FABP5+ lipid-loaded macrophages process tumor-derived unsaturated fatty acid signal to suppress T-cell antitumor immunity.

BACKGROUND & AIMS: Tumour-associated macrophages (TAMs) contribute to hepatocellular carcinoma (HCC) progression. However, while the pro-tumour and immunosuppressive roles of lipid-loaded macrophages are well established, the mechanisms by which lipid metabolism enhances the tumour-promoting effects in TAMs remain unclear. METHODS: Single-cell RNA sequencing was performed on mouse and human HCC tumour samples to elucidate the landscape of HCC TAMs. Macrophages were stimulated with various long-chain unsaturated fatty acids (UFAs) to assess immunosuppressive molecules expression in vitro. Additionally, in vivo and in vitro studies were conducted using mice with macrophage-specific deficiencies in fatty acid-binding protein 5 (FABP5) or peroxisome proliferator-activated receptor (PPAR). RESULTS: Single-cell RNA sequencing identified a subpopulation of FABP5+ lipid-loaded TAMs characterized by enhanced immune checkpoint blocker ligands and immunosuppressive molecules in an oncogene-mutant HCC mouse model and human HCC tumours. Mechanistically, long-chain UFAs released by tumour cells activate PPARvia FABP5, resulting in TAM immunosuppressive properties. FABP5 deficiency in macrophages decreases immunosuppressive molecules expression, enhances T-cell-dependent antitumor immunity, diminishes HCC growth, and improves immunotherapy efficacy. CONCLUSIONS: This study demonstrates that UFAs promote tumourigenesis by enhancing the immunosuppressive tumour microenvironment via FABP5-PPAR signaling and provides a proof-of-concept for targeting this pathway to improve tumour immunotherapy.

Clinical Features and Unusual Heterozygous Mutations in Patients with Renal Hypokalemia.

BACKGROUND: Renal hypokalemia is associated with mutation. This study aimed to investigate the clinical features and pathogenic mutations in patients with renal hypokalemia. METHODS: The patients with hypokalemia were enrolled, and the renal function, thyroid function, renin-aldosterone system, urinary potassium excretion, and exome sequencing were performed. The correlation between the clinical phenotypes and causative genes was assessed. RESULTS: Five patients with hypokalemia were enrolled and diagnosed as tubular hypokalemia. The patients with common clinical manifestations were difficult to differentiate based on atypical laboratory findings. The results of the genetic analysis were as follows: both patient 1 and patient 2 were heterozygous for the c.C625T mutation of the KCNJ1 gene, which is responsible for Bartter syndrome. Patient 3 was heterozygous for the c.G298A mutation of the ATP6V1B1 gene, which is responsible for renal tubular acidosis. Patient 4 had a compound heterozygous mutation of c.G893A of the BSND gene, responsible for Bartter syndrome, and c.1029+5G>A, the ATP6V0A4 gene responsible for distal renal tubular acidosis. Patient 5 had Gitelman syndrome and carried the compound heterozygous mutations c.C1963T and c.G2029A of the SLC12A3 gene. All the above loci were known heterozygous mutations. CONCLUSIONS: The unusual heterozygous mutations were identified in five renal hypokalemia patients. Molecular diagnosis of tubular hypokalemia was conducive to accurate diagnosis and treatment.

Optimization of Selenization Condition for Efficiency CIGSe Solar Cells Based on Postselenization of CuInGa Precursors.

The degree of selenization of the CIGSe absorbers is controlled by regulating the parameters of the selenization reaction. The structure, element distribution, phase composition of the CIGSe absorbers, and the performances of the solar cells with different selenization degrees are studied. Insufficient selenization will lead to residual Cu2Se phase on the surface and insufficient Na diffusion, which will affect the VCu+ on the surface and the recombination at the front interface. However, excessive selenization will make the MoSe2 layer thicken at the back interface of the CIGSe/Mo, resulting in the increase of the series resistance and the enhancement of the recombination at the back interface. The appropriate selenization degree is conducive to inhibiting the recombination at the front and back interfaces. Improved device performances can be obtained by optimizing the selenization degree of the absorbers.

