ERO1L induces macrophage infiltration to potentiate KrasG12D and inflammation-induced pancreatic tumorigenesis via oxidative protein folding of CCL2 Authors: Junfeng Zhang, Chengyu Hu, Yannan Xu, Macheng Lu, Yahui Ma, Ting Yang, Jing Li, Shu-Heng Jiang, Chengkai Jiang, Pengcheng Chen, Li Yao, Rong Hua Published: 2026 ABSTRACT AbstractPancreatitis is a significant risk factor for pancreatic cancer, with Kras mutations being the most common in this malignancy. Inflammation accelerates tumorigenesis driven by Kras mutations. However, the interactions between inflammation and Kras mutation in the context of pancreatic tumorigenesis remain unclear. In this study, we discovered that under pancreatitis conditions, the endoplasmic reticulum stress-related pathways are downregulated in Kras mutant mice compared to wild-type mice, while endoplasmic reticulum oxidoreductase 1 alpha (ERO1L) is abnormally upregulated. Kras mutation downregulates the ER-associated E3 ubiquitin ligase HRD1, resulting in reduced ubiquitination and degradation of HIF1α, which transcriptionally induces the expression of ERO1L. Inhibition of ERO1L activity or pancreas-specific knockout of the Ero1l gene can delay the progression of pancreatic carcinogenesis. Mechanistically, ERO1L enhances the protein expression of the chemokine CCL2 through its role in oxidative folding, thereby promoting the recruitment of macrophages. Thus, this study reveals that Kras mutation upregulates ERO1L to promote inflammation and pancreatic tumorigenesis, providing new insights and potential targets for the treatment of pancreatic cancer.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13578-026-01601-3.Keywords: Pancreatic ductal adenocarcinoma, Unfolded protein response, ER stress, ERO1α, SYVN1 Introduction Pancreatic cancer is one of the most malignant tumors of the digestive system, characterized by its difficulty in early diagnosis, rapid progression, and poor prognosis [1]. Therefore, elucidating the mechanisms underlying the development of pancreatic cancer and its tumor microenvironment is crucial for seeking early diagnostic and intervention strategies to improve patient outcomes [2]. Numerous clinical studies have confirmed that both chronic pancreatitis and acute pancreatitis are significant risk factors for pancreatic cancer [3]. Patients with chronic pancreatitis have a 16-fold increased risk of developing pancreatic cancer. Clinically, approximately 5% of chronic pancreatitis patients may progress to pancreatic cancer after a 20-year disease course [4]. The Kirsten rat sarcoma viral oncogene homolog (Kras) gene is one of the most commonly mutated genes in pancreatic cancer, with a mutation rate exceeding 90%. Kras mutations lead pancreatic cells into an uncontrolled proliferative state, undergoing processes such as acinar-to-ductal metaplasia (ADM) and pancreatic intraepithelial neoplasia (PanIN), gradually advancing from chronic inflammation and precancerous lesions to invasive cancer. The Kras mutation can promote tumorigenesis, but it typically requires a long latency period. However, in the presence of inflammation or tissue damage, the carcinogenic effects of the Kras mutation can be significantly accelerated [5]. For instance, the KrasG12D/+; Pdx1-Cre mouse model (referred to as KC mice) combined with cerulein-induced pancreatitis is highly suitable for studying pancreatic tumorigenesis [6]. Endoplasmic reticulum stress (ERS) refers to a cellular state induced by adverse environmental factors, such as hypoxia, fluctuations in pH, or deficiencies in essential nutrients like glucose or amino acids [7]. These conditions impede the proper folding of peptide chains within the endoplasmic reticulum, leading to the accumulation of unfolded or misfolded proteins and placing the cell under stress. The presence of unfolded or misfolded proteins activates specific receptors on the endoplasmic reticulum membrane, which in turn triggers related protein kinases or initiates signal transduction pathways that regulate gene expression, culminating in various biological effects. This phenomenon is known as the unfolded protein response (UPR) [8]. The UPR maintains cellular homeostasis through three primary mechanisms: reducing protein synthesis, enhancing protein folding and secretion, and promoting the degradation of endoplasmic reticulum-associated proteins. However, if the stress persists for an extended period or is excessively severe, the cellular environment cannot be restored in a timely manner, leading to apoptosis as a consequence of ERS [9]. Endoplasmic Reticulum Oxidoreductase 1 Alpha (ERO1L) is a flavin adenine dinucleotide-dependent enzyme that exhibits enhanced expression under conditions of endoplasmic reticulum stress, particularly pronounced during hypoxic stress [10, 11]. Previously, we have shown that ERO1L is induced by ER stress and hypoxia and promotes pancreatic cancer progression via enhancing the Warburg effect [12]. Proteins must form disulfide bonds within the endoplasmic reticulum before being secreted. The formation of these disulfide bonds is facilitated by the oxidative processes carried out by the Protein Disulfide Isomerase (PDI) family. Upon oxidizing proteins, PDIs themselves undergo reduction, necessitating the role of ERO1L to oxidize PDIs, thereby sustaining the protein folding process [13, 14]. Moreover, ERO1L can modulate many target proteins via oxidative folding, such as CXCL1, CXCL2, G-CSF, and MHC class I heavy chains [15–18]. In this study, based on bioinformatics analysis, we discovered that under conditions of cerulein-induced pancreatitis, KrasG12D mutation inhibits