Clinicopathologic and prognostic implications of Golgi Phosphoprotein 3 in colorectal cancer: A meta-analysis

Background Golgi Phosphoprotein 3 (GOLPH3) has been implicated in the development of colorectal cancer (CRC). Nevertheless, the clinicopathological and prognostic roles of GOLPH3 in CRC remain undefined. We thus did a meta-analysis to assess GOLPH3 association with the clinicopathological characteristics of patients and evaluate the prognostic significance of GOLPH3 in CRC. Methods An electronic search for relevant articles was conducted in the PubMed, Cochrane Library, Web of Science, Medline, Embase, CNKI, and WanFang databases. Two independent reviewers searched all the literature and finished the data extraction and quality assessment. Odds ratio (OR) or hazard ratio (HR) with 95% confidence interval (CI) were used to assess estimates. Stata software (version12.0) was employed to analyze the data. Results A total of 8 published studies were eligible (N = 723 participants). Meta-analysis revealed that GOLPH3 was found to be highly expressed in tumor tissues compared to that of adjacent colorectal tissues (OR, 2.63), and overexpression of GOLPH3 had significant relationship with advanced clinical stage (OR, 3.42). GOLPH3 expression was not correlated with gender (OR, 0.89), age (OR, 0.95), positive lymphatic metastasis (OR, 1.27), tumor size (OR, 1.12), poor differentiation of tumor (OR, 0.56) or T stage (OR, 0.70). Moreover, GOLPH3 overexpression was not associated with worse overall survival (OS) (HR = 1.14, 95% CI: 0.42–1.86, P>0.05) and disease-free survival (DFS) (HR = 0.80, 95% CI:-0.26–1.86, P>0.05). Conclusions GOLPH3 overexpression is correlated with tumor stage, which is an adverse clinicopathological characteristic of CRC. But, GOLPH3 can not serve as a useful biomarker in evaluating the progression of CRC.


Introduction
Colorectal cancer (CRC) is the third leading malignancy worldwide and the second most deadly cancer [1] Patients suffering from CRC in the early stages (I-III) usually have a long survival. Once patients develop malignant metastasis, the 5-year survival rate is less than 13% [2]. Much of the morbidity and mortality of CRC is due to patients being diagnosed in the late stages, where therapeutic intervention is less effective [3]. Additionally, certain biomarkers are important in tumor progression. In this regard, numerous studies have been focused on finding predictive biomarkers for CRC with a special interest in evaluating prognostic significance [4,5].
GOLPH3, also called GPP34, is located in chromosome 15 p13 [6]. GOLPH3 protein is mainly enriched in the outer membrane on the opposite side of the Golgi vesicle, which also exists in the tubules, vesicles, endohedral chamber and membrane [7]. GOLPH3 is an oncogene that is known to be upregulated in breast cancer, esophageal cancer, glioma, prostate cancer and other cancers [8][9][10][11][12]. GOLPH3 can regulate the function of Golgi protein glycosyltransferase, leading to the abnormal secretion of glycosylated protein, which can affect the growth, adhesion, migration, invasion and immunity of tumor cells [13]. Researches have indexed that GOLPH3 is very relevant to the occurrence, development and prognosis of CRC and may serve as a representative molecular marker [14,15]. The relationship between the expression of GOLPH3 and the clinicopathological features and prognosis of CRC remains inconsistent and controversial. Therefore, a meta-analysis was conducted to determine the authenticity of GOLPH3 in regard to the clinicopathological features and prognosis of CRC, providing further evidence for clinical practice.

Search strategy
Manual retrieval of PubMed, Cochrane Library, Web of Science, Medline, Embase, CNKI and Wanfang Database (last search updated to December 31, 2020). The following literature were searched from Medical Subject Headings (MeSH) with the following keywords:("Colorectal Neoplasm" or "Colorectal Cancer" or "Colorectal Tumor" or "CRC") and ("Golgi phosphoprotein 3" or "GOLPH3" or "GPP34 protein"). We applied no language restrictions. In addition, references in other related articles were also scanned to find eligible studies.

