The value of lipid metabolism-related genes in pancreatic cancer immunotherapy and drug prediction

Introduction

Pancreatic cancer holds a significant position in the global cancer burden. According to the latest epidemiological data, this disease has become the 12th most common malignant tumor worldwide and the seventh leading cause of cancer-related deaths (1). Among histological types, pancreatic ductal adenocarcinoma (PDAC) accounts for over 90% of cases, with its molecular heterogeneity and stromal fibrosis characteristics further increasing treatment challenges (2-4). The pathogenesis of pancreatic cancer is complex, resulting from the combined effects of multiple factors including genetic, environmental, and lifestyle components. Current research typically categorizes risk factors into: non-modifiable factors: including genetic susceptibility, age, and gender; intervenable lifestyle factors: including smoking, obesity and metabolic syndrome, diabetes, alcohol consumption and diet; and disease-related factors: including chronic pancreatitis and Helicobacter pylori infection, as well as environmental and occupational exposures.

Disorders of lipid metabolism play a crucial role in tumor cell proliferation, survival, invasion, metastasis, and the formation of the tumor microenvironment. These mutations can reprogram lipid metabolism, altering fatty acid uptake, oxidation, and synthesis to provide membrane structural components and energy for tumors, thereby promoting tumor growth and progression (5-8). In recent years, with the advancement of multi-omics technologies, researchers have employed multidisciplinary approaches such as transcriptomics, lipidomics, and pathology to investigate alterations in lipid metabolic pathways in pancreatic cancer and their potential relationship with tumorigenesis and progression. For example, abnormalities in the cholesterol synthesis pathway mediated by squalene epoxidase (SQLE) can promote tumor cell proliferation while enhancing signal transduction of receptor tyrosine kinases such as the epidermal growth factor receptor (EGFR) through the formation of lipid rafts (9,10). Additionally, the hypoxic microenvironment induces hypoxia-inducible factor 1-alpha (HIF-1α)-mediated upregulation of the fatty acid oxidation-related enzyme carnitine palmitoyltransferase 1A (CPT1A), promoting tumor metastasis (11,12). The DNA demethylase ten-eleven translocation 3 (TET3) was identified as a key regulator of lipogenic metabolism in PDAC, and its elevated expression correlates with poor patient prognosis (13). Current research not only elucidates potential mechanisms of lipid metabolic reprogramming in pancreatic cancer but, more importantly, provides breakthroughs for clinical translation. Possible strategies include: (I) inhibiting fatty acid synthesis: fatty acid synthase inhibitors can reduce lipid droplet accumulation and induce apoptosis (14,15). (II) Modulating cholesterol metabolism: statins, which specifically inhibit 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR), lower cholesterol levels and exhibit synergistic effects with chemotherapy in preclinical models (9,16). A novel lipid-lowering agent bempedoic acid markedly enhances the suppression of de novo lipogenesis by gemcitabine plus nab-paclitaxel in pancreatic cancer cells (17). (III) Targeting lipid signaling pathways: antibodies against platelet-derived growth factor C (PDGFC) can block its induced lipid metabolic reprogramming and suppress PDAC progression (18). (IV) Combination immunotherapy: blocking CD36-mediated fatty acid uptake can reverse T-cell exhaustion and enhance immunotherapy efficacy (19,20).

Nevertheless, the specific molecular mechanisms between lipid metabolism and pancreatic cancer remain incompletely understood. Future studies are warranted to identify which lipid metabolism-associated genes and pathways influence pancreatic cancer tumorigenesis and progression, and how these genes may be leveraged to develop novel diagnostic markers and therapeutic targets.

This study aims to systematically explore the mechanism of lipid metabolism-related genes (LMRGs) in pancreatic cancer through comprehensive bioinformatics analysis, establish a prognostic scoring model and verify it, and explore potential avenues for clinical application. We present this article in accordance with the TRIPOD reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-507/rc).

Methods

Data preparation

The messenger ribonucleic acid (mRNA) expression profiles and corresponding clinical information of pancreatic cancer tissues were obtained from the public database The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/), while mRNA expression data of healthy pancreatic tissues were acquired from the Genotype-Tissue Expression (GTEx, https://gtexportal.org/home/) database as the training set. The batch effects in mRNA expression data between the tumor group and control group of the training set were removed using the limma package. For the test set, standardized mRNA expression profiles and clinical data of pancreatic cancer PACA-CA were selected from the International Cancer Genome Consortium (ICGC, https://dcc.icgc.org/) database to validate the prognostic model. Cases with incomplete data were excluded.

LMRGs were retrieved from the Gene Set Enrichment Analysis (GSEA) database (https://www.gsea-msigdb.org/) using search terms “metabolic” or “metabolism” and “lipid”. After reviewing gene set descriptions, we screened human studies related to lipid metabolism, lipid metabolism regulation, or lipid transport pathways. The filtered gene sets were merged and duplicate genes were removed to obtain the final gene list for subsequent analysis.

Construction and validation of prognostic LMRGs

In the training set, differential expression analysis of mRNA data between tumor and control groups was conducted using the threshold criteria of |log₂ fold change (FC)| >2 and P<0.05, thereby identifying lipid metabolism-associated differentially expressed genes (DEGs) in pancreatic cancer.

Univariate Cox regression analysis was employed to further screen these DEGs for their prognostic value. A P value <0.05 was considered statistically significant. To refine the set of lipid metabolism-related DEGs and enhance model stability, the least absolute shrinkage and selection operator (LASSO) Cox regression method was applied, using 10-fold cross-validation to optimize the λ value and reduce overfitting. Using the identified genes and their corresponding coefficients, a prognostic prediction formula for pancreatic cancer was established: Risk Score = Σ (Coefficient × Each Gene’s Expression). Patients were stratified into high-risk and low-risk groups based on the median risk score. The Kaplan-Meier method was used for survival analysis to compare prognosis between the two groups. The prognostic power of the risk model was evaluated using time-dependent receiver operating characteristic (ROC) curves, generated with the survival ROC R package. The PACA-CA dataset served as the validation set to further assess the predictive performance of the model.

Additionally, immunohistochemical (IHC) results from the Human Protein Atlas (HPA) database were utilized to further evaluate the differential expression of the gene signatures.

The correlation between risk scores and clinical variables

The Wilcoxon signed-rank test was employed to analyze the correlation between the risk score and various clinical characteristics, aiming to clarify the distributional differences of the risk score across distinct clinical subgroups. Subsequently, both univariate and multivariate Cox regression analyses were performed, incorporating the risk score alongside other clinical variables, to investigate whether the risk score holds independent prognostic value for pancreatic cancer patients.