the ERS response. Notably, ERO1L was found to be abnormally overexpressed. Conditional knockout or pharmacological inhibition of ERO1L effectively suppressed pancreatic cancer development. Further mechanistic investigations demonstrated that ERO1L promotes the production of the chemokine CCL2 through oxidative protein folding, thereby facilitating macrophage infiltration and contributing to pancreatic tumorigenesis. Materials and methods Data mining of GEO datasets GSE65146, which includes expression data from cerulein-treated wild-type mice and KrasG12D-mutated mice at various stages of pancreatic regeneration following inflammatory injury, was utilized to identify key factors involved in KrasG12D and inflammation-driven pancreatic tumorigenesis [19]. This dataset is accessible through the GEO database (https://www.ncbi.nlm.nih.gov/geo/). In brief, the raw data and corresponding metadata were downloaded, followed by preprocessing that involved quality control, normalization, and filtering. Statistical analyses were conducted to identify differentially expressed genes (DEGs), which were then visualized using heatmaps. Additionally, the DEGs were annotated through Gene Ontology enrichment analysis. DEGs were annotated as fold change ≥ 2 and an adjusted p-value < 0.05. All steps were performed using R/Bioconductor. Cell culture and reagents Pancreatic cancer cell lines, including AsPC-1 (RRID: CVCL_0152), Capan-2 (RRID: CVCL_0026), MiaPaCa-2 (RRID: CVCL_0428), SW1990 (RRID: CVCL_1723), CFPAC-1 (RRID: CVCL_1119), PANC1 (RRID: CVCL_0480), and BxPC-3 (RRID: CVCL_0186) were obtained from the Cell Resource Center at the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences (Shanghai, China) or ATCC. THP-1 cells (RRID: CVCL_0006) were reported previously [20]. These cells were cultured following protocols established by the American Type Culture Collection (ATCC), with growth medium supplemented with 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) streptomycin-penicillin (Sigma-Aldrich, Shanghai, China). All cell lines were authenticated before each experiment. The cultures were maintained in a humidified incubator at 37 °C with 5% CO2. EN460 (E1698) and AZD6244 (S1008) were purchased from Selleck (Shanghai, China), and methyl methanethiosulfonate (MMTS) was obtained from Pierce (#23011). Prior to the experiments, the cells were verified based on their morphology and checked to confirm that they were free from Mycoplasma contamination. Clinical specimens and immunohistochemistry (IHC) IHC analysis and its scoring methodology have been described previously [21]. To investigate the association between ERO1L expression and macrophage infiltration, we evaluated CD68+ macrophages using immunohistochemistry (IHC) techniques. Additionally, F4/80+ macrophages, CD86+ macrophages, and CD206+ macrophages in pancreatic tissues from KC mice were analyzed. The IHC procedure was performed as follows: after de-paraffinization, antigen retrieval was conducted by boiling the slides in 10 mM citrate buffer (pH 6.0) at 100 °C for 15 min. Endogenous peroxidase activity was blocked by immersing the slides in a 0.3% H2O2-methanol solution at room temperature for 20 min. Subsequently, the slides were incubated with 5% horse serum to prevent non-specific binding. Primary antibodies, including anti-F4/80 (28463-1-AP; Proteintech; RRID: AB_2881149), anti-CD68 (28058-1-AP; Proteintech; RRID: AB_2881049), anti-CD86 (Cat# 20018, Cell Signaling Technology, RRID: AB_3698949), anti-CD206 (Cat# 76220, Cell Signaling Technology, RRID: AB_3697062) were applied. The following day, the slides were incubated with HRP-conjugated secondary antibodies for 1 h at room temperature, followed by the addition of 3,3’-diaminobenzidine tetrahydrochloride (DAB) for color development. Finally, the slides were counterstained with hematoxylin. The number of indicated macrophages per mm² was scored, and these scores were independently evaluated by two senior pathologists who were blinded to the clinical and pathological data, as well as the experimental group assignments. All research protocols involving the use of human materials were approved by the ethics committee at Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, and written informed consent was obtained from all patients prior to their participation. Gene knockdown or overexpression Pancreatic cancer cells expressing short hairpin RNAs (shRNAs) targeting ERO1L were generated following previously established protocols. To achieve the overexpression of HRD1, wild-type (WT) ERO1L, and its inactive mutant C394A, specific plasmids were constructed by cloning the respective genes into the pCDH-CMV-MCS-EF1-Puro vector. This resulted in the generation of three distinct overexpression plasmids: pCDH-HRD1 for HRD1, pCDH-ERO1L for wild-type ERO1L, and pCDH-ERO1L-C394A for the inactive mutant. For the transfection of these plasmids into pancreatic cancer cells, Lipofectamine® RNAiMAX reagent (ThermoFisher Scientific, #13778030, USA) was utilized, following the manufacturer’s instructions. For ectopic expression of mutant Kras, human mutant K-ras4BG12V and wild type K-ras4B cDNAs were synthesized and inserted into the lentiviral transfer vector pCDH-CMV-MCS-EF1-Puro (Shanghai Generay Biotech Co., Ltd., Shanghai, China). Lentiviral packaging was performed in 293T cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA), and viral titers were determined according to standard packaging protocols. BxPC-3 cells were infected with 1 × 106 recombinant lentivirus-transducing units in the presence of 6 µg/mL polybrene (Sigma-Aldrich, H9268). Two days post-infection, transduced