The investigation began with a high-resolution survey of the bladder tumor landscape. The team integrated four public single-cell RNA sequencing datasets covering five normal bladder and nineteen bladder cancer samples, annotating cells into ten distinct lineages. Within the malignant compartment, LRG1 expression was markedly elevated, and non-negative matrix factorization revealed a specific tumor subcluster defined by high LRG1 output. When the researchers projected the marker genes of these subclusters onto bulk RNA sequencing cohorts from The Cancer Genome Atlas and independent datasets, the LRG1 meta-program was significantly associated with poorer overall survival. Clinically, the signal was even more compelling: serum LRG1 concentrations were substantially higher in patients with hematogenous metastasis than in those without, and histological staining showed dense LRG1 accumulation precisely in the vascular niche of metastatic tumors, where it colocalized with the endothelial marker CD31 while mural cell coverage, marked by alpha-SMA, was conspicuously reduced.

To test whether LRG1 was merely a bystander or an active driver, the researchers turned to genetically engineered mice lacking the Lrg1 gene. When MB49 bladder cancer cells were implanted into these animals, both subcutaneous and orthotopic tumors grew significantly more slowly than in wild-type controls, and lung metastases in a tail-vein model were drastically attenuated. Crucially, the tumor vasculature itself underwent a transformation. Microvessel density fell while average vessel diameter increased, a morphological hallmark of so-called vascular normalization. Pericyte coverage of the vessel wall improved, basement membranes became continuous and uniform rather than fragmented, and scanning electron microscopy revealed a pronounced reduction in the intraluminal endothelial inclusions that riddle leaky, pathological microvessels. Notably, the knockout mice showed no macroscopic abnormalities in major organs, suggesting that LRG1 is a pathological rather than a housekeeping factor and may be a relatively safe therapeutic target.

The pivotal clue to the mechanism came from single-cell transcriptomics performed on tumors overexpressing LRG1. After quality control retained 13,454 high-quality cells, unsupervised clustering identified six major lineages, and subclustering of the neutrophil compartment revealed four distinct subpopulations. One of these, designated C1, scored dramatically higher for a gene signature associated with neutrophil extracellular traps, and its proportion was markedly enriched in LRG1-overexpressing tumors compared with controls. Independent analysis of human data reinforced the picture: a NET-associated neutrophil subset defined by expression of NCF2 was expanded in tumor tissue relative to adjacent normal tissue, and gene set variation analysis across three bulk RNA sequencing cohorts showed a strong positive correlation between the LRG1 tumor meta-program and the abundance of this NET-forming subset.

Functional experiments then established LRG1 as both a recruiter and an activator of neutrophils. Conditioned medium from LRG1-overexpressing bladder cancer cells attracted freshly isolated human neutrophils far more effectively than control medium, while LRG1 knockdown reversed the effect, and flow cytometry confirmed that these differences were not explained by altered neutrophil survival. In vivo, depleting neutrophils with an anti-Ly6G antibody abolished the tumor-promoting effect of LRG1 overexpression entirely. Serological measurements in patients told a consistent story: circulating levels of cell-free DNA, myeloperoxidase, and neutrophil elastase, all canonical markers of NETosis, were elevated in bladder cancer patients and correlated positively with LRG1, with cell-free DNA showing a correlation coefficient of 0.59. Most strikingly, recombinant human LRG1 alone was sufficient to induce NET formation in purified neutrophils, upregulating MPO and citrullinated histone H3, the molecular fingerprints of NETs.

What do these NETs actually do to the vasculature? The answer, according to the study, is a two-faced assault. In vitro, intact NETs accelerated endothelial tube formation, increasing total tube length, capillary-like structures, and junctional nodes, while simultaneously inducing hyperpermeability in endothelial monolayers as measured by fluorescent dextran flux across cell layers. Bladder cancer cells were physically trapped within NET networks, an adhesion that DNase I digestion completely abolished, and exposure to intact NETs significantly enhanced cancer cell migration. In mice, an LPS-induced model of systemic NETosis accelerated tumor growth, but co-administration of DNase I, which degrades the extracellular DNA scaffolds of NETs, abrogated this effect. Immunofluorescence confirmed that NET accumulation stripped mural cells from tumor vessels, and DNase I treatment restored pericyte coverage and vascular integrity, mirroring in reverse the normalization seen in Lrg1-deficient animals.

