Next Article in Journal
Cancer Extracellular Matrix Proteins Regulate Tumour Immunity
Previous Article in Journal
Current Advances and Hurdles in Chimeric Antigen Receptor Technology
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Cancers
Volume 12
Issue 11
10.3390/cancers12113330
cancers-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

Prognostic Factors Involved in the Epithelial–Mesenchymal Transition Process in Colorectal Cancer Have a Preponderant Role in Oxidative Stress: A Systematic Review and Meta-Analysis

by
Eva Parisi
1,
Anabel Sorolla
2,3,
Robert Montal
1,4,
Rita González-Resina
1,
Anna Novell
1,
Antonieta Salud
1,4,† and
Maria Alba Sorolla
1,*,†
1
Research Group of Cancer Biomarkers, Biomedical Research Institute (IRBLleida), 25198 Lleida, Spain
2
Harry Perkins Institute of Medical Research, QEII Medical Centre, Nedlands, WA 6009, Australia
3
Centre for Medical Research, The University of Western Australia, Crawley, WA 6009, Australia
4
Department of Medical Oncology, Arnau de Vilanova University Hospital, 25198 Lleida, Spain
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this paper as last authors.
Cancers 2020, 12(11), 3330; https://doi.org/10.3390/cancers12113330
Submission received: 25 October 2020 / Revised: 6 November 2020 / Accepted: 9 November 2020 / Published: 11 November 2020
(This article belongs to the Section Systematic Review or Meta-Analysis in Cancer Research)
Browse Figures
Versions Notes

Abstract

:

Simple Summary

Metastasis is responsible for most of the deaths related to cancer patients. One of the hypotheses that explains the initiation of metastasis is a process called epithelial-to-mesenchymal transition (EMT), in which tumor cells change shape and acquire more aggressive properties that allows them to escape from the tumor and invade other organs. This also occurs in colorectal cancer (CRC), one of the most diagnosed types of cancer worldwide. During the past years, many scientists have discovered that certain molecules or biomarkers participating in this EMT process are able to predict the severity of the cancer and this is helping clinicians to manage treatments. Nevertheless, we think that all this information needs a detailed revision because a lot of biomarkers have been described but have not been analyzed whether they interact with each other in the same mechanism or not. Herein, we performed a bibliographic revision on this topic and identified a great number of biomarkers participating in oxidative stress, a cellular phenomenon that could have a role on the patient’s prognosis because its presence or absence on the patient’s tumor or blood had an influence on survival. Our findings suggest that oxidative stress deserves further study to understand metastasis better and to predict prognosis in a more efficient way.

Abstract

Epithelial-to-mesenchymal transition (EMT) is one of the most accepted mechanisms leading to metastasis, which is responsible for most of the cancer-related deaths. In order to identify EMT-related biomarkers able to predict clinical outcomes in colorectal cancer (CRC), a systematic review and meta-analysis of prognostic factors associated to overall survival (OS) and progression free survival (PFS) was conducted. The systematic literature search included studies from June 2014 to June 2019 available at PubMed and Scopus databases. Meta-analysis was performed for those markers appearing in minimum three works with a total number of 8656 participants. The rest were enlisted and subjected to functional enrichment. We identified nine clinical biomarkers and 73 EMT-related molecular biomarkers associated to OS and/or PFS in CRC. The significant enrichment of biomarkers found involved in cellular oxidoreductase activity suggests that ROS generation plays an active role in the EMT process. Clinical practice needs new biomarkers with a reliable prognostic value able to predict clinical outcomes in CRC. Our integrative work supports the role of oxidative stress in tumorigenesis and EMT progress highlighting the importance of deciphering this specific mechanism to get a better understanding of metastasis.

Graphical Abstract

1. Introduction

Colorectal cancer (CRC) is the second most commonly diagnosed cancer type worldwide and is the fourth leading cause of cancer deaths [1]. The prognosis of patients with CRC has greatly improved due to advances in early detection and treatment. However, 30% of patients who undergo curative resection die within a few years after surgery due to metastasis, mainly in the liver [2]. The metastatic process begins with the transition of tumor cells from epithelial characteristics to mesenchymal features (epithelial–mesenchymal transition or EMT). In particular, the EMT program induces disruption of cell adhesions, loss of apical-basal polarity, drastic remodeling of the cytoskeleton and acquisition of mesenchymal cells-related abilities such as the increase in the migratory capacity, leading to invasiveness. This is accompanied by a high resistance to apoptosis and an increase in the production of extracellular matrix components [3,4]. Biomarkers that allow clinicians to distinguish these tumors with high metastatic capacity would help for the therapeutic decision-making process.
The tumor–node–metastasis (TNM) pathological classification is the recommended prognostic tool for CRC [5]. Other factors with strong impact in prognosis are: poorly differentiated or high histological grade, vascular or lymphatic or perineural invasion, intestinal obstruction or perforation at diagnosis and elevated preoperative serum carcinoembryonic antigen (CEA) [6]. In fact, some of these factors serve as indicators of risk assessment in early CRC, which has been categorized into two differentiated groups: low-risk and high-risk groups. The low-risk group is composed by stage I and a fraction of stage II patients who generally undergo curative surgery without adjuvant treatment. The high-risk group encompasses a portion of stage II and the total of stage III patients, who are often treated with adjuvant chemotherapy after tumor resection. Nevertheless, there is an intense debate about the real benefit of using this categorization to guide chemotherapy administration, especially in high-risk stage II patients [7,8]. Moreover, these prognostic factors fail to precisely predict patient’s outcomes due to the wide range of OS rates observed across different stages [9]. Similarly, an individual risk of recurrence after treatment cannot be accurately predicted due to a high variability among individuals [10].
It is known that tumor genetic aberrations such as allelic loss of chromosome 18 [11], microsatellite instability-deficient mismatch repair (MSI/dMMR) [12] and mutations in KRAS [13] and in TP53 [14] have a detrimental effect on prognosis. Regarding EMT biomarkers, low expression of E-cadherin [15] and high expression of N-cadherin [16], Slug and Vimentin [17] have been linked to poorer prognosis in CRC. Multiple molecular signatures have proven to be useful in the stratification of patients according to risk of recurrence [18,19]. However, a lack of large-scale validation and low feasibility of its integration in the clinical practice are common concerns.
There is a high amount of information being generated about prognostic factors in CRC during the past years. However, they offer partial information, about just one or a few biomarkers, without integrating it with previous findings. Thus, the elaboration of an integrative scheme that could amalgamate published data with a quantified effect on overall survival (OS) and progression free survival (PFS) would allow scientists to see the whole scenario.
Herein, we present a comprehensive five-year retrospective systematic review and meta-analysis about prognostic biomarkers related to the EMT pathway. This study focuses on those processes such as oxidative stress signaling and non-coding biology that unravels novel promising prognostic biomarkers.

2. Results

2.1. Search Results

A total of 3357 unique indexed citations published between June 2014 and June 2019 were identified. Of these, 58 were selected for evaluation according to the defined eligibility criteria assessed in titles and abstracts. Finally, 39 studies with a total of 8656 participants were included in the meta-analysis and data was extracted after full text assessment. In total, 1219 studies were excluded because they were not about CRC, 824 presented exclusively preclinical data, 137 were related to other pathologies or processes, 117 were reviews, 68 lacked HR determination, 5 did not perform multivariate models and 6 had incomplete data as they were letters to the editor, communications in congresses, etc. The flowchart of the systematic review is presented in Figure 1.

2.2. Review of Eligible Studies

Among the 39 publications included, 3 were multicentric and the rest (36) were unicentric studies. All the studies had at least one discovery cohort that was an internal retrospective collection or a repository dataset (The Cancer Genome Atlas (TCGA): https://www.cancer.gov/about-nci/organization/ccg/research/structural-genomics/tcga; and Gene Expression Omnibus (GEO): https://www.ncbi.nlm.nih.gov/geo/). Six studies included validation cohorts, from either TCGA or internal collections. OS, PFS, cancer specific survival (CSS), recurrence-free survival (RFS), metastasis-free survival (MFS) and disease-free survival (DFS) were determined as predicted endpoints. From the total, 22 studies only assessed OS or CSS and two only determined PFS or RFS. Meanwhile, 15 works assessed both OS and PFS.
Pathological staging followed TNM (American Joint Committee on Cancer (AJCC) [20]/Union for International Cancer Control (UICC) [21]) or Dukes [22] classifications. TNM staging was the chosen system in 37 studies, whereas Dukes classification was used in two articles. Most of the works (27) included patients of stage I to IV disease while in five studies, patients cancers were categorized as I to III stages. One study recruited only stage II patients and six studies included metastatic patients exclusively (stage IV).
The cited prognostic biomarkers were determined by protein content: by immunohistochemistry (IHC), immunocytochemistry (ICC) or flow cytometry. Others were assessed by expression profiling such as real-time quantitative PCR (RT-qPCR), in situ hybridization (ISH) or a microarray. Finally, Sanger sequencing was used for polymorphisms identification. The type of sample analyzed was a primary tumor piece in 35 articles and circulating tumor cells (CTCs) in four works. All the information about included studies is specified in Table 1.

2.3. Meta-Analysis of the EMT-Related Prognostic Biomarkers in CRC

A meta-analysis was performed for each of the potential EMT-related prognostic biomarkers associated with OS, PFS or both in CRC patients, found in any of the 39 included studies, and found in at least three independent articles. Overall, our findings showed that nine clinical biomarkers were detected to be relevant for prognosis: Pathological tumor extension (pT), pathological node involvement (pN), pathological metastasis (pM), pathological staging (pStage), lymphovascular invasion (LVI), histological grade (HG), tumor size, tumor location and chemotherapy administration. In addition, we identified four molecular prognostic biomarkers that were CEA, CA19.9, Ki67 and E-cadherin. The results from the meta-analysis are summarized in Table 2.

2.3.1. Clinical Prognostic Biomarkers

There were a major number of studies predicting OS than PFS, which was the reason why we found more clinical biomarkers associated to the OS than to PFS. Pathological factors such as pT, pN, pM, pStage, LVI and HG were identified as strongly significantly affecting both OS and PFS (Figure 2A,B and Figure 3A,B), while chemotherapy was only found to be significantly associated to OS. Tumor size and tumor location were not significantly associated to any survival endpoint (Figure 2C). The test of heterogeneity confirmed homogeneity in pStage and HG in predicting both OS (Figure 2A) and PFS (Figure 3A,B), while meta-analysis of pT and LVI were homogenous only in predicting PFS (Figure 3A,B). In these cases, the symmetric distribution of funnel plot shapes demonstrated a low risk of publication bias. In contrast, variables such as pN, pM, tumor location, tumor size and chemotherapy showed heterogeneity in their prediction, suggesting a high risk of publication bias (Figure 2A–C and Figure 3A). Fixed or random effect models were applied accordingly.

2.3.2. Molecular Prognostic Biomarkers

CEA was the unique molecular biomarker identified as an independent prognostic factor of both OS and PFS that was statistically significant (Figure 2D and Figure 3B), while E-cadherin was only significantly associated to OS. CA19.9 and Ki67 were not associated to survival (Figure 2D). In contrast to what happened with clinical biomarkers, meta-analysis results from molecular biomarkers were homogeneous for almost all variables cases except for CA19.9 levels in predicting OS (Figure 2D). Therefore, the most widely used type of effect model was the fixed-effect with results showing a low risk of publication bias as evidenced by the symmetry of their funnel plots.

