The prognostic value of over-expressed TrkB in solid tumors: a systematic review and meta-analysis

It is reported recently Tropomyosin-related receptor Kinase B (TrkB) plays key roles in the anoikis resistance during the processes of tumorigenesis and metastasis. However, its prognostic significance for clinical patients remains inconclusive. In order to establish a correct and practicable link between increased TrkB and prognostication of human solid tumors, a meta-analysis was performed in this article. A systematic literature research in the electronic databases PubMed, Embase and Web of Science was performed to identify eligible studies. A fixed-effects meta-analytical model was employed to correlate TrkB expression with OS, DFS and clinicopathological features. A total of 11 studies covering 1516 patients with various solid tumors were recruited in this meta-analysis. TrkB over-expression was associated with poorer OS and poorer DFS in multivariate analysis. Additionally, the pooled odds ratios (ORs) indicated that TrkB over-expression was associated with large tumor size, lymph node metastasis, distant metastasis and a higher clinical stage. Overall, these results indicated that TrkB over-expression in patients with solid tumors might be related to poor prognosis and serve as a potential predictive marker of poor clinicopathological prognosis factor.


INTRODUCTION
Anoikis is a form of programmed cell death that occurs in anchorage-dependent cells when they detach from the surrounding extracellular matrix [1][2][3]. Anoikis acts as an important defense for the organism by preventing detached cells' re-adhesion to new matrices in incorrect locations and their dysplastic growth [4]. Failure to execute the anoikis program and resistance to anoikis evolve as a hallmark of metastatic cancers which enables cancer cells to disseminate to distant organs through systemic circulation [5][6][7].
Tropomyosin-related receptor Kinase B (TrkB) is a tyrosine kinase receptor for brain-derived neutrophic factor (BDNF), which triggers several intracellular signals [8][9][10]. TrkB was originally defined as a specific suppressor of caspase-associated anoikis of nonmalignant epithelial cells. High expression of TrkB promoted tumorigenesis and metastasis. The vast bulk of the evidence from studies on solid tumor cells of liver, lung, breast and ovary indicated that anoikis suppression induced by over-expressed TrkB best owed enhanced metastatic capacity on various tumor cells [11][12][13][14]. Therefore, TrkB is expected to be a potential drug target [15,16]. the correlation between the expression of TrkB and patient survival remains inconclusive. Therefore, it is necessary to analyze the data of TrkB systematically in human solid tumors to draw a reasonable conclusion about its prognostic significance.

Research Paper www.impactjournals.com/oncotarget
In this study, we conducted a meta-analysis to investigate the significance of over-expressed TrkB in the prediction of prognosis of corresponding patients. The results showed that increased TrkB expression in patients with solid tumors related to poor prognosis and served as a potential prediction of poor clinicopathological prognosis factor.

Study search information
The initial search identified 315 publications, of which, 18 studies were of acceptable relevance. However, four of these studies were excluded because the survival curve was based on NTRK2 expression, and three were excluded because of the absence of information about clinicopathological characteristics. Ultimately, 11 studies met the eligibility criteria and were included in the current meta-analysis ( Figure 1).

Correlations of TrkB expression and OS
The pooled hazard ratio (HR) revealed that overexpressed TrkB was significantly associated with poor overall survival (OS) for cancer victims in multivariate analysis (HR: 1.76, 95% CI: 1.48-2.10; Figure 2). However, no significant heterogeneity (I 2 = 31.6%, P = 0.147) was observed when using a fixed-effects model to analyze the pooled HR of the OSs.

Correlations of TrkB expression and DFS
A significant correlation between over-expressed TrkB and disease-free survival (DFS) was also observed in the patients with solid tumors in multivariate analysis (HR: 2.20, 95% CI: 1.63-2.96; Figure 3) in the fixedeffects model without significant heterogeneity (I 2 = 0.0%, P = 0.419).

Correlations of TrkB expression and clinicopathological parameters
The clinical and pathological parameters collected from the eligible studies were presented in Table 2. Meanwhile, pooled results of the correlations were identified between over-expressed TrkB and clinicopathological features of patients with solid tumors.      Table 3).

Publication bias
We constructed funnel plots and performed Begg's test to assess publication bias. As a result, the shape of the funnel plot for the OS, DFS and clinicopathological parameters seemed symmetrical in the multivariate analysis method (Figure 4, Figure 5 and Supplementary Figure 1).