Functional enrichment analysis

To investigate the potential mechanisms of differentially expressed LMRGs, multidimensional enrichment analysis was performed based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) databases. The GO analysis encompassed three dimensions: biological process (BP), reflecting the physiological activities in which genes participate; molecular function (MF), characterizing the activity of gene products; and cellular component (CC), localizing the subcellular distribution of gene products. A P value <1 was considered statistically significant. GSEA was conducted between high-risk and low-risk score groups (FDR <0.25, P<0.05).

Immune cell infiltration analysis

The tumor microenvironment of pancreatic cancer is a crucial factor contributing to the poor efficacy of immunotherapy in pancreatic cancer. Therefore, we employed CIBERSORT, CIBERSORT-ABS, QUANTISEQ, MCP-counter, XCELL, TIMER, and EPIC algorithms to evaluate the differences in immune cell infiltration levels between high-risk and low-risk groups based on LMRGs. Given the pivotal role of immune checkpoint molecules in tumor immunotherapy, this study further analyzed the association between the risk model and 79 immune checkpoint genes reported in the literature (21). The Wilcoxon rank-sum test was used to compare the expression of genes between the high-risk and low-risk groups, with a P value <0.05 considered statistically significant. Additionally, the Tumor Immune Estimation Resource (TIMER) database (https://cistrome.shinyapps.io/timer/) was utilized to analyze the relationship between immune cells and LMRGs, enhancing our understanding of the interplay among LMRGs, immune cells, and pancreatic cancer.

Prediction of candidate drugs

The DSigDB is a global repository for identifying target drug substances associated with DEGs. The DSigDB database in the online platform Enrichr (https://maayanlab.cloud/Enrichr/) was utilized to predict potential drug sensitivity. Six LMRGs were uploaded, and drug prediction results were obtained, with a P value <0.05 considered statistically significant.

This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Statistical analysis

Bioinformatics analyses were performed using RStudio and its corresponding packages. Differential expression analysis was conducted using the limma package. LASSO Cox regression modeling was performed using the glmnet package. Survival analysis was carried out with the survminer package. Differential expression analysis was implemented using the limma package (22). Key gene screening was achieved through LASSO Cox regression modeling with the glmnet package. Survival analysis was performed using the survival and survminer packages for Kaplan-Meier curve plotting. GSEA was conducted using the clusterProfiler package based on gene sets from the MSigDB database to identify significantly enriched pathways. GO enrichment analysis was similarly performed using the clusterProfiler package (23,24). Data visualization was completed using ggplot2.

Results

Identification and functional enrichment analysis of DEGs

The flow chart of the whole analysis is shown in Figure 1. We obtained RNA-seq data and clinical information from 177 pancreatic cancer samples and 167 normal samples from TCGA and GTEx databases, respectively. The TCGA pancreatic cancer data and GTEx normal sample RNA-seq data were merged and batch effects were removed. Based on 234 LMRGs acquired from the GSEA database (Table S1), differential expression analysis (|log₂FC| >2, P<0.05) between tumor and control groups identified 17 DEGs, including 8 upregulated and 9 downregulated genes (Figure 2A,2B).

Figure 1 The flow chart of overall study. DEGs, differentially expressed genes; GO, Gene Ontology; GSEA, Gene Set Enrichment Analysis; GTEx, genotype-tissue expression; ICGC, International Cancer Genome Consortium; KEGG, Kyoto Encyclopedia of Genes and Genomes; KM, Kaplan-Meier; LASSO, least absolute shrinkage and selection operator; RNA, ribonucleic acid; ROC, receiver operating characteristic; TCGA, The Cancer Genome Atlas.

Figure 2 Identification and functional enrichment analysis of lipid metabolism-related DEGs. (A) Volcano plot of DEGs, depicting upregulated (red), downregulated (blue), and non-significant (gray) genes. (B) Heat map of DEGs, showing gene expression levels. (C-F) Chord diagrams illustrating the associations between DEGs and GO in terms of biological processes, cellular components, molecular functions and KEGG pathways. DEGs, differentially expressed genes; FC, fold change; FDR, false discovery rate; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Subsequently, we performed KEGG and GO functional enrichment analyses on these 17 DEGs. The chord diagram illustrates the results of the GO and KEGG enrichment analyses (Figure 2C-2F). The most significantly enriched BPs were lipid transport, regulation of lipid metabolic process, and regulation of inflammatory response, which aligned with our gene screening criteria. For CCs, the most enriched terms were alpha-ketoacid dehydrogenase complex, RNA polymerase II transcription regulator complex, and oxidoreductase complex. Regarding MFs, phospholipase A2 activity, Hsp90 protein binding, nuclear receptor activity, and ligand-activated transcription factor activity showed the strongest enrichment. The GO enrichment results demonstrated that DEGs were associated with lipid transport, lipid metabolism, inflammatory response, and redox reactions. KEGG pathway analysis revealed significant enrichment in arachidonic acid metabolism, linoleic acid metabolism, ether lipid metabolism, and alpha-linolenic acid metabolism.

Establishment and validation of LMRGs

In the training set, univariate Cox regression analysis was performed on 17 DEGs, yielding 9 survival-related DEGs (Figure 3). Subsequently, LASSO Cox regression analysis was conducted using these 9 genes to identify and integrate 6 key LMRGs associated with pancreatic cancer prognosis (Figure S1). A LMRGs risk score was calculated based on their correlation coefficients (Table 1).

Figure 3 Forest plots for the results of univariate regression analysis. CI, confidence interval.

Risk stratification was performed in the training cohort using the median risk score as the cutoff. The risk score distribution and survival status plot (Figure 4A,4B) demonstrated denser clustering of deceased patients in the high-risk group, indicating reliable prognostic utility of LMRGs for pancreatic cancer survival prediction.

Figure 4 Prognostic performance of the 6-gene signature model in TCGA cohort. (A) Distribution of patient survival status stratified by the median risk score. (B) Distribution of OS time and status of PDAC patients. (C) Kaplan-Meier survival curves comparing high-risk and low-risk groups. (D) Time-dependent ROC curves at 1, 3, 5 years. (E) The 6 genes expression between high-risk groups and low-risk groups of PDAC patients. AUC, area under the curve; OS, overall survival; PDAC, pancreatic ductal adenocarcinoma; ROC, receiver operating characteristic; TCGA, The Cancer Genome Atlas.

The prognostic performance of the risk score model was evaluated using Kaplan-Meier analysis and ROC curves. Kaplan-Meier survival curves (Figure 4C) revealed significantly longer overall survival (OS) in the low-risk group compared to the high-risk group (P=0.02). ROC analysis (Figure 4D) showed 1-, 3-, and 5-year area under the curve (AUC) values of 0.623, 0.698, and 0.720, respectively, in the training cohort. A heatmap visualizing expression patterns of the six signature genes in TCGA pancreatic cancer patients (Figure 4E) demonstrated upregulated expression of AHR, PTGS2, TNFRSF21, and VDR in high-risk individuals.