2.4. EMT-Related Molecular Biomarkers in CRC

In total, we found 73 EMT-related molecular biomarkers associated with prognosis in CRC patients, as represented in Figure 4.
Almost all biomarkers found were proteins. Remarkably, we also found the presence of several non-coding RNAs including two microRNA (miRNA): miR-490-3p and miR-139-5p and four long-non-coding RNA (lncRNA): HOTAIR, GAPLINC, NNT-AS1 and PANDAR (Figure 4).
Of them all, four (CEA, CA19.9, ki67 and E-cadherin) were mentioned in a previous section and analyzed through meta-analysis, and the remaining 69 were only mentioned in one or two independent works. In addition, 24 were independent prognostic factors for OS and PFS, while 48 and 1 were exclusively associated to OS or PFS, respectively (Figure 4).
Regarding the biomarker determination, 69 of them were determined through mRNA or protein levels, while five consisted of specific gene polymorphisms. In order to simplify all the information, Figure 4 shows two bar diagrams containing the 69 biomarkers represented according to its LN (HR), ranging from 5.49 to −2.41 for OS, and from 1.48 and −0.61 for PFS, for the specific comparison of elevated/high levels versus normal/low levels.

2.5. Functional Enrichment Analysis

Functional enrichment was applied to our biomarker list in order to map genes into known functional information and detect enriched terms. Overall results showed a significant enrichment in several terms regarding GO and MIRNA databases, as seen in the Manhattan plot (Figure 5A). A total of 100, 19 and 14 GO terms were enriched for biological process (BP), molecular function (MF) and cellular compartment (CC), respectively. Additionally, six terms were found in MIRNA databases, containing both miRNA and lncRNA. Detailed classification revealed that oxidoreductase activity, superoxide-generating NADPH oxidase activity and coenzyme binding were the main enriched MF processes (Figure 5B) with a combined percentage of 63%. BP was mainly concentrated in superoxide anion generation, multicellular organismal process and cell population proliferation, with a percentage of 34% (Figure 5C). Additionally, extracellular components, NADPH oxidase and oxidoreductase complexes were the main CC, and their percentages were as high as 64% (Figure 5D). Detailed information about each biomarker categorization is illustrated in Figure S1.

3. Discussion

Seven clinical biomarkers, not strictly related to the EMT pathway, were found to be significantly associated to survival: pT, pN, pM, pTNM, LVI, HG and chemotherapy. The local extent of the tumor and the presence of involved nodes are known factors influencing survival. Population-based data of 109,953 N0 patients confirmed that patients with T1–2 cancers had an increased 5-year survival compared to T3, and T3 higher than T4 [61]. This tendency was also maintained in N2. Another study with data from 50,042 patients is also aligned with our findings, showing better survival on those patients with lesions categorized as T1–2 compared to T3–T4, both in N1 and N2 [62]. Additionally, it seems that T stage has a preponderant role on survival where lesions growing into the peritoneum (T4a) have a better prognosis than those invading adjacent tissues (T4b), regardless of the N category [61]. Another work supporting this idea suggests that TNM should be reconsidered by T stage weighting as this affects CRC survival more significantly than the N stage [63]. The presence of metastasis at diagnosis is observed in around 25% of newly diagnosed CRC cases, and 20% of CRC patients will develop distant metastasis during disease course [64]. Distant metastasis is strongly associated to poor outcome in CRC, being the primary cause of treatment failure and consequent death [65]. HG reflects the morphology and proliferative capacity of the primary tumor and has been repeatedly described as an independent prognostic factor in CRC though multivariate model analysis [66]. HG has classically been divided in three distinct categories: well (G1), moderate (G3) and poor (G3) differentiated tumors. G3 tumors are more aggressive and more prone to acquire EMT-like features. Some authors highlight the assessment of the dedifferentiation phenotype and desmoplastic environment, which provides a more individualized outcome prediction than conventional grading and staging systems [67].
The administration of chemotherapy is significantly associated with higher survival. In particular, adjuvant chemotherapy provides significantly higher disease free survival benefit by reducing the rate of recurrence rate and by translating into long-term OS in resected II and III-staged patients [68,69]. Additionally, neoadjuvant chemotherapy provides a benefit in those patients with initially unresectable liver metastasis, impacting both OS and PFS [70].
Regarding LVI, we found that this factor was a significant predictor of both OS and PFS. LVI is considered to be an strong stage-independent prognostic factor and influences decisions regarding the administration of adjuvant chemotherapy in CRC patients with stage II tumors [71]. The presence of tumor cells within the endothelium-lined lymphatic of vascular channels is a very common feature in CRC, accounting for 10-89.5% of the cases [72], which precedes lymph node metastasis and systemic dissemination of cancer cells [73]. In contrast, other authors suggest that vascular invasion rather than lymphatic invasion is responsible for distant recurrences [74].
We did not find significant associations between tumor size and tumor location with survival. There is a great controversy in the literature on these variables. Tumor size, defined as the widest horizontal diameter of tumors, is not involved in the AJCC TNM staging system for CRC as it is in other cancers, and some studies reported no effects on survival [6,75]. Conversely, a more recent study including data from 300,386 patients, concludes that tumor size predicts long-term survival in colon cancer patients subjected to colectomy [76]. Similarly, a study deploying multivariate analysis on 3971 stage I–III CRC patients with curative resection has identified that a tumor size greater than 4 cm is an independent risk factor for CSS [77]. This same study indicates that the tumor side location has a differential impact on OS. Small tumors (≤4 cm) placed on the right side of the colon presented worse prognosis than the ones placed at other locations. There is, in fact, a general agreement in considering that right-sided tumors have worse prognosis than the left-sided counterparts, irrespectively of tumor staging [78,79]. In contrast, other studies suggest that the prognostic value of tumor location depends on other confounding factors such as an elevated systemic inflammatory response and high CD3+ immune infiltrate at the tumor margin and within cancer cell nests, especially in resected I–III CRC patients [80].
Regarding molecular prognostic biomarkers, we found that elevated levels or expression of CEA, CA19.9, Ki67 and loss of E-cadherin expression were associated to poor OS, and CEA levels inversely correlate with PFS. CEA is the most used tumor marker in CRC and is involved in cell adhesion and cancer progression shown to target adherens junctions in CRC cell lines [81]. Interestingly, preoperative serum CEA levels correlates with the CEA-cell associate molecule 1 expression in tumors, which induces EMT and tumor angiogenesis in hepatocellular carcinoma [82]. At the clinical level, preoperative serum CEA levels over 5 ng/mL are significantly associated to decreased disease free survival in CRC patients [83] and postoperative levels positively predict recurrence and survival [84]. The tumor antigen CA19.9 is a tetrasaccharide carbohydrate synthetized by the gastrointestinal epithelium and is considered an established serum biomarker for monitoring treatment efficacy in pancreatic cancer patients [85]. Although, the determination of CA19.9 together with CEA adds value to the prognosis, it is still insufficient to manage CRC patients [86]. Preclinical data from CRC cell lines shows a relationship between CA19.9 antigen exposure and metastatic potential through an EMT-related process. In particular, the responsible enzymes for CA19.9 synthesis, fucosyl-transferases, enhance TGF-β signaling resulting in CRC cell migration and invasion, potentiating cancer cell adhesion to endothelial cells by upregulation of Sialyl Lewis antigens [87]. In many cancers such as breast and gastric cancer, nuclear Ki67 positively correlates with tumor grading and is a reliable indicator of tumor recurrence risk [88]. In CRC, however, there is a discrepancy in considering Ki67 as a prognostic marker. While some authors claim that Ki67 labeling index is an independent prognostic factor indicating poor prognosis [89,90], others defend that Ki67 expression is associated with a favorable outcome [91]. Ki67 expression often positively correlates with EMT-related factors such as survivin, vimentin and N-cadherin, thus promoting tumor aggressiveness [92,93]. Finally, E-cadherin is a calcium-dependent glycoprotein localized in the adherens junctions, playing a role in cell adhesion and in maintaining the epithelial morphology [15]. In cancer, loss of E-cadherin caused by mutations, proteolytic cleavage or gene promoter silencing is responsible for invasiveness, anoikis resistance and metastasis dissemination through the EMT program [94]. In CRC, low expression of E-cadherin [16] and the existence of some specific polymorphisms in E-cadherin gene [38] are considered as independent prognostic factors of increased survival.
The impact of non-coding RNAs in CRC prognosis is being deeply explored. Remarkably, the HR values of miRNA and lncRNA were the highest among our entire biomarker list. miRNA are involved in almost all aspects of cancer biology and are considered tumor suppressors or oncogenes depending on the cellular context in which they are expressed [95]. They are former members of the regulatory networks of EMT program like TGF-β/ZEB axis and involved in Notch, Wnt and p53 signaling pathways [96]. They are generally found overexpressed in CRC tissue [97]. Although miR-139-5p has been found upregulated in blood and tissue of recurrent CRC patients [49], another work suggests an inhibitory role of this miRNA on EMT in CRC cell lines [98]. We found that downregulation of miR 490-3p in CRC tissue correlated with poor prognosis, which is supported by the tumor suppression role of miR 490-3p in repressing migration and invasion through partial TGF-β signaling described by others [99].
Regarding the lncRNAs described in our study, they strongly predicted OS when upregulated. Several studies support the role of lncRNAs in the regulation of tumor progression and metastasis through the regulation of the EMT, acting as EMT promoters or EMT suppressors [100]. Specifically, the four lncRNAs identified in our systematic review, PANDAR, HOTAIR, NNT-AS1 and GAPLINC, are independent factors of poor OS when elevated, and HOTAIR also predicts poor PFS. Additionally, HOTAIR has been shown to promote EMT through the activation of Wnt/Notch signaling and its upregulation constitutes a prognostic factor in esophageal squamous cell carcinoma [101].
It is worth mentioning that the EMT pathway is represented in The Consensus Molecular Subtypes of Colorectal Cancer described in 2015 by Guinney et al. [102]. In his work, authors demonstrated the presence of four different molecular subtypes (CMS1–4) in CRC. In particular, CMS4 tumors had upregulation of genes implicated in EMT and signatures associated with the activation of TGF-β signaling, angiogenesis, matrix remodeling pathways and complement inflammatory system. This subtype is also enriched for downregulated miRNAs (e.g., hsa-mir-148a, the miR-192 and miR-200 families), and such downregulation is associated with suppression of EMT, matrix remodeling and TGF-β-associated signatures. This could explain the higher aggressivity of CMS4 tumors compared to the other CMS subtypes. At the clinical level, the CMS classification possesses a significant prognostic value in metastatic CRC and according to Mooi et al [103] this seems to be independent from the first-line treatment. In contrast, the FIRE-3 trial demonstrated that CMS classification is predictive for outcomes in CMS4, favoring FOLFIRI plus cetuximab-treated tumors compared to FOLFIRI plus bevacizumab-treated cancers when they are RAS wild type [104].
Our work encourages the analysis of epithelial-like markers such as E-cadherin, which loss of expression has an impact on both OS and PFS in patients treated with oxaliplatin or irinotecan-based chemotherapy, as described in [42]. In addition, the evaluation of mesenchymal-like markers is also worth of attention. In this regard, we suggest the determination of vimentin and N-cadherin expression as they impact on OS, as previously reported [16,31]. However, not adjuvant treatment was specified in these works. Furthermore, the evaluation of EMT promoters such as TWIST, ZEB and SNAIL family members sounds sensible. Related to these markers, it seems that the detection of specific polymorphisms rather than their expression predicts survival. In particular, TWIST1 polymorphisms predict survival in patients with metastatic CRC receiving first-line bevacizumab plus oxaliplatin-based chemotherapy [38]. To summarize, we thought that the identification of a multiple marker EMT signature based on IHC, qRT-PCR or SNP sequencing methods as routine testing in the anatomy pathology laboratory is highly valuable to predict OS and PFS in CRC patients.
Results from functional enrichment of the molecular biomarker list revealed consistent roles around oxidative stress. Interestingly, oxidoreductase activity accounted for the major GO molecular function annotated and superoxide anion generation for the most demanded BP. Moreover, we would like to emphasize a significant presence of biomarkers in the extracellular compartment or being part of the NADPH oxidase complex. Aerobic respiration generates reactive oxygen species (ROS), which at normally low concentrations are necessary for several cellular processes such as signal transduction, enzyme activation, gene expression, disulfide bond formation and caspase activity control [105]. However, when the antioxidant defense of the cell is overwhelmed and oxidative exacerbates, cell damage takes place, a fact that is considered to be a central event in the physiopathology of several disorders, including cardiovascular and neurodegenerative diseases [106]. In cancer, ROS promotes cellular proliferation, evasion of apoptosis and anoikis, tissue invasion, metastasis and angiogenesis [105]. The NAPDH oxidase complex, where the catalytic component isoforms are Nox1–Nox5 and Duox1–2, is one of the main sources of intracellular ROS when activated. NOX family members regulate redox signaling that ultimately leads to angiogenesis, as reviewed in [107]. In particular, Nox1 upregulates VEGF expression and thus activates VEGF receptors (VEGFR1 and VEGFR2), and hydrogen peroxide production enhances matrix metalloproteinase activity [108]. In addition, it has been shown that Nox4 is a critical regulator of the ROS-mediated DNA damage response induced by oncogenic H-RasVal12, one of the most frequent mutated oncogenes in CRC [109]. ROS is also directly linked to the EMT process through the activation of NF-κB, HIF-1α, TGF-β and extracellular matrix remodeling proteins such as integrins and MMPs, all highly sensitive to the redox status [110]. Overall, a detailed knowledge of redox factors on tumors will add value in the prediction of prognosis with more accuracy.
The main limitation of this meta-analysis is the high diversity of patient’s clinical status across all included studies, which ranges from I to IV stages, and the differences in the treatment management of patients. Although this fact adds variability in our meta-analysis, the extraction of data only from those articles containing HR values adjusted for the individual characteristics of the corresponding study ensures consistency.

