Abstract
Objective
Chronic pancreatitis (CP) is a progressive inflammatory disease characterized by persistent immune dysregulation and pancreatic injury. This study aimed to systematically investigate the expression characteristics of ferroptosis-related immune genes in CP and to identify key genes associated with disease status.
Methods
Two public transcriptomic datasets were analyzed to identify differentially expressed genes between CP and normal pancreatic tissues. Ferroptosis-related genes were obtained from the FerrDb database and intersected with differentially expressed genes. Functional enrichment, protein–protein interaction, and immune infiltration analyses were performed. A ferroptosis-related gene signature was constructed using LASSO regression and internally validated. A cerulein-induced CP mouse model was used for experimental validation by histological analysis and qRT-PCR.
Results
Nineteen ferroptosis-related differentially expressed genes were identified and enriched in oxidative stress–related pathways. Immune infiltration analysis suggested potential alterations in immune cell composition in CP; however, only plasma cells showed statistically significant differences between groups. The Ferro-Score model demonstrated excellent discrimination performance in the training dataset (AUC = 1.0). However, given that the model was evaluated only within the same dataset used for model construction, this result may reflect potential overfitting. Therefore, the predictive performance of the model should be interpreted with caution. The experimental verification at the mRNA level confirmed the differential expression of ZFP36 and BNIP3 in pancreatic tissue.
Conclusion
This study identifies distinct ferroptosis-related immune gene signatures in CP and provides preliminary insights into the association between ferroptosis-related immune pathways and CP.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12876-026-04794-6.
Keywords: Chronic pancreatitis, Ferroptosis, Immune infiltration, ZFP36, BNIP3
Introduction
Chronic pancreatitis (CP) is a progressive inflammatory disease characterized by persistent pancreatic inflammation, irreversible acinar cell injury, and progressive fibrosis, ultimately leading to exocrine and endocrine dysfunction [1, 2]. Epidemiological studies indicate that the global incidence of CP ranges from approximately 5 to 14 cases per 100,000 individuals per year, with a steadily increasing disease burden reported in both Western countries and Asia [3]. The prevalence of CP is particularly high in certain regions with metabolic disorders and alcohol-related risk factors, underscoring its growing public health impact. Increasing evidence suggests that CP is not merely a consequence of repeated tissue injury, but rather a chronic immune-mediated process driven by sustained inflammatory signaling and immune cell infiltration [4]. Aberrant activation and accumulation of immune cells, together with a dysregulated inflammatory microenvironment, play a critical role in perpetuating pancreatic damage and fibrotic remodeling [5, 6]. Although immune dysregulation has been recognized as a central feature of CP pathogenesis, the molecular mechanisms linking immune responses to pancreatic tissue injury remain incompletely understood.
Ferroptosis is a recently characterized form of regulated cell death driven by iron-dependent lipid peroxidation, which is mechanistically distinct from apoptosis, necroptosis, and autophagy [7]. Growing evidence suggests that ferroptosis is closely involved in the initiation and progression of various inflammatory and metabolic disorders, as well as organ fibrosis, by modulating oxidative stress and inflammatory signaling pathways [8]. In recent years, ferroptosis has been implicated in pancreatic diseases, particularly in acute pancreatic injury and pancreatic cancer [9]. However, existing studies have largely focused on acute injury models or tumor biology, while the potential involvement of ferroptosis in the chronic inflammatory and fibrotic processes of CP has not been systematically explored.
In this context, recent studies have demonstrated that metabolic reprogramming plays a critical role in shaping the immune microenvironment, influencing immune cell activation, inflammatory signaling, and therapeutic responses [10].
Accumulating experimental and theoretical studies indicate that ferroptosis is tightly linked to immune regulation, influencing immune cell activation, cytokine release, and inflammatory homeostasis [11]. Conversely, immune-derived inflammatory signals can reshape cellular susceptibility to ferroptosis, suggesting a bidirectional regulatory relationship [12]. Despite this emerging concept, several critical gaps remain in the context of CP. First, the global expression landscape of ferroptosis-related genes in CP has not been comprehensively characterized. Second, the association between ferroptosis-related genes and immune cell infiltration in CP lacks quantitative evaluation. Third, an integrated bioinformatics framework for identifying key ferroptosis-related immune genes with potential biological and clinical relevance in CP is still absent.