cell populations were selected with 500 ng/mL puromycin (Sigma-Aldrich, P8833). Plate colony formation assay To assess the impact of HRD1 overexpression on the proliferation of pancreatic cancer cells, a plate colony formation assay was performed. Briefly, cells were seeded at a density of 1,000 cells per well in a 6-well plate and incubated at 37 °C with 5% CO2 for 9–12 days to facilitate colony formation. The culture medium was refreshed every 3 days to maintain optimal growth conditions. At the end of the incubation period, colonies were fixed using methanol and subsequently stained with 0.1% crystal violet for visualization. Real-time qPCR analysis Total RNA was isolated from pancreatic cancer cells using RNAiso Plus (Takara, Japan). Following RNA extraction, complementary DNA (cDNA) was synthesized using the PrimeScript RT-PCR Kit (TaKaRa), adhering to the manufacturer’s guidelines. For quantitative real-time PCR (qRT-PCR), SYBR Premix Ex Taq (Takara, Japan) was employed on a 7500 Real-time PCR system (Applied Biosystems, Inc., USA) to quantify gene expression levels. The relative expression of target genes was determined using the 2-ΔΔCt method, with ACTB (β-actin) as an internal control to normalize the data. The specific sequences of the primers were shown as follows: ERO1L forward, 5′-GACTGTGCTGTCAAACCATGT-3′; ERO1L reverse, 5′-TCCAAGTCGTTCAGCTTGTTC-3′; SYVN1 forward, 5′-CTTCACCGTTTTTCGGGATGA-3′; SYVN1 reverse, 5′-CCAGGAGGAACATAAGAGAGACA-3′; CCL2 forward, 5′-CAGCCAGATGCAATCAATGCC-3′; CCL2 reverse, 5′-TGGAATCCTGAACCCACTTCT-3′; HSPA5 forward, 5′-GAAAGAAGGTTACCCATGCAGT-3′; HSPA5 reverse, 5′-CAGGCCATAAGCAATAGCAGC-3′; DDIT3 forward, 5′-GAACGGCTCAAGCAGGAAATC-3′; DDIT3 reverse, 5′-TTCACCATTCGGTCAATCAGAG-3′; β-actin forward, 5′-CATGTACGTTGCTATCCAGGC-3′, β-actin reverse, 5′-CTCCTTAATGTCACGCACGAT-3′. Western blot analysis Whole-cell extracts were prepared from pancreatic cancer cells using a lysis buffer (Beyotime, Shanghai, China) that included proteinase inhibitors. The protein concentration in the extracts was measured using a BCA Protein Assay Kit (Pierce Biotechnology, USA). Subsequently, the protein samples were separated via SDS-PAGE on gels ranging from 6% to 12% and then transferred onto a nitrocellulose membrane (Millipore, MA, USA). To prevent nonspecific binding, the membrane was blocked with a solution of 5% skimmed milk. Following this, the membrane was incubated with primary antibodies targeting ERO1L (1:1,000, Abcam, ab177156; RRID: AB_2941804), HIF1α (1:1,000, Abcam, ab113642; RRID: AB_10862695), HRD1 (1:1,000; Proteintech, 13473-1-AP; RRID: AB_2287023), and β-actin (1:2,000, Abcam, ab8226; RRID: AB_306371). The following day, the membranes were treated with species-specific secondary antibodies (ThermoFisher Scientific, USA). Finally, the blots were visualized using the Odyssey imaging system (LI-COR Biosciences, Lincoln, NE, USA). Co-immunoprecipitation (Co-IP) experiments 293T cells that were engineered to overexpress exogenous Flag-HIF1α, His-Ub, and HA-HRD1 were lysed to obtain cell extracts. These lysates were then incubated overnight at 4 °C with Flag beads (MCE, HY-K0207). After incubation, the supernatant was removed, and the beads were washed three times with PBS to eliminate any unbound proteins. Following the washes, 1× SDS loading buffer was added to the beads, which were subsequently heated at 100 °C for 10 min to release the proteins that were bound to the beads. The eluted proteins were then analyzed using Western blot with the following antibodies: anti-Flag (1:5,000; Proteintech, 20543-1-AP; RRID: AB_11232216), anti-His (1:1,000; Proteintech, 10001-0-AP; RRID: AB_11232228), anti-HA (1:5,000; Proteintech, 51064-2-AP; RRID: AB_11042321), anti-CCL2 (1:1,000; Cell Signaling Technology, Cat# 81559, RRID: AB_3271583), and anti-PDI (1:1,000; Proteintech, 11245-1-AP; RRID: AB_2298937). Chemokine antibody assay Conditional medium obtained from sh-Ctrl or sh-ERO1L-1 Capan-2 cells was analyzed using the Human Chemokine Antibody Array Kit C1 (Raybiotech, AAH-CHE-1-2) according to the manufacturer’s instructions. Briefly, the arrays were first treated with a blocking buffer to prevent nonspecific binding. Subsequently, 1 ml of undiluted conditional medium was added and incubated at 4 °C overnight. Following this, the arrays were incubated with a cocktail of biotin-conjugated antibodies for 2 h at room temperature. After three washing steps, the membranes were treated with HRP-Streptavidin for 2 h at room temperature. A chemiluminescent substrate was then applied, and the membranes were exposed to X-ray film for 15 min before processing. Quantitative analysis of the array data was performed using Array Vision Evaluation 8.0 (GE Healthcare Life Science). ELISA analysis of CCL2 The protein concentration of CCL2 in the conditional medium obtained from shCtrl or shERO1L Capan-2 cells was measured using the Human CCL2/MCP-1 Quantikine ELISA Kit (R&D Systems, DCP00; RRID: AB_2894843), following the instructions provided by the manufacturer. THP-1 cell migration assay THP-1 cells (1 × 104) were placed in the upper section of Transwell chambers, with the lower section containing RPMI 1640 medium along with the indicated treatments. Conditioned medium from pancreatic cancer cells was added to the lower section. Following a 12-h incubation of the THP-1 cells, the migrated cells were treated with 2 µM Calcein-AM (MCE, HY-D0041) for 15 min and subsequently observed under a fluorescence microscope. Measurement of Ras activity Ras activity was measured using the Ras GTPase ELISA Kit (Abcam, ab134640) according to the manufacturer’s instructions. Briefly, Raf-RBD–GST