4. Materials and Methods

4.1. Eligibility Criteria for Study Inclusion

We included all the observational studies published from June 2014 to June 2019 involving CRC patients. In order to be included, studies had to evaluate prognostic biomarkers related to the EMT process and to predict patient’s OS and/or PFS. We included biomarker information extracted by any technique (protein, gene expression and sequencing) and from any matrix: Frozen tissue, paraffin tissue and blood. All works contained the adjusted hazard ratio (HR) with its 95% confidence interval (CI) for each biomarker in association with either OS or PFS or both or equivalent, evaluated through multivariate Cox logistic regression. We only included articles published in English. We excluded reviews and original articles containing only preclinical data and studies researching about biomarkers associated to treatment, diagnosis or toxicity. This systematic review had no previously registered protocol.

4.2. Literature Search and Systematic Review Procedure

The search strategy was performed in PUBMED and SCOPUS. The search included the following terms:
1. “Colonic Neoplasms” [Mesh] OR “Sigmoid Neoplasms” [Mesh] AND “Biomarkers” [Mesh] OR “Biomarkers, Tumor” [Mesh] AND “Epithelial-Mesenchymal Transition” [Mesh] OR “Snail Family Transcription Factors” [Mesh];
2. “Colonic Neoplasms” OR “Sigmoid Neoplasms” AND “Biomarkers” OR “Biomarkers, Tumor” AND “EMT” OR “EMT pathway”. Duplicates were removed.
The systematic search of literature was assessed by three investigators (RG, AN and MAS). Disagreements were solved by consensus. Finally, data extraction and synthesis were performed by three investigators (EP, RG and MAS) and included: article identifier, author, year, work design, study cohorts, number of patients in each cohort grouped by the clinical stage, included variables in the multivariate logistic regression model with its adjusted HR with CI for each one of the prognostic biomarkers cited by each study, technique of biomarker determination, predicted variable (OS and/or PFS and/or equivalent) and treatment (neo-adjuvant and/or adjuvant treatment).
This systematic review and meta-analysis followed the published Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [111].

4.3. Statistical Analysis of the Meta-Analysis

Hazard ratios (HRs) with the corresponding 95% CI obtained through multivariate logistic regression models were extracted from publications. If needed, the reciprocal HR and CI values were calculated in order to be annealed with the magnitude comparison for each variable. Calculations of Log (HR) and standard error (SE) were performed. The meta-analysis was based on the inverse variance method between the results of at least three independent studies for each variable.
Heterogeneity of included studies was calculated using Higgin’s I2 index [112] and chi2 tests. Heterogeneous studies were considered when I2 was more than 50% and chi2 p-value < 0.05. In case of discrepancy between both tests, chi2 p-value was prioritized. The random effect model was applied on heterogeneous studies. Otherwise, the fixed effect model was used when the studies were homogeneous.
Meta-analysis results were illustrated in forest plots, reporting both weighted and pooled effect from individual studies with their corresponding global HR, 95% CI, and p-value.
Possible publication bias was evaluated through visual inspection of funnel plots (i.e., an asymmetrical distribution). The statistical analysis of meta-analysis was performed using Review Manager Software (RevMan-v5.3; Cochrane, Oxford, UK).

4.4. GO Profiler Analysis

In those biomarkers in which meta-analysis could not be performed due to a lack of HR estimations available, functional enrichment was executed in order to interpret the resulting biomarker list. The statistical enrichment analysis was carried out by g:Profiler (version e99_eg46_p14_f929183; link: https://biit.cs.ut.ee/gprofiler/gost) and the analysis parameters were as follows: a specific organism was chosen (Homo sapiens (human)) and MIRNA and GO analyses (GO molecular function (GO:MF), GO cellular component (GO:CC) and GO biological process (GO:BP)) were carried out sequentially. The statistical domain scope was used only for annotated genes. The significance threshold in the g:Profiler analysis was the g:SCS multiple testing correction method applying significance threshold of 0.05 [113].

5. Conclusions

This work identified 9 clinical and 73 EMT-related molecular biomarkers associated to CRC prognosis described in the last 5 years. Apart from the classical ones, novel molecular markers implicated in the EMT process were being considered factors with promising prognostic value. Emerging biomarkers involved in oxidoreductase activity suggest a critical role of ROS in tumorigenesis, particularly in angiogenesis, which is one of most targeted processes in CRC. Although antiangiogenics have become indispensable for the treatment of CRC, more research is needed to identify and validate predictive biomarkers of efficacy. Unraveling ROS mechanisms could provide this valuable information. Furthermore, epigenetic regulation through non-coding RNA in EMT represents a complex framework of interactions that warrant further exploration to understand the process as a whole.

Supplementary Materials

The following are available online at https://www.mdpi.com/2072-6694/12/11/3330/s1, Figure S1: Detailed biomarker categorization.

Author Contributions

A.N. and M.A.S. were responsible of the conceptualization; E.P., R.G.-R., A.N. and M.A.S. performed the systematic review and meta-analysis. E.P. performed the GO profiling analysis; and A.S. (Anabel Sorolla) and R.M. made a critical review and edited the manuscript. Finally, M.A.S. and A.S. (Antonieta Salud) wrote the manuscript and were responsible of project administration and funding. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Merck Salud Foundation, 2018 project award, given to A.S. (Antonieta Salud) and the University and Research Grants Management Agency (AGAUR) of Catalonia Government emergent research group recognition award (2017SRG1620) given to M.A.S.; and by the Raine Medical Research Foundation (grant number: RPG-004-19) and by Cancer Council Western Australia Collaborative Cancer Funding Scheme (grant ID: 814) given to A.S. (Anabel Sorolla).

Acknowledgments

The authors thank Montse Martínez, statistical expert, for her critical review of the manuscript and valuable comments and suggestions.

