How to Explain the Contradiction of microRNA 200 c Expression and Survival in Solid Tumors ? : a Meta-analysis

MicroRNAs (miRNAs) are 21-23 nucleotides long non-coding RNA molecules that regulate various biological processes by modulating the expression of their targeted mRNAs (Rana, 2007; Nilsen, 2007; Filipowicz et al., 2008). Downregulation of global miRNAs can be found in tumors compared with normal tissues by miRNA expression profiling analysis which reveal the potential corelation of miRNAs and tumorigenesis (Lu et al., 2005; Thomsom et al., 2006; Xu et al., 2013). MiRNA 200 family consists of five miRNAs including miRNA-200a, -200b, -200c, -141 and -429 and therein miRNA 200c is widely investigated in recent years. It is documented that the suppression of miRNA200c by interleukin 6 (IL6) drives transformation and tumorigenesis (Rokavec et al., 2012). Increasing evidence also demonstrate that miRNA 200c plays an critical role in epithelial-mesenchymal transition (EMT) which enhance the migration and invasion of tumor cells and eventually lead to metastasis (Mongroo et al., 2010; Dykxhoorn 2010). Moreover, restoration of miRNA 200c to certain solid tumors showed the ability to increase the sensitivity to chemo drug or radiation by various mechanisms (Cochrane et al., 2009; Cittelly et al., 2012; Kopp et al., 2012; Chang et al., 2013; Shi et al., 2013). However, the prognostic relevance of miRNA 200c


Introduction
MicroRNAs (miRNAs) are 21-23 nucleotides long non-coding RNA molecules that regulate various biological processes by modulating the expression of their targeted mRNAs (Rana, 2007;Nilsen, 2007;Filipowicz et al., 2008).Downregulation of global miRNAs can be found in tumors compared with normal tissues by miRNA expression profiling analysis which reveal the potential corelation of miRNAs and tumorigenesis (Lu et al., 2005;Thomsom et al., 2006;Xu et al., 2013).
MiRNA 200 family consists of five miRNAs including  and therein miRNA 200c is widely investigated in recent years.It is documented that the suppression of miRNA200c by interleukin 6 (IL6) drives transformation and tumorigenesis (Rokavec et al., 2012).Increasing evidence also demonstrate that miRNA 200c plays an critical role in epithelial-mesenchymal transition (EMT) which enhance the migration and invasion of tumor cells and eventually lead to metastasis (Mongroo et al., 2010;Dykxhoorn 2010).Moreover, restoration of miRNA 200c to certain solid tumors showed the ability to increase the sensitivity to chemo drug or radiation by various mechanisms (Cochrane et al., 2009;Cittelly et al., 2012;Kopp et al., 2012;Chang et al., 2013;Shi et al., 2013).
However, the prognostic relevance of miRNA 200c

Search strategy
Pubmed was searched in November 2013.The following keywords were combined: "microRNA 200c" and "cancer".No language and time restrictions were made.

Inclusion criteria
In order to be eligible, studies had to: (i) discuss the relevance of microRNA 200c expression and OS, (ii) provide sufficient data for extracting or estimating HR and its 95% CI.

Exclusion criteria
Studies were excluded from the analysis if: (i) the articles were not written in English, (ii) the articles were review or letters, (iii) studies did not investigate solid tumors , (iv) primary data for meta-analysis could not be extracted or calculated as lacking essential information.

Data extraction
The primary data were hazard ratio (HR) and its 95% confidence interval (CI) of OS.Three reviewers (HY Wang, J Shen, CP Jiang) independently extracted the primary data and baseline characteristics of included studies.Only the Kaplan-Meier survival curves but not HR and its 95% CI were given in some included articles.As for these articles, methods according to the work of Parmar, Williamson and Tierney were used to calculate HR (Parmar et al., 1998;Williamson et al., 2002;Tierney et al., 2007).The baseline characteristics included first author, publication year, tumor type, study size, sample site, methods to detect microRNA 200c expression, percentage of patients with high miRNA 200c expression and HR estimation methods.Disagreements were resolved by discussion.

