The relationship among Girdin DNA methylation, its high expression, and immune infiltration in hepatocellular carcinoma: Clues from in silico analysis

Hepatocellular carcinoma (HCC) is the sixth most commonly diagnosed cancer and the third leading cause of cancer-related deaths worldwide [1]. More than 840,000 new cases of HCC were diagnosed globally in 2018, resulting in approximately 700,000 deaths [2]. Although surgery-dominated, comprehensive individual therapy, including chemotherapy and targeted therapy, has improved the survival of HCC patients to a certain extent, tumor recurrence and resistance to treatment remain key reasons for poor prognosis [3,4]. This highlights the need to identify novel high-precision biomarkers to improve the early diagnosis and treatment of HCC patients.

Girders of actin filaments (Girdin) is a multiple functionally protein (∼216 kDa) consisting of a C-terminal domain with signal-transducing function, a coiled-coil oligomerization domain, and an N-terminal hook domain that can bind to cytoskeletal microtubules [5]. Studies have shown that Girdin possesses diverse biological functions, including as an enhancer of AKT phosphorylation and Gα-binding vesicle-associated protein [5–7]. We have previously demonstrated that Girdin plays a key role in the formation and function of invadopodia in HCC [8]; however, no study to date has explored the relationship among Girdin DNA methylation, its high expression, and immune infiltration in HCC.

Here, The Cancer Genome Atlas (TCGA) database was used to compare Girdin mRNA expression between HCC tissues and normal tissues, to determine the relationship between Girdin DNA methylation and expression, and to analyze the association between Girdin methylation/expression and HCC prognosis. Then, the Gene Expression Omnibus (GEO) and International Cancer Genome Consortium (ICGC) databases, as well as a meta-analysis in combination with TCGA data, were used to verify the correlation between Girdin expression and HCC prognosis. Given the large number of reports on immune cell infiltration in HCC tissues in recent years, the Tumor IMmune Estimation Resource (TIMER) database was also used to explore the correlation between Girdin expression and HCC immune infiltration. Finally, Gene Ontology (GO) enrichment analysis was used to identify the biological processes in which Girdin may participate in HCC.

TCGA database (https://gdc-portal.nci.nih.gov/) was searched to obtain the Girdin mRNA and DNA methylation data, as well as the clinical data, of HCC patients [9]. The clinical data included pathological staging, grade, overall survival (OS), progression-free survival (PFS), and family history of cancer. The GEO (https://www.ncbi.nlm.nih.gov/geo/) and ICGC (https://dcc.icgc.org/) databases (a comprehensive gene expression database and a tumor genomics database, respectively), both of which are based on sequencing and array technologies, were also searched to extract the Girdin mRNA expression data as well as the survival data for HCC patients [10,11]. A total of 699 HCC cases (TCGA: 380; GEO-GSE76427: 115; ICGC- LIRI-JP: 203) and 243 normal liver cases (TCGA: 89; GEO-GSE76427: 104; ICGC- LIRI-JP: 50) were included for further analysis [10,11]. The present study conformed with the access rules and release guidelines of TCGA, GEO, and ICGC.

The data from TCGA, GEO, and ICGC databases were meta-analyzed to evaluate the significance of Girdin expression for HCC prognosis. Heterogeneity among the included studies was determined by the I²-value obtained from the Cochrane Q test and the P-value obtained from the chi-square test. If there was heterogeneity (I² ≥ 50% or P<0.05), the results were summarized using a random-effects model. Otherwise, a fixed-effect model was used for analysis. The ´meta’ R package (R version 4.0.0) was used to perform the meta-analysis.

The TIMER database (http://timer.cistrome.org/) was used to comprehensively explore the immunological, clinical, and genomic features of the tumors [12]. Additionally, the TIMER database and its algorithm were employed to analyze the correlation between the abundance of six types of immune cells (CD4+ T cells, CD8+ T cells, B cells, neutrophils, dendritic cells, and macrophages) and Girdin expression in HCC tissues and compare the prognostic data among groups with different levels of Girdin expression and immune cell abundance.

