN.

A significant finding of our study was that lncRNA GAS5 could bind to miR-222-3p and PTEN was a target gene of miR-222-3p. The biological prediction site (https://cm.jefferson.edu/rna22/) revealed that miR-222-3p contains the binding sites of PTEN in its 3′ UTR and of lncRNA GAS5. Likely explanations for the results are that miR-222 targets PTEN in CRC and miR-222-3p can regulate the expression of PTEN.13, 14 Furthermore, Pickard and Williams have suggested that lncRNA GAS5 can be negatively regulated by the anti-tumor gene PTEN. In support, a group of Chinese researchers showed that miR-222 has a negative relationship with GAS5 in patients with glioma. According to the RNA IP assay conducted in our study, the expression of lncRNA GAS5 immunoprecipitated with Ago2 was higher than that immunoprecipitated with IgG. Taken together, it is reasonable enough to conclude that lncRNA GAS5 is involved in the regulation of PTEN and miR-222-3p.

Moreover, our study also suggested that HCT116 and SW480 cell migration and invasion tended to be enhanced but autophagy tended to be reduced with upregulated miR-222-3p or downregulated GAS5. Our results were supported by the concept proposed by Liu et al. that miR-222-3p overexpression heightens the proliferative and invasive potential in endometrial carcinoma. In addition, a prior study demonstrated that HCT116 and SW480 cells treated with overexpressed GAS5 show lighter and smaller tumor size, suggesting inhibited cell viability, proliferation, and progression. Consistently, the overexpression of GAS5 was found to inhibit the proliferation of CRC cells, and it might also act as a promising prognostic biomarker in CRC. Li et al. also demonstrated that upregulated GAS5 expression contributed to suppressed proliferation, migration, and invasion as well as accelerated apoptosis of CRC cells. Moreover, our results demonstrated that the underexpression of GAS5 led to upregulated miR-222-3p in addition to downregulated Beclin1, LC3B, and PTEN.

The overexpression of miR-222 has been detected in CRC, and the inhibition of miR-222 was found to suppress proliferation and colony formation of CRC cells. A lack of PTEN expression is positively associated with the aggressive capacity of CRC, whereas adenovirus-mediated PTEN plays a tumor-suppressive role in CRC.21, 22 Therefore, a decreased PTEN level might exercise restraint on cell apoptosis but contribute to proliferation and invasion through the Akt/PI3K- or the Akt/PKB-signaling pathway.13, 23 Additionally, LC3B, a member of the LC3 gene family, has been reported to interfere with CRC tumor growth. Autophagy is an intricate metabolism procedure, and inhibited autophagy is conducive to cancer treatment.24, 25 Han et al. and colleagues have proven that an increased Beclin1 level has a bearing on CRC metastasis and unsatisfactory prognosis. Similarly, overexpressed Beclin1 in CRC might give rise to autophagy. Therefore, this idea is consistent with the observations from the current study that CRC cell migration, invasion, and autophagy were promoted in HCT116 and SW480 cells after transfection with si-lncRNA GAS5. Furthermore, in the current study, the xenograft experiment in nude mice demonstrated that overexpression of GAS5 inhibits the tumorigenesis of CRC cells in vivo. Consistently, GAS5 was found to be capable of suppressing tumorigenesis in melanoma and non-small-cell lung cancer.

As stated above, the current study suggested that lncRNA GAS5 inhibits CRC cell migration and invasion and promotes autophagy through the miR-222-3p/GAS5/PTEN-signaling pathway. These findings might offer a novel approach and new insights for future CRC therapies. However, the shortage of sufficient data from in vivo animal experiments is noteworthy, and, therefore, additional attention should be focused on establishing an orthotopic tumor model as a good approach toward further support of our findings.

Materials and Methods

Ethics Statement

The animal experiments were performed with the approval of the Guide for the Care and Use of Laboratory Animal by International Committees. All efforts were made to minimize suffering of the animals.