Statistical methods
LogHR and standard error (SE) were calculated by using the software designed by Matthew Sydes and Jayne Tierney (Medical Research Council Clinical Trials Unit, London, UK) (Tierney et al., 2007).The pooled HR was gained using fixed or random-effects models according to the heterogeneity between studies.Heterogeneity was evaluated with the Cochran's Q test as well as the I 2 index and was defined as p < 0.10 or I 2 > 50% (Higgins et al., 2003).Publication bias was evaluated by Begg's funnel plots and Egger's test.
Forrest plots showed the pooled HR and HR>1 indicated worse survival outcome.All calculations were conducted by use of Review Manager Version 5.2 (The Cochrane Collaboration, Software Update, Oxford, UK) or Stata software (Stata Corporation, College Station, Texas).

Characteristics of eligible studies
A total of 166 articles were yielded after the literature search.After title reading, abstract reading and full-text reviewing, five studies were included (Figure 1).Eligible studies investigated colorectal cancer, gastric cancer, ovarian cancer, pancreatic cancer and endometrial cancer.The sample size ranged from 34 to 182 patients.The included studies were conducted between 2010 and 2013.The main features of these studies were listed in Table 1.

Serum miRNA 200c expression and survival
Two of the five included studies evaluated the association of serum miRNA 200c expression and OS (Valladares-Ayerbes et al., 2012;Toiyama et al., 2013).HR and its 95% CI were given in one study enrolling colorectal cancer patients (Toiyama et al., 2013).In the other study focused on gastric cancer patients, HR and its 95% CI were extracted from the given Kaplan-Meier survival curve (Tierney et al., 2007;Valladares-Ayerbes et al., 2012).Fixed-effects model was used and the pooled HR (high vs low) was 2.16 (95%CI: 1.32-3.52)(I 2 =0%, p=0.002) (Figure 2A, Table 2).The result showed an increased mortality in patients with high expression level of miRNA 200c in serum.

Publication bias
We used Begg's funnel plots and Egger's test to evaluate the publication bias of the included studies.As shown in figure 3, the funnel plots were almost symmetric.The p value of Egger's regression intercepts in the metaanalysis of tissue microRNA 200c was 0.731.Therefore, there was no significant publication bias exists in our meta-analysis.