DISCUSSION
TrkB is a tyrosine kinase receptor for brainderived neutrophic factor (BDNF), which triggers several intracellular signals. TrkB is over-expressed in various human malignancies and growing evidence demonstrated its association with tumour cell proliferation, invasion and metastasis [28][29][30][31]. TrkB is also an anoikis resistance related gene, whose over-expression may enhance a resistance of detachment-induced apoptosis. Consequently, knockdown of TrkB provided inhibition of growth or invasion and decrease of apoptosis in oral squamous cell carcinoma (OSCC) and endometrial carcinoma cells [25,32]. However, thus far, no meta-analyses have been performed to evaluate the prognostic value of TrkB in cancer victims. This meta-analysis systematically estimates the association between TrkB expression and prognostic value of solid tumors.
In this meta-analysis, we evaluated survival data from 1516 solid tumor patients from 11 eligible studies, which met the inclusion criteria, were organized according to OS and DFS. For all studies, TrkB expression was detected by IHC. By meta-analysis of the 11 studies, we identified the pool HRs which indicated that TrkB was a factor in poor prognosis in various cancers. Because there is no significant heterogeneity among our included studies, so we did not perform further subgroup analyses.  Additionally, no publication bias was observed. These results indicated over-expressed TrkB was positive related to poor OS and DFS in solid tumor patients. In addition, pooled results of the correlations were identified between over-expressed TrkB and clinicopathological features of patients with solid tumors. The results showed the expression of TrkB was positively associated with large tumor size, lymph node metastasis, distant metastasis and a higher clinical stage. We can explain this result by TrkB's ability to enhance anoikis resistance, promote tumour cell proliferation, invasion and metastasis. Because of its involvement in these processes, TrkB is likely to be causally involved in tumor progression and, consequently, increased levels of TrkB would be expected to indicate a poor prognosis. This meta-analysis was properly conducted, however, further analysis with several limitations would be considered in the future. first, need more trials to analysis; second, some of the survival data were extracted from Kaplan-Meier curves and might be less reliable than a direct analysis of variance; third, we need to search more non-English publications. In addition, the possible existence of unpublished studies could also result in potential publication bias. In general, concerning these limitations mentioned above, we must interpret these results with adequate caution.
In conclusion, our meta-analysis indicated that overexpressed TrkB was positive related to poor OS and DFS in solid tumor patients. Over-expressed TrkB could be served as a potential marker for poor clinicopathological prognostic factors in patients with various solid tumors. However, additional studies related to specific tumor types are necessary to illuminate the clinical utility of increased TrkB in solid tumors.

Study inclusion and exclusion criteria
The collected studies included in this meta-analysis had to meet the following criteria: (1) a pathological diagnosis of cancer was made; (2) TrkB expression in patients with any type of tumor was measured via immunohistochemistry; (3) associations of TrkB expression with OS, DFS or clinicopathological features were described; (4) HRs and 95% confidence intervals (CIs) were reported or could be calculated (based on the information in the paper); and (5) when the same author reported repeated results from the same population, the most complete report was included. The exclusion criteria for this meta-analysis were as follows: (1) unpublished papers; (2) laboratory articles, reviews and letters; (3) non-English language articles; (4) overlapping articles or ones with duplicate data; (5) articles with only animal experiments; (6) studies without information about survival curves; and (7) TrkB expression in patients with any type of tumor was analyzed only using RT-PCR method.

Data extraction and quality assessment
All data were extracted independently by two investigators (Chunze Zhang and Dan Gao). For each eligible study, the following characteristics were extracted: first author's name, publication year, region, type of cancer, number of patients, patients' ages, follow-up times, detection methods, staining position, cut-off values, survival data (including OS and DFS) and clinicopathological parameters, such as tumor differentiation, tumor size, lymph node metastasis and TNM stage. For studies that presented only Kaplan-Meier curves was used to extract the survival data. The Newcastle-Ottawa-Scale (NOS) was adopted to assess the study quality of each individual study. The NOS score ranged from 0 to 9, and studies with NOS score ≥ 7 were defined as high-quality studies.

Statistical analysis
This meta-analysis was performed using Stata 12.0 (Stata Corporation, College Station, TX, USA) software. Pooled estimates of HRs and their 95% CIs were used to estimate the association between TrkB expression and patients' survival. The chisquared test (Cochrane' s Q test) and I-squared statistical test were used to analyze the heterogeneity between studies. When the result of a Q-test (I 2 > 50% or P < 0.05) indicated heterogeneity, the randomeffects model was used for the meta-analysis. Otherwise, a fixed-effects model was used. HR with its value over 1.0 indicated poor prognosis patients with increased TrkB expression. Funnel plots were used to graphically represent the publication bias. Begg's (rank correlation) test was adopted to confirm the publication bias.