IRE1α-XBP1s axis regulates SREBP1-dependent MRP1 expression to promote chemoresistance in non-small cell lung cancer cells.

BACKGROUND: Inositol-requiring enzyme 1 (IRE1) is an endoplasmic reticulum (ER)-resident transmembrane protein that senses ER stress and mediates an essential arm of the unfolded protein response (UPR). IRE1 reduces ER stress by upregulating the expression of multiple ER chaperones through activation of X-box-binding protein 1 (XBP1). Emerging lines of evidence have revealed that IRE1-XBP1 axis serves as a multipurpose signal transducer during oncogenic transformation and cancer development. In this study, we explore how IRE1-XBP1 signaling promotes chemoresistance in lung cancer. METHODS: The expression patterns of UPR components and MRP1 were examined by Western blot. qRT-PCR was employed to determine RNA expression. The promoter activity was determined by luciferase reporter assay. Chemoresistant cancer cells were analyzed by viability, apoptosis. CUT & Tag (Cleavage under targets and tagmentation)-qPCR analysis was used for analysis of DNA-protein interaction. RESULTS: Here we show that activation of IRE1α-XBP1 pathway leads to an increase in MDR-related protein 1 (MRP1) expression, which facilitates drug extrusion and confers resistance to cytotoxic chemotherapy. At the molecular level, XBP1-induced c-Myc is necessary for SREBP1 expression, and SREBP1 binds to the MRP1 promoter to directly regulate its transcription. CONCLUSIONS: We conclude that IRE1α-XBP1 had important role in chemoresistance and appears to be a novel prognostic marker for lung cancer.

Spatial Continuity and Nonequal Importance in Salient Object Detection With Image-Category Supervision.

Due to the inefficiency of pixel-level annotations, weakly supervised salient object detection with image-category labels (WSSOD) has been receiving increasing attention. Previous works usually endeavor to generate high-quality pseudolabels to train the detectors in a fully supervised manner. However, we find that the detection performance is often limited by two types of noise contained in pseudolabels: 1) holes inside the object or at the edge and outliers in the background and 2) missing object portions and redundant surrounding regions. To mitigate the adverse effects caused by them, we propose local pixel correction (LPC) and key pixel attention (KPA), respectively, based on two key properties of desirable pseudolabels: 1) spatial continuity, meaning an object region consists of a cluster of adjacent points; and 2) nonequal importance, meaning pixels have different importance for training. Specifically, LPC fills holes and filters out outliers based on summary statistics of the neighborhood as well as its size. KPA directs the focus of training toward ambiguous pixels in multiple pseudolabels to discover more accurate saliency cues. To evaluate the effectiveness of our method, we design a simple yet strong baseline we call weakly supervised saliency detector with Transformer (WSSDT) and unify the proposed modules into WSSDT. Extensive experiments on five datasets demonstrate that our method significantly improves the baseline and outperforms all existing congeneric methods. Moreover, we establish the first benchmark to evaluate WSSOD robustness. The results show that our method can improve detection robustness as well. The code and robustness benchmark are available at https://github.com/Horatio9702/SCNI.

Insight into Mocha Diffusion Occurring during Ethanol Droplet Spreading on Acrylic Paint.

Mocha diffusion, a significant interfacial phenomenon in pottery painting, remains insufficiently understood in the documented literature regarding its dynamics. This study experimentally investigates Mocha patterns by quantitatively dripping ethanol droplets onto an acrylic paint surface. Results indicate that the spreading radius increases with the ethanol mass fraction, while the fractal period decreases. The fractal dimension of all Mocha patterns approximates 1.4087. Marangoni flow, generated by the volatilization of ethanol, is crucial for the growth and fractal formation of the "dendrites" in this spreading. The scaling analysis is used to interpret the spreading dynamics. This work encourages the interface science community to develop a comprehensive theory for the dynamics of Mocha diffusion and highlights the potential of this intriguing decorative technique.