GO enrichment analysis was performed using the ´org.Hs.eg.db’ and ´enrichplot’ R packages based on the mRNA sequencing expression profile for HCC in TCGA database [13]. Two-tailed P- and q-values <0.05 were considered statistically significant.

Statistical analysis was performed using R software (version 4.0.0). The HCC-related mRNA sequencing data extracted from the different databases were all normalized using the ´limma’ package [14]. HCC tissues were divided into high and low Girdin expression groups according to the median value of Girdin mRNA expression from each database. The HCC specimens downloaded from TCGA database were also divided into Girdin hypermethylation and hypomethylation groups based on the median value of the Girdin DNA methylation level. The ´plyr’, ´reshape2’, ´ggpubr’, and ´ggplot2’ R packages were used to map the box and bubble plots. The difference in survival between the high-risk and the low-risk groups was assessed by the Kaplan–Meier method and compared using the log-rank test. A one-way analysis of variance (ANOVA) followed by a Newman–Keuls post-hoc test was used for comparisons among multiple groups. Two-tailed P-values <0.05 were considered statistically significant.

The analysis of the RNA-sequencing data for 380 HCC and 89 normal tissue specimens from TCGA database showed that Girdin expression was significantly higher in HCC tissue specimens than in normal tissues (Figure 1A). HCC patients in the high Girdin expression group had poorer OS (hazard ratio [HR] = 1.773, 95% confidence interval [CI] = 1.322–2.378, P<0.001; Figure 1B) and PFS (HR = 1.716, 95% CI = 1.333–2.209, P<0.001; Figure 1C) compared with that of patients in the low Girdin expression group. The ICGC-LIRI-JP data set, including 203 HCC cases and 50 normal liver cases from Japan, showed that high Girdin expression was closely related to poorer OS (HR = 1.873, 95% CI: 1.060–3.308, P=0.031; Figure 1D). In contrast, in the GEO-GSE76427 data set (115 HCC cases and 104 normal liver cases from Singapore), no correlation was found between high Girdin expression and OS (HR = 1.572, 95% CI: 0.506–4.882, P=0.433; Supplementary Figure S1). Given these contradictory results, we then performed a meta-analysis of the above three data sets to assess the association between Girdin expression and OS and obtain more objective conclusions. As there was no statistical heterogeneity among the three data sets (P=0.962, I² = 0%), the fixed-effects model was chosen to estimate the combined HR and 95% CI. High Girdin expression was significantly correlated with poorer OS (HR = 1.781, 95% CI: 1.382–2.297, P<0.0001; Figure 1E), suggesting that it could be used as a predictor of poor OS.

We next sought to identify the cause of the up-regulation of Girdin expression in HCC through analyzing the methylation status of its CpG sites. There were nine CpG sites spanning the whole Girdin gene in TCGA database, including the promoter region [15]. Girdin expression had no correlation with the overall DNA methylation profile or with that of the other eight CpG sites (Supplementary Figure S2A–I); however, the methylation of Girdin cg03188526 was correlated with Girdin mRNA expression (r = 0.13, P=0.0098; Figure 2B) and cg03188526 was highly methylated when compared with the other eight CpG sites (Figure 2A). We further explored the correlation between the methylation status of the nine CpG sites and OS or PFS in HCC patients. Hypermethylation at cg03188526 (P<0.001) and cg03128421 (P=0.002) were correlated with poorer OS, whereas hypermethylation at cg17470531 (P=0.002) and cg00427800 (P<0.001) was associated with better OS (Figure 3A–D). Meanwhile, cg03128421 hypermethylation was associated with poorer PFS (P=0.016, Figure 3E). The remaining sites were not significantly correlated with OS or PFS (Supplementary Figure S3A–M).

Finally, we assessed the correlation between Girdin expression or its DNA methylation and the clinicopathological characteristics of the HCC patients from TCGA database. The results revealed that, in HCC patients, high Girdin expression was positively associated with high histological grade and TNM stage, while the overall Girdin DNA hypermethylation status was negatively associated with high histological grade but positively associated with high TNM stage (Table 1).

