ZHOU Wen, WANG Qing-song

1. Department of Histology and Embryology,Hainan Medical University,Haikou 571199,China

2. Department of Biochemistry and Molecular Biology,Hainan Medical University,Haikou 571199,China

Keywords:Claudin-10 Primary malignant tumor of digestive system Meta-analysis Clinicopathological characteristics Prognosis

ABSTRACT

1. Introduction

Claudins proteins are a family of transmembrane proteins involved in the formation of intercellular tight junctions and intercellular ion channels [1], maintenance of cell polarity [2], and regulation of epithelial-mesenchymal transition [3]. A total of 27 members of the claudins family have been found [4], all of which have 4 transmembrane helical domains, a short intracellular amino terminus and a carbonyl terminus, and 2 extracellular loops [5].Most claudin proteins have a PDZ domain at the carboxyl terminus,which is very important for the binding of claudin proteins to other backbone proteins [6]. Recent studies have found that the abnormal expression of claudin protein is closely related to the occurrence and development of malignant tumors [7, 8], and its expression is different in different tumor tissues, which can affect tumor cell behavior and prognosis [9].

Claudin-10 is a member of the claudins protein family, encoded by the CLDN10 gene on chromosome 13q32.1, with high expression in kidney, lung, pancreas, sweat gland, and salivary gland, and low expression in other tissues [10]. Recent studies have found that Claudin-10 protein is highly expressed in various primary malignant tumors such as liver cancer [11], lung cancer [12], and thyroid cancer [13], and is related to the prognosis of patients. Some studies have found that inhibiting the expression of claudin-10 can reduce the invasiveness of tumor cells such as liver cancer and melanoma [14, 15], and high expression of claudin-10 indicates poor prognosis [11]. Another study showed that in lung adenocarcinoma[12], clear cell renal cell carcinoma [16], and ovarian cancer [17], high expression of claudin-10 is beneficial to patient prognosis and is an independent predictor of patient prognosis. In primary malignancies of digestive system, the changes of claudin-10 protein expression and its relationship with the prognosis of patients are controversial.Therefore, this study conducted a systematic review and metaanalysis of the correlation between the expression of claudin-10 and the clinicopathological characteristics and prognosis of patients with primary digestive cancers, so as to provide a basis for future research.

2. Materials and methods

2.1 Inclusion and exclusion criteria

2.1.1 Type of study

Retrospective study.

2.1.2 Research Objects

The experimental group is the postoperative specimens of patients with confirmed primary malignant solid tumors of the digestive system and high expression of claudin-10; the control group is the specimens of patients with primary malignant solid tumors of the digestive system and low expression of claudin-10.

2.1.3 Exposure factors

High expression of claudin-10.

2.1.4 Outcome measures

Odds ratio (OR), hazard ratio (HR) and 95% confidence interval(CI).

2.1.5 Inclusion criteria

Studies are included if they meet all the following requirements:① Published journal articles; ② Chinese or English literature; ③Research subjects are patients with primary malignant tumors; ④Exposure factor is claudin-10 expression; ⑤ Provided correlation analysis of claudin-10 expression and clinicopathological characteristics; or survival analysis was performed, and survival analysis data between the two groups of claudin-10 high expression group and low expression group were provided, including HR and p value, or survival curve and P value.

2.1.6 Exclusion criteria

Studies are excluded if one of the following criteria is met:①Research without using patient tissue specimens, such as cell line experiments, animal experiments; ②Research using online public databases; ③The research which object is patients with nondigestive system tumors or metastatic tumors; ④ The study without information on preoperative treatment, or with the patients having received neoadjuvant therapy before surgery; ⑤ Literatures in languages other than Chinese and English; ⑥ Repeated publication or no data extraction; ⑦ Conference abstracts, reviews, and dissertations.

2.2 Literature search strategy and registration

This study was registered on the PROSPERO platform (number CRD42021295353). PubMed, Web of Science, Ovid MEDLINE,CNKI, CQVIP, WanFang Data databases were searched to collect studies on the correlation between claudin-10 expression and malignant solid tumors. The retrieval time limit is from the establishment of the database to November 30, 2021. References of included studies were manually searched to supplement access to relevant literature.

Chinese search terms include: tight junction protein claudin-10,CLDN10, malignant tumor, cancer; English search terms include:claudin-10, CLDN10, cancer, malignant tumor, malignant neoplasm,carcinoma.

