Transcriptional level of PLODs in cancer
Firstly, we analyzed the expression levels of three PLOD genes in different kinds of human cancer using GEPIA2 database. Three PLOD genes showed a relatively up-regulated expression pattern in most of the cancer types(Figure1A,Figure1B and Figure1C). Secondly We use the GEPIA2 database to compare PLODs mRNA expression between PAAD and normal pancreatic tissue. (Figuer. 2A, Figure 2B and Figure 2C). In the GEO database.we compared the mRNA expression of PLODs between PAAD and normal pancreatic tissue., using the GSE16515 dataset and the GSE15471 dataset, respectively. PLODs expression was significantly upregulated in PAAD when compared to normal pancreatic tissue. ( Figure. 2D, Figure 2E and Figure 2F are GSE16515 Figure. 2G, Figure 2H and Figure. 2I are GSE15471).
Transcription level of PLODs in the pancreatic cancer cell line
The expression distribution of PLODs genes in different cell lines of pancreatic cancer is shown in the CCLE database. The horizontal axis in the figure represents the expression of the gene, the longitudinal coordinates are different cell lines, the size of the circular dot represents the expression volume, and different colors also represent the expression volume.Figure 3A shows the high expression of PLOD1 in Panc02.13, PANC-1, and BxPC-3 in pancreatic cancer cancer cell lines. Figure 2B shows the high expression of PLOD2 in PANC-1, Panc02.13, and SU86.86 in pancreatic cancer cell lines. Figure 3C shows the high PLOD3 expression of Panc02.13, M I A P A C a-2, and PANC-1 in pancreatic cancer cell lines.
Protein expression levels of PLOD in the human protein map
To further study the expression of PLOD at the protein levels in pancreatic cancer, we further validate their expression levels using the Human Protein Atlas (HPA) database. PLOD1 is moderately expressed in pancreatic cancer tissue and is not expressed in normal pancreatic tissue (Figure 4A). PLOD2 and PLOD3 were highly expressed in pancreatic cancer and were not expressed in normal pancreatic tissue (Figure 4B, Figure 4C).
Prognostic value of PLODs in patients with PAAD
To explore the prognostic value of PLODs in PAAD patients, we analyzed the Kaplan-Meier Plotter database based on mRNA expression of individual members of the PLODs family. The OS curves of the three PLOD members are demonstrated in Figure5A, respectively.Notably, high transcriptional levels of PLOD1 (P= 0.033) and PLOD2 (P= 0.0062), were markedly associated with shorter OS in PAAD patients.The prognostic roles of differentially expressed PLOD members in the RFS of PAAD patients were also explored. It was found that high transcriptional levels of PLOD1 (P= 0.005) and PLOD2 (P=0.00086) were remarkably associated with shorter RFS in PAAD patients. High expression levels of PLOD3 mRNA were not significantly associated with OS and RFS in patients with PAAD (as shown in Figure 5B). In predicting the outcome of Normal and Tumor, the predictive ability of variable PLOD family genes is more accurate.Based on the TCGA and GTEx datasets, the ROC curve shows area under PLOD1 curve (AUC) of 0.975, confidence interval (CI) of 0.957-0.993, area (AUC) of 0.873 under PLOD2 curve, confidence interval (CI) of 0.832-0.914, area (AUC) of 0.964 under PLOD3 curve, and confidence interval (CI) of 0.943-0.986 (shown in Figure 6). In short, the PLODs family genes have strong sensitivity and specificity in pancreatic cancer.
Baseline table of the relationship between PLOD family genes and clinicopathological features of pancreatic cancer
Expression of PLODS in pancreatic cancer with KRAS, TP53, CDKN2A, SMAD4 mutations
We analyzed the expression level of PLOD family genes in KRAS,TP53,CDKN2A and SMAD4 mutant pancreatic cancer.Compared with pancreatic cancer without mutation, the expression level of PLOD1 in TP53 mutant pancreatic cancer was higher(Figure D). The expression of PLOD2 was increased in KRAS mutant pancreatic carcinoma(Figure B). However, the expression level of PLOD3 in SMAD4 wild type pancreatic cancer is increased(Figure I).