Multi-view heterogeneous graph learning with compressed hypergraph neural networks.

Multi-view learning is an emerging field of multi-modal fusion, which involves representing a single instance using multiple heterogeneous features to improve compatibility prediction. However, existing graph-based multi-view learning approaches are implemented on homogeneous assumptions and pairwise relationships, which may not adequately capture the complex interactions among real-world instances. In this paper, we design a compressed hypergraph neural network from the perspective of multi-view heterogeneous graph learning. This approach effectively captures rich multi-view heterogeneous semantic information, incorporating a hypergraph structure that simultaneously enables the exploration of higher-order correlations between samples in multi-view scenarios. Specifically, we introduce efficient hypergraph convolutional networks based on an explainable regularizer-centered optimization framework. Additionally, a low-rank approximation is adopted as hypergraphs to reformat the initial complex multi-view heterogeneous graph. Extensive experiments compared with several advanced node classification methods and multi-view classification methods have demonstrated the feasibility and effectiveness of the proposed method.

Pancreatic STAT5 activation promotes KrasG12D-induced and inflammation-induced acinar-to-ductal metaplasia and pancreatic cancer.

OBJECTIVE: Pancreatic ductal adenocarcinoma (PDAC) is a highly lethal malignancy because it is often diagnosed at a late-stage. Signal transducer and activator of transcription 5 (STAT5) is a transcription factor implicated in the progression of various cancer types. However, its role in KRAS-driven pancreatic tumourigenesis remains unclear. DESIGN: We performed studies with LSL-Kras G12D; Ptf1a-Cre ERT (KCERT) mice or LSL-KrasG12D; LSL-Trp53R172H ; Pdx1-Cre (KPC) mice crossed with conditional disruption of STAT5 or completed deficiency interleukin (IL)-22. Pancreatitis was induced in mice by administration of cerulein. Pharmacological inhibition of STAT5 on PDAC prevention was studied in the orthotopic transplantation and patient-derived xenografts PDAC model, and KPC mice. RESULTS: The expression and phosphorylation of STAT5 were higher in human PDAC samples than control samples and high levels of STAT5 in tumour cells were associated with a poorer prognosis. The loss of STAT5 in pancreatic cells substantially reduces the KRAS mutation and pancreatitis-derived acinar-to-ductal metaplasia (ADM) and PDAC lesions. Mechanistically, we discovered that STAT5 binds directly to the promoters of ADM mediators, hepatocyte nuclear factor (HNF) 1ß and HNF4α. Furthermore, STAT5 plays a crucial role in maintaining energy metabolism in tumour cells during PDAC progression. IL-22 signalling induced by chronic inflammation enhances KRAS-mutant-mediated STAT5 phosphorylation. Deficiency of IL-22 signalling slowed the progression of PDAC and ablated STAT5 activation. CONCLUSION: Collectively, our findings identified pancreatic STAT5 activation as a key downstream effector of oncogenic KRAS signalling that is critical for ADM initiation and PDAC progression, highlighting its potential therapeutic vulnerability.

Exploring the Mechanism of Asiatic Acid against Atherosclerosis Based on Molecular Docking, Molecular Dynamics, and Experimental Verification.

