Regulation of Colorectal Carcinoma Stemness, Growth, and Metastasis by an miR-200c-Sox2–Negative Feedback Loop Mechanism

Cancer cells are considered to be capable of unlimited proliferation and self-renewal, similar to stem cells. Moreover, a small number of cancer cells express stem cell markers and possess the stem cell–like ability to sustain tumor growth, metastasis, and recurrence. Previous studies suggested that carcinoma cells most likely transform into cancer stem cell (CSC)-like cells due to genetic alterations or microenvironment changes (1–3). Stemness-associated genes such as Sox2, KLF4, Bmi1, and Oct4 are involved in the maintenance of the stemness of embryonic stem cells, and more recent studies have demonstrated their expression in some human tumors, governing tumorigenesis and metastases (4–6). However, the functions of stemness-associated genes in control of CRC tumorigenesis and metastases remain elusive.

Growing evidence indicates that miRNAs are aberrantly expressed in many human cancers and involved in the initiation, development, and metastasis of cancers. Several miRNAs are emerging as important regulators of the proliferation and metastases of CSCs. The reduction of miR-34a in CD44⁺ prostatic CSCs assists prostate cancer development and metastasis (7). The metastasis and maintenance of cancer-initiating cells that rely on KLF4-Numb–like signaling are suppressed by miR-296 in lung cancer (8). Recent findings have noted a connection between miRNAs and CSCs in CRCs. MiR-93 is downregulated during the course of colon CSC differentiation into colon cancer cells and depresses the proliferation of colon CSCs (9). MiR-328 maintains a CSC-like side population (SP) phenotype in CRCs (10). MiR-200c is the predominant member of the miR-200 family, which suppresses epithelial–mesenchymal transition (EMT), and the downregulation of miR-200 family members in some tumors promotes invasion and metastasis (11–15). The significance of miR-200c in cancer biologic processes is becoming apparent. Moreover, there has been no published data about the role of miR-200c in CRC stemness. Therefore, we examined its expression and roles in stemness sustainment and metastasis in CRCs.

In addition, we identified Sox2 as a novel target of miR-200c. Together with Oct4 and NANOG, Sox2 encodes homeodomain proteins that are required for the early development and propagation of undifferentiated embryonic stem cells (16). Sox2 is involved in the proliferation or initiation of several cancers such as glioblastoma, gastric, and breast cancer (17–19). Sox2 is essential for the transformation of foregut basal progenitor cells to the esophagus and forestomach and cooperates with inflammatory signaling pathways to induce the formation of carcinomas (20).

In summary, miR-200c and its target may be involved in tumorigenesis and CSC-like cell transformation, but there is no direct evidence that the network including miR-200c and Sox2 is implicated in CRC stemness and metastasis. We therefore attempted to focus on the miR-200c-Sox2–related mechanism that transforms CRC stemness and modulates proliferation and metastasis.

Human CRC specimens were collected at Nanfang Hospital, Southern Medical University (Guangzhou, China), with written consent. Surgically removed tissues were immediately frozen in liquid nitrogen and stored at −80°C. The use of clinical materials for research purposes was approved by the Southern Medical University Institutional Board. The CRC cell lines SW480, SW620, HCT116, Lovo, and HT29 were obtained from the American Type Culture Collection (ATCC) and authenticated according to the recommendation of the ATCC. All cells were cultured in RMPI-1640 medium containing 10% FBS (Hyclone) in 5% CO₂ at 37°C.

Lentiviral constructs expressing (Lenti-miR microRNA precursor clone collection; System Biosciences) or repressing (miRZips lentiviral-based microRNA inhibition; System Biosciences) miR-200c were packaged using the pPACKH1 Lentivector Packaging Kit (System Biosciences) and were used to infect CRC cells to establish cells constitutively expressing or repressing miR-200c, respectively. Then, the miR-200c-overexpressing cell lines were transfected with Sox2 vectors not containing its 3′-untranslated region (UTR).