Selection criteria
Inclusion criteria are as follows: (1) immunohistochemical detection of GOLPH3 expression; (2) histological diagnosis of the colorectal tumor; (3) Patient clinical information is complete; (4) Language limit: Chinese or English. Exclusion criteria were:(1) Repeat published; (2) The types of studies do not match (editorials, letters, and so on); (3) The patient's clinical data information is incomplete; (4) studies not conducted in humans.

Data extraction and quality assessment
All articles are screened independently by two independent reviewers to ensure science and fairness. The extracted information included the author, year of publication, country, number of patients, age, pathologic differentiation degree, gender, lymphatic metastasis, tumor stage, tumor size, T stage, hazard ratio (HR) and corresponding 95% CI for OS and DFS. Different studies had different definitions of positive and negative GOLPH3. Accordingly, this paper adopted the definition in the original literature. HRs (95% CI) were extracted directly from certain studies via univariate and multivariate analyses. If not reported HR and 95% CI, their values were estimated in the Kaplan-Meier curve [16]. The Newcastle-Ottawa Scale (NOS) [17] tool was used to assess the quality of literature. Any disagreement was decided through a discussion of both parties or by consultation with a third team member.

Statistical analysis
Review Manager5.2 and Stata12.0 software were used for statistical analysis. The correlation between GOLPH3 overexpression and clinicopathological characteristics of CRC was analyzed by OR and 95% CI. Correlation between GOLPH3 overexpression and OS or DFS in CRC was analyzed using HR and 95% CI. According to heterogeneity, this meta-analysis was performed using fixed-effects or random-effects model analyses. HR and 95% CI were extracted using the Engauge Digitizer for references that did not provide HR but provided Kaplan-Meier survival curves instead. The I 2 test were used to quantify the potential heterogeneity that may exist among the analyzed studies. In the absence of significant heterogeneity, the data were analyzed by fixed-effect model, otherwise by random effects model.

Literature search
The process of article identification, inclusion, and exclusion is shown in Fig 1. Based on the defined criteria, 284 potential studies were initially identified. In the following studies, 42 replications and 196 unrelated studies were excluded, while 46 studies had full commentary. Subsequently, 38 articles were excluded as they were reviews (14), letters or conference abstracts (5), cell studies (15), and animal studies (4). Finally, 8 articles [18][19][20][21][22][23][24][25] were included. All studies were published between 2013 and 2018 and included a total of 723 CRC patients. The sample size ranged from 40 to 148 patients, who were all from Asia. Additionally, 4 of the studies [18][19][20][21] were published in English while 4 were [22][23][24][25] in Chinese. All studies were retrospective studies, with males accounting for 55-62.2%. GOLPH3 expression was determined by immunohistochemistry in all studies, and the thresholds for distinguishing GOLPH3 levels were found to be significantly different in each study. According to the Newcastle-Ottawa Scale, all studies scored above 5 (Table 1).

GOLPH3 expression and clinicopathological parameter
All studies reported GOLPH3 expression in CRC and reported data for at least four clinicopathological parameters ( Table 2). More specifically, 7 studies reported on gender; 7 studies reported on age; 7 studies reported on tumor stage; 6 studies reported on tumor differentiation; 8 studies reported on lymphatic metastasis; 6 studies reported on tumor size; and 4 studies reported on T stage (Table 2). When the data were combined separately, the high expression of GOLPH3 was significantly correlated with the clinicopathological parameters of the tumor. Specifically, the results were as follows: gender (male vs female) (OR = 0.89,95% CI = 0.69-1. 16 Fig 4B). Four studies [18-20,

