UALCAN (http://ualcan.path.uab.edu/analysis.html), a comprehensive and interactive web resource, provides easy access to publicly available cancer OMICS data (TCGA and MET500) (14). Our study obtained LPPs level in the “Expression” links using the “TCGA analysis” module and the glioma and dataset.

GEPIA (http://gepia.cancer-pku.cn/index.html) is a developed interactive web server for analyzing the RNA sequencing expression data of tumors and normal samples from the TCGA and the GTEx project (15). In this study, we performed the pathological stage analysis and multiple gene comparison analysis of LPPs. And the correlation between PLPP4 and disease-free survival (DFS) was also calculated.

Kaplan–Meier Plotter (https://kmplot.com/analysis/) is a useful prognostic biomarker assessment tool that explored the effect of 54k genes on survival in 21 cancer types using the databases from GEO, EGA, and TCGA to analyze the prognostic value of PLPP4 in PAAD overall survival.

MethSurv (https://biit.cs.ut.ee/methsurv/), a web tool for survival analysis based on CpG methylation patterns, was applied to explore the prognostic value of single CpG methylation of PLPP4 in glioma and PAAD patients (16).

SurvivalMeth (http://bio-bigdata.hrbmu.edu.cn/survivalmeth/) was used to analyze the DNA methylation of PLPP4 signature on glioma and PAAD prognosis (17).

EMBL-EBI (https://elixir-europe.org/about-us/who-we-are/nodes/embl-ebi), a flexible pipeline for single cell RNA-seq analysis that integrates many existing tools for filtering and mapping reads, quantifying expression, clustering, finding marker genes and variable genes (18).

Timer web server (https://cistrome.shinyapps.io/timer/) is a comprehensive resource for systematical analysis of the infiltration of different immune cells and their clinical impact across diverse cancer types (19, 20).

The human glioma cell lines LN229, U251, U87MG, SHG-44, T98G, and PAAD cell lines AsPC-1, BxPC-3, PANC-1, HPAF-11, Hs766T were obtained from ATCC (The American Type Culture Collection). All cells were cultured in DMEM medium (Life Technologies, Carlsbad, CA, US) supplemented with penicillin (100 U/ml), streptomycin (100 mg/ml), and 10% fetal bovine serum (FBS, Life Technologies) and were grown under a humidified atmosphere of 5% CO₂ at 37°C.

According to the reference, based on the PLPP4 cDNA sequence in GenBank (NM_001030059.2). The interference sequence of target protein was shRNA-PLPP4: 5’-GGAGTGATGAACTCGGAAAATG-3’, and the negative control sequence was 5’-ACTACCGTTGTTATAGGTG-3’ (13); two DNA single strands expressing shRNA-PLPP4 complement connected to pGV248⁃SC1⁃PLPP4 (purchased from Shanghai Jikai Biotechnology Co., LTD.), and constructed lentivirus recombinant vector pGV248⁃SC1⁃PLPP4. HEK293T cells are then used for packaging the virus. When the degree of LN229 and PANC-1 cells fusion reached 85%, 3x10 ⁴ /mL cell suspension was prepared and inoculated on 6-well plates. When the cell density reached 30-40%, the shRNA-Con group and shRNA-PLPP4 group were transfected with negative control lentivirus and shRNA-PLPP4 lentivirus, respectively. Cells were collected when the degree of cell fusion was about 80%, and the interference effect was detected.

PLPP4 gene was synthesized by chemical synthesis method and connected to pcDNA3.1. The PLPP4 gene (BC132787.1 CDS region) with KpnI/XhoI enzyme cutting site was synthesized by chemical synthesis method, and the primers were F: 5’-GGGGTACCATGCGGGAGCTGG-3’ and R: 5’-CCCTCGAGTCACAGATCCTCTTCAG-3’, and connected to pcDNA3.1 through KpnI/XhoI after enzyme digestion. After sequencing, the vector and overexpressed plasmid were transfected into glioma cells LN229 and pancreatic cancer PANC-1 cells by liposo2000 (Invitrogen, Carlsbad, CA, USA), respectively.