In this study, we aimed to systematically characterize the expression patterns of ferroptosis-related immune genes in CP. By integrating multiple transcriptomic datasets and applying comprehensive bioinformatics analyses, we sought to identify key genes associated with ferroptosis and immune regulation in CP. Furthermore, an experimental CP model was employed to provide preliminary biological validation, thereby offering new insights into the potential role of ferroptosis-related immune mechanisms in CP pathogenesis.
Materials and methods
Data acquisition and preprocessing
Publicly available gene expression datasets related to CP were retrieved from the Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/gds/) database. Based on the availability of CP samples and normal pancreatic tissue samples, we selected two independent datasets: GSE77858 and GSE143754. GSE77858 (platform: GPL7264, Agilent) contains 5 samples of CP and 3 samples of normal pancreatic tissue, while GSE143754 (platform: GPL17586, Affymetrix) includes 6 samples of CP and 9 control samples. Samples diagnosed as pancreatic ductal adenocarcinoma were excluded to minimize potential confounding effects of malignant transformation. Raw expression matrices were downloaded and processed using R software. Probe identifiers were annotated to gene symbols according to the corresponding platform annotation files, and probes mapping to the same gene symbol were averaged. Expression data were normalized within each dataset to ensure intra-dataset comparability. Given the substantial cross-platform differences between GSE77858 (GPL7264, Agilent) and GSE143754 (GPL17586, Affymetrix), including differences in probe design, signal processing, and dynamic range, direct merging of raw expression matrices may introduce significant batch effects. Therefore, differential expression analyses were performed independently within each dataset to avoid artificial bias introduced by cross-platform integration. A comprehensive list of ferroptosis-related genes was obtained from the FerrDb database(http://www.zhounan.org/ferrdb/current/), including drivers, suppressors, and markers of ferroptosis. In this study, these categories were integrated into a unified gene set. After removing duplicate entries, a nonredundant ferroptosis gene set was generated for subsequent analyses. The complete list of ferroptosis-related genes has been provided in Supplementary Table S1.
The two datasets were generated from different microarray platforms (GPL7264, Agilent; GPL17586, Affymetrix), which differ in probe design, signal processing, and dynamic range. Direct merging without correction would introduce substantial batch effects. Therefore, a cross-platform integration strategy based on independent DEG identification and gene intersection was applied, without direct merging of expression matrices, to minimize batch effects.
Identification of differentially expressed genes
Differential gene expression analysis between CP and normal control samples was performed separately for each dataset using the limma package in R. Linear models were fitted to the normalized expression data, and empirical Bayes moderation was applied to improve variance estimation. For GSE77858, genes with |log2 fold change (FC)| > 0.5 and P < 0.05 were considered differentially expressed. For GSE143754, genes with |log2FC| > 1 and P < 0.05 were considered differentially expressed. P values were derived from limma moderated t-tests. Because of the limited sample size and cross-platform heterogeneity, nominal P values were used for the initial DEG screening, while the false discovery rate (FDR) calculated using the Benjamini–Hochberg method was used as a reference for result interpretation. Because the two datasets were generated from different platforms and showed different signal distributions, slightly different fold-change thresholds were applied to avoid excessive false-positive or false-negative findings. The differentially expressed genes (DEGs) in the two datasets are presented in the supplementary materials. DEGs were identified independently in each dataset, and the resulting gene lists were subsequently integrated by removing duplicated gene symbols. This strategy was adopted to reduce platform-driven artifacts and to prioritize genes consistently dysregulated across independent cohorts, thereby enhancing the robustness of candidate gene selection.
Although formal meta-analysis or merged-matrix approaches can be applied to multi-dataset studies, these methods require stringent cross-platform normalization and may still be influenced by residual batch effects. In this study, we adopted a conservative integration strategy based on independent DEG identification followed by overlap analysis, which has been widely used in cross-platform transcriptomic studies to identify reproducible gene signatures.
Screening of ferroptosis-related differentially expressed genes
To identify genes associated with ferroptosis in CP, the unified DEG list was intersected with the curated ferroptosis gene set obtained from FerrDb. Genes present in both datasets were defined as ferroptosis-related differentially expressed genes (FR-DEGs) (Supplementary Table S2). Visualization of differential expression patterns was performed using volcano plots and heatmaps generated with R-based visualization packages. The overlap between DEGs and ferroptosis-related genes was illustrated using Venn diagrams to facilitate intuitive interpretation of gene set intersections.