Subsequently, the prognostic value of LMRGs was validated using an independent cohort of 179 pancreatic cancer cases from ICGC-PACA. Risk scores were calculated for each patient using the established formula. Patients were stratified into high- and low-risk groups based on the median risk score in the ICGC cohort. Prognostic analysis demonstrated results consistent with the TCGA cohort (Figure 5). Kaplan-Meier survival curves (Figure 5C) revealed significantly longer OS in the low-risk group compared to the high-risk group (P=0.02). ROC analysis (Figure 5D) showed 1-, 3-, and 5-year AUC values of 0.587, 0.579, and 0.616, respectively, in the validation cohort. Expression patterns of the six signature genes in ICGC pancreatic cancer patients (Figure 5E) aligned with those observed in the TCGA database.

Figure 5 Prognostic performance of the 6-gene signature model in ICGC cohort. (A) Distribution of patient survival status stratified by the median risk score. (B) Distribution of OS time and status of PDAC patients. (C) Kaplan-Meier survival curves comparing high-risk and low-risk groups. (D) Time-dependent ROC curves at 1, 3, 5 years. (E) The 6 genes expression between high-risk groups and low-risk groups of PDAC patients. AUC, area under the curve; ICGC, International Cancer Genome Consortium; OS, overall survival; PDAC, pancreatic ductal adenocarcinoma; ROC, receiver operating characteristic.

Risk score is an independent indicator for the prognosis of pancreatic cancer

To demonstrate that the lipid metabolism-related risk score serves as an independent prognostic indicator, univariable Cox regression analysis was performed incorporating clinical characteristics to assess its association with prognosis in the TCGA training cohort. The univariable analysis (Figure 6A) revealed that age, risk score, and N stage were statistically significant predictors of survival (all P<0.05). These three variables were subsequently incorporated into a multivariable Cox regression model, which confirmed the risk score, N stage, and age as independently associated with prognosis (Figure 6B). These results demonstrate that the LMRGs-based risk score is an independent prognostic indicator for pancreatic cancer patients.

Figure 6 Correlation between the risk score and clinical characteristics. (A) Univariate analysis showed that age, risk score and N stage were statistically significant survival predictors. (B) Cox regression confirmed that risk score, N stage and age were independently associated with prognosis. (C) The association between risk score and age. (D) The association between risk score and gender. (E) The association between risk score and tumor stage. (F) The association between risk score and T stage. (G) The association between risk score and N stage. CI, confidence interval; N, node; T, tumor.

Risk score and clinical characteristics

The correlation between risk scores and clinical characteristics in pancreatic cancer patients was analyzed using the Wilcoxon signed-rank test. The results demonstrated that the risk score based on LMRGs showed significant differences in T stage, but no significant differences in age, gender, tumor grade, or N stage (Figure 6C-6G). Therefore, a high-risk score may indicate stronger tumor local invasiveness.

GSEA analysis

To further elucidate the molecular mechanisms of LMRGs, GSEA analysis was performed between high-risk and low-risk groups based on risk scores in the training cohort of pancreatic cancer patients. The results demonstrated (top 5 pathways shown) that ABC transporters, alpha-linolenic acid metabolism, butanoate metabolism, fatty acid biosynthesis, and pyruvate metabolism were upregulated in the high-risk group, while basal transcription factors, drug metabolism-other enzymes, non-homologous end-joining, ribosome, and selenocompound metabolism were downregulated in the high-risk group (Figure 7).

Figure 7 GSEA analysis. (A) The upregulated pathways in the high-risk group compared to the low-risk group (top 5 pathways shown). (B) The down-regulation pathways in the high-risk group (top 5 pathways shown). GSEA, Gene Set Enrichment Analysis.

Protein expression

The immunohistochemistry results from the HPA database further validated the differential expression of six LMRGs between normal pancreatic tissue and pancreatic cancer tissue. Since the VDR gene is predominantly highly expressed in the parathyroid glands, intestines (colon, small intestine), kidneys, and skin with minimal pancreatic expression, no VDR immunohistochemistry data was available in the HPA database. The protein immunohistochemistry results for the other five genes were largely consistent with our previous analyses, showing significantly enhanced expression of AHR, PTGS2, and TNFRSF21 in tumor tissues (Figure 8).

Figure 8 Immunohistochemistry of the 5 proteins in normal tissue and pancreatic cancer tumor tissue based on the HPA (brown 3,3'-diaminobenzidine signal, blue hematoxylin counterstain). (A) AGT expression was not detected in pancreatic cancer tissues (image available from https://www.proteinatlas.org/ENSG00000135744-AGT/cancer/pancreatic+cancer#img) and normal pancreatic tissues (image available from https://www.proteinatlas.org/ENSG00000135744-AGT/tissue/pancreas#img). (B) The expression of AHR in pancreatic cancer tissues (image available from https://www.proteinatlas.org/ENSG00000106546-AHR/cancer/pancreatic+cancer#img) is higher than that in normal pancreatic tissues (image available from https://www.proteinatlas.org/ENSG00000106546-AHR/tissue/pancreas#img). (C) The expression of PLA2G6 in pancreatic cancer tissues (image available from https://www.proteinatlas.org/ENSG00000184381-PLA2G6/cancer/pancreatic+cancer#img) is lower than that in normal pancreatic tissues (image available from https://www.proteinatlas.org/ENSG00000184381-PLA2G6/tissue/pancreas#img). (D) The expression of PTGS2 in pancreatic cancer tissues (image available from https://www.proteinatlas.org/ENSG00000073756-PTGS2/cancer/pancreatic+cancer#img) is higher than that in normal pancreatic tissues (image available from https://www.proteinatlas.org/ENSG00000073756-PTGS2/tissue/pancreas#img). (E) The expression of TNFRSF21 in pancreatic cancer tissues (image available from https://www.proteinatlas.org/ENSG00000146072-TNFRSF21/cancer/pancreatic+cancer#img) is higher than that in normal pancreatic tissues (image available fromhttps://www.proteinatlas.org/ENSG00000146072-TNFRSF21/tissue/pancreas#img). HPA, human protein atlas.

Immune microenvironment

According to analytical methods including TIMER, CIBERSORT, CIBERSORT-ABS, QUANTISEQ, XCELL, EPIC, and MCP-counter, the tumor-infiltrating immune cells in pancreatic cancer were quantified, and the relationship between high-risk and low-risk groups and immune infiltration was displayed in a heatmap (Figure 9). The CIBERSORT results showed that the proportion of CD8⁺ T cells was higher in the low-risk group, while the proportion of M0 macrophages was higher in the high-risk group. The QUANTISEQ results revealed that the high-risk group had higher proportions of M0 macrophages and neutrophils. The XCELL results indicated that the low-risk group exhibited higher proportions of endothelial cells and B cells.

Figure 9 The infiltration of immune cells between the high-risk group and the low-risk group.