Total RNA from tissues or cells was extracted using the RNA Isolation Kit-miRNeasy Mini Kit (Qiagen, USA) according to the manufacturer’s instructions (21). According to the manufacture’s protocol, the messenger RNA (mRNA) was reverse transcribed from the total mRNA using the Revert Aid First Strand cDNA Synthesis Kit (Thermo, USA). Complementary DNA (cDNA) was amplified and quantified on a CFX96 system (BIORAD, USA) using iQ SYBR Green (BIO-RAD, USA). The upstream primers of PLPP4 were 5’-TTTGGATCCGTTCCAGAGAG-3’and the downstream primers were 5’-CAGGGGTGTGAGGAAAGAAA-3’. The upstream primer of β-actin was 5’ -CATGGGCCAGAAGgACTC-3’, and the downstream primer was 5’-AAGGTCTGGAGCCAGATC-3’. Amplification conditions: 95°C for 2 min, 95°C for 30s, 60°C for 30s, 72°C for 30s, 35 cycles. Relative fold expression was calculated using the comparative threshold cycle (2-ΔΔCt) method.

Total protein of LN229 cells and PANC-1 cells in each treatment group was extracted. The exact amount of protein was taken for SDS-PAGE electrophoresis, then transferred to NC membrane, containing 50 g/L skim milk powder sealed for 1h. Rabbit anti-PLPP4 antibody (1:1000, PA5-116155, Invitrogen) and mouse anti-β-actin (1:2000, 66009-1, Wuhan Sanying Biotechnology Co., LTD., China) were added and incubated at 4°C overnight. TBST was washed 3 times, 10 min each, and HRP labeled goat anti-rabbit IgG (1:4000), rabbit anti-mouse IgG (1:8000) were added, and incubated at room temperature for 1.5 h. TBST was washed 3 times, 10 min each, ECL substrate was colored, and the absorbance (A) value of the protein bands was scanned and analyzed.

For cell counting kit-8 analysis, cells (3 × 10 ³ were seeded into 96 well plates, and the specific staining process and methods were performed according to the previous study (22). Each group is provided with 3 parallel holes. The cell growth was observed at 0, 1, 2, 3, 4, and 5 days, and 10 μL CCK-8 was added to the culture for 4h. The A value of each well was detected by the absorbance microplate reader.

Con group (LN229 cells or PANC-1 cells), vector (pcDNA3.1), overexpression, shRNA-Con, shRNA-PLPP4 cells were digested with trypsin, and the counting cells were inoculated on 6-well plates. Scratch with pipette tip when cells fuse. The plate was rinsed twice with a fresh medium to remove adherent cells. Cell migration was observed and photographed at 0 and 24h. Cell migration at the wound edge was quantified and presented as the mean ± SD.

Take 100 μLMatrigel, and 300 μL precooled serum-free medium and mix well, add 50 μL of the above prepared matrix glue into Transwell™ well, and set at room temperature for 1h. The number of logarithmic growth cells in each group was adjusted to 2×10 ⁵ cells/ml, and 200 μL of the above cell suspension was added into the upper chamber of Transwell™. Complete medium 600 μL was added to the lower chamber of the culture plate and cultured for 24h. Fixed with 40 g/L paraformaldehyde for 15 min, stained with 2 g/L crystal violet for 20 min, counted five different fields at 100 times, and calculated the average value.