fusion protein coated on a glutathione plate captures active Ras (RAS-GTP) from cell extracts while inactive Ras does not. After binding, samples are probed with anti-Ras antibody, followed by HRP-conjugated secondary antibody and chemiluminescent detection. Cell lysates were prepared with Complete Lysis/Binding Buffer supplemented with protease inhibitors. After incubation at 4 °C (binding) and 20–25 °C (antibody steps), luminescence was detected. Animal experiments Pdx1-Cre and LSL-KrasG12D/+ mice were preserved in our laboratory [22, 23]. Ero1lflox/flox mice (Strain S-CKO-11363) were sourced from Cyagen. The LSL-KrasG12D/+; Pdx-1-Cre (KC) mice were produced by breeding LSL-KrasG12D/+ mice with Pdx-1-Cre mice. To establish the pancreatitis model, KC, KEC mice, and their littermates received six hourly intraperitoneal injections of cerulein at a dose of 50 µg/kg body weight (Selleck, Cat. S9690) three times per week for four weeks. For EN460 treatment, a dosage of 10 mg/kg EN460 was administered intraperitoneally three times weekly for a total duration of four weeks. Mice were randomized into DMSO or EN460 group. Mice of specific genotypes were chosen for the study, with both female and male mice included. All mice were housed in a specific pathogen-free environment, kept on a 12-hour light-dark cycle, with a temperature maintained at 23 ± 1 °C and relative humidity between 40 and 75%. All mice were given standard chow and access to water throughout the study period. All animal experiments received Institutional Animal Care and Use Committee approval. Sample sizes were determined based on pilot experiments, and power analysis using the G*Power software (Version. 3.1.9.7). Statistical analysis The results are expressed as mean ± standard deviation (SD). All experiments were repeated at least three times. Comparisons between two groups were conducted using a two-tailed unpaired t-test, while a one-way ANOVA followed by Tukey’s multiple comparisons was employed for three or more groups. Statistical analyses were performed using GraphPad Prism (GraphPad Software Inc., San Diego, CA). The correlation between genes was assessed using Spearman’s correlation coefficient. A P value of less than 0.05 was deemed statistically significant. Results Kras mutation modulates ERS-related pathways and ERO1L Wild-type C57BL/6J mice subjected to cerulein exhibit histological changes that follow a self-limiting process, which can be categorized into three distinct phases: the inflammatory phase (3–36 h), the regenerative phase (2–6 days), and the recovery phase (7–14 days) [19]. In contrast, KC mice demonstrate a different trajectory; initially, they display inflammatory changes similar to those observed in wild-type mice, but this is swiftly followed by the emergence of ADM and PanIN, which continue to progress without returning to normal tissue architecture. Utilizing GSE65146, which contains expression data of cerulein-treated wild-type mice and KrasG12D-mutated mice at different times of inflammatory injury, we analyzed DEGs between the two groups of mice at two time points: the regeneration phase of pancreatitis (60 h) and the recovery phase (14 days). Pathway analysis revealed that the upregulated DEGs were predominantly associated with pathways related to inflammatory responses, leukocyte chemotaxis, and wound healing, which aligns with the sustained inflammation observed in KrasG12D-mutated mice (Fig. 1A and Table S1-2). Conversely, the downregulated DEGs were primarily concentrated in pathways linked to cellular response to unfolded protein, endoplasmic reticulum unfolded protein response, and ERAD pathway (Fig. 1B and Table S3-4). With a particular focus on ERS-related genes (Fig. 1C), we identified the intersection of these genes and found that the majority were downregulated, while ERO1L remained unchanged (Fig. 1D). Given that ERO1L can promote tumor progression by sustaining redox balance and disulfide-bond formation in secretory/membrane proteins, and can modulate immune processes [24, 25], we further measured its protein expression pattern. To achieve this, we generated mouse models by employing KC mice combined with cerulein-induced pancreatitis and harvested pancreas tissues at the regenerative phase (60 h) and the recovery phase (14 days) (Fig. 1E). Different from the observation at mRNA level, we revealed that ERO1L protein levels were markedly increased in KC mice at both two time points through IHC analysis (Fig. 1F), suggesting a potential regulatory role of Kras mutation on ERO1L at post-transcriptional level. Therefore, we selected ERO1L as the candidate gene for further investigation. Kras mutation modulates ERS-related pathways and ERO1L. A, B GO annotation of upregulated (A) and downregulated DEGs (B) identified by KC mice versus WT mice at the regenerative phase (60 h) and the recovery phase (14 days). DEGs were annotated as fold change ≥ 2 and an adjusted p-value < 0.05. C A schematic model showing genes involved in ER stress and unfolded protein response. D Expression pattern of ERS-related genes at the regenerative phase and the recovery phase. E KC or wild-type mice were sacrificed at time points ranging from 60 to 14 days following two consecutive cerulein injections. F IHC staining showing ERO1L expression in pancreas tissues harvested from the regenerative phase and the recovery phase. Scale bar: 50 μm ERO1L expression is regulated by Kras-HRD1-HIF1α axis Our prior research has reported that ERO1L expression is most pronounced under hypoxic conditions and is regulated by HIF1α [12]. Ubiquitination is one of the most common regulatory