2.3 Literature selection and data extraction

Two researchers independently screened literature, extracted data and cross-checked. If there is any disagreement, an agreement will be reached through discussion and consultation. In literature screening, the title was first read, and after excluding obviously irrelevant literature, the abstract and full text were further read to determine whether or not to include. Data extraction includes:①Basic information of included studies: research title, first author,publication journal, research location, research type, tumor type,sample size, follow-up time; ②Key elements of bias risk assessment;③Claudin-10 expression The detection method, evaluation method and cut-off point; ④Outcome indicators and outcome measures,including clinicopathological characteristics and odds ratio (OR)of claudin-10 expression, 95% confidence interval (confidence interval, CI) and p value, hazard ratio (HR), 95% CI and p value. If the literature did not report the HR value, but reported the Kaplan-Meier survival curve and the p value of the log-rank test, the survival rate at different time points was extracted from the survival curve using Engauge Digitizer version 4.1, and the HR and 95% CI were calculated [14].

2.4 Risk of bias assessment of included studies

The risk of bias of the included studies was independently assessed by 2 investigators and discrepancies were resolved by discussion. The risk of bias in included studies was assessed using the Newcastle-Ottawa Scale (NOS). The scale consists of three parts: the selection of research subjects (4 points), the comparability between the experimental group and the control group (2 points),and the measurement of outcome indicators (3 points). A total score of 0-5 was considered to be low-quality literature, and 6-9 was considered high-quality literature. Risk of bias assessment was done independently by 2 investigators, and if the scores were inconsistent,agreement was reached through discussion.

2.5 Statistical processing

Meta-analysis was performed using RevMan 5.3. HR and 95% CI were used to evaluate the effect of high claudin-10 expression on the prognosis of primary malignant tumors. The relationship between high claudin-10 expression and clinicopathological parameters of primary malignant tumors, including gender, age, TNM stage, T stage, tumor differentiation, tumor size, lymph node metastasis,and distant metastasis, was evaluated by OR and 95% CI. The χ2test was used to analyze the heterogeneity among the results of the included studies, and the I2value was used to judge the size of the heterogeneity among the studies. When I2<50%, P>0.10, there was no statistical heterogeneity among the studies, and a fixed-effect model was used for Meta-analysis; when I2≥50%, P≤0.10, the cause of heterogeneity was further analyzed, and a random-effect model was used for meta-analysis. Subgroup analysis and sensitivity analysis were used if there was significant clinical heterogeneity, or only descriptive analysis was performed. Publication bias analysis was performed using Stata 15.1.

3. Results

3.1 Literature search and selection

A total of 557 related literatures were obtained from the initial search. After screened through inclusion and exclusion criteria, 6 retrospective studies were included, including 512 cases of primary malignant tumors. The literature selection process and results are shown in Figure 1.

Figure 1 PRISMA flow chart for literature search and selection

3.2 Characteristics of included studies and risk of bias assessment results

The characteristics of the included studies and the results of the risk of bias assessment are shown in Table 1. The NOS score of the studies were ≥6 and the risks of bias were low (Figure 2).

3.3 Meta-analysis results

3.3.1 Differential expression of claudin-10 in primary digestive cancers, adjacent tissues and normal tissues

5 studies reported the expression of claudin-10 in patients with primary digestive cancers and controls. Tumor types include esophageal cancer [20, 22], colorectal cancer [19, 21], and hepatocellular carcinoma [11]. Among them, there were 411 tumor patients and 256 control groups.

Due to the heterogeneity in the control group, we performed a subgroup analysis, as shown in Figure 3. 1 study used paracancerous tissue as a control, and the results showed that there was astatistically significant difference in the expression of claudin-10 in the malignant tumor tissue and the paracancerous tissue [OR=3.18,95%CI (1.91,5.27), P<0.01]. 4 studies used normal tissues from nontumor patients as controls, and there was no statistical heterogeneity among the studies (I2=49%, P=0.12). Using a fixed-effect model, the results showed that the high expression rate of claudin-10 in primary malignant tumor tissues was significantly different from that in normal tissues [OR=3.27, 95%CI (1.98, 5.39), P <0.01]. There was no statistical heterogeneity among the 5 studies (I2=32%, P=0.21).Using a fixed-effect model, the results showed that, compared with the control group, claudin-10 was highly expressed in primary digestive system malignant tumor tissues with statistically significant difference [OR=3.22, 95%CI (2.26, 4.60), P<0.01].