Functional enrichment analysis
Functional enrichment analyses were conducted to explore the biological significance of the identified FR-DEGs. Gene Ontology (GO) enrichment analysis was performed to assess overrepresented biological processes, cellular components, and molecular functions, while Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was used to identify enriched signaling pathways. In addition, gene set enrichment analysis (GSEA) was applied to evaluate the coordinated expression changes of predefined gene sets associated with ferroptosis-related biological functions. Statistical significance was determined based on adjusted p-values, and ferroptosis-relevant functional terms were further curated according to biological relevance.
Protein-protein interaction (PPI) network construction and hub gene identification
PPI analysis of FR-DEGs was conducted using the STRING database to evaluate potential functional associations among encoded proteins. Interaction data were imported into Cytoscape software for network visualization and analysis. The CytoHubba plugin was applied to identify hub genes within the PPI network using multiple topological algorithms, and genes with high connectivity scores were considered candidate hub genes for subsequent analyses. The top 10 hub genes identified are presented in Supplementary Table S3.
Immune infiltration analysis
The relative abundance of immune cell populations was estimated using CIBERSORTx with the LM22 signature matrix, along with complementary analyses using xCell and MCP-counter algorithms. Given that CIBERSORTx was originally developed based on blood-derived immune signatures and has been primarily validated in tumor tissues, its application in fibrotic pancreatic tissue may introduce potential bias. Therefore, the immune infiltration analysis in this study was considered exploratory, and no strict filtering based on CIBERSORTx P-values was applied. Differences in immune cell composition between CP and control samples were assessed, and correlations between FR-DEG expression levels and significantly altered immune cell types were evaluated using correlation analysis to explore potential associations between ferroptosis-related genes and immune infiltration patterns. The complete correlation coefficient matrix is provided in Supplementary Table S4.
Construction and validation of the ferroptosis-related gene signature
Candidate genes for model construction were selected by integrating FR-DEGs, hub genes identified from the PPI network, and genes significantly correlated with immune cell infiltration. Least absolute shrinkage and selection operator (LASSO) regression was applied to reduce model complexity and select key variables. A ferroptosis-related gene signature was subsequently constructed using multivariable logistic regression, and a Ferro-Score model was established. Model performance and stability were evaluated through internal validation procedures, including cross-validation, calibration analysis, and decision curve analysis (DCA).
Establishment of the chronic pancreatitis model and experimental grouping
Male C57BL/6 mice (6–8 weeks old) were obtained from the Guangdong Medical Laboratory Animal Center. All animals were housed under specific pathogen-free conditions with a controlled environment (12-h light/dark cycle, temperature of 22–25 °C, and relative humidity of 50–60%) and were allowed free access to standard chow and water throughout the experimental period. After one week of acclimatization, mice were randomly assigned to different experimental groups.
A CP model was established by repeated intraperitoneal injections of cerulein, as previously described with minor modifications [13]. Briefly, mice in the CP group received intraperitoneal injections of cerulein at a dose of 50 µg/kg, administered repeatedly over a period of 6 weeks to induce chronic pancreatic injury. Control mice received intraperitoneal injections of an equivalent volume of phosphate-buffered saline (PBS) according to the same schedule. The injection route, dosage, and modeling duration were selected based on established protocols to ensure stable induction of CP.
Mice were randomly divided into two groups (n = 6 per group): the control group and the CP group. At the end of the modeling period, mice were euthanized within 24 h after the final injection by intraperitoneal administration of an overdose of sodium pentobarbital (200 mg/kg). Pancreatic tissues were rapidly harvested, with portions fixed in 4% paraformaldehyde for histological analysis and the remaining tissues snap-frozen in liquid nitrogen and stored at − 80 °C for subsequent molecular analyses.
The study was approved by the Animal Ethics Committee of Guangdong Medical Laboratory Animal Center (Approval No.D202601-8). The study was conducted in compliance with the ARRIVE guidelines.
