We downloaded RNA sequencing (RNA-seq, Illumina Inc, San Diego, CA, USA) data of 701 patients, including 532 cases of low-grade glioma (LGG) and 169 cases of GBM, from the UCSC Xena database for further analysis [20]. Downloading from the CGGA website, enabled a dataset comprising 1013 cases of glioma patients to be obtained (625 cases of LGG and 388 cases of GBM). Information from the Gene Expression Profiling Interactive Analysis (GEPIA) database was used to analyse the of PLEKHA4 in various cancers [21]. The Gliovis database [22] was used to assess the correlation between PLEKHA4 expression and various clinical features including WHO tumour grade, sex, methylation status of the O(6)-methylguanine-DNA methyltransferase (MGMT) gene promoter, sequence deletions on chromosomes (Chr) 1p/19q, mutation status of the gene encoding isocitrate dehydrogenase (IDH), and different glioma molecular subtypes (mesenchymal, classical, and proneural). We used representative immunohistochemical (IHC) staining images of PLEKHA4 in gliomas obtained from The Human Protein Atlas (HPA) online database (http://www.proteinatlas.org).

Kaplan–Meier survival analysis was used to determine the relationship between the expression levels of PLEKHA4 and overall survival (OS) in patients with glioma. The R package rms was used to build a nomogram based on the Cox method and the prognostic values of the features of interest was assessed. Risk scores were obtained from the analysis and receiver operating characteristic (ROC) curves were plotted to show their predictive values for TCGA and CGGA glioma.

PLEKHA4 gene co-expression data were obtained from the LinkedOmics database (http://linkedomics.org/). The AmiGO 2 database (http://amigo.geneontology.org/) was searched for both non-canonical and canonical Wnt signalling pathway information. Enrichment analysis was performed to assess the functional enrichment of 100 genes that were positively associated with PLEKHA4 in gliomas. GSEA was conducted using GSEA software (version 3.0, Broad Institute Inc, Cambridge, MA, USA) [23].

We downloaded the TCGA pan-cancer (N = 10,535, G = 60,499) database, which consists of a uniformly normalized dataset, and extracted expression data for the PLEKHA4 (ENSG00000105559) gene from each sample. We obtained RNAss (RNA stemness scores) and DNAss (DNA stemness scores) by calculating epigenetic and transcriptomic features for each tumour, respectively, as previously reported by Malta et al. [24]. Finally, we integrated the stemness scores and gene expression data of the samples to perform correlation analyses using the Sangerbox tool (http://sangerbox.com/).

We used the R package ESTIMATE to calculate the stromal, immune, and ESTIMATE (version 1.0.13, https://bioinformatics.mdanderson.org/estimate/) scores of each patient. CIBERSORT (https://cibersort.stanford.edu/) was used to assess the immune infiltration scores [25].

PLEKHA4 expression and drug sensitivity were investigated using the GDSC and Clinical Trial Reporting Program (CTRP) databases [26, 27].

Human glioma (LN-229 and U-118MG) cell lines were obtained from Meison Cell Corporation (Jinhua, Zhejiang, China). All cells were tested for STR analysis and mycoplasma contamination. When the LN-229 and U-118MG cells, cultures in six-well plates, reached 70–80% confluency, they were respectively transfected with siRNA (32330, Genepharma Inc, Suzhou, Jiangsu, China), using Lipofectamine 3000 transfection reagent (Thermo Fisher Scientific, Waltham, MA, USA). The siRNA sequences are shown (Supplementary Table 1).

Total RNA was extracted from cell samples using Trizol (T9424, Sigma-Aldrich Inc, Springfield, MO, USA). Next, the concentration of RNA was detected using a NanoDropOne spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Furthermore, the RNA was reverse transcribed using a cDNA synthesis kit (RR036A, Takara Bio Inc, Yamanashi, Japan). Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was performed using TBGreen (RR430A, Takara Bio Inc, Yamanashi, Japan) with a LightCycler96 system (version 1.1, Roche Diagnostics Inc, Basel, Switzerland). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used to normalize PLEKHA4 levels; primer sequences for the RT-qPCR are shown (Supplementary Table 1).

Approximately 4000 LN-229 and U-118MG cells were first inoculated on 96-well plates, and then transfected with PLEKHA4-siRNA the next day. To assess the effect of PLEKHA4 on the malignant biological behaviour of glioma cells, the effect of PLEKHA4 knockdown was examined by continuously monitoring cell numbers using the IncuCyte system (20$\times$ objective), during which period images were acquired every 3 h for 96 h, following which they were analysed.

