Xiao Yu

Pancreatic cancer is characterized by early metastasis and high aggressiveness, highlighting the need for novel molecular biomarkers to improve prognostic assessment and therapeutic strategies. In this study, we constructed a migrasome-related prognostic model through multiple cohorts using 101 machine learning algorithms, and applied Shapley Additive Explanations (SHAP) to interpret the model’s feature contributions. Six key genes — tetraspanin 5 (TSPAN5), bone morphogenetic protein 1 (BMP1), integrin subunit alpha 3 (ITGA3), integrin subunit alpha 5 (ITGA5), Wnt family member 11 (WNT11), and tetraspanin 2 (TSPAN2) —were identified, mainly enriched in extracellular matrix (ECM)–receptor interaction and cell adhesion pathways. Further analyses showed that these genes were preferentially expressed in tumor cells and fibroblasts, and were associated with an altered immune microenvironment. Drug sensitivity prediction and molecular docking indicated that the monocarboxylate transporter 1 (MCT1) inhibitor AZD-3965 may have therapeutic potential in this context. In general, our findings suggest that migrasome-related genes may contribute to prognostic stratification of pancreatic cancer and point to mechanisms of stroma–immune crosstalk, thereby offering exploratory avenues for personalized treatment.

Pancreatitis is a severe and increasingly prevalent disease that affects the digestive system. Early detection and accurate diagnosis of this condition are crucial for reducing mortality rates and improving patient outcomes. Therefore, the development of novel diagnostic markers is essential for enhancing clinical management and advancing the understanding of pancreatitis. The initial phase involved applying the ssGSEA method to extract hypoxia scores from these samples. Subsequently, a thorough differential expression analysis was performed, complemented by functional assessments and various machine learning techniques designed to pinpoint critical diagnostic genes relevant to pancreatitis. From this, a robust diagnostic model was constructed and validated using a series of machine learning strategies. To further validate our results, molecular docking studies were conducted to determine the binding affinities between the identified markers and standard medications such as omeprazole and lansoprazole. Additionally, the ssGSEA methodology was leveraged to compute immune cell scores within the pancreatitis samples, thus enriching the analysis of the relationships between significant diagnostic genes and various immune cell types. Finally, the experiments of ELISA and qRT‐PCR were used to verify the expression of key target genes. Through WGCNA, we identified a total of 50 genes associated with hypoxic conditions within the pancreatitis samples. Further investigations, including differential expression analysis and machine learning techniques, revealed six significant diagnostic markers for pancreatitis: RAP1GDS1, TOP2A, ADK, POLL, CD44, and CD4. The diagnostic model we developed exhibited a high accuracy level in predicting pancreatitis onset, while molecular docking analyses indicated that these six key diagnostic genes hold promise as drug targets. Moreover, the ssGSEA algorithm confirmed the relationships between these diagnostic markers and a range of immune cell populations. Ultimately, the expression levels of the identified key genes were rigorously validated through experimental techniques, reinforcing the credibility of our findings.

Acute pancreatitis (AP) is a disease characterized by local and systemic inflammation with an increasing incidence worldwide. Receptor-interacting serine/threonine protein kinase 3 (RIPK3), mixed-lineage kinase domain-like protein (MLKL), and innate immune cell macrophages have been reported to be involved in the pathogenesis of AP. However, the mechanisms by which RIPK3 and MLKL regulate pancreatic injury, as well as the interactions between injured pancreatic acinar cells and infiltrating macrophages in AP, remain poorly defined. In the present study, experimental pancreatitis was induced in C57BL/6J, Ripk3^-/- and Mlkl^-/- mice by cerulein plus lipopolysaccharide in vivo, and primary pancreatic acinar cells were also isolated to uncover cellular mechanisms during cerulein stimulation in vitro. The results showed that MLKL and its phosphorylated protein p-MLKL were upregulated in the pancreas of the mouse AP model and cerulein-treated pancreatic acinar cells, independent of its canonical upstream molecule Ripk3, and appeared to function in a cell death-independent manner. Knockout of Mlkl attenuated AP in mice by reducing the polarization of pancreatic macrophages toward the M1 phenotype, and this protective effect was partly achieved by reducing the secretion of CXCL10 from pancreatic acinar cells, whereas knockout of Ripk3 did not. In vitro neutralization of CXCL10 impaired the pro-M1 ability of the conditioned medium of cerulein-treated pancreatic acinar cells, whereas in vivo neutralization of CXCL10 reduced the polarization of pancreatic macrophages toward M1 and the severity of AP in mice. These findings suggested that targeting the MLKL-CXCL10-macrophage axis might be a promising strategy for the treatment of AP.