Asiatic acid (AA) is a pentacyclic triterpene derived from the traditional medicine Centella asiatica. It is known for its anti-inflammatory, antioxidant, and lipid-regulating properties. Though previous studies have suggested its potential therapeutic benefits for atherosclerosis, its pharmacological mechanism is unclear. The objective of this study was to investigate the molecular mechanism of AA in the treatment of atherosclerosis. Therefore, network pharmacology was employed to uncover the mechanism by which AA acts as an anti-atherosclerotic agent. Furthermore, molecular docking, molecular dynamics (MD) simulation, and in vitro experiments were performed to elucidate the mechanism of AA's anti-atherosclerotic effects. Molecular docking analysis demonstrated a strong affinity between AA and PPARγ. Further MD simulations demonstrated the favorable stability of AA-PPARγ protein complexes. In vitro experiments demonstrated that AA can dose-dependently inhibit the expression of inflammatory factors induced by lipopolysaccharide (LPS) in RAW264.7 cells. This effect may be mediated through the PPARγ/NF-κB signaling pathway. This research underscores anti-inflammation as a crucial biological process in AA treatments for atherosclerosis, with PPARγ potentially serving as a key target.

High-throughput microalgae sorting based on the deterministic lateral displacement technique.

Microalgae are a group of photosynthetic organisms that can grow autotrophically, performing photosynthesis to synthesize abundant organic compounds and release oxygen. They are rich in nutritional components and chemical precursors, presenting wide-ranging application prospects. However, potential contamination by foreign strains or bacteria can compromise their analytical applications. Therefore, the obtaining of pure algal strains is crucial for the subsequent analysis and application of microalgae. This study designed a deterministic lateral displacement (DLD) chip with dual input and dual outlet of equal width for the separation of Haematococcus pluvialis and Chlorella vulgaris. Optimal separation parameters were determined through a series of experiments, resulting in a purity of 99.80 % for Chlorella vulgaris and 94.58 % for Haematococcus pluvialis, with recovery rates maintained above 90 %, demonstrating high efficiency. This study provides a reliable foundation for future research and applications of microalgae, which holds considerable significance for the subsequent analysis and utilization of microalgae.

Heterogeneous graph convolutional network for multi-view semi-supervised classification.

This paper proposes a novel approach to semantic representation learning from multi-view datasets, distinct from most existing methodologies which typically handle single-view data individually, maintaining a shared semantic link across the multi-view data via a unified optimization process. Notably, even recent advancements, such as Co-GCN, continue to treat each view as an independent graph, subsequently aggregating the respective GCN representations to form output representations, which ignores the complex semantic interactions among heterogeneous data. To address the issue, we design a unified framework to connect multi-view data with heterogeneous graphs. Specifically, our study envisions multi-view data as a heterogeneous graph composed of shared isomorphic nodes and multi-type edges, wherein the same nodes are shared across different views, but each specific view possesses its own unique edge type. This perspective motivates us to utilize the heterogeneous graph convolutional network (HGCN) to extract semantic representations from multi-view data for semi-supervised classification tasks. To the best of our knowledge, this is an early attempt to transfigure multi-view data into a heterogeneous graph within the realm of multi-view semi-supervised learning. In our approach, the original input of the HGCN is composed of concatenated multi-view matrices, and its convolutional operator (the graph Laplacian matrix) is adaptively learned from multi-type edges in a data-driven fashion. After rigorous experimentation on eight public datasets, our proposed method, hereafter referred to as HGCN-MVSC, demonstrated encouraging superiority over several state-of-the-art competitors for semi-supervised classification tasks.

Identification of ClpB, a molecular chaperone involved in the stress tolerance and virulence of Streptococcus agalactiae.

Bacterial ClpB is an ATP-dependent disaggregate that belongs to the Hsp100/Clp family and facilitates bacterial survival under hostile environmental conditions. Streptococcus agalactiae, which is regarded as the major bacterial pathogen of farmed Nile tilapia (Oreochromis niloticus), is known to cause high mortality and large economic losses. Here, we report a ClpB homologue of S. agalactiae and explore its functionality. S. agalactiae with a clpB deletion mutant (∆clpB) exhibited defective tolerance against heat and acidic stress, without affecting growth or morphology under optimal conditions. Moreover, the ΔclpB mutant exhibited reduced intracellular survival in RAW264.7 cells, diminished adherence to the brain cells of tilapia, increased sensitivity to leukocytes from the head kidney of tilapia and whole blood killing, and reduced mortality and bacterial loads in a tilapia infection assay. Furthermore, the reduced virulence of the ∆clpB mutant was investigated by transcriptome analysis, which revealed that deletion of clpB altered the expression levels of multiple genes that contribute to the stress response as well as certain metabolic pathways. Collectively, our findings demonstrated that ClpB, a molecular chaperone, plays critical roles in heat and acid stress resistance and virulence in S. agalactiae. This finding provides an enhanced understanding of the functionality of this ClpB homologue in gram-positive bacteria and the survival strategy of S. agalactiae against immune clearance during infection.