Genetic variation analysis of the PAAD PLODs
The cBioPortal online tool is used to analyze genetic variations in members of the PLOD family in patients with PAAD. As shown in Figure 8A, 43 (29%) PAAD patients exhibited significant alterations in the three PLOD genes, including amplification, deep deletion, truncating mutation,missensemutation , and transcriptional upregulation.Specifically, the percentage changes in the genetic alterations of PLOD1, PLOD2, and PLOD3 among PAAD patients were 16, 10, and 10%, respectively (Figure 8B).
PLOD family member expression was associated with PAAD immune infiltration levels
To explore the immune microenvironment, the relationship of the levels of immune infiltration and the expression of PLODs in PAAD was analyzed by TIMER database. Results show that all members of the PLOD family were not correlated with tumor purity. The level of PLOD1/2 expression was significantly positively associated with the infiltration levels of CD8 + T cells, neutrophils, macrophages, and dendritic cells (Figure 8A,Figure 8B). PLOD3 mRNA expression was negatively associated with the infiltration level of CD8 + T cells (Figure 8C).
Analysis of the genes associated with the PLOD family in PAAD
To further validate the function of PLODs-related molecules in PAAD, we analyzed the mRNA sequencing data of 178 PAAD patients in TCGA using the functional module of the LinkedOmics database. As shown in the volcanic map,the PAAD samples had 3,340, 4,829 and 2,251 genes with significant positive correlation with PLOD1, PLOD2 and PLOD3) (pink and red dots, respectively. While 3338,3803, and 5205 genes showed a significant negative correlation with PLOD1, PLOD2 and PLOD3 in the PAAD sample, respectively (Fig. 10A,, 10B and 10C). To find the molecules associated with PLODs in pancreatic cancer, 121 genes were shown positively correlated with PLOD1, 2 and 3 (Figure 10D) and 237 were negatively correlated with PLOD1, 2, 3 (Figure 10E).We found that RALA is a concurrent regulatory molecule of PLODs in pancreatic cancer (Figure 11A-C).
GO enrichment and KEGG analysis of PLOD1/2/3
To further explore the biological function of these PLODs interaction genes in PAAD, we construct the GO and KEGG pathways using DAVID.Figure 12 Results of positive co-expression gene biological processes (BP) show that these genes are mainly involved in cell division, response to hypoxia, intercell adhesion, and chromosomal separation. For GO cell composition (CC) analysis, the term apparent enrichment is spindle, sticky spots, egg fissure.Significantly enriched molecular function (MF) terms include L-ascorbic acid binding, oxidoretase activity, protocollagen-lysine 5-dioxygenase activity, and microtubule-binding protein.KEGG pathway analysis revealed extracellular matrix − receptor interaction, focal adhesion, and enrichment of cancer pathways.Figure 13 Results of negative coexpression gene biological processes (BP) show that branched chain amino acids, cardiac conduction function regulation, blood pressure regulation, and drug metabolism. For GO cell composition (CC) analysis, the mitochondrial matrix, neuron ocellular cell membrane, intracellular, myogenic fibers. Significantly enriched molecular function (MF) terms include nucleic acid binding, catalytic activity, monooxygenase activity, and hydrolylase activity. KEGG pathway analysis showed the degradation of valine, leucine, iseucine, metabolic pathway, reabsorption of proximal tubulate bicarbonate, and gastric acid secretion.
In conclusion, the results show that PLODs is primarily involved in tumor-related regulatory mechanisms such as local adhesion, adhesion spots, oxidoretase activity, tumor pathways, and lysine degradation.
Functional protein interaction network of the PLOD family genes
We began with the STRING database looking for interacting protein networks of the PLOD family. The enclosure of the network includes 10 functional partners with the highest interactive confidence scores, namely COL5A2, COL5A1, COL1A1, COL1A2, COLGALT1, COL3A1, COL4A1, COL4A2, COL12A1 and COL2A1.The inner shell includes 20 other functional partners, all with interactive confidence scores higher than 0.9. These biological process functions are mainly involved in hydroxylysine biosynthesis processes, negative regulation of post-translational protein modification, collagen synthesis processes, peptidtide-lysine hydroxylation.Molecular functions include primary collagen-lysine 5-dioxygenase activity, extracellular matrix structural components given to tensile strength, and primary collagen galactosyltransferase activity. KEGG pathway analysis showed protein digestion and absorption, ECM-receptor interaction, and adhesion spot enrichment, and lysine degradation. In conclusion, the above results indicate that the PLOD family genes are primarily involved in regulating collagen metabolism and extracellular matrix composition.