The prognostic and clinicopathological significance of desmoglein 2 in human cancers: a systematic review and meta-analysis

Introduction

Cancer is the leading cause of death worldwide, and with increasing incidence, is becoming a severe public heath burden. Despite many advances in our understanding of this disease, it is still uncommon for cancer patients to be diagnosed at an early stage, and instead, the majority of patients are initially diagnosed at an advanced stage. Additionally, the prognosis of these cancer patients remains poor due to limited efficacy of conventional treatments such as chemoradiotherapy. Therefore, it is of great value to identify factors closely related to cancer initiation and progression that can be used as effective biomarkers for early diagnosis, prognosis evaluation, and treatment guidance in cancer patients (Siegel et al., 2021).

Desmoglein 2 (DSG2), a trans-membrane glycoprotein belonging to the cadherin superfamily, mediates cell-cell adhesion via intact desmocollin to form desmosome sand and plays an important role in maintaining epithelial tissue integrity. Recently, several studies reported that aberrant expression of DSG2 was associated with tumorigenesis and cancer progression in a variety of malignancies. However, the exact role of DSG2 in cancer remains unclear and appears to be complicated, with data indicating it can function as both a tumor suppressor and an oncogene in different cancer types. For example, Chang et al. (2021) reported that high expression of DSG2 increased cancer cell invasiveness and led to worse prognosis in breast cancer, whilst the opposite was observed in another study, which found that knockdown of DSG2 in anaplastic thyroid cancer cells increased cell migration and distant metastasis in xenografts (Lee et al., 2020). On the issue of relation between prognosis and DSG2 expression, existing studies also showed different results. In NSCLC, Cai et al. (2017) and Jin et al. (2020) reported that high expression of DSG2 was associated with unfavorable prognosis and a negative correlation between DSG2 expression and prognosis was also found by Han et al. (2018) in hepatocellular carcinoma. However, in esophageal squamous cell carcinoma, DSG2 expression showed no significant association with prognosis (Fang et al., 2014) and in ovarian serous carcinoma, Chen et al. (2018) reached the reverse conclusion that low expression of DSG2 was related to poor prognosis. Taken together, the relationship between DSG2 expression and prognosis in cancer patients is inconclusive and is yet to be systematically reviewed. Thus, this study aimed to define the prognostic value of DSG2 in patients with cancer.

Methods

The study protocol was registered on the PROSPERO database (http://www.crd.york.ac.uk/PROSPERO) with ID: CRD42020202201. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) was followed (Moher et al., 2009).

Results

Search results

We identified 116 studies from PubMed, 0 from the Cochrane Library, 212 from Embase, 181 from Web of Science and 151 from Scopus database. Twelve studies were also found through the references. We had a total of 358 articles after removing duplication. First, after reading the title and abstract, 279 articles were excluded for irrelevance. Then, we read the full text of the remaining 79 articles, among which 41 in vivo studies, 21 conference abstracts, four reviews and two bioinformatics analyses were excluded. Ultimately, 11 studies (Barber et al., 2014; Cai et al., 2017; Chen et al., 2018; Fang et al., 2014; Han et al., 2018; Jin et al., 2020; Ormanns et al., 2015; Qin et al., 2020; Sun et al., 2020; Xu et al., 2020; Yashiro, Nishioka & Hirakawa, 2006) containing 1,488 cases were selected for systematic review, and 10 studies were included in the meta-analysis (Fig. 1).

Association between DSG2 expression and OS

In total, 10 studies containing 1,074 patients were included in the meta-analysis to assess the association between DSG2 expression and OS (Cai et al., 2017; Chen et al., 2018; Fang et al., 2014; Han et al., 2018; Jin et al., 2020; Ormanns et al., 2015; Qin et al., 2020; Sun et al., 2020; Xu et al., 2020; Yashiro, Nishioka & Hirakawa, 2006). The pooled HR suggested no statistical significance (HR = 1.02, 95% CI [0.75–1.40], I² = 88%, random-effect model) (Fig. 2). Considering that the cancer types were different, we further conducted subgroup analysis to explore the association between DSG2 expression and OS in different cancer types. Studies were divided into three subgroups based on cancer types: NSCLC (Cai et al., 2017; Jin et al., 2020; Sun et al., 2020), digestive system cancer (Fang et al., 2014; Han et al., 2018; Ormanns et al., 2015; Xu et al., 2020; Yashiro, Nishioka & Hirakawa, 2006) and female reproductive system cancer (Chen et al., 2018; Qin et al., 2020). Barber’s study was not included in the meta-analysis because it only reported RFS (Barber et al., 2014). The results of the meta-analysis showed that high DSG2 expression was significantly associated with poor OS in NSCLC (HR = 1.54, 95% CI [1.11–2.15], I² = 81%, random-effect model). In digestive system cancer, DSG2 expression was not statistically correlated with OS (HR = 0.68, 95% CI [0.37–1.24], I² = 84%, random-effect model). In female reproductive system cancer, the result also indicated no statistical significance (HR = 1.66, 95% CI [0.33–8.22], I² = 86%, random-effect model).