The deepest mechanistic insight concerns how an extracellular protein persuades a neutrophil to self-destruct into a NET in the sterile, metabolically stressed environment of a tumor. Using pull-down assays followed by mass spectrometry, the team identified annexin A2, or ANXA2, as a direct binding partner of LRG1, an interaction confirmed by co-immunoprecipitation in primary human neutrophils and engineered cell lines. Truncation mutant experiments pinpointed the leucine-rich repeat domain of LRG1, amino acids 93 through 282, as essential for this binding. ANXA2, it turns out, functions as a chaperone that ferries activated Akt, phosphorylated at serine 473, into mitochondria, and this trafficking depends on ANXA2’s own phosphorylation at tyrosine 24. LRG1 binding obstructed ANXA2 phosphorylation and abolished the mitochondrial accumulation of active Akt. Engineered phosphomimetic ANXA2-Y24D mutants restored mitochondrial Akt localization, whereas phospho-dead Y24A mutants worsened the damage.

The downstream consequences for neutrophil metabolism were catastrophic. Recombinant LRG1 treatment triggered a massive surge in mitochondrial reactive oxygen species, a reduction in mitochondrial mass, severe depolarization of the mitochondrial membrane potential, and compromised ATP synthesis. Transmission electron microscopy provided direct visual confirmation: mitochondria in LRG1-treated neutrophils displayed swollen organelles, electron-dense matrices, and disorganized cristae, ultrastructural aberrations that the ANXA2-Y24D mutant partially ameliorated. This places the study squarely within an emerging paradigm in which mitochondrial, rather than purely cytosolic, reactive oxygen species drive NETosis, oxidizing mitochondrial DNA to yield highly immunogenic and pro-angiogenic extracellular traps. In effect, LRG1 acts as a metabolic reprogramming factor that short-circuits the energy machinery of innate immune cells, converting them into agents of vascular destruction.

The translational implications may prove the most consequential part of the story. In a retrospective cohort of twenty-nine bladder cancer patients treated with neoadjuvant cisplatin plus PD-1 blockade, non-responders showed synchronous upregulation of both LRG1 and the NET marker H3Cit, with nine of twelve non-responders classified as LRG1-high and ten of twelve as NETs-high, compared with far lower proportions among responders whose baseline characteristics were otherwise balanced. In mice, Lrg1 deficiency synergized with both cisplatin and anti-PD-1 therapy independently: cisplatin-treated knockout tumors showed increased DNA double-strand breaks and apoptosis, indicating deeper drug penetration, while anti-PD-1-treated knockout mice displayed augmented infiltration of CD8-positive cytotoxic T cells expressing elevated levels of Granzyme B. The authors propose that the LRG1-ANXA2-NETosis axis likely cooperates with the known endothelial TGF-beta/ALK1 pathway, jointly explaining the profoundly disorganized vasculature of bladder cancer, and they acknowledge that questions remain, including the functional role of tumor-specific glycosylation of LRG1 and the need for neutrophil-specific PAD4 knockout models to provide definitive genetic evidence. If those questions are answered, dismantling this axis could convert therapy-refractory, vessel-chaotic tumors into ones that drugs and immune cells can finally reach.

Subject of Research: LRG1-driven neutrophil mitochondrial dysfunction and NETosis in bladder cancer angiogenesis

Article Title: LRG1 Drives Pathological Angiogenesis by Disrupting Neutrophil Mitochondrial Homeostasis in Bladder Cancer

Article References: Chen, D., Zhang, C., Xue, S., Zeng, Y., Cui, H., Wang, C., Feng, J., Yan, L., & Zang, Y. (2026). LRG1 Drives Pathological Angiogenesis by Disrupting Neutrophil Mitochondrial Homeostasis in Bladder Cancer. Advanced Science, 13(56), Article e76604. https://doi.org/10.1002/advs.76604

Image Credits: AI Generated

DOI: 10.1002/advs.76604

Keywords: bladder cancer, LRG1, angiogenesis, neutrophil extracellular traps, NETosis, mitochondrial reactive oxygen species, ANXA2, vascular normalization, tumor microenvironment, single-cell RNA sequencing, immunotherapy resistance, pericyte coverage