R software (v.4.0.5, RStudio Inc., Boston, MA, USA) was used to statistically analyse the datasets from TCGA and CGGA databases. Wilcoxon rank sum and chi-square tests for continuous and categorical variables, respectively. The correlation between clinical characteristics and OS was assessed using multivariate and univariate Cox regression analyses. Correlations between PLEKHA4 expression and immune infiltration and tumour stemness scores (DNAss and RNAss) were calculated and assessed using Spearman’s correlation coefficient. Results with a false discovery rate (FDR) 0.25 and p 0.05 were considered statistically significant.

The results of the pan-cancer analysis revealed that PLEKHA4 expression levels in different tumours of most cancer types varied compared to levels in normal tissues (Fig. 1A). The transcript levels of PLEKHA4 in LGG and GBM were significantly higher than those in normal brain tissues (Fig. 1B).

An assessment of the correlation between PLEKHA4 expression and molecular characteristics of patients with glioma showed higher PLEKHA4 expression levels in gliomas without Chr 1p/19q co-deletions, mesenchymal subtypes, IDH wild-type (WT) gliomas, and MGMT non-methylated gliomas (Fig. 1C). Furthermore, analysis of the in-situ expression of PLEKHA4 using the HPA database, showed higher PLEKHA4 levels in both GBM and LGG (Fig. 1D).

Fig. 1.

Pleckstrin homology containing family A, number 4 (PLEKHA4) expression is increased in gliomas. (A) Expression pattern of PLEKHA4 mRNA in The Cancer Genome Atlas (TCGA) tumours, with normal tissues matched to the Genotype–Tissue Expression (GTEx) database. (B) PLEKHA4 expression levels in TCGA gliomas and normal tissues, with normal tissues matched to those in the GTEx database. (C) Correlation of PLEKHA4 expression with various clinical features, including World Health Organization (WHO) tumour grade, isocitrate dehydrogenase (IDH) mutation status, 1p/19q co-deletion status, methylation status of O(6)-methylguanine-DNA methyltransferase (MGMT) gene promoter and different glioma molecular subtypes. (D) PLEKHA4 protein staining in different grades of glioma tissues from Human Protein Atlas. The scale bar represents 200μm and 50μm. (ns, no significance, p $\ge$ 0.05, *p 0.05, **p 0.01, and ***p 0.001). LGG, low-grade glioma; GBM, glioblastoma; CGGA, Chinese Glioma Genome Atlas; TCGA, The Cancer Genome Atlas.

DNA methylation can regulate gene expression without affecting the sequence of genes. Therefore, identifying the methylation status of PLEKHA4 can help elucidate the phenomenon of increased PLEKHA4 mRNA expression in gliomas. The PLEKHA4 gene has seven methylation sites, and in LGGs, six of them (cg01093065, cg06339706, cg06705122, cg15549821, cg18689573, and cg23002721) were negatively correlated with PLEKHA4 mRNA expression levels, whereas one (cg25261547) was positively correlated (Fig. 2A). The six negatively correlated methylation sites had a lower methylation level in GBM than they did in LGG (Fig. 2B,C), whereas the PLEKHA4 mRNA expression level was higher in GBM than it was in LGG (Fig. 1B). In addition, hypermethylation at cg06339706 and cg06705122 sites significantly affected the prognosis of patients with LGG, who showed a longer survival time than patients without hypermethylation (Fig. 2D,E).

Fig. 2.

Methylation analysis and its prognostic significance in glioma. (A,B) Heat map showing methylation sites in LGG and GBM. (C) Comparison of the methylation degree of pleckstrin homology containing family A, number 4 (PLEKHA4) between LGG and GBM. (D,E) Kaplan–Meier curves show that hypermethylation of CpG sites (cg06339706 and cg06705122) corresponds to longer survival in patients with LGG (**p 0.01, ***p 0.001 and **** p 0.0001). LGG, low-grade glioma; GBM, glioblastoma.

Dichotomous grouping of patients from the TCGA and CGGA databases showed that high PLEKHA4 expression was mostly associated with a worse OS (Fig. 3A,B). Univariate and Cox multifactor regression analysis showed that PLEKHA4 expression with tumour grade (LGG vs GBM), MGMT promoter status (methylated vs unmethylated), Chr7 gain and Chr10 loss status (yes vs no), Chr 1p/19q co-deletion (yes vs no), IDH status (Mut vs WT), and age were independent prognostic factors (Tables 1,2). The results suggest that gliomas expressing high levels of PLEKHA4 may have a worse prognosis compared to low levels of PLEKHA4 expression.

Fig. 3.

Prognostic value analysis of pleckstrin homology containing family A, number 4 (PLEKHA4) in glioma. Prognostic value of PLEKHA4 in glioma examined using (A) The Cancer Genome Atlas (TCGA) and (B) Chinese Glioma Genome Atlas (CGGA) database.