Discussion
In miRNA 200 family, miRNA 200c was one of the most studied members and showed great potential in cancer therapy.Matjaz Rokavec reported that the supression of miRNA 200c by IL6 activated a feedforward inflammatory signaling circuit which drives transformation and tumorigenesis of normal cells and loss of IL6 and conservation of miRNA 200c impairs tumorigenesis significantly (Rokavec et al., 2012).
Together with the involvement of tumorigenesis, the corelation of miRNA 200c and EMT was discovered.It is well known that EMT of cancer cells diminishes the epithelial characteristics such as supression of E-cadherin, cytokeratin, occludin and desmoplakin.Meanwhile, EMT increases mesenchymal attributes like expression of vimentin, matrix metalloproteinase family members and E-cadherin transcriptional repressors ZEB1 and ZEB2 (Bonnomet et al., 2010).These characteristics endow cancer cells with enhanced ability of migration and invasion and trigger the cascade of metastasis.Moreover, the evidence of molecular links between EMT and stemness of cancer cells further suggested that interventions to EMT progress may provide new thrapeutic strategies for cancer (Scheel et al., 2012).
It was reported that miRNA 200c targets ZEB1 and ZEB2 (Hurteau et al., 2007;Cochrane et al., 2009).Reintroduction of miRNA 200c restored the expression of E-cadherin and significantly inhibited the migration and invasion in cancer cell line researches (Cochrane et al., 2009).In addition, Florian Kopp reported that miRNA 200c targeted TrkB and Bmi1 so that increasing the sensitivity of breast cancer cells to doxorubicin (Kopp et al., 2012).Another study demonstrated that RhoE was also a target of miRNA 200c and the sensitivity to DDP could be improved by miRNA 200c (Chang et al., 2013).Besides, it was documented that miR-200c targeted class III b-tubulin gene (TUBB3), increasing sensitivity to taxanes in vitro (Cittelly et al., 2012).Radiosensitivity of a non-small cell lung cancer cell line A549 was also increased by applying miRNA 200c (Shi et al., 2013).
These results indicated a protective role of miRNA 200c in solid tumors.Our meta-analysis showed the pooled HR (high vs low) for tissue miRNA 200c and OS is 0.62 [0.42-0.91](HR [95%CI]).This result is in accordance with the verified protective efficacy of miRNA 200c in cancer.
However, contradiction emerged when it came to miRNA 200c expression in serum.Our result of metaanalysis for serum miRNA 200c expression and OS manifested that high expression level in serum correlated with worse survival.The pooled HR and its 95%CI were 2. 16 [1.32-3.52].How to explain this inconformity?Firstly, because of the small sample size, this inconformity could be occasional.Further studies focusing on this topic are necessary to validate it.Even so, it is still valuable to discuss the potential reason of this inconformity.In one of our included studies, Yuji Toiyama studied both serum and tissue miRNA 200c expression (Toiyama et al., 2013).Their results demonstrated that only serum miRNA 200c expression possess the ability to predict OS (HR [95%CI]: 2. 43 [1.26-4.68]),but the result of miRNA 200c expression in tumor tissue was not significant (HR [95%CI]: 0.56 [0.28-1.10]).Yuji Toiyama also found that, compared with the primary tumor, the matched liver metstases had higher expression level of miRNA 200c.Hence, the origin of serum miRNA 200c was conjectured to be the metastatic site.On the other side, we put a hypothesis forward that miRNA 200c in serum may be a mirror of circulating tumor cells (CTCs).Chang and colleagues using model systems showed that around 1×10 6 tumor cells per gram tumor tissue could be introduced daily into the blood stream (Chang et al., 2000).Though 85% CTCs become broken and disappear within 5 minutes, they still could be vital contributors of serum miRNA 200c (Berezovskaya et al., 2005).In other words, higher level of miRNA 200c may represent more tumor cells entering the circulation.It was documented that CTCs predict worse survival in various cancers (Mocellin et al., 2006;Rahbari et al., 2010;Wang et al., 2011;Ma et al., 2012;Zhang et al., 2012).The result of our meta-analysis could be well explained if the origin of serum miRNA 200c is CTCs.Though these were speculative, we believed the correlation of serum miRNA 200c, metastasis and CTCs may be a key to figure out the inconformity.Besides, the possibility of the increased active secretion of miRNA 200c needs to be discussed.Kosaka and colleagues proved that the secretion of miRNAs was regulated by a ceramide-dependent pathway (Kosaka et al., 2010).The dysfunction of the regulation pathway may also lead to the higher expression level of serum miRNA 200c, while no related studies could be found to date.Eagerly, further studies are needed to prove the hypotheses and ravel out the puzzle.
Limitations of the present meta-analysis need to be discussed.Firstly, HRs and 95%CI of some included studies were extracted.Log (HR) and se (log (HR)) were then calculated by the software provided by Matthew Sydes and Jayne Tierney.Potential biases may relate to this process.Secondly, only five studies fit the included criteria and they investigated five different cancers.More studies in future could be added into our meta-analysis to validate the present results and subgroup analysis should be conducted according to the tumor type.At last, our meta-analysis only used published data.Updated individual patient data were not obtained.If those data were added to our analysis, the accuracy and determinacy could be better.
In conclusion, our meta-analysis suggest that both serum and tissue miRNA 200c expression could predict OS in solid tumors.Low expression of miRNA 200c in tumor tissue predict worse OS while low expression of miRNA 200c in serum are associated with better survival in solid tumors.To figure out the reason of the opposite results, more studies especially those focused on the origin of serum miRNA 200c are needed in future.

Flowchart of the Selection Figure 2. Forrest Plots of Estimated Hazard Ratios (HRs) for (A) Serum miRNA200c Expression and OS, and (B) Tumor Tissue miRNA200c Expression and OS DOI
:http://dx.doi.org/10.7314/APJCP.2014.15.8.3687How to Explain the Contradiction of microRNA 200c Expression and Survival in Solid Tumors?: a Meta-analysis