Role of metastasis-associated protein 1 in prognosis of patients with digestive tract cancers: A meta-analysis

Objectives Metastasis-associated protein 1 (MTA1) is a transcriptional regulator and significantly associated with prognosis of patients with cancer. However, its role as a potential prognostic marker in digestive tract cancer (DTC) is controversial. In this study, a meta-analysis was conducted to evaluate the MTA1 expression as a predictor of clinicopathology and survival of patients with DTC. Methods We searched PubMed, Ovid, Web of Science and Cochrane databases using multiple search strategies for eligible studies. STATA 11.0 software was used to pool the data and analyze the association, odds ratios (ORs) and 95% confidence intervals (CIs) were used to measure the strength of the association. Furthermore, the Newcastle-Ottawa scale was used to evaluate the quality of eligible studies. Results MTA1 overexpression was strongly associated with depth of invasion (OR = 1.88, 95%CI: 1.05–3.37, P = 0.03), lymph node metastasis (OR = 2.30, 95%CI: 1.76–3.01, P<0.001), vascular invasion (OR = 2.02, 95%CI: 1.40–2.91, P<0.001) and TNM stage (OR = 2.78, 95%CI: 1.63–4.74, P<0.001), and was related to 1- (RR = 1.84, 95%CI: 1.18–2.89, P = 0.008), 3- (RR = 1.74, 95%CI: 1.32–2.30, P<0.001) and 5-year (RR = 1.64, 95%CI: 1.18–2.27, P = 0.003) OS. Further, MTA1 was associated with 1- (RR = 4.16, 95%CI: 1.35–12.81, P = 0.01), 3- (RR = 1.90, 95%CI: 1.02–3.53, P = 0.04) and 5- (RR = 2.17, 95%CI: 1.41–3.32, P<0.001) year DFS. In subgroup analyses based on study quality and tumor type, MTA1 overexpression was obviously related to clinical parameters, such as lymph node metastasis and TNM stage, and was also associated with prognosis of patients with gastrointestinal or esophageal cancer. Conclusions MTA1 expression is strongly correlated with metastasis-related variables, and represents a promising prognostic factor in DTC.


Results
Introduction Digestive tract cancers (DTCs) are a heterogeneous group of gastrointestinal (GI) cancers as well as hepatobiliary and pancreatic tumors. DTCs are important causes of cancer-related deaths worldwide [1][2]. Data from the Global Cancer Statistics, 2012 [1] indicate that colorectal cancer (CRC), gastric cancer (GC), and esophageal cancer (EC) rank fourth, sixth, and tenth among all DTCs. DTC increases the risk of lymph node metastasis and distant metastasis. Lymph node status and distant metastasis are included in tumor staging, which is the most useful indicator in predicting outcomes. However, adequate evidence reminds us of the inability of conventional staging criteria to differentiate prognostic features of DTC. Metastasis is a multi-step process encompassing dissemination of primary cancer cells and subsequent colonization at distant sites [3], and is the overwhelming cause of morbidity and mortality inpatients with cancer. Therefore, elucidation of the mechanism and development of new strategies to prevent metastasis are essential to combating cancers.
Several factors are associated with the prognosis of cancer and regulation of metastasis. Metastasis-associated proteins (MTAs), especially MTA1, play prominent roles.MTA1 belongs to MTA family (consisting of MTA1, MTA2 and MTA3) that is associated with the nucleosome remodeling and histone deacetylation (NuRD) complex, which regulates transcription via histonedeacetylation and chromatin remodeling [4]. MTA1 was first reported in metastatic rat breast adenocarcinoma cell lines, where it was highly expressed compared with poorly metastatic cell lines. It plays a key role as a tumor invasion and metastasis-related gene [5]. Toh et al. [6] found that higher mRNA levels of MTA1 were closely related to depth of invasion and lymph node metastasis and a tendency toward a higher rate of lymphatic involvement. Song et al. [7] found that overexpression of MTA1protein is an independent prognostic risk factor, and is associated with shorter disease-free survival and lower 5-year survival rate.
To the best of our knowledge, only one meta-analysis reviewed the prognostic significance of MTA1 in solid tumors. The prognostic value of MTA1 in DTC is inconclusive and controversial. Therefore, we conducted a meta-analysis to investigate the role of MTA1 expression in the prognosis and survival of patients with DTC.

Search strategy
Two investigators independently searched PubMed, Ovid, Web of Science, Cochrane databases for studies published until Jul 2016. The search terms used were:("MTA1" OR "Metastasis-associated protein 1") AND ("esophagus" OR "esophageal" OR "oesophagus" OR "gullet" OR "esophago-cardiac" OR "colon" OR "colorectal" OR "rectal" OR "anal" OR "pancreas" OR "pancreatic" OR "liver" OR "hepatic" OR " biliary duct" OR "bile duct" OR "gastric" OR "stomach" OR "cardia" OR "digestive tract") AND ("carcinoma" OR "cancer" OR "tumour" OR "neoplasm" OR "tumor" OR"malignancy"). The full texts of the studies were retrieved to determine their eligibility for inclusion in the meta-analysis.

Data extraction and assessment
All the pertinent data were extracted independently from each eligible study by two investigators (Guo-dong Cao, Bo Chen). Any disagreement was resolved through discussion until a consensus was reached. The following data were extracted: first author's name, year of publication, total number of patients, clinicopathological parameters, and survival time. Two researchers independently evaluated the quality of eligible studies using the Newcastle-Ottawa scale [8].