Total RNA, including miRNA from the tissue samples and cultured cells, was extracted using TRIzol (Takara) according to established protocols. The expression of miR-200c and RNU6B (U6 snRNA, a reference gene) was analyzed by quantitative reverse transcription PCR (qRT-PCR) using the All-in-One miRNA qRT-PCR Detection Kit (GeneCopeia). The PrimeScript RT Reagent Kit (Perfect Real Time, Takara) was used to generate cDNA for the detection of Sox2, Bmi1, Oct4, CD133, CD166, and β-catenin mRNA. Their expression levels were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA. The relative mRNA or miR-200c levels were calculated using the comparative C_t method (ΔΔC_t). Primer sequences for qRT-PCR are listed in Supplementary Table S1.

Protein extracts were obtained using a lysis buffer and then quantified by a bicinchoninic acid (BCA) protein assay (KeyGen Biotech). Equivalent amounts of cell lysates were separated using SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane (Roche Applied Sciences). The membrane was then blocked in TBST solution containing 5% non–fat milk and incubated with the primary antibody at 4°C overnight, followed by the appropriate second antibody. The bands were visualised using Pierce ECL Western Blotting Substrate (Thermo Scientific).

The proliferation, plate colony formation, migration, and invasion of transfected CRC cells were determined as previously described (21).

For soft assays, the base agar layer (1.32%) was prepared in 6-well plates, and a single-cell solution of 3 × 10⁴ cells/mL was prepared in the top agar solution (0.66%). For sphere assays, cells were cultured and suspended in serum-free DMEM/F12 medium (Hyclone) supplemented with basic fibroblast growth factor (20 ng/mL), EGF (20 ng/mL), leukemia-inhibitory factor (10 ng/mL; Invitrogen), and insulin (25 mg/mL; Sigma). The number of colonies and colospheres was observed for more than 2 weeks.

Luciferase reporter plasmids were generated by ligating oligonucleotides containing the wild-type (Wt) or mutant (Mut) putative target site of the Sox2 3′UTR into the psi-CHECK2 vectors (Promega). Cells were cotransfected with the luciferase vectors and the pre-miR-200c or the negative control pre-miRNA, using Lipofectamine 2000 (Invitrogen). After 48 hours, cells were lysed, and luciferase activity was determined using the Dual Luciferase Reporter Assay Kit (Promega) according to the manufacturer's instructions. To generate an miR-200c promoter vector, a 667-bp fragment containing the 2 binding sites of Sox2 was PCR-amplified and inserted into a PGL3- luciferase reporter vector (Promega). In addition, some mutation vectors related to the Sox2-binding site were constructed. These PGL3-derived vectors, the control pRL-TK Renilla plasmids (Promega), and Sox2-expressing vectors were cotransfected into HEK293 or SW480 cells using Lipofectamine 2000 Reagent (Invitrogen). Luciferase activities were assayed as the aforementioned methods. The primer sequences used for PCR amplification of plasmid construction are listed in Supplementary Table S2.

Cells were cultured, fixed with 1% formaldehyde, washed, harvested, and lysed, followed by sonication to produce chromatin of primarily mononucleosomal size. Immunoprecipitation was performed overnight with anti-Sox2 or anti-IgG antibodies. Protein–DNA complexes were recovered using protein G agarose beads, washed, and then eluted. Cross-links were reversed at 65°C overnight, and DNA was purified using reagents provided in the EZ-ChIP Chromatin Immunoprecipitation Kit (Millipore). The immunoprecipitated DNA was amplified by PCR for sequences containing Sox2-binding sites.

Four- to 6-week-old athymic BALB/c nude mice were obtained from the Central Laboratory of Animal Science at Southern Medical University and housed in laminar flow cabinets under specific pathogen-free conditions. All experimental procedures involving animals were performed in accordance with animal protocols approved by the Animal Care and Use Committee of Southern Medical University. Xenograft tumors were generated by subcutaneous injection of 2 × 10⁶ cells. From the seventh day after injection, the size of the tumor was measured as described previously (22). For orthotropic metastasis assays, nude mice were anesthetized, and their caeca were exteriorized by laparotomy. The subcutaneous tumors were cut into small masses and embedded into the mesentery at the tail end of the caecum. The gut was reposited to the abdominal cavity and subsequently closed with surgical sutures. Six weeks later, the mice were sacrificed, and all organs were resected for biopsy.

