PPL Expression is up-regulated in OV patients’ tissues
The PPL mRNA expression from GTEx projects and TCGA samples was compared using the GEPIA2 initially. The obvious differences were observed in fig 1a. (P＜0.001). Then, the oncomine database was further used to analyze PPL mRNA levels in normal tissues and tumor tissues in a number of studies, which showed a higher PPL expression in OV tumors tissues than that in normal ovary (P＜0.05, fig. 1b).Therefore, we surmised PPL up-regulation may serve as one of the tumor drivers in OV. However, the PPL expression profile may indicate the different roles of PPL in pan-cancer. A differential PPL mRNA expression was shown between normal tissues and tumor tissues across the different cancers (fig. 1c). Besides the differences between OV tissue and normal ovary in transcriptome level of PPL, the examination of the PPL protein level also suggested the consistent trend, showing the PPL protein level in OV tissues was higher compared with normal ovary tissues. Representative images of PPL staining in OV and normal ovary were shown in figure 1d from human protein atlas data, while figure 1e indicated the non-uniform differential of PPL protein level between the different types of cancers and normal tissues.
Elevation in PPL is associated with poor prognosis in OV patients
To investigate the prognostic implication of PPL in OV, Kaplan-Meier analysis was conducted based on the TCGA data. The patients were grouped according to the median value of PPL mRNA expression. The results displayed the group with high PPL expression had a significant shorter OS compared with the low group (OS: Cox P = 0.011, HR = 1.4, Fig. 2a). Furthermore, the similar results in OV were shown in the survival data depending on Kaplan-Meier plotter online tool (Fig. 2b). What’s more, high PPL expression also showed a strong correlation with poor survival in pancreatic ductal adenocarcinoma, bladder cancer, and uterine corpus endometrial carcinoma (Fig. 2a-b). However, the inconsistent founding was shown in adrenocortical carcinoma (ACC), sarcoma (SARC), and breast cancer, etc, indicating the low expression of PPL was correlated to poor survival.
Role of PPL in an independent OV cohort
The clinicopathological features of 42 recruited OV patients were listed in Table 1. The expression of PPL was tested in the 42 tumor samples and 10 normal ovaries, respectively. The mRNA expression and protein of PPL in OV tumors were both higher than those in normal ovary (P=0.001, P=0.001, fig. 3a-b, Table 2). Representative images of PPL protein staining were shown in figure 3b PPL protein was mainly expressed in cell membrane and cytoplasm. Among the 42 OV tumors, the positive expression rate of PPL was 90.47%. Among them, there were 17 cases with moderate staining, of which 7 cases were from the patient with FIGO Stage 2, 4 cases were from the patient with FIGO Stage 3 and 6 cases were from the patient with FIGO stage 4. The negative for PPL protein was only in 4 cases of OV.
Further, the 42 patients were divided into low and high group according to the median value of PPL mRNA expression. Kaplan-Meier analysis indicated that the PFS and OS of patients with PPL high group were obviously shorter compared with PPL low group (P = 0.008, P = 0.033, fig. 3c).
Evaluation and estimation of nomogram
Multivariable analysis for survival in TCGA was performed with PPL expression index included to detect if PPL plays a prognostic-associated role in OV patients. Results indicated that, when tumor residual size, grade, stage and age were included, a significant association between OS and PPL mRNA levels (P = 0.002, HR =1.3, 95%CI = 1.09-1.5, Fig.4a), suggesting that for OV patients PPL mRNA expression may be a potential independent prognostic biomarker figure 4b showed nomogram based on multivariate cox regression analysis. Depending on clinical information as well as genes, patients 3- and 5-year survival can be accurately predicted, and patients’ prognosis can be visually predicted by nomograms. The model had a strong prediction power because of the predicted values were quite close to observed results of 3- or 5-year survival probabilities (fig. 4c).
PPL co-expression networks in OV
To gain a deep understanding of PPL biological functions in OV, the function fig 5A indicated the showed negative and positive correlation with PPL. The figure 5b presented top 50 significantly correlated genes (either negatively or positively). The figure 5c showed the STRING online database derived PPL network and its co-expression genes. A total of 502 PPL expression correlated essential genes were identified and listed in supplement table 1. By method MCC in CytoHubb, we found 10 hub genes (LAMC2, PXN, LAMA3, LAMB3, LAMA5, ITGA3, TLN1, ACTN4, ACTN1, ITGB4, Fig. 5d), of which the survival map of these ten genes in OV was shown in figure 5e.
Significant GSEA annotated geneontology_Biological_Process_no Redundant term indicated that PPL co-expressed genes mainly participate in Ras protein signal transduction, cell-cell signaling by wnt, cell cycle checkpoint, postreplication repair, DNA-templated transcription, termination, etc. KEGG pathway analysis indicated the existence of enrichment in Wnt signaling pathway, MAPK signaling pathway, Hippo signaling pathway, EGFR tyrosine kinase inhibitor resistance, PI3K-Akt signaling pathway, etc. (fig. 5f).
PPL is correlated with immune infiltration level in OV
More importantly, we further assessed the underlying relationships of the mutants of PPL with immune infiltrates in OV microenvironment, a significant correlation was observed between PPL CNV and infiltrating levels of neutrophils and macrophages cells (fig. 6a). TIMER database results indicated a significant correlation between PPL expression and infiltrating levels of CD4+ T cell macrophages, neutrophils, and dendritic cells (fig.6b). Furthermore, Kaplan-Meier analysis results presented a correlation between poor survival outcomes in OV and lower infiltration levels of Dendritic cell (fig. 6c). The affection of PPL expression on various immune cells including neutrophil cell, dendritic cell, CD4+T cell, macrophage, CD8+T cell and B cell was evaluated using multivariable hazards models based on TIMER database (Table 3). CD4+T cell, Macrophage cell, and PPL expression were correlated with OS, based on Cox results . From above all, we surmised that PPL may have an impact on patients’ survival through immune infiltration interaction in OV.