The role of PLOD family genes in liver hepatocellular carcinoma: from mechanisms to therapeutic potential.

Introduction
Liver hepatocellular carcinoma (LIHC) is the most common and aggressive form of liver cancer, accounting for approximately 75% of all primary liver cancers globally [1–3]. It ranks as the third leading cause of cancer-related mortality, with its incidence steadily rising in many regions [4, 5]. The poor prognosis associated with LIHC is largely due to late-stage diagnosis, as patients often remain asymptomatic until the disease has reached an advanced stage [6, 7]. Furthermore, treatment options for LIHC are limited, with the current standard of care, including surgical resection, liver transplantation, and systemic therapies like sorafenib failing to provide long-term survival benefits for many patients [8, 9]. This highlights the critical need for novel diagnostic biomarkers, prognostic indicators, and targeted therapies to improve clinical outcomes. The complexity of LIHC arises from its molecular heterogeneity, which includes dysregulated signaling pathways, immune evasion, and alterations in the tumor microenvironment [10, 11]. Identifying specific molecular targets that can improve early detection, predict clinical outcomes, and provide effective therapeutic strategies is essential for advancing the management of this malignancy. Among the genes implicated in cancer progression, the prolyl 4-hydroxylase (PLOD) family has gained attention due to its involvement in collagen biosynthesis and extracellular matrix (ECM) remodeling [12, 13].
The PLOD family consists of three genes, including PLOD1, PLOD2, and PLOD3 that encode prolyl 4-hydroxylases, enzymes responsible for hydroxylating proline residues in collagen and other proteins of the ECM [13]. This post-translational modification stabilizes collagen fibers and facilitates the formation of mature ECM structures, which are essential for maintaining tissue integrity and supporting cellular functions [14]. Recent studies have revealed that PLOD enzymes not only contribute to ECM remodeling but also influence cancer progression through mechanisms such as tumor cell migration, invasion, angiogenesis, and regulation of the hypoxic response [15, 16]. As a result, PLOD enzymes have been implicated in the development and metastasis of several cancers, including breast cancer, gastric cancer, and colorectal cancer, where their overexpression has been linked to poor prognosis and aggressive disease features [17–19]. PLOD1 has been shown to promote tumor cell migration and invasion by stabilizing collagen fibers in the ECM [20]. In various cancers, including non-small cell lung cancer (NSCLC) and breast cancer, overexpression of PLOD1 has been associated with poor prognosis [21, 22]. PLOD2, similarly, plays a significant role in the progression of multiple cancers, including ovarian, gastric, and colorectal cancers [23, 24]. By regulating collagen biosynthesis, PLOD2 affects tumor cell adhesion, migration, and the formation of a supportive stromal environment [24]. PLOD3, though less studied than PLOD1 and PLOD2, has been identified as a key factor in the progression of pancreatic cancer [25]. In pancreatic cancer, PLOD3 expression supports ECM integrity and tumor invasiveness [25]. PLOD3 is also involved in the regulation of the hypoxia-inducible factor (HIF) pathway, which is crucial for tumor adaptation to low-oxygen conditions, thereby contributing to tumor survival and growth in the hostile tumor microenvironment [26, 27].
Despite growing evidence for the involvement of PLOD family genes in cancer biology, their role in LIHC remains inadequately explored. While the expression and function of PLOD genes have been implicated in various cancers, their specific diagnostic, prognostic, and therapeutic significance in LIHC has not been systematically examined. This presents a critical gap in our understanding of LIHC. Given the importance of ECM remodeling in LIHC progression, we hypothesize that PLOD genes may serve as key molecular players in the pathogenesis of LIHC. Their roles in regulating collagen production, modulating tumor cell behavior, and enhancing the invasive potential of tumor cells suggest that these genes may have diagnostic and prognostic value in LIHC, as well as potential as therapeutic targets. In this study, we aim to explore the diagnostic, prognostic, and therapeutic potential of PLOD family genes in LIHC. Using a combination of in silico analysis [28] and in vitro experimental [29, 30] validation, we investigate the correlation between PLOD gene expression and clinical outcomes in LIHC.