The immunohistochemistry procedure and scoring of PLPP4 expression were performed as previously described (23). The glioma and PAAD tissue were cut into slices of 8um and placed in an oven at 65°C for 2 h. The sections were incubated in 3% hydrogen peroxide solution for 10 minutes and washed with PBS 3 times. The sections were placed in boiling EDTA repair solution for 20 min for antigen repair. The slices were washed with PBS 3 times, 5 minutes each, and sealed with 5% BSA for 20 minutes. The slides were incubated overnight at 4°C in a humidified chamber with anti-PLPP4 antibodies (Novus Biologicals: NBP2-14545) diluted 1:100 in PBS. Wash with PBS 3 times, 5 min each time, add rabbit secondary antibody and incubate at 37°C for 90 min. DAB staining, hematoxylin re-staining, microscopic observation, and section analysis to evaluate PLPP4 expression.

For glioma and PAAD patients of The Cancer Genome Atlas (TCGA) database, tumoral RNA-seq data were downloaded from the Genomic Data Commons (GDC) data portal (TCGA) and glioma and PAAD of the tumors also had mRNA expression data of paired normal tissue samples. All values are presented as the mean ± standard deviation (SD). Significant differences were determined using GraphPad 5.0 software (USA). The student’s t-test was used to determine significant differences between two groups. One-way ANOVA was used to determine statistical differences between multiple groups. The chi-square test was used to analyze the relationship between PLPP4 expression and clinicopathological characteristics. Survival curves were plotted using the Kaplan-Meier method and compared by log-rank test. Univariate and multivariate Cox regression analyses were performed using IBM SPSS statistical software (version 24.0), and the clinical value of PLPP4 for survival and prognosis of the two cancers was realized by ROC curve. P < 0.05 was considered significant. All the experiments were repeated three times.

We first analyzed the expression levels of LPPs proteins, including PLPP1-7, in the high throughput paired cancers’ RNA expression profile datasets from TCGA and found that the expression levels of PLPP1, PLPP2, PLPP3, PLPP5, PLPP6, and PLPP7 were expressed in both cancer tissues and normal tissues, except PLPP4 ( Figure 1A ). PLPP1 was differentially elevated in lymphoid neoplasm diffuse large B-cell lymphoma (DLBC), acute myeloid leukemia (LAML), and thymoma (THYM). PLPP2 was higher in many cancers like adrenocortical carcinoma (ACC), cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), pancreatic adenocarcinoma (PAAD), and so on, while the level of PLPP2 was lower in kidney chromophobe (KICH), pheochromocytoma and paraganglioma (PCPG), skin cutaneous melanoma (SKCM), and so on. PLPP3, PLPP4, and PLPP6 were differentially elevated in some cancers, particularly PLPP4 showed “with” and “without” expression in cancer or tumor tissues (T) compared to the respective adjacent normal tissues (N). Conversely, the expression level of PLPP7was decreased to some cancer tissues compared with those in the normal tissues ( Figure 1A ), indicating that different members of the PLPP family have oncogenic or tumor-suppressive roles in the development of cancers. Relatively independently, there was no significant difference in PLPP5 expression between cancer tissues and normal tissues. Next, we want to focus on the analysis of PLPP4 in different cancers because of the specificity of PLPP4 expression between cancer and the respective adjacent normal tissues. Furthermore, A comprehensive analysis of the molecular characteristics of PLPP4 was further performed. As a result, the mutation ratio on PLPP4 was highest in SKCM, and there were showed mutations in LUAD, UCEC, PAAD, and so on ( Figure 1B ). Therefore, our findings indicated that increased expression of PLPP4 was positively associated with advanced clinicopathological features in many cancers.

Subsequent analysis of PLPP4 expression in different types of cancer datasets from TCGA and ArrayExpress showed that PLPP4 expression was upregulated in BLCA, BRCA, GBMLGG, HNSC, LUAD, LUSC, OV, PAAD, TGCT, THYM, and UCS tissues, and down-regulated in KIRC and KIRP compared that in the respective adjacent normal tissues and was further increased in LGG and PAAD tissues ( Figures 2A–E ). We then assessed the correlation between the expression of PLPP4 and the pathological stage of BLCA, BRCA, GBMLGG, HNSC, KIRC, KIRP, LUAD, LUSC, OV, PAAD, TGCT, THYM, and UCS patients. We found a significant correlation between the expression of PLPP4 and pathological stage in BLCA, LUAD, OV, and PAAD patients ( Figures 3A –L ). These data suggested that PLPP4 played significant roles in developing BLCA, LUAD, OV, and PAAD.

