Single-cell and Spatial Transcriptomic Mapping Reveals MSEC-myoCAF Interactions Driving Lymph Node Metastasis in Colorectal Cancer

Single-cell and Spatial Transcriptomic Mapping Reveals MSEC-myoCAF Interactions Driving Lymph Node Metastasis in Colorectal Cancer | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Single-cell and Spatial Transcriptomic Mapping Reveals MSEC-myoCAF Interactions Driving Lymph Node Metastasis in Colorectal Cancer Xinxing Li, Jinran Wu, Xiaomao Yin, Jiahui Yin, Yan Wang, Jiexuan Wang, and 14 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6803290/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Lymph node metastasis (LNM) is a major determinant of poor prognosis in colorectal cancer (CRC), yet the cellular and spatial mechanisms driving metastatic initiation remain poorly defined. Here, we performed integrative single-cell RNA sequencing (scRNA-seq; 84,735 cells) and spatial transcriptomics (ST; 20,859 spots) on paired primary tumors and metastatic lymph nodes from five CRC patients. We identified a metastatic stem-like epithelial cell population (MSECs), marked by LGR5, AXIN2, UBE2H, and LAMC2, enriched at invasive fronts and metastatic niches. MSECs co-localized with myofibroblastic cancer-associated fibroblasts (myoCAFs) and communicated via a COL1A1-SDC4 axis that activated NOTCH signaling and promoted epithelial–mesenchymal transition (EMT). Functional assays demonstrated that UBE2H or LAMC2 knockdown impaired CRC cell migration, invasion, and EMT, while their high expression predicted poor disease-free survival. Spatial analysis further revealed that myoCAFs formed a stromal barrier facilitating immune exclusion and metastatic expansion. These findings highlight the MSEC–myoCAF crosstalk as a key driver of CRC metastasis and nominate UBE2H, LAMC2, and COL1A1-SDC4 as potential therapeutic targets. Biological sciences/Cancer/Gastrointestinal cancer/Colorectal cancer Biological sciences/Cancer/Gastrointestinal cancer/Colorectal cancer/Rectal cancer Biological sciences/Cancer/Cancer microenvironment Biological sciences/Cancer/Tumour heterogeneity Colorectal cancer Metastatic stem-like epithelial cells Myofibroblastic CAFs Tumor microenvironment Spatial transcriptomics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Colorectal cancer (CRC) remains one of the leading causes of cancer-related mortality worldwide, with lymph node metastasis (LNM) representing a pivotal prognostic marker and a major therapeutic obstacle [1] . Despite notable progress in targeted therapies and precision medicine, the 5-year survival rate for patients with stage III CRC has stagnated around 65% [2] , underscoring the urgent need to elucidate the molecular and cellular mechanisms that underpin metastatic progression. Metastasis is a complex, multistep cascade encompassing tumor cell invasion, intravasation, survival in circulation, extravasation, and eventual colonization at distant sites. In CRC, lymph nodes are typically the first destination of metastatic spread [3] . A deeper mechanistic understanding of LNM is crucial for the development of effective interventions that can limit dissemination and improve clinical outcomes. Recent studies have drawn attention to the critical roles of tumor cell stemness and stromal remodeling in CRC metastasis [4] . Cancer stem cells (CSCs)—a specialized subpopulation endowed with self-renewal capability and phenotypic plasticity—are proposed to initiate metastasis by generating heterogeneous progenies adapted to diverse microenvironments [5] . These cells are notoriously resistant to conventional therapies and are believed to be key drivers of tumor recurrence and distant metastasis [6-7] . However, despite their functional relevance, the spatial dynamics of CSCs and their interactions with the surrounding stromal components within the tumor microenvironment (TME) remain poorly characterized, hampering the development of CSC-targeted therapies. The TME is composed of a heterogeneous array of cell types, including immune cells, endothelial cells, and cancer-associated fibroblasts (CAFs), all of which modulate tumor behavior and contribute to metastasis. Among them, myofibroblast-like CAFs (myoCAFs) have emerged as key mediators of metastatic niche formation [8] . These cells actively secrete extracellular matrix (ECM) proteins, cytokines, and growth factors that facilitate epithelial–mesenchymal transition (EMT)—a fundamental biological process that enhances epithelial cell motility and invasiveness [9-10] . Despite this growing understanding, it remains unclear whether specific stem-like epithelial populations cooperate spatially with CAFs to drive LNM in CRC. Elucidating the spatial architecture and functional interactions within the TME is essential for discovering novel mechanisms of metastasis and identifying actionable therapeutic targets. The advent of spatial multi-omics technologies has enabled unprecedented resolution in mapping cell–cell communication and tissue architecture within intact tumor specimens [11-16] . These platforms allow for simultaneous analysis of gene expression, protein localization, and cellular spatial proximity, providing a holistic view of the TME and its role in cancer progression [17] . Leveraging these advances, we hypothesized that metastatic progression in CRC is orchestrated by stem-like epithelial cells that interact with adjacent myoCAFs through spatially organized signaling pathways. This hypothesis is supported by frequent co-localization of CSCs and CAFs at the invasive margins of tumors, suggesting potential cooperative interactions. To rigorously test this hypothesis, we conducted an integrative study combining single-cell RNA sequencing (scRNA-seq), spatial transcriptomics (ST), and functional validation assays. Our goal was to identify metastatic epithelial subpopulations and delineate their spatial and molecular crosstalk with myoCAFs in the context of lymph node metastasis. This multi-modal approach enabled us to map high-resolution cellular landscapes while preserving spatial context, thereby uncovering critical pathways that govern CRC dissemination. Ultimately, our findings provide novel insights into the metastatic trajectory of CRC, particularly its spread to mesenteric lymph nodes, and highlight promising therapeutic targets for early intervention in metastatic disease. Results A single-cell transcriptomic atlas of primary and metastatic CRC lesions To comprehensively characterize the cellular landscape of primary colorectal tumors and matched metastatic lymph nodes we performed scRNA-sequsing the 10x Genomics platform on ten tissue samples obtained from five CRC patients with confirmed lymph node metastasis. Each patient contributed a pair of samples: one from the primary tumor and one from the corresponding metastatic lymph node. The experimental workflow and detailed clinical-pathological features of the cohort are summarized in Figure 1A and Supplementary Table 1 . Following data preprocessing—including rigorous quality control and batch effect correction—we integrated the transcriptomes across all samples and conducted dimensionality reduction using UMAP. This analysis resolved 20 transcriptionally distinct clusters across the entire dataset, visualized in two-dimensional space ( Figure 1B, Figure S2A ). In total, 84,735 high-quality single cells were retained for downstream analyses. Based on canonical lineage markers, we annotated the clusters into 10 major cell types, comprising both non-immune and immune populations. The four non-immune lineages included epithelial cells (EPCAM, CDH1, KRT18), fibroblasts (COL1A1, COL1A2, COL3A1), endothelial cells (PECAM1, CDH5, VWF), and smooth muscle cells (ACTA2, TAGLN, MYH1). The six immune lineages were represented by B cells (CD19, CD79A, CD79B), T cells (CD3G, CD3D, NKG7), neutrophils (CSF3R, S100A8, S100A9), mast cells (CPA3, MS4A2, KIT), monocytes (CD14, CD300E, VCAN), and macrophages (CD68, C1QA, C1QB) ( Figure 1C, 1F, and Figure S1C, D ). We next quantified the relative abundance of these major cell types across primary tumor and lymph node tissues, revealing distinct compositional differences between the two anatomical sites ( Figure 1E ). This high-resolution atlas provides a foundational framework for dissecting the cellular programs and microenvironmental remodeling events associated with CRC metastasis. Metastatic stem-like epithelial cells as potential initiators of CRC dissemination As CRC originates from epithelial cells, we performed high-resolution UMAP-based clustering on epithelial subsets, identifying nine transcriptionally distinct subclusters ( Figure 2A ). Marker gene expression patterns for each cluster are shown in Figure 2B . Quantitative analysis across sample origins revealed that cluster C6 was predominantly enriched in primary tumor samples, whereas clusters C3 and C4 were significantly enriched in both primary tumors and corresponding metastatic lymph nodes, suggesting a potential role in metastatic progression ( Figures 2C, S2A , and S3B–C ). To further characterize the biological nature of these subpopulations, we applied copy number variation (CNV) analysis, a commonly used method in scRNA-seq to infer malignant transformation and track tumor evolution [18-19] . Among all epithelial clusters, C6 exhibited the lowest CNV burden, consistent with normal colonic epithelial identity, while C3 and C4 displayed markedly elevated CNV levels, supporting their classification as malignant cell populations [20] ( Figure 2D and Figure S3D ). We next evaluated the differentiation potential of epithelial clusters using CytoTRACE [21] , which revealed that clusters C3 and C4 harbored the highest stemness scores and least differentiation, indicating a poorly differentiated, stem-like state ( Figure 2E and Figure S2D ). These putative metastatic stem-like epithelial cells (MSECs) also exhibited elevated EMT and proliferation scores ( Figure S2C, E ), reinforcing their aggressive phenotype. Pseudotime trajectory analysis placed MSECs at the apex of the differentiation hierarchy, suggesting they give rise to other epithelial lineages within the tumor ( Figure 2F ). RNA velocity analysis further demonstrated directional flow of transcriptomic states originating from