Quantitative values of all experiments are expressed as the mean ± SD. Differences among/between sample groups were analyzed by one-way ANOVA or the independent-samples t test. Relationships between miR-200c expression and clinicopathologic characteristics were tested using Fisher exact test. Differences were considered significant if P < 0.05: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

The average expression level of miR-200c was significantly decreased in 30 of 34 CRC specimens (P < 0.001) compared with their normal counterparts (Fig. 1A and B), with a 9.35-fold decrease in the colorectal cancer tissue samples. Furthermore, to investigate the clinicopathologic significance of miR-200c expression in patients with CRCs, the median relative expression level of miR-200c in the 34 CRC samples was recommended as the cutoff point for dividing miR-200c level into a low-expression group and a high-expression group. Correlation analysis showed that the miR-200c expression level was reversely correlated to tumor size (P < 0.05), serosal invasion (P < 0.05), lymph metastasis (P < 0.05), and tumor–node–metastasis (TNM) classification (P < 0.001; Supplementary Table S3). Relative miR-200c expression was significantly lower in highly metastatic CRC cell lines SW620 and Lovo, compared with the low metastasis cell lines SW480, HT29, and HCT116 (Fig. 1C). We reasoned that expression of miR-200c is negatively correlated with the metastatic potential of CRCs and that miR-200c may suppress CRC growth.

Potential targets of miR-200c were analyzed by bioinformatic algorithms. Sox2, Bmi1, and KLF4 were the predicted targets of miR-200c in TargetScan, Pictar, and microRNA.org in unison. Bioinformatic analysis indicated a putative miR-200c target site in the Sox2 3′UTR (Fig. 1D). MiRNA is inversely correlated with its target genes in expression. The basal expression of miR-200c was inversely correlated with Sox2 mRNA in the CRC cell lines (P = 0.037, r = −0.900), whereas the Bmi1 and KLF4 mRNAs were not inversely correlated with miR-200c in these cells (P = 0.144, r = −0.396; P = 0.136, r = −0.404; Fig. 2A). Accordingly, we chose Sox2 as a preferential, predicted target of miR-200c. Sox2 expression in colorectal cancer (91.2%, 31 of 34) was significantly more frequent than that in the corresponding normal tissue (Fig. 2B). Spearman correlation analyses demonstrated that Sox2 and miR-200c were again inversely related in expression (P = 0.01, r = −0.870; Fig. 2C). We also quantified the protein level of Sox2 in the CRC cell lines (Fig. 2D).

There was scarcely any expression of miR-200c in the HEK293 cells (23). First, we cotransfected the luciferase plasmids Wt Sox2 3′UTR or Mut Sox2 3′UTR and pre-miR-200c into the HEK293 cells and compared these cells with cells transfected with negative control pre-miRNA. We observed that the exogenous expression of miR-200c significantly decreased the luciferase activity of the Wt Sox2 3′UTR but not the Mut Sox2 3′UTR (P < 0.001; Fig. 2E). To determine whether endogenous miR-200c levels regulated Sox2 expression, SW620 and HCT116 cells were transfected with the Wt Sox2 3′UTR or Mut Sox2 3′UTR. The luciferase activity of Wt Sox2 3′UTR was lower in the HCT116 cells (P < 0.001; Fig. 2E), but the luciferase activity of Mut Sox2 3′UTR had no significant difference in HCT116 and SW620 cells. Therefore, Sox2 is a direct target of miR-200c, and the negative regulation of Sox2 by miR-200c is due to the miR-200c–binding sites in the Sox2 3′UTR. A recent publication reports that the miR-200 family promotes murine neural progenitor cell-cycle exit and differentiation by directly targeting Sox2 (24).

