Prognostic role of c‐Jun activation domain‐binding protein‐1 in cancer: A systematic review and meta‐analysis

Abstract c‐Jun activation domain‐binding protein‐1 (Jab1) is aberrantly overexpressed in multiple cancers and plays an oncogenic role in cancer progression. We examined the association between Jab1 expression and prognosis in patients with cancer by conducting a meta‐analysis. A comprehensive search strategy was performed using the PubMed, Web of Science, Ovid and EMBASE in July 2020. Eligible studies were enrolled according to definite criteria. Twenty‐seven studies involving 2609 patients were enrolled in this meta‐analysis. A significant association between high Jab1 expression and poor overall survival (pooled hazard ratio [HR] 2.344, 95% confidence interval [CI]: 2.037‐2.696) was observed. Subgroup analyses of the type of cancer, sample size, follow‐up period, Jab1 detection method and preoperative treatment did not alter the significance. On pooling data from Cox multivariate analyses, high Jab1 expression was found to be an independent prognostic indicator for overall survival. In addition, high Jab1 expression was found to be associated with advanced clinicopathological features such as clinical stage, lymphatic metastasis, histological grade and distant metastasis in cancers. Our meta‐analysis is the first to demonstrate that high Jab1 expression may be a promising indicator of poor prognosis and has an independent prognostic value for overall survival in patients with cancer.

that dysregulation of Jab1/COPS5 contributes to tumorigenesis by functionally interacting with several tumour-related proteins, such as the cyclin-dependent kinase inhibitor 1B (p27) and 1C (p57), p53, SMAD4/7 and programmed death-ligand 1 (PD-L1). [4][5][6][7][8][9][10] Based on data from the Gene Expression Profiling Interactive Analysis database (http://gepia.cance r-pku.cn/index.html) and other related articles, Jab1/COPS5 overexpression has been reported in various cancers, including hepatocellular carcinoma (HCC), breast cancer, non-small cell lung cancer and nasopharyngeal carcinoma [11][12][13][14][15][16] In addition, an increasing number of studies have revealed that Jab1 overexpression is associated with poor prognosis and clinicopathological characteristics in a variety of cancers, such as HCC, breast cancer and colorectal cancer. 12,14,17 Hence, we expect Jab1 to act as a potential prognostic biomarker in cancer. Although numerous studies have assessed the prognostic value of Jab1 in multiple cancers, there are controversial results regarding overall survival (OS) or some clinicopathological characteristics. Thus, we performed a systematic review and meta-analysis to evaluate the prognostic value of Jab1 expression in multiple cancers.

| Search strategy
In the present study, up-to-date databases including PubMed, Web of Science, Ovid and EMBASE were used for the selection of publications. We performed a publication search using the following keywords: 'c-Jun activation domain-binding protein-1 or Jab1 or constitutive photomorphogenic-9 signalosome or CSN5 or COP9 Signalosome Subunit 5 or COPS5' and 'cancer or carcinoma or sarcoma or tumour or neoplasia or malignancy' and 'prognosis or outcome or survival or mortality or "hazard ratio" or HR'. The latest literature search was conducted on 15 July 2020.

| Study selection criteria
The inclusion criteria for this study were as the following:  incomplete study would be excluded. Study selection was performed by two authors independently, and the divergence was solved by discussion in the group.

| Quality assessment
Study quality was measured using the Newcastle-Ottawa quality assessment scale (NOS). This scale contains a maximum score of 9 points for the assessment of each included study with 4 points in the selection, quality, 2 points on the comparability and 3 points on the quality of outcome and follow-up. A study with a NOS score lower than 6 was considered to be of inferior quality and was excluded from this metaanalysis. Study quality assessment was performed by two authors independently, and the divergence was solved by discussion in the group.

| Data extraction
For each study, necessary information was extracted from the cor-

| Statistical analysis
In this meta-analysis, we used Stata 14.0 (STATA Corporation, College Station, TX, USA) to analyse the data. HRs with their 95% CIs were used to assess the association between Jab1 expression and patient prognosis. An HR > 1 indicated a poor prognosis in patients. For the pooled results, we applied the chi-square-based Q test and Higgins I 2 statistic to evaluate heterogeneity. A P value lower than 0.05 and I 2 value larger than 50% implied that there was significant heterogeneity. The fixed-effects model was used when no significant heterogeneity was observed; otherwise, a random-effect model was applied.
Publication bias was evaluated using Egger's and Begg's tests. When significant publication bias was observed, 'trim and fill' analysis was used to detect and adjust for publication bias in the meta-analysis results. 19,20 Besides, in the process of data extraction from Kaplan-Meier curve, the Engauge Digitizer (Version 4.0, http://engau ge-digit izer.softw are.infor mer.com) was used.