Materials and methods

Cell culture
Seven LIHC cell lines (HepG2, Hep3B, Huh7, SNU-182, SNU-449, SNU-387, and PLC/PRF/5) and five normal liver tissue cell lines (THLE-2, LO2, HL-7702, WRL-68, and QSG-7701) were purchased from ATCC (American Type Culture Collection). The cell lines were cultured in RPMI-1640 medium (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (FBS) (Thermo Fisher Scientific), 1% penicillin-streptomycin (Thermo Fisher Scientific), and maintained at 37 °C in a humidified atmosphere with 5% CO₂. The cells were subcultured when they reached 80–90% confluence and were used for experiments at passages between 4 and 10. Short Tandem Repeat (STR) profiling [31] was performed to authenticate all LIHC and normal liver cell lines prior to experimentation. STR analysis confirmed that the genetic profiles of all cell lines matched their respective reference profiles available in the ATCC and DSMZ databases, with no evidence of cross-contamination or misidentification. Only authenticated and mycoplasma-free cell lines were used for subsequent assays.

RT-qPCR for gene expression analysis
Total RNA was extracted from all cell lines using the PureLink™ RNA Mini Kit (Thermo Fisher Scientific) following the manufacturer’s instructions [32, 33]. The RNA was quantified using a Nanodrop spectrophotometer (Thermo Fisher Scientific) to ensure purity and concentration. Reverse transcription of 1 µg of RNA was performed using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. RT-qPCR was performed using the PowerUp™ SYBR™ Green Master Mix (Thermo Fisher Scientific). The reaction was run in a StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific) The relative gene expression was calculated using the ΔΔCT method with GAPDH as the housekeeping gene. The primer sequences used for PLOD1, PLOD2, PLOD3, and GAPDH were as follows:

GAPDH-F 5'-ACCCACTCCTCCACCTTTGAC-3',

GAPDH-R 5'-CTGTTGCTGTAGCCAAATTCG-3'

PLOD1-F: 5'-GCCGTTTGTGTCCCTGTTCTTC-3'

PLOD1-R: 5'-ATGCTGTGCCAGGAACTCTTCC-3'

PLOD2-F: 5'-GACAGCGTTCTCTTCGTCCTCA-3'

PLOD2-R: 5'-CTCCAGCCTTTTCGTGGTGACT-3'

PLOD3-F: 5'-CGAGTGTGAGTTCTACTTCAGCC-3'

PLOD3-R: 5'-CCAGAAGTTGGACCACAGCTTG-3'

Gene expression validation, pathway, survival, and enrichment analyses of PLOD genes in LIHC using GSCA database
The GSCA database (http://bioinfo.life.hust.edu.cn/GSCA/) [34, 35] was used to analyze the expression levels, clinical significance, and involvement of PLOD1, PLOD2, and PLOD3 in LIHC. Gene expression data for these genes were retrieved from the TCGA LIHC cohort to compare their levels between LIHC tumors and normal liver tissues. The database was also used to examine the correlation between PLOD gene expression and different pathological stages of LIHC. Oncogenic pathway analysis was conducted to evaluate the role of PLOD genes in key processes. Additionally, Kaplan-Meier survival analysis was performed to assess the relationship between PLOD expression and patient survival outcomes. Gene Set Enrichment Analysis (GSEA) [36, 37] in GSCA was utilized to investigate the enrichment of PLOD genes in relevant biological pathways in LIHC.

Promoter methylation analysis of PLOD genes using GSCA and UALCAN databases
The methylation status of PLOD1, PLOD2, and PLOD3 in LIHC and normal tissues was examined using the GSCA database [34]. Additionally, Spearman correlation analysis was performed via the GSCA database to assess the relationship between the methylation and expression levels of the PLOD genes. Moreover, promoter methylation levels of PLOD genes were further validated using the UALCAN database (https://ualcan.path.uab.edu/) [38]. This database provides detailed methylation data across multiple cancer types.

Mutational and copy number variations (CNV) analysis of PLOD genes using cBioPortal database
Mutational analysis of PLOD1, PLOD2, and PLOD3 was conducted using the cBioPortal database (http://www.cbioportal.org/) [39]. This database provides extensive data on genetic alterations, including mutations and CNVs, across various cancer types. For our study, we utilized the database to evaluate the frequency and types of mutations in PLOD genes in LIHC samples. CNV analysis was also carried out using the BioPortal database to assess the presence of heterozygous and homozygous amplifications or deletions of PLOD1, PLOD2, and PLOD3.