mechanisms for HIF1α [26]. Interestingly, there was also a notable downregulation of the endoplasmic reticulum-associated degradation (ERAD) pathway related to Kras mutation (Fig. 1B). Therefore, we tested whether ubiquitination of HIF1α is implicated in the regulation of ERO1L in the context of Kras mutation. To explore this, we utilized UbiBrowser (http://ubibrowser.ncpsb.org/) to predict potential E3-ubiquitin ligases for HIF1α. The result yielded 1 known E3 ligase in literature (VHL) and 25 predicted E3 ligases in M.musculus (Table S5). We then intersected these targets with ERS-related DEGs across the two time points, identifying only one E3-ubiquitin ligase, namely SYVN1 (Fig. 2A). The SYVN1 gene encodes a protein known as synoviolin, also referred to as HMG-CoA reductase degradation 1 homolog (HRD1), which is an E3-ubiquitin ligase. HRD1 is a transmembrane protein that forms a complex with SEL1L, allowing it to recognize and ubiquitinate target proteins, thus participating in the ERAD pathway [27–29]. Indeed, overexpression of HRD1 resulted in a significant downregulation of both HIF1α and ERO1L protein levels in Capan-2 and MiaCaPa-2 cells (Fig. 2B). The long-term proliferation assay showed that HRD1 overexpression inhibited the proliferative capacity of Capan-2 and MiaCaPa-2 cells (Fig. 2C). Given that HRD1 participates in the ERAD pathway and is tightly coupled to ERS responses, we sought to exclude the possibility that the observed phenotypes resulted from non-specific transfection-induced stress. Transient plasmid transfection achieved a modest (~ 2-fold) upregulation of HRD1 (Fig. 2A) without inducing canonical ER stress markers (HSPA5/BiP or DDIT3/CHOP) or affecting short-term cell growth (48 h) (Supplementary Fig. 1). Compared to WT mice, KC mice displayed reduced HRD1 expression in the context of pancreatitis at both the regenerative phase and the recovery phase (Fig. 2D). ERO1L expression is regulated by Kras-HRD1-HIF1α axis. A Venn diagram illustrating the overlapping genes associated with E3-ubiquitin ligases and the downregulated DEGs observed at the regenerative phase (60 h) and recovery phase (14 days). B Western blotting analysis of HIF1α and ERO1L protein in Capan-2 and MiaPaCa-2 cells after overexpression of HRD1. C Plate colony formation assay showed the effect of HRD1 overexpression on the cell proliferation of Capan-2 and MiaPaCa-2 cells. D IHC staining showing HRD1 expression in pancreas tissues harvested from the regenerative phase and the recovery phase. Scale bar: 50 μm. E Co-transfect HEK293T cells with His-Ubiquitin, Flag-HIF1α, and HA-HRD1 plasmids. Immunoprecipitate HIF1α using anti-His antibody and detect ubiquitination of HIF1α via western blot with anti-HA antibody. F In MiaPaCa-2 cells, the effect of HRD1 knockdown on HIF1α ubiquitination. G In Capan-2 cells, the effect of HRD1 overexpression on the protein expression of HIF1α under MG132 treatment. *P < 0.05, ****P < 0.0001 To elucidate the mechanism by which HRD1 regulates HIF1α, we co-transfected His-Ubiquitin, Flag-HIF1α, and HA-HRD1 expression plasmids into HEK293T cells. By performing immunoprecipitation of HIF1α using an anti-His antibody and subsequently detecting the ubiquitination of HIF1α through Western blotting with an anti-HA antibody, we found that HRD1 can indeed modify HIF1α through ubiquitination (Fig. 2E). Then, we knocked down endogenous HRD1 in MiaPaCa-2 cells and observed a significant reduction in the ubiquitination level of endogenous HIF1α, supporting an endogenous role for HRD1 in HIF1α ubiquitination (Fig. 2F). Moreover, we overexpressed HRD1 in Capan-2 cells and treated the cells with the proteasome inhibitor MG132. Under MG132 treatment, the HRD1-overexpression–induced decrease in HIF1α was rescued toward the level seen in vector-control cells, indicating that HRD1 promotes proteasome-dependent degradation of HIF1α (Fig. 2G). These findings suggest that HRD1 can downregulate HIF1α protein levels via ubiquitination, thereby leading to the subsequent suppression of ERO1L expression. KRAS suppresses HRD1 expression at the mRNA level via transcriptional repression To explore how KRAS mutation regulates HRD1 expression, we first compared HRD1 mRNA levels in KRAS-mutant cells (Capan-2, MiaCaPa-2, SW1990, AsPC-1, CFPAC-1, and PANC1) versus KRAS-wild-type cells (BxPC-3) by qRT-PCR. HRD1 mRNA was much lower in KRAS-mutant cells (Fig. 3A). To test whether KRAS signaling controls HRD1 transcription, we treated KRAS-mutant cells with a MEK inhibitor (selumetinib/AZD6244) to block downstream KRAS signaling. The result showed that MEK inhibition consistently increased HRD1 mRNA levels (Fig. 3B). Moreover, we overexpressed human mutant K-ras4BG12V and wild type K-ras4B in BxPC-3 cells. K-ras4BG12V exhibited much higher Kras activity and decreased HRD1 mRNA level compared to wild type K-ras4B (Fig. 3C). Blocking Kras activity with AZD6244 further abolished the K-ras4BG12V-driven decrease in HRD1 mRNA in BxPC-3 cells (Fig. 3D), indicating that Kras primarily downregulates HRD1 at least at the transcriptional level. Together, these findings support a model in which KRAS–MEK–ERK signaling suppresses HRD1 transcription. KRAS suppresses HRD1 expression at the mRNA level via transcriptional repression. A Real-time qPCR analysis of HRD1 mRNA levels in KRAS-mutant cells (Capan-2, MiaCaPa-2, SW1990, AsPC-1, CFPAC-1, and PANC1), KRAS-wild-type cells (BxPC-3), and the non-malignant HPDE6c7 cells. B Real-time qPCR analysis of HRD1 mRNA levels in KRAS-mutant cells upon AZD6244 treatment. C Measurement of Ras activity in BxPC-3 cells expressing wild type K-ras4B and K-ras4BG12V. D In BxPC-3 cells, real-time qPCR analysis of the effects of K-ras4BG12V on the mRNA level of HRD1 in the presence or absence of AZD6244 treatment. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 Targeting ERO1L mitigates KrasG12D and inflammation-induced pancreatic tumorigenesis To further confirm whether ERO1L acts as a driving factor in this process, we constructed Ero1lflox/flox transgenic mice and crossed them with KrasLSL−G12D/+ mice and Pdx1-Cre mice to generate KrasLSL−G12D/+; Ero1lflox/flox; Pdx1-Cre (KEC mice) (Fig. 4A). At week 8, both KC and KEC mice were injected with cerulein, and the treatment continued for 4 weeks (Fig. 4B). By week 12, pancreas tissues were harvested and we observed that the pancreatic tissue of KEC mice exhibited reduced areas of PanIN lesions compared to KC mice (Fig. 4C). EN460 has been identified as a pharmacological inhibitor of ERO1L, with demonstrated selectivity and potent inhibitory efficacy across diverse cellular contexts [30, 31]. To further investigate the functions of ERO1L, we administered cerulein to the KC mice at week 8 while simultaneously treating them with EN460 (Fig. 4D). Upon analysis at week 12, we noted a significant presence of PanIN lesions in the control group, while only a sparse number of PanIN lesions were detected in the EN460-treated group (Fig. 4E). Across these in vivo experiments, both male and female mice were included, and sex did not significantly influence the observed phenotypes (Supplementary Fig. 2). Together, these data support that pharmacological inhibition of ERO1L is also effective to mitigate KrasG12D and inflammation-induced pancreatic tumorigenesis. Targeting ERO1L mitigates KrasG12D and inflammation-induced pancreatic tumorigenesis. A A schematic model for generation of KC and KEC mice. B Cerulein treatment schedule in KC and KEC mice (n = 7 per group). C The PanIN lesions within the pancreas of KC and KEC mice were analyzed by H&E staining. D Treatment schedule of cerulein and EN460 in KC and KEC mice (n = 9 per group). E KC mice were treated with cerulein and the effect of EN460 on the development of PanIN lesions were analyzed by H&E staining. ***P < 0.001, ****P < 0.0001 ERO1L induces macrophage infiltration Macrophage infiltration is essential for inflammation-induced pancreatic cancer, as macrophage deletion compromises the malignant transformation in the context of Kras mutation and pancreatic injury. Consistently, we found that either ERO1L knockdown or pharmacological inhibition with EN460 markedly reduced the infiltration of F4/80+ macrophages (Fig. 5A and B). Moreover, ERO1L deficiency or inhibition reduced the frequency of M2-polarized macrophages without significantly impacting M1 polarization (Fig. 5A and B). In clinical settings, we analyzed the correlation between ERO1L expression and immune cell populations using the TCGA database and revealed a significant relationship between ERO1L expression and macrophage infiltration (Fig. 5C). To further substantiate this connection, we utilized a pancreatic cancer cohort (n = 62) to examine the expression patterns of ERO1L alongside the macrophage marker CD68. As a result, ERO1L expression score was closely associated with CD68+ macrophage infiltration in pancreatic cancer (Fig. 5D). Moreover, we harvested the conditioned medium (CM) from sh-ERO1L and sh-Ctrl pancreatic cancer cells. In vitro Transwell experiments demonstrated that the CM from sh-ERO1L cells significantly enhanced THP-1 cell migration compared to the CM from sh-Ctrl cells (Fig. 5E). Collectively, these data suggest a connection between ERO1L and macrophage infiltration in pancreatic cancer. ERO1L induces macrophage infiltration. A IHC staining showing F4/80+, CD86+, and CD206+ macrophage infiltration in the KC and KEC pancreas tissues harvested from the recovery phase. Scale bar: 50 μm. B In the recovery phase, F4/80+, CD86+, and CD206+ macrophage infiltration in the DMSO-treated or EN460-treated KC pancreas tissues were analyzed by IHC method. Scale bar: 50 μm. C Analyzing the correlation between ERO1L expression and various immune cells using the TCGA database. D Representative IHC staining images of ERO1L and CD68 in the pancreatic cancer specimens from Ren Ji cohort, scale bar: 100 μm. The correlation curve shows the correlation between ERO1L H-score and macrophage marker (CD68) in pancreatic cancer tissues. E Chemotaxis assay and statistical results of human macrophages (THP-1) treated with conditioned medium from sh-ERO1L and shCtrl pancreatic cancer cells (Capan-2 and MiaPaCa-2). ****P < 0.0001 ERO1L regulates CCL2 through oxidative folding To determine whether ERO1L regulates macrophage infiltration through the modulation of chemokine expression, we employed a chemokine array to assess the differences in the secretion levels of various chemokines between sh-ERO1L-1 and sh-Ctrl Capan-2 cells. The results indicated that the secretion level of CCL2 was significantly suppressed in the ERO1L knockdown group (Fig. 6A). We further focused on CCL2, as it plays a key role in promoting macrophage infiltration in pancreatic cancer. Real-time qPCR revealed that ERO1L did not influence the expression level of CCL2 at mRNA level (Fig. 6B). However, ELISA experiments revealed a significant reduction in the secretion level of CCL2 after ERO1L knockdown (Fig. 6C), which implies a regulatory role of ERO1L in CCL2 expression at the post-transcriptional level. ERO1L regulates CCL2 through oxidative folding. A Chemokine array analysis of conditioned medium from sh-ERO1L and shCtrl Capan-2 cells. B Real-time qPCR analysis of CCL2 mRNA expression after ERO1L