Upregulation of neuronal ER-phagy improves organismal fitness and alleviates APP toxicity.

ER-phagy, a selective autophagy targeting the endoplasmic reticulum (ER) for lysosomal degradation through cargo receptors, plays a critical role in ER quality control and is linked to various diseases. However, its physiological and pathological roles remain largely unclear due to a lack of animal model studies. This study establishes Drosophila as an in vivo ER-phagy model. Starvation triggers ER-phagy across multiple fly tissues. Disturbing ER-phagy by either globally upregulating or downregulating ER-phagy receptors, Atl or Rtnl1, harms the fly. Notably, moderate upregulation of ER-phagy in fly brains by overexpressing Atl or Rtnl1 significantly attenuates age-associated neurodegenerations. Furthermore, in a Drosophila model of Alzheimer's disease expressing human amyloid precursor protein (APP), impaired ER-phagy is observed. Enhancing ER-phagy in the APP-expressing fly brain facilitates APP degradation, significantly alleviating disease symptoms. Therefore, our findings suggest that modulating ER-phagy may offer a therapeutic strategy to treat aging and diseases associated with ER protein aggregation.

Misclassification in Weakly Supervised Object Detection.

Weakly supervised object detection (WSOD) aims to train detectors using only image-category labels. Current methods typically first generate dense class-agnostic proposals and then select objects based on the classification scores of these proposals. These methods mainly focus on selecting the proposal having high Intersection-over-Union with the true object location, while ignoring the problem of misclassification, which occurs when some proposals exhibit semantic similarities with objects from other categories due to viewing perspective and background interference. We observe that the positive class that is misclassified typically has the following two characteristics: 1) It is usually misclassified as one or a few specific negative classes, and the scores of these negative classes are high; 2) Compared to other negative classes, the score of the positive class is relatively high. Based on these two characteristics, we propose misclassification correction (MCC) and misclassification tolerance (MCT) respectively. In MCC, we establish a misclassification memory bank to record and summarize the class-pairs with high frequencies of potential misclassifications in the early stage of training, that is, cases where the score of a negative class is significantly higher than that of the positive class. In the later stage of training, when such cases occur and correspond to the summarized class-pairs, we select the top-scoring negative class proposal as the positive training example. In MCT, we decrease the loss weights of misclassified classes in the later stage of training to avoid them dominating training and causing misclassification of objects from other classes that are semantically similar to them during inference. Extensive experiments on the PASCAL VOC and MS COCO demonstrate our method can alleviate the problem of misclassification and achieve the state-of-the-art results.

UMCGL: Universal Multi-View Consensus Graph Learning With Consistency and Diversity.

Existing multi-view graph learning methods often rely on consistent information for similar nodes within and across views, however they may lack adaptability when facing diversity challenges from noise, varied views, and complex data distributions. These challenges can be mainly categorized into: 1) View-specific diversity within intra-view from noise and incomplete information; 2) Cross-view diversity within inter-view caused by various latent semantics; 3) Cross-group diversity within inter-group due to data distribution differences. To this end, we propose a universal multi-view consensus graph learning framework that considers both original and generative graphs to balance consistency and diversity. Specifically, the proposed framework can be divided into the following four modules: i) Multi-channel graph module to extract principal node information, ensuring view-specific and cross-view consistency while mitigating view-specific and cross-view diversity within original graphs; ii) Generative module to produce cleaner and more realistic graphs, enriching graph structure while maintaining view-specific consistency and suppressing view-specific diversity; iii) Contrastive module to collaborate on generative semantics to facilitate cross-view consistency and reducing cross-view diversity within generative graphs; iv) Consensus graph module to consolidate learning a consensual graph, pursuing cross-group consistency and cross-group diversity. Extensive experimental results on real-world datasets demonstrate its effectiveness and superiority.