Prognostic significance of PLOD genes using KM plotter tool
The KM plotter tool (http://kmplot.com/analysis/) [40, 41] was used to assess the prognostic significance of PLOD1, PLOD2, and PLOD3 in terms of overall survival (OS) and relapse-free survival (RFS) in LIHC patients. This tool integrates gene expression data with clinical outcomes, providing Kaplan-Meier survival curves for various genes across multiple cancers, including LIHC. For our study, expression levels of PLOD1, PLOD2, and PLOD3 were correlated with OS and RFS data from LIHC patients. The tool calculates hazard ratios (HRs) along with 95% confidence intervals (CIs) and log-rank p-values to determine the statistical significance of gene expression in relation to survival outcomes.

miRNA-mRNA network analysis
A miRNA-mRNA regulatory network was constructed using the miRNET database (http://www.mirnet.ca/) [42], which allows for the analysis of miRNA-target interactions and their regulatory roles in cancer. This database was used to identify potential miRNAs that regulate the expression of PLOD genes in LIHC. Based on predicted interactions, three specific miRNAs (hsa-miR-503-5p, hsa-miR-195-5p, and hsa-miR-193a-3p) were selected for further investigation. These miRNAs were chosen due to their potential regulatory effects on PLOD1, PLOD2, and PLOD3 expression in liver cancer. The expression levels of these miRNAs were analyzed using the UALCAN database (http://ualcan.path.uab.edu/) [38], which provides access to expression data from The Cancer Genome Atlas (TCGA). UALCAN was used to compare the expression of hsa-miR-503-5p, hsa-miR-195-5p, and hsa-miR-193a-3p between LIHC tissues and normal tissues. This database was selected for its user-friendly interface and its ability to access and analyze large-scale data sets related to gene and miRNA expression across multiple cancer types, including LIHC.
To further validate the expression of these miRNAs, RT-qPCR was performed on seven LIHC and five normal control cell lines. For RT-qPCR, the TaqMan™ Universal PCR Master Mix, No AmpErase® UNG (Thermo Fisher Scientific) was used. The probes for hsa-miR-503-5p, hsa-miR-195-5p, and hsa-miR-193a-3p were obtained from the TaqMan™ MicroRNA Assays (Thermo Fisher Scientific). The qPCR reactions were performed on a StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific). Relative miRNA expression levels were calculated using the ΔΔCT method, with U6 as the internal reference gene.

Correlation analysis between PLOD gene expression, immune infiltrates, and drug sensitivity using GSCA, CTRP, and GDSC databases
The GSCA [34, 43] was used to analyze the correlation between PLOD gene expression and various immune cell types in LIHC. Spearman’s correlation analysis was performed to assess the relationship between PLOD1, PLOD2, and PLOD3 expression and immune cell infiltrates. For drug sensitivity analysis, the CTRP (http://www.broadinstitute.org/ctrp/) [44] and GDSC (https://www.cancerrxgene.org/) [45] databases were used to examine the correlation between PLOD gene expression and the sensitivity of LIHC cell lines to various therapeutic agents. Spearman’s correlation was applied to evaluate the relationship between PLOD expression and drug resistance.

PLOD2/3 gene knockdown
siRNAs targeting PLOD2 (Thermo Fisher Scientific Silencer™ Select Human PLOD2 Assay ID: HSS108124) and PLOD3 (Thermo Fisher Scientific Silencer™ Select Human PLOD3 Assay ID: s17135) were used. Transfection was performed according to manufacturer’s protocol with 50 nM final concentration using Lipofectamine™ RNAiMAX (Thermo Fisher Scientific). Knockdown efficiency was confirmed by RT-qPCR and Western blot. After 48 h, the knockdown efficiency was confirmed by RT-qPCR and Western blotting.

Western blotting
Proteins were extracted from cells using RIPA Lysis and Extraction Buffer (Thermo Fisher Scientific) supplemented with protease and phosphatase inhibitors (Thermo Fisher Scientific). Protein concentrations were determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific). Equal amounts of total protein (30 µg per lane) were separated by SDS-PAGE and transferred onto PVDF membranes. Membranes were blocked with 5% non-fat dry milk prepared in TBS containing 0.1% Tween-20 (TBS-T) for 1 h at room temperature. Primary antibodies used included PLOD2 (Thermo Fisher Scientific, Cat. No. 21214-1-AP), PLOD3 (Thermo Fisher Scientific, Cat. No. PA5-51546), E-cadherin (Thermo Fisher Scientific, Cat. No. MA5-14458), Vimentin (Thermo Fisher Scientific, Cat. No. MA5-11883), and GAPDH (Thermo Fisher Scientific, Cat. No. MA5-15738). Membranes were incubated with primary antibodies overnight at 4 °C, followed by incubation with HRP-conjugated secondary antibodies for 1 h at room temperature. Protein bands were visualized using the Pierce™ ECL Western Blotting Substrate (Thermo Fisher Scientific, Cat. No. 32209) and detected using a chemiluminescent imaging system. GAPDH was used as the internal loading control.