The miR-200c expression lentiviruses were stably transduced into SW480 and SW620 cells generating sub-cell lines SW480/miR-200c and SW620/miR-200c, which persistently overexpressed mature miR-200c (P < 0.05; Supplementary Fig. S1A); their controls were infected with lentiviruses containing empty vectors. Meanwhile, miR-200c was permanently knocked down in SW480 and HCT116 by the anti-miR-200c lentiviruses producing the SW480/zip-200c and HCT116/zip-200c sub-cell lines (P < 0.05; Supplementary Fig. S1B), and scramble was used as a control. We observed an inverse change in Sox2 mRNA and protein expression when the level of the miR-200c was altered (P < 0.05; Supplementary Fig. S1C and S1D). These results suggest that miR-200c regulates Sox2 expression through mRNA degradation and translation silencing. The cotransfection of miR-200c/Sox2 (not containing its 3′UTR) rescued Sox2 expression (Supplementary Fig. S1C).

The upregulation of miR-200c restrained cell proliferation compared with control cells, as measured by CCK-8 cell proliferation assays (P < 0.05; Fig. 3A) and plate colony formation assay (P < 0.001; Fig. 3B). However, reducing miR-200c in CRC cell lines yielded the opposite effect (P < 0.05; Fig. 3A and P < 0.001; Supplementary Fig. S2A). This evidence indicates that miR-200c could impede the proliferation of CRC cells in vitro.

The soft agar assays showed that overexpression of miR-200c caused a marked reduction of anchorage-independent growth ability, as indicated by the reduction in colony number and volume in soft agar (P < 0.001; Fig. 3C). Inversely, the clonogenicity in soft agar was observably increased in SW480 and HCT116 cells depleted of endogenous miR-200c in contrast with scramble-transfected cells (P < 0.001; Supplementary Fig. S2B). Therefore, miR-200c is involved in the regulation of tumorigenesis in CRC cells.

Migration and invasion assays showed that miR-200c overexpression markedly repressed the motility and invasiveness of SW480 and SW620 cells as compared with their control cells (P < 0.001; Fig. 3D and E). However, the knockdown of miR-200c significantly enhanced the migration and invasion of HCT116 and SW480 cells, compared with scramble-transfected cells (P < 0.001; Supplementary Fig. S2C and S2D).Therefore, miR-200c dramatically attenuated the migratory and invasive abilities of CRC cells.

As an indication of the function of miR-200c in CRC cell lines, we attempted to determine the mechanisms of its effects. Further analyses revealed that Sox2 overexpression increased CRC cell proliferation, migration, and invasion (Supplementary Fig. S3). However, the knockdown of Sox2 decreased the growth rate, colony formation, and migration of SW620 (25). The rescue of Sox2 expression almost completely restored the proliferation, tumorigenesis, migration, and invasion in SW480/miR-200c and SW620/miR-200c cells (Fig. 3). These results indicate that Sox2 is one of the significant functional targets of miR-200c in CRC cells.

To assess the effect of miR-200c on tumor growth in vivo, SW620/control, SW620/miR-200c, and SW620/miR-200c/Sox2 were injected subcutaneously on the hind limb of nude mice. Tumor growth in the SW620/miR-200c group was slower than that in the SW620/control and the SW620/miR-200c/Sox2 groups (P < 0.01). Tumors in mice injected with SW620/miR-200c/Sox2 cells showed no significant difference from SW620/control cells (P > 0.05; Fig. 4A and B). Immunostaining confirmed that the cell proliferation index Ki-67 and the CSC biomarkers β-catenin, CD133, and CD44 were downregulated by miR-200c and restored by Sox2 (Supplementary Fig. S4A–S4C).