Publication bias and sensitivity analysis

To detect the publication bias among included studies, we mapped the funnel plot in RevMan software (Fig. 3). In the funnel plot, the points on both sides were evenly distributed, indicating on significant publication bias among the included studies. We also did the sensitivity analysis to detect the robustness of our result. When we omitted the included studies one by one in RevMan software, the result of meta-analysis did not dramatically change, showing that our result was stable and robust (Supplemental File 1).

Discussion

Cell-cell anchoring junctions are important for tumor cell biology and have been gaining increased attention (Wu et al., 2019). It has been reported that cell-cell interactions via E-cadherin can protect cancer cells from ferroptosis, an iron-dependent programmed cell death pathway, whilst integrin-mediated anchoring can lead to metabolic reprogramming to support cancer cell survival (Endo et al., 2020). DSG2, a key factor that mediates cell-cell junctions, was also reported to regulate cancer cell behavior and could be a biomarker for prognosis in cancer patients. In this study, we performed a systematic review and meta-analysis that included 11 studies with 1,488 patients in total. The combined data from all 10 included studies showed no correlation between DSG2 expression and prognosis. Considering the heterogeneity in the 10 included studies for the meta-analysis, which might lead to unreliable conclusions, we conducted further subgroup analysis. We assessed the association between DSG2 expression and the prognosis of cancers in NSCLC, digestive system cancer and female reproductive system cancer. In NSCLC, high DSG2 expression was significantly associated with unfavorable survival. Together with the pooled OS and OR of the indicated clinicopathological features, our analyses suggested DSG2 may function as a promoter of tumor progression, which is supported by our mining of the TCGA database. Furthermore, the TCGA database mining also indicated that high DSG2 expression was associated poor prognosis in NSCLC (Supplemental File 2). In digestive system cancers, there was no statistically significant association between DSG2 and survival. There was only one study in male reproductive system cancer and two studies in female reproductive system cancers, thus, due to insufficient evidence; we have not yet been able to draw convincing conclusions in these two subgroups.

The function of DSG2 is not limited to the cell junction (Chidgey & Dawson, 2007). According to recent findings, the mechanisms of DSG2 in regulating cancer development and the response to anticancer treatment are associated with cell cycle changes and molecular interactions involved in signal transduction (Hao, Liu & Ma, 2021; Ungewiß et al., 2018). In NSCLC, Cai et al. (2017) reported that DSG2 knockdown inhibited NSCLC xenograft growth in nude mice. Knockdown of DSG2 reduced NSCLC cell proliferation and arrested them at the G1 phase of the cell cycle. In addition, high DSG2 expression was also reported to promote cancer cell growth and migration, and decrease the response sensitivity to EGFR targeting therapy in NSCLC cancer cells. A further study revealed that it was mediated via DSG2-activated Src and the downstream Rac1-PAK1 signaling pathway by binding and stabilizing EGFR in the cell membrane in NSCLC cancer cells (Jin et al., 2020). Therefore, we consider that its effect on the cell cycle and interaction with EGFR contribute to the unfavorable prognosis of NSCLC. Unlike lung cancer, the role of DSG2 in digestive cancers was not well defined and previous reports were even paradoxical. For instance, Han et al. (2018) reported that DSG2 expression was significantly higher in hepatocellular carcinoma tumor tissues than in matched noncancerous tissues and positively correlated with tumor size, tumor stage and unfavorable prognosis, while Xu et al. (2020) illustrated that the expression of DSG2 proteins was significantly lower in extrahepatic cholangiocarcinoma tissues than in normal tissues and that negative DSG2 expression was an independent poor prognosis factor. In anaplastic thyroid cancers, Lee demonstrated that DSG2 knockdown induced cell motility and invasiveness through the cMet/Src/Rac1 signaling axis and that targeting cMet inhibited DSG2 knockdown-induced distant metastasis in mouse xenografts (Lee et al., 2020). All these studies indicated a differing role for DSG2 in cancer development, progression and response to anticancer therapy, and thus further studies are needed to uncover the mechanisms at play.

Admittedly, there are limitations to our systematic review and meta-analysis. First, the number of included studies was relatively limited. Second, the sample size of each included study was also limited, and all eligible studies were retrospective studies. Third, it was difficult to determine a standard expression cut-off value because of different cancer types and diverse detection methods in the included studies, which may also lead to heterogeneity among the included studies. Therefore, further high-quality clinical studies are required to attain a stable conclusion.

Conclusion

The association between DSG2 expression and prognosis was not consistent between different types of cancers. Our study demonstrated that in NSCLC, high DSG2 expression was associated with poor prognosis. Therefore, DSG2 can potentially be a biomarker of the prognosis of NSCLC.

Supplemental Information