Objective Long non-coding RNAs (lncRNAs) are emerging as key molecules in cancers, yet their potential molecular mechanisms are not well understood. The objective of this study is to examine the expression and functions of lncRNAs in the development of colorectal cancer (CRC).
Methods LncRNA expression profiling of CRC, adenoma and normal colorectal tissues was performed to identify tumour-related lncRNAs involved in colorectal malignant transformation. Then, we used quantitative reverse transcription PCR assays to measure the tumour-related lncRNA and to assess its association with survival and response to adjuvant chemotherapy in 252 patients with CRC. The mechanisms of CCAL function and regulation in CRC were examined using molecular biological methods.
Results We identified colorectal cancer-associated lncRNA (CCAL) as a key regulator of CRC progression. Patients whose tumours had high CCAL expression had a shorter overall survival and a worse response to adjuvant chemotherapy than patients whose tumours had low CCAL expression. CCAL promoted CRC progression by targeting activator protein 2α (AP-2α), which in turn activated Wnt/β-catenin pathway. CCAL induced multidrug resistance (MDR) through activating Wnt/β-catenin signalling by suppressing AP-2α and further upregulating MDR1/P-gp expression. In addition, we found that histone H3 methylation and deacetylases contributed to the upregulation of CCAL in CRC.
Conclusions Our results suggest that CCAL is a crucial oncogenic regulator involved in CRC tumorigenesis and progression.
- COLORECTAL CANCER
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Significance of this study
What is already known on this subject?
Long non-coding RNAs (lncRNAs) may predict the clinical outcomes of colorectal cancer (CRC) and may be used as diagnostic or prognostic markers.
What are the new findings?
We identified colorectal cancer-associated lncRNA (CCAL) as a key regulator of CRC progression. Overexpression of CCAL correlated with poor overall survival and predicted a poor response to adjuvant chemotherapy of patients with CRC.
We demonstrated that CCAL promoted CRC progression and induced multidrug resistance (MDR) by targeting activator protein 2α (AP-2α), which in turn activated Wnt/β-catenin pathway and upregulated MDR1/P-gp expression.
How might it impact on clinical practice in the foreseeable future?
LncRNA-CCAL may be as a biomarker for prognosis and therapeutic outcome in patients with CRC.
Colorectal cancer (CRC) is the second leading cause of cancer-related mortality in the Western world.1 Up to 90% of patients can be cured by surgery if the CRC is detected at an early stage. Unfortunately, patients with CRC are very often diagnosed at advanced stages, and the prognosis is poor.2 ,3 Chemotherapies are routinely used to treat patients with CRC with advanced disease. However, the response is often not satisfactory due to the lack of predictive markers to identify patients who are more likely to benefit from such treatments.4 ,5 Although remarkable progress has been made in past decades, the molecular mechanisms underlying CRC carcinogenesis remain to be elucidated.
Long non-coding RNAs (lncRNAs) represent a subgroup of non-coding RNAs of more than 200 nucleotides in length.6 ,7 The lncRNAs have emerged as a new aspect of biology, with evidence suggesting that they are frequently cell-type specific, contribute important functions to numerous systems and may interact with known cancer genes.7–9 A handful of studies10–19 have found that lncRNAs may be important players in cancer biology, typically resulting in aberrant expression of gene products that contribute to the progression of a variety of human tumours.20–23 Moreover, lncRNA expression may confer clinical information about tumour outcomes and have utility as diagnostic or prognostic markers.11 ,14 ,16 ,24 Nevertheless, the clinical significance and biological mechanisms of lncRNAs in the progression of CRC remain largely unknown.
In the present study, differentially expressed lncRNAs from normal colorectal tissues, adenoma and CRC were profiled to study their potential roles in tumorigenesis and progression of CRC. A colorectal cancer-associated lncRNA (CCAL) that displayed a remarkable trend of increasing expression levels from normal colorectal to adenoma and carcinoma tissues was identified and selected for further validation and functional analysis.
Experimental methods, including statistical analysis are described in detail in the online supplementary information.