Cell Culture

CRC HCT116 and SW480 cells from American Type Culture Collection (Manassas, VA, USA) were cultured in DMEM containing 10% fetal bovine serum (FBS) at 37°C in 5% CO2. Normal human colon mucosal epithelial cell line NCM460 was purchased from INCELL (San Antonio, TX, USA). After adherence to the wall, the cells were digested using 0.25% trypsin for sub-culture. Cells in the logarithmic growth phase were selected for subsequent experiments.

Dual-Luciferase Reporter Gene Assay

The putative binding sites of miR-222-3p, lncRNA GAS5, and PTEN were predicted in the online bioinformatics prediction website (https://cm.jefferson.edu/rna22/), and sequences containing active sites were obtained. GAS5 full length and the 3′ UTR of PTEN were amplified and cloned on pmirGLO Luciferase vector (E1330, Promega, Madison, WI, USA) as pGAS5-WT and pPTEN-WT, respectively. Putative miR-222-3p-binding sites in the 3′ UTR of lncRNA GAS5 and PTEN were predicted, followed by site-directed mutation. pGAS5-MUT and pPTEN-MUT vectors were constructed with the pRL-TK vector (E2241, Promega, Madison, WI, USA), expressing renilla luciferase used as the internal reference. miR-222-3p mimic and miR-222-3p empty vectors were separately co-transfected with luciferase reporter gene vector into HCT116 and SW480 cells (CRL-1415, American Type Culture Collection, Manassas, VA, USA). Luciferase activity was measured at 560 nm (relative light unit [RLU] of Firefly luciferase) and 465 nm (RLU of Renilla luciferase) using the Dual-Luciferase Reporter Gene Assay Kit (GM-040502A, Qcbio Science & Technologies, Shanghai, China). Luciferase activity = RLUFirefly luciferase/RLURenilla luciferase.

RNA Pull-Down Assay

HCT116 and SW480 cells were transfected with 50 nM WT-bio-miR-222-3p and MUT-bio-miR-222-3p marked by biotin. At 48 h after transfection, cells were collected and washed with PBS. Specific lysis buffer (Ambion, Austin, TX, USA) was introduced for cell incubation for 10 min, followed by centrifugation (14,000 × g) with the supernatant obtained. M-280 streptavidin magnetic beads (S3762, Sigma-Aldrich, St. Louis, MO, USA) pre-coated with RNase-free BSA and yeast tRNA (TRNABAK-RO, Sigma-Aldrich, St. Louis, MO, USA) were later incubated with the protein lysis. After 3 h of incubation at 4°C, the beads were washed twice with pre-cooled lysis buffer, three times with low-salt buffer, and once with high-salt buffer. The binding RNA was purified using TRIzol, and lncRNA GAS5 was examined by qRT-PCR.

RNA IP

HCT116 and SW480 cells were treated with lysis buffer (25 mM Tris-HCl [pH 7.4], 150 mM NaCl, 0.5% NP-40, 2 mM EDTA, 1 mM NaF, and 0.5 mM dithiothreitol) supplemented with a mixture of RNasin (TaKaRa Biotechnology, Dalian, Liaoning, China) and protease inhibitor (B14001a, Roche, USA). The lysis buffer was centrifuged for 30 min (12,000 × g). The supernatant was obtained and added with anti-human Ago2 magnetic beads (BMFA-1, Biomarker Technologies, Beijing, China), and anti-IgG magnetic beads were added in the control group. After 4 h of incubation at 4°C, the beads were washed three times with washing buffer (50 mM Tris-HCl, 300 mM NaCl [pH 7.4], 1 mM MgCl2, and 0.1% NP-40). RNA was extracted from the magnetic beads using TRIzol, and lncRNA GAS5 was determined by qRT-PCR.

Cell Grouping and Transfection

Cells were assigned into the following seven groups: control (cells transfected without any sequence), empty vector (cells transfected with empty vector), si-lncRNA GAS5 (cells transfected with si-lncRNA GAS5), oe-lncRNA GAS5 (cells transfected with lncRNA GAS5 plasmid), miR-222-3p mimic (cells transfected with miR-222-3p mimic), miR-222-3p inhibitor (cells transfected with miR-222-3p inhibitor), and si-lncRNA GAS5 + miR-222-3p mimic (cells co-transfected with si-lncRNA GAS5 and miR-222-3p mimic). The si-lncRNA GAS5, oe-lncRNA GAS5, miR-222-3p mimic, and miR-222-3p inhibitor were all purchased from Guangzhou Ribo Biotechnology (Guangzhou, Guangdong, China).