MSECs, supporting their role as progenitor cells ( Figure 2G ). Differential gene expression analysis identified UBE2H and LAMC2 as specifically upregulated in MSECs, alongside canonical stemness-associated markers such as LGR5, AXIN2, SOX9, CD44, and ALCAM ( Figure 2H ). Notably, UBE2H and LAMC2 expression levels were strongly correlated with stemness metrics across the epithelial compartment. Pathway enrichment analysis revealed that WNT, EMT, and partial EMT (pEMT) signaling pathways were significantly activated in MSECs ( Figure 2I and Figure S2B ), further linking them to metastatic potential. Taken together, these findings define MSECs as a transcriptionally and functionally distinct epithelial subpopulation enriched in metastatic lesions, endowed with high stemness, EMT capacity, and tumor-initiating features. These results suggest that MSECs may serve as the cellular origin of CRC metastasis. Overexpression of LAMC2 and UBE2H confers a metastatic phenotype in CRC UBE2H (ubiquitin-conjugating enzyme E2 H) belongs to the E2 family of ubiquitin-conjugating enzymes and plays a pivotal role in regulating protein ubiquitination, a process essential for controlling cellular adhesion, migration, and invasion—key hallmarks of metastatic progression [22] . Within the MSECs, we observed a strong positive correlation between UBE2H and LAMC2 expression ( Figure 3A , Figure S4A ). Spatial transcriptomics of primary CRC lesions further revealed a distinct epithelial subpopulation with high co-expression of both genes, predominantly localized at the invasive front ( Figure 3B , Figure S4D ). To evaluate their functional relevance, we silenced UBE2H and LAMC2 expression using small interfering RNA (siRNA) in two CRC cell lines, SW1116 and SW480. Knockdown of either gene significantly impaired CRC cell proliferation as measured by in vitro assays ( Figure 3C , Figure S4E ). To validate these findings in clinical specimens, immunohistochemistry (IHC) was performed on CRC tissues with and without lymph node metastasis. Both UBE2H and LAMC2 showed elevated protein expression in lymph node-positive tumors ( Figure 3D ). Further in vitro functional assays demonstrated that knockdown of either UBE2H or LAMC2 led to a marked reduction in cell migration and invasion capabilities ( Figure 3E , Figure S4F ). Mechanistically, gene silencing resulted in increased expression of the epithelial marker E-cadherin, and decreased levels of mesenchymal markers N-cadherin, Vimentin, and the EMT-inducing transcription factor Snail, indicating reversal of the EMT phenotype ( Figure 3F , Figure S4G ). Consistently, pathway enrichment analysis of our scRNA-seq data confirmed a strong positive association between LAMC2, UBE2H, and EMT-related signaling pathways ( Figure 3G , Figure S4B ). Kaplan–Meier survival analysis of CRC patient cohorts revealed that elevated expression of UBE2H and LAMC2 was significantly associated with reduced disease-free survival (DFS) ( Figure 3H , Figure S4C ). Collectively, these results establish UBE2H and LAMC2 as key mediators of the metastatic phenotype in CRC. Their co-expression within MSECs not only defines a transcriptional state of high stemness and invasiveness but also contributes functionally to enhanced tumor aggressiveness, suggesting their potential as biomarkers and therapeutic targets in metastatic CRC. Spatial transcriptomics reveals localization and interactions of metastatic stem-like cells Spatial context is essential for decoding tumor architecture and cellular interactions, yet conventional scRNA-seq lacks positional information. To overcome this limitation, we performed spatial transcriptomics (ST-seq) on five CRC samples with confirmed lymph node metastasis (LN+) to map in situ gene expression patterns and spatial organization. Based on hematoxylin and eosin (H&E) staining, spatial spots were annotated into tumor, stromal, and epithelial regions ( Figure 4A ). CNV analysis revealed the highest CNV burden in tumor-designated regions, consistent with their malignant identity ( Figure S5A, B ). To determine cellular composition within these regions, we applied Robust Cell Type Decomposition (RCTD) [23-24] , a reference-based deconvolution algorithm. As expected, stromal regions were enriched with endothelial cells, while epithelial and tumor regions were dominated by epithelial cells. Notably, fibroblasts were highly concentrated along the tumor–stroma boundary, suggesting a spatially defined fibroblast population at the invasive front ( Figure 4A ), potentially shaping the local microenvironment. Across five CRC tissue sections, we obtained transcriptomic data from 20,859 spatial barcodes ( Figure 4B ) and evaluated EMT scores at tumor margins [25] . EMT scoring revealed significantly higher EMT activity at the tumor front compared to the tumor core ( Figure 4C ), supporting the notion that MSECs preferentially localize at the invasive edge—a finding that corroborates our scRNA-seq results ( Figure 4D, Figure S5D, E ). To explore intercellular interactions within this spatial context, we conducted ligand-receptor interaction analysis and identified significant enrichment of the COL1A1–SDC4 axis in fibroblas-epithelial communication ( Figure 4E ). This interaction was predominantly observed at the tumor–stroma interface, where MSECs and myofibroblast-like CAFs (myoCAFs) co-localize. Notably, both COL1A1 and SDC4 have previously been implicated in promoting colorectal cancer metastasis by facilitating tumor cell adhesion, motility, invasion, and maintenance of stem-like traits [26-27] . In summary, ST-seq analysis provided critical spatial insight into the localization of MSECs within the tumor microenvironment. These cells are enriched at the tumor front, where they interact with surrounding myoCAFs through pro-metastatic signaling pathways such as COL1A1–SDC4, reinforcing their role in metastatic progression. Enhanced interactions between myoCAFs and metastatic stem-like cells promote lymph node metastasis in CRC Cancer metastasis is not solely driven by tumor-intrinsic mechanisms but is profoundly influenced by stromal components within the TME, particularly CAFs [28-29] . Recent advances in single-cell RNA sequencing have enabled high-resolution characterization of CAF heterogeneity across multiple malignancies [30-31] . In our dataset, fibroblasts were subdivided into five transcriptionally distinct populations based on specific marker gene expression, and their proportional distributions were analyzed across sample types ( Figure 5A-C ). Further classification revealed two dominant CAF phenotypes: epithelial-like CAFs (eCAFs) and myoCAFs ( Figure 5D, E ), the latter being associated with pro-metastatic functions. To elucidate the molecular crosstalk between CAFs and MSECs, we performed ligand-receptor interaction analysis. The COL1A1-SDC4 ligand-receptor pair emerged as a dominant signaling axis between myoCAFs and MSECs, consistent with our findings from spatial transcriptomic analysis ( Figure 5G ). Notably, COL1A1-SDC4-mediated signaling was significantly more pronounced in myoCAF-MSEC interactions compared to those involving other tumor epithelial populations (Figure S6G ). In addition, we observed that MSECs exhibited the highest frequency and strength of fibroblast-mediated intercellular interactions. Among these, NOTCH signaling—an established pathway implicated in tumor progression and stemness—was markedly activated within MSECs engaged by myoCAFs ( Figure 5F, I and Figure S6E ). Based on these findings, we hypothesized that MSECs and myoCAFs are spatially co-localized within the tumor niche. Using RCTD-based spatial mapping, we observed partial but significant co-localization of these two cell populations in the primary tumor ( Figure 5H, Figure S6F ). This spatial association was further validated through multiplex immunohistochemistry (mIHC), which demonstrated significantly elevated COL1A1 and SDC4 co-expression in lymph node metastasis-positive (Tumor/LN + ) samples compared to non-metastatic (Tumor/LN - ) counterparts ( Figure 5J ). Collectively, these findings reveal an intensive, spatially organized interaction between MSECs and myoCAFs, primarily mediated through the COL1A1-SDC4 axis and associated with NOTCH activation, supporting a cooperative role in facilitating CRC lymph node metastasis. Distinct fibroblast states determined by spatial proximity to tumor epithelial cells CAFs are pivotal components of the solid tumor microenvironment, known to modulate tumor growth, invasion, and metastasis [32-33]. Given their dynamic interactions with tumor epithelial cells and their impact on clinical outcomes, we further dissected the molecular features of fibroblasts in CRC based on spatial localization relative to the tumor epithelium. Using spatial transcriptomics (ST) data, we mapped tumor epithelial regions and quantified the proximity of fibroblasts to epithelial compartments to assess their spatial organization ( Figure 6A, Figure S7A ). We observed that fibroblast localization significantly influenced the composition of neighboring cells within theTME [34] . Specifically, immune cell density—particularly B cells and T cells—increased with greater fibroblast distance from the tumor epithelium, whereas tumor cell density was highest in proximity to fibroblasts and declined with increasing distance ( Figure 6B, Figure S7B ), suggesting spatially dependent immune modulation. Fibroblasts were subsequently stratified into two categories based on their median spatial distance from the tumor epithelium: epithelial-adjacent fibroblasts and epithelial-distant fibroblasts. Gene expression profiling revealed that epithelial-distant fibroblasts expressed markers such as ZG16, OLFM4, CA2, PLA2G2A, DUOX2, IGHA1, PPBP, GUCA2A, FABP1, and CEACAM7, predominantly enriched in fibroblast-rich peripheral zones of the TME ( Figure 6C ). In contrast, epithelial-adjacent fibroblasts showed elevated expression of FN1, COL5A1, IGHG1, AEBP1, COL1A2, THBS2, COL1A1, COL3A1, MMP2, and C3, localizing mainly to central tumor regions with dense fibroblast infiltration ( Figure 6D, Figure S7C ). To further explore the functional implications of this spatial dichotomy, we computed enrichment scores for epithelial-distant and epithelial-adjacent