CCAL, a novel lncRNA involved in colorectal malignant transformation
To search for potential protein-coding RNAs and lncRNAs involved in colorectal malignant transformation, we globally analysed the protein-coding RNA and lncRNA expression profiles of normal colorectal tissues, colorectal adenoma and CRC tissues. The comparison of the protein-coding RNA and lncRNA expression levels among the three groups is shown in online supplementary tables S1 and S2. We focused on 34 protein-coding RNAs and 9 lncRNAs whose expression decreased or increased with the progression from normal mucosa to adenoma to carcinoma. Then, we used gene coexpression networks to cluster the transcripts into phenotypically relevant coexpression modules.25 ,26 Coexpression networks of protein-coding RNAs and lncRNAs were constructed (see online supplementary figure S1). Because coexpression modules may correspond to biological pathways and many functions of protein-coding RNAs can be found in National Center for Biotechnology Information (NCBI) RefSeq,27 we focused on the coexpression modules that had a high proportion of protein-coding RNAs in the CRC coexpression network. A non-annotated lncRNA, CCAL, in CRC was identified by this method. In the cancer coexpression network, CCAL is connected to one lncRNA and 11 protein-coding genes that are enriched for gene products involved in tumour cell differentiation, proliferation, apoptosis and drug resistance (see online supplementary figure S1). Next, we performed a rapid amplification of cDNA ends analysis to identify the 5′ and 3′ ends of the CCAL transcript. The transcription start and termination sites and sequences of full-length CCAL cDNA are presented in online supplementary figure S2. Using the Coding Potential Calculator for annotating human lncRNA genes,28 we classified CCAL as an lncRNA, as the transcript had no protein-coding potential (see online supplementary figure S3). The expression pattern of CCAL (figure 1A) was validated using quantitative reverse transcription PCR (qRT-PCR) analysis (figure 1B). The transcript for CCAL was located in the nucleus and in the cytoplasm of LoVo, LoVo/5-Fu and NCM460 cells (figure 1C). The qRT-PCR analysis revealed significantly higher CCAL expression in LoVo cells than that in the normal human colon epithelial cell line, NCM460 cells. Moreover, the expression level of CCAL in LoVo/5-FU cell line was significantly higher than that in LoVo cells (figure 1C).
High CCAL expression and prognosis
We hypothesised that CCAL might act as an oncogenic lncRNA, and if so, CCAL overexpression should be a frequent event in CRC. Therefore, we used qRT-PCR to determine the CCAL expression levels in tumour and paired non-tumour tissues obtained from 252 patients with CRC. A summary of clinical information for these patients is shown in online supplementary table S3.
To analyse the association between CCAL expression and the clinicopathological characteristics of patients with CRC, we used the median ratio between CCAL expression in tumour to that in non-tumour samples (T:N) to dichotomise the CRC cases. We found that high T:N CCAL expression ratio was associated with a shorter survival (p=0.007; figure 1D). This association could be due to the changes of CCAL expression levels in the tumour tissue, the surrounding non-tumorous tissue or a combination of both. To distinguish between these possibilities, we separately analysed CCAL expression in tumours and paired non-tumorous tissues. High CCAL expression levels in tumours were associated with a shorter survival (p=0.003; figure 1E). No significant association between CCAL expression in non-tumorous tissue and survival was observed (p=0.55; figure 1F). In the univariate analysis, high CCAL expression in tumours (HR, 1.86; 95% CI 1.18 to 2.93; p=0.008), tumor node metastasis staging (HR, 2.60; 95% CI 1.67 to 4.06; p<0.001) and lymph node metastasis (HR, 2.63; 95% CI 1.68 to 4.12; p<0.001) were significantly associated with survival (see online supplementary table S4). In a multivariate Cox regression model, high CCAL expression levels in tumours were associated with a poor survival prognosis (HR, 2.25; 95% CI 1.35 to 3.74; p=0.002) independent of the other clinical covariates (see online supplementary table S4).