Table 1 Characteristics of 6 studies included in metaanalysis

Figure 2 Risk of bias table

Sensitivity analysis by excluding individual studies one by one found that the results did not change significantly, which indicated that the study results were reliable.

3.3.2 Correlation between claudin-10 expression and gender in patients with primary digestive cancers

4 studies reported the expression of claudin-10 in patients of different genders with primary digestive malignancies, including colorectal cancer [19, 21] and esophageal cancer [20, 22]. There were 205 male patients and 107 female patients. There was no statistical heterogeneity among the studies (I2=0%, P=0.81), and a fixed-effects model was used for meta-analysis. The results showed that there was no statistical difference in the expression rate of claudin-10 in patients of different genders [OR=1.62, 95%CI (0.91, 2.90), P=0.10],see Figure 4A. Sensitivity analysis was performed by excluding studies one by one, and the results did not change significantly,indicating that the study results were reliable.

3.3.3 Correlation between claudin-10 expression and age in patients with primary digestive cancers

4 studies reported the association of claudin-10 with age in patients with primary digestive malignancies, including colorectal cancer [19,21] and esophageal cancer [20, 22]. However, the cut-off values for age groups were not consistent, so a subgroup analysis was used. There was no statistical heterogeneity between studies (I2=0%, P=0.42),and a fixed-effects model was used. Meta-analysis results showed that claudin-10 expression was not associated with patient age[OR=1.72, 95%CI (0.97, 3.07), P=0.06], see Figure 4B. Sensitivity analysis showed that the findings were reliable.

3.3.4 Correlation between claudin-10 expression and the size of primary malignancies of digestive system

Figure 3 Meta-analysis results for the correlation between claudin-10 expression and primary digestive cancers

Figure 4 Meta-analysis results for the correlation between claudin-10 expression and gender or age of patients

3 groups of studies reported the correlation between claudin-10 and the size of malignant tumor. Huang [11] studied hepatocellular carcinoma and the patients were divided into groups ≤5cm and>5cm; Jiang [19] and Pu [21] studied colorectal cancer and the patients were divided into groups <5cm and ≥5cm. The cutoff values of tumor size in the three groups of studies were not completely consistent, and subgroup analysis was used. Subgroup analysis showed that in patients with primary colorectal cancer, the high expression rate of claudin-10 in the tumor size ≥5cm group was higher than that in the <5cm group, and the difference was statistically significant [OR=6.21, 95%CI (2.44, 15.83), P<0.01], see Figure 5A.

There was significant statistical heterogeneity among the three groups of studies (I2= 84%, P< 0.01), and a random-effects model was used. Meta-analysis results showed that there was no significant difference in the expression of claudin-10 in tumors of different sizes [OR=2.70, 95%CI (0.50, 14.59), P=0.25]. Sensitivity analysis showed that after excluding the study of Huang [11], there was no statistical heterogeneity between the other two groups of studies (I2=0%, P=0.86). Using a fixed-effect model, the results showed that the expression of claudin-10 was significantly correlated with tumor size [OR=6.21, 95%CI (2.44, 15.83), P<0.01].

3.3.5 Correlation between the expression of claudin-10 and the differentiation degree of primary digestive cancers

5 studies analyzed the relationship between claudin-10 and the degree of differentiation of primary digestive system malignancies,including esophageal cancer [20, 22], hepatocellular carcinoma [11],and colorectal cancer [19, 21]. A total of 411 tumor patients were included, of which 129 were in the well-differentiated group and 282 in the moderately or poorly differentiated group. There was no statistical heterogeneity among the studies (I2=0%, P=0.63),and a fixed-effect model was used. The results of meta-analysis showed that the high expression rate of claudin-10 in the moderately and poorly differentiated group was lower than that in the welldifferentiated group, and the difference was statistically significant[OR=0.56, 95%CI (0.32, 0.99), P=0.05], see Figure 5B. The results suggest that the expression of claudin-10 is positively correlated with the differentiation degree of primary malignant tumors of digestive system. Sensitivity analysis showed that the results of the study were reliable.