Conflicts of Interest

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

References

  1. Brenner, H.; Kloor, M.; Pox, C.P. Colorectal cancer. Lancet 2014, 383, 1490–1502. [Google Scholar] [CrossRef]
  2. Cunningham, D.; Atkin, W.; Lenz, H.J.; Lynch, H.T.; Minsky, B.; Nordlinger, B.; Starling, N. Colorectal cancer. Lancet 2010, 64, 104–117. [Google Scholar] [CrossRef]
  3. Cao, H.; Xu, E.; Liu, H.; Wan, L.; Lai, M. Epithelial-mesenchymal transition in colorectal cancer metastasis: A system review. Pathol. Res. Pract. 2015, 211, 557–569. [Google Scholar] [CrossRef] [PubMed]
  4. Tsai, J.H.; Yang, J. Epithelial-mesenchymal plasticity in carcinoma metastasis. Genes Dev. 2013, 27, 2192–2206. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  5. Edge, S.B.; Compton, C.C. The american joint committee on cancer: The 7th edition of the AJCC cancer staging manual and the future of TNM. Ann. Surg. Oncol. 2010, 17, 1471–1474. [Google Scholar] [CrossRef] [PubMed]
  6. Compton, C.C.; Fielding, L.P.; Burgart, L.J.; Conley, B.; Cooper, H.S.; Hamilton, S.R.; Hammond, M.E.H.; Henson, D.E.; Hutter, R.V.P.; Nagle, R.B.; et al. Prognostic factors in colorectal cancer: College of American Pathologists consensus statement. Archiv. Pathol. Lab. Med. 2000, 124, 979–994. [Google Scholar]
  7. Bender, U.; Rho, Y.S.; Barrera, I.; Aghajanyan, S.; Acoba, J.; Kavan, P. Adjuvant therapy for stages ii and iii colon cancer: Risk stratification, treatment duration, and future directions. Curr. Oncol. 2019, 26, S43–S52. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  8. Fotheringham, S.; Mozolowski, G.A.; Murray, E.M.A.; Kerr, D.J. Challenges and solutions in patient treatment strategies for stage II colon cancer. Gastroenterol. Rep. 2019, 7, 151–161. [Google Scholar] [CrossRef] [Green Version]
  9. Labianca, R.; Nordlinger, B.; Beretta, G.D.; Mosconi, S.; Mandalà, M.; Cervantes, A.; Arnold, D. Early colon cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann. Oncol. 2013, 24 (Suppl. 6), vi64–vi72. [Google Scholar] [CrossRef]
  10. Zafar, S.N.; Hu, C.Y.; Snyder, R.A.; Cuddy, A.; You, Y.N.; Lowenstein, L.M.; Volk, R.J.; Chang, G.J. Predicting Risk of Recurrence After Colorectal Cancer Surgery in the United States: An Analysis of a Special Commission on Cancer National Study. Ann. Surg. Oncol. 2020, 27, 2740–2749. [Google Scholar] [CrossRef]
  11. Jen, J.; Kim, H.; Piantadosi, S.; Liu, Z.-F.; Levitt, R.C.; Sistonen, P.; Kinzler, K.W.; Vogelstein, B.; Hamilton, S.R. Allelic Loss of Chromosome 18q and Prognosis in Colorectal Cancer. N. Engl. J. Med. 1994, 331, 213–221. [Google Scholar] [CrossRef] [PubMed]
  12. Kawakami, H.; Zaanan, A.; Sinicrope, F.A. Microsatellite Instability Testing and Its Role in the Management of Colorectal Cancer. Curr. Treat. Options Oncol. 2015, 16, 30. [Google Scholar] [CrossRef] [PubMed]
  13. Deng, Y.; Wang, L.; Tan, S.; Kim, G.P.; Dou, R.; Chen, D.; Cai, Y.; Fu, X.; Wang, L.; Zhu, J.; et al. KRAS as a predictor of poor prognosis and benefit from postoperative FOLFOX chemotherapy in patients with stage II and III colorectal cancer. Mol. Oncol. 2015, 9, 1341–1347. [Google Scholar] [CrossRef] [PubMed]
  14. Russo, A.; Bazan, V.; Iacopetta, B.; Kerr, D.; Soussi, T.; Gebbia, N. TP53-CRC Collaborative Study Group. The TP53 colorectal cancer international collaborative study on the prognostic and predictive significance of p53 mutation: Influence of tumor site, type of mutation, and adjuvant treatment. J. Clin. Oncol. 2005, 23, 7518–7528. [Google Scholar] [CrossRef] [PubMed]
  15. Christou, N.; Perraud, A.; Blondy, S.; Jauberteau, M.O.; Battu, S.; Mathonnet, M. E-cadherin: A potential biomarker of colorectal cancer prognosis. Oncol. Lett. 2017, 13, 4571–4576. [Google Scholar] [CrossRef] [Green Version]
  16. Yan, X.; Yan, L.; Liu, S.; Shan, Z.; Tian, Y.; Jin, Z. N-cadherin, a novel prognostic biomarker, drives malignant progression of colorectal cancer. Mol. Med. Rep. 2015, 12, 2999–3006. [Google Scholar] [CrossRef] [Green Version]
  17. Toiyama, Y.; Yasuda, H.; Saigusa, S.; Tanaka, K.; Inoue, Y.; Goel, A.; Kusunoki, M. Increased expression of slug and vimentin as novel predictive biomarkers for lymph node metastasis and poor prognosis in colorectal cancer. Carcinogenesis 2013, 34, 2548–2557. [Google Scholar] [CrossRef] [Green Version]
  18. Salazar, R.; Roepman, P.; Capella, G.; Moreno, V.; Simon, I.; Dreezen, C.; Lopez-Doriga, A.; Santos, C.; Marijnen, C.; Westerga, J.; et al. Gene expression signature to improve prognosis prediction of stage II and III colorectal cancer. J. Clin. Oncol. 2011, 29, 17–24. [Google Scholar] [CrossRef] [Green Version]
  19. Wang, L.; Shen, X.; Wang, Z.; Xiao, X.; Wei, P.; Wang, Q.; Ren, F.; Wang, Y.; Liu, Z.; Sheng, W.; et al. A molecular signature for the prediction of recurrence in colorectal cancer. Mol. Cancer 2015, 14, 1–10. [Google Scholar] [CrossRef] [Green Version]
  20. Weiser, M.R. AJCC 8th Edition: Colorectal Cancer. Ann. Surg. Oncol. 2018, 25, 1454–1455. [Google Scholar] [CrossRef] [Green Version]
  21. Brierley, J.D.; Gospodarowicz, M.K.; UICC; Brierley, J.; Gospodarowicz, M.K. TNM Classification of Malignant Tumours, 8th ed.; John Wiley & Sons: Oxford, UK; Hoboken, NJ, USA, 2016; 253p. [Google Scholar]
  22. Dukes, C. The classification of cancer of the rectum. J. Pathol. Bacteriol. 1932, 35, 323–332. [Google Scholar] [CrossRef]
  23. Barbazan, J.; Muinelo-Romay, L.; Vieito, M.; Candamio, S.; Diaz-Lopez, A.; Cano, A.; Gomez-Tato, A.; De Los Angeles Casares De Cal, M.; Abal, M.; Lopez-Lopez, R. A multimarker panel for circulating tumor cells detection predicts patient outcome and therapy response in metastatic colorectal cancer. Int. J. Cancer 2014, 135, 2633–2643. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  24. Gao, Z.H.; Lu, C.; Wang, Z.N.; Song, Y.X.; Zhu, J.L.; Gao, P.; Sun, J.X.; Chen, X.W.; Wang, M.X.; Dong, Y.L.; et al. ILEI: A novel marker for epithelial-mesenchymal transition and poor prognosis in colorectal cancer. Histopathology 2014, 65, 527–538. [Google Scholar] [CrossRef]
  25. Sugimachi, K.; Yokobori, T.; Iinuma, H.; Ueda, M.; Ueo, H.; Shinden, Y.; Eguchi, H.; Sudo, T.; Suzuki, A.; Maehara, Y.; et al. Aberrant Expression of Plastin-3 Via Copy Number Gain Induces the Epithelial–Mesenchymal Transition in Circulating Colorectal Cancer Cells. Ann. Surg. Oncol. 2014, 21, 3680–3690. [Google Scholar] [CrossRef] [PubMed]
  26. Wu, Z.E.H.; Wang, X.L.; Tang, H.M.; Jiang, T.; Chen, J.; Lu, S.U.; Qiu, G.Q.; Peng, Z.H.; Yan, D.W. Long non-coding RNA HOTAIR is a powerful predictor of metastasis and poor prognosis and is associated with epithelial-mesenchymal transition in colon cancer. Oncol. Rep. 2014, 32, 395–402. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  27. Zhou, Y.; Wu, J.; Fu, X.; Du, W.; Zhou, L.; Meng, X.; Yu, H.; Lin, J.; Ye, W.; Liu, J.; et al. OTUB1 promotes metastasis and serves as a marker of poor prognosis in colorectal cancer. Mol. Cancer 2014, 13, 258. [Google Scholar] [CrossRef] [Green Version]
  28. Dawson, H.; Grundmann, S.; Koelzer, V.H.; Galván, J.A.; Kirsch, R.; Karamitopoulou, E.; Lugli, A.; Inderbitzin, D.; Zlobec, I. Tyrosine kinase receptor B (TrkB) expression in colorectal cancers highlights anoikis resistance as a survival mechanism of tumour budding cells. Histopathology 2015, 66, 715–725. [Google Scholar] [CrossRef]
  29. Kahlert, C.; Lerbs, T.; Pecqueux, M.; Herpel, E.; Hoffmeister, M.; Jansen, L.; Brenner, H.; Chang-Claude, J.; Bläker, H.; Kloor, M.; et al. Overexpression of SIX1 is an independent prognostic marker in stage I-III colorectal cancer. Int. J. Cancer 2015, 137, 2104–2113. [Google Scholar] [CrossRef]
  30. Liu, G.; Zhu, J.; Yu, M.; Cai, C.; Zhou, Y.; Yu, M.; Fu, Z.; Gong, Y.; Yang, B.; Li, Y.; et al. Glutamate dehydrogenase is a novel prognostic marker and predicts metastases in colorectal cancer patients. J. Transl. Med. 2015, 13, 144. [Google Scholar] [CrossRef] [Green Version]
  31. Xiao, S.; Liu, L.; Lu, X.; Long, J.; Zhou, X.; Fang, M. The prognostic significance of bromodomain PHD-finger transcription factor in colorectal carcinoma and association with vimentin and E-cadherin. J. Cancer Res. Clin. Oncol. 2015, 141, 1465–1474. [Google Scholar] [CrossRef]
  32. Busch, E.L.; Keku, T.O.; Richardson, D.B.; Cohen, S.M.; Eberhard, D.A.; Avery, C.L.; Sandler, R.S. Evaluating markers of epithelial-mesenchymal transition to identify cancer patients at risk for metastatic disease. Clin. Exp. Metastasis 2016, 33, 53–62. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  33. Gao, T.; Wang, M.; Xu, L.; Wen, T.; Liu, J.; An, G. DCLK1 is up-regulated and associated with metastasis and prognosis in colorectal cancer. J. Cancer Res. Clin. Oncol. 2016, 142, 2131–2140. [Google Scholar] [CrossRef] [PubMed]
  34. Hashimoto, M.; Kobayashi, T.; Tashiro, H.; Arihiro, K.; Kikuchi, A.; Ohdan, H. h-Prune is associated with poor prognosis and epithelial–mesenchymal transition in patients with colorectal liver metastases. Int. J. Cancer 2016, 139, 812–823. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  35. Li, Q.; Wei, P.; Huang, B.; Xu, Y.; Li, X.; Li, Y.; Cai, S.; Li, D. MAEL expression links epithelial-mesenchymal transition and stem cell properties in colorectal cancer. Int. J. Cancer 2016, 139, 2502–2511. [Google Scholar] [CrossRef]
  36. Li, C.; Chen, J.; Wang, Y.; Song, G.; Xiao, C.; Yan, D.; Zhong, L.; Sun, X.; Wang, X.; Yu, F.; et al. High-level expression of P21-Cdc/Rac-actived kinase 7 is closely related to metastatic potential and poor prognosis of colon carcinoma. Oncotarget 2016, 7, 46042–46055. [Google Scholar] [CrossRef] [Green Version]
  37. Luo, Y.; Ye, G.Y.; Qin, S.L.; Mu, Y.F.; Zhang, L.; Qi, Y.; Qiu, Y.E.; Yu, M.H.; Zhong, M. High expression of Rab3D predicts poor prognosis and associates with tumor progression in colorectal cancer. Int. J. Biochem. Cell Biol. 2016, 75, 53–62. [Google Scholar] [CrossRef]
  38. Matsusaka, S.; Zhang, W.; Cao, S.; Hanna, D.L.; Sunakawa, Y.; Sebio, A.; Ueno, M.; Yang, D.; Ning, Y.; Parekh, A.; et al. TWIST1 Polymorphisms Predict Survival in Patients with Metastatic Colorectal Cancer Receiving First-Line Bevacizumab plus Oxaliplatin-Based Chemotherapy. Mol. Cancer Ther. 2016, 15, 1405–1411. [Google Scholar] [CrossRef] [Green Version]
  39. Ning, Y.; Zhang, W.; Hanna, D.L.; Yang, D.; Okazaki, S.; Berger, M.D.; Miyamoto, Y.; Suenaga, M.; Schirripa, M.; El-Khoueiry, A.; et al. Clinical relevance of EMT and stem-like gene expression in circulating tumor cells of metastatic colorectal cancer patients. Pharmacogenomics J. 2018, 18, 29–34. [Google Scholar] [CrossRef]
  40. Satelli, A.; Batth, I.S.; Brownlee, Z.; Rojas, C.; Meng, Q.H.; Kopetz, S.; Li, S. Potential role of nuclear PD-L1 expression in cell-surface vimentin positive circulating tumor cells as a prognostic marker in cancer patients. Sci. Rep. 2016, 6, 28910. [Google Scholar] [CrossRef]
  41. Shen, Z.Y.; Fang, Y.; Zhen, L.; Zhu, X.J.; Chen, H.; Liu, H.; Jiang, B.; Li, G.X.; Deng, H.J. Analysis of the predictive efficiency of S100P on adverse prognosis and the pathogenesis of S100P-mediated invasion and metastasis of colon adenocarcinoma. Cancer Genet. 2016, 209, 143–153. [Google Scholar] [CrossRef]
  42. Teraoku, H.; Morine, Y.; Ikemoto, T.; Saito, Y.; Yamada, S.; Yoshikawa, M.; Takasu, C.; Higashijima, J.; Imura, S.; Shimada, M. Role of thrombospondin-1 expression in colorectal liver metastasis and its molecular mechanism. J. Hepatobiliary Pancreat. Sci. 2016, 23, 565–573. [Google Scholar] [CrossRef] [PubMed]