CCAL expression in colorectal epithelial cells
To further confirm the differential expression of CCAL between CRC and non-tumoral tissues, we used in situ hybridisation to visualise CCAL expression in tumour and paired non-tumour tissues. CCAL was expressed at higher levels in the tumour cells of human CRC samples than in normal tissues (figure 1G). These results suggested that CCAL overexpression plays a role in colorectal carcinogenesis.
CCAL expression levels and therapeutic outcome
To examine whether CCAL expression can predict response to adjuvant therapy, we analysed the associations between CCAL expression and the therapeutic outcomes in patients with stage II and stage III CRC treated with adjuvant chemotherapy. Patients with stage I or stage IV disease were excluded since stage I patients typically do not receive adjuvant therapy and stage IV patients often received palliative treatment. The chemotherapy regimens (with or without leucovorin, levamisole or cisplatin) were primarily fluorouracil based. Kaplan-Meier analyses revealed that high CCAL expression was associated with a poor prognosis in stage II (p=0.0007) and stage III (p=0.0004) patients (figure 2). For individuals who received adjuvant therapy, high CCAL expression was associated with a poor therapeutic outcome in the patients with stage II and stage III cancer (p=0.042), and patients with stage II cancer alone (p=0.044) or patients with stage III cancer alone (p=0.027, figure 2). Multivariate Cox regression revealed that high CCAL expression predicted poor outcomes (HR, 1.92; 95% CI 1.09 to 3.37; p=0.023) and that adjuvant chemotherapy was associated with favourable survival (HR, 0.39; 95% CI 0.27 to 0.75; p<0.001) independent of other clinical covariates (see online supplementary table S5). Therefore, CCAL expression emerged as an independent predictor of the response to adjuvant chemotherapy.
The biological functions of CCAL
To evaluate the oncogenic properties and effects of CCAL on CRC, we established CRC cell lines with CCAL stable overexpression or knockdown (see online supplementary figure S4A–H). Knockdown CCAL expression in Lv-LoVo-CCAL-shRNA1 cells leads to a significant decrease in tumour cell proliferation. On the other hand, CCAL overexpression in Lv-LoVo-CCAL cells leads to a significant increase in tumour cell proliferation (figure 3A, B). The cell cycle and apoptosis assays indicated that CCAL suppressed cell apoptosis and led to G1 arrest (figure 3C, D). In addition, CCAL overexpression promoted migration and invasion in the Lv-LoVo-CCAL cells measured by wound healing and transwell assays. In contrast, knockdown CCAL in Lv-LoVo-CCAL-shRNA1 cells resulted in attenuated migration and invasion of tumour cells (figure 4A, B). Similar results were obtained using the HCT116 cells (see online supplementary figures S5A–D and S6A,B).
To determine the effects of CCAL on tumorigenesis in vivo, Lv-LoVo-CCAL-shRNA1 cells, Lv-LoVo-CCAL cells or appropriate control cells were subcutaneously injected into nude mice. Knockdown CCAL significantly decreased tumour growth in vivo compared with the shRNA controls. In contrast, xenograft tumours from Lv-LoVo-CCAL cells grew significantly faster than the tumours from control cells (figure 5A, B). Using qRT-PCR and in situ hybridisation analysis, we confirmed the CCAL knockdown or overexpression in the xenograft tumours generated from Lv-LoVo-CCAL-shRNA1 cells or Lv-LoVo-CCAL cells, respectively (figure 5C, D).
To further investigate the in vivo effects of CCAL, tumour cell proliferation and apoptosis were assessed using proliferation-related nuclear antigen ki67 immunoreactivity and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay. As shown in figure 5E, F, upregulation of CCAL promoted tumour cell proliferation and deceased apoptosis, and downregulation of CCAL inhibited tumour cell proliferation and promoted apoptosis. Taken together, these data indicate that CCAL plays an important role in CRC progression.
Histone H3 methylation and deacetylation is involved in the upregulation of CCAL
To understand the mechanism of CCAL regulation in CRC, we performed the promoter luciferase transcriptional reporter arrays. We found that CCAL transcription is predominantly controlled by a region spanning −2000 to +50 nucleotides relative to the transcription start site (see online supplementary figure S7). We observed that CCAL upregulation in LoVo cells was associated with decreased histone H3 methylation and increased histone H3 acetylation across the CCAL promoter region when compared with the normal human colon epithelial cell line NCM460 (figure 6A, B).