To test the effect of miR-200c on the metastasis of CRC in vivo, tiny tumor masses of subcutaneous tumors were transplanted into the mouse cecal subserosa. Sixty percent (3 of 5) of mice in the SW620/control group or the SW620/miR-200c/Sox2 group had hepatic metastatic lesions, and the hepatic metastatic lesions in the 2 groups showed no significant differences. However, no mice in the SW620/miR-200c group had hepatic metastatic lesions (Fig. 4C). The number of hepatic metastatic lesions in mice of the SW620/miR-200c group was obviously reduced compared with the SW620/control group (P < 0.01), whereas enhanced expression of Sox2 increased the number of hepatic metastases of mice and blocked the inhibitory effects of miR-200c (Supplementary Fig. S4D). No metastatic nodules were discovered in the other organs in any group. These results collectively indicate that miR-200c upregulation profoundly suppresses tumor growth and metastasis in vivo and is partly reversed by Sox2.

It has been reported that Sox2 is vital for the maintenance of the self-renewal of embryonic stem cells and neural stem cells (NSC) as well as the tumorigenicity of glioblastoma tumor-initiating cells (17, 26). In other words, Sox2 plays an integral role in maintaining the stemness of adult cells as well as CSCs. Thus, we hypothesized that Sox2 regulator miR-200c may have a connection with stemness maintenance in CRC cells. Loss of miR-200c results in the acquisition of stemness, which is characterized by an increased stem cell sphere-forming capacity (Supplementary Fig. S5A) and an increased expression of CRC stem cell markers such as CD166, CD133, and β-catenin (Fig. 5A). Most essentially, loss of miR-200c causes SW480 and HCT116 cells to reverse differentiate to CSC-like cells, represented by the cells transitioning from spindle-shaped cells to round cells (Supplementary Fig. S5B).

Overexpressing Sox2 in CRC cells accompanied a decrease of miR-200c (Supplementary Figs. S1A and S1C and S5C). To further investigate the potential relationship between miR-200c and Sox2, we relied on bioinformatics. While analyzing a 2-kb region upstream of the transcription start site (TSS) of miR-200c using UCSC, TESS, and TFSEARCH, we noted that there were 2 Sox2 transcription factor–binding sites (TFBS) located within the miR-200c promoter. For convenience, the 2 TFBSs were named A and B (Fig. 5B). A reduction of the wild-type miR-200c promoter luciferase activity was observed upon upregulation of Sox2 in the HEK293 and SW480 cell lines (P < 0.001), and a similar effect was observed when B was mutated alone (P < 0.001). However, Sox2 overexpression did not result in further reduction of the luciferase activity when A was mutated alone or A and B were mutated together (Fig. 5C). These data indicate that Sox2 binds to specific promoter TFBS of miR-200c and inhibits transcription. The chromatin immunoprecipitaion (ChIP) analyses revealed that Sox2 is most significantly bound to TFBS A within the miR-200c promoter (Fig. 5D), indicating a direct association of Sox2 with the miR-200c promoter. These findings establish a negative feedback loop controlling Sox2 and miR-200c expression that regulates CRC stemness. This negative feedback loop mechanism was verified once again by the increase in miR-200c in the Sox2-overexpressing cells, causing suppression of Sox2 (Supplementary Fig. S5D).

Sox2 cooperates or interacts with some genes to govern embryonic stem cell differentiation (16, 27). KIT is a Sox2 enhancer–interacting gene assayed by 4C-seq (circular chromosome conformation capture) in human embryonic stem cells (28). Previous research indicated that KIT promotes proliferation and invasion by activating the PI3K–AKT pathway in some human carcinomas (29, 30). For this reason, we wondered whether miR-200c participates in the growth and metastasis of CRCs through the PI3K–Akt pathway. The upregulation of miR-200c restrained the phosphorylation of PI3K and AKT (p-PI3K and p-AKT) compared with control and blank cells; the restrained of p-PI3K and p-AKT by miR-200c was restored by Sox2 (Fig. 6A). On the other hand, the downregulation of miR-200c caused the opposite effect. Treatment with the PI3K–Akt inhibitor LY294002 could abrogate the activities of p-PI3K and p-AKT by inhibiting miR-200c (Fig. 6B). No change in the total protein amounts of PI3K and AKT (T-PI3K and T-AKT) was observed under any circumstance. In brief, the PI3K–AKT pathway has been considered relevant to the maintenance of CRC stemness, growth, and metastasis.