Given the recent success of immune checkpoint inhibitors such as camrelizumab in treating HCC, we comprehensively analyzed the relationship between Girdin expression and immune infiltration in HCC tissue using the TIMER database. In tumors, Girdin expression was positively correlated with immune infiltration, including by dendritic cells (r = 0.338, P<0.001; Figure 4A), macrophages (r = 0.316, P<0.001; Figure 4B), neutrophils (r = 0.224, P<0.001; Figure 4C), CD8+ T cells (r = 0.198, P<0.001; Figure 4D), CD4+ T cells (r = 0.127, P<0.001; Figure 4E), and B cells (r = 0.164, P<0.001; Figure 4F). We further analyzed the relationship between Girdin expression and the expression of markers for the above-mentioned immune cells. Girdin expression was correlated with the CD8+ T cell markers CD8A (r = 0.225, P<0.001) and CD8B (r = 0.145, P<0.001); the B-cell markers CD79A (r = 0.125, P<0.05) and CD79B (r = 0.099, P<0.05); and the dendritic cell marker CD209 (r = 0.307, P<0.001) (Supplementary Figure S4A–E).

Given that Girdin expression in HCC was closely associated with the infiltration of a variety of immune cells and that elevated Girdin expression was closely related to poor OS, we further clarified whether the adverse effect of high Girdin expression on OS was partially mediated by immune infiltration using the Outcome module in TIMER. As indicated by the Kaplan–Meier survival curve in Figure 5, high Girdin expression was associated with worse prognosis in macrophage-enriched HCC specimens (HR = 1.316, 95% CI: 0.103–1.075, P=0.008), but not in those enriched in the other five immune cell types (Supplementary Figure S5A–E), suggesting that the adverse effects of high Girdin expression on the OS of HCC patients may be partially mediated by macrophage infiltration.

To better clarify the function of Girdin in HCC, as well the underlying mechanisms, GO enrichment analysis was performed using the ‘org.Hs.eg.db’ and ‘enrichplot’ R packages based on the mRNA sequencing expression profile for HCC derived from TCGA database. For biological processes (BP), we found that Girdin was mainly involved in nuclear cell mitosis and cell microtubule skeletal organization (Figure 6A). Interestingly, Girdin was also involved in the molecular functions (MF) of opsonin binding and complement binding, both of which are important in cellular and humoral immunity (Figure 6B).

In the present study, we performed a meta-analysis on 684 HCC samples to systematically analyze the relationship between Girdin expression and the prognosis of HCC patients. We demonstrated that Girdin expression was higher in HCC tissues than in normal tissues and that high Girdin expression was correlated with worse clinical outcome for HCC patients, which was consistent with previously reported results [16–18]. Studies have also indicated that a close relationship exists between dysregulated Girdin expression and poor prognosis in a variety of tumors, including breast cancer and lung cancer [19,20]. The oncogenic effect of Girdin has also been confirmed in several tumors [21,22]. Girdin expression can promote the proliferation, resistance to apoptosis, invasion, and metastasis of tumor cells [17,23,24], all of which are associated with worse clinical outcomes. In terms of the mechanism, Girdin can act as a signal transduction platform to amplify signaling by receptor tyrosine kinases, such as the EGFR and integrins, or by nonreceptor tyrosine kinases, such as Src, which, in turn, stimulates PI3K/AKT signaling [25–27]. We have previously shown that DLG5, Girdin, and TKS5 can interact and regulate DLG5-dependent invadopodia formation in HCC cells [8]. Furthermore, scutellarin suppresses HCC cell invasion by inhibiting Girdin activity [28]. Consistent with these observations, GO enrichment analysis indicated that Girdin was mainly involved in microtubule cytoskeleton organization during cell division and mitosis.

Interestingly, we found that the degree of methylation at cg03188526 was positively correlated with Girdin mRNA expression, and that Girdin DNA cg03188526 was highly methylated in HCC tissues. DNA methylation represents an important means of modifying gene expression and is closely related to tumor occurrence and development [29]. We noticed that the cg03188526 site is located in the gene body of Girdin. Unlike the hypermethylation of gene promoter regions, which is often closely associated with gene silencing [30], gene body methylation is not thought to be related to gene silencing, but may instead promote gene expression and alternative splicing [31,32]. Consequently, we reasoned that the cg03188526 methylation level would affect Girdin expression or its alternative splicing. In keeping with this, we found that the cg03188526 methylation level in HCC was also correlated with the prognosis of HCC patients.