  43. Woischke, C.; Blaj, C.; Schmidt, E.M.; Lamprecht, S.; Engel, J.; Hermeking, H.; Kirchner, T.; Horst, D. CYB5R1 links epithelial-mesenchymal transition and poor prognosis in colorectal cancer. Oncotarget 2016, 7, 31350–31360. [Google Scholar] [CrossRef] [PubMed]
  44. Yang, P.; Chen, T.; Xu, Z.; Zhu, H.; Wang, J.; He, Z. Long noncoding RNA GAPLINC promotes invasion in colorectal cancer by targeting SNAI2 through binding with PSF and NONO. Oncotarget 2016, 7, 42183–42194. [Google Scholar] [CrossRef] [PubMed]
  45. Yokota, M.; Kojima, M.; Higuchi, Y.; Nishizawa, Y.; Kobayashi, A.; Ito, M.; Saito, N.; Ochiai, A. Gene expression profile in the activation of subperitoneal fibroblasts reflects prognosis of patients with colon cancer. Int. J. Cancer 2016, 138, 1422–1431. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  46. Chen, Z.; Zhu, Y.; Dong, Y.; Zhang, P.; Han, X.; Jin, J.; Ma, X. Overexpression of TrpC5 promotes tumor metastasis via the HIF-1α-Twist signaling pathway in colon cancer. Clin. Sci. 2017, 131, 2439–2450. [Google Scholar] [CrossRef]
  47. Forse, C.L.; Rahimi, M.; Diamandis, E.P.; Assarzadegan, N.; Dawson, H.; Grin, A.; Kennedy, E.; O’Connor, B.; Messenger, D.E.; Riddell, R.H.; et al. HtrA3 stromal expression is correlated with tumor budding in stage II colorectal cancer. Exp. Mol. Pathol. 2017, 103, 94–100. [Google Scholar] [CrossRef]
  48. Lu, M.; Liu, Z.; Li, B.; Wang, G.; Li, D.; Zhu, Y. The high expression of long non-coding RNA PANDAR indicates a poor prognosis for colorectal cancer and promotes metastasis by EMT pathway. J. Cancer Res. Clin. Oncol. 2017, 143, 71–81. [Google Scholar] [CrossRef]
  49. Miyoshi, J.; Toden, S.; Yoshida, K.; Toiyama, Y.; Alberts, S.R.; Kusunoki, M.; Sinicrope, F.A.; Goel, A. MiR-139-5p as a novel serum biomarker for recurrence and metastasis in colorectal cancer. Sci. Rep. 2017, 7. [Google Scholar] [CrossRef]
  50. Wang, Q.; Yang, L.; Hu, X.; Jiang, Y.; Hu, Y.; Liu, Z.; Liu, J.; Wen, T.; Ma, Y.; An, G.; et al. Upregulated NNT-AS1, a long noncoding RNA, contributes to proliferation and migration of colorectal cancer cells in vitro and in vivo. Oncotarget 2017, 8, 3441–3453. [Google Scholar] [CrossRef] [Green Version]
  51. Yusup, A.; Huji, B.; Fang, C.; Wang, F.; Dadihan, T.; Wang, H.J.; Upur, H. Expression of trefoil factors and TWIST1 in colorectal cancer and their correlation with metastatic potential and prognosis. World J. Gastroenterol. 2017, 23, 110–120. [Google Scholar] [CrossRef]
  52. Zhao, J.; Ou, B.; Han, D.; Wang, P.; Zong, Y.; Zhu, C.; Liu, D.; Zheng, M.; Sun, J.; Feng, H.; et al. Tumor-derived CXCL5 promotes human colorectal cancer metastasis through activation of the ERK/Elk-1/Snail and AKT/GSK3β/β-catenin pathways. Mol. Cancer 2017, 16, 1–15. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  53. Cho, S.Y.; Kim, J.S.; Eun, H.S.; Kang, S.H.; Lee, E.S.; Kim, S.H.; Sung, J.K.; Lee, B.S.; Jeong, H.Y.; Moon, H.S. Expression of NOX Family Genes and Their Clinical Significance in Colorectal Cancer. Dig. Dis. Sci. 2018, 63, 2332–2340. [Google Scholar] [CrossRef] [PubMed]
  54. Hu, X.; Li, Y.Q.; Li, Q.G.; Ma, Y.L.; Peng, J.J.; Cai, S.J. ITGAE Defines CD8+ Tumor-Infiltrating Lymphocytes Predicting a better Prognostic Survival in Colorectal Cancer. EBioMedicine 2018, 35, 178–188. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  55. Konishi, Y.; Kawamata, F.; Nishihara, H.; Homma, S.; Kato, Y.; Tsuda, M.; Kohsaka, S.; Einama, T.; Liu, C.; Yoshida, T.; et al. Tumor budding and human chorionic gonadotropin-β expression correlate with unfavorable patient outcome in colorectal carcinoma. Med. Oncol. 2018, 35, 104. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  56. Liu, M.; Zhang, X.; Cai, J.; Li, Y.; Luo, Q.; Wu, H.; Yang, Z.; Wang, L.; Chen, D. Downregulation of TRIM58 expression is associated with a poor patient outcome and enhances colorectal cancer cell invasion. Oncol. Rep. 2018, 40, 1251–1260. [Google Scholar] [CrossRef] [PubMed]
  57. Wang, W.; Wan, L.; Wu, S.; Yang, J.; Zhou, Y.; Liu, F.; Wu, Z.; Cheng, Y. Mesenchymal marker and LGR5 expression levels in circulating tumor cells correlate with colorectal cancer prognosis. Cell. Oncol. 2018, 41, 495–504. [Google Scholar] [CrossRef] [PubMed]
  58. Wang, B.; Mujun, Y.; Cheng, C.; Jiang, H.; Jiang, K.; Shen, Z.; Yingjiang, Y.; Wang, S. Decreased expression of miR-490-3p in colorectal cancer predicts poor prognosis and promotes cell proliferation and invasion by targeting RAB14. Int. J. Oncol. 2018, 53, 1247–1256. [Google Scholar] [CrossRef] [PubMed]
  59. Yang, Z.; Zhang, C.; Qi, W.; Cui, C.; Cui, Y.; Xuan, Y. Tenascin-C as a prognostic determinant of colorectal cancer through induction of epithelial-to-mesenchymal transition and proliferation. Exp. Mol. Pathol. 2018, 105, 216–222. [Google Scholar] [CrossRef]
  60. Zhu, B.; Wang, Y.; Wang, X.; Wu, S.; Zhou, L.; Gong, X.; Song, W.; Wang, D. Evaluation of the correlation of MACC1, CD44, Twist1, and KiSS-1 in the metastasis and prognosis for colon carcinoma. Diagn. Pathol. 2018, 13, 45. [Google Scholar] [CrossRef] [Green Version]
  61. Gunderson, L.L.; Jessup, J.M.; Sargent, D.J.; Greene, F.L.; Stewart, A.K. Revised TN categorization for colon cancer based on national survival outcomes data. J. Clin. Oncol. 2010, 28, 264–271. [Google Scholar] [CrossRef] [Green Version]
  62. Greene, F.L.; Stewart, A.K.; Norton, H.J.; Cohen, A.M. A new TNM staging strategy for node-positive (stage III) colon cancer: An analysis of 50,042 patients. Ann. Surg. 2002, 236, 416–421. [Google Scholar] [CrossRef] [PubMed]
  63. Li, J.; Guo, B.C.; Sun, L.R.; Wang, J.W.; Fu, X.H.; Zhang, S.Z.; Poston, G.; Ding, K.F. TNM staging of colorectal cancer should be reconsidered by T stage weighting. World J. Gastroenterol. 2014, 20, 5104–5112. [Google Scholar] [CrossRef] [PubMed]
  64. Rahbari, N.N.; Carr, P.R.; Jansen, L.; Chang-Claude, J.; Weitz, J.; Hoffmeister, M.; Brenner, H. Time of Metastasis and Outcome in Colorectal Cancer. Ann. Surg. 2019, 269, 494–502. [Google Scholar] [CrossRef] [PubMed]
  65. Vatandoust, S.; Price, T.J.; Karapetis, C.S. Colorectal cancer: Metastases to a single organ. World J. Gastroenterol. 2015, 21, 11767–11776. [Google Scholar] [CrossRef]
  66. Marzouk, O.; Schofield, J. Review of histopathological and molecular prognostic features in colorectal cancer. Cancers 2011, 3, 2767–2810. [Google Scholar] [CrossRef] [Green Version]
  67. Ueno, H.; Shinto, E.; Kajiwara, Y.; Fukazawa, S.; Shimazaki, H.; Yamamoto, J.; Hase, K. Prognostic impact of histological categorisation of epithelial-mesenchymal transition in colorectal cancer. Br. J. Cancer 2014, 111, 2082–2090. [Google Scholar] [CrossRef]
  68. Sargent, D.; Sobrero, A.; Grothey, A.; O’Connell, M.J.; Buyse, M.; Andre, T.; Zheng, Y.; Green, E.; Labianca, R.; O’Callaghan, C.; et al. Evidence for cure by adjuvant therapy in colon cancer: Observations based on individual patient data from 20,898 patients on 18 randomized trials. J. Clin. Oncol. 2009, 27, 872–877. [Google Scholar] [CrossRef]
  69. Shah, M.A.; Renfro, L.A.; Allegra, C.J.; André, T.; De Gramont, A.; Schmoll, H.J.; Haller, D.G.; Alberts, S.R.; Yothers, G.; Sargent, D.J. Impact of patient factors on recurrence risk and time dependency of oxaliplatin benefit in patients with colon cancer: Analysis from modern-era adjuvant studies in the adjuvant colon cancer end points (ACCENT) database. J. Clin. Oncol. 2016, 34, 843–853. [Google Scholar] [CrossRef]
  70. Barone, C.; Nuzzo, G.; Cassano, A.; Basso, M.; Schinzari, G.; Giuliante, F.; D’Argento, E.; Trigila, N.; Astone, A.; Pozzo, C. Final analysis of colorectal cancer patients treated with irinotecan and 5-fluorouracil plus folinic acid neoadjuvant chemotherapy for unresectable liver metastases. Br. J. Cancer 2007, 97, 1035–1039. [Google Scholar] [CrossRef] [Green Version]
  71. Harris, E.I.; Lewin, D.N.; Wang, H.L.; Lauwers, G.Y.; Srivastava, A.; Shyr, Y.; Shakhtour, B.; Revetta, F.; Washington, M.K. Lymphovascular invasion in colorectal cancer: An interobserver variability study. Am. J. Surg. Pathol. 2008, 32, 1816–1821. [Google Scholar] [CrossRef] [Green Version]
  72. Jiang, H.H.; Zhang, Z.Y.; Wang, X.Y.; Tang, X.; Liu, H.L.; Wang, A.L.; Li, H.G.; Tang, E.J.; Lin, M. Bin Prognostic significance of lymphovascular invasion in colorectal cancer and its association with genomic alterations. World J. Gastroenterol. 2019, 25, 2489–2502. [Google Scholar] [CrossRef] [PubMed]
  73. Barresi, V.; Reggiani Bonetti, L.; Vitarelli, E.; Di Gregorio, C.; Ponz De Leon, M.; Barresi, G. Immunohistochemical assessment of lymphovascular invasion in stage i colorectal carcinoma: Prognostic relevance and correlation with nodal micrometastases. Am. J. Surg. Pathol. 2012, 36, 66–72. [Google Scholar] [CrossRef] [PubMed]
  74. Fujii, T.; Sutoh, T.; Morita, H.; Yajima, R.; Yamaguchi, S.; Tsutsumi, S.; Asao, T.; Kuwano, H. Vascular invasion, but not lymphatic invasion, of the primary tumor is a strong prognostic factor in patients with colorectal cancer. Anticancer Res. 2014, 34, 3147–3151. [Google Scholar] [PubMed]
  75. Crozier, J.E.M.; McMillan, D.C.; McArdle, C.S.; Angerson, W.J.; Anderson, J.H.; Horgan, P.G.; McKee, R.F. Tumor size is associated with the systemic inflammatory response but not survival in patients with primary operable colorectal cancer. J. Gastroenterol. Hepatol. 2007, 22, 2288–2291. [Google Scholar] [CrossRef]
  76. Saha, S.; Shaik, M.; Johnston, G.; Saha, S.K.; Berbiglia, L.; Hicks, M.; Gernand, J.; Grewal, S.; Arora, M.; Wiese, D. Tumor size predicts long-term survival in colon cancer: An analysis of the National Cancer Data Base. Am. J. Surg. 2015, 209, 570–574. [Google Scholar] [CrossRef] [PubMed]
  77. Li, X.; An, B.; Ma, J.; He, B.; Qi, J.; Wang, W.; Qin, C.; Zhao, Q. Prognostic value of the tumor size in resectable colorectal cancer with different primary locations: A retrospective study with the propensity score matching. J. Cancer 2019, 10, 313–322. [Google Scholar] [CrossRef] [Green Version]
  78. Loupakis, F.; Yang, D.; Yau, L.; Feng, S.; Cremolini, C.; Zhang, W.; Maus, M.K.H.; Antoniotti, C.; Langer, C.; Scherer, S.J.; et al. Primary tumor location as a prognostic factor in metastatic colorectal cancer. J. Natl. Cancer Inst. 2015, 107, dju427. [Google Scholar] [CrossRef]
  79. Wang, C.B.; Shahjehan, F.; Merchea, A.; Li, Z.; Bekaii-Saab, T.S.; Grothey, A.; Colibaseanu, D.T.; Kasi, P.M. Impact of tumor location and variables associated with overall survival in patients with colorectal cancer: A mayo clinic colon and rectal cancer registry study. Front. Oncol. 2019, 9, 76. [Google Scholar]
  80. Patel, M.; McSorley, S.T.; Park, J.H.; Roxburgh, C.S.D.; Edwards, J.; Horgan, P.G.; McMillan, D.C. The relationship between right-sided tumour location, tumour microenvironment, systemic inflammation, adjuvant therapy and survival in patients undergoing surgery for colon and rectal cancer. Br. J. Cancer 2018, 118, 705–712. [Google Scholar] [CrossRef] [Green Version]
  81. Bajenova, O.; Chaika, N.; Tolkunova, E.; Davydov-Sinitsyn, A.; Gapon, S.; Thomas, P.; O’Brien, S. Carcinoembryonic antigen promotes colorectal cancer progression by targeting adherens junction complexes. Exp. Cell Res. 2014, 324, 115–123. [Google Scholar] [CrossRef]