Cell proliferation (CCK8 assay)
Cell proliferation was assessed using the CellTiter 96® AQueous One Solution Cell Proliferation Assay (Thermo Fisher Scientific). Cells were seeded in 96-well plates at a density of 5,000 cells per well and incubated overnight. After treatment, 20 µL of CCK8 reagent was added, and cells were incubated for 2 h. Absorbance was measured at 450 nm using a microplate reader (BioTek Instruments).

Colony formation assay
For colony formation, 500 cells were seeded in 6-well plates and allowed to adhere overnight. After transfection, cells were cultured for 10–14 days, with medium replenished every 3–4 days. Colonies were fixed with 4% paraformaldehyde (Thermo Fisher Scientific) and stained with crystal violet solution (Thermo Fisher Scientific) for 30 min.

Wound healing assay
To assess cell migration, a wound healing assay was performed. Cells were seeded to confluence in 6-well plates, and a scratch was made using a 200 µL pipette tip. After washing with PBS, cells were incubated in serum-free RPMI-1640 medium (Thermo Fisher Scientific). Images were taken at 0 and 24 h using a light microscope (Leica) to observe the closure of the wound.

Diagnostic validation of PLOD gene expression in a Chinese LIHC cohort
To validate the diagnostic relevance of PLOD family genes in an independent Chinese cohort, we utilized the Gene Expression Omnibus (GEO) dataset GSE84402 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE84402), which contains paired 14 paired LIHC tissues and adjacent non-tumor liver samples collected from Chinese patients. The microarray expression matrix generated on the Illumina HumanHT-12 v4.0 platform was downloaded, quantile-normalized, and provided as log₂-normalized signal intensities. Probes were annotated to official gene symbols based on the manufacturer’s annotation file, and when multiple probes mapped to the same gene, their average expression value was used. Expression levels of PLOD1, PLOD2, and PLOD3 were extracted, and differential expression between tumour and adjacent normal tissues was assessed using Student’s t-test. Diagnostic performance was evaluated by constructing receiver operating characteristic (ROC) curves and calculating the area under the curve (AUC) to quantify discriminative ability. All analyses were performed using Python (v3.10) with the pandas, numpy, scikit-learn, and matplotlib packages.

Statistical analysis
Data were presented as mean ± standard deviation (SD). All experiments were performed in triplicate. For comparisons between two groups, statistical significance was determined using the student’s t-test. For multiple comparisons, one-way ANOVA followed by Tukey’s post hoc test was applied. Correlations between gene expression and immune infiltrates, drug sensitivity, and other clinical factors were assessed using Spearman’s correlation coefficient. Survival analyses were performed using Kaplan-Meier curves, and differences between groups were evaluated using the log-rank test. P-values less than 0.05 were considered statistically significant. All statistical analyses were performed using GraphPad Prism (version 9.5.1) software.

Results

Expression analysis and diagnostic performance of PLOD genes
The expression of PLOD1, PLOD2, and PLOD3 was analyzed in LIHC and normal control cell lines using RT-qPCR across seven LIHC and five normal cell lines. PLOD1, PLOD2, and PLOD3 exhibited significantly (p-value < 0.05) higher expression in LIHC cell lines compared to normal controls, with a distinct separation in median expression levels between the two groups (Fig. 1A). Next, based on the RT-qPCR data, the ROC curve analysis was performed to evaluate the diagnostic potential of each PLOD gene. PLOD3 demonstrated the highest AUC (0.8), indicating its strongest potential as a diagnostic marker (Fig. 1B). PLOD1 (0.75) and PLOD2 (0.7) also showed moderate diagnostic accuracy, suggesting moderate discriminatory power (Fig. 1B).