Furthermore, we integrated data on survival status, survival time, and seven characteristics in a sample of 621 TGCA gliomas to create a nomogram to assess the prognostic significance of these characteristics (Fig. 4A). The overall C-index of the model for the TCGA cohort was 0.865 (95% confidence interval [CI], 0.844–0.887; p = 4.096 $\times$ 10${}^{-243}$). We integrated data on survival time, survival status and five characteristics in a sample of 983 CGGA gliomas to create a nomogram to assess the prognostic significance of these characteristics (Fig. 4B). The overall C-index of the model for the CGGA cohort was 0.759 (95% CI, 0.740–0.778; p = 8.297 $\times$ 10${}^{-158}$) . The line graph predicted the 1-, 3-, and 5-year OS of patients with glioma (Fig. 4C,E). In addition, the risk score was calculated based on the results of the Cox regression analysis and the plotted ROC curves showed area under the curve (AUC) values of 0.90, 0.94, and 0.94 at 1, 3, and 5 years, respectively in TCGA glioma (Fig. 4D). In addition, AUC values for the CGGA glioma cohort were 0.80, 0.86, and 0.88 at 1, 3, and 5 years, respectively (Fig. 4F).

Fig. 4.

Prognostic prediction of glioma in patients using pleckstrin homology containing family A, number 4 (PLEKHA4) expression. Construction of nomogram to predict the overall survival (OS) for glioma patient data, (A) The Cancer Genome Atlas (TCGA) and (B) the Chinese Glioma Genome Atlas (CGGA). Calibration curves show the OS of (C) TCGA and (E) CGGA glioma cohorts. Risk scores were obtained from analysis and receiver operating characteristic (ROC) curves show predictive value with (D) TCGA and (F) CGGA glioma. IDH, isocitrate dehydrogenase; MGMT, O(6)-methylguanine DNA methyltransferase

The analysis of co-expressed genes showed that 10,510 and 9605 genes were positively and negatively correlated with PLEKHA4 expression, respectively. The top 50 genes positively associated with PLEKHA4 were selected based on correlation, following which heat maps were constructed (Fig. 5A). In addition, enrichment analysis of the top 100 genes that showed a positive correlation with PLEKHA4 in gliomas was conducted using Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) databases. The results were visualized using bubble plots.

KEGG pathway enrichment revealed that the co-expressed genes were enriched in pathways and processes such as tuberculosis, human immunodeficiency virus 1 infection, phagocytosis, NOD-like receptor signalling, influenza A, leukocyte trans-endothelial migration, and antigen processing and presentation (Fig. 5B). The GO analysis of BPs suggested that the co-expressed genes were involved in processes, such as response to biotic stimulus, defence response, immune effector process, defence response to other organism, innate immune response, response to cytokines, cell activation, response to interferon gamma, and cytokine-mediated signalling pathway (Fig. 5C). GO cellular component (CC) analysis indicated that the co-expressed genes were involved in components, such as the Golgi apparatus; anchoring junction; cell substrate junction; phagocytic vesicle; and peptidase inhibitor complex (Fig. 5D). Furthermore, the GO molecular function (MF) analysis suggested that the co-expressed genes were involved in the functions including identical protein binding; cysteine type endopeptidase activity involved in the apoptotic process; phospholipase a2 inhibitor; deoxycytidine deaminase; cytidine deaminase; and phospholipase inhibitor activities; S100 protein binding; and card domain binding (Fig. 5E).

Fig. 5.

Functional analysis of pleckstrin homology containing family A, number 4 (PLEKHA4) in gliomas. (A) LinkedOmics database exploration of genes co-expressed with PLEKHA4 in gliomas. (B–E) Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of genes co-expressed with PLEKHA4 in gliomas.

Signalling pathway analysis using a GSEA tool based on the KEGG database revealed that high PLEKHA4 expression was associated with janus kinase/signal transducer and activator of transcription (JAK/STAT) signalling, regulation of actin cytoskeleton, apoptosis, chemokine signalling, and cancer (Fig. 6A). Annotation of GO BP terms suggested that high PLEKHA4 expression may be associated with regulation of T cell activation, and the JUN N-terminal kinase (JNK) cascade, as well as the positive regulation of canonical Wnt signalling pathway (Fig. 6B). Annotation of GO CC terms suggested that high PLEKHA4 expression may be associated with protein complexes involved in cell adhesion, actin filament, actin filament bundle, and actin cytoskeleton (Fig. 6C). Annotation of GO MF terms suggested that high PLEKHA4expression levels may be associated with cysteine type endopeptidase regulator activity involved in the apoptotic process, integrin binding, and laminin binding (Fig. 6D, Table 3).