To determine whether CCAL is downregulated by histone methylation and deacetylation in vivo, we measured histone H3 methylation and acetylation levels across the CCAL promoter in 30 primary CRC tissues and their matched normal colon tissues. We found that histone H3 methylation levels in the CCAL promoter were significantly decreased in the CRC samples compared with normal colon tissues (figure 6C). However, histone H3 acetylation levels in the CCAL promoter were significantly increased in the CRC samples compared with normal colon tissues (figure 6D). We also analysed the possible CpG islands in the promoter (2000 nucleotides) of CCAL. The possible CpG island region is shown in online supplementary figure S8. Seven sets of primers were used for methylation-specific PCR analysis (see online supplementary table S6). The results indicated that the methylation level in the CpG island region of CCAL was significantly lower in the CRC tissue samples compared with the corresponding normal tissue samples (figure 6E). Collectively, these data revealed that histone H3 methylation and deacetylation may contribute to the upregulation of CCAL in the CRC.
CCAL downregulates AP-2α expression via the ubiquitin-proteasome pathway
Several recent studies have demonstrated that many lncRNAs regulate molecular pathways via their interactions with proteins.17 ,29 ,30 To test whether CCALs affect the biological behaviours of CRC cells in a similar way, we sought to identify the proteins that are associated with CCAL using an RNA-pull-down assay. Among all of the proteins identified by mass spectrometry, only AP-2α was detected by western blotting from three independent RNA pull-down assays (figure 7A). We further performed RNA immunoprecipitation with an antibody against AP-2α using cell extracts from the LoVo cells. We observed more CCAL enrichment using the AP-2α antibody than a non-specific antibody (IgG control) (figure 7B). We further tested the ability of the AP-2α protein to bind to CCAL using an electrophoretic mobility shift assay. The results clearly indicated a mobility shift for the AP-2α protein (figure 7C). AP-2α binds to CCAL with 1:1 stoichiometry, as determined by electrophoretic mobility shift assay (figure 7D). Deletion-mapping analyses identified a 513-nt region at the 3′ end of CCAL (CCAL-3′-513) that is required for its association with AP-2α (figure 7E). Together, these results demonstrate a specific association between CCAL and AP-2α.
We sought to determine the functional relevance of the association between CCAL and AP-2α. We detected a downregulation of the AP-2α protein upon CCAL overexpression in LoVo and HCT116 cells (figure 7F). However we did not observe a reduction in AP-2α mRNA levels (see online supplementary figure S9A, B). Similar results were obtained in vivo by western blot and immunohistochemistry (IHC) analysis using the xenograft tumours, which had knockdown CCAL or CCAL overexpression (figure 7G, H). No reduction in AP-2α mRNA levels was observed in vivo (see online supplementary figure S9C). To determine whether CCAL reduced AP-2α protein stability, we treated CCAL-overexpressing Lv-LoVo-CCAL and control LoVo cells with the protein synthesis inhibitor cycloheximide or the proteasome inhibitor MG-132. As shown in figure 7I, J, only MG-132 abolished the downregulation of AP-2α protein levels in Lv-LoVo-CCAL cells overexpressing CCAL. These findings indicate that the ubiquitin-proteasome pathway might be required for CCAL-mediated reduction of AP-2α protein.