Among 79 immune checkpoint-related genes analyzed for association with the risk score, we identified 13 genes demonstrating significantly differential expression between high- and low-risk groups: BTNL9, CD70, CD80, CD160, CD276, CEACAM1, KIR3DL1, KIR3DL3, PDCD1, TNFRSF9, TNFSF4, TNFSF9, and TNFSF14 (Figure S2). Notably, eight genes—CD70, CD80, CD160, CEACAM1, KIR3DL3, TNFRSF9, TNFSF4, and TNFSF9—exhibited significantly elevated expression in the high-risk group. This expression pattern suggests that high-risk patients may possess an “immunosuppressive” or “immune-desert” tumor microenvironment.

TIMER database was utilized to analyze the correlation between LMRGs within the risk score model and the level of immune infiltration in pancreatic cancer. The results demonstrated that AGT showed positive correlations with CD8⁺ T cells, CD4⁺ T cells, macrophages, neutrophils, and dendritic cells; AHR was positively correlated with B cells, CD8⁺ T cells, macrophages, neutrophils, and dendritic cells; PLA2G6 exhibited positive correlations with CD8⁺ T cells and CD4⁺ T cells but a negative correlation with dendritic cells; PTGS2 was positively associated with B cells, CD8⁺ T cells, macrophages, neutrophils, and dendritic cells; TNFRSF21 displayed positive correlations with B cells, CD8⁺ T cells, and dendritic cells; and VDR showed positive correlations with B cells, CD8⁺ T cells, macrophages, neutrophils, and dendritic cells (Figure S3).

Prediction of candidate drugs

Using Enrichr and the selected six LMRGs, we obtained results from the drug prediction database DSigDB. A total of 1,391 drugs were identified, among which 1,179 had P<0.05 and 1,150 had adjusted P<0.05. The prediction results showed that prolinedithiocarbamate and isoliquiritigenin were the two drugs with the most significant gene interactions, while aspirin and resveratrol interacted with the majority of the genes (Table 2).

Discussion

Studies have shown that obesity and metabolic syndrome can lead to adipose tissue expansion and dysfunction, thereby causing local and systemic lipid metabolism abnormalities (25). This metabolic disorder not only promotes ectopic fat deposition in the pancreas but also creates favorable conditions for the occurrence and progression of pancreatic cancer by altering the adipose tissue microenvironment (26,27). Based on these findings, this study focuses on the mechanistic roles of LMRGs in pancreatic cancer, aiming to provide a theoretical foundation for developing novel therapeutic targets through systematic analysis of the expression patterns of these genes and their associations with clinical prognosis.

Among the six LMRGs identified as significantly associated with pancreatic cancer prognosis in this study, AGT is expressed in the liver as a precursor of angiotensin and serves as a key component of the renin-angiotensin system (RAS), functioning as a potent regulator of blood pressure, fluid, and electrolyte homeostasis. A study by Wu et al. on colon cancer and lipid metabolism-related regulated cell death genes demonstrated that AGT, as a signature gene of lipid metabolism and regulated cell death, plays a critical role in the development of colon adenocarcinoma and is associated with tumor immune infiltration (28). Our study found that AGT was significantly overexpressed in pancreatic cancer tissues, and immune infiltration analysis suggested positive correlations between AGT expression and various immune cells. Its overexpression may exert complex effects through lipid metabolism pathways and immune regulation, warranting further investigation into its role and mechanisms in pancreatic cancer.

The AHR mediates biological responses to planar aromatic hydrocarbons and plays pivotal roles in diverse BPs. In lipid metabolism, AHR influences lipid homeostasis by regulating gene expression in the liver and adipose tissue. Multiple studies have shown that under high-fat diet (HFD) conditions, AHR activation promotes hepatic steatosis and obesity, while AHR-knockout mice exhibit resistance to diet-induced obesity and fatty liver (29-31). In PDAC, suppression of AHR expression sensitizes PDAC cells to gemcitabine through the ELAVL1-DCK pathway (32). In Kras-Trp53 mutant pancreatic cancer mouse models, AHR maintains reactive oxygen species (ROS) balance, cell proliferation, invasion, and cancer regression via the AhR-Jdp2-Nrf2 axis (33). Additionally, AHR can influence pancreatic cancer progression by modulating the Treg/Th17 balance (34,35).

PLA2G6 mediates phospholipid metabolism through its catalytic activity and plays a central role in maintaining cellular membrane homeostasis. Its activity directly affects membrane phospholipid composition and fluidity, sustains mitochondrial function, and participates in regulating multiple cellular signaling pathways. Aberrant PLA2G6 expression is closely associated with lipid synthesis, storage, and catabolism, particularly in C16:0 and C18:1 fatty acid metabolism (36). Under oxidative stress conditions, PLA2G6 significantly reduces the accumulation of oxidatively modified lipids in pancreatic β-cells by reprogramming mitochondrial membrane phospholipid metabolism to facilitate clearance and renewal of damaged phospholipids (37). In pancreatic cancer—a tumor with highly reprogrammed metabolism—PLA2G6’s role as a “homeostasis maintainer” may confer protective functions. PLA2G6 could indirectly influence immune cell infiltration or function, or by maintaining redox balance in cancer cells to reduce pro-inflammatory factor production, thereby creating a microenvironment less favorable for tumor progression.

PTGS2 (also known as COX-2), or cyclooxygenase-2, is a key enzyme in prostaglandin biosynthesis (e.g., prostaglandin E2, PGE2), catalyzing the conversion of arachidonic acid to prostaglandins. PTGS2 is the core enzyme of the COX/PGE2 signaling axis, and its product PGE2 promotes tumor microenvironment remodeling, angiogenesis, and immune escape by activating downstream receptors (e.g., prostaglandin E2 receptor 4, EP4) (38-40). Aberrant PTGS2 expression is closely associated with dysregulated lipid metabolism. Studies indicate that PTGS2 can reduce hepatic lipid droplet accumulation and alleviate hepatic steatosis by downregulating lipid synthesis-related genes (41), though its role in pancreatic fat infiltration requires further validation. In pancreatic cancer, PTGS2 may promote tumorigenesis and progression through multiple mechanisms:

• Lipid metabolic reprogramming: The PTGS2/PGE2 axis drives pancreatic cancer cells to rely on lipid synthesis to support proliferation and chemoresistance (42,43). Consistent with this, clinical data associate elevated PTGS2 expression with poor prognosis in patients (44,45).
• Inflammation and tumorigenesis: PTGS2-mediated chronic inflammation is a critical driver of pancreatic cancer, with its overexpression exacerbating tumor-associated inflammation via transcription factors such as nuclear factor-κB (NF-κB) (45,46). Targeting this gene or its downstream pathways (e.g., COX-2/PGE2/EP4) may represent a potential therapeutic strategy.