Discussion
CRC is a very common gastrointestinal malignancy. However, due to improvements in living standards as well as changes in diet, the incidence and mortality of CRC rise annually, which seriously influences quality of life. For the treatment of CRC, surgery is considered radical treatment, while adjuvant treatment is comprised of radiotherapy and chemotherapy. Once the recurrence and metastasis will increase the difficulty of treatment. Accordingly, to clarify the pathogenesis of colorectal cancer is of great significance for the early diagnosis, treatment and prognosis of colorectal cancer [26]. During malignant tumor proliferation, the activation of various proto-oncogenes or the inactivation of tumor suppressor genes is often accompanied, leading to the abnormal expression of proteins in tumor cells [27,28]. Therefore, studying protein expression changes in tumor genesis is important in tumor targeted therapy and prognosis.
The GOLPH3 gene has been shown to promote tumorigenesis. GOLPH3 important for maintaining the maintenance of the Golgi ribbon structure, vesicle transport and Golgi glycosylation [29,30]. The role of GOLPH3 in tumorigenesis may be related to the following cellular activities: regulation of Golgi-to-plasma membrane transport and acceleration of malignant secretory phenotypes [31,32]; control of the internalization and circulation of key signaling molecules or increase of the glycosylation of cancer-associated glycoproteins [33,34]; and the influence of DNA damage response and maintenance of genomic stability [35,36]. GOLPH3 can promote the proliferation and inhibit the apoptosis of cancer cells via activation of the Wnt signaling pathway and can enhance the expression of β-catenin [37,38]. Further studies have shown that GOLPH3 can activate AKT and AKT activated Wnt signaling through GSK-3β [39]. GOLPH3 is associated with the prognosis of CRC patients receiving 5-FU-based adjuvant chemotherapy following surgery and may serve as a potential predictor of 5-FU chemical sensitivity [21]. Centromere protein H interacts with GOLPH3 to inhibit the malignant phenotype of CRC in the mTOR signaling pathway and attenuate mTORC1 and mTORC2, thereby hindering the occurrence and development of CRC [40]. Numerous studies have confirmed that the expression level of GOLPH3 protein in various solid tumor tissues, such as esophageal cancer, gastric cancer, lung cancer, liver cancer and prostate cancer, is significantly higher than that in normal tissues.

PLOS ONE
The implications of GOLPH 3 in CRC The results of this study demonstrated that the overall expression of GOLPH3 was found to be significantly higher than that of para-carcinoma tissue, which signifies that the expression of GOLPH3 in colorectal carcinoma was negatively associated with cell apoptosis. In addition, this study also found that the expression of GOLPH3 in CRC was related to clinical stage

PLOS ONE
The implications of GOLPH 3 in CRC rather than gender, age, differentiation degree, lymphatic metastasis, tumor size and T stage. The positive expression rate of GOLPH3 in patients with advanced stage was observed to be significantly higher than that in patients at an early stage [18,19]. Moreover, studies [20,41] have demonstrated that downregulation of GOLPH3 expression can inhibit the proliferation, invasion and migration of colorectal carcinoma, indicating that GOLPH3 plays an important role in the development of CRC. This study also analyzed the expression of GOLPH3 in CRC along with its clinical significance, providing guidance for the clinical treatment of CRC. As a result, methods to inhibit CRC may be determined by studying the related pathway of GOLPH3. This study also showed that the OS and DFS of CRC patients with positive expressions of GOLPH3 does not differ from that of patients with negative expressions. In light of the results of this study, it is suggested that the high expression of GOLPH3 is a factor of poor clinical stage in patients with CRC. As a result, GOLPH3 cannot be used as a prognostic indicator of CRC, and GOLPH3 may be not involved in the occurrence, development, invasion, metastasis and recurrence of CRC, which may be due to a small sample size.
Although we strived to conduct a comprehensive and scientific meta-analysis, this study has some limitations. First, all included studies were retrospective and were conducted in Asia with small sample sizes, which may lead to potential selection bias. Additionally, a question was put forward concerning the external validity of the results as well as their applicability to patients in Western countries. Therefore, prospective, large-sample, multinational studies should be performed in the future. Second, in regard to the eligible studies, the cut-off values of high expression/positive expression of GOLPH3 were different, and this inconsistency may serve as one of the reasons for the results described in this paper. Hence, further multicenter studies using standardized methods should be conducted. Third, the survival data of 4 studies were extracted from a Kaplan-Meier curve, which may affect the reliability of the prognosis. Finally, sensitivity analysis and publication bias cannot be accurately assessed due to the small number of literatures. However, random effects models were designed in order to reduce this outcome. Given the above limitations, the corresponding results should be interpreted with caution; likewise, the conclusions of this meta-analysis should be carefully drawn.

Conclusions
This study demonstrated that the overexpression of GOLPH3 was found to be positively correlated with tumor stage, which is an adverse clinicopathological characteristic of CRC. Furthermore, it was not correlated with other adverse clinicopathological characteristics, such as lymphatic metastasis, tumor size, poor differentiation of tumor or T stage. As a result, GOLPH3 can not serve as a prognostic biomarker for CRC patients.