Cells were seeded in a 24-well plate. When cell confluence reached nearly 50%–60%, HCT116 and SW480 cells were subjected to transfection according to the instructions of the Lipofectamine 2000 kit (Invitrogen, Carlsbad, CA, USA). Lipofectamine 2000 (1 μL) and serum-free culture medium (50 μL) were allowed to stand in a sterile Eppendorf (EP) tube at room temperature for 5 min. RNA to be transfected (20 pmol) and serum-free culture medium (50 μL) were placed in another sterile EP tube. The complex of RNA and liposome was obtained by mixing the solution in the above two tubes and allowing to stand at room temperature for 20 min. The mixture was added to the culture dish containing cells to be transfected and cultured at 37°C with 5% CO2. The previous culture medium was replaced by complete culture medium 6–8 h later.

qRT-PCR

CRC HCT116 and SW480 cells were subjected to total RNA extraction using the miRNeasy Mini Kit (QIAGEN, Duesseldorf, North Rhine-Westphalia, Germany). An amount of 5 μL RNA sample was diluted 20 times with RNase-free ultrapure water. Optical density (OD) values were measured at 260 and 280 nm using a UV spectrophotometer to determine the concentration and purity of extracted RNA. RNA was considered sufficiently pure for the subsequent experiments if the OD260/OD280 value was between 1.7 and 2.1. The total RNA was reversely transcribed into the cDNA template according to the instructions of the Reverse Transcription Kit (Beijing TransGen Biotech, Beijing, China), using a PCR amplification instrument. Primers of GAS5, miR-222-3p, PTEN, Beclin1, and LC3B were designed and synthesized by Sangon Biotech, Shanghai, China (Table 1). The reverse transcription system was conducted according to the instructions of the EasyScript First-Strand cDNA Synthesis SuperMix (AE301-02, Beijing TransGen Biotech, Beijing, China). Reverse transcription was performed on a PCR instrument (9700, Beijing Dingguo Changsheng Biotechnology, Beijing, China). Real-time fluorescent qPCR was conducted based on the instructions of the SYBR Premix Ex Taq II Kit (TaKaRa Biotechnology, Dalian, Liaoning, China). Real-time qRT-PCR was performed using an ABI7500 qPCR instrument (Applied Biosystems, Foster City, CA, USA) with glyceraldehyde phosphate dehydrogenase (GAPDH) serving as the internal control. The relative expression of the target gene was calculated based on the 2-ΔΔCt method.

Table 1

Primer Sequences for qRT-PCR

Gene	Primer Sequence
U6	F: 5′-AAAGCAAATCATCGGACGACC-3′
	R: 5′-GTACAACACATTGTTTCCTCGGA-3′
GAPDH	F: 5′-TGTGGGCATCAATGGATTTGG-3′
	R: 5′-ACACCATGTATTCCGGGTCAAT-3′
PTEN	F: 5′-TGGATTCGACTTAGACTTGACCT-3′
	R: 5′-GGTGGGTTATGGTCTTCAAAAGG-3′
LC3B	F: 5′-AGCAGCATCCAACCAAAATC-3′
	R: 5′-CTGTGTCCGTTCACCAACAG-3′
Beclin1	F: 5′-ACCGTGTCACCATCCAGGAA-3′
	R: 5′-GAAGCTGTTGGCACTTTCTGT-3′
miR-222-3p	F: 5′-GGGGAGCTACATCTGGCT-3′
	R: 5′-TGCGTGTCGTGGAGTC-3′
lncRNA GAS5	F: 5′-CCTGTGAGGTATGGTGCTGG-3′
	R: 5′-CTGTGTGCCAATGGCTTGAG-3′

F, forward; R, reverse; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; PTEN, phosphatase and tensin homolog; LC3B, light-chain 3B; miR-222-3p, microRNA-222-3p; lncRNA, long non-coding RNA; GAS5, growth arrest-specific 5.