Expression validation and clinical significance of PLOD genes
The expression levels of PLOD were validated in the TCGA LIHC cohort using the GSCA database. Results indicated significant (p-value < 0.05) upregulation of all three PLOD genes in LIHC tumors compared to normal tissues (Fig. 2A). Further analysis across different pathological stages of LIHC revealed that PLOD2 and PLOD3 genes showed progressively higher expression from stage I to IV (Fig. 2B). While PLOD1 showed opposite results (Fig. 2B). The heatmap demonstrated that PLOD1, PLOD2, and PLOD3 were increasingly expressed in advanced stages of LIHC, suggesting their involvement in tumor progression and disease severity. Furthermore, the role of PLOD genes in the activation and deactivation of oncogenic pathways was analyzed using the GSCA database. PLOD1, PLOD2, and PLOD3 were found to significantly (p-value < 0.05) activate key oncogenic pathways associated with EMT, which is crucial for tumor metastasis (Fig. 2C). Moreover, the overall survival (OS) analysis showed that higher expression of PLOD1, PLOD2, and PLOD3 was significantly (p-value < 0.05) associated with poorer survival outcomes in LIHC patients (Fig. 2D). GSEA analysis via the GSCA database revealed that up-regulated PLOD1, PLOD2, and PLOD3 genes were significantly (p-value < 0.05) enriched in LIHC (Fig. 2E).

Promoter methylation analysis of PLOD genes
The methylation status of PLOD1, PLOD2, and PLOD3 in LIHC and normal tissues was examined using the GSCA database. Significant (p-value < 0.05) differential methylation was observed across all three genes, with tumors showing significantly lower methylation compared to normal tissues. PLOD1 exhibited a p-value of 2.22e − 16, PLOD2 showed a p-value of 2.22e − 16, and PLOD3 had a p-value of 2.22e − 16, all indicating strong statistical significance (Fig. 3A). Moreover, the Spearman correlation analysis was conducted to assess the relationship between methylation and expression levels of the PLOD genes via the GSCA database. A negative correlation was observed between PLOD1, PLOD2, and PLOD3 methylation and expression in LIHC (Fig. 3B). To further validate these findings, promoter methylation levels of PLOD genes were analyzed using the UALCAN database. Consistent with GSCA results, PLOD1, PLOD2, and PLOD3 exhibited significantly (p-value < 0.05) lower promoter methylation levels (hypomethylation) in LIHC tissue samples compared to normal liver tissues (Fig. 3C).

Mutational and CNV analysis of PLOD genes
Mutational analysis was conducted using the cBioPortal database to evaluate the frequency and type of mutations in PLOD genes across LIHC samples. PLOD1, PLOD2, and PLOD3 exhibited mutations in 17%, 33%, and 50% of samples, respectively (Fig. 4A). The majority of these mutations were missense mutations, as shown in the classification summary (Fig. 4B). The somatic mutation rates for PLOD1, PLOD2, and PLOD3 were reported as 0.27%, 0.82%, and 0.55%, respectively (Fig. 4C). The analysis of heterozygous CNVs in LIHC revealed amplification in PLOD1 and PLOD3, as shown in Fig. 4D. PLOD1 exhibited the highest CNV percentage at 37%, indicating significant copy number alterations in these genes (Fig. 4D). In contrast, PLOD2 showed lower CNV values (Fig. 4D). For homozygous CNVs, PLOD1 and PLOD3 exhibited amplification, while deletions were less common (Fig. 4E).

Prognostic significance of PLOD genes
The OS and relapse-free survival (RFS) of LIHC patients were assessed based on the expression levels of PLOD1, PLOD2, and PLOD3 using the KM plotter tool. Figure 5A shows the OS analysis for PLOD1, PLOD2, and PLOD3. High expression levels of all three PLOD genes were significantly (p-value < 0.05) associated with poorer OS (Fig. 5A). Specifically, high PLOD1 expression resulted in a hazard ratio (HR) of 1.88 (95% CI: 1.33–2.65, log-rank p = 0.00028), high PLOD2 expression showed an HR of 2.34 (95% CI: 1.62–3.38, log-rank p = 3.3e-06), and high PLOD3 expression had an HR of 1.65 (95% CI: 1.15–2.37, log-rank p = 0.0061) (Fig. 5A). Figure 5B presents the RFS analysis for PLOD1, PLOD2, and PLOD3. Similarly, higher expression of PLOD1, PLOD2, and PLOD3 was associated with a worse RFS. The HR for PLOD1 was 1.71 (95% CI: 1.22–2.41, log-rank p = 0.0017), for PLOD2 it was 1.76 (95% CI: 1.25–2.49, log-rank p = 0.0011), and for PLOD3 it was 1.88 (95% CI: 1.32–2.67, log-rank p = 0.00033) (Fig. 5B).