fibroblast signatures within our scRNA-seq dataset. Projection of these scores revealed clustering patterns that were highly consistent with our prior CAF subtyping ( Figure 6E ). Notably, epithelial-adjacent fibroblasts closely aligned with myoCAF populations, while epithelial-distant fibroblasts were more closely associated with eCAFs. These results indicate that fibroblast populations exhibit distinct spatially dependent states, which correspond to divergent molecular programs and functional phenotypes. Proximity to tumor epithelial cells appears to dictate CAF identity and behavior, potentially influencing tumor-stroma crosstalk, immune cell exclusion, and metastatic progression in CRC. Macrophages exhibit immunosuppressive and tumor-promoting phenotypes in CRC We identified a total of 3,021 macrophages across the single-cell RNA sequencing dataset and further classified them into seven transcriptionally distinct subpopulations ( Figure 7A ). Comparative analysis revealed that primary tumors harbored a higher abundance of DCSTAMP + , LYPD2 + , and CCDC141 + macrophages, whereas CCL20 + macrophages were relatively less frequent in the primary site compared to metastatic lesions ( Figure 7B ). To delineate macrophage differentiation trajectories, we conducted RNA velocity analysis. This revealed that BCL11B + macrophages likely serve as the progenitor population, giving rise to two major branches: one differentiating toward CCL20 + and P2RY6 + macrophages, and the other toward CCDC141 + and LYPD2 + subsets ( Figure 7C ). This trajectory was further validated by pseudotime analysis using Monocle2, which consistently positioned BCL11B + macrophages at the root of the lineage tree, generating multiple transcriptionally distinct endpoints ( Figure 7D ). Gene expression profiling demonstrated that BCL11B + macrophages exhibited low expression of inflammation- and immunity-associated genes, indicative of an immature or quiescent state. In contrast, CCL20 + , P2RY6 + , and CCDC141 + macrophages showed robust upregulation of pro-inflammatory and immune-modulatory genes, albeit with distinct transcriptional programs, suggesting divergent functional roles within the tumor microenvironment ( Figure 7E ). Interestingly, all macrophage subsets expressed elevated levels of immune checkpoint molecules. Specifically, CCL20 + macrophages upregulated TIGIT and CTLA4, while P2RY6 + cells showed high expression of CD274 (PD-L1) and VSIR (VISTA). LYPD2 + and CCDC141 + subsets displayed the highest levels of multiple checkpoint molecules, consistent with their terminal differentiation phenotypes. These data indicate that macrophages may contribute to immune evasion through checkpoint-mediated suppression. To further assess functional polarization, we computed M1/M2 macrophage polarization scores [35] . The majority of macrophages displayed an M2-like immunosuppressive phenotype, with the exception of BCL11B + macrophages, which showed low scores for both M1 and M2 signatures (Figure 7F-G). This suggests that most macrophage subsets in the CRC microenvironment adopt a pro-tumorigenic, immune-suppressive state, potentially contributing to metastatic progression and therapeutic resistance. Discussion Intratumoral heterogeneity in CRC plays a fundamental role in shaping tumor progression and metastatic potential. While scRNA-seq has offered valuable insights into the cellular diversity and transcriptomic states of primary CRC and liver metastases [36] , the molecular and spatial mechanisms governing lymph node metastasis remain largely unresolved. In particular, the specific epithelial subpopulations responsible for initiating metastasis have not been clearly defined. In this study, we integrated scRNA-seq with ST to characterize MSECs within both primary CRC tumors and paired metastatic lymph nodes, revealing their spatial distribution, temporal dynamics, and potential functional significance. CSCs have been widely implicated as critical drivers of metastatic dissemination and tumor recurrence, with mounting evidence indicating that only a subset of malignant cells possess the capacity to propagate tumors at distant sites [6,37-39] . For example, ablation of LGR5 + CSCs in orthotopically xenografted CRC organoid models suppresses tumor growth, whereas their re-emergence restores proliferative capacity during early dissemination [40] . In our study, MSECs were localized predominantly along tumor-invasive boundaries and exhibited molecular features consistent with stemness and metastatic competence. Notably, we observed significant overexpression of UBE2H and LAMC2 in MSECs, two genes previously associated with aggressive tumor phenotypes. UBE2H, a ubiquitin-conjugating enzyme, has been shown to be upregulated in CRC and linked to poor prognosis and metastatic progression, consistent with reports in lung adenocarcinoma [22] . Similarly, LAMC2 has been implicated in promoting local invasion and lymphatic spread in CRC and may serve as a prognostic biomarker [41] . Functional assays further demonstrated that silencing UBE2H and LAMC2 markedly reduced CRC cell proliferation, migration, invasion, and EMT activity, underscoring their pro-metastatic roles. Both scRNA-seq and ST data confirmed the enrichment of UBE2H + /LAMC2 + MSECs within primary tumors, suggesting that lymph node metastases may arise from a spatially distinct, transcriptionally defined subpopulation within the primary site. Although previous studies have utilized scRNA-seq to profile distant metastases in CRC [42-44] , lymphatic dissemination often constitutes the earliest step in the metastatic cascade. Our study provides the first spatially resolved single-cell atlas of this early metastatic event, revealing the molecular and metabolic characteristics of disseminating tumor cells. Importantly, our work was conducted using patient-matched primary CRC tumors and metastatic lymph nodes, enabling a clinically relevant reconstruction of tumor evolution. These findings offer new insights into the cellular origins of lymph node metastasis in CRC and suggest potential molecular targets for intercepting early-stage metastatic progression. Metastatic progression involves a cascade of genetic, epigenetic, and transcriptional alterations that endow tumor cells with enhanced invasive and adaptive capacities [45] . Recent advances in high-resolution sequencing technologies have enabled deeper exploration of these evolutionary trajectories. Our study identifies MSECs as a key tumor subpopulation driving metastasis, and underscores the critical role of tumor-microenvironment interactions in this process [46] . Notably, MSECs were preferentially localized at the tumor-host interface, a region often enriched for early disseminating tumor cells. The surrounding microenvironment at this invasive front appears to modulate intercellular communication, reducing direct interactions between tumor cells and stromal components [47] . In squamous cell carcinoma, for example, tumor-specific clusters at the invasive margin engage with multiple stromal lineages to suppress local invasion. Similarly, our CellChat analysis of scRNA-seq data revealed that the NOTCH signaling pathway is highly active in MSECs, particularly through communication with myoCAFs. The NOTCH pathway has a context-dependent role in tumorigenesis and is known to promote tumor stem cell proliferation and self-renewal [48-50] . Additionally, we identified a critical interaction mediated by the SDC4-COL1A1 ligand-receptor pair, which enhances CRC invasiveness and metastatic competence. These findings suggest that therapeutic targeting of the SDC4-COL1A1 axis or its downstream NOTCH signaling may represent a promising strategy for metastatic CRC intervention. Macrophages are essential components of the tumor immune microenvironment and exhibit multifaceted roles in tumor initiation, progression, and metastatic dissemination. In our dataset, macrophages within lymph node metastatic CRC displayed predominantly M2-like immunosuppressive phenotypes, characterized by elevated expression of immune checkpoint and other immunoregulatory genes. This macrophage subset contributes to immune evasion, extracellular matrix (ECM) remodeling, and promotion of metastasis [51-54] . Spatial transcriptomics-based CellChat analysis further confirmed that SDC4-COL1A1 interactions were spatially enriched between tumor epithelial cells and adjacent fibroblasts, reinforcing the notion that the TME is spatially organized and functionally compartmentalized. Moreover, we found that myoCAFs consistently localize in proximity to tumor epithelial cells, a pattern that was robustly observed across both scRNA-seq and spatial datasets. Collectively, these findings underscore the spatial heterogeneity of the TME and its influence on tumor cell behavior, particularly in the metastatic cascade of CRC. There are several limitations to this study. First, the substantial inter-patient heterogeneity poses challenges in identifying universal features across samples. Second, the sample size is limited—we performed scRNA-seq and spatial transcriptomics on only five CRC patients—which restricts the generalizability of our findings. To strengthen the robustness of our conclusions, we applied an integrative strategy combining both technologies to improve data confidence and spatial resolution. Third, while early dissemination of CRC can occur via both hematogenous and lymphatic routes, the detection of early hematogenous metastasis remains technically difficult and inconsistently characterized. As a result, we focused on lymphatic dissemination in this study, while the molecular mechanisms underlying early hematogenous spread warrant future investigation. Our results delineate the spatiotemporal characteristics of MSECs, demonstrating their spatial enrichment at the tumor–host interface and close association with myoCAFs at both single-cell and spatial levels. Importantly, we found that fibroblasts exhibit distinct transcriptional states depending on their spatial proximity to tumor epithelial cells, highlighting the dynamic evolution of CRC cells during early metastatic dissemination. These findings lay a foundation for further exploration of early dissemination mechanisms and offer potential prognostic and therapeutic implications. In summary, we conducted scRNA-seq and spatial transcriptomics to systematically investigate the cellular heterogeneity and tumor ecosystems of paired primary and metastatic lesions in CRC. Through the integrated analysis of 84,735 single cells from five patients, we identified nine epithelial subpopulations, among which clusters C3 and C4 were enriched for stem-like features and metastatic potential. MSECs were characterized by elevated expression of UBE2H and LAMC2, which promoted EMT and metastatic dissemination and were associated with shorter progression-free survival. Spatial analysis confirmed that MSECs localize to the tumor periphery, where they engage myoCAFs via the SDC4-COL1A1 ligand-receptor axis, shaping an immunosuppressive microenvironment and limiting immune infiltration. Additionally, tumor-associated macrophages within metastatic lymph nodes exhibited M2-like polarization and expressed immunoregulatory molecules such as PD-L1 and VISTA, further facilitating immune evasion. Together, this study presents a comprehensive single-cell and spatial atlas of lymph node metastatic CRC, identifying MSECs as key drivers of dissemination. We demonstrate that UBE2H and LAMC2 promote the metastatic phenotype of CRC cells, and that the SDC4-COL1A1-NOTCH axis mediates critical tumor-stroma interactions. These findings provide new mechanistic insights into lymphatic dissemination in CRC and suggest actionable therapeutic targets to prevent metastatic progression. Methods Ethical statement The study was reviewed and approved by the Institutional Review Board of Tongji Hospital, Tongji University. All patients provided written informed consent prior to the collection of samples. Human specimens This study involved five patients who underwent surgerical treatment at the Department of General Surgery, Tongji Hospital, Tongji University. A total of ten fresh surgical specimens were collected, consisting of five primary tumors and five paired metastatic lymph nodes. The presence of lymph node metastases was confirmed by pathologists through intraoperative cytologic assessments and postoperative analysis of paraffin-embedded sections. Clinical data, including demographic and clinicopathologic characteristics of the tumors and staging at diagnosis according to the American Joint Committee on Cancer System, are detailed in Supplementary Table 1 , Cell culture Two CRC cell lines, SW1116 and SW480, were obtained from the Chinese Academy of Sciences (Shanghai, China). The cells were cultured in 5% CO 2 atmosphere at 37℃ using DMEM or RPMI-1640 medium (Corning) supplemented with 1% penicillin-streptomycin and 10% fetal bovine serum (FBS). The cell lines were confirmed to be mycoplasma-free and subjected to monthly authentication via PCR analysis. Preparation of single-cell suspensions Prior to tissue processing, adipose tissue and visible blood vessels were removed. Fresh normal mucosa and CRC tissues were rinsed with ice-cold PBS, and minced into small fragments. Normal mucosal tissues were incubated in 10 mL of EDTA-containing buffer (15 mM HEPES, 5 mM EDTA, 1 mM DTT and 10% FBS-supplemented PBS) for 1 h at 37°C. Tumor tissues were treated with 10 mL of DTT (65 mM)-supplemented PBS (with 10% FBS) for 15 min at 37'C with continuous shaking. Following incubation, EDTA and DTT were removed by washing with PBS twice. The tissue fragments were then digested with collagenase VIII (0.38 mg/ml) and DNase I (0.1 mg/ml) in complete RPMI-1640 medium (with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin) for 1 h at 37°C. Following digestion, the mixture was vigorously shaken for 5 min and cells were mechanically dissociated using 21-gauge syringes. The resulting cell suspension were filtered through a 100 µm filter, pelleted and washed twice with PBS to prepare freshly isolated cell suspensions for scRNA-seq and flow cytometry staining. Single-cell RNA sequencing Freshly prepared cell suspensions were processed immediately according to the manufacturer’s protocol of the 10X Chromium 3’ v3 kit (10x Genomics, Pleasanton, CA). Libraries were prepared and sequenced on the NovaSeq 6000 platform (Illumina, Inc., San Diego, CA) at GENERGY BIO (Shanghai, China). Dimension reduction and clustering analysis Data were scaled using the top 2000 most variable genes via FindVariableFeatures function in the Seurat v3 R package. Principal component analysis (PCA) was performed using variable genes, followed by neighbor identification with FindNeighbors to facilitate graph-based clustering. Cell subtypes were delineated using FindClusters, and UMAP (uniform manifold approximation and projection) visualization was employed to illustrate cell distribution. To mitigate the batch effects, the Harmony algorithm was utilized prior to clustering. Cells were first partitioned into epithelial, stromal, myeloid, T, and B cell categories, followed by a more granular classification into distinct subtypes across samples. In the first step, clusters were scored for the predefined gene signatures [3], including epithelial cells (EPCAM, KRT8, KRT18), stromal cells (COL1A1, COL1A2, COL6A1, COL6A2, VWF, PLVAP, CDH5, S100B), myeloid cells (CD68, XCR1, CLEC9A, CLEC10A, CD1C, S100A8, S100A9, TPSAB1, and OSM), T cells (NKG7, KLRC1, CCR7, FOXP3, CTLA4, CD8B, CXCR6, and CD3D), and B cells (MZB1, IGHA1, SELL, CD19, and AICDA). Signature scores were calculated as the mean log 2 (LogNormalizedUMI+1) across all genes in each signature. Each cluster was assigned to the compartment of its maximal score and all cluster assignments were manually verified to ensure the accurate partitioning. In the second step, the Harmony algorithm was applied before clustering analysis to correct for batch effects, followed by further application of FindNeighbors and FindClusters in Seurat to obtain cell subtypes. Additionally, we defined 58 cell types in CRC based on the gene signatures and known lineage markers for each cell type. InferCNV analysis Initial copy number variations (CNVs) for each region were estimated using inferCNV R package (version 1.0.4). The CNV for each cell type was calculated according to expression level derived from single-cell sequencing data, employing a cutoff of 0.1. Genes were sorted according to their chromosomal locations, and a moving average of gene expression was computed using a window size of 101 genes. The expression values were then centered to zero by subtracting the mean. Epithelial cells were selected as malignant cells, while other cells were considered as normal cells. De-noising procedures were applied to refine the CNV profiles, resulting in the final CNV datasets for anlaysis. Cell Cycle analysis To calculate cell cycle scores for G1, S and G2/M phases, the Cyclone function of the scran R package (version 1.14.3) was applied. Cells were classified as follows: G1 phase (if the G1 score is above 0.5 and greater than the G2/M score); G2/M phase (if the G2/M score is above 0.5 and greater than the G1 score); S phase (if neither score exceeds 0.5). The results were visualized using UMAP to show the proportions of different cell cycle across various clusters or subgroups. Monocle2 Pseudotime Analysis The developmental pseudotime was determined using the Monocle2 package (version 2.9.0). The raw counts were first converted from Seurat object into Cell Data Set object by the importCDS function of Monocle. The differentialGeneTest function was employed to select ordering genes (with q-value < 0.01), which were likely to be informative for ordering cells along the pseudotime trajectory. Dimensional reduction clustering analysis was performed using the reduceDimension function, followed by trajectory inference with the orderCells function using default parameters. Changes in gene expression over pseudotime were visualized with plot_genes_in_pseudotime function, allowing for the tracking of developmental changes throughout the pseudotime trajectory. scVelo Analysis To perform the RNA velocity analysis, the spliced reads and unspliced reads were recounted using the Python script velocyto.py (https://github.com/velocyto-team/velocyto.py) on the output folder from Cell Ranger. Following this, the likelihood-based dynamical model and the velocity graph were constructed using scVelo ( https://scvelo.readthedocs.io/). This process enabled to calculate the RNA velocities, which include rates of transcription, splicing and degradation for individual single cells. The resulting velocity fields were then projected onto the UMAP of Seurat, allowing for visualization of cellular dynamics in the context of the identified clusters. CytoTRACE Analysis CytoTRACE (version 0.3.3) was used to predict the differentiation states from scRNA-seq data. The raw count data was first loaded, and the CytoTRACE function was utilized with the enableFast = TRUE parameter to determine the recommended time order of cells. Visualization of the results was achieved using the plotCytoTRACE function. scMetabolism Analysis scMetabolism (version 0.2.1) was performed to quantify the metabolic activity from scRNA-seq data. The metabolic pathway activity for each cell was assessed using VISIONalgorithm. This analysis leveraged preset preset metabolic gene sets, including 85 KEGG pathways and 82 REACTOME pathways, to score the metabolic activity comprehensively across the sampled cells. Cell subtype similarity analysis To evaluate the similarity of cell subtypes, the following steps were performed: (1) the top 1,000 highly variable genes were identified across different cell subtypes; (2) the mean expression values of these top 1,000 highly variable genes were calculated for each cell subtype; (3) the hierarchical clustering was performed using a distance metric defined by (1-Pearson correlation coefficient)/2. Differential expression analysis To identify differentially expressed genes for each cell subtype, the functions “FindAllMarkers” (multiple condition comparisons) and “FindMarkers” (two condition comparisons) from the Seurat package were used with default parameters. Genes were considered as differentially expressed if they met the criteria of a p-value 0.25. Plasmid constructions and transduction Human UBE2H and LAMC2 were amplified using the reverse-transcribed cDNA from SW1116 and other cell lines. The siRNA targeting UBE2H and LAMC2 were purchased from Sangon Biotech. The specific siRNA sequences are as follows: siUBE2H: Forward Primer TTATTGGCCTATCCTAACCCCA Reverse Primer TGCTTGTATTCTTCTGGTCGGT siLAMC2: Forward Primer GACAAACTGGTAATGGATTCCGC Reverse Primer TTCTCTGTGCCGGTAAAAGCC 10x Visium spatial RNA-seq data analysis The FASTQ files were processed and aligned to GRch38 huma reference genome using Space Ranger software (version 2.0.1) from 10x Genomics, with unique molecular identifier (UMI) counts summarized for each barcode. To distinguish tissue overlaying spots from the background, tissue overlaying spots were detected according to the images. The filtered UMI count matrix was then analyzed using Seurat (version 4.1.0) R package. Bright-field imaging Bright-field images were captured by a whole slide scanner (Panoramic MIDI FL, 3DHISTECH) at a 20× resolution. H&E-stained sections of each sample were carefully reviewed by two experienced pathologists to confirm the pathology. Following this, a trained pathologist manually annotated the images to identify distinct regions, including tumor, stromal, and normal regions. ST barcoded microarray slide Information: The library preparation slides utilized in this study were obtained from the ST team (https://www.spatialtranscriptomics.com). Each slide features an array of spots measuring 55 μm in diameter, spaced 100 μm apart from center to center, covering a total area of 6.5×6.5 mm². Each slide contains four capture zones, with each zone comprising approximately 5,000 unique gene expression spots. This design enables high-resolution spatial transcriptomics analysis, allowing for in-depth examination of gene expression patterns within tissue samples, which is crucial for understanding tissue architecture and cellular interactions. RCTD Analysis To infer the cell-type composition of each spot, RCTD (version 1.1.0) was employed. The analysis utilized the default parameters in the creat. RCTD function, with specific modifications: a minimum number of cell > 1 per cell type and a minimum of unique molecular identifier (UMI) count > 1 per pixel. Additionally, the doublet_mode parameter was configured to FALSE in the run.RCTD function. CellChat Analysis The cell communication analysis was conducted using the CellChat R package (version 1.1.3). The processs began by importing the normalized expression matrix to create a cellchat object using the create CellChat function. Subsequently, the data was preprocessed through several embedded functions with the default parameters: identify Over Expressed Genes, identify Over Expressed Interactions and project Data. Potential ligand-receptor interactions were assessed using the compute Commun Prob, filter Communication (with a minimum of 10 cells) and compute Commun Prob Pathway functions. Finally, the cell communication network was aggregated using the aggregate Net function, allowing for a comprehensive understanding of the interactions between different cell types within the tissue. Cell Communication Analysis Cell communication analysis was conducted using CellPhoneDB (version 4.1.0) to identify biologically relevant ligand-receptor interactions from scRNA-seq data. A ligand or receptor was considered “expressed” in a specific cell type if at least 10% of the cells exhibited non-zero read counts for the corresponding gene. Statistical significance was assessed by randomly shuffling cluster labels, generating a null distribution through 1,000 permutations, and calculating p-values based on a normal distribution curve derived from permuted interaction scores. Cell-cell communication networks were constructed by linking cell types where ligands were expressed in one type and corresponding receptors in another, and these networks were visualized using the Igraph (version 1.2.4.1) and Circlize (version 0.4.8) R packages. In spatial communication analysis, intra-spots represented closely interacting cells of the same type, inter-spots included other cell types within two layers of proximity, and distal-spots indicated minimal interaction. This classification helped elucidate spatial communication dynamics within the tissue. Additionally, spatial co-localization analysis was performed by selecting a specific cell type as the center spot, calculating Euclidean distances, and generating density plots to infer co-localization patterns at varying distances. Abbreviations CRC: Colorectal Cancer; LNM: Lymph Node Metastasis; TME: Tumor Microenvironment; scRNA-seq: Single-Cell RNA sequencing; ST: Spatial Transcriptomics; MSECs: Metastatic Stem-like Epithelial Cells; EMT: Epithelial-Mesenchymal Transition; myoCAFs: myofibroblast fibroblasts; DFS: Disease-Free Survival; CSCs: Cancer stem cells; ECM: Extracellular matrix; CAFs: Cancer-Associated Fibroblasts; IHC: Immunohistochemistry; mIHC: Multiplex immunohistochemical; UBE2H: Uiquitin-conjugating enzyme E2 H; LAMC2: Laminin subunit gamma-2; COL1A1: Collagen Type I Alpha 1 Chain; SDC4: Syndecan-4; UMAP: Manifold Approximation and Projection; CNV: copy number variation; siRNA: small interfering RNA; H&E: hematoxylin and eosin; RCTD: Robust Cell Type Decomposition. Declarations Funding This work was financially supported by the following grants: Jiangsu Provincial Research Hospital Project (YJXYY202204-ZD06), Jiangsu Commission of Health (K2024011), Postdoctoral Research Funding Project of Jiangsu Province (2021K012A), Science and Technology Bureau of Nantong City (JC2023108), Macao Science and Technology Development Fund (0133/2024/RIA2 and 0069/2021/AFJ), Shanghai Science and Technology Commission (23Y11902400), Shanghai Municipal Health Commission (202340231), Tongji University Scientific Research Project (22120240386), Ganquan New Star Talent Program of Shanghai Tongji Hospital (HBRC2104), Shanghai Tongji Hospital Research Funding (RCOD2102, ITJ(ZD)2308, and GJPY2111). Data availability The datasets generated and/or analyzed during the current study are in the process of being uploaded and will be made available to the journal and the public upon request as soon as the deposition is complete. Ethics approval and consent to participate Human colorectal cancer tumor and lymph node samples were collected from patients at Shanghai Tongji Hospital in accordance with institutional ethical guidelines under protocol number 2021-116, approved by the Ethics Committee of Shanghai Tongji Hospital. Informed consent was obtained from all participants prior to sample collection. Competing interests The authors declare no conflicts of interest and no competing financial interests. All participants provided written informed consent. Author Contributions Xinxing Li and Feng Wang conceived and supervised the study. Jinran Wu, Xiaomao Yin, and Jiexuan Wang performed the single-cell and spatial transcriptomic experiments. Lin Zhu, Xin Yang, and Jingjing Qian were responsible for sample processing, quality control, and library preparation. Wenqiang Wang, Liangchen Zhu, Xuan Dai, and Zekun Zhao contributed to clinical sample collection and annotation. Dongsheng Li, Siyuan Yin, and Runqi Hong performed bioinformatics analysis and data interpretation. Kai Xu, Zhiqian Hu, and Io Nam Wong assisted with computational modeling and pathway analysis. Yan Wang and Jiahui Yin contributed to figure preparation and literature review. Jinran Wu and Xiaomao Yin drafted the manuscript with input from all authors. Xinxing Li, Io Nam Wong, and Feng Wang critically revised the manuscript and provided overall project coordination. All authors read and approved the final manuscript. References CHEN K, COLLINS G, WANG H, et al. Pathological Features and Prognostication in Colorectal Cancer [J]. Current oncology (Toronto, Ont), 2021, 28(6): 5356-83. SIEGEL R L, MILLER K D, FUCHS H E, et al. Cancer statistics, 2022 [J]. CA: a cancer journal for clinicians, 2022, 72(1): 7-33. BENSON A B, VENOOK A P, AL-HAWARY M M, et al. Colon Cancer, Version 2.2021, NCCN Clinical Practice Guidelines in Oncology [J]. Journal of the National Comprehensive Cancer Network : JNCCN, 2021, 19(3): 329-59. FRANK M H, WILSON B J, GOLD J S, et al. Clinical Implications of Colorectal Cancer Stem Cells in the Age of Single-Cell Omics and Targeted Therapies [J]. Gastroenterology, 2021, 160(6): 1947-60. BAYIK D, LATHIA J D. Cancer stem cell-immune cell crosstalk in tumour progression [J]. Nature reviews Cancer, 2021, 21(8): 526-36. BATLLE E, CLEVERS H. Cancer stem cells revisited [J]. Nature medicine, 2017, 23(10): 1124-34. EUN K, HAM S W, KIM H. Cancer stem cell heterogeneity: origin and new perspectives on CSC targeting [J]. BMB reports, 2017, 50(3): 117-25. WU J, JING X, DU Q, et al. Disruption of the Clock Component Bmal1 in Mice Promotes Cancer Metastasis through the PAI-1-TGF-β-myoCAF-Dependent Mechanism [J]. Advanced science (Weinheim, Baden-Wurttemberg, Germany), 2023, 10(24): e2301505. SHIBUE T, WEINBERG R A. EMT, CSCs, and drug resistance: the mechanistic link and clinical implications [J]. Nature reviews Clinical oncology, 2017, 14(10): 611-29. LAMBERT A W, WEINBERG R A. Linking EMT programmes to normal and neoplastic epithelial stem cells [J]. Nature reviews Cancer, 2021, 21(5): 325-38. CAñELLAS-SOCIAS A, CORTINA C, HERNANDO-MOMBLONA X, et al. Metastatic recurrence in colorectal cancer arises from residual EMP1(+) cells [J]. Nature, 2022, 611(7936): 603-13. WEI C, SUN W, SHEN K, et al. Delineating the early dissemination mechanisms of acral melanoma by integrating single-cell and spatial transcriptomic analyses [J]. Nature communications, 2023, 14(1): 8119. QUAH H S, CAO E Y, SUTEJA L, et al. Single cell analysis in head and neck cancer reveals potential immune evasion mechanisms during early metastasis [J]. Nature communications, 2023, 14(1): 1680. ZHANG S, FANG W, ZHOU S, et al. Single cell transcriptomic analyses implicate an immunosuppressive tumor microenvironment in pancreatic cancer liver metastasis [J]. Nature communications, 2023, 14(1): 5123. KARIMI E, YU M W, MARITAN S M, et al. Single-cell spatial immune landscapes of primary and metastatic brain tumours [J]. Nature, 2023, 614(7948): 555-63. LIU Y, HE M, TANG H, et al. Single-cell and spatial transcriptomics reveal metastasis mechanism and microenvironment remodeling of lymph node in osteosarcoma [J]. BMC medicine, 2024, 22(1): 200. ZILIONIS R, ENGBLOM C, PFIRSCHKE C, et al. Single-Cell Transcriptomics of Human and Mouse Lung Cancers Reveals Conserved Myeloid Populations across Individuals and Species [J]. Immunity, 2019, 50(5): 1317-34.e10. PURAM S V, TIROSH I, PARIKH A S, et al. Single-Cell Transcriptomic Analysis of Primary and Metastatic Tumor Ecosystems in Head and Neck Cancer [J]. Cell, 2017, 171(7): 1611-24.e24. LAWSON D A, KESSENBROCK K, DAVIS R T, et al. Tumour heterogeneity and metastasis at single-cell resolution [J]. Nature cell biology, 2018, 20(12): 1349-60. XING H, LIU D, LI J, et al. TERTp Mutation and its Prognostic Value in Glioma Patients Under the 2021 WHO Classification: A Real-World Study [J]. Cancer medicine, 2025, 14(2): e70533. GULATI G S, SIKANDAR S S, WESCHE D J, et al. Single-cell transcriptional diversity is a hallmark of developmental potential [J]. Science (New York, NY), 2020, 367(6476): 405-11. YEN M C, WU K L, LIU Y W, et al. Ubiquitin Conjugating Enzyme E2 H (UBE2H) Is Linked to Poor Outcomes and Metastasis in Lung Adenocarcinoma [J]. Biology, 2021, 10(5). CABLE D M, MURRAY E, ZOU L S, et al. Robust decomposition of cell type mixtures in spatial transcriptomics [J]. Nature biotechnology, 2022, 40(4): 517-26. CHEN J, LIU W, LUO T, et al. A comprehensive comparison on cell-type composition inference for spatial transcriptomics data [J]. Briefings in bioinformatics, 2022, 23(4). BYERS L A, DIAO L, WANG J, et al. An epithelial-mesenchymal transition gene signature predicts resistance to EGFR and PI3K inhibitors and identifies Axl as a therapeutic target for overcoming EGFR inhibitor resistance [J]. Clinical cancer research : an official journal of the American Association for Cancer Research, 2013, 19(1): 279-90. LI R, LIU X, HUANG X, et al. Single-cell transcriptomic analysis deciphers heterogenous cancer stem-like cells in colorectal cancer and their organ-specific metastasis [J]. Gut, 2024, 73(3): 470-84. LI, X., PAN, J., LIU, T., YIN, W., MIAO, Q., ZHAO, Z., GAO, Y., ZHENG, W., LI, H., DENG, R., HUANG, D., QIU, S., ZHANG, Y., QI, Q., DENG, L., HUANG, M., TANG, P. M., CAO, Y., CHEN, M., YE, W. & ZHANG, D. 2023. Novel TCF21(high) pericyte subpopulation promotes colorectal cancer metastasis by remodelling perivascular matrix. Gut, 72 , 710-721. LI, X., SUN, Z., PENG, G., XIAO, Y., GUO, J., WU, B., LI, X., ZHOU, W., LI, J., LI, Z., BAI, C., ZHAO, L., HAN, Q., ZHAO, R. C. & WANG, X. 2022. Single-cell RNA sequencing reveals a pro-invasive cancer-associated fibroblast subgroup associated with poor clinical outcomes in patients with gastric cancer. Theranostics, 12 , 620-638. LIN, S., MA, L., MO, J., ZHAO, R., LI, J., YU, M., JIANG, M. & PENG, L. 2024. Immune cell senescence and exhaustion promote the occurrence of liver metastasis in colorectal cancer by regulating epithelial-mesenchymal transition. Aging (Albany NY), 16 , 7704-7732. LIOTTA, L. A. & KOHN, E. C. 2001. The microenvironment of the tumour-host interface. Nature, 411 , 375-9. LIU, X., QIN, J., NIE, J., GAO, R., HU, S., SUN, H., WANG, S. & PAN, Y. 2023. ANGPTL2+cancer-associated fibroblasts and SPP1+macrophages are metastasis accelerators of colorectal cancer. Front Immunol, 14 , 1185208. PETITPREZ F, DE REYNIèS A, KEUNG E Z, et al. B cells are associated with survival and immunotherapy response in sarcoma [J]. Nature, 2020, 577(7791): 556-60. SHEN T, LIU J L, WANG C Y, et al. Targeting Erbin in B cells for therapy of lung metastasis of colorectal cancer [J]. Signal transduction and targeted therapy, 2021, 6(1): 115. MA, C., YANG, C., PENG, A., SUN, T., JI, X., MI, J., WEI, L., SHEN, S. & FENG, Q. 2023. Pan-cancer spatially resolved single-cell analysis reveals the crosstalk between cancer-associated fibroblasts and tumor microenvironment. Mol Cancer, 22 , 170. AZIZI E, CARR A J, PLITAS G, et al. Single-Cell Map of Diverse Immune Phenotypes in the Breast Tumor Microenvironment [J]. Cell, 2018, 174(5): 1293-308.e36. WANG F, LONG J, LI L, et al. Single-cell and spatial transcriptome analysis reveals the cellular heterogeneity of liver metastatic colorectal cancer [J]. Science advances, 2023, 9(24): eadf5464. YIN W, WANG J, JIANG L, et al. Cancer and stem cells [J]. Experimental biology and medicine (Maywood, NJ), 2021, 246(16): 1791-801. BABAEI G, AZIZ S G, JAGHI N Z Z. EMT, cancer stem cells and autophagy; The three main axes of metastasis [J]. Biomedicine & pharmacotherapy = Biomedecine & pharmacotherapie, 2021, 133: 110909. LAMBERT A W, PATTABIRAMAN D R, WEINBERG R A. Emerging Biological Principles of Metastasis [J]. Cell, 2017, 168(4): 670-91. SHIMOKAWA M, OHTA Y, NISHIKORI S, et al. Visualization and targeting of LGR5(+) human colon cancer stem cells [J]. Nature, 2017, 545(7653): 187-92. CHENG L, LI X, DONG W, et al. LAMC2 regulates the proliferation, invasion, and metastasis of gastric cancer via PI3K/Akt signaling pathway [J]. Journal of cancer research and clinical oncology, 2024, 150(5): 230. SHEN T, LIU J L, WANG C Y, et al. Targeting Erbin in B cells for therapy of lung metastasis of colorectal cancer [J]. Signal transduction and targeted therapy, 2021, 6(1): 115. LI R, LIU X, HUANG X, et al. Single-cell transcriptomic analysis deciphers heterogenous cancer stem-like cells in colorectal cancer and their organ-specific metastasis [J]. Gut, 2024, 73(3): 470-84. LIU X, QIN J, NIE J, et al. ANGPTL2+cancer-associated fibroblasts and SPP1+macrophages are metastasis accelerators of colorectal cancer [J]. Frontiers in immunology, 2023, 14: 1185208. GUI P, BIVONA T G. Evolution of metastasis: new tools and insights [J]. Trends in cancer, 2022, 8(2): 98-109. ZHANG Z, WANG Y, ZHANG J, et al. COL1A1 promotes metastasis in colorectal cancer by regulating the WNT/PCP pathway [J]. Molecular medicine reports, 2018, 17(4): 5037-42. ZHENG H, LIU H, GE Y, et al. Integrated single-cell and bulk RNA sequencing analysis identifies a cancer associated fibroblast-related signature for predicting prognosis and therapeutic responses in colorectal cancer [J]. Cancer cell international, 2021, 21(1): 552. CHU X, TIAN W, NING J, et al. Cancer stem cells: advances in knowledge and implications for cancer therapy [J]. Signal transduction and targeted therapy, 2024, 9(1): 170. LI Y, LIN C, CHU Y, et al. Characterization of Cancer Stem Cells in Laryngeal Squamous Cell Carcinoma by Single-cell RNA Sequencing [J]. Genomics, proteomics & bioinformatics, 2024, 22(4). SHI Q, XUE C, ZENG Y, et al. Notch signaling pathway in cancer: from mechanistic insights to targeted therapies [J]. Signal transduction and targeted therapy, 2024, 9(1): 128. DAKAL T C, BHUSHAN R, XU C, et al. Intricate relationship between cancer stemness, metastasis, and drug resistance [J]. MedComm, 2024, 5(10): e710. MANTOVANI A, LOCATI M. Tumor-associated macrophages as a paradigm of macrophage plasticity, diversity, and polarization: lessons and open questions [J]. Arteriosclerosis, thrombosis, and vascular biology, 2013, 33(7): 1478-83. QIAN B Z, POLLARD J W. Macrophage diversity enhances tumor progression and metastasis [J]. Cell, 2010, 141(1): 39-51. RUFFELL B, COUSSENS L M. Macrophages and therapeutic resistance in cancer [J]. Cancer cell, 2015, 27(4): 462-72. Additional Declarations There is NO Competing Interest. Supplementary Files S1.jpg Suppl. Figure 1 S2.jpg Suppl. Figure 2 S3.jpg Suppl. Figure 3 S4.jpg Suppl. Figure 4 S5.jpg Suppl. Figure 5 S6.jpg Suppl. Figure 6 S7.jpg Suppl. Figure 7 Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6803290","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":466572639,"identity":"7eb4968e-4f0f-43da-bdc9-fbc1a8e0907d","order_by":0,"name":"Xinxing Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1klEQVRIiWNgGAWjYDACCQglx8bM//BBQkUNMVqYwZQxPzsPs8GDM8eI15I4s5+HTfJhCzNhHfyz+49J87YdZtxwmPdYRWIDGwN/e3cCfkvuHGY2nHEmjdngMF/ajcQdMgwSZ85uwG/NjWTGBx8qbNgMDjOY3Ug8w8ZgIJGLX4v8jWSGAwkGEjwgLQWJbcyEtRhAbZGQbOYxYyBKi+GNZGOQXwz4mdmSJRLOHOMh6Be5G4nPQCFW38Z/+ODHHxU1cvztvQS8jw54SFM+CkbBKBgFowArAACEJEX60H6rLgAAAABJRU5ErkJggg==","orcid":"","institution":"Department of General Surgery, Tongji Hospital, School of Medicine, Tongji University","correspondingAuthor":true,"prefix":"","firstName":"Xinxing","middleName":"","lastName":"Li","suffix":""},{"id":466572640,"identity":"8ffbcdaa-f351-4d9e-822d-0997e33dc48a","order_by":1,"name":"Jinran Wu","email":"","orcid":"","institution":"Research Center of Clinical Medicine, Affiliated Hospital of Nantong University, Medical School of Nantong University","correspondingAuthor":false,"prefix":"","firstName":"Jinran","middleName":"","lastName":"Wu","suffix":""},{"id":466572641,"identity":"a935d5f9-6d88-443e-9177-3aff80a27fb2","order_by":2,"name":"Xiaomao Yin","email":"","orcid":"https://orcid.org/0000-0003-1045-659X","institution":"Department of General Surgery, Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Xiaomao","middleName":"","lastName":"Yin","suffix":""},{"id":466572642,"identity":"9c3c26d4-3738-4946-863e-7c00b317bdb5","order_by":3,"name":"Jiahui Yin","email":"","orcid":"","institution":"Information Institute of Science and Technology, Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Jiahui","middleName":"","lastName":"Yin","suffix":""},{"id":466572643,"identity":"691bec65-0f7a-4e0d-ad48-0abb63e21606","order_by":4,"name":"Yan Wang","email":"","orcid":"","institution":"Shanghai Key Laboratory of Multidimensional Information Processing, East China Normal University","correspondingAuthor":false,"prefix":"","firstName":"Yan","middleName":"","lastName":"Wang","suffix":""},{"id":466572644,"identity":"9b1ba103-2cde-45c1-ae7d-d3be519951d2","order_by":5,"name":"Jiexuan Wang","email":"","orcid":"","institution":"Department of General Surgery, Tongji Hospital, School of Medicine, Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Jiexuan","middleName":"","lastName":"Wang","suffix":""},{"id":466572645,"identity":"f4e884e5-ee36-4ecf-8931-12968fca571f","order_by":6,"name":"Lin Zhu","email":"","orcid":"","institution":"Department of General Surgery,Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Lin","middleName":"","lastName":"Zhu","suffix":""},{"id":466572646,"identity":"f7db20d4-781c-4efc-af67-e126969e74f3","order_by":7,"name":"Xin Yang","email":"","orcid":"","institution":"Department of General Surgery,Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Yang","suffix":""},{"id":466572647,"identity":"b739339d-534f-4f5e-8930-1a8bdf8b11f4","order_by":8,"name":"Jingjing Qian","email":"","orcid":"","institution":"Department of General Surgery,Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Jingjing","middleName":"","lastName":"Qian","suffix":""},{"id":466572648,"identity":"0e13f7f0-0254-428b-abb1-e8e1b0bbf9d9","order_by":9,"name":"Wenqiang Wang","email":"","orcid":"","institution":"Department of General Surgery,Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Wenqiang","middleName":"","lastName":"Wang","suffix":""},{"id":466572649,"identity":"c8defe96-5e2c-4373-a339-afe54395bc80","order_by":10,"name":"Kai Xu","email":"","orcid":"","institution":"Department of General Surgery,Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Kai","middleName":"","lastName":"Xu","suffix":""},{"id":466572650,"identity":"732f8255-22f2-413e-9429-6834bf05e8d4","order_by":11,"name":"Liangchen Zhu","email":"","orcid":"","institution":"Department of General Surgery,Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Liangchen","middleName":"","lastName":"Zhu","suffix":""},{"id":466572651,"identity":"bb57f1ec-45df-422a-b873-5fd083da87e5","order_by":12,"name":"Zhiqian Hu","email":"","orcid":"","institution":"Department of General Surgery, Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Zhiqian","middleName":"","lastName":"Hu","suffix":""},{"id":466572652,"identity":"6b55ce8f-84df-4a84-ac19-49923907878a","order_by":13,"name":"Xuan Dai","email":"","orcid":"","institution":"Department of General Surgery, Tongji Hospital, School of Medicine, Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Xuan","middleName":"","lastName":"Dai","suffix":""},{"id":466572653,"identity":"0a231caf-6f6e-4a5c-b977-d073c2ad551a","order_by":14,"name":"Siyuan Yin","email":"","orcid":"","institution":"School of Public Health, Nantong University, Nantong","correspondingAuthor":false,"prefix":"","firstName":"Siyuan","middleName":"","lastName":"Yin","suffix":""},{"id":466572654,"identity":"c7215965-1db6-45d0-a550-34bd31880c71","order_by":15,"name":"Zekun Zhao","email":"","orcid":"","institution":"Department of General Surgery,Tongji Hospital, Medical College of Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Zekun","middleName":"","lastName":"Zhao","suffix":""},{"id":466572655,"identity":"37bd93a8-d1d5-4501-8a16-2d8e0f6b5488","order_by":16,"name":"Runqi Hong","email":"","orcid":"","institution":"Department of General Surgery, Tongji Hospital, School of Medicine, Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Runqi","middleName":"","lastName":"Hong","suffix":""},{"id":466572656,"identity":"321fe1c0-7ca6-4171-b95b-8cf67d48d3a0","order_by":17,"name":"Dongsheng Li","email":"","orcid":"","institution":"Department of General Surgery, Tongji Hospital, School of Medicine, Tongji University","correspondingAuthor":false,"prefix":"","firstName":"Dongsheng","middleName":"","lastName":"Li","suffix":""},{"id":466572657,"identity":"d8d14852-0486-429f-8f84-3542d6cc7b1c","order_by":18,"name":"Io Wong","email":"","orcid":"","institution":"Faculty of Medicine, Macau University of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"Io","middleName":"","lastName":"Wong","suffix":""},{"id":466572658,"identity":"174afd9b-15e8-4509-81da-171ca05a5954","order_by":19,"name":"Feng Wang","email":"","orcid":"","institution":"Research Center of Clinical Medicine, Affiliated Hospital of Nantong University, Medical School of Nantong University, Nantong","correspondingAuthor":false,"prefix":"","firstName":"Feng","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2025-06-02 14:47:57","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6803290/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6803290/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":84003694,"identity":"34e6bf32-e96b-4901-8576-4676c58a17bc","added_by":"auto","created_at":"2025-06-05 15:09:45","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":7897246,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"Fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/6ba97ec3a69ffa46708bf8bc.jpg"},{"id":84003693,"identity":"626a7eb0-f11f-4417-99a4-9d267d13396b","added_by":"auto","created_at":"2025-06-05 15:09:45","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":3981994,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"Fig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/c5c39c7bc8412459b07928ee.jpg"},{"id":84003695,"identity":"fe4db8a0-0cc3-497e-a33d-8707d3640955","added_by":"auto","created_at":"2025-06-05 15:09:45","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":11485010,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"Fig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/00c9497b45bed1f767928ccc.jpg"},{"id":84004218,"identity":"4f081164-97ab-40cf-8ba4-73a8fe55c2d9","added_by":"auto","created_at":"2025-06-05 15:17:45","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":20135806,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"Fig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/c972b9d5396c55ee1ff45517.jpg"},{"id":84005164,"identity":"a1104176-f29a-4d5b-8864-c085e20fd84e","added_by":"auto","created_at":"2025-06-05 15:25:45","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":10891313,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"Fig5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/40908746f05e4d81849851e7.jpg"},{"id":84005672,"identity":"204c0034-5435-4dce-b2dd-6cd928f1ccf5","added_by":"auto","created_at":"2025-06-05 15:33:45","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":10805763,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"Fig6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/1d6c6ed1c67c6057ac416721.jpg"},{"id":84004221,"identity":"b8f2eec6-c51a-42a1-967a-96a23f64c534","added_by":"auto","created_at":"2025-06-05 15:17:46","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":7015585,"visible":true,"origin":"","legend":"\u003cp\u003eLegend not included with this version.\u003c/p\u003e","description":"","filename":"Fig7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/be29a66dbe543f48ec9bc3ad.jpg"},{"id":84006996,"identity":"1ba75e85-7872-4ae1-817e-cf62a7c7337e","added_by":"auto","created_at":"2025-06-05 15:42:08","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":73603052,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/4fa651aa-5a4d-43f0-9392-5e31e81b0767.pdf"},{"id":84003704,"identity":"eb734068-1117-4da0-bad7-0ce7c188547a","added_by":"auto","created_at":"2025-06-05 15:09:46","extension":"jpg","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":8235628,"visible":true,"origin":"","legend":"Suppl. Figure 1","description":"","filename":"S1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/26259df62aefaa80f2960475.jpg"},{"id":84003706,"identity":"d488e273-c048-44d8-8480-8efbdb8fe18f","added_by":"auto","created_at":"2025-06-05 15:09:46","extension":"jpg","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":8300583,"visible":true,"origin":"","legend":"Suppl. Figure 2","description":"","filename":"S2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/c7f8a63b2985bc5e49dab353.jpg"},{"id":84003702,"identity":"954b0f5c-9e1a-49e4-acd8-9d8149f845ec","added_by":"auto","created_at":"2025-06-05 15:09:46","extension":"jpg","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":13658113,"visible":true,"origin":"","legend":"Suppl. Figure 3","description":"","filename":"S3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/83130cf090d89611befb0343.jpg"},{"id":84003703,"identity":"ed365dbb-b22d-43e7-b74d-d5f25da8862d","added_by":"auto","created_at":"2025-06-05 15:09:46","extension":"jpg","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":11268433,"visible":true,"origin":"","legend":"Suppl. Figure 4","description":"","filename":"S4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/e15995bccb18b4380fd32e0f.jpg"},{"id":84003705,"identity":"60f815fb-96cd-4fbd-99df-afc7ccf91cd7","added_by":"auto","created_at":"2025-06-05 15:09:46","extension":"jpg","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":14307523,"visible":true,"origin":"","legend":"Suppl. Figure 5","description":"","filename":"S5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/5f9f688c04c3aa9e6fb61174.jpg"},{"id":84003701,"identity":"c561b1bb-4bb3-459b-b35c-3fef6ddd69d9","added_by":"auto","created_at":"2025-06-05 15:09:46","extension":"jpg","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":9008193,"visible":true,"origin":"","legend":"Suppl. Figure 6","description":"","filename":"S6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/7e75cd9ee25133b74378fdf4.jpg"},{"id":84003699,"identity":"84188b3f-03eb-4849-94a1-60bf68fe893b","added_by":"auto","created_at":"2025-06-05 15:09:46","extension":"jpg","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":8956222,"visible":true,"origin":"","legend":"\u003cp\u003eSuppl. Figure 7\u003c/p\u003e","description":"","filename":"S7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6803290/v1/f0b4962566768f1a76d07a6e.jpg"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Single-cell and Spatial Transcriptomic Mapping Reveals MSEC-myoCAF Interactions Driving Lymph Node Metastasis in Colorectal Cancer","fulltext":[{"header":"Introduction","content":"\u003cp\u003eColorectal cancer (CRC) remains one of the leading causes of cancer-related mortality worldwide, with lymph node metastasis (LNM) representing a pivotal prognostic marker and a major therapeutic obstacle\u003cstrong\u003e\u003csup\u003e[1]\u003c/sup\u003e\u003c/strong\u003e. Despite notable progress in targeted therapies and precision medicine, the 5-year survival rate for patients with stage III CRC has stagnated around 65%\u003cstrong\u003e\u003csup\u003e[2]\u003c/sup\u003e\u003c/strong\u003e, underscoring the urgent need to elucidate the molecular and cellular mechanisms that underpin metastatic progression. Metastasis is a complex, multistep cascade encompassing tumor cell invasion, intravasation, survival in circulation, extravasation, and eventual colonization at distant sites. In CRC, lymph nodes are typically the first destination of metastatic spread\u003cstrong\u003e\u003csup\u003e[3]\u003c/sup\u003e\u003c/strong\u003e. A deeper mechanistic understanding of LNM is crucial for the development of effective interventions that can limit dissemination and improve clinical outcomes.\u003c/p\u003e\n\u003cp\u003eRecent studies have drawn attention to the critical roles of tumor cell stemness and stromal remodeling in CRC metastasis\u003cstrong\u003e\u003csup\u003e[4]\u003c/sup\u003e\u003c/strong\u003e. Cancer stem cells (CSCs)\u0026mdash;a specialized subpopulation endowed with self-renewal capability and phenotypic plasticity\u0026mdash;are proposed to initiate metastasis by generating heterogeneous progenies adapted to diverse microenvironments\u003cstrong\u003e\u003csup\u003e[5]\u003c/sup\u003e\u003c/strong\u003e. These cells are notoriously resistant to conventional therapies and are believed to be key drivers of tumor recurrence and distant metastasis\u003cstrong\u003e\u003csup\u003e[6-7]\u003c/sup\u003e\u003c/strong\u003e. However, despite their functional relevance, the spatial dynamics of CSCs and their interactions with the surrounding stromal components within the tumor microenvironment (TME) remain poorly characterized, hampering the development of CSC-targeted therapies.\u003c/p\u003e\n\u003cp\u003eThe TME is composed of a heterogeneous array of cell types, including immune cells, endothelial cells, and cancer-associated fibroblasts (CAFs), all of which modulate tumor behavior and contribute to metastasis. Among them, myofibroblast-like CAFs (myoCAFs) have emerged as key mediators of metastatic niche formation\u003cstrong\u003e\u003csup\u003e[8]\u003c/sup\u003e\u003c/strong\u003e. These cells actively secrete extracellular matrix (ECM) proteins, cytokines, and growth factors that facilitate epithelial\u0026ndash;mesenchymal transition (EMT)\u0026mdash;a fundamental biological process that enhances epithelial cell motility and invasiveness\u003cstrong\u003e\u003csup\u003e[9-10]\u003c/sup\u003e\u003c/strong\u003e. Despite this growing understanding, it remains unclear whether specific stem-like epithelial populations cooperate spatially with CAFs to drive LNM in CRC. Elucidating the spatial architecture and functional interactions within the TME is essential for discovering novel mechanisms of metastasis and identifying actionable therapeutic targets.\u003c/p\u003e\n\u003cp\u003eThe advent of spatial multi-omics technologies has enabled unprecedented resolution in mapping cell\u0026ndash;cell communication and tissue architecture within intact tumor specimens\u003cstrong\u003e\u003csup\u003e[11-16]\u003c/sup\u003e\u003c/strong\u003e. These platforms allow for simultaneous analysis of gene expression, protein localization, and cellular spatial proximity, providing a holistic view of the TME and its role in cancer progression\u003cstrong\u003e\u003csup\u003e[17]\u003c/sup\u003e\u003c/strong\u003e. Leveraging these advances, we hypothesized that metastatic progression in CRC is orchestrated by stem-like epithelial cells that interact with adjacent myoCAFs through spatially organized signaling pathways. This hypothesis is supported by frequent co-localization of CSCs and CAFs at the invasive margins of tumors, suggesting potential cooperative interactions.\u003c/p\u003e\n\u003cp\u003eTo rigorously test this hypothesis, we conducted an integrative study combining single-cell RNA sequencing (scRNA-seq), spatial transcriptomics (ST), and functional validation assays. Our goal was to identify metastatic epithelial subpopulations and delineate their spatial and molecular crosstalk with myoCAFs in the context of lymph node metastasis. This multi-modal approach enabled us to map high-resolution cellular landscapes while preserving spatial context, thereby uncovering critical pathways that govern CRC dissemination. Ultimately, our findings provide novel insights into the metastatic trajectory of CRC, particularly its spread to mesenteric lymph nodes, and highlight promising therapeutic targets for early intervention in metastatic disease.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eA single-cell transcriptomic atlas of primary and metastatic CRC lesions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo comprehensively characterize the cellular landscape of primary colorectal tumors and matched metastatic lymph nodes we performed scRNA-sequsing the 10x Genomics platform on ten tissue samples obtained from five CRC patients with confirmed lymph node metastasis. Each patient contributed a pair of samples: one from the primary tumor and one from the corresponding metastatic lymph node. The experimental workflow and detailed clinical-pathological features of the cohort are summarized in \u003cstrong\u003eFigure 1A\u0026nbsp;\u003c/strong\u003eand \u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eFollowing data preprocessing—including rigorous quality control and batch effect correction—we integrated the transcriptomes across all samples and conducted dimensionality reduction using UMAP. This analysis resolved 20 transcriptionally distinct clusters across the entire dataset, visualized in two-dimensional space (\u003cstrong\u003eFigure 1B, Figure S2A\u003c/strong\u003e). In total, 84,735 high-quality single cells were retained for downstream analyses. Based on canonical lineage markers, we annotated the clusters into 10 major cell types, comprising both non-immune and immune populations. The four non-immune lineages included epithelial cells (EPCAM, CDH1, KRT18), fibroblasts (COL1A1, COL1A2, COL3A1), endothelial cells (PECAM1, CDH5, VWF), and smooth muscle cells (ACTA2, TAGLN, MYH1). The six immune lineages were represented by B cells (CD19, CD79A, CD79B), T cells (CD3G, CD3D, NKG7), neutrophils (CSF3R, S100A8, S100A9), mast cells (CPA3, MS4A2, KIT), monocytes (CD14, CD300E, VCAN), and macrophages (CD68, C1QA, C1QB) (\u003cstrong\u003eFigure 1C, 1F, and Figure S1C, D\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eWe next quantified the relative abundance of these major cell types across primary tumor and lymph node tissues, revealing distinct compositional differences between the two anatomical sites (\u003cstrong\u003eFigure 1E\u003c/strong\u003e). This high-resolution atlas provides a foundational framework for dissecting the cellular programs and microenvironmental remodeling events associated with CRC metastasis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMetastatic stem-like epithelial cells as potential initiators of CRC dissemination\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAs CRC originates from epithelial cells, we performed high-resolution UMAP-based clustering on epithelial subsets, identifying nine transcriptionally distinct subclusters (\u003cstrong\u003eFigure 2A\u003c/strong\u003e). Marker gene expression patterns for each cluster are shown in \u003cstrong\u003eFigure 2B\u003c/strong\u003e. Quantitative analysis across sample origins revealed that cluster C6 was predominantly enriched in primary tumor samples, whereas clusters C3 and C4 were significantly enriched in both primary tumors and corresponding metastatic lymph nodes, suggesting a potential role in metastatic progression (\u003cstrong\u003eFigures 2C, S2A\u003c/strong\u003e, and \u003cstrong\u003eS3B–C\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eTo further characterize the biological nature of these subpopulations, we applied copy number variation (CNV) analysis, a commonly used method in scRNA-seq to infer malignant transformation and track tumor evolution\u003cstrong\u003e\u003csup\u003e[18-19]\u003c/sup\u003e\u003c/strong\u003e. Among all epithelial clusters, C6 exhibited the lowest CNV burden, consistent with normal colonic epithelial identity, while C3 and C4 displayed markedly elevated CNV levels, supporting their classification as malignant cell populations\u003cstrong\u003e\u003csup\u003e[20]\u0026nbsp;\u003c/sup\u003e\u003c/strong\u003e(\u003cstrong\u003eFigure 2D\u003c/strong\u003e and \u003cstrong\u003eFigure S3D\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eWe next evaluated the differentiation potential of epithelial clusters using CytoTRACE\u003cstrong\u003e\u003csup\u003e[21]\u003c/sup\u003e\u003c/strong\u003e, which revealed that clusters C3 and C4 harbored the highest stemness scores and least differentiation, indicating a poorly differentiated, stem-like state (\u003cstrong\u003eFigure 2E\u003c/strong\u003e and \u003cstrong\u003eFigure S2D\u003c/strong\u003e). These putative metastatic stem-like epithelial cells (MSECs) also exhibited elevated EMT and proliferation scores (\u003cstrong\u003eFigure S2C, E\u003c/strong\u003e), reinforcing their aggressive phenotype. Pseudotime trajectory analysis placed MSECs at the apex of the differentiation hierarchy, suggesting they give rise to other epithelial lineages within the tumor (\u003cstrong\u003eFigure 2F\u003c/strong\u003e). RNA velocity analysis further demonstrated directional flow of transcriptomic states originating from MSECs, supporting their role as progenitor cells (\u003cstrong\u003eFigure 2G\u003c/strong\u003e). Differential gene expression analysis identified UBE2H and LAMC2 as specifically upregulated in MSECs, alongside canonical stemness-associated markers such as LGR5, AXIN2, SOX9, CD44, and ALCAM (\u003cstrong\u003eFigure 2H\u003c/strong\u003e). Notably, UBE2H and LAMC2 expression levels were strongly correlated with stemness metrics across the epithelial compartment. Pathway enrichment analysis revealed that WNT, EMT, and partial EMT (pEMT) signaling pathways were significantly activated in MSECs (\u003cstrong\u003eFigure 2I\u003c/strong\u003e and \u003cstrong\u003eFigure S2B\u003c/strong\u003e), further linking them to metastatic potential.\u003c/p\u003e\n\u003cp\u003eTaken together, these findings define MSECs as a transcriptionally and functionally distinct epithelial subpopulation enriched in metastatic lesions, endowed with high stemness, EMT capacity, and tumor-initiating features. These results suggest that MSECs may serve as the cellular origin of CRC metastasis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOverexpression of LAMC2 and UBE2H confers a metastatic phenotype in CRC\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUBE2H (ubiquitin-conjugating enzyme E2 H) belongs to the E2 family of ubiquitin-conjugating enzymes and plays a pivotal role in regulating protein ubiquitination, a process essential for