CCAL activates the Wnt/β-catenin pathway by suppressing AP-2α
Aberrant Wnt/β-catenin signalling plays an important role in CRC.31 ,32 Recently, Li and Dashwood33 have demonstrated that AP-2α can attenuate β-catenin/TCF-4 interactions in CRC. To test whether CCAL-mediated downregulation of AP-2α plays a role in Wnt/β-catenin signalling, we cotransfected the plasmids pcDNA-AP2a-HA and pcDNA-β-catenin-His into the LoVo cells. Qualitative IP results using the whole cell lysates showed that AP-2α, APC and β-catenin can form a complex (figure 8A). In addition, we found that overexpression of CCAL downregulated AP-2α which, in turn, increased β-catenin expression and decreased expression of c-myc, cyclin D1 and E-cadherin (figure 8B, C). On the other hand, knockdown CCAL upregulated AP-2α expression, decreased β-catenin expression and led to increased expression of c-myc, cyclin D1 and E-cadherin (figure 8B, C). Consistent with these results, the expression levels of β-catenin, c-myc, cyclin D1 and E-cadherin were increased in the LoVo-cell xenograft tumours that overexpressed CCAL. The expression levels of β-catenin, c-myc, cyclin D1 and E-cadherin were decreased in the xenograft tumours with knockdown CCAL (figure 8D). Together, these data indicate that the ability of CCAL to enhance proliferation, cell-cycle progression, invasion and migration is in large part attributable to its ability to inhibit AP-2α, which subsequently activates the Wnt/β-catenin signalling pathway.
CCAL modulates MDR1/P-glycoprotein expression by activating the Wnt signalling pathway
Multidrug resistance (MDR) constitutes a major obstacle to successful chemotherapy in patients with CRC. We sought to explore the potential mechanisms of MDR using the newly established MDR cell line LoVo/5-FU, which exhibits drug resistance to common chemotherapeutics. We found that the expression level of CCAL was significantly higher in LoVo/5-FU cells than that in LoVo cells and the normal human colon epithelial cell line NCM460 (figure 8E). Knockdown CCAL expression in Lv-LoVo/5-FU cells increased the sensitivity to 5-FU. In contrast, CCAL overexpression in Lv-LoVo/5-FU cells attenuated the sensitivity to 5-FU. In the rescue experiments, knockdown AP-2α partially reversed the effects of CCAL inhibition on the sensitivity to 5-FU (figure 8F).
It is well documented that MDR has frequently been associated with elevated expression levels of the MDR1/P-gp (P-gp, encoded by the MDR1) gene in several types of cancer.34 We hypothesised that CCAL might induce MDR through activating Wnt/β-catenin signalling and upregulating MDR1/P-gp expression. To test this hypothesis, the expression levels of AP-2α, β-catenin and MDR1/P-gp were analysed in LoVo/5-FU cells with either knockdown CCAL or CCAL overexpression. Overexpression of CCAL downregulated AP-2α expression levels, increased β-catenin expression and led to upregulated expression of MDR1/P-gp, which have been associated with MDR in CRC (figure 8G, H). AP-2α-siRNA can partially reverse the effects of CCAL-shRNA1 on the decreased expression of MDR1/P-gp through activating the Wnt/β-catenin signalling pathway (figure 8G, H). Together, these data suggest that lncRNA-CCAL regulates CRC progression and MDR through activation of Wnt/β-catenin signalling by suppressing AP-2α and leads to upregulation of MDR1/P-gp expression (figure 8I).
In this study, we compared the lncRNA profiles of CRC, adenoma and normal colorectal tissues using the Arraystar LncRNA Expression Microarray. Only one lncRNA, CCAL, displayed a remarkable trend of increased expression levels with the progression from normal colorectal tissues to adenoma and carcinoma. We validated the CCAL expression patterns in the CRC clinical samples. In addition, we showed that CCAL expression was associated with patient survival and predicted the response to adjuvant chemotherapy. Our findings suggested that CCAL plays an important role during CRC tumorigenesis. In this regard, our data contribute to a growing body of literature supporting the importance of non-annotated lncRNA species in the field of cancer research.8 ,11 ,12 ,14–16 ,35 ,36
In the current study, we demonstrated an association between CCAL expression levels and CRC prognosis or therapeutic outcome. A robust association of high CCAL expression in tumours with poor survival was confirmed in 252 CRC samples. The association was independent of other clinical covariates, indicating that CCAL expression may be a useful prognostic biomarker to help identify patients at a higher risk of CRC progression. In addition, the patients whose tumours had increased CCAL expression had a poor response to adjuvant chemotherapy. These results indicate that CCAL status in tumours may be a useful tool for estimating prognosis of a patient with CRC and for selecting patients who are likely to benefit from adjuvant chemotherapy to prevent relapse.