Western Blot Assay

CRC HCT116 and SW480 cells were treated with radioimmunoprecipitation assay (RIPA) lysis buffer to collect total protein, followed by the determination of protein concentration by the bicinchoninic acid (BCA) method, and 10% separation gel and 5% spacer gel were prepared using the SDS-PAGE gel kit. After electrophoresis, the protein bands on the gel were transferred to a nitrocellulose membrane using the wet transfer method. The membrane was blocked in 5% BSA at room temperature for 1 h and incubated with the following antibodies (all purchased from Abcam, Cambridge, MA, USA) at 4°C overnight: rabbit polyclonal primary antibody PTEN (ab31392, 1:1,000), LC3B (ab48394, 1 μg/mL), Beclin1 (ab62557, 1 μg/mL), Akt (ab8805, 1:500), phosphorylated-Akt (p-Akt) (ab8932, 1:1,000), cleaved caspase-3 (ab49822, 1:500), and GAPDH (ab37168, 1 μg/mL). The next day, after washing 3 times (10 min each time) with PBS and Tween-20 (PBST), the membrane was incubated with secondary antibody goat anti-rabbit polyclonal antibody (ab205718, diluted to 1:5,000 using 5% skim milk). Then, the membrane was washed by PBST three times (15 min each time) and developed in the Bio-Rad gel imaging system (MG8600, Beijing Thmorgan Biotech, Beijing, China) using a developing agent. IPP7.0 software (Media Cybernetics, Silver Springs, MD, USA) was used in quantitative analysis, with GAPDH as the internal control. Relative protein expression = gray valuetarget band/gray valueinternal control.

Scratch Test

CRC HCT116 and SW480 cells were seeded in a 6-well plate with 5 × 105 cells/well. Transfection was performed when cells grew to adherence. Then 24 h later, a thin wound was created along the ruler on the surface of cells using a sterile pipette tip (200 μL). Images were acquired under an inverted microscope (CX23, Olympus, Tokyo, Japan). After 48 h of culture, the well plate was removed for photography, and the wound-healing rate was measured using the images acquired.

Transwell Assay

Matrigel (Becton Dickinson, NJ, USA) was diluted with pre-cooled serum-free DMEM at a ratio of 10:1 and mixed. Each apical chamber was coated with 100 μL diluted Matrigel and placed at room temperature for 2 h. The chambers were washed with 200 μL serum-free 1640 medium. After 24 h of transfection, cells were digested in each group before re-suspension with serum-free DMEM. The number of cells was counted and diluted to 3 × 105 cells/mL. Amounts of 100 μL cell suspension were added to the apical chambers, and 600 μL DMEM supplemented with 10% serum as a chemotactic factor was added in the basolateral chambers, according to the instructions of the Transwell chambers. Crystal violet was applied for cell staining. The cells penetrating the membrane were counted in three randomly selected fields of vision.

Electron Microscopy Detection

CRC HCT116 and SW480 cells in the logarithmic growth phase were transfected for 48 h. EBSS was added as starvation inducer to treat the cells for 30 min. The cells were washed once with PBS, centrifuged for 10 min, and collected into a centrifuge tube. Glutaraldehyde was used to fix the cells for 2–4 h. At 4°C, cells were washed with 0.1 mol/L PBS for 1 h during which the solution was replaced three times. At 0°C–4°C, cells were immersed in osmic acid fixative solution for 1–2 h. After soaking in acetone (30%, 70%, and 90%, 10 min/time) and in pure acetone another three times (10 min/time), cells were placed in the capsule supplemented with embedding medium. Polymerization was conducted in an oven at 37°C for 12 h and at 60°C for 24 h. The ultratome was introduced to obtain semi-thin sections, and a light microscopy was applied for examination and localization, followed by preparation of ultrathin sections (40–50 nm). Images were acquired under a transmission electron microscope (JEM-1011, JEOL Group, Beijing, China).