H&E staining
Pancreatic tissues were fixed in 4% paraformaldehyde (Servicebio, Wuhan, China; Cat. No. G1101) at room temperature, followed by dehydration through a graded ethanol series and embedding in paraffin. Paraffin-embedded tissues were sectioned at a thickness of 4 μm using a rotary microtome (Leica RM2235, Leica Microsystems, Germany). Tissue sections were deparaffinized, rehydrated, and stained with hematoxylin and eosin using a commercial H&E staining kit (Solarbio, Beijing, China; Cat. No. G1120) according to the manufacturer’s instructions. Stained sections were examined under a light microscope (Olympus BX53, Olympus Corporation, Japan). Histopathological evaluation focused on alterations in acinar architecture, inflammatory cell infiltration, and stromal changes. All histological assessments were performed in a blinded manner.
qRT-PCR analysis
Total RNA was extracted from pancreatic tissues using TRIzol reagent (Invitrogen, Thermo Fisher Scientific, USA; Cat. No. 15596026) according to the manufacturer’s protocol. RNA concentration and purity were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). Complementary DNA (cDNA) was synthesized using a reverse transcription kit (PrimeScript™ RT Reagent Kit, Takara, Japan; Cat. No. RR037A). Quantitative real-time PCR was performed using SYBR® Premix Ex Taq™ II (Takara, Japan; Cat. No. RR820A) on an ABI 7500 Fast Real-Time PCR System (Applied Biosystems, USA). Gene-specific primers were designed and synthesized for ZFP36 and BNIP3, with GAPDH serving as the internal reference gene. The primer sequences (5′–3′) were as follows: ZFP36, forward: CACCCCAAGTACAAAACGGA, reverse: AAAGACAGGTGAGTCTGACC; BNIP3, forward: GACACCACAAGATACCAACAG, reverse: GCGCTTCGGGTGTTTAAAAAGG. Relative gene expression levels were calculated using the 2^−ΔΔCt method.
Statistical analysis
Statistical analyses for experimental data were performed using SPSS software (version 26.0; IBM Corp., Armonk, NY, USA). Continuous variables were expressed as mean ± standard deviation. Prior to analysis, data distribution was assessed for normality. Comparisons between two groups were conducted using Student’s t-test for normally distributed data or the Mann–Whitney U test for non-normally distributed data. All statistical tests were two-tailed, and a P value < 0.05 was considered statistically significant.
Results
Differentially expressed genes in chronic pancreatitis identified from GEO datasets
To identify genes differentially expressed in CP, differential expression analyses were independently performed in the GSE77858 and GSE143754 datasets to minimize cross-platform bias. As shown in Fig. 1A, volcano plot analysis of GSE77858 revealed a distinct distribution of differentially expressed genes between CP and control samples using the predefined cutoff criteria. Using the predefined criteria for GSE77858 (|log2FC| > 0.5, P < 0.05), a total of 81 DEGs were identified, including both upregulated and downregulated genes. The corresponding heatmap (Fig. 1B) demonstrated clear expression pattern differences between CP and control groups, indicating good separation and consistency among samples.
Fig. 1.
Identification of differentially expressed genes (DEGs) in chronic pancreatitis (CP). A Volcano plot of DEGs between CP and control samples in the GSE77858 dataset. Red dots represent significantly upregulated genes, green dots represent significantly downregulated genes, and black dots indicate genes without significant differential expression. B Heatmap showing the expression patterns of DEGs identified in GSE77858. Rows represent genes and columns represent individual samples. C Volcano plot of DEGs between CP and control samples in the GSE143754 dataset .D Heatmap illustrating the expression profiles of DEGs in GSE143754. Samples are annotated by group, and color gradients indicate relative gene expression levels
According to the preset GSE143754 standard (|log2FC| > 1, P < 0.05), compared with normal pancreatic tissue, there are a large number of significantly altered genes in CP tissues, and a total of 414 differentially expressed genes were identified (Fig. 1C). Hierarchical clustering heatmap analysis (Fig. 1D) further illustrated distinct transcriptional profiles between the two groups, with samples clustering according to disease status.
Identification and functional characterization of ferroptosis-related differentially expressed genes in chronic pancreatitis
To identify ferroptosis-related genes associated with CP, differentially expressed genes from the GEO datasets were intersected with the ferroptosis-related gene set obtained from the FerrDb database. As shown in Fig. 2A, a total of 19 genes were shared between the DEGs and ferroptosis-related genes and were defined as FR-DEGs. Functional enrichment analysis of these FR-DEGs was subsequently performed. GO analysis demonstrated that the FR-DEGs were enriched in biological processes related to cellular stress responses, including response to oxidative stress, response to metal ion, cellular response to reactive oxygen species, and glutathione metabolic process (Fig. 2B). KEGG pathway enrichment analysis showed that these genes were involved in multiple pathways, such as the MAPK signaling pathway, autophagy-related pathways, FoxO signaling pathway, and glutathione metabolism (Fig. 2C). In addition, a protein–protein interaction (PPI) network was constructed for the FR-DEGs using the STRING database (Fig. 2D). Based on the PPI network, the CytoHubba plugin in Cytoscape was used to rank genes according to topological characteristics, and the top 10 hub genes were identified as NR4A1, KLF2, DUSP1, ZFP36, ALB, BNIP3, CHAC1, LAMP2, PSAT1, and MGST1.