In this paper, we developed a sensitive, label-free and facile fluorescent strategy for detecting streptavidin (SA) based on double-hairpin DNA-templated copper nanoparticles (CuNPs) and terminal protection of small molecule-linked DNA. Herein, a special DNA hairpin probe was designed and synthesized, which contained two poly T single-stranded loops and a nick point in the middle of the stem. Inspired by the concept of the terminal protection interaction, the specific binding of SA to the biotinylated DNA probe can prevent the exonuclease degradation and keep the integrity of DNA probe, which can be used for synthesizing fluorescent CuNPs as a template. Conversely, the DNA probe would be digested by exonucleases and therefore, would fail to form CuNPs without SA. After systematic optimization, the detection range of SA concentration is from 0.5 to 150 nM with a low detection limit of 0.09 nM. Additionally, the proposed method was also successfully applied in the biological samples. Finally, the proposed method is sensitive, effective and simple, and can be potentially applied for predicting diseases and discovering new drugs.

Background: Pancreatic cancer (PC) is one of the most lethal malignancies and carries a dismal mortality and morbidity. Four types of RNA modification (namely m6A, m1A, APA and A-to-I) could be catalyzed by distinct enzymatic compounds (“writers”), mediating numerous epigenetic events in carcinogenesis and immunomodulation. We aim to investigate the interplay mechanism of these writers in immunogenomic features and molecular biological characteristics in PC.Methods: We first accessed the specific expression pattern and transcriptional variation of 26 RNA modification writers in The Cancer Genome Atlas (TCGA) dataset. Unsupervised consensus clustering was performed to divide patients into two RNA modification clusters. Then, based on the differentially expressed genes (DEGs) among two clusters, RNA modification score (WM_Score) model was established to determine RNA modification-based subtypes and was validated in International Cancer Genome Consortium (ICGC) dataset. What’s more, we manifested the unique status of WM_Score in transcriptional and post-transcriptional regulation, molecular biological characteristics, targeted therapies and immunogenomic patterns.Results: We documented the tight-knit correlations between transcriptional expression and variation of RNA modification writers. We classified patients into two distinct RNA modification patterns (WM_Score_high and _low), The WM_Score_high subgroup was correlated with worse prognosis, Th2/Th17 cell polarization and oncogenic pathways (e.g. EMT, TGF-β, and mTORC1 signaling pathways), whereas the WM_Score_low subgroup associated with favorable survival rate and Th1 cell trend. WM_Score model also proved robust predictive power in interpreting transcriptional and post-transcriptional events. Additionally, the potential targeted compounds with related pathways for the WM_Score model were further identified.Conclusions: Our research unfolds a novel horizon on the interplay network of four RNA modifications in PC. This WM_Score model demonstrated powerful predictive capacity in epigenetic, immunological and biological landscape, providing a theoretical basis for future clinical judgments of PC.

Immune responses are an integral part of the pathogenesis of pancreatitis. Studies applying the mouse model of pancreatitis induced by partial ligation of the pancreatic duct to explore the pancreatic immune microenvironment are still lacking. The aim of the present study is to explore the macrophage profile and associated regulatory mechanisms in mouse pancreatitis, as well as the correlation with human chronic pancreatitis (CP). In the present study, the mouse model of pancreatitis was induced by partial ligation of the pancreatic duct. Mice in the acute phase were sacrificed at 0, 4, 8, 16, 32, 72 h after ligation, while mice in the chronic phase were sacrificed at 7, 14, 21, 28 days after ligation. We found that the pancreatic pathological score, expression of TNF-α and IL-6 were elevated over time and peaked at 72h in the acute phase, while in the chronic phase, the degree of pancreatic fibrosis peaked at day 21 after ligation. Pancreatic M1 macrophages and pyroptotic macrophages showed a decreasing trend over time, whereas M2 macrophages gradually rose and peaked at day 21. IL-4 is involved in the development of CP and is mainly derived from pancreatic stellate cells (PSCs). The murine pancreatitis model constructed by partial ligation of the pancreatic duct, especially the CP model, can ideally simulate human CP caused by obstructive etiologies in terms of morphological alterations and immune microenvironment characteristics.