TNFRSF21 (also known as CD358 or DR6) regulates cell fate through dual signaling mechanisms: on one hand, it activates NF-κB transcription factor to promote survival signals, while on the other, it induces apoptosis via the mitogen-activated protein kinase 8 (MAPK8/JNK1) cascade. In pancreatic cancer, TNFRSF21 may function through the following mechanisms:

(I) Immunomodulatory effects: high TNFRSF21 expression correlates with CD8⁺ T-cell infiltration and may facilitate pancreatic cancer progression by suppressing immune responses. A related family member, TNFR2 (TNFRSF1B), has been shown to drive pancreatic tumorigenesis through dual effects (immune suppression and partial tumor growth acceleration), suggesting TNFRSF21 may participate in immune escape via similar pathways (47,48). (II) Apoptosis regulation: tumor necrosis factor (TNF) receptor family members typically modulate the survival/death balance by activating NF-κB or caspase pathways, though the precise mechanism of TNFRSF21 in pancreatic cancer remains to be elucidated.

The VDR, the nuclear receptor for calcitriol, primarily regulates target gene transcription and plays a central role in calcium homeostasis. VDR is expressed in pancreatic β-cells and participates in glucose homeostasis regulation through the vitamin D/VDR signaling pathway (49,50). Studies demonstrate that VDR deficiency leads to impaired glycogen storage and abnormal carbohydrate metabolism in mice, which may indirectly affect lipid utilization (51). Our analysis reveals that high VDR expression significantly correlates with poor prognosis in pancreatic cancer patients. Additionally, Li et al. showed that VDR overexpression promotes M2 macrophage polarization and recruitment via C-C motif chemokine ligand 20 (CCL20) secretion, thereby accelerating pancreatic cancer progression (52).

Based on the risk score established from the above six LMRGs, pancreatic cancer patients in the TCGA cohort were divided into high-risk and low-risk subgroups with significant prognostic differences using the median cutoff method. Kaplan-Meier analysis demonstrated that the low-risk group had significantly better survival rates, and the ROC curve also confirmed the predictive ability of the risk score model. Overall, LMRGs-based risk score holds significant clinical value in evaluating pancreatic cancer prognosis. The results from the ICGC cohort further validated that the LMRG-based risk score is a reliable prognostic indicator.

Previous studies have also employed bioinformatic approaches to construct prognostic models for pancreatic cancer. A study on ferroptosis-related genes in pancreatic cancer developed a risk score based on 14 genes, with ROC curves at 12, 18, and 24 months all achieving approximately 0.8. This signature of 14 ferroptosis-related genes was also found to be significantly associated with immune infiltration and immune checkpoint blockade proteins (53). Recent studies have also investigated the association between LMRGs and pancreatic cancer. Yuan Shu et al. characterized distinct lipid metabolic subtypes in pancreatic cancer and demonstrated that in the validation cohort, the AUC values for predicting 1‑, 3‑, and 5year survival were 0.65, 0.81, and 0.79, respectively (54). These results suggest that a high lipid metabolism score may indicate advanced disease progression (54). A study by Wu et al. developed a predictive model for pancreatic cancer based on macrophage-related genes and lipid metabolism-related genes, the AUC values for predicting 1‑, 2‑, and 3year survival were 0.67, 0.66, and 0.71, respectively (55). The AUC values in our study (ranging from 0.58 to 0.72) indicate moderate predictive performance, which currently limits the clinical utility. Further refinement of the model—such as incorporating immune-related genes in an integrated analysis—is warranted to improve its robustness and translational potential.

GSEA analysis revealed significant lipid metabolic reprogramming features in high-risk pancreatic cancer patients. Multiple key lipid metabolism pathways (including α-linolenic acid metabolism, fatty acid biosynthesis, etc.) were significantly upregulated in the high-risk group. Aberrantly activated lipid metabolism may accelerate tumor progression by: (I) providing abundant energy substrates for rapidly proliferating tumor cells; (II) activating pro-survival signaling pathways; and (III) inducing the expression of chemotherapy efflux pumps. The high-risk group simultaneously exhibited significant downregulation of basal transcriptional regulatory mechanisms and DNA damage repair pathways (particularly non-homologous end joining), potentially suggesting: (I) mutation accumulation due to increased genomic instability; (II) oncogene aberrant expression caused by transcriptional dysregulation; and (III) tumorigenesis and progression resulting from defective DNA damage repair. Therefore, LMRGs may influence pancreatic cancer prognosis through these pathways. In the tumor microenvironment of low-risk pancreatic cancer patients, CD8⁺ T cell infiltration levels were significantly higher, consistent with previous studies reporting the association between CD8⁺ T cells and favorable prognosis (56). As key effector cells in anti-tumor immune responses, the enrichment of CD8⁺ T cells may reflect stronger immune surveillance capacity in low-risk patients. High-risk patients showed significant upregulation of multiple immune checkpoint-related genes, suggesting their tumor microenvironment may be in a highly immunosuppressive state. This could be related to LMRGs upregulation, where adipocyte-released chemokines recruit various immune cell subsets and stromal immune infiltration, collectively shaping an immunosuppressive microenvironment (25). High-risk pancreatic cancer is associated with immunosuppression, which may offer potential targets for future mechanistic investigations. However, whether this relationship is causal or represents a vicious cycle remains to be elucidated through further experimental research. Additionally, the elevated expression of immune checkpoint molecules suggests that these patients may exhibit sensitivity to immune checkpoint inhibitor therapy, warranting further exploration of its clinical translational potential in future studies.

Among the drugs screened through pharmacological prediction analysis, derivatives of prolinedithiocarbamate have demonstrated antitumor activity. In breast cancer, the manganese(II) prolinedithiocarbamate (MnProDtc) complex effectively inhibits tumor cell proliferation (57), while copper(II) prolinedithiocarbamate complexes have also been synthesized and evaluated as highly effective, low-toxicity anticancer drug candidates with fewer side effects compared to traditional drugs like cisplatin (58). Isoliquiritigenin, a natural flavonoid, exhibits antioxidant, anti-inflammatory, and antitumor properties. Studies suggest it may treat obesity induced by HFDs through lipid metabolism regulation (59), and has shown antitumor effects in lung cancer and melanoma cell lines (60,61). Aspirin, a nonsteroidal anti-inflammatory drug, can block the cell cycle by inhibiting the glycogen synthase kinase-3β (GSK-3β) pathway, thereby reducing pancreatic cancer cell proliferation (62). Zhou et al. demonstrated that aspirin enhances gemcitabine’s effects by suppressing the phosphoinositide 3-kinase/protein kinase B/mammalian target of rapamycin (PI3K/AKT/mTOR) pathway and reversing epithelial-mesenchymal transition, thereby decreasing pancreatic cancer cell proliferation, migration, and invasion while increasing apoptosis (63). Epidemiological data also indicate that long-term aspirin use reduces pancreatic cancer risk (64), particularly in diabetic patients (65). Resveratrol, a natural polyphenol, exhibits dual effects in lipid metabolism regulation and antitumor activity. In animal studies, resveratrol alleviates endoplasmic reticulum stress, upregulates peroxisome proliferator-activated receptor γ (PPARγ) expression (essential for glucolipid metabolism), inhibits lipid peroxidation, and effectively blocks acrolein-triggered ferroptosis cascades (66). In nude mouse pancreatic cancer models, resveratrol suppresses tumor growth by inhibiting NAF-1 (CISD2, CDGSH iron-sulfur domain-containing protein 2) expression (67). Additionally, resveratrol improves high-fat-diet-induced hepatic lipid metabolism disorders by regulating m6A mRNA methylation (68), suggesting its potential to intervene in pancreatic cancer’s abnormal lipid metabolism through epigenetic pathways. Nevertheless, the drug predictions are based solely on retrospective bioinformatic analyses, allowing only speculation regarding the potential mechanisms of treatment response. At this stage, the findings lack rigorous experimental validation, which limits their clinical translatability. Further experimental studies and prospective clinical validation are required in the future.