  82. Yoshikawa, M.; Morine, Y.; Ikemoto, T.; Imura, S.; Higashijima, J.; Iwahashi, S.; Saito, Y.; Takasu, C.; Yamada, S.; Ishikawa, D.; et al. Elevated preoperative serum cea level is associated with poor prognosis in patients with hepatocellular carcinoma through the epithelial-mesenchymal transition. Anticancer Res. 2017, 37, 1169–1175. [Google Scholar] [CrossRef] [Green Version]
  83. Destri, G.L.; Rubino, A.S.; Latino, R.; Giannone, F.; Lanteri, R.; Scilletta, B.; Di Cataldo, A. Preoperative carcinoembryonic antigen and prognosis of colorectal cancer. An independent prognostic factor still reliable. Int. Surg. 2015, 100, 617–625. [Google Scholar] [CrossRef] [Green Version]
  84. Bhatti, I.; Patel, M.; Dennison, A.R.; Thomas, M.W.; Garcea, G. Utility of postoperative CEA for surveillance of recurrence after resection of primary colorectal cancer. Int. J. Surg. 2015, 16, 123–128. [Google Scholar] [CrossRef] [PubMed]
  85. Poruk, E.K.; Gay, Z.D.; Brown, K.D.; Mulvihill, J.; Boucher, M.K.; Scaife, L.C.; Firpo, A.M.; Mulvihill, J.S. The Clinical Utility of CA 19-9 in Pancreatic Adenocarcinoma: Diagnostic and Prognostic Updates. Curr. Mol. Med. 2013, 13, 340–351. [Google Scholar] [CrossRef] [PubMed]
  86. Stiksma, J.; Grootendorst, D.C.; Van Der Linden, P.W.G. CA 19-9 as a marker in addition to CEA to monitor colorectal cancer. Clin. Colorectal Cancer 2014, 13, 239–244. [Google Scholar] [CrossRef] [PubMed]
  87. Hirakawa, M.; Takimoto, R.; Tamura, F.; Yoshida, M.; Ono, M.; Murase, K.; Sato, Y.; Osuga, T.; Sato, T.; Iyama, S.; et al. Fucosylated TGF-β receptors transduces a signal for epithelial-mesenchymal transition in colorectal cancer cells. Br. J. Cancer 2014, 110, 156–163. [Google Scholar] [CrossRef] [Green Version]
  88. Li, L.T.; Jiang, G.; Chen, Q.; Zheng, J.N. Predic Ki67 is a promising molecular target in the diagnosis of cancer (Review). Mol. Med. Rep. 2015, 11, 1566–1572. [Google Scholar] [CrossRef] [Green Version]
  89. Luo, Z.W.; Zhu, M.G.; Zhang, Z.Q.; Ye, F.J.; Huang, W.H.; Luo, X.Z. Increased expression of Ki-67 is a poor prognostic marker for colorectal cancer patients: A meta analysis. BMC Cancer 2019, 19, 123. [Google Scholar] [CrossRef] [Green Version]
  90. Tong, G.; Zhang, G.; Liu, J.; Zheng, Z.; Chen, Y.; Niu, P.; Xu, X. Cutoff of 25% for Ki67 expression is a good classification tool for prognosis in colorectal cancer in the AJCC-8 stratification. Oncol. Rep. 2020, 43, 1187–1198. [Google Scholar] [CrossRef] [Green Version]
  91. Melling, N.; Kowitz, C.M.; Simon, R.; Bokemeyer, C.; Terracciano, L.; Sauter, G.; Izbicki, J.R.; Marx, A.H. High Ki67 expression is an independent good prognostic marker in colorectal cancer. J. Clin. Pathol. 2016, 69, 209–214. [Google Scholar] [CrossRef] [Green Version]
  92. Gasinska, A.; Jaszczynski, J.; Adamczyk, A.; Janecka-Widla, A.; Wilk, W.; Cichocka, A.; Stelmach, A. Biomarkers of epithelial-mesenchymal transition in localized, surgically treated clear-cell renal cell carcinoma. Folia Histochem. Cytobiol. 2019, 56, 195–206. [Google Scholar] [CrossRef] [PubMed]
  93. Hemalatha, A.; Suresh, T.N.; Harendra Kumar, M.L. Expression of vimentin in breast carcinoma, its correlation with Ki67 and other histopathological parameters. Indian J. Cancer 2013, 50, 189–194. [Google Scholar] [CrossRef] [PubMed]
  94. Onder, T.T.; Gupta, P.B.; Mani, S.A.; Yang, J.; Lander, E.S.; Weinberg, R.A. Loss of E-cadherin promotes metastasis via multiple downstream transcriptional pathways. Cancer Res. 2008, 68, 3645–3654. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  95. Schetter, A.J.; Okayama, H.; Harris, C.C. The role of MicroRNAs in colorectal cancer. Cancer J. 2012, 18, 244–252. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  96. Zaravinos, A. The regulatory role of MicroRNAs in EMT and cancer. J. Oncol. 2015, 2015, 865816. [Google Scholar] [CrossRef] [Green Version]
  97. Luo, X.; Burwinkel, B.; Tao, S.; Brenner, H. MicroRNA signatures: Novel biomarker for colorectal cancer? Cancer Epidemiol. Biomarkers Prev. 2011, 20, 1272–1286. [Google Scholar] [CrossRef] [Green Version]
  98. Li, Q.; Liang, X.; Wang, Y.; Meng, X.; Xu, Y.; Cai, S.; Wang, Z.; Liu, J.; Cai, G. MiR-139-5p Inhibits the Epithelial-Mesenchymal Transition and Enhances the Chemotherapeutic Sensitivity of Colorectal Cancer Cells by Downregulating BCL2. Sci. Rep. 2016, 6, 1–10. [Google Scholar] [CrossRef]
  99. Xu, X.; Chen, R.; Li, Z.; Huang, N.; Wu, X.; Li, S.; Li, Y.; Wu, S. MicroRNA-490-3p inhibits colorectal cancer metastasis by targeting TGFβR1. BMC Cancer 2015, 15, 1023. [Google Scholar] [CrossRef] [Green Version]
  100. Gugnoni, M.; Ciarrocchi, A. Long noncoding RNA and epithelial mesenchymal transition in cancer. Int. J. Mol. Sci. 2019, 20, 1924. [Google Scholar] [CrossRef] [Green Version]
  101. Ge, X.S.; Ma, H.J.; Zheng, X.H.; Ruan, H.L.; Liao, X.Y.; Xue, W.Q.; Chen, Y.B.; Zhang, Y.; Jia, W.H. HOTAIR, a prognostic factor in esophageal squamous cell carcinoma, inhibits WIF-1 expression and activates Wnt pathway. Cancer Sci. 2013, 104, 1675–1682. [Google Scholar] [CrossRef]
  102. Guinney, J.; Dienstmann, R.; Wang, X.; De Reyniès, A.; Schlicker, A.; Soneson, C.; Marisa, L.; Roepman, P.; Nyamundanda, G.; Angelino, P.; et al. The consensus molecular subtypes of colorectal cancer. Nat. Med. 2015, 21, 1350–1356. [Google Scholar] [CrossRef] [PubMed]
  103. Mooi, J.K.; Wirapati, P.; Asher, R.; Lee, C.K.; Savas, P.; Price, T.J.; Townsend, A.; Hardingham, J.; Buchanan, D.; Williams, D.; et al. The prognostic impact of consensus molecular subtypes (CMS) and its predictive effects for bevacizumab benefit in metastatic colorectal cancer: Molecular analysis of the AGITG MAX clinical trial. Ann. Oncol. 2018, 29, 2240–2246. [Google Scholar] [CrossRef] [PubMed]
  104. Stintzing, S.; Wirapati, P.; Lenz, H.J.; Neureiter, D.; Fischer von Weikersthal, L.; Decker, T.; Kiani, A.; Kaiser, F.; Al-Batran, S.; Heintges, T.; et al. Consensus molecular subgroups (CMS) of colorectal cancer (CRC) and first-line efficacy of FOLFIRI plus cetuximab or bevacizumab in the FIRE3 (AIO KRK-0306) trial. Ann. Oncol. 2019, 30, 1796–1803. [Google Scholar] [CrossRef] [Green Version]
  105. Sosa, V.; Moliné, T.; Somoza, R.; Paciucci, R.; Kondoh, H.; LLeonart, M.E. Oxidative stress and cancer: An overview. Ageing Res. Rev. 2013, 12, 376–390. [Google Scholar] [CrossRef] [PubMed]
  106. Sorolla, M.A.; Reverter-Branchat, G.; Tamarit, J.; Ferrer, I.; Ros, J.; Cabiscol, E. Proteomic and oxidative stress analysis in human brain samples of Huntington disease. Free Radic. Biol. Med. 2008, 45, 667–678. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  107. Prieto-Bermejo, R.; Hernández-Hernández, A. The importance of NADPH oxidases and redox signaling in angiogenesis. Antioxidants 2017, 6, 32. [Google Scholar] [CrossRef] [Green Version]
  108. Arbiser, J.L.; Petros, J.; Klafter, R.; Govindajaran, B.; McLaughlin, E.R.; Brown, L.F.; Cohen, C.; Moses, M.; Kilroy, S.; Arnold, R.S.; et al. Reactive oxygen generated by Nox1 triggers the angiogenic switch. Proc. Natl. Acad. Sci. USA 2002, 99, 715–720. [Google Scholar] [CrossRef] [Green Version]
  109. Weyemi, U.; Lagente-Chevallier, O.; Boufraqech, M.; Prenois, F.; Courtin, F.; Caillou, B.; Talbot, M.; Dardalhon, M.; Al Ghuzlan, A.; Bidart, J.M.; et al. ROS-generating NADPH oxidase NOX4 is a critical mediator in oncogenic H-Ras-induced DNA damage and subsequent senescence. Oncogene 2012, 31, 1117–1129. [Google Scholar] [CrossRef] [Green Version]
  110. Jiang, J.; Wang, K.; Chen, Y.; Chen, H.; Nice, E.C.; Huang, C. Redox regulation in tumor cell epithelial–mesenchymal transition: Molecular basis and therapeutic strategy. Signal Transduct. Target. Ther. 2017, 2, 17036. [Google Scholar] [CrossRef] [Green Version]
  111. Stewart, L.A.; Clarke, M.; Rovers, M.; Riley, R.D.; Simmonds, M.; Stewart, G.; Tierney, J.F. Preferred reporting items for a systematic review and meta-analysis of individual participant data: The PRISMA-IPD statement. JAMA 2015, 313, 1657–1665. [Google Scholar] [CrossRef]
  112. Higgins, J.P.T.; Thompson, S.G. Quantifying heterogeneity in a meta-analysis. Stat. Med. 2002, 21, 1539–1558. [Google Scholar] [CrossRef] [PubMed]
  113. Raudvere, U.; Kolberg, L.; Kuzmin, I.; Arak, T.; Adler, P.; Peterson, H.; Vilo, J. G:Profiler: A web server for functional enrichment analysis and conversions of gene lists (2019 update). Nucleic Acids Res. 2019, 41, W191–W198. [Google Scholar] [CrossRef] [PubMed] [Green Version]
Cancers 12 03330 g001 550
Figure 1. Study flow diagram. Systematic revision workflow from search to included studies for meta-analysis. Exclusion criteria of discarded studies are specified.
Figure 1. Study flow diagram. Systematic revision workflow from search to included studies for meta-analysis. Exclusion criteria of discarded studies are specified.
Cancers 12 03330 g001
Cancers 12 03330 g002a 550Cancers 12 03330 g002b 550Cancers 12 03330 g002c 550Cancers 12 03330 g002d 550
Figure 2. Meta-analysis of EMT-clinical and molecular prognostic biomarkers predicting OS in CRC. (A) Forest and funnel plots of pathological T, N and M; (B) forest and funnel plots of pathological stage, lymphovascular invasion and histological grade; (C) forest and funnel plots of tumor size, tumor location and chemotherapy administration and (D) forest and funnel plots of CEA, CA19.9, Ki67 and E-cadherin. Red rombs represent specific HR for each study and their size determines individual weight. Black rombs represent HR global value and their size represents 95%CI. All Studies and Subgroups are referenced in Table 1.
Figure 2. Meta-analysis of EMT-clinical and molecular prognostic biomarkers predicting OS in CRC. (A) Forest and funnel plots of pathological T, N and M; (B) forest and funnel plots of pathological stage, lymphovascular invasion and histological grade; (C) forest and funnel plots of tumor size, tumor location and chemotherapy administration and (D) forest and funnel plots of CEA, CA19.9, Ki67 and E-cadherin. Red rombs represent specific HR for each study and their size determines individual weight. Black rombs represent HR global value and their size represents 95%CI. All Studies and Subgroups are referenced in Table 1.
Cancers 12 03330 g002aCancers 12 03330 g002bCancers 12 03330 g002cCancers 12 03330 g002d
Cancers 12 03330 g003a 550Cancers 12 03330 g003b 550
Figure 3. Meta-analysis of EMT-clinical and molecular prognostic biomarkers predicting PFS in CRC. (A) Forest and funnel plot of pathological T, N and M and stage and (B) forest and funnel plot of lymphovascular invasion, histological grade and CEA. Red rombs represent specific HR for each study and their size determines individual weight. Black rombs represent HR global value and their size represents 95%CI. All Studies and Subgroups are referenced in Table 1.
Figure 3. Meta-analysis of EMT-clinical and molecular prognostic biomarkers predicting PFS in CRC. (A) Forest and funnel plot of pathological T, N and M and stage and (B) forest and funnel plot of lymphovascular invasion, histological grade and CEA. Red rombs represent specific HR for each study and their size determines individual weight. Black rombs represent HR global value and their size represents 95%CI. All Studies and Subgroups are referenced in Table 1.
Cancers 12 03330 g003aCancers 12 03330 g003b
Cancers 12 03330 g004 550
Figure 4. Classification of EMT-related prognostic molecular biomarkers in CRC. (A) Molecular biomarkers predicting OS and (B) molecular biomarkers predicting PFS.