Fig. 2.
Functional enrichment and protein–protein interaction analysis of ferroptosis-related differentially expressed genes (FR-DEGs). A Venn diagram showing the intersection between differentially expressed genes (DEGs) and ferroptosis-related genes from the FerrDb database, identifying 19 FR-DEGs. B Gene Ontology (GO) enrichment analysis of FR-DEGs, with dot size representing the number of genes enriched in each term and color indicating statistical significance. C Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of FR-DEGs. D Protein–protein interaction (PPI) network constructed for FR-DEGs using the STRING database, visualized in Cytoscape. Nodes represent proteins and edges indicate predicted or known interactions
Immune infiltration analysis and correlation with ferroptosis-related genes in CP
Immune cell infiltration analysis was performed to characterize the immune landscape in CP samples. As shown in Fig. 3A, the relative proportions of 22 immune cell types were estimated across control and CP samples and visualized using stacked bar plots. Differences in immune cell composition between the two groups were further evaluated. Box plot analysis (Fig. 3B) demonstrated that the proportion of plasma cells was significantly increased in CP samples compared with controls (P < 0.05), whereas no statistically significant differences were observed for most other immune cell types. To further explore the association between ferroptosis-related differentially expressed genes (FR-DEGs) and immune cell infiltration, Spearman correlation analysis was performed between gene expression levels and immune cell proportions. As shown in Fig. 3C, a limited number of gene–immune cell pairs exhibited statistically significant correlations. Specifically, NR4A1 and DUSP1 showed positive correlations with CD8⁺ T cells (r = 0.56 and r = 0.59, respectively), while DUSP1 was also positively correlated with T follicular helper cells (r = 0.54). In addition, PSAT1 demonstrated a positive correlation with regulatory T cells (Tregs) (r = 0.52). Most other correlations were weak to moderate (|r| < 0.4) and did not reach statistical significance. These findings suggest that the associations between ferroptosis-related genes and immune cell infiltration are limited and should be interpreted with caution.
Fig. 3.
Immune infiltration landscape and correlation analysis between ferroptosis-related genes and immune cells. A Stacked bar plot showing the relative proportions of 22 immune cell types in control and CP samples. B Box plots comparing immune cell proportions between control and CP groups; (C) Heatmap illustrating correlations between the expression levels of ferroptosis-related differentially expressed genes (FR-DEGs) and immune cell proportions. Color intensity represents the strength and direction of correlation, and values indicate correlation coefficients
Construction and evaluation of the ferroptosis-related gene signature in chronic pancreatitis
To construct and evaluate a ferroptosis-related gene signature, six candidate genes (NR4A1, KLF2, DUSP1, ZFP36, BNIP3, and PSAT1) were included in LASSO regression analysis with five-fold cross-validation. The coefficient trajectories of these candidate genes across a range of log-transformed penalty parameters are shown in Fig. 4A, illustrating the shrinkage of regression coefficients with increasing regularization strength. Cross-validation analysis was conducted to identify the optimal penalty parameter, and the minimum mean squared error was observed at the selected lambda value (Fig. 4B). At this lambda, two genes, ZFP36 and BNIP3, retained non-zero coefficients and were selected for subsequent modeling. A logistic regression model was then constructed based on these two genes and visualized as a nomogram (Ferro-Score), which assigns points to each variable and integrates them to estimate the predicted probability (Fig. 4C). Internal validation of the model was performed using bootstrap resampling with 5,000 repetitions, and calibration analysis demonstrated the relationship between predicted probabilities and observed outcomes (Fig. 4D). Decision curve analysis was further conducted to depict the net benefit of the Ferro-Score model across a range of threshold probabilities (Fig. 4E). The Ferro-Score model achieved an AUC of 1.0 in the training dataset (Supplementary Fig. 1). There may be a risk of overfitting. The results should be interpreted with caution.
Fig. 4.