The clinical backdrop makes the finding urgent. Surgical resection remains the only curative option for cholangiocarcinoma, yet because early-stage disease produces few symptoms, only 30 to 40 percent of patients are candidates for an operation. Even among those who undergo complete tumor removal, recurrence is common and the five-year overall survival rate falls below 40 percent. Current guidelines recommend six months of adjuvant capecitabine after curative surgery, while for advanced disease the combination of cisplatin and gemcitabine, known as CisGem, stands as the first-line standard of care. But chemotherapy resistance is high, and patients receiving CisGem experience progression-free survival and overall survival of only about 8.0 and 11.7 months, respectively. Targeted agents and immunotherapies have begun to reshape the landscape, with the IDH-1 inhibitor ivosidenib demonstrating activity in the ClarIDHy phase III trial for patients with IDH-1-mutant tumors, and pembrolizumab indicated for tumors deficient in mismatch repair proteins or showing high microsatellite instability. Still, the field suffers from a shortage of specific biomarkers and effective therapeutic targets.

CD39, also known as ectonucleoside triphosphate diphosphohydrolase 1, or ENTPD1, is a membrane-bound enzyme that hydrolyzes extracellular ATP and ADP into anti-inflammatory adenosine. It is normally expressed in organs such as the bladder, brain, lung, spleen, thymus, and placenta, primarily on endothelial and immune cells. Its abundance varies dramatically across immune populations: more than 90 percent of B lymphocytes, monocytes, and neutrophils in peripheral blood carry it, roughly 20 to 30 percent of CD4-positive T cells including memory T cells and regulatory T cells express it, fewer than 5 percent of CD8-positive T cells do, and only 2 to 5 percent of natural killer cells display it on their surface. Crucially, CD39 has been found overexpressed in a range of solid tumors, including colorectal cancer, lung cancer, pancreatic cancer, head and neck cancer, and chronic lymphocytic leukemia, and high expression has been linked to poor outcomes in clear cell renal cell carcinoma, non-small-cell lung cancers, hepatocellular carcinoma, ovarian carcinoma, and gastric cancers. Its role in cholangiocarcinoma, however, had remained unexplored until now.

To fill that gap, the researchers began with a systematic search for genes that behave abnormally in bile duct cancer. RNA sequencing of 12 paired tumor and adjacent non-tumor samples from their institution identified 3,383 upregulated and 3,258 downregulated genes. They then cross-referenced these results with RNA-seq data from the TCGA dataset, comprising 9 pairs, and a GEO dataset, also 9 pairs, and found 21 genes abnormally expressed across all three cohorts. Ten upregulated candidates were selected for validation by quantitative reverse transcription PCR in 24 paired tumor and normal tissue samples. Only one survived the scrutiny: CD39 was significantly upregulated in tumor tissue, and both RT-qPCR and western blotting confirmed that its levels were especially elevated in tumors from patients with lymph node metastasis.

The clinical significance became even clearer when the team turned to immunohistochemistry in a much larger cohort of 180 cholangiocarcinoma patients. Consistent with the molecular data, CD39 was highly expressed in tumor tissue compared with adjacent non-tumor tissue, with a P value below 0.001. Aberrant expression was also observed in patients with lymph node metastasis compared with those without, again at P < 0.001. Survival analysis revealed that patients with high CD39 expression had significantly worse overall survival, with P < 0.001. High expression correlated with a higher rate of lymph node metastasis, P = 0.007, and with more advanced TNM stage, P = 0.004. Most importantly, multivariable Cox proportional hazards regression established CD39 expression as an independent prognostic factor for overall survival, meaning its predictive power held even after accounting for other clinical variables.