miRNA-mRNA network analysis
A miRNA-mRNA regulatory network was constructed using the miRNET database. This analysis revealed that three specific miRNAs (hsa-miR-503-5p, hsa-miR-195-5p, and hsa-miR-193a-3p) simultaneously regulate the expression of PLOD genes in LIHC (Fig. 6A). These miRNAs were selected for further investigation due to their potential involvement in the regulation of PLOD gene expression in liver cancer. The expression levels of these miRNAs were analyzed using the UALCAN database (Fig. 6B). The results indicated that hsa-miR-195-5p and hsa-miR-193a-3p were significantly (p-value < 0.05) upregulated, while hsa-miR-503-5p was significantly (p-value < 0.05) down-regulated in LIHC tissue compared to normal tissues (Fig. 6B). Moreover, Boxplots in Fig. 6C show the relative expression levels of hsa-miR-503-5p, hsa-miR-195-5p, and hsa-miR-193a-3p across LIHC and normal control cell lines, analyzed using RT-qPCR. miRNAs including, hsa-miR-195-5p and hsa-miR-193a-3p were significantly (p-value < 0.05) upregulated, while hsa-miR-503-5p was significantly (p-value < 0.05) down-regulated in LIHC cell lines compared to normal controls (Fig. 6C). Next, the AUC values for the miRNAs were as follows: hsa-miR-503-5p (AUC = 0.94), hsa-miR-195-5p (AUC = 1), and hsa-miR-193a-3p (AUC = 0.97) (Fig. 6D). Kaplan-Meier survival curves were generated using UALCAN database to assess the effect of these miRNAs on patient survival. Aberrantly expressed hsa-miR-503-5p, hsa-miR-195-5p and hsa-miR-193a-3p were linked to poor prognosis, with p-values of > 0.05 (Fig. 6E), suggesting that these miRNAs may be weak prognostic biomarkers in LIHC.

Correlation between PLOD gene expression, immune infiltrates, and drug sensitivity in LIHC
The correlation between PLOD gene expression and various immune cell types was analyzed using the GSCA database. PLOD1 and PLOD2 expression showed significant (p-value < 0.05) negative correlations with immune infiltrates, particularly with CD4_T, CD4_naive, and Tfh cells (Fig. 7A). Notably, PLOD2 showed weak correlations with immune infiltrates, suggesting its less pronounced involvement in the immune response in LIHC (Fig. 7A). Figure 7B provides a detailed heatmap of the correlation between PLOD gene expression and immune cell types in LIHC. Significant (p-value < 0.05) correlations were observed between PLOD expression and various immune cells, including gamma-delta T cells and CD4 + naive T cells, with PLOD3 showing the strongest correlation across multiple immune cell types (Fig. 7B). Furthermore, Fig. 7C-D presents the correlation between PLOD gene expression and drug sensitivity based on the CTRP and GDSC databases. PLOD1, PLOD2, and PLOD3 exhibited significant (p-value < 0.05) positive correlations with almost all drug (such as AR-42 and GW-441756 s), highlighting drug resistance against these drugs (Fig. 7C-D).

PLOD2/3 knockdown, cell proliferation, colony formation, and wound healing assays
To investigate the functional role of PLOD genes in LIHC, we performed gene knockdown and assessed the impact on cellular behaviors in both HepG2 and Hep3B cell lines. The expression of PLOD2, and PLOD3 was effectively silenced using siRNA, as confirmed by both mRNA (Figs. 8A and 9A) and protein expression analysis (Figs. 8B and 9B, Supplementary data Fig. 1 [replicate 2], and Supplementary data Table 1). The knockdown of PLOD2 and PLOD3 led to a significant (p***-value < 0.001) reduction in cellular proliferation, as observed in the proliferation assays for both transfected HepG2 (Fig. 8C) and Hep3B (Fig. 9C) cells. Specifically, the proliferation rate of PLOD2 and PLOD3 knockdown cells was significantly (p***-value < 0.001) decreased to approximately 50% of the control groups, indicating that these genes play a crucial role in promoting cell growth (Figs. 8C and 9C). In colony formation assays, both PLOD2 and PLOD3 knockdown resulted in a marked (p***-value < 0.001) reduction in the number of colonies formed by HepG2 and Hep3B cells (Figs. 8-D-E and 9D-E). This decrease in colony formation further supports the conclusion that PLOD2 and PLOD3 are involved in the tumorigenic potential of LIHC cells. The wound healing assays demonstrated that PLOD2 and PLOD3 knockdown significantly (p***-value < 0.001) impaired the ability of HepG2 and Hep3B cells to migrate, as shown by slower wound closure in the knockdown groups (Figs. 8F-G and 9F-G). This finding suggests that PLOD2 and PLOD3 are involved in the migratory potential of LIHC cells, which is a key feature of cancer metastasis.