To investigate the clinical correlation of PLPP4 with survival in BLCA, BRCA, GBMLGG, HNSC, KIRC, KIRP, LUAD, LUSC, OV, PAAD, TGCT, THYM, and UCS patients, these datasets from TCGA, ArrayExpress, and Kaplan-Meier Plotter were further analyzed. The results revealed that GBMLGG, LUAD, and PAAD patients with high expression of PLPP4 exhibited shorter overall survival rates compared with the corresponding patients with “low” or “none” expression of PLPP4 ( Figures 4C , G, J ). Meanwhile, there was no significant difference in the clinical correlation of PLPP4 with survival in other cancers ( Figure 4 ). Collectively, these results from publicly available cancer datasets suggest that the overexpression of PLPP4 correlates with poor prognosis and progression status in glioma, LUAD, and PAAD patients.

The PLPP4 mRNA was used for prognostic analysis in SurvExpress. The prognostic value of DNA methylation of PLPP4 in glioma, LUAD, and PAAD was analyzed by MethSurv. The heat maps of DNA methylation of PLPP4 are displayed in Figures 5A –C . Among them, cg04121368 of PLPP4in glioma, LUAD, and PAAD showed the highest DNA methylation level. And overall, we found that 15CpGs of PLPP4 in glioma, 3CpGs of PLPP4 in PAAD were significantly associated with prognosis in GLIOMA and PAAD patients, but no significant association in LUAD patients ( Table 1 ). The DNA methylation level was significantly associated with survival probability in glioma and PAAD patients. However, no statistically significant association in LUAD was found between the high- and low-risk groups ( Figures 5D –F ). Therefore, our results indicated that PLPP4 might be implicated in the development and progression of glioma and PAAD.

Univariate and Multivariate Cox regression analyses were applied to identify overall survival-related differentially expressed genes and establish a prognostic gene signature whose performance was evaluated by Kaplan-Meier curve, receiver operating characteristic (ROC). PLPP4 was significantly associated with overall survival based on the univariate Cox regression model (P < 0.001, Table 2 ). Multivariate Cox analysis showed that PLPP4 expression level (HR=1.041, 95%CI: 0.561~1.774, P=0.011), WHO grade, age, and lymph node invasion and metastasis influenced poor factors prognosis of glioma patients ( Table 2 ). Based on the expression level of PLPP4, we performed a subgroup analysis of tumor grade, age, race, gender, recurrence, and metastasis for glioma. Further, we analyzed the impact of different subgroups on survival and prognosis ( Figures 6A –D ). We found that tumor grade, age, race, and gender subgroups impacted the prognosis of glioma (all <0.001). ROC curve analysis was used to analyze the diagnostic effect of PLPP4 expression on glioma patients. The analysis results showed that the area under ROC curve analysis (AUC) of PLPP4 expression in predicting glioma was 0.723 (95%CI: 0.696 ~ 0.749, P < 0.001), its sensitivity and specificity were 89.8% and 51.3%, as shown in Figure 6E . In general, the PLPP4 signature performed well at predicting the overall survival of glioma.