Statistical analysis
All the statistical analyses were performed using the STATA software (version 11.0, StataCorp LP, College Station, TX, USA).The crude odds ratios (OR) and95% confidence intervals (CI) were used to estimate the strength of association between MTA1 and clinicopathological parameters. Risk ratios (RR) and 95% CIs were used to estimate the association of MTA1 status with the overall survival (OS) and disease-free survival (DFS). I 2 value, which indicated the percentage of total variation across studies, was used to assess statistical heterogeneity. Random-effects models (I 2 >50% or P<0.10) were used if significant heterogeneity was detected. Otherwise, fixed-effects models were used. Begg's rank correlation and Egger's weighted regression were used to determine potential publication bias. P value less than0.05 indicates statistically significant publication bias.

Correlation of MTA1 overexpression with OS and DFS
Survival time was extracted from Kaplan-Meier survival curves analyzed by the Enguage Digitizer software. In the present study, as shown in Table 4, MTA1 expression was not only clearly linked to OS, but also showed significant association with DFS. DTC patients with    Fig 4B) year DFS.

Subgroup analyses
In order to further investigate the relationship between MTA1 and prognosis of DTC, all the eligible studies were divided into several subgroups according to the quality of each study and tumor type (Table 5). High-quality studies were divided into high quality studies subgroup. And according to tumor type, we investigated MTA1 expression in patients with gastrointestinal cancers (GI cancers) or EC.

Sensitivity analysis and publication bias
In order to test the robustness of RR estimates in OS, sensitivity analysis was conducted by individually excluding studies and analyzing the effects of the remaining studies. Sensitivity analysis (S1 Fig) indicated that the RR estimates were relatively reliable and credible as no point estimate of the omitted study fell outside the 95% CI.
Begg's rank correlation and Egger's weighted regression methods were used to statistically assess publication bias. As shown in Fig 7A and 7B, neither Begg's (P = 0.35) nor Egger's (P = 0.13) test provided a clear evidence of publication bias. No publication bias was detected in the current study. The results reported in this article are credible.

Discussion
Depth of invasion (T), lymph node metastasis (N), and the presence of distant metastasis (M)-TNM stage was considered as the most important prognostic factors for DTC, such as GC [21]. However, large clinical case studies suggest that patients at similar pathological  [22][23]. MTA1 is a founding member of this family and was first identified as a metastasis-associated tumor gene differentially  expressed in rat metastatic tumors [24]. MTA1 overexpression has been identified in many cancers. However, the molecular functions of MTA1 were unclear until it was identified as an integral component of the NuRD complex [25]. Luo et al. [26] conducted a meta-analysis to further investigate the role of MTA1 in solid tumors, and confirmed that MTA1 expression was significantly associated with prognosis of solid cancers. Currently, the clinical and prognostic value of MTA1 in DTC is unknown. Ning et al. [27] reviewed the expression and clinical significance of MTA family, and concluded that MTA1 expression was correlated with invasion and lymph node metastasis in GI cancer. However, the prognostic value of MTA1 expression in DTC is unclear and controversial. Several studies found that MTA1-positive expression was not correlated with OS in patients with esophageal squamous cell carcinoma and breast cancer [10,28].
In our study, we investigated the overexpression of MTA1 and clinicopathological parameters in DTC. The results demonstrate that MTA1-positive expression increased the risk of stomach wall invasion (OR = 1.88, 95%CI: 1.05-3.37, P = 0.03), lymph node-positive metastasis (OR = 2.30, 95%CI: 1.76-3.01, P<0.001) and vascular invasion (OR = 2.02, 95%CI: 1.40-   was tightly associated with clinicopathological parameters and 1-, 3-, 5-year OS in GI cancer and EC. From a clinical perspective, MTA1 over-expression was strongly and independently correlated with depth of invasion, lymph node metastasis, vascular invasion and TNM stage. Tumor tissues expressing MTA1 show deeper invasion into the lymphatic network under the mucosa. Vascular invasion leads to advanced tumor stages, and shortens the OS of patients with DTC. Previous studies suggested that the MTA1 gene acted as a transcriptional regulator, in conjunction with other components of NURD to mediate transcriptional repression and the association of repressor molecules with chromatin [23,[29][30]. For example, MTA1 protein physically interacts with HDAC1 [31]. The two proteins are the key components of NuRD complex, which contains histone deacetylase. Histonedeacetylation alters chromatin structure and transcriptional control. Toh et al. [9] observed that MTA1 expression in ESCC was associated with the activity of H4 histone deacetylase. Tumor suppressor genes including p53, p21 and Bcl-2 are regulated by histone acetylation [32][33].  The limitations of this meta-analysis are as follows: (1) A few eligible non-English publications were excluded; (2) IHC assessments of MTA1 were still discordant; and (3) The number of articles was inadequate. Nonetheless, the meta-analysis has several advantages: (1) This study is the first of its kind to investigate the association between MTA1 overexpression and clinicopathological parameters in DTC; (2) The study successfully evaluated the association of MTA1 expression with the OS/DFS of DTC patients. This work was supported by grants from the National Natural Science Foundation of China (NO: 81602425) and the Natural Science Foundation of Anhui Province (NO: 1508085QH152,1608085MH163). The funders had no role in study design, data collection and analysis, manuscript preparation, or submission for publication.
In conclusion, MTA1 expression is significantly associated with clinicopathological parameters, DFS and OS in DTC patients. It may play an independent role in predicting aggressive tumor behavior and poor prognosis. The results of the meta-analysis suggest that MTA1 is a potential target for anticancer therapy. Further investigations are needed to identify the mechanisms underlying the role of MTA1.