This study provides an in-depth investigation into the potential mechanisms of LMRGs in pancreatic cancer. Functional enrichment analysis revealed that these genes may influence the tumor microenvironment by regulating fatty acid metabolism and gene repair pathways, thereby promoting or suppressing pancreatic cancer progression. These findings offer new theoretical insights into understanding metabolic reprogramming in pancreatic cancer. Immune cell infiltration analysis based on risk scores demonstrated significant differences in the immune microenvironment between high-risk and low-risk groups, suggesting that lipid metabolism may regulate immune cell function and contribute to tumor immune evasion. This discovery provides a potential strategy for identifying pancreatic cancer patients who may benefit from immunotherapy. By predicting drugs targeting key LMRGs, this study explored the possible pharmacological mechanisms of these drugs for pancreatic cancer, presenting new possibilities for the treatment of pancreatic cancer. The drug sensitivity analysis further validated the correlation between LMRGs and therapeutic response. This research primarily relies on retrospective data from public databases and lacks experimental validation (e.g., cell or animal models) to elucidate the biological functions and regulatory mechanisms of key genes. Future studies should validate these findings in independent, multicenter prospective clinical cohorts and optimize the model by integrating experimental and clinical data to explore its association with drug treatment responses. The current limitations include the need for further investigation into the precise roles of lipid metabolism in immune cell infiltration and immunotherapy efficacy.

Conclusions

The risk score based on six LMRGs may hold predictive value for the prognosis of pancreatic cancer. The activation of an immunosuppressive microenvironment is associated with high-risk pancreatic cancer. The identified candidate drugs exhibit potential for personalized treatment of pancreatic cancer. These findings provide an important theoretical foundation for developing novel combination therapeutic strategies targeting the regulation of lipid metabolism, though extensive experimental validation is still required.

Acknowledgments

Acknowledgments to the GTEx, TCGA and ICGC databases for providing researchable patient data.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-507/rc