More than one hundred sixty patients from the TCGA PAAD dataset were included in subsequent survival analyses. The expression of PLPP4 in PAAD with age, sex, lymph node invasion and metastasis, TNM stage, and other clinical traits was analyzed by univariate Cox regression analysis. The results indicated that the stage of PAAD (HR 0.481, P=0.034) and the expression of PLPP4 (HR 1.767, P=0.007) were associated with prognostic survival. Multivariate Cox regression analysis showed that age (HR 0.366, P=0.008), stage (HR 1.614, P=0.031), and expression of PLPP4 (HR 1.766, P=0.008) were correlated with prognosis and survival. This suggests that PLPP4 overexpression is an independent poor prognostic factor for PAAD ( Figure 7 and Table 3 ). Based on the expression level of PLPP4, we performed a subgroup analysis of PAAD based on tumor grade, age, race, gender, recurrence, and metastasis. Further, we analyzed the impact of different subgroups on survival and prognosis ( Figure 7A–E ). The subgroup analysis of tumor grade (P=0.002) and race (P=0.026) was significant ( Figures 7A –C ). Subsequently, we conducted ROC analyses to assess how PLPP4 could behave in predicting prognosis. As shown in Figure 7F , the area under the ROC curve (AUC) of the PLPP4 risk score model performed on overall survival in the training cohort was 0.853 (95%CI: 0.814-0.892, P < 0.001), and its sensitivity and specificity were 97.8% and 70.7%. Consistently, in the prediction model of progression-free survival predicted in the training cohort, PLPP4 signature risk score also showed a powerful ability. In general, PLPP4 signature performed well at predicting the overall survival of pancreatic cancer.

PLPP4 is widely expressed in various tumor cells and plays a role in tumor genesis and invasion. PLPP4 promotes proliferation and tumorigenesis via activating the influx of intracellular Ca²⁺ in lung adenocarcinoma (ADC) tissues (13, 24). To explore the biological roles of PLPP4 in glioma, immunohistochemical analysis of PLPP4 expression in glioma tissues was further examined ( Figure 8A ). As shown in Figure 8A , PLPP4 expression was primarily detected in the cytoplasm, and the staining intensity of PLPP4 was increased in glioma tissues compared with normal tissues. Therefore, our findings indicated that high expression of PLPP4 was positively associated with advanced clinicopathological features in glioma patients. Then we first examined PLPP4 expression levels in glioma cell lines by western blot, and LN229 of the glioma cell line was used for the further experiment ( Figure 8B ). As shown in Figures 8C , D , the mRNA and protein levels of PLPP4 were differentially increased via overexpression compared with vector.

In contrast, the mRNA and protein levels of PLPP4 were differentially decreased via sh-RNA compared with sh-con. To investigate the effects of overexpression or silencing of PLPP4 on proliferation, migration, and invasion of glioma LN229 cells, we first used CCK-8 cell proliferation assay to confirm that overexpression of PLPP4 enhanced cell proliferation and increased with time (P<0.05). Knocking down PLPP4 inhibited cell proliferation ( Figure 8E ). Colony formation assays revealed that silencing PLPP4 dramatically inhibited the colony-forming ability of LN229 cell; while PLPP4 overexpression promoted the growth of LN229 cell ( Figure 8F ). The scratch test determined the effect of PLPP4 on the migration of LN229. The results showed that overexpression of PLPP4 significantly increased the mobility of LN229 cells, while silencing of PLPP4 significantly decreased the mobility of LN229 cells ( Figure 8G ). Recent studies have found that PLPP4 can promote the invasion and migration in LUAD and gastric cancer (13, 25). Our results also showed that overexpression of PLPP4 promoted the invasiveness of glioma LN229 cells, while the number of cells in the shRNA-PLPP4 group was reduced, and the invasiveness was reduced by 55.2% through the transwell assay ( Figure 8H ). The results showed that inhibiting PLPP4 expression could effectively reverse the proliferation, migration, and invasion of glioma LN229 cells. Taken together, high expression of PLPP4 correlated with advanced proliferation.

Further, we analyzed the correlation between PLPP4 and other genes, and the heatmap showed genes that are positively correlated with PLPP4 and genes which are negatively associated with PLPP4 in glioma ( Figures 9A , B ). Correlogram analysis also found CD59, APOD, C4A, and QSOX1 were upregulated, while SYMD2, NAPA, COX19, and GSK3A were down-regulated ( Figures 9C –J ). These data were consistent with the heatmap results.