Construction and performance evaluation of the Ferro-Score. A LASSO regression coefficient profiles of candidate genes plotted against log-transformed lambda values. B Cross-validation curve for LASSO regression showing the relationship between mean squared error and log(lambda). C Nomogram (Ferro-Score) based on ZFP36 and BNIP3 for estimating the predicted probability. D Calibration curve of the Ferro-Score model generated by bootstrap resampling, comparing predicted probabilities with observed outcomes. E Decision curve analysis (DCA) evaluating the net benefit of the Ferro-Score model across different threshold probabilities
Experimental validation of ferroptosis-related genes in a chronic pancreatitis mouse model
To validate the bioinformatics findings at the experimental level, a cerulein-induced CP mouse model was established. Representative H&E–stained sections of pancreatic tissues are shown in Fig. 5A. Compared with the control group, pancreatic tissues from the CP group exhibited altered histological features, including changes in acinar architecture, increased interstitial areas, and inflammatory cell infiltration. To further examine the expression of key ferroptosis-related genes, quantitative real-time PCR analysis was performed. As shown in Fig. 5B, the mRNA expression level of ZFP36 in pancreatic tissues was significantly increased in the CP group compared with the control group (P < 0.01), whereas the expression level of BNIP3 was significantly decreased in the CP group (P < 0.01).
Fig. 5.
Experimental validation of ferroptosis-related genes in a CP mouse model. A Representative H&E-stained pancreatic tissue sections from control and CP mice. Scale bars: 50 μm (upper panels) and 20 μm (lower panels). B Quantitative real-time PCR analysis of ZFP36 and BNIP3 mRNA expression levels in pancreatic tissues from control and CP groups
Discussion
In this study, we systematically characterized the expression landscape of ferroptosis-related immune genes in CP by integrating multi-dataset transcriptomic analyses with experimental validation. We identified 19 FR-DEGs, which were predominantly enriched in oxidative stress–related biological processes and signaling pathways. In addition, an exploratory immune infiltration analysis suggested potential alterations in immune cell composition, with a significant increase in plasma cells in CP. A ferroptosis-related gene signature based on ZFP36 and BNIP3 was further constructed and internally validated. These findings collectively provide a comprehensive overview of ferroptosis-associated transcriptional alterations and their potential links to immune dysregulation in CP.
Ferroptosis has been increasingly recognized as a critical form of regulated cell death involved in inflammatory diseases, metabolic disorders, and organ fibrosis [14, 15]. While previous studies have primarily focused on its role in acute pancreatic injury and pancreatic cancer, its involvement in CP remains insufficiently explored. In the present study, functional enrichment analyses revealed that FR-DEGs were mainly associated with oxidative stress responses, glutathione metabolism, MAPK signaling, FoxO signaling, and autophagy-related pathways. These pathways are well-established contributors to pancreatic injury, inflammatory signaling, and fibrotic remodeling [16, 17], suggesting that ferroptosis-related processes may be embedded within the broader pathological network of CP.
Emerging evidence indicates that metabolic reprogramming plays a critical role in shaping the immune microenvironment [10]. As ferroptosis is fundamentally driven by lipid peroxidation and oxidative stress, it is closely linked to cellular metabolic states. Therefore, our findings support the concept that ferroptosis-related pathways may influence immune cell infiltration through metabolic–immune crosstalk. This relationship provides a potential mechanistic framework linking oxidative stress, ferroptosis, and immune dysregulation in CP.
Consistent with the immune-mediated nature of CP [18, 19], our immune infiltration analysis suggested a shift in immune cell composition, particularly an increase in plasma cells. This observation is in line with previous studies indicating enhanced humoral immune responses in chronic pancreatic inflammation. Furthermore, several FR-DEGs, including NR4A1, DUSP1, ZFP36, BNIP3, and PSAT1, exhibited moderate correlations with T cell subsets and regulatory immune populations. These findings are consistent with the notion that ferroptosis and immune regulation are interconnected processes, in which oxidative stress and lipid peroxidation can influence immune cell activation, while immune-derived inflammatory signals may modulate cellular susceptibility to ferroptosis [14, 20].
Among the identified genes, ZFP36 and BNIP3 were selected as key components of the ferroptosis-related gene signature. ZFP36 is an RNA-binding protein that regulates the stability of inflammatory cytokine mRNAs and plays an important role in immune homeostasis, whereas BNIP3 is involved in mitochondrial stress responses, hypoxia signaling, and metabolic regulation [21]. However, it is important to note that both genes are not classical ferroptosis regulators. Their altered expression observed in this study may reflect broader inflammatory and stress-response processes rather than direct ferroptosis-specific mechanisms. Therefore, their roles in CP should be interpreted as potential mediators linking oxidative stress, metabolism, and immune regulation, rather than definitive drivers of ferroptosis.