In this study, we evaluated the expression of miR-200c in CRC cell lines and clinical tissues. The findings indicate that a lower level of miR-200c is significantly associated with the growth and metastasis of CRC as well as a more advanced stage. MiR-200c abnormalities have been found in various tumors. Expression of miR-200c was significantly lower in esophageal adenocarcinoma in comparison to Barrett esophageal epithelium (31). MiRNA microarray analyses of noninvasive and invasive bladder urothelial carcinoma indicated that miR-200c was reduced in invasive lesions (32). These data may indicate that downregulation of miR-200c has potential as a cancer progression and metastasis biomarker. Overexpression of miR-200c and miR-141 in PC-3 cells impaired proliferation and survival (33). AC3E+ cells expressing miR-200 had obviously less tumorigenic potential in vivo (34). Increasing miR-200c in the highly metastatic MDA-MB-231 cells displayed remarkably reduced motility (35). These previous observations depicted that miR-200c is tumor repressor. Our data are consistent with previous evidence. However, Liu and colleagues reported that miR-200c was upregulated in non–small cell lung cancer (NSCLC) tissues compared with adjacent normal lung tissues (36). This inconsistent expression phenotype of miR-200c in tumors may be due to carcinoma heterogeneity. In the context of tumor development and progression, tumor cells display phenotypic and functional heterogeneity that might result in differential expression of some genes in different types of tumors or in different areas in the same tumor.

Functionally, inhibition of miR-200c reinforced proliferation, migration, and invasion in CRC cells. Correspondingly, overexpression of miR-200c attenuated proliferation, migration, and invasion. Furthermore, overexpression of miR-200c displayed significant suppression of growth and metastasis in vivo. Our data provide evidence to support the role of miR-200c in suppressing CRC growth and metastasis. However, our findings have yielded a contradiction to a recent study on the role of miR-200c in CRC cell growth; the miR-200c expression levels in the SW480 and HCT1116 cell lines in our study did not coincide with the data reported previously (37). These contrary findings may be due, in part, to the different source organization and empirical method models. With regard to the role of miR-200c in tumor growth, several recent studies have reported its conflicting roles (23, 38–40).

Sox2 was known to be a stem cell factor and was identified as a target of miR-200c. In our study, Sox2 enhanced CRC cell proliferation, migration, and invasion. MiR-200c reduced the expression of Sox2, which caused the restoration of the growth and metastasis inhibited by miR-200c in vitro and in vivo. Knockdown of miR-200c strongly increased the CRC stemness phenotype. We propose that Sox2 may be necessary to maintain a stemness phenotype or that miR-200c influences stemness in a direct modulation manner. Cancer stem cell theory posits that a fraction of cancer cells maintain typical stemness properties, including the capacities of self-renewal, differentiation, and tumorigenicity, and are thought to be crucial for the maintenance of uncontrolled tumor proliferation as well as metastasis. We propose that suppressing miR-200c, leading to an increased expression of Sox2, maintains the stemness of CRC cells. On the basis of these observations, we speculate that miR-200c restrains cell proliferation and metastasis by eliminating a stem cell phenotype. Our findings are also supported by data showing that ZEB1 promotes tumor cell dissemination and tumorigenicity by suppressing stemness-inhibiting miRNAs, including miR-200c, miR-203, and miR-183 (39).