knockdown. C ELISA detection of chemokine CCL2 concentration in the conditioned medium from Capan-2 and MiaPaCa-2 cells after ERO1L knockdown. D Western blot analysis of the effect of ERO1L knockdown on the redox state of CCL2. E Western blot analysis of the effect of overexpressing either wild type ERO1L or the inactive mutant C394A on the redox state of CCL2 in AsPC-1 cells. F co-IP experiment showed the ERO1L-PDI and ERO1L-CCL2 interactions in Capan-2 and MiaPaCa-2 cells **P < 0.01, ****P < 0.0001 Oxidative protein folding, which involves the formation of disulfide bonds, is one of the most prevalent post-translational modifications and is crucial for the proper conformation and function of secretory proteins and cell surface molecules [32]. ERO1L and protein disulfide isomerase (PDI) play significant roles in the oxidative folding of cytokines and certain receptors, thereby influencing the secretion dynamics of cytokines and the expression levels of receptors on the cell surface. This, in turn, impacts the infiltration and activity of immune cells, such as CD8+ T cells. Given that CCL2 contains intramolecular disulfide bonds and that the formation of these bonds is closely related to its biological activity, we hypothesized that ERO1L may influence the oxidative folding of CCL2. To test this hypothesis, we conducted Western blotting to assess the impact of ERO1L on the redox state of CCL2. To maintain a non-reducing environment, we treated the cells with methyl methanethiosulfonate (MMTS) [33]. Then, the expression levels of oxidized CCL2 in both sh-ERO1L-1 and sh-Ctrl Capan-2 cells were measured. The results indicated that the ratio of oxidized CCL2 to reduced CCL2 in the ERO1L knockdown group was significantly lower than that in the control group (Fig. 6D). To further validate this finding, we overexpressed ERO1L and its inactive mutant C394A in AsPC-1 cells. Notably, the ratio of oxidized CCL2 to reduced CCL2 increased with ERO1L overexpression, whereas no such increase was observed with the inactive mutant C394A (Fig. 6E). To determine whether ERO1L directly interacts with CCL2 or acts indirectly via PDI family proteins, we performed co-IP experiment. The result showed a robust association between ERO1L and PDI, whereas direct ERO1L-CCL2 interaction was not detected. These findings suggest that ERO1L modulates CCL2 secretion predominantly through PDI-mediated oxidative folding rather than direct binding to CCL2 (Fig. 6F). Discussion Pancreatic cancer remains one of the most aggressive malignancies, characterized by late diagnosis and poor prognosis. Understanding the molecular mechanisms that underpin its development is essential for identifying potential therapeutic targets and improving patient outcomes. The findings presented in the present study elucidate the pivotal role of ERO1L in KrasG12D mutation and inflammation-driven pancreatic tumorigenesis. ERO1L expression is significantly upregulated in the context of KrasG12D mutations during pancreatitis. The persistence of ERO1L under inflammatory conditions suggests that it may play a crucial role in mediating the transition from chronic inflammation to cancer. The observation that ERO1L facilitates oxidative protein folding, particularly for proteins involved in immune responses, highlights its importance in maintaining the cellular environment conducive to tumor progression. The relationship between ERO1L and the Kras mutation appears to be multifaceted. ERO1L not only supports the proper folding of proteins necessary for cellular functions but also seems to be intricately linked to the hypoxic response, which is prevalent in tumor microenvironments [34, 35]. The upregulation of ERO1L under hypoxia, driven by HIF1α, further indicates that Kras mutations may enhance cellular adaptations to stress, thereby promoting tumorigenesis. Furthermore, the regulation of ERO1L by the Kras-HRD1-HIF1α axis points to a complex interplay between oncogenic signaling and the cellular stress response. The downregulation of HIF1α via HRD1-mediated ubiquitination may represent a feedback mechanism that regulates ERO1L expression, thereby influencing the oxidative environment within the ER and affecting overall protein homeostasis. Chronic inflammation, as seen in conditions like pancreatitis, significantly increases the risk of developing pancreatic cancer [36]. Our findings demonstrate that ERO1L is abnormally overexpressed during cerulein-induced pancreatitis, a model that mimics the inflammatory environment associated with pancreatic cancer. The role of ERO1L in facilitating macrophage infiltration through the regulation of chemokines, particularly CCL2, highlights its involvement in the inflammatory milieu that supports tumor progression (Fig. 7). Previously, our loss-of-function and gain-of-function studies demonstrated that ERO1L promotes PDAC cell proliferation in vitro and facilitates tumor growth in vivo. Specifically, ectopic expression of wild-type ERO1L but not the catalytically inactive mutant ERO1LC394A promotes tumor growth [12]. ERO1L can also enhance cancer metastasis by activating the epithelial–mesenchymal transition program [37]. Together with our current findings, these results indicate that ERO1L plays versatile oncogenic roles in PDAC across different stages of disease. Mechanism model of ERO1L and pancreatic tumorigenesis. Under pancreatitis, Kras mutation can inhibit the endoplasmic reticulum stress response and downregulate the expression of HRD1. The reduced HRD1 decreases the ubiquitination and degradation of HIF1α, resulting in increased HIF1α protein levels and enhanced ERO1L expression. ERO1L