We identified the specificity of PLPP4 expression in glioma by single-cell sequencing analysis. PLPP4 was mainly expressed in oligodendrocyte precursor cells, a type of glial cell, and immune cells also detected the PLPP4 expression ( Figures 10A , B ). To further explore the research mechanism, we determined the co-expression of PLPP4 and other genes ( Figures 10C –F ). Positive correlation genes, like CD59 and APOD, had high co-expression, while negative correlation genes, like SMTD2 and GSK3A, had low co-expression. APOD and PLPP4 were mainly expressed in oligodendrocyte precursor cells. Most of these genes were related to the regulation of apoptosis signaling pathway, inflammatory process, and cellular metabolic pathway (26, 27). CD59 is highly expressed in several cancer cell lines and tumor tissues and regulates the function, infiltration, and phenotypes of various immune cells in the tumor microenvironment (28). APOD may be a potential therapeutic target for tumor angiogenesis by suppressing PI3K-Akt-eNOS signaling, an essential pathway regulating angiogenesis (29, 30).

PLPP4 is widely expressed in a variety of tumor cells and plays a role in tumor genesis and invasion. Another cancer with a low survival rate, PAAD, was analyzed in this project. First, immunohistochemistry showed that the expression of PLPP4 protein in PAAD tissues was significantly higher than that in normal pancreatic cancer tissues ( Figure 11A ). As the expression of PLPP4 was detected in various pancreatic cancer cell lines ( Figure 11B ), we constructed PLPP4-stably suppressing PANC-1 cells by endogenously knocking down PLPP4 via retroviral infection; meanwhile, we transfected of the constructed plasmid (vector, over-expression) into pancreatic cancer PANC-1 cells ( Figures 11C , D ). CCK-8 assays were carried out, and the results showed downregulation of PLPP4 decreased viability in PANC-1 cells ( Figure 11E ). Overexpression of PLPP4 significantly enhanced the proliferation, migration, and invasion of PANC-1 cells. Inhibition of PLPP4 expression could down-regulate the proliferation, migration, and invasion of PANC-1 cells ( Figures 11F –H ). Taken together, these findings demonstrated that silencing PLPP4 inhibits the proliferation ability of PANC-1 cells.

Further, we analyzed the correlation between PLPP4 and other genes, and the heatmap showed genes that are positively correlated with PLPP4 and genes which are negatively associated with PLPP4 in PAAD ( Figures 12A , B ). Correlogram analysis also found PTK7, HTRA1, AEBP1, and MMP14 were upregulated, while PDCD4, SYBU, ANO5, and SLAIN1 were down-regulated ( Figures 12C –J ). These data were consistent with the heatmap results. Among them, PTK7 promotes cell migration by regulating Wnt signaling pathway (31, 32). MMP14 acts as a positive regulator of cell growth and migration via activation of MMP15 (33). PDCD4 (programmed cell death protein 4) inhibits tumor promoter-induced neoplastic transformation (34, 35). These analyses could explain the short survival and high mortality of PAAD.

PLPP4 has been involved in cancer-related inflammation and the infiltration of immune cells, thus affecting the clinical outcome of cancers. Therefore, the TIMER database was used to comprehensively analyze the correlation between PLPP4 and immune cell infiltration. The results are presented in Figure 13 . PLPP4 was positively correlated with infiltration of two immune cell types (macrophages and dendritic cells; all P < 0.05). Our results further verify that overexpression of PLPP4 is observed in the glioma and PAAD tissues and cells and positively correlates with advanced clinicopathological features and poor prognosis in glioma and PAAD patients. Silencing PLPP4 inhibits the proliferation and tumorigenicity of glioma and PAAD cells in vitro. In addition, our findings reveal that DNA methylation of PLPP4 is related to survival probability in glioma and PAAD patients. Taken together, our findings indicate that PLPP4 plays an important role in the progression of glioma and PAAD and suggest that PLPP4 may serve as a potential target for human glioma and PAAD treatment.