Peer Review File: Available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-507/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-507/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All the original data are obtained from the database and no ethical approval is required.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Huang J, Lok V, Ngai CH, et al. Worldwide Burden of, Risk Factors for, and Trends in Pancreatic Cancer. Gastroenterology 2021;160:744-54. [Crossref] [PubMed]
2. McGuigan A, Kelly P, Turkington RC, et al. Pancreatic cancer: A review of clinical diagnosis, epidemiology, treatment and outcomes. World J Gastroenterol 2018;24:4846-61. [Crossref] [PubMed]
3. Ruze R, Song J, Yin X, et al. Mechanisms of obesity- and diabetes mellitus-related pancreatic carcinogenesis: a comprehensive and systematic review. Signal Transduct Target Ther 2023;8:139. [Crossref] [PubMed]
4. Sherman MH, Beatty GL. Tumor Microenvironment in Pancreatic Cancer Pathogenesis and Therapeutic Resistance. Annu Rev Pathol 2023;18:123-48. [Crossref] [PubMed]
5. Zhang M, Xiang R, Glorieux C, et al. PLA2G2A Phospholipase Promotes Fatty Acid Synthesis and Energy Metabolism in Pancreatic Cancer Cells with K-ras Mutation. Int J Mol Sci 2022;23:11721. [Crossref] [PubMed]
6. Zhu L, Ji J, Ma J, et al. Differential Effects of Dietary Macronutrients on the Development of Oncogenic KRAS-Mediated Pancreatic Ductal Adenocarcinoma. Cancers (Basel) 2022;14:2723. [Crossref] [PubMed]
7. Zhang R, Peng X, Du JX, et al. Oncogenic KRASG12D Reprograms Lipid Metabolism by Upregulating SLC25A1 to Drive Pancreatic Tumorigenesis. Cancer Res 2023;83:3739-52. [Crossref] [PubMed]
8. Wu Z, Duan W, Xiong Y, et al. NeuroD1 drives a KAT2A-FDFT1 signaling axis to promote cholesterol biosynthesis and hepatocellular carcinoma progression via histone H3K27 acetylation. Oncogene 2025;44:4017-31. [Crossref] [PubMed]
9. Gabitova-Cornell L, Surumbayeva A, Peri S, et al. Cholesterol Pathway Inhibition Induces TGF-β Signaling to Promote Basal Differentiation in Pancreatic Cancer. Cancer Cell 2020;38:567-583.e11. [Crossref] [PubMed]
10. Xu R, Song J, Ruze R, et al. SQLE promotes pancreatic cancer growth by attenuating ER stress and activating lipid rafts-regulated Src/PI3K/Akt signaling pathway. Cell Death Dis 2023;14:497. [Crossref] [PubMed]
11. Jin X, Dai L, Ma Y, et al. Implications of HIF-1α in the tumorigenesis and progression of pancreatic cancer. Cancer Cell Int 2020;20:273. [Crossref] [PubMed]
12. Li Y, Liang X, Che G, et al. Molecular Classification of Genes Associated with Hypoxic Lipid Metabolism in Pancreatic Cancer. Biomolecules 2022;12:1533. [Crossref] [PubMed]
13. Liu S, Kang S, Lin N, et al. The TET3/GATA6 Axis Drives Lipid Metabolism and Therapeutic Vulnerabilities in Pancreatic Ductal Adenocarcinoma. Adv Sci (Weinh) 2025;12:e01774. [Crossref] [PubMed]
14. Kim DH, Song NY, Yim H. Targeting dysregulated lipid metabolism in the tumor microenvironment. Arch Pharm Res 2023;46:855-81. [Crossref] [PubMed]
15. Zhou S, Guo Q, Chen A, et al. Integrated bioinformatics analysis identifies PCSK9 as a prognosticator correlated with lipid metabolism in pancreatic adenocarcinoma. World J Surg Oncol 2024;22:256. [Crossref] [PubMed]
16. Li Y, Amrutkar M, Finstadsveen AV, et al. Fatty acids abrogate the growth-suppressive effects induced by inhibition of cholesterol flux in pancreatic cancer cells. Cancer Cell Int 2023;23:276. [Crossref] [PubMed]
17. Liu G, Yang D, Meng J, et al. The combination of gemcitabine and albumin-bound paclitaxel effectively inhibits de novo lipogenesis in pancreatic cancer cells by targeting the AMPK/SREBP1 pathway. Biochem Pharmacol 2025;232:116721. [Crossref] [PubMed]
18. Shi YH, Liu ZD, Ma MJ, et al. Platelet-Derived Growth Factor C Facilitates Malignant Behavior of Pancreatic Ductal Adenocarcinoma by Regulating SREBP1 Mediated Lipid Metabolism. Adv Sci (Weinh) 2024;11:e2407069. [Crossref] [PubMed]
19. Cao S, Saw PE, Shen Q, et al. Reduction-responsive RNAi nanoplatform to reprogram tumor lipid metabolism and repolarize macrophage for combination pancreatic cancer therapy. Biomaterials 2022;280:121264. [Crossref] [PubMed]
20. Ye L, Li Y, Zhang S, et al. Exosomes-regulated lipid metabolism in tumorigenesis and cancer progression. Cytokine Growth Factor Rev 2023;73:27-39. [Crossref] [PubMed]
21. Hu FF, Liu CJ, Liu LL, et al. Expression profile of immune checkpoint genes and their roles in predicting immunotherapy response. Brief Bioinform 2021;22:bbaa176. [Crossref] [PubMed]
22. Ritchie ME, Phipson B, Wu D, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015;43:e47. [Crossref] [PubMed]
23. Yu G, Wang LG, Han Y, et al. clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 2012;16:284-7. [Crossref] [PubMed]
24. Xu S, Hu E, Cai Y, et al. Using clusterProfiler to characterize multiomics data. Nat Protoc 2024;19:3292-320. [Crossref] [PubMed]
25. Brocco D, Florio R, De Lellis L, et al. The Role of Dysfunctional Adipose Tissue in Pancreatic Cancer: A Molecular Perspective. Cancers (Basel) 2020;12:1849. [Crossref] [PubMed]
26. Lee-Rueckert M, Canyelles M, Tondo M, et al. Obesity-induced changes in cancer cells and their microenvironment: Mechanisms and therapeutic perspectives to manage dysregulated lipid metabolism. Semin Cancer Biol 2023;93:36-51. [Crossref] [PubMed]
27. Khodashahi R, Beiraghdar F, Ferns GA, et al. The Role of Local Angiotensin II/Angiotensin Type 1-receptor Mechanisms in Adipose Tissue Dysfunction to Promote Pancreatic Cancer. Curr Cancer Drug Targets 2024;24:1187-94. [Crossref] [PubMed]
28. Wu M, Zhou Y, Pei D, et al. Unveiling the role of AGT in lipid metabolism and regulated cell death in colon cancer. Neoplasia 2024;54:101009. [Crossref] [PubMed]
29. Gourronc FA, Markan KR, Kulhankova K, et al. Pdgfrα-Cre mediated knockout of the aryl hydrocarbon receptor protects mice from high-fat diet induced obesity and hepatic steatosis. PLoS One 2020;15:e0236741. [Crossref] [PubMed]
30. Pohjanvirta R, Karppinen I, Galbán-Velázquez S, et al. Effects of a high-fat diet and global aryl hydrocarbon receptor deficiency on energy balance and liver retinoid status in male Sprague-Dawley rats. J Nutr Biochem 2021;95:108762. [Crossref] [PubMed]
31. Xu X, Sun S, Liang L, et al. Role of the Aryl Hydrocarbon Receptor and Gut Microbiota-Derived Metabolites Indole-3-Acetic Acid in Sulforaphane Alleviates Hepatic Steatosis in Mice. Front Nutr 2021;8:756565. [Crossref] [PubMed]
32. Stukas D, Jasukaitiene A, Bartkeviciene A, et al. Targeting AHR Increases Pancreatic Cancer Cell Sensitivity to Gemcitabine through the ELAVL1-DCK Pathway. Int J Mol Sci 2023;24:13155. [Crossref] [PubMed]
33. Wuputra K, Tsai MH, Kato K, et al. Jdp2 is a spatiotemporal transcriptional activator of the AhR via the Nrf2 gene battery. Inflamm Regen 2023;43:42. [Crossref] [PubMed]
34. Bartkeviciene A, Jasukaitiene A, Zievyte I, et al. Association between AHR Expression and Immune Dysregulation in Pancreatic Ductal Adenocarcinoma: Insights from Comprehensive Immune Profiling of Peripheral Blood Mononuclear Cells. Cancers (Basel) 2023;15:4639. [Crossref] [PubMed]