Figure 4. Classification of EMT-related prognostic molecular biomarkers in CRC. (A) Molecular biomarkers predicting OS and (B) molecular biomarkers predicting PFS.
Cancers 12 03330 g004
Cancers 12 03330 g005 550
Figure 5. Profiling analysis of EMT-related molecular prognostic biomarkers in CRC. (A) Manhattan Plot of the significantly changed terms enriched by GO and MIRNA databases. Circular graphics about the percentages of the specific –log10 (padjust) values of the most represented categories over the total, for: (B) molecular function of MF, (C) biological process or BP and (D) cellular compartment or CC.
Figure 5. Profiling analysis of EMT-related molecular prognostic biomarkers in CRC. (A) Manhattan Plot of the significantly changed terms enriched by GO and MIRNA databases. Circular graphics about the percentages of the specific –log10 (padjust) values of the most represented categories over the total, for: (B) molecular function of MF, (C) biological process or BP and (D) cellular compartment or CC.
Cancers 12 03330 g005
Table
Table 1. Characteristics of included studies for meta-analysis. All studies have included colorectal cancer (CRC) patients with described epithelial–mesenchymal transition (EMT)-related clinical and molecular biomarkers predicting overall survival (OS) and/or progression free survival (PFS).
Table 1. Characteristics of included studies for meta-analysis. All studies have included colorectal cancer (CRC) patients with described epithelial–mesenchymal transition (EMT)-related clinical and molecular biomarkers predicting overall survival (OS) and/or progression free survival (PFS).
PMIDFirst AuthorYearPredicted EndpointIncluded Variables in Multivariate Logistic ModelDesignValitation CohortTotal Number of PatientsClinical StageSample/TechniqueTissue Analyzed
24752533Barbazan [23]2014OS, PFS4-week CTC marker model, baseline CTC marker model, ECOG performance status, lung metastasisUnicentricNo50TNM: IV: 50CTC/RTqPCRColon and rectum
24738665Gao [24]2014OSILEI expresssion, TNM stage, peritumoral deposits, lymphatic invasion, venous invasionUnicentricNo194TNM: I: 55; II: 59, III: 74, IV: 6Tumor/IHCColon and rectum
24217791Sugimachi [25]2014OSPlastin 3 expression, pT, pN, lymphatic invasion, venous invasion, Dukes stage, CEA and CA19.9 levelsUnicentricNo177Dukes: A–B: 101; C–D: 76Tumor and blood/RTqPCRColon and rectum
24840737Wu [26]2014OS, MFSHOTAIR gene expression, pN, pM, AJCC stage, HGUnicentricNo120TNM: I: 13, II: 44, III: 49, IV: 14Tumor/RTqPCRColon
25431208Zhou [27]2014OS, PFSOTUB1 expression, age, gender, tumor location, tumor size, pT, pN, pM, CEA levels, therapyUnicentricNo260TNM: I: 61, II: 63, III: 76, IV: 60Tumor/IHCColon and rectum
25382057Dawson [28]2015OSmTrkB expression, pT, pN, adjuvant therapyUnicentricNo211TN: I–III: 182, IV: 22, UNK: 7Tumor/IHCColon and rectum
25951369Kahlert [29]2015OSSIX1 expression, KRAS and MSI status, age, gender, HG, UICC stage, therapyMulticentricYes
Internal cohort: 817 patients		945TNM: Cohort 1: I: 42,II: 45, III: 41.Cohort 2: I: 188, II: 314, III: 315Tumor/IHC, RTqPCRColon and rectum
25947346Liu [30]2015OS, PFSGDH expression, HG, pT, pN, pM, venous invasion, nervous invasionUnicentricNo104TNM: I/II: 42, III/IV: 62Tumor/IHCColon and rectum
25936636Yan [16]2015OSN-cadherin and E-cadherin expression, age, gender, tumor location, HG, tumor size, pT, pN, pMUnicentricNo102TNM: I–II–III: 53, IV: 49Tumor/IHC, RTqPCRColon and rectum
25716692Xiao [31]2015OSBTPF expression, vimentin expression, E-cadherin expression, age, gender, tumor size, HG, UICC stage, pN, pM, recurrenceUnicentricNo105TNM: I–II: 61, III–V: 44Tumor/IHC, RTqPCRColon and rectum
26507436Busch [32]2016OSE-cadherin and Snail expressionMulticentricNo190TNM:
Local: 100 Regional: 66 Distant: 24		Tumor/IHCColon and rectum
27520310Gao [33]2016OSDCLK1 expression, TNM stage, pMUnicentricNo71TNM: I: 11, II: 17, III: 32, IV: 11Tumor/IHC RTqPCRColon and rectum
27037526Hashimoto [34]2016OS, DFSh-Prune expression, pN, pM, hepatectomy type, CEA and CA19-9 levelsUnicentricNo87TNM: IV: 87Tumor/IHCColon and rectum
27537253Li [35]2016OS, PFSKi67 and MAEL expression, pT, pN, HGUnicentricNo185TNM: I–II: 105, III: 80Tumor/IHCColon and rectum
27323857Li [36]2016OS, MFSPAK7 and Ki67 expression, age, gender, location, pT, pN, pM, AJCC stage, vascular invasion, HGUnicentricNo203TNM: I: 24, II: 81, III: 80, IV: 18Tumor/IHC, RTqPCRColon
27046094Luo [37]2016OSRab3D expression, pT, pN, pM and CEA levelsUnicentricNo300TNM: I: 48, II: 92, III: 103, IV: 7Tumor/IHCColon and rectum
26983880Matsusaka [38]2016OS, PFSTWIST1, SNAIL, ZEB1 and E-cadherin gene polymorphismsUnicentricNo220TNM: Beva cohort: IV: 143 Cetu cohort: IV: 77Tumor/Sanger SeqColon and rectum
27503579Ning [39]2016OSALDH1, PI3KCA and AKT2 gene expression in CTC’sUnicentricNo78TNM: IV: 78CTC/RTqPCRColon and rectum
27363678Satelli [40]2016OS, PFSCTC counts, PD-L1 expression in CTC’sUnicentricNo62TNM: IV: 62CTC/Flow cytometry, ICCColon
26975699Shen [41]2016OSS100P expression, CEA and CA19-9 levels, tumor size, HG, TNM stageUnicentricoNo125TNM: I/II: 50 III/IV: 75Tumor/IHC RTqPCRColon
27404020Teraoku [42]2016OSHG, THBS1, Ki67 and E-cadherin expressionUnicentricNo94TNM: IV: 94Tumor/IHCColon and rectum
27120783Woischke [43]2016CSSCYB5R1 expression, age, gender, location, pT, HGUnicentricYes: TCGA cohort: 457 patients678TNM: Dis cohort: I: 1, II: 35, III: 177, IV: 8Tumor/IHCColon and rectum
27259250Yang [44]2016OSGAPLINC gene expression, tumor size, pT, pN, HGUnicentricNo180TNM: I: 13; II: 96, III: 66, IV: 3Tumor/ISHColon and rectum
26370611Yokota [45]2016RFSCTGF-CALD1-INHBA-TAGLN gene expression, PDLIM5-MAGL1-SPTBN-TAGLN gene expression and Dukes stageUnicentricYes: Internal cohort: 113 patients339Dukes: Dis cohort: A: 41, B: 94, C: 91 TNM: Val cohort: I: 18, II: 45, III: 42, IV: 8Tumor and SPF/RTqPCRColon
28864720Chen [46]2017OS, MFSTrpC5 expression, AJCC stage, HGUnicentricNo127TNM: I: 17, II: 52, III: 45, IV: 13Tumor/IHCColon
28716573Forse [47]2017OS
DFS		sHtrA3 expression, gender, pT, pN, tumor location, lymphovascular invasionUnicentricNo172TNM: II: 172Tumor/IHCColon and rectum
27629879Lu [48]2017OSPANDAR gene expression, pT, TNM stageUnicentricNo124TNM: I/II: 46 III/IV: 78Tumor/RTqPCRColon and rectum
28262692Miyoshi [49]2017RFSmiR-139-5p gene expression, pT, pNMulticentricYes: Internal cohorts 1 and 2 = 111 and 139 patients497TNM: Dis cohorts: MC: III: 100; TCGA: II: 42, III: 105; Val cohorts: Cohort 1: II: 60, III: 50 Cohort 2: I: 12, II: 67, III: 60Tumor and blood/RTqPCRColon and rectum
27966450Wang [50]2017OSNNT-AS1 gene expression, TNM stage, vascular invasionUnicentricNo70TNM: I/II: 36 III/IV: 34Tumor/RTqPCRColon and rectum
28104986Yusup [51]2017OSTFF1, TFF3, TWIST1 expression, age, gender, histopathological type, HG, pT, pN, pM, TNM stageUnicentricNo75TNM: I: 15, II: 27, III:27, IV: 6Tumor/IHCColon and rectum
28356111Zhao [52]2017OS, DFSCXCL5 expression, Dukes stageUnicentricNo78TNM I: 15, II: 43, III: 24, IV: 6Tumor/IHC, RTqPCRColon and rectum
29781053Cho [53]2018OSNOX1, NOX2, NOX3, NOX4, NOX5, DUOX1, DUOX2 gene expressionUnicentricNo458TNM: I: 76, II: 178, III: 129, IV: 65, UNK: 10Tumor/microaray mRNA expressionColon and rectum
30100393Hu [54]2018OS, PFSITGAE and TIL expression, TNM stage, pT, pN, pM, perineural invasion, vascular invasion, CEA levelsUnicentricYes: Internal cohort: 276 patients1154TNM: Val cohort: I: 21, II: 81, III: 132, IV: 42Tumor/IHCColon and rectum
29892782Konishi [55]2018OShCGBb expression, vascular invasion, tumor buddingUnicentricNo80TNM: I–II: 53, III: 27Tumor/IHCColon and rectum
29956813Liu [56]2018OSTRIM58 expression, age, gender, therapy, HG, pT, pN, pMUnicentricYes: Internal cohort: 152 patients313TNM: Val cohort: I: 26, II: 63, III: 59, IV:1, UNK: 3Tumor/RTqPCRColon and rectum
29949050Wang [57]2018OS, PFSTotal CTC count, mesenchymal CTC count, LGR5 gene expression in CTCUnicentricNo66TNM: I: 31, II: 15, III–IV: 20CTC/ISHColon and rectum
29916545Wang [58]2018OSmiR-490-3p and RAB14 gene expression, age, gender, tumor size, HG, TNM stage, pN, pMUnicentricNo50TNM: I–II: 23, III–IV: 27Tumor/RTqPCRColon and rectum
30170017Yang [59]2018OS, PFSTenascin-C expression, age, pT, pN, pMUnicentricNo100TNM: I: 8, II: 38, III: 21, IV: 33Tumor/IHCColon and rectum
30021598Zhu [60]2018OSMACC1, CD44, TWIST1 and KISS-1 expression, pN, TNM stageUnicentricNo212TNM: I: 69, II: 67; III: 76Tumor/IHCColon
Abbreviations: AJCC: American Joint Committee on Cancer; CSS: cancer specific survival; CTC: circulating tumor cells; DFS: disease free survival; Dis cohort: Discovery cohort; GEO: Gene Expression Omnibus; HG: histological grade; IHC: immunohistochemistry; ICC: immunocytochemistry, MFS: metastasis free survival; OS: overall survival; PFS: progression free survival; RFS: recurrence free survival; SPF: subperitoneal fibroblasts; TCGA: The Cancer Genome Atlas; TNM: Tumor extension-Node-Metastasis; UICC: Union for International Cancer Control; UNK: unknown; Val cohort: Validation cohort.
Table
Table 2. Meta-analysis overall results. Table includes statistical results from meta-analysis, heterogeneity test, number of studies/cohorts, number of participants and effect-model applied.
Table 2. Meta-analysis overall results. Table includes statistical results from meta-analysis, heterogeneity test, number of studies/cohorts, number of participants and effect-model applied.
Clinical and Histological MarkersPredicted VariableHR (95%CI)Z (p-Value)I2 (%)Heterogeneity p-ValueNumber of Studies/CohortsNumber of ParticipantsStatistical Effect Model
pT (T3–T4 vs. T1–T2)OS1.74 (1.07–2.83)0.0394<0.0114/142449Random
PFS1.53 (1.12–2.08)<0.0100.976/61196Fixed
pN (N1–N2 vs. N0)OS2.33 (1.72–3.14)<0.0172<0.0119/193453Random
PFS2.79 (1.70–4.56)<0.0184<0.0110/111757Random
pM (M1 vs. M0)OS3.48 (2.49–4.85)<0.0154<0.0114/141968Random
PFS3.30 (2.11–5.18)<0.01560.037/7113Random
pStage (III–IV vs. I–II)OS2.73 (2.37–3.14)<0.01280.1021/233769Fixed
PFS3.19 (2.21–4.61)<0.0100.556/71054Fixed
Lymphovascular Invasion (present vs. absent)OS1.63 (1.12–2.37)0.0171<0.019/101638Random
PFS1.53 (1.19–1.95)<0.01490.124/4836Fixed
Histological Grade (high vs. intermediate-low)OS1.70 (1.47–1.97)<0.0170.3714/152694Fixed
PFS2.29 (1.61–3.24)<0.0100.884/4635Fixed
Tumor location (left vs. right)OS0.79 (0.33–1.88)0.6740.022/3578Random
Tumor location (rectum vs. colon)OS1.75 (0.85–1.88)0.13790.0083/3583Random
Tumor size (>2 vs. ≤2 cm)OS1.00 (0.70–1.43)0.99550.056/6822Random
Chemotherapy (yes vs. no)OS0.46 (0.27–0.76)<0.01610.054/41440Random
CEA (elevated vs. normal)OS1.33 (1.10–1.61)<0.0100.556/61225Fixed
PFS1.48 (1.11–1.99)<0.0100.653/3623Fixed
CA19.9 (elevated vs. normal)OS1.21 (0.90–1.62)0.21560.103/3389Random
E–cadherin (low vs. high)OS1.84 (1.21–2.78)<0.0100.954/4491Fixed
Ki67 (high vs. low)OS1.30 (0.95–1.76)0.1000.553/3482Fixed
Abbreviations: CI: confidence interval; HR: hazard ratio; OS: overall survival; PFS: progression free survival.
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Parisi, E.; Sorolla, A.; Montal, R.; González-Resina, R.; Novell, A.; Salud, A.; Sorolla, M.A. Prognostic Factors Involved in the Epithelial–Mesenchymal Transition Process in Colorectal Cancer Have a Preponderant Role in Oxidative Stress: A Systematic Review and Meta-Analysis. Cancers 2020, 12, 3330. https://doi.org/10.3390/cancers12113330