| Data selection and characteristics of studies
We used the designed search strategy to collect 498 studies. After excluding duplicates, 152 articles remained. Based on the inclusion and exclusion criteria described above, 27 studies were finally enrolled. The detailed process of study selection is shown in Figure 1.
In Table 1, the characteristics of each enrolled study are presented. In

| Association of Jab1 expression with prognosis
As shown in Figure

| Independent prognostic value of Jab1 expression in OS
Among the 27 included studies, a total of 14 studies used a Cox proportional hazard model for multivariate analysis, which enrolled 13 types of cancer with 1356 patients. As shown in Figure 4, we observed a significant association between Jab1 overexpression and poor OS in patients with multiple cancers (pooled HR 2.190, 95%CI: 1.853-2.587). A fixed-effects model was used because low heterogeneity was detected among studies (χ 2 = 8.76, df = 13, P = 0.791; . This result indicated that Jab1 overexpression may act as an independent prognostic factor for OS in human cancers.

| Association between Jab1 expression and clinicopathological characteristics of cancers
In the present meta-analysis, clinicopathological characteristics including clinical staging, size or invasion extent of the primary tumour, lymphatic metastasis, histological grade and distant metastasis were analysed for the correlation with Jab1 expression ( Figure 5 and

| Sensitivity analysis and publication bias
Sensitivity analysis was performed to further assess the stability of the obtained results, although no significant heterogeneity was observed in the analysis of the association between Jab1 expression and OS. As shown in Figure 6A, exclusion of any individual study did not alter the significance of the association between Jab1 expression and OS. Similarly, sensitivity analysis was applied to studies that performed Cox multivariate analysis. Excluding any individual study did not influence the significance of the independent prognostic value of Jab1 expression in OS ( Figure 6B).
For publication bias, Begg's funnel plot and Egger's linear regression tests were performed and both showed significant evidence of publication bias among our pooled results. The analysis of the association between Jab1 expression and OS revealed P < 0.001 in Begg's test and P < 0.001 in Egger's test, as shown in Figure 7A. The analysis of the independent prognostic value of Jab1 expression in OS revealed P = 0.001 in Begg's test and P < 0.001 in Egger's test ( Figure 7B). Thus, for the analysis of the association between Jab1 expression and OS, we carried out nonparametric 'trim-and-fill' analysis as shown in Figure 7C, which detected nine trimmed studies, and the estimated pooled HR and 95% CI were 2.114 (1.852-2.413) after adjustment. For the analysis of the independent prognostic value of Jab1 expression in OS, we found that five studies were trimmed and the estimated pooled HR with its 95% CI was 2.046 (1.750-2.392) after adjustment ( Figure 7D).
Sensitivity analysis and assessment of publication bias were also conducted to analyse the association between Jab1 expression and clinicopathological characteristics of cancers. For the results of sensitivity analysis, exclusion of any individual study did not alter the significance of the association between Jab1 expression and each analysed clinicopathological characteristic, except for distant metastasis ( Figure S1). For the results of publication bias assessment, neither Begg's funnel plot nor Egger's linear regression test found significant publication bias in any analysis of each clinicopathological characteristic ( Figure S2).

| D ISCUSS I ON
As one of the major public health problems worldwide, cancer causes a large number of deaths and creates a significant economic burden every TA B L E 2 Subgroup analysis of association between overall survival and high Jab1 expression   The stability of these results was verified by the sensitivity analysis.
No alteration of the pooled results was observed by excluding any individual study. However, significant publication bias was found, thus nonparametric 'trim-and-fill' analysis was used for detection and estimation. The estimated pooled HRs and 95% CIs showed no significant alteration in these results after the trim-and-fill adjustment.
Therefore, our results provide robust evidence that high Jab1 expression is significantly associated with poor OS and may act as an independent prognostic indicator of OS in patients with cancer.
In addition, we collected other survival data, such as RFS, DFS and PFS, from five studies and found that high Jab1 expression is associated with poor RFS, DFS and PFS in patients with colorectal cancer, non-small cell lung cancer and osteosarcoma respectively, although Jab1 expression was reported to not be correlated with It is noteworthy that there were several limitations to our study.
First, only articles written in English were included, most of which were from China and Japan. Data corresponding to non-Asian patients are limited. Second, overestimation of the prognostic value of Jab1 was inevitable because a small number of publications reported negative results. Third, the HRs and 95% CIs in several studies were only obtained by extraction and calculation. Fourth, the cut-off values of high and low expression of Jab1 were diverse among studies.
Hence, more well-designed, large-scale studies are warranted to confirm our results.
In summary, this meta-analysis is the first to demonstrate that Jab1 is a promising indicator of prognosis in human cancers. Our study found that high Jab1 expression is associated with poor OS and advanced clinicopathological characteristics and possesses an independent prognostic value for OS in patients with cancer. In the future, more further studies of Jab1 in cancers should be conducted to investigate the underlying biological mechanisms, novel therapeutic targets and clinical translational applications.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest. Writing-review & editing (lead).

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.