Increasing evidence indicates that miRNAs play crucial roles in carcinoma cell proliferation and metastasis. Previous studies showed that miR-200c–inhibited CRC cell migration and invasion correlated with loss of the EMT phenotype (37). In the present study, we found that the miR-200c-Sox2 negative feedback loop mechanism was involved in the regulation of CRC cell proliferation and metastasis. Sox2, occupying the promoters of miR-200c, in turn decreased expression of miR-200c; such a loop might well lead to the reactivation of Sox2. In a previous report, a miR-200c-Sox2 feedback loop contributed to cell-cycle exit and the neuronal differentiation of neural stem/progenitor cells (24). During induced pluripotent stem cell (iPSC) reprogramming in mouse embryonic fibroblasts, Oct4 specifically activates the mir-141/200c cluster and Sox2 specifically activates the mir-200a/b/429 cluster by binding to their promoters (41). Altogether, miR-200c might be regulated by a distinct set of transcriptional activations/repressions of Sox2. We noted a reciprocal regulation of miR-200c and Sox2 in adult tumor cells paradigms that might be different from the pluripotent/multipotent stem/progenitor cell situation.

In the EMT process, miR-200s enhance E-cadherin expression by directly targeting ZEB1 and ZEB2, which encode transcriptional repressors of E-cadherin (11, 12, 42, 43). Apart from EMT, the inflammatory signaling harnessed by miR-200c is helpful in epithelial cell transformation and mammary cell tumorigenesis (44). The PI3K–Akt signaling pathway controls fundamental cellular processes such as cell survival, growth, proliferation, cell repair, cell migration, and angiogenesis and is constitutively activated in several cancer types, including CRCs. The Sox2 enhancer-interacting gene KIT could activate multiple downstream signaling cascades, including the PI3K–AKT pathway (45, 46). Consequently, we speculated that miR-200c would negatively suppress Sox2 to inhibit the activation of the PI3K–AKT pathway. Here, we reported that overexpression of miR-200c results in low levels of p-PI3K and p-AKT and that this phenomenon could be rescued by Sox2. Luo and colleagues have reported that the miR-20a/miR-200c-PTEN-AKT axis regulates EMT-associated CSC enrichment of epithelial ovarian cancer (EOC) cells (47). In this axis, miR-200c expression was significantly higher in EOC spheroid cells, which produce large numbers of CSC cells, and activated the PI3K–AKT pathway. In contrast, we presented evidence indicating that the downregulation of miR-200c confers the characteristics of CSC cells and stimulates the PI3K–AKT pathway in CRCs. The PI3K–AKT pathway can be triggered through various mechanisms, including increasing the expression of its positive regulators such as EGF receptor (EGFR) and RAS and decreasing the expression of its negative regulators such as PTEN. Our data indicate that the induction of the PI3K–AKT pathway was dependent on the increased expression of Sox2. PI3K activity was inhibited by LY294002 in the enterocyte-like differentiation program of colon CSC cells, indicating that PI3K signaling is, in part, necessary for the maintenance of colon CSC cells (48). In this regard, it is reasonable to assume that PI3K–AKT pathway activation is relevant to the maintenance of CRC stemness, growth, and metastasis when miR-200c is suppressed.

In summary, in addition to functioning as a tumor suppressor by blocking the PI3K-Akt signaling pathway through an miR-200c-Sox2 negative feedback loop mechanism, miR-200c also acts as a stemness inhibitor (Fig. 6C). Although miRNA-based therapeutics are still in their infancy, our findings indicate that miR-200c may be a promising therapeutic target for the inhibition of CRC growth and metastasis.

No potential conflicts of interest were disclosed.

Conception and design: Y.-X. Lu, X.-N. Li

Development of methodology: M. Zhou, Y. Liu, C. Zhang, M. Hong

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y.-X. Lu, L. Yuan, M. Zhou, Y. Liu, M. Hong

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y.-X. Lu, J.-P. Li, X.-N. Li

Writing, review, and/or revision of the manuscript: Y.-X. Lu, M. Zhou, J.-P. Li, L. Zheng, X.-N. Li

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y.-X. Lu, X.-L. Xue, M. Zhou, C. Zhang, X.-N. Li

Study supervision: X.-N. Li

This work is supported by the National Natural Science Foundation of China (Nos 81272758 and 81302158) and the Science and Technology Project of Guangdong Province (2009B030803041).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.