Acridine Orange Staining

The transfected cells (2 × 105 cells/well) were inoculated in a confocal culture dish until the cells’ adherence to the wall. Next, cells were incubated with acridine orange with a final concentration of 1 mg/L for 10 min, and they were observed and quantified by laser confocal fluorescence microscopy at excitation wavelength 488 nm and emission wavelength 510 nm (Leica, Mannheim, Germany).

Flow Cytometry

HCT116 and SW480 cell apoptosis after 24/48 h of culture and transfection was examined using the Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide (PI) double-staining kit (556547, Shanghai Solja Technology, Shanghai, China). The experimental procedures are indicated below. Deionized water was used to dilute 10× binding buffer to 1× binding buffer. HCT116 and SW480 cells in each group were collected after centrifugation at room temperature for 5 min. Subsequently, pre-cooled 1× PBS was applied to re-suspend the cells, followed by centrifugation for 5–10 min. After washing, cells were re-suspended by adding 300 μL 1× binding buffer. An amount of 5 μL Annexin V-FITC was mixed before incubation was conducted at room temperature for 15 min in the dark. An amount of 5 μL PI was added, and the mixture was ice bathed for 5 min in the dark. Flow cytometry (Cube6, Partec Group, Germany) was used to detect FITC at 480 and 530 nm and PI at excitation wavelengths of over 575 nm. The experiment was independently repeated three times.

Xenografts in Nude Mice

HCT116 and SW480 cells with strong growth viability were collected, and the concentration of HCT116 and SW480 cells was adjusted to 5 × 106 cells/mL. A total of 35 male nude mice, of specific pathogen-free grade aged 4 weeks old and weighing 14–16 g, was purchased from Shanghai SLAC Laboratory Animal (Shanghai Experimental Animal Center of Chinese Academy of Sciences, Shanghai, China) and included in this study. The nude mice were randomly assigned to the following 7 groups (n = 5 for each group): control (without treatment), empty vector (delivered with empty vector), si-lncRNA GAS5 (delivered with si-lncRNA GAS5), oe-lncRNA GAS5 (delivered with lncRNA GAS5 plasmid), miR-222-3p mimic (delivered with miR-222-3p mimic), miR-222-3p inhibitor (delivered with miR-222-3p inhibitor), and si-lncRNA GAS5 + miR-222-3p mimic (co-delivered with si-lncRNA GAS5 and miR-222-3p mimic).

After the nude mice underwent anesthesia via pentobarbital sodium, the cells were inoculated via the back of the right hind leg of the nude mice at a density of 1 × 106 cells (200 μL). All mice were fed under the same environment. The tumors were recorded once every 4 days, and the length and width of the tumors were recorded in detail. The tumor size was calculated according to the following equation: volume = length × width2/2. On the 20th day, the nude mice were sacrificed, and the tumors were dissected, with 3 tumors collected from each group.

Immunohistochemistry

Immunochemistry was performed with 4 μm formalin-fixed paraffin-embedded xenograft sections to evaluate PTEN levels, as described previously. The VECTASTAIN ABC Elite Kit and the ImmPACT DAB kit (Vector Laboratories, CA, USA) were used for visualization. The antibodies used were anti-PTEN (ab228466, 1:200; Abcam, Cambridge, MA, USA). Images were obtained using a Nikon Eclipse 80i microscope and analyzed by the NIS elements BR 3.22.11 software.

Statistical Analysis

All data were analyzed using SPSS 21.0 software (IBM, Armonk, NY, USA). Measurement data were expressed as the mean ± SD. Measurement data subject to a normal distribution between two groups were compared using the t test. Comparisons among multiple groups were assessed by one-way ANOVA, and p < 0.05 was considered statistically significant.

Author Contributions

L.L., H.-J.W., and T.M. designed the study. C.L., X.-H.Y., and Q.-S.W. collated the data, carried out data analyses, and produced the initial draft of the manuscript. B.J. and J.-F.Z. contributed to drafting the manuscript. All authors read and approved the final submitted manuscript.

Conflicts of Interest

The authors declare no competing interests.