We investigated the mechanisms by which CCAL exerts its function and modulates malignant CRC phenotypes in vitro and in vivo. Our data clearly indicated that silencing CCAL expression inhibited CRC cell proliferation, cell migration and invasion and induced cell apoptosis in vitro and in vivo. The CCAL transcript was found to be associated with AP-2α to promote AP-2α protein ubiquitination and subsequent degradation. We also found that the ability of CCAL to enhance proliferation, cell-cycle progression, invasion and migration is in large part attributed to its ability to inhibit AP-2α and subsequent activation of the Wnt/β-catenin signalling pathway. These findings provide additional evidence that CCAL plays an important role in CRC tumorigenesis and progression.
Recent studies have demonstrated the functional roles of lncRNAs,7–9 and provided insights into the molecular mechanisms by which lncRNAs function in a variety of human tumours.10–17 However, the mechanisms regulating lncRNA expression in CRC have not been thoroughly elucidated. Whether the epigenetic regulatory factors, such as DNA methylation or histone acetylation, can regulate the expression of lncRNAs remains largely unknown. A recent report found that histone acetylation is the key factor controlling lncRNA-LET expression during cancer progression.17 Our results revealed that H3 methylation and deacetylation play a critical role in the regulation of CCAL expression in CRC. Our results, along with recent studies,37 ,38 highlight the relationship between epigenetic regulation of lncRNAs and provide a novel field for lncRNA study in the future.
Overcoming MDR is critical for developing effective therapeutic drugs for CRC.39 Several studies have found that miRNAs might play a key role in MDR.40 MDR is mediated by the overexpression of MDR1/P-gp. The transcriptional activation of the MDR1 gene is a highly regulated complex event and is associated with several signalling pathways.41 However, the mechanisms by which lncRNAs regulate MDR in human cancer are not clear. Our present results indicate that knockdown CCAL upregulated AP-2α expression levels and decreased β-catenin expression and subsequently led to decreased expression of MDR1/P-gp. AP-2α-siRNA can partially reverse the effects of CCAL-shRNA1 on the decreased expression of MDR1/P-gp through activating the Wnt/β-catenin signalling pathway. These results provide the new mechanistic insights into the regulation of MDR in CRC by CCAL and explain the clinical association of high CCAL expression with poor adjuvant chemotherapeutic outcome in the patients with CRC observed in this study. To our knowledge, this is the first study reporting the effects of lncRNAs on MDR1/P-gp expression and the MDR phenotype, which highlights the association of lncRNAs with MDR in cancer and opens up a new field for lncRNA study.
Taken together, our results indicate that CCAL is an oncogenic lncRNA that promotes the tumorigenesis, MDR and progression of CRC. This finding suggests that lncRNAs may be important targets for tumour therapy.
The authors thank the patients and clinicians for their contributions to this study.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online supplement
YM, YY and FW contributed equally.
Correction notice This article has been corrected since it published Online First. The labels in figure 6 have been corrected.
Contributors HQ and HW: study concept and design, analysis and interpretation of data, drafting of the manuscript, statistical analysis, obtained funding, study supervision. YM, YY, FW, PZ and HW: study concept and design, acquisition of data, analysis and interpretation of data, drafting of the manuscript, statistical analysis. M-PM, QW, ZY, HZ, WL and NC: material support, acquisition of data.
Funding This work is supported by grants from the National Natural Science Foundation of China (No.81372615; No.81200264; No. 81472262; No. 81230057), the National High Technology Research and Development Program (863 Program; No. 2014AA020803), the Shanghai Science and Technology Development Fund (No.12140902300 and No.12410707400), and the Shanghai Health System Outstanding Young Talent Training Plan (No. XYQ2013118).
Competing interests None declared.
Patient consent Obtained.
Ethics approval The Committee for Ethical Review of Research Involving Human Subjects of the Shanghai Tenth People's Hospital Affiliated to Tongji University (Shanghai, China)
Provenance and peer review Not commissioned; externally peer reviewed.