Multilayer regulatory landscape and new regulators identification for bud dormancy release and bud break in Populus.

For trees originating from boreal and temperate regions, the dormancy-to-active transition, also known as bud dormancy release and bud break, are crucial processes that allow trees to reactive growth in the spring. The molecular mechanisms underlying these two processes remain poorly understood. Here, through integrative multiomics analysis of the transcriptome, DNA methylome, and proteome, we gained insights into the reprogrammed cellular processes associated with bud dormancy release and bud break. Our findings revealed multilayer regulatory landscapes governing bud dormancy release and bud break regulation, providing a valuable reference framework for future functional studies. Based on the multiomics analysis, we have determined a novel long intergenic noncoding RNA named Phenology Responsive Intergenic lncRNA 1 (PRIR1) plays a role in the activation of bud break. that the molecular mechanism of PRIR1 has been preliminary explored, and it may partially promote bud break by activating its neighbouring gene, EXORDIUM LIKE 5 (PtEXL5), which has also been genetically confirmed as an activator for bud break. This study has revealed a lncRNA-mediated regulatory mechanism for the control of bud break in Populus, operating independently of known regulatory pathways.

Genomic evidence for evolutionary history and local adaptation of two endemic apricots: Prunus hongpingensis and P. zhengheensis.

Apricot, belonging to the Armeniaca section of Rosaceae, is one of the economically important crop fruits that has been extensively cultivated. The natural wild apricots offer valuable genetic resources for crop improvement. However, some of them are endemic, with small populations, and are even at risk of extinction. In this study we unveil chromosome-level genome assemblies for two southern China endemic apricots, Prunus hongpingensis (PHP) and P. zhengheensis (PZH). We also characterize their evolutionary history and the genomic basis of their local adaptation using whole-genome resequencing data. Our findings reveal that PHP and PZH are closely related to Prunus armeniaca and form a distinct lineage. Both species experienced a decline in effective population size following the Last Glacial Maximum (LGM), which likely contributed to their current small population sizes. Despite the observed decrease in genetic diversity and heterozygosity, we do not observe an increased accumulation of deleterious mutations in these two endemic apricots. This is likely due to the combined effects of a low inbreeding coefficient and strong purifying selection. Furthermore, we identify a set of genes that have undergone positive selection and are associated with local environmental adaptation in PHP and PZH, respectively. These candidate genes can serve as valuable genetic resources for targeted breeding and improvement of cultivated apricots. Overall, our study not only enriches our comprehension of the evolutionary history of apricot species but also offers crucial insights for the conservation and future breeding of other endemic species amidst rapid climate changes.

The mitochondrial stress-induced protein carboxyl-terminal alanine and threonine tailing (msiCAT-tailing) promotes glioblastoma tumorigenesis by modulating mitochondrial functions.