Proposed pathophysiological model and validation of PLOD gene–mediated EMT pathway in LIHC
The upregulation of PLOD genes (PLOD1, PLOD2, and PLOD3) in LIHC is proposed to promote tumor progression through the activation of the EMT pathway (Fig. 10A). Elevated PLOD expression enhances collagen hydroxylation and ECM remodeling, leading to increased matrix stiffness and disruption of epithelial integrity. These microenvironmental changes trigger EMT, characterized by the loss of epithelial markers such as E-cadherin and induction of mesenchymal markers such as Vimentin, facilitating increased motility, invasiveness, and acquisition of stem-like traits. As depicted in the model, this PLOD–ECM remodeling–EMT axis promotes enhanced cellular plasticity, aggressive migratory behavior, and metastatic dissemination, ultimately contributing to tumor progression and poor prognosis in LIHC patients (Fig. 10A). To experimentally validate this model, we analyzed EMT-related protein markers following PLOD1/2 gene knockdown in HepG2 cells. Western blot analysis revealed that siRNA-mediated silencing of PLOD2/3 markedly altered EMT marker expression. Compared with control cells, E-cadherin expression markedly increased, whereas Vimentin expression decreased, indicating suppression of EMT activation (Fig. 10B, Supplementary data Fig. 1, and Supplementary data Table 2). GAPDH levels remained stable, confirming equal protein loading. Collectively, these results substantiate that PLOD upregulation activates the EMT pathway, while its knockdown reverses this process, highlighting a crucial mechanistic link between PLOD gene activity, EMT regulation, and tumor aggressiveness in LIHC.

Diagnostic validation of PLOD genes in a Chinese LIHC cohort
Validation using the Chinese LIHC dataset GSE84402 demonstrated that all three PLOD genes were significantly upregulated in tumor tissues compared with matched adjacent non-tumor liver samples. PLOD1, PLOD2, and PLOD3 showed markedly higher expression in LIHC (all p < 0.5), consistent with findings from TCGA and cell line analyses. Among the three members, PLOD3 exhibited the highest diagnostic accuracy, with an AUC of 1.0 (Fig. 11A), indicating strong discriminatory ability between tumor and normal tissues. PLOD1 and PLOD2 also showed robust diagnostic performance, with AUC values of 0.98 and 0.99, respectively (Fig. 11B). ROC curve analysis demonstrated clear separation between tumor and normal samples across all genes. Overall, these results confirm that PLOD1, PLOD2, and particularly PLOD3, are consistently overexpressed in Chinese LIHC patients, supporting their potential utility as diagnostic biomarkers across independent cohorts.