Additionally, the TIMER database was used to reveal for the first time that Girdin expression in HCC is correlated with the infiltration of a variety of immune cell types, and that the adverse effects of high Girdin expression on the OS of HCC patients may be partially mediated by macrophage infiltration of the tumor microenvironment. With the progress of immunotherapy, research has increasingly focused on the putative correlations between tumors and immunity [33]. The level of tumor infiltration by immune cells is correlated with tumor growth, progression, and patient outcome [34], and studies have also reported that a close relationship exists between immune infiltration and HCC occurrence and development [35,36]. However, no study has analyzed the association between Girdin expression and immune cell infiltration. Here, we showed that the expression of Girdin in HCC was correlated with moderate to low levels of infiltration of dendritic cells, macrophages, neutrophils, CD8+ T cells, CD4+ T cells, and B cells. Importantly, we found that high Girdin expression was correlated with worse prognosis in macrophage-enriched HCC specimens, but not in those enriched with the other five immune cell types. Recent studies have demonstrated that there is a positive correlation between the expression of programmed cell death 1 ligand 1 (PD-L1) and tumor cell-intrinsic osteopontin (OPN), key drivers of macrophage infiltration and immune escape in HCC, and that both were associated with HCC development [37]. GO enrichment analysis indicated that Girdin is involved in opsonin binding and complement binding in HCC. Recent studies have shown that high levels of complement activation and the infiltration of tumor-associated macrophages can maintain chronic inflammation, promote an immunosuppressive microenvironment, induce angiogenesis, and increase the motility and metastatic potential of cancer cells [38,39]. The present study is the first to indicate that the adverse effects of high Girdin expression on the OS of HCC patients may be partially mediated by macrophage infiltration.

The present study had several limitations. First, although our in silico results suggested that methylation at a specific site in the Girdin gene body and Girdin expression are correlated with poor prognosis in HCC, cell- and animal-based experiments are needed to validate these findings. Second, we could not verify that a correlation exists between Girdin DNA methylation and PFS in the GEO-GSE76427 and ICGC-LIRI-JP data sets because they do not include information for PFS and Girdin DNA methylation. Third, given that this was an exploratory study and a significance level of 5% was selected, no adjustment for multiple testing was performed [40]. However, it is worth noting that, irrespective of the study type, not adjusting for multiple comparisons may result in an inflation of the false-positive rate [40].

In the present study, we found that high Girdin expression and cg03188526 hypermethylation were correlated with poor prognosis for HCC patients, while hypermethylation at the cg03188526 site in the Girdin gene body may explain, at least in part, its high expression in HCC. Finally, the adverse effect of high Girdin expression on the OS of HCC patients may be partially mediated by macrophage infiltration of the tumor microenvironment.

The data sets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

The authors declare that there are no competing interests associated with the manuscript.

This work was supported by grants from the National Natural Science Foundation of China [grant numbers 81660399 and 81860423]; the Medical Leading Talent Project of Yunnan Province [grant number L201622]; the Yunnan Provincial Clinical Center of Hepato-biliary-pancreatic Diseases [no specific number] (to L.W.); the Applied Fundamental Research Joint Program of Science and Technology Department of Yunnan Province and Kunming Medical University (Youth PhD Fund) and the Leading Academic and Technical Young and Mid-aged Program of Kunming Medical University [grant number 60118260108] (to Y.K.).

C.Z., Y.K., X.L., X.L., and H.L. designed the study. C.Z., Y.K., and X.L. analyzed the data and wrote the manuscript. H.L. and R.S collected the data. L.W. revised the manuscript. All authors approved the final version of the manuscript for submission.

GEO

Gene Expression Omnibus

HCC

hepatocellular carcinoma

ICGC

International Cancer Genome Consortium

OPN

osteopontin

OS

overall survival

PFS

progression-free survival

PD-L1

programmed cell death 1 ligand 1

TCGA

The Cancer Genome Atlas

TIMER

Tumor IMmune Estimation Resource