3.3.6 Correlation between claudin-10 expression and lymph node metastasis of primary digestive cancers

Figure 5 Meta-analysis results for the association between claudin-10 expression, clinicopathological features and prognosis

Three studies analyzed the association of claudin-10 with lymph node metastasis in primary digestive malignancies, including colorectal cancer [19, 21] and esophageal cancer [22]. A total of 250 tumor patients were included, including 97 patients with lymph node metastasis (LN+) and 153 patients without lymph node metastasis (LN-). There was statistical heterogeneity among the studies (I2=62%, P=0.07), and a random-effect model was used.Meta-analysis results showed that in primary malignant tumors of the digestive system, the high expression rate of claudin-10 in the group with lymph node metastasis was higher than that in the group without lymph node metastasis, and the difference was statistically significant [OR=3.41, 95%CI (1.09, 10.65), P=0.03], see Figure 5C.Sensitivity analysis showed that the findings were reliable.

3.3.7 Correlation between claudin-10 expression and distant metastasis of primary digestive cancers

2 studies analyzed the relationship between claudin-10 and distant metastasis of primary digestive cancers, both of which were colorectal cancer [19, 21]. A total of 110 tumor patients were included,of which 9 cases have distant metastasis (M1) and 101 cases have no distant metastasis group (M0). There was no statistical heterogeneity among the studies (I2=0%, P=0.92), and a fixed-effects model was used. Meta-analysis results showed that there was no significant difference in the high expression rate of claudin-10 between the group with distant metastasis and the group without distant metastasis [OR=4.51, 95%CI (0.55, 36.70), P=0.16], see Figure 5D.The results suggest that the expression of claudin-10 is not related to the distant metastasis of primary colorectal cancer. Sensitivity analysis showed that the results of the study were reliable.

3.3.8 Correlation between claudin-10 expression and T stage and clinical stage of primary digestive cancers

Only 1 study [22] analyzed the relationship between claudin-10 and T stage and clinical stage of primary malignancies of digestive system. A total of 140 patients with primary esophageal cancer were included, of which 62 in T1-2 stage, 78 cases in stage T3-4, 52 cases in stage I-II, and 88 cases in stage III-IV. The results showed that the expression of claudin-10 was not related to T stage of primary esophageal cancer [OR=1.26, 95%CI (0.58, 2.71), P=0.56], nor was it related to clinical stage [OR=1.38, 95%CI (0.63, 3.01), P=0.42].Since there was only 1 study, no meta-analysis was performed.

3.3.9 Correlation between claudin-10 expression and prognosis of primary digestive cancers

3 studies analyzed the relationship between claudin-10 and the prognosis of primary digestive malignancies, including liver cancer[11, 18] and colorectal cancer [19]. There was significant statistical heterogeneity between studies (I2=97%, P<0.01), and a randomeffect model was used for meta-analysis. Meta-analysis results showed that there was no significant correlation between the prognosis of malignant solid tumors and the expression of claudin-10[HR=1.99, 95%CI (0.93, 4.23), P=0.08], see Figure 5E. Sensitivity analysis showed that the results of the study were reliable.

3.4 Publication bias analysis

Egger's test results showed that, with the test level of 0.05, there was no obvious publication bias in correlation study of the expression difference of claudin-10 in the primary digestive cancer and the control, as well as the correlation of claudin-10 expression with age,tumor size, degree of differentiation, lymph node metastasis and prognosis in the other correlation study. There was publication bias in the correlation study between claudin-10 expression and gender(P<0.05), see Table 2. The number of studies involving association of Claudin-10 with distant metastasis, T stage, and clinical stage was too small to be tested by Egger's test. The number of literatures included in this study was less than 10, so no funnel plot was drawn.

Table 2 Publication bias analysis(Egger’stest)

4. Discussion

This study included 6 literatures, 512 cases of primary malignancies of digestive system including hepatocellular carcinoma, esophageal cancer and colorectal cancer. Among them, the correlation between claudin-10 and tumor differentiation was investigated in 5 studies[11, 19-22], with a total of 411 tumor patients, and there was no statistical heterogeneity among the studies; the correlation between claudin-10 and lymph node metastasis was observed in 3 studies including colorectal cancer [19, 21] and esophageal cancer [22], with a total of 250 cases. There was a certain statistical heterogeneity between studies (P=0.07), and the heterogeneity may be related to the different stages of tumor patients in different studies. Previous studies have shown that claudin-10 is expressed at a low level in normal tissues of the digestive system [10]. This study showed that,compared with normal tissues, claudin-10 was highly expressed in primary digestive cancers, and claudin-10 expression was positively correlated with tumor size, differentiation degree and lymph node metastasis. Therefore, claudin-10 may be a potential marker for primary digestive system malignancies.