AMA Style

Parisi E, Sorolla A, Montal R, González-Resina R, Novell A, Salud A, Sorolla MA. Prognostic Factors Involved in the Epithelial–Mesenchymal Transition Process in Colorectal Cancer Have a Preponderant Role in Oxidative Stress: A Systematic Review and Meta-Analysis. Cancers. 2020; 12(11):3330. https://doi.org/10.3390/cancers12113330

Chicago/Turabian Style

Parisi, Eva, Anabel Sorolla, Robert Montal, Rita González-Resina, Anna Novell, Antonieta Salud, and Maria Alba Sorolla. 2020. "Prognostic Factors Involved in the Epithelial–Mesenchymal Transition Process in Colorectal Cancer Have a Preponderant Role in Oxidative Stress: A Systematic Review and Meta-Analysis" Cancers 12, no. 11: 3330. https://doi.org/10.3390/cancers12113330

APA Style

Parisi, E., Sorolla, A., Montal, R., González-Resina, R., Novell, A., Salud, A., & Sorolla, M. A. (2020). Prognostic Factors Involved in the Epithelial–Mesenchymal Transition Process in Colorectal Cancer Have a Preponderant Role in Oxidative Stress: A Systematic Review and Meta-Analysis. Cancers, 12(11), 3330. https://doi.org/10.3390/cancers12113330

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop