Spatial Sequence Research Articles

Cancer-Associated Endocrine Cells Participate in Pancreatic Carcinogenesis

Background & AimsThe pancreas is composed of endocrine and exocrine parts, and its interlacing structure indicates potential interaction between endocrine and exocrine cells. Although the tumor microenvironment of pancreatic ductal adenocarcinoma (PDAC) has been well characterized, the role of pancreatic endocrine cells during carcinogenesis is relatively understudied. MethodsWe depicted the changes of endocrine cells in PDAC by single-cell transcriptome sequencing, spatial transcriptome sequencing and multiplex immunohistochemistry. After that, the interaction between pancreatic carcinogenesis and endocrine changes was explored in orthotopic transplantation mice, KC mice and KPC mice. Finally, we proved the mechanism of the interaction between endocrine and exocrine parts of the pancreas through islet isolation, co-culture in vitro and co-injection in vivo. ResultsWe found that pancreatic endocrine cells displayed significantly different transcriptomic characteristics and increased interaction with exocrine part in PDAC. Specifically, among all the changes, pancreatic polypeptide positive (PPY+) cells showed a sharp increment accompanied with the progression of the cancer lesion, which might be derived from the transdifferentiation of α and β cells. Interestingly, it was proved that PDAC cells were able to induce the transdifferentiation of pancreatic α cells and β cells into GCG+PPY+ and INS+PPY+ double-positive cells, which further promoted carcinogenesis and development of PDAC in a paracrine-dependent manner and formed a reciprocal interaction. ConclusionsOur study systematically maps the alteration of pancreatic endocrine cells in PDAC and elucidates the potential endocrine-exocrine interaction mechanisms during PDAC carcinogenesis. Meanwhile, we first time define and characterize cancer-associated endocrine cells (CAEs), thereby further broadening the composition of PDAC microenvironment.

Decoder-seq: a technology for high sensitivity, high resolution, and low-cost spatial RNA sequencing.

Decoder-seq: a technology for high sensitivity, high resolution, and low-cost spatial RNA sequencing.

Medulloblastoma Spatial Transcriptomics Reveals Tumor Microenvironment Heterogeneity with High-Density Progenitor Cell Regions Correlating with High-Risk Disease.

The tumor microenvironment (TME) of medulloblastoma (MB) influences progression and therapy response, presenting a promising target for therapeutic advances. Prior single-cell analyses have characterized the cellular components of the TME but lack spatial context. To address this, we performed spatial transcriptomic sequencing on sixteen pediatric MB samples obtained at diagnosis, including two matched diagnosis-relapse pairs. Our analyses revealed inter- and intra-tumoral heterogeneity within the TME, comprised of tumor-associated astrocytes (TAAs), macrophages (TAMs), stromal components, and distinct subpopulations of MB cells at different stages of neuronal differentiation and cell cycle progression. We identified dense regions of quiescent progenitor-like MB cells enriched in patients with high-risk (HR) features and an increase in TAAs, TAMs, and dysregulated vascular endothelium following relapse. Our study presents novel insights into the spatial architecture and cellular landscape of the medulloblastoma TME, highlighting spatial patterns linked to HR features and relapse, which may serve as potential therapeutic targets.

Unfolding the mysteries of heterogeneity from a high-resolution perspective: integration analysis of single-cell multi-omics and spatial omics revealed functionally heterogeneous cancer cells in ccRCC.

The genomic landscape of clear cell renal cell carcinoma (ccRCC) has a considerable intra-tumor heterogeneity, which is a significant obstacle in the field of precision oncology and plays a pivotal role in metastasis, recurrence, and therapeutic resistance of cancer. The mechanisms of intra-tumor heterogeneity in ccRCC have yet to be fully established. We integrated single-cell RNA sequencing (scRNA-seq) and transposase-accessible chromatin sequencing (scATAC-seq) data from a single-cell multi-omics perspective. Based on consensus non-negative matrix factorization (cNMF) algorithm, functionally heterogeneous cancer cells were classified into metabolism, inflammatory, and EMT meta programs, with spatial transcriptomics sequencing (stRNA-seq) providing spatial information of such disparate meta programs of cancer cells. The bulk RNA sequencing (RNA-seq) data revealed high clinical prognostic values of functionally heterogeneous cancer cells of three meta programs, with transcription factor regulatory network and motif activities revealing the key transcription factors that regulate functionally heterogeneous ccRCC cells. The interactions between varying meta programs and other cell subpopulations in the microenvironment were investigated. Finally, we assessed the sensitivity of cancer cells of disparate meta programs to different anti-cancer agents. Our findings inform on the intra-tumor heterogeneity of ccRCC and its regulatory networks and offers new perspectives to facilitate the designs of rational therapeutic strategies.

The Heterogeneity of Form and Meaning in Chinese Political Discourse Chunk Constructions and its Interpretation in English Translation

The trend of "Translating China" emphasizes the importance of understanding and disseminating the source language. This paper discusses the representation of spatial chunk construction sequences, such as the "4±0" chunk pattern, in political discourse and their heterogeneity in form and meaning, as well as the cognitive interpretation and processing paradigms of translators under the theory of spatial-temporal thinking differences between English and Chinese. The study finds that: 1) The "4±0" chunk pattern and its chunk construction sequences both exhibit strong spatial characteristics in their chunk-like and discrete features; 2) The preference for semantic pairing within the "4±0" chunk pattern is driven by the spatial preferences in Chinese, leading to differences in semantic background; 3) The strong spatial thinking in Chinese influences the dynamic pairing of chunks in the "4±0" sequences, resulting in significant heterogeneity of form and meaning, which creates translation bottlenecks; 4) The spatial-temporal differences between English and Chinese drive the deep interpretation and reprocessing of micro-macro semantic logic within and between chunks in the source language, while constraining the representation of temporal characteristics in the target language translation.

ScHolography: a computational method for single-cell spatial neighborhood reconstruction and analysis

Spatial transcriptomics has transformed our ability to study tissue complexity. However, it remains challenging to accurately dissect tissue organization at single-cell resolution. Here we introduce scHolography, a machine learning-based method designed to reconstruct single-cell spatial neighborhoods and facilitate 3D tissue visualization using spatial and single-cell RNA sequencing data. scHolography employs a high-dimensional transcriptome-to-space projection that infers spatial relationships among cells, defining spatial neighborhoods and enhancing analyses of cell–cell communication. When applied to both human and mouse datasets, scHolography enables quantitative assessments of spatial cell neighborhoods, cell–cell interactions, and tumor-immune microenvironment. Together, scHolography offers a robust computational framework for elucidating 3D tissue organization and analyzing spatial dynamics at the cellular level.

Dentate Gyrus Morphogenesis is Regulated by an Autism Risk Gene Trio Function in Granule Cells.

Autism Spectrum Disorders (ASDs) are reported as a group of neurodevelopmental disorders. The structural changes of brain regions including the hippocampus were widely reported in autistic patients and mouse models with dysfunction of ASD risk genes, but the underlying mechanisms are not fully understood. Here, we report that deletion of Trio, a high-susceptibility gene of ASDs, causes a postnatal dentate gyrus (DG) hypoplasia with a zigzagged suprapyramidal blade, and the Trio-deficient mice display autism-like behaviors. The impaired morphogenesis of DG is mainly caused by disturbing the postnatal distribution of postmitotic granule cells (GCs), which further results in a migration deficit of neural progenitors. Furthermore, we reveal that Trio plays different roles in various excitatory neural cells by spatial transcriptomic sequencing, especially the role of regulating the migration of postmitotic GCs. In summary, our findings provide evidence of cellular mechanisms that Trio is involved in postnatal DG morphogenesis.

Advances in Single-Cell Transcriptome Sequencing and Spatial Transcriptome Sequencing in Plants.

"Omics" typically involves exploration of the structure and function of the entire composition of a biological system at a specific level using high-throughput analytical methods to probe and analyze large amounts of data, including genomics, transcriptomics, proteomics, and metabolomics, among other types. Genomics characterizes and quantifies all genes of an organism collectively, studying their interrelationships and their impacts on the organism. However, conventional transcriptomic sequencing techniques target population cells, and their results only reflect the average expression levels of genes in population cells, as they are unable to reveal the gene expression heterogeneity and spatial heterogeneity among individual cells, thus masking the expression specificity between different cells. Single-cell transcriptomic sequencing and spatial transcriptomic sequencing techniques analyze the transcriptome of individual cells in plant or animal tissues, enabling the understanding of each cell's metabolites and expressed genes. Consequently, statistical analysis of the corresponding tissues can be performed, with the purpose of achieving cell classification, evolutionary growth, and physiological and pathological analyses. This article provides an overview of the research progress in plant single-cell and spatial transcriptomics, as well as their applications and challenges in plants. Furthermore, prospects for the development of single-cell and spatial transcriptomics are proposed.

MDB-106. MEDULLOBLASTOMA TUMOR MICROENVIRONMENTS SHOW INTRA- AND INTER-TUMORAL HETEROGENEITY ON SPATIAL TRANSCRIPTOMICS

Abstract BACKGROUND Medulloblastoma defines a heterogeneous set of neuroepithelial neoplasms of the posterior fossa. Four groups of medulloblastoma are recognized: wingless (WNT), sonic hedgehog (SHH), group 3, and group 4. The tumor microenvironment (TME) influences tumor progression and response to therapy and has emerged as a growing target for the study of novel therapeutic approaches. METHODS We show spatial gene expression architecture through 10X Visium Spatial Sequencing on 16 formalin-fixed paraffin-embedded (FFPE) samples representing all molecular groups of medulloblastoma. Spatial transcriptomics data is correlated with clinical parameters, including M-stage at diagnosis, tumor histological classification, and outcomes data including survival and relapse status. RESULTS Spatial sequencing demonstrates both intra- and inter-tumoral TME heterogeneity across molecular subgroups. Unsupervised clustering and uniform manifold approximate projection illustrate clusters of distinct spatial gene expression representing cell states from different components of the TME, including tumor-associated astrocytes (TAA), macrophages (TAM), and vascular endothelium. Medulloblastoma cells constitute the majority of cells and express marker genes representing different stages of neuronal differentiation and cell cycle. Pathway analysis reveals the enrichment of pathways associated with cell motility (RHO GTPase, MET, MAPK signaling), heat shock response, growth factor signaling, and adaptive immune response (TLR, MHC-II signaling). In addition, molecular pathways known in medulloblastoma, including NF-kB and TP53 regulatory pathways, were also enriched. Neighborhood-enrichment analysis reveals the co-localization of TAMs and TAAs in close spatial relationship with mitotic progenitor-like medulloblastoma cells. Additionally, these two cell types constitute a greater proportion of the TME in patients at relapse compared to initial diagnosis. CONCLUSION Spatial transcriptomics enables new insight into the heterogeneity of the medulloblastoma TME. Notably, our analysis elucidates the spatial association of TAMs and TAAs with proliferating tumor cells and an increased abundance of these cell types following relapse. These initial results support the role of TAMs and TAAs in medulloblastoma progression.

Outlier detection in temporal and spatial sequences via correlation analysis based on graph neural networks

Outlier detection is essential for identifying patterns that deviate from expected normal representations in data. Real-world challenges such as the lack of labeled data, noise, and high dimensionality significantly impact the effectiveness of existing methods. Understanding the temporal or spatial correlation of normal data is crucial for detecting or forecasting anomalous patterns. In this paper, we introduce UOSC-GNN, a novel Unsupervised Outlier detection architecture by Sequential Correlation analysis with Graph Neural Network. The architecture includes a deviation generation module to measure the variance between expected and actual states of sequential data. This module incorporates a Generic Feature Extraction component to extract intrinsic features from outliers and normal instances tailored to specific tasks, and an Expected State Estimator component based on Graph Neural Network to learn sequential patterns. To ensure the forecast of outliers with high confidence and improve alarm accuracy, an Outlier Probability Assessment module is introduced. This module combines a rule-based index derived from expert knowledge and a statistical index calculated based on generated deviations. Our method is evaluated on two real-world tasks: medical imaging analysis utilizing spatial-related data and early fault detection of instruments leveraging temporal-related data. The results show that our method triggers alarms about 70 min earlier than the best models on the run-to-failure bearing dataset and achieves an accuracy of 92.82% and a sensitivity of 88.51% on the wireless capsule endoscopy image dataset, outperforming traditional outlier detection algorithms consistently.

DIPG-49. SPATIAL TRANSCRIPTOMIC SEQUENCING OF A DIPG PATIENT TUMOUR RESOURCE

Abstract BACKGROUND Diffuse Intrinsic Pontine Glioma (DIPG), the most common subtype of Diffuse Midline Glioma (DMG), is a brainstem high-grade glioma arising in young children and always proves fatal. Major challenges include the disease’s intratumoural genetic and cellular heterogeneity as well as its diffusely invasive phenotype. Research of the last few years has highlighted that interactions between cancer cells, the nervous system and immune cells crucially contribute to tumour cell invasion and disease malignancy. Hence, a better understanding of cancer-promoting interactions is a key requirement to elucidate novel strategies for the treatment of this devastating disease. METHOD We performed spatial transcriptomic sequencing on 10 sequential tumour-infiltrated DIPG patient brainstem regions, using the 10x Genomics Visium platform. RESULTS Spatially-informed analysis and cell type clustering overall highlighted the molecular intratumoural heterogeneity present in the disease. We assessed cell cluster interactions as well as receptor-ligand (RL) pairing, illustrating communication between neuronal, immune cell and tumour clusters. We further determined respective RL pairs involved in such. To address the question whether invasive edges genetically differ from the tumour core, we evaluated genes that are significantly correlated with distal brainstem regions compared to tumour proximal sites. Results revealed gene signatures that are highly associated with neuronal signalling and communication, validating previous findings emphasising the importance of neuron-to-glioma interactions driving DIPG disease progression. CONCLUSION DIPG is a lethal paediatric brainstem glioma for which more durable therapeutic modalities remain to be found. In-depth analysis of this disease is often hampered due to the lack of patient tissue suitable for laboratory studies. Our patient autopsy specimen, spanning the entire tumour-infiltrated brainstem provided a platform to study this aggressive disease in a spatio-temporal manner. This analysis elucidates how DIPG is driven by the tumour microenvironment, including associated receptor-ligand interactions. This has the potential to ultimately discover new treatment targets.

Deciphering the molecular landscape of human peripheral nerves: implications for diabetic peripheral neuropathy.

Diabetic peripheral neuropathy (DPN) is a prevalent complication of diabetes mellitus that is caused by metabolic toxicity to peripheral axons. We aimed to gain deep mechanistic insight into the disease process using bulk and spatial RNA sequencing on tibial and sural nerves recovered from lower leg amputations in a mostly diabetic population. First, our approach comparing mixed sensory and motor tibial and purely sensory sural nerves shows key pathway differences in affected nerves, with distinct immunological features observed in sural nerves. Second, spatial transcriptomics analysis of sural nerves reveals substantial shifts in endothelial and immune cell types associated with severe axonal loss. We also find clear evidence of neuronal gene transcript changes, like PRPH, in nerves with axonal loss suggesting perturbed RNA transport into distal sensory axons. This motivated further investigation into neuronal mRNA localization in peripheral nerve axons generating clear evidence of robust localization of mRNAs such as SCN9A and TRPV1 in human sensory axons. Our work gives new insight into the altered cellular and transcriptomic profiles in human nerves in DPN and highlights the importance of sensory axon mRNA transport as an unappreciated potential contributor to peripheral nerve degeneration.

Optimizing the design of spatial genomic studies

Spatial genomic technologies characterize the relationship between the structural organization of cells and their cellular state. Despite the availability of various spatial transcriptomic and proteomic profiling platforms, these experiments remain costly and labor-intensive. Traditionally, tissue slicing for spatial sequencing involves parallel axis-aligned sections, often yielding redundant or correlated information. We propose structured batch experimental design, a method that improves the cost efficiency of spatial genomics experiments by profiling tissue slices that are maximally informative, while recognizing the destructive nature of the process. Applied to two spatial genomics studies—one to construct a spatially-resolved genomic atlas of a tissue and another to localize a region of interest in a tissue, such as a tumor—our approach collects more informative samples using fewer slices compared to traditional slicing strategies. This methodology offers a foundation for developing robust and cost-efficient design strategies, allowing spatial genomics studies to be deployed by smaller, resource-constrained labs.

Spatial whole exome sequencing reveals the genetic features of highly-aggressive components in lung adenocarcinoma

In invasive lung adenocarcinoma (LUAD), patients with micropapillary (MIP) or solid (SOL) components had a significantly poorer prognosis than those with only lepidic (LEP), acinar (ACI) or papillary (PAP) components. It is interesting to explore the genetic features of different histologic subtypes, especially the highly aggressive components.Based on a cohort of 5,933 patients, this study observed that in different tumor size groups, LUAD with MIP/SOL components showed a different prevalence, and patients with ALK alteration or TP53 mutations had a higher probability of developing MIP/SOL components. To control individual differences, this research used spatial whole-exome sequencing (WES) via laser-capture microdissection of five patients harboring these five coexistent components and identified genetic features among different histologic components of the same tumor. In tracing the evolution of components, we found that titin (TTN) mutation might serve as a crucial intratumor potential driver for MIP/SOL components, which was validated by a cohort of 146 LUAD patients undergoing bulk WES. Functional analysis revealed that TTN mutations enriched the complement and coagulation cascades, which correlated with the pathway of cell adhesion, migration, and proliferation.Collectively, the histologic subtypes of invasive LUAD were genetically different, and certain trunk genotypes might synergize with branching TTN mutation to develop highly aggressive components.

Optimization study of pruning strategy based on KNN trajectory similarity query

With the development of GPS positioning technology, a large amount of spatiotemporal trajectory data has been generated. Due to the complex structure of trajectory data, which has irregular spatial shapes and continuous temporal sequence attributes, querying massive trajectory data poses certain challenges. The pruning strategy of existing trajectory similarity query methods is not very effective. Even after pruning operations, there are still a large number of trajectories that need to undergo distance calculations to confirm whether they are similar trajectories. This paper proposes multiple local pruning optimization schemes, which maximally reduce the number of trajectories in the candidate set. Specifically, it starts with region pruning in the index space, eliminating index spaces with distances greater than the maximum similarity distance from the query trajectory. Then, it performs distance pruning between trajectories, removing trajectories that do not meet the distance conditions. Finally, it adds a termination condition: when the distance between the current index space and the query trajectory is greater than the maximum similarity distance and the number of trajectories in the result set is K, the loop is exited, and the query process ends. Tests on real datasets demonstrate that the KTSS method outperforms current algorithms of the same type.

Spatially and Single‐Cell Resolved Profiling of RNA Life Cycle and Epitranscriptomics

AbstractThe intricate network of cell functions relies on gene expression programs, where the whole RNA life cycle from DNA to protein is subjected to extensive transcriptional and post‐transcriptional gene regulation events. Established bulk RNA sequencing methods provide an averaged, transcriptome‐wide quantification of the RNA life cycle, including transcription, processing, translation, transport, and degradation through RNA‐protein interactions. Furthermore, numerous studies using bulk epitranscriptomic profiling unveiled that dynamic RNA modifications (e. g., N6‐Methyladenosine), add another layer of gene regulations. However, many regulatory events are cell‐type specific, subcellularly localized, and subjected to cell‐cell communications within the native tissue environment. Thanks to the advances in single‐cell sequencing, spatial sequencing, and highly multiplexed imaging methods, we can routinely measure single‐cell and spatial transcriptomics. Yet more comprehensive methods to profile every step of the RNA life cycle with single‐cell and spatial information are still lacking. In this review, we will summarize and compare early explorations in developing state‐of‐the‐art methods for spatially and single‐cell resolved mapping of RNA kinetics, translation, RNA‐protein interactions, and epitranscriptomics. It is promising that these new techniques will greatly facilitate our understanding of the RNA‐centered regulation landscapes in different cell types and how the post‐transcriptional regulations are interconnected within cellular and tissue architecture.

Short-Term and Imminent Rainfall Prediction Model Based on ConvLSTM and SmaAT-UNet.

Short-term precipitation forecasting methods are mainly divided into statistical forecasting, numerical model-based forecasting, and radar image extrapolation techniques. The two methods based on statistical prediction and numerical model have the disadvantages of being unstable and generating large errors. Therefore, this study proposes the use of deep learning for radar image extrapolation for precipitation forecasting, in particular by developing algorithms for ConvLSTM and SmaAT-UNet. The ConvLSTM model is a fusion of a CNN (Convolutional Neural Network) and LSTM (Long Short-Term Memory network), which solves the challenge of processing spatial sequence data, which is a task that traditional LSTM models cannot accomplish. At the same time, SmaAT-UNet enhances the traditional UNet structure by incorporating the CBAM (Convolutional Block Attention Module) attention mechanism and replacing the standard convolutional layer with depthwise separable convolution. This innovative approach aims to improve the efficiency and accuracy of short-term precipitation forecasting by improving feature extraction and data processing techniques. Evaluation and analysis of experimental data show that both models exhibit good predictive ability, with the SmaAT-UNet model outperforming ConvLSTM in terms of accuracy. The results show that the performance indicators of precipitation prediction, especially detection probability (POD) and the Critical Success index (CSI), show a downward trend with the extension of the prediction time. This trend highlights the inherent challenges of maintaining predictive accuracy over longer periods of time and highlights the superior performance and resilience of the SmaAT-UNet model under these conditions. Compared with the statistical forecasting method and numerical model forecasting method, its accuracy in short-term rainfall forecasting is improved.

Quantitative Analysis Method of the Organizational Characteristics and Typical Types of Landscape Spatial Sequences Applied with a 3D Point Cloud Model

(1) Background: Sequential landscape changes give people experiences of dynamic beauty, and the key to creating spatial sequences lies in the organization of spatial changes. The purpose of this study is to use a 3D point cloud model to achieve a refined description of spatial sequences’ organizational characteristics from information acquisition to description and explore the quantitative interpretation methods of typical sequence organizational characteristics. (2) Methods: The proposed model contains three main steps: data acquisition and extraction, characteristic index system construction and data processing, and quantitative characterization analysis. A typology research method that combines quantitative induction with qualitative statistical verification is proposed. (3) Results: Seventy-two spatial sequence point cloud models of study cases are obtained; 4 indicators are established; 3 typical organization types are summarized, namely fluctuating, reversal and moderate type; and the characterization factors and threshold intervals for each sequence organization type are analyzed to validate the type classification result. (4) Conclusions: This research improves the accuracy of spatial data, the comprehensiveness of sequence organization characterization factors, and the reliability of classification results. It supplements the existing spatial sequence theoretical knowledge system and provides parameters that can be referred to in practical design.

Analysis of Intracellular Communication Reveals Consistent Gene Changes Associated with Early-Stage Acne Skin.

A comprehensive understanding of the intricate cellular and molecular changes governing the complex interactions between cells within acne lesions is currently lacking. Herein, we analyzed early papules from six subjects with active acne vulgaris, utilizing single-cell and high-resolution spatial RNA sequencing. We observed significant changes in signaling pathways across seven different cell types when comparing lesional skin samples (LSS) to healthy skin samples (HSS). Using CellChat, we constructed an atlas of signaling pathways for the HSS, identifying key signal distributions and cell-specific genes within individual clusters. Further, our comparative analysis revealed changes in 49 signaling pathways across all cell clusters in the LSS- 4 exhibited decreased activity, whereas 45 were upregulated, suggesting that acne significantly alters cellular dynamics. We identified ten molecules, including GRN, IL-13RA1 and SDC1 that were consistently altered in all donors. Subsequently, we focused on the function of GRN and IL-13RA1 in TREM2 macrophages and keratinocytes as these cells participate in inflammation and hyperkeratinization in the early stages of acne development. We evaluated their function in TREM2 macrophages and the HaCaT cell line. We found that GRN increased the expression of proinflammatory cytokines and chemokines, including IL-18, CCL5, and CXCL2 in TREM2 macrophages. Additionally, the activation of IL-13RA1 by IL-13 in HaCaT cells promoted the dysregulation of genes associated with hyperkeratinization, including KRT17, KRT16, and FLG. These findings suggest that modulating the GRN-SORT1 and IL-13-IL-13RA1 signaling pathways could be a promising approach for developing new acne treatments.

Distinct roles of TREM2 in central nervous system cancers and peripheral cancers

Glioblastomas (GBM) are incurable central nervous system (CNS) cancers characterized by substantial myeloid cell infiltration. Whether myeloid cell-directed therapeutic targets identified in peripheral non-CNS cancers are applicable to GBM requires further study. Here, we identify that the critical immunosuppressive target in peripheral cancers, triggering receptor expressed on myeloid cells-2 (TREM2), is immunoprotective in GBM. Genetic or pharmacological TREM2 deficiency promotes GBM progression invivo. Single-cell and spatial sequencing reveals downregulated TREM2 in GBM-infiltrated myeloid cells. TREM2 negatively correlates with immunosuppressive myeloid and Tcell exhaustion signatures in GBM. We further demonstrate that during GBM progression, CNS-enriched sphingolipids bind TREM2 on myeloid cells and elicit antitumor responses. Clinically, high TREM2 expression in myeloid cells correlates with better survival in GBM. Adeno-associated virus-mediated TREM2 overexpression impedes GBM progression and synergizes with anti-PD-1 therapy. Our results reveal distinct functions of TREM2 in CNS cancers and support organ-specific myeloid cell remodeling in cancer immunotherapy.