Only 1 study [22] analyzed the relationship between claudin-10 and the T stage and clinical stage of primary malignancies of digestive system (esophageal cancer), and found that claudin-10 expression was not associated with T stage, nor is related to clinical stage. 2 studies [19, 21] including 110 cases analyzed the relationship between claudin-10 and distant metastasis of primary colorectal cancer. There was no statistical heterogeneity among the studies. Meta-analysis results showed that claudin-10 expression is not related to distant metastasis of primary colorectal cancer. However, due to the small number of included studies and the small sample size, the correlation between claudin-10 expression and T stage, distant metastasis and clinical stage of primary digestive malignancies is still unclear, and further research is needed.

This study also analyzed the relationship between claudin-10 and the prognosis of primary digestive cancers. There are still few studies on the relationship between claudin-10 and the prognosis of primary digestive system malignant tumors. A total of 3 literatures were included, involving 250 tumor patients. The results showed that the high expression of claudin-10 was not associated with the prognosis of patients, and there was great heterogeneity among the studies. The reasons for the large heterogeneity are: (1) The detection methods of claudin-10 are different. Cheung[18] used RT-PCR, Jiang[19] and Huang[11] used immunohistochemistry. There are differences in the type and source of antibodies, and the standard for setting the critical value of claudin-10 expression is not uniform; (2) The outcome indicators of prognosis are not uniform, Cheung[18] used DFS,Jiang[19] and Huang[11] used OS; (3)The included studies involved different tumor types, including liver cancer [11, 18] and colorectal cancer [19], and the aggressiveness of different types of tumors is different. Due to the large heterogeneity of the included literature and the small sample size, the correlation between claudin-10 and prognosis is still unclear.

The specific mechanism of Claudin-10 in the occurrence and development of digestive tumors is still unclear. Studies have found [14] that claudin-10 promotes the survival, migration and invasion of liver cancer cells, and increases the expression of matrix metalloproteinase (MMP). Li [17] used gene enrichment analysis and found that TGF-β and Wnt signaling pathways that regulate epithelial to mesenchymal transformation (EMT) are closely related to claudin-10 expression. Claudins have been found to be closely related to the occurrence and development of digestive malignancies[8, 23]. For example, Claudin-1 is highly expressed in primary colon cancer [24], and in vitro study show that claudin-1 inhibits the expression of E-cadherin by up-regulating the transcription factor ZEB-1 (Zinc Finger E-box binding homeobox-1), thereby promoting the invasion and metastasis of colon cancer [25]. Knockout of claudin-2 resulted in a significant decrease in the invasion and metastasis of colorectal cancer cells HT-29, and the PDZ-binding domain played an important role in the promotion of colorectal cancer invasion and metastasis by claudin-2 [26]. In contrast to claudin-1 and claudin-10, claudin-7 is highly expressed in the normal colon, but significantly decreased in patients with primary colorectal cancer, and claudin-7 can promote mesenchymal to epithelial transformation (MET) to inhibit tumor growth [27]. The claudin-18.2 antibody treatment of gastric cancer has entered the clinical trial stage [28].

This study has certain limitations: (1) There are limited studies related to claudin-10 and the clinicopathology and prognosis of malignant tumors, and the sample size that can be included is only 512 cases, and the small sample size may lead to sampling errors; (2) In the correlation analysis of claudin-10 with prognosis,the heterogeneity of the included literature was large, mainly due to differences in claudin-10 detection methods and inconsistent prognostic outcome indicators; (3) All the included studies were in China; (4) Except for Cheung's [18] study, which is a prospective study, the others are all retrospective studies, so selection bias is inevitable. These factors may have influenced the results of this study.

In conclusion, this study show that claudin-10 is highly expressed in primary digestive cancers, and its expression is positively correlated with tumor size, pathological differentiation and lymph node metastasis, and may be a potential tumor marker.

Author contributions

Zhou Wen and Wang Qing-song were jointly responsible for the study design, literature search, and data extraction. Zhou Wen analyzed data and wrote the manuscript. Wang Qingsong provided guidance for topic selection.

Conflicts of interest

All authors have no conflicts of interest and disputes.