Discussion
Our study provides a comprehensive analysis of the PLOD family genes (PLOD1, PLOD2, and PLOD3) in liver hepatocellular carcinoma (LIHC) [46, 47], examining their expression profiles, clinical significance, genetic alterations, and functional roles. We compare and contrast our findings with existing literature to highlight novel insights into the involvement of these genes in LIHC.
Consistent with previous studies, we observed significant upregulation of PLOD1, PLOD2, and PLOD3 in LIHC cell lines compared to normal controls. This aligns with findings in other cancers, such as lung adenocarcinoma (LUAD), where elevated expression of PLOD genes was associated with poor prognosis [17, 48, 49]. Our ROC curve analysis further supports the diagnostic potential of these genes, with PLOD3 demonstrating the highest AUC of 0.8, indicating its strong discriminatory power. This finding is in agreement with studies suggesting the utility of PLOD3 as a diagnostic biomarker in various cancers [17, 23].
Our survival analyses revealed that high expression levels of PLOD1, PLOD2, and PLOD3 are associated with poorer OS and (RFS in LIHC patients. These results corroborate earlier studies by Yang et al., who found that elevated expression of PLOD1 and PLOD3 was linked to shorter OS and disease-free survival in HCC patients [50]. Notably, our study extends these findings by also highlighting the prognostic significance of PLOD2 in LIHC, a gene that has been less extensively studied in this context. Our mutational and CNV analyses revealed that PLOD1, PLOD2, and PLOD3 exhibit various genetic alterations in LIHC, including amplifications and mutations. These alterations may contribute to the dysregulated expression of these genes in tumors. Additionally, our promoter methylation analysis indicated that lower methylation levels are associated with higher expression of PLOD genes, suggesting epigenetic regulation. These findings are consistent with previous research by Yang et al., who reported genetic alterations and methylation changes in PLOD genes in HCC [50].
In our miRNA-mRNA network analysis, we identified three miRNAs (hsa-miR-503-5p, hsa-miR-195-5p, and hsa-miR-193a-3p) that regulate the expression of PLOD genes in LIHC. This is a novel aspect of our work, as the role of these specific miRNAs in regulating PLOD genes has not been extensively studied in the context of LIHC. Previous studies have suggested that miRNAs such as hsa-miR-195-5p regulate key oncogenic pathways in various cancers, including breast and gastric cancer [51, 52], but their direct interaction with PLOD genes had not been established. We found that hsa-miR-195-5p and hsa-miR-193a-3p were significantly upregulated, while hsa-miR-503-5p was downregulated in LIHC tissues and cell lines. The AUC values of these miRNAs suggest that they may serve as reliable diagnostic biomarkers for LIHC, particularly hsa-miR-195-5p, which exhibited a perfect AUC of 1. This result is a significant contribution to understanding the regulatory networks controlling PLOD expression and opens up potential therapeutic avenues for targeting these miRNAs to modulate PLOD gene activity.
Functional assays demonstrated that knockdown of PLOD2 and PLOD3 significantly reduced cell proliferation, colony formation, and migration in HepG2 and Hep3B cells. These results align with studies by Li et al., who reported that PLOD2 enhances HCC metastasis via BIRC3, and by Ding et al., who found that PLOD3 promotes liver metastasis in colorectal cancer [18]. Our work extends these findings by providing evidence that PLOD2 and PLOD3 are essential for the proliferative and migratory abilities of liver cancer cells.
The correlation between PLOD gene expression and immune cell infiltrates in LIHC is an area of increasing interest [49]. Our findings suggest that PLOD1 and PLOD2 expression negatively correlates with several immune cell types, including CD4 + T cells and Tfh cells, while PLOD3 shows stronger associations with immune infiltration. This is consistent with studies that highlight the complex interplay between cancer cells and the immune microenvironment [53–55]. In particular, PLOD expression may influence immune cell recruitment and function, contributing to immune evasion in LIHC. Additionally, the significant positive correlation between PLOD genes and drug resistance, observed in both the CTRP and GDSC databases, suggests that PLOD genes may be involved in mediating resistance to chemotherapy and targeted therapies. Similar findings have been reported in other cancers, where PLOD expression is associated with resistance to various therapeutic agents [56, 57].
This study provides several novel insights into the role of PLOD genes in LIHC. First, we present comprehensive data on the expression, genetic alterations, methylation status, and prognostic significance of PLOD1, PLOD2, and PLOD3 in LIHC. Second, our miRNA-mRNA network analysis highlights previously unexplored regulatory mechanisms that control PLOD gene expression. Third, we establish a potential link between PLOD gene expression and immune microenvironment modulation, as well as drug resistance, which opens new avenues for therapeutic targeting. Taken together, our findings not only deepen our understanding of the molecular mechanisms driving LIHC but also provide a foundation for the development of novel diagnostic and therapeutic strategies targeting PLOD genes.

Conclusion
In conclusion, our study underscores the critical roles of PLOD1, PLOD2, and PLOD3 in the progression and prognosis of LIHC. These genes may serve as valuable biomarkers for diagnosis and prognosis and potential therapeutic targets in LIHC treatment. Future studies should focus on validating miRNA-PLOD interactions, developing targeted therapies for PLOD inhibition, and conducting large-scale clinical trials to assess their prognostic value and therapeutic potential in LIHC.

Supplementary Information