35. Walcheck MT, Schwartz PB, Carrillo ND, et al. Aryl Hydrocarbon Receptor Knockout Accelerates PanIN Formation and Fibro-Inflammation in a Mutant Kras -Driven Pancreatic Cancer Model. Pancreas 2024;53:e670-80. [Crossref] [PubMed]
36. Chen C, Lou MM, Sun YM, et al. Serum metabolomic characterization of PLA2G6-associated dystonia-parkinsonism: A case-control biomarker study. Front Neurosci 2022;16:879548. [Crossref] [PubMed]
37. Hsu YH, Dumlao DS, Cao J, et al. Assessing phospholipase A2 activity toward cardiolipin by mass spectrometry. PLoS One 2013;8:e59267. [Crossref] [PubMed]
38. Nakanishi M, Rosenberg DW. Epithelial Cell-specific Deletion of Microsomal Prostaglandin E Synthase-1 Does Not Influence Colon Tumor Development in Mice. J Cancer Prev 2021;26:304-8. [Crossref] [PubMed]
39. He J, Lin X, Meng F, et al. A Novel Small Molecular Prostaglandin Receptor EP4 Antagonist, L001, Suppresses Pancreatic Cancer Metastasis. Molecules 2022;27:1209. [Crossref] [PubMed]
40. Wang Q, Morris RJ, Bode AM, et al. Prostaglandin Pathways: Opportunities for Cancer Prevention and Therapy. Cancer Res 2022;82:949-65. [Crossref] [PubMed]
41. He X, Hu Y, Liu W, et al. Deciphering the Effective Constituents and Mechanisms of Portulaca oleracea L. for Treating NASH via Integrating Bioinformatics Analysis and Experimental Pharmacology. Front Pharmacol 2021;12:818227. [Crossref] [PubMed]
42. Oberle R, Kührer K, Österreicher T, et al. The HDL particle composition determines its antitumor activity in pancreatic cancer. Life Sci Alliance 2022;5:e202101317. [Crossref] [PubMed]
43. Yin X, Xu R, Song J, et al. Lipid metabolism in pancreatic cancer: emerging roles and potential targets. Cancer Commun (Lond) 2022;42:1234-56. [Crossref] [PubMed]
44. Renaldi K, Simadibrata M, Rahadiani N, et al. Prognostic Value of COX-2, NF-κB, and Sp1 Tissue Expressions in Pancreatic Ductal Adenocarcinoma: A Systematic Review and Meta-analysis. Turk J Gastroenterol 2021;32:956-70. [Crossref] [PubMed]
45. Renaldi K, Simadibrata M, Rahadiani N, et al. The Expressions of NF-κB, COX-2, Sp1, and c-Jun in Pancreatic Ductal Adenocarcinoma and Their Associations with Patient Survival. Pathophysiology 2023;30:92-109. [Crossref] [PubMed]
46. Mittal R, Sharma S, Mittal A, et al. Recent Progress in Selective COX-2 Inhibitor Formulations and Therapeutic Applications - A Patent Review (2012-2022). Mini Rev Med Chem 2023;23:2130-41. [Crossref] [PubMed]
47. Zhang X, Lao M, Xu J, et al. Combination cancer immunotherapy targeting TNFR2 and PD-1/PD-L1 signaling reduces immunosuppressive effects in the microenvironment of pancreatic tumors. J Immunother Cancer 2022;10:e003982. [Crossref] [PubMed]
48. Gao Z, Zhang Q, Chen H, et al. TNFR2 promotes pancreatic cancer proliferation, migration, and invasion via the NF-κB signaling pathway. Aging (Albany NY) 2023;15:8013-25. [Crossref] [PubMed]
49. Lau SL, Stokes RA, Ng B, et al. Metabolic changes in vitamin D receptor knockout mice. PLoS One 2022;17:e0267573. [Crossref] [PubMed]
50. Wu J, Atkins A, Downes M, et al. Vitamin D in Diabetes: Uncovering the Sunshine Hormone's Role in Glucose Metabolism and Beyond. Nutrients 2023;15:1997. [Crossref] [PubMed]
51. Das A, Jawla N, Meena V, et al. Lack of vitamin D signalling shifts skeletal muscles towards oxidative metabolism. J Cachexia Sarcopenia Muscle 2024;15:67-80. [Crossref] [PubMed]
52. Li H, Ruan Y, Liu C, et al. VDR promotes pancreatic cancer progression in vivo by activating CCL20-mediated M2 polarization of tumor associated macrophage. Cell Commun Signal 2024;22:224. [Crossref] [PubMed]
53. Jiang P, Yang F, Zou C, et al. The construction and analysis of a ferroptosis-related gene prognostic signature for pancreatic cancer. Aging (Albany NY) 2021;13:10396-414. [Crossref] [PubMed]
54. Shu Y, Huang H, Gao M, et al. Lipid Metabolism-Related Gene Markers Used for Prediction Prognosis, Immune Microenvironment, and Tumor Stage of Pancreatic Cancer. Biochem Genet 2024;62:931-49. [Crossref] [PubMed]
55. Wu L, Liang F, Chen C, et al. Identification of prognostic and therapeutic biomarkers associated with macrophage and lipid metabolism in pancreatic cancer. Sci Rep 2025;15:14584. [Crossref] [PubMed]
56. Zhu K, Liu X, Deng W, et al. Identification of a chromatin regulator signature and potential candidate drugs for bladder cancer. Hereditas 2022;159:13. [Crossref] [PubMed]
57. Jarre S, Raya I. Synthesis, characterization, molecular docking studies of Mn(II)Prolinedithiocarbamate and its potential as anticancer agents. Mol Divers 2024;28:889-900. [Crossref] [PubMed]
58. Irfandi R, Raya I, Ahmad A, et al. Anticancer potential of Cu(II)prolinedithiocarbamate complex: design, synthesis, spectroscopy, molecular docking, molecular dynamic, ADMET, and in-vitro studies. J Biomol Struct Dyn 2023;41:12938-50. [Crossref] [PubMed]
59. Zhao L, Li M, Zhu Q, et al. Isoliquiritigenin Ameliorates High-Fat Diet-Induced Obesity in Mice by Activating Brown Adipose Tissue. Int J Mol Sci 2025;26:1616. [Crossref] [PubMed]
60. Zeng X, Zhao M, Yao H. Anti-lung Cancer, Anti-microbial, Anti-α-glucosidase, Anti-sorbitol Dehydrogenase, and in silico Studies of Wogonoside and Isoliquiritigenin as Natural Compounds. J Oleo Sci 2023;72:919-27. [Crossref] [PubMed]
61. Kwon MJ, Raut PK, Jang JH, et al. Isoliquiritigenin Induces Apoptosis via ROS-Mediated Inhibition of p38/mTOR/STAT3 Pathway in Human Melanoma Cells. Biomol Ther (Seoul) 2025;33:378-87. [Crossref] [PubMed]
62. Nakabayashi R, Fujihara S, Iwama H, et al. Effect of Aspirin on G(0)/G(1) Cell Cycle Arrest and microRNA Signatures in Pancreatic Adenocarcinoma Cells. Anticancer Res 2022;42:4037-48. [Crossref] [PubMed]
63. Zhou H, Yun X, Shu Y, et al. Aspirin increases the efficacy of gemcitabine in pancreatic cancer by modulating the PI3K/AKT/mTOR signaling pathway and reversing epithelial mesenchymal transition. Oncol Lett 2023;25:101. [Crossref] [PubMed]
64. Brasky TM, Jager LR, Newton AM, et al. Use of Nonsteroidal Anti-Inflammatory Drugs and Pancreatic Cancer Risk in the Women's Health Initiative. Cancer Epidemiol Biomarkers Prev 2024;33:1203-10. [Crossref] [PubMed]
65. Tan JT, Mao X, Cheng HM, et al. Aspirin is associated with lower risk of pancreatic cancer and cancer-related mortality in patients with diabetes mellitus. Gut 2025;74:603-12. [Crossref] [PubMed]
66. Zhang X, Jiang L, Chen H, et al. Resveratrol protected acrolein-induced ferroptosis and insulin secretion dysfunction via ER-stress- related PERK pathway in MIN6 cells. Toxicology 2022;465:153048. [Crossref] [PubMed]
67. Qin T, Cheng L, Xiao Y, et al. NAF-1 Inhibition by Resveratrol Suppresses Cancer Stem Cell-Like Properties and the Invasion of Pancreatic Cancer. Front Oncol 2020;10:1038. [Crossref] [PubMed]
68. Wu J, Li Y, Yu J, et al. Resveratrol Attenuates High-Fat Diet Induced Hepatic Lipid Homeostasis Disorder and Decreases m(6)A RNA Methylation. Front Pharmacol 2020;11:568006. [Crossref] [PubMed]