The rapid and sustained proliferation in cancer cells requires accelerated protein synthesis. Accelerated protein synthesis and disordered cell metabolism in cancer cells greatly increase the risk of translation errors. ribosome-associated quality control (RQC) is a recently discovered mechanism for resolving ribosome collisions caused by frequent translation stalls. The role of the RQC pathway in cancer initiation and progression remains controversial and confusing. In this study, we investigated the pathogenic role of mitochondrial stress-induced protein carboxyl-terminal terminal alanine and threonine tailing (msiCAT-tailing) in glioblastoma (GBM), which is a specific RQC response to translational arrest on the outer mitochondrial membrane. We found that msiCAT-tailed mitochondrial proteins frequently exist in glioblastoma stem cells (GSCs). Ectopically expressed msiCAT-tailed mitochondrial ATP synthase F1 subunit alpha (ATP5α) protein increases the mitochondrial membrane potential and blocks mitochondrial permeability transition pore (MPTP) formation/opening. These changes in mitochondrial properties confer resistance to staurosporine (STS)-induced apoptosis in GBM cells. Therefore, msiCAT-tailing can promote cell survival and migration, while genetic and pharmacological inhibition of msiCAT-tailing can prevent the overgrowth of GBM cells. Highlights: The RQC pathway is disturbed in glioblastoma (GBM) cellsmsiCAT-tailing on ATP5α elevates mitochondrial membrane potential and inhibits MPTP openingmsiCAT-tailing on ATP5α inhibits drug-induced apoptosis in GBM cellsInhibition of msiCAT-tailing impedes overall growth of GBM cells.

A WOX homolog disrupted by a transposon led to the loss of spines and contributed to the domestication of lettuce.

The loss of spines is one of the most important domestication traits for lettuce (Lactuca sativa). However, the genetics and regulation of spine development in lettuce remain unclear. We examined the genetics of spines in lettuce using a segregating population derived from a cross between cultivated and wild lettuce (Lactuca serriola). A gene encoding WUSCHEL-related homeobox transcription factor, named as WOX-SPINE1 (WS1), was identified as the candidate gene controlling the spine development in lettuce, and its function on spines was verified. A CACTA transposon was found to be inserted into the first exon of the ws1 allele, knocking out its function and leading to the lack of spines in cultivated lettuce. All lettuce cultivars investigated have the nonfunctional ws1 gene, and a selection sweep was found at the WS1 locus, suggesting its important role in lettuce domestication. The expression levels of WS1 were associated with the density of spines among different accessions of wild lettuce. At least two independent loss-of-function mutations in the ws1 gene caused the loss of spines in wild lettuce. These findings provide new insights into the development of spines and facilitate the exploitation of wild genetic resources in future lettuce breeding programs.

Fe-CGS Effectively Inhibits the Dynamic Migration and Transformation of Cadmium and Arsenic in Soil.

The iron-modified coal gasification slag (Fe-CGS) material has excellent performance in purifying heavy-metal-contaminated water due to its good surface properties and adsorption capacities. However, it is unclear whether it can provide long-term simultaneous stabilization of Cd and As in composite-contaminated soils in extreme environments. This study investigated the long-term stabilization of Cd and As in acidic (JLG) and alkaline (QD) soils by simulating prolonged heavy rainfall with the addition of Fe-CGS. Multiple extraction methods were used to analyze the immobilization mechanisms of Cd and As in soil and their effects on bioavailability. The results indicate that the stabilization efficiency was related to the dosage of Fe-CGS. The concentrations of Cd and As in the JLG soil leachate were reduced by 77.6% (2.0 wt%) and 87.8% (1.0 wt%), respectively. Additionally, the availability of Cd and As decreased by 46.7% (2.0 wt%) and 53.0% (1.0 wt%), respectively. In the QD soil leachate, the concentration of Cd did not significantly change, while the concentration of As decreased by 92.3% (2.0 wt%). Furthermore, the availability of Cd and As decreased by 22.1% (2.0 wt%) and 40.2% (1.0 wt%), respectively. Continuous extraction revealed that Fe-CGS facilitated the conversion of unstable, acid-soluble Cd into oxidizable Cd and acid-soluble Cd. Additionally, it promoted the transformation of both non-specifically and specifically adsorbed As into amorphous iron oxide-bound and residual As. Fe-CGS effectively improved the soil pH, reduced the bioavailability of Cd and As, and blocked the migration of Cd and As under extreme rainfall leaching conditions. It also promoted the transformation of Cd and As into more stable forms, exhibiting satisfactory long-term stabilization performance for Cd and As.