promotes the secretion of CCL2, facilitating macrophage infiltration, and thereby promoting pancreatic tumorigenesis Macrophages are key players in the tumor microenvironment, often promoting tumor growth and metastasis through various mechanisms, including the secretion of pro-inflammatory cytokines and chemokines [38–41]. The ability of ERO1L to modulate CCL2 secretion via oxidative folding suggests that it plays a critical role in shaping the immune landscape of pancreatic lesions. Our data showing reduced macrophage infiltration upon ERO1L knockdown or inhibition with EN460 further support the hypothesis that ERO1L facilitates a pro-tumorigenic environment through its effects on immune cell dynamics. The mechanistic investigations reveal that ERO1L influences the oxidative folding of CCL2, which is essential for its biological activity. The formation of disulfide bonds within CCL2 is crucial for its stability and function, and ERO1L’s role in this process highlights its importance in the secretion and activity of chemokines that recruit immune cells. The observed decrease in the oxidized form of CCL2 in ERO1L-deficient cells may correlate with impaired macrophage migration, highlighting the notion that ERO1L is integral to the production of functional chemokines in the context of pancreatic cancer. Given the important role of ERO1L in promoting KrasG12D and inflammation-driven pancreatic tumorigenesis, targeting ERO1L may offer a novel therapeutic strategy. Pharmacological inhibition of ERO1L, as demonstrated with EN460, effectively mitigates disease progression, suggesting that ERO1L is a promising target for therapeutic intervention. Additionally, understanding the interactions between ERO1L, Kras mutations, and the inflammatory response could lead to the development of combination therapies that address both tumor growth and the immune landscape. In summary, ERO1L emerges as a critical player in the intersection of Kras mutations and inflammation in pancreatic tumorigenesis. Its role in oxidative protein folding, chemokine regulation, and macrophage infiltration highlights its potential as a therapeutic target in pancreatic cancer. Further studies exploring the detailed mechanisms by which ERO1L contributes to tumor progression and immune modulation will be essential for developing effective strategies to combat this devastating disease. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Author contributions Junfeng Zhang: Conceptualization, Data curation, formal analysis, validation, investigation, visualization, methodology, funding acquisition, writing–original draft. Chengyu Hu: Data curation, formal analysis, investigation, visualization, methodology. Yannan Xu: Validation, investigation, methodology. Macheng Lu: Investigation, methodology. Yahui Ma: Investigation, methodology. Ting Yang: Formal analysis, investigation, methodology. Jing Li: Validation, Resources. Shu-Heng Jiang: Investigation, methodology, Writing–review and editing. Chengkai Jiang: Data curation, investigation, writing–review and editing. Pengcheng Chen: Conceptualization, data curation, investigation, supervision, writing–original draft, writing–review and editing. Li Yao: Conceptualization, investigation, data curation, supervision, writing–original draft, writing–review and editing. Rong Hua: Conceptualization, resources, supervision, funding acquisition, project administration, writing–review and editing. Funding This work was supported by grant from National Natural Science Foundation of China (82372622 to Rong Hua), Jiading District Health Commission Project, Shanghai, China (2024-KY-ZD-05 to Jun-Feng Zhang), the University level Scientific Fund of Shanghai University of Medicine and Health Sciences (SSF-24-16-02 to Jun-Feng Zhang; SSF-24-16-01 to Ting Yang), Natural Science Research Project of Jiading District, Shanghai, China (JDKW-2023-0042 to Pengcheng Chen; JDKW-2022-0016 to Ting Yang), Pudong New Area science and technology development Fund project, Shanghai, China (PJK 2024-Y41 to Li Yao), Jiading district medical key discipline construction project (XK202401), 2024 National Clinical Key Specialty Construction Project (10000015Z155080000004), Jiading District Central Hospital Scientific Research Innovation Exploration Program Shanghai, China (JZXKCTS-2024-06 to Chengkai Jiang), and Renji Hospital Science and Technology Projects, Shanghai, China (grant number: RJTJ25-MS-095 to Jun-Feng Zhang). Data availability The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. All data supporting the conclusions of this article are included within the article and additional files. Declarations Ethics approval and consent to participate All research protocols involving the use of human materials were approved by the ethics committee at Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University (KY2024-102-B). All animal experimental protocols were performed following the Ethical Animal Care and Use Committee of Shanghai Jiao Tong University (202401641). 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. Junfeng Zhang, Chengyu Hu, Yannan Xu and Macheng Lu have contributed equally to this work. Contributor Information Chengkai Jiang, Email: luckyk26@163.com. Pengcheng Chen, Email: cpc0315@126.com. Li Yao, Email: ylwy992421@126.com. Rong Hua, Email: 13611657722@sina.cn. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Data Availability Statement The datasets used and analyzed during the current study are available from the corresponding author on reasonable request. All data supporting the conclusions of this article are included within the article and additional files.