Having established the correlation, the researchers moved to causation. They first profiled CD39 expression across human cholangiocarcinoma cell lines, finding it low in the RBE line and high in TFK-1, and used these two lines for gain- and loss-of-function experiments. Cell Counting Kit-8 assays showed that knocking down CD39 inhibited proliferation, while overexpressing it significantly boosted proliferative capacity. Transwell assays told a parallel story for motility: silencing CD39 suppressed the migratory and invasive abilities of TFK-1 cells, whereas forcing its expression enhanced migration and invasion in RBE cells. The pattern extended into living organisms. In cell-line-derived xenograft models using immunodeficient B-NDG mice, tumor volume and weight rose significantly when CD39 was overexpressed and fell when it was knocked down. The number of metastatic nodules in the liver increased in the overexpression group and decreased in the knockdown group, directly implicating CD39 in the metastatic spread of the disease.

The mechanistic breakthrough came from co-immunoprecipitation followed by mass spectrometry. Applying stringent filters, including confidence scores of at least 100 in the mass spectrometric assays and absence from corresponding anti-sense control groups, the team identified nine candidate proteins that physically interact with CD39. Western blotting confirmed that one of them, Annexin A2, or ANXA2, specifically binds CD39. ANXA2 is a calcium-mediated phospholipid-binding protein known to regulate cell survival and communication with the extracellular environment, and it has been implicated in the metastasis of gastric, colorectal, prostate, and breast cancers. Notably, the interaction with CD39 did not change the total amount of ANXA2 in the cells; instead, it altered its phosphorylation state. Knocking down CD39 suppressed phosphorylation of ANXA2 at the tyrosine 24 residue, a modification previously linked to malignant transformation, epithelial-mesenchymal transition, tumor invasion, and metastasis.

Downstream of ANXA2, the team traced the signal into two well-known cancer-promoting programs. Gene Set Enrichment Analysis of RNA-seq data from both tumor tissue and cell lines showed that the PI3K/AKT pathway and the epithelial-mesenchymal transition, or EMT, process were significantly influenced in cases with high CD39 expression. Western blotting confirmed the bioinformatics: knocking down CD39 reduced the phosphorylation of PI3K, AKT, and mTOR at the S2448 site without changing their total protein levels, while overexpression enhanced their phosphorylation. CD39 levels also reshaped the EMT machinery, with knockdown lowering the expression of Claudin-1, Twist-1, Vimentin, and N-cadherin while raising E-cadherin, and overexpression producing the opposite pattern. The decisive proof came from a rescue experiment in which the researchers mutated the Tyr24 site of ANXA2, substituting alanine for tyrosine. This single amino acid change reversed nearly everything CD39 had done: the proliferative boost measured by CCK-8, the migratory and invasive gains seen in transwell assays, the EMT marker shifts, and the PI3K/AKT phosphorylation all collapsed once the phosphorylation site was eliminated.

The authors are candid about the limitations of their work. The study is retrospective, and the prognostic role of CD39 will need validation in a prospective multi-center cohort. The number of samples and cell lines was modest, and larger cohorts would strengthen the statistical power and generalizability of the findings. It also remains unclear exactly how the CD39/ANXA2 axis influences cholangiocarcinoma progression, and how CD39 modulates ANXA2 phosphorylation and the downstream signaling cascades requires further investigation. Even so, the convergence of evidence from human tissue, cell culture, animal models, and mechanistic biochemistry makes a compelling case. CD39 stands out as both a biomarker that could identify patients at highest risk and a therapeutic target whose blockade might starve bile duct tumors of a key growth and spread pathway. For a cancer with so few options, that dual promise is a rare and welcome development.

Subject of Research: The role of CD39 expression in the progression, metastasis, and prognosis of cholangiocarcinoma

Article Title: The aberrant expression of CD39 is associated with the dismal outcome of patients with cholangiocarcinoma

Article References: Zheng, B., Shen, S., Wang, J., Sun, W., Wan, W., Liu, J., Zhang, D., Ni, X., Suo, T., Liu, H., Ni, X., & Liu, H. (2025). The aberrant expression of CD39 is associated with the dismal outcome of patients with cholangiocarcinoma. Clinical Cancer Bulletin, 4(1), Article 1. https://doi.org/10.1007/s44272-025-00028-0

Image Credits: AI Generated

DOI: 10.1007/s44272-025-00028-0

Keywords: CD39, cholangiocarcinoma, ANXA2, PI3K/AKT signaling, EMT, phosphorylation, prognostic biomarker, biliary tract cancer, ENTPD1, metastasis, immunohistochemistry, therapeutic target