Based on these candidate genes, we constructed a Ferro-Score model, which demonstrated excellent discrimination performance in the training dataset. However, this model was only internally validated, and the AUC value of 1.0 suggests a potential risk of overfitting. Therefore, the model should be regarded as an exploratory tool reflecting disease-associated molecular patterns rather than a clinically applicable predictive model. External validation using independent datasets is required to further assess its robustness and generalizability.
Several methodological considerations should also be noted. Due to cross-platform differences between the included datasets, we adopted a conservative analysis strategy by performing differential expression analysis independently within each dataset and subsequently integrating the results. Although this approach reduces platform-driven bias and enhances reproducibility, residual heterogeneity cannot be completely excluded.
Importantly, the immune infiltration analysis in this study has inherent limitations. The CIBERSORTx LM22 signature matrix was originally developed based on blood-derived immune cell profiles and has been more extensively validated in tumor tissues. In contrast, CP is characterized by extensive fibrosis and stromal remodeling, which may compromise the accuracy of computational deconvolution. In addition, due to the lack of available confidence metrics (e.g., CIBERSORTx P-values), we were unable to perform quality filtering of samples. Notably, only plasma cells showed statistically significant differences between groups, while most other immune cell types did not reach significance. Therefore, these results should be interpreted as exploratory and hypothesis-generating rather than definitive evidence of immune alterations in CP. Future studies using tissue-specific deconvolution methods, single-cell RNA sequencing, or experimental validation are warranted.
It should also be emphasized that this study does not provide direct evidence of ferroptotic cell death in CP. Our conclusions are primarily based on transcriptomic data and predefined ferroptosis-related gene sets, which reflect associations rather than causality. Moreover, the experimental validation was limited to mRNA expression analysis, and key ferroptosis markers such as GPX4, ACSL4, and lipid peroxidation indicators were not assessed. Therefore, further studies incorporating protein-level validation and functional assays are required to clarify the role of ferroptosis in CP.
In addition, the ferroptosis-related gene set used in this study included drivers, suppressors, and markers from the FerrDb database. While this approach provides a comprehensive overview, it may introduce bias, as some identified genes may represent indirect associations rather than functional regulators. Future studies focusing on functionally validated ferroptosis regulators are needed.
In addition, this study has several limitations. These include a small sample size, lack of external validation, and the absence of mechanistic experiments. Moreover, the cell type-specific effects and causal relationships between ferroptosis and immune responses remain to be clarified. Future research should include larger cohort studies, functional validations, and cell-specific analyses to further elucidate the role of ferroptosis-related immune mechanisms in CP.
Conclusion
In summary, this study systematically characterized the expression profile of ferroptosis-related immune genes in CP by integrating transcriptomic analyses, immune infiltration assessment, predictive modeling, and experimental validation. The findings highlight distinct ferroptosis-associated molecular features and their associations with immune dysregulation in CP, and identify ZFP36 and BNIP3 as candidate genes associated with ferroptosis-related processes. These results provide novel insights into the potential involvement of ferroptosis-related immune mechanisms in the pathogenesis of CP.
Supplementary Information
Acknowledgements
Not applicable.
Authors’ contributions
Xianrong Du: Conceptualization, Methodology, Supervision, Writing – Original Draft, Writing – Review & Editing. Xiaoyu Cai: Data Curation, Bioinformatic Analysis, Formal Analysis, Writing – Review & Editing. Lin Zhong: Experimental Investigation, Validation, Animal Studies, Data Interpretation, Writing – Review & Editing. All authors have read and approved the final manuscript and agree to its submission to this Journal.
Funding
This study did not receive any funding in any form.
Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The study was approved by the Animal Ethics Committee of Guangdong Medical Laboratory Animal Center (Approval No.D202601-8). The study was conducted in compliance with the ARRIVE guidelines.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Lucocq J, Hughes M. Chronic pancreatitis. Surg (Oxford). 2025;43(1):57–63. [Google Scholar]
- 2.Lin Z, et al. Navigating chronic pancreatitis pain: a pathophysiological and therapeutic overview. Front Physiol. 2025;16:1622845. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Li T, et al. Global and regional burden of pancreatitis: epidemiological trends, risk factors, and projections to 2050 from the global burden of disease study 2021. BMC Gastroenterol. 2024;24(1):398. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Zhang Y, et al. Immune cells and immune cell-targeted therapy in chronic pancreatitis. Front Oncol. 2023;13:1151103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sabu A, et al. Targeted nanotheranostics mitigates radiation-induced fibrosis to promote immune infiltration and enhance radio-chemotherapy in pancreatic ductal adenocarcinoma. Biomater Adv. 2025;182:214677. [DOI] [PubMed] [Google Scholar]
- 6.Li N, et al. Mild microwave hyperthermia promotes mitotic catastrophe, induces time-delayed cGAS-STING activation and restores sensitivity to anti-PDL1 therapy in Pan02 pancreatic cancer model. Oncoimmunology. 2025;14(1):2602216. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Dixon SJ, et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell. 2012;149(5):1060–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dong G, et al. Schisanhenol Inhibits MPTP/MPP(+)-Induced Ferroptosis in Dopaminergic Neurons Via Nrf2/TrxR1/GPX4 Pathway against Parkinson’s Disease. Neurochem Res. 2025;51(1):26. [DOI] [PubMed] [Google Scholar]
- 9.Huang CS, et al. Synergistic Induction of Ferroptosis by GPX4 Inhibitor and GSH-Depleted Nanoparticles Effectively Reverses Gemcitabine Resistance in Pancreatic Ductal Adenocarcinoma PDX Model. Adv Healthc Mater. 2025;15(11):e03694. [DOI] [PubMed]
- 10.Zhang P, et al. Metabolic reprogramming signature predicts immunotherapy efficacy in lung adenocarcinoma: Targeting SLC25A1 to overcome immune resistance. Chin J Cancer Res. 2025;37(6):1000–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gharbi W, et al. Functional genetic variants in immune-checkpoint molecules and susceptibility to Nasopharyngeal carcinoma in a Tunisian case-control study. Mol Biol Rep. 2025;53(1):230. [DOI] [PubMed] [Google Scholar]
- 12.Liu N, et al. Glycolysis inhibition-based breakdown of ferroptosis defenses to achieve ferroptosis and immune cascades for antitumor therapy. Biomaterials. 2025;329:123939. [DOI] [PubMed] [Google Scholar]
- 13.Yan Z, et al. The Effects of Da-Chai-Hu Decoction Alleviating Pancreatic Exocrine Dysfunction by Inhibiting Endoplasmic Reticulum Stress of Acinar Cell in Mice with Chronic Pancreatitis. J Inflamm Res. 2025;18:18009–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Stockwell BR, et al. Ferroptosis: A Regulated Cell Death Nexus Linking Metabolism, Redox Biology, and Disease. Cell. 2017;171(2):273–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Youssef H, et al. Neurotoxic and neurobehavioral impacts of silica nanoparticles on brain tissue of albino rats with the potential ameliorative efficacy of liposomal curcumin. J Mol Histol. 2025;57(1):10. [DOI] [PubMed] [Google Scholar]
- 16.Amer J, et al. Antagonizing IL-6 receptor restores pancreatic tissue resident NK cells activation and ameliorates pancreatic injury in the mouse model of MASH. Front Pharmacol. 2025;16:1611637. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hu C, et al. The unique pancreatic stellate cell gene expression signatures are associated with the progression from acute to chronic pancreatitis. Comput Struct Biotechnol J. 2021;19:6375–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Xie W, et al. From stem cells to nanomedicine: A multimodal approach targeting pancreatic fibrosis via MFGE8-dependent ANXA1-SMAD2/3 axis. Int J Biol Macromol. 2025;338(Pt 1):149698. [DOI] [PubMed] [Google Scholar]
- 19.Li C, et al. Chaihu Guizhi Ganjiang decoction attenuates chronic pancreatitis by suppressing acinar cell ferroptosis via regulating p53/SLC7A11/GPX4 pathway. J Ethnopharmacol. 2026;358:121001. [DOI] [PubMed] [Google Scholar]
- 20.Liu N, et al. Glycolysis inhibition-based breakdown of ferroptosis defenses to achieve ferroptosis and immune cascades for antitumor therapy. Biomaterials. 2026;329:123939. [DOI] [PubMed] [Google Scholar]
- 21.Wei X, et al. Identification of Subtypes and a Delayed Graft Function Predictive Signature Based on Ferroptosis in Renal Ischemia-Reperfusion Injury. Front Cell Dev Biol. 2022;10:800650. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.