Pancreatic cancer is one of the malignancies having the poorest prognosis due to late diagnoses and lack of efficient treatment regimens. The identification of potential miRNA-targeted gene axes could act as targets for developing novel treatment strategies. Herein, it was assessed that miR-488 expression was markedly downregulated within pancreatic carcinoma. Higher expression of miR-488 was shown to be linked to better prognosis rates of pancreatic carcinoma as per online data. Within two pancreatic tumor cells, MIA PaCa-2 and PANC-1, miR-488 overexpression significantly suppressed malignant cytological behavior by inhibiting cell viability, enhancing cell apoptosis, and inducing cell cycle G2/M-phase arrest. Moreover, miR-488 overexpression also decreased the protein levels of cell cycle regulators, including cyclin A, cyclin B, CDK1, and CDK2. miR-488 directly targets ERBB2 (receptor tyrosine-protein kinase2) to suppress the expression of ERBB2 by targeting its 3ʹUTR. ERBB2 knockdown in MIA PaCa-2 and PANC-1 cell lines suppressed, but miR-488 inhibition enhanced the cancer cell biological malignant behavior; the effects of miR-488 inhibition on pancreatic cancer cells were significantly reversed by ERBB2 knockdown. NF-κB suppressed the expression of miR-488 transcriptionally via targeting its promoter region, consequentially repressing the tumor-suppressive effects of miR-488 upon pancreatic tumor cells. Thus, an NF-κB/miR-488/ERBB2 axis modulating pancreatic cancer cell malignancy and tumor growth through cell cycle signaling was conclusively demonstrated.

Pancreatic cancer (PC) is one of the most lethal malignancies, the mortality and morbidity of which have been increasing over the past decade. Ferroptosis, a newly identified iron-dependent cell death pattern, can be induced by iron chelators and small lipophilic antioxidants. Nonetheless, the prognostic significance of ferroptosis-related lncRNAs in PC remains to be clarified. We obtained the lncRNA expression matrix and clinicopathological information of PC patients from The Cancer Genome Atlas (TCGA) and the International Cancer Genome Consortium (ICGC) datasets in the current study. Firstly, we conducted Pearson correlation analysis to delve into the ferroptosis-related lncRNAs, and univariate Cox analysis was implemented to examine the prognostic values in PC patients. Twenty-three prognostic ferroptosis-related lncRNAs were confirmed and loaded into the least absolute shrinkage and selection operator Cox (LASSO-Cox) analysis, then a ferroptosis-related lncRNA prognostic marker (Fe-LPM) was established in the TCGA dataset. Risk scores of patients were calculated and segregated PC patients into low-risk and high-risk subgroups in each dataset. The prognostic capability of Fe-LPM was also confirmed in the ICGC dataset. Gene set enrichment analysis (GSEA) revealed that several ferroptosis-related pathways were enriched in low-risk subgroups. Furthermore, we adopted a multivariate Cox regression to establish a nomogram based on risk score, age, pathological T stage and primary therapy outcome. A competing endogenous RNA (ceRNA) network was also created relied on four of the twenty-three ferroptosis-related lncRNAs. In conclusion, the eight Fe-LPM can be utilized to anticipate the overall survival (OS) of PC patients, which are meaningful to guiding clinical strategies in PC.

Immunotherapy has made great progress in hepatocellular carcinoma (HCC), yet there is still a lack of biomarkers for predicting response to it. Cancer stem cells (CSCs) are the primary cause of the tumorigenesis, metastasis, and multi-drug resistance of HCC. This study aimed to propose a novel CSCs-related cluster of HCC to predict patients' response to immunotherapy. Based on RNA-seq datasets from The Cancer Genome Atlas (TCGA) and Progenitor Cell Biology Consortium (PCBC), one-class logistic regression (OCLR) algorithm was applied to compute the stemness index (mRNAsi) of HCC patients. Unsupervised consensus clustering was performed to categorize HCC patients into two stemness subtypes which further proved to be a predictor of tumor immune microenvironment (TIME) status, immunogenomic expressions and sensitivity to neoadjuvant therapies. Finally, four machine learning algorithms (LASSO, RF, SVM-RFE and XGboost) were applied to distinguish different stemness subtypes. Thus, a five-hub-gene based classifier was constructed in TCGA and ICGC HCC datasets to predict patients' stemness subtype in a more convenient and applicable way, and this novel stemness-based classification system could facilitate the prognostic prediction and guide clinical strategies of immunotherapy and targeted therapy in HCC.