Fig. 6.

Gene set enrichment analysis (GSEA) of pleckstrin homology containing family A, number 4 (PLEKHA4) expression levels. (A–D) Kyoto Encyclopedia of Genes and Genomes (KEGG); Gene Ontology (GO) biological process (BP); GO cellular component (CC); and GO molecular function (MF) analyses were performed using the GSEA tool. (E) PLEKHA4 and top 60 Wnt pathway-related genes are shown in the heat map.

We further loaded 151 Wnt signalling pathway-related genes (Supplementary Table 2), which were identified by searching the AmiGO 2 database. The heat map displayed the top 60 genes associated with PLEKHA4 expression among these genes, indicating a potential regulatory role for PLEKHA4 in the activities of the Wnt signalling pathway in glioma (Fig. 6E).

RNAss and DNAss were calculated based on transcriptomic and epigenetic feature sets, respectively, and an analysis of their correlation with PLEKHA4 expression in pan-cancer showed relatively higher correlation with glioma than with other cancers (Fig. 7A,C). Meanwhile, PLEKHA4 expression was significantly correlated with glioma cell stemness (Fig. 7B,D).

Fig. 7.

Analysis of the correlation between pleckstrin homology containing family A, number 4 (PLEKHA4) expression and tumour stemness. Correlation between PLEKHA4 expression and DNA and RNA stemness scores (DNAss and RNAss, respectively) in (A–D) TCGA gliomas.

The correlation between high PLEKHA4 expression and T cell activation prompted us to further investigate the relationship between PLEKHA4 expression and the immune microenvironment in gliomas. The results indicated that increased PLEKHA4 expression was associated with higher immune, stromal, and ESTIMATE scores (Fig. 8A–C). CIBERSORT analysis indicated that regulatory T cells (Treg), resting memory CD4${}^{+}$ T cells, M2 and M1 macrophages, naive B cells, $\gamma$$\delta$T cells, activated dendritic cells, and neutrophils were enriched in the group showing high PLEKHA4 expression (Fig. 8D). Correlation analysis also confirmed that PLEKHA4 expression was significantly associated with increases in resting memory CD4${}^{+}$ T cells, Treg, M2 and M1 macrophages, naive B cells, gamma delta T cells, activated dendritic cells, and neutrophil immune cells (Fig. 8E).

Fig. 8.

The expression levels of Pleckstrin homology containing family A, number 4 (PLEKHA4) were correlated with the immune microenvironments of gliomas. (A–C) Correlation between PLEKHA4 expression and immune, stromal, and ESTIMATE scores. (D) Violin plot depicting CIBERSORT scores of 22 types of immune cells in high versus low expression groups. (E) Correlation analysis of PLEKHA4 expression levels using CIBERSORT score of 22 types of immune cells in The Cancer Genome Atlas (TCGA) glioma. (*p 0.05, **p 0.01, ***p 0.001, and ****p 0.0001).

The results of the drug sensitivity analysis showed that PLEKHA4 expression was possibly negatively correlated with the sensitivity of various small molecule kinase inhibitors, such as selumetinib, trametinib, and dabrafenib, in the GDSC database (Fig. 9A). The CTRP drug sensitivity analysis also showed a correlation with small molecule kinase inhibitors, while PLEKHA4 expression was negatively correlated with virofenib drug sensitivity (Fig. 9B).

Fig. 9.

Drug sensitivity analysis. Correlation between pleckstrin homology containing family A, number 4 (PLEKHA4) expression and (A) Genomics of Drug Sensitivity in Cancer (GDSC) and (B) Cancer Treatment Response Portal (GTRP) drug sensitivity in pan-cancer. FDR, false discovery rate.

The results of analyses utilized to investigate the potential role played by PLEKHA4 in glioma progression were validated vis conventional experimental methods. We transfected LN-229 and U-118MG cells with siRNA and verified PLEKHA4 knockdown efficiency using RT-PCR assay (Fig. 10A,B). The association between PLEKHA4 and the malignant biological behaviour of glioma cells was assessed via image analysis, which indicated that knockdown of PLEKHA4 expression can inhibit glioma cell proliferation (Fig. 10C–F).

Fig. 10.

The effect of pleckstrin homology containing family A, number 4 (PLEKHA4) expression on glioma cell proliferation. (A,B) Comparison of the efficiency of PLEKHA4 knockdown using small interfering RNA (siRNA) sequences in LN-229 and U-118MG cell lines. IncuCyte automated bright-field imaging of cell proliferation in (C,D) LN-299 and (E,F) U-118MG glioma cells treated with siRNAs targeting different regions of PLEKHA4 (siPLEKHA4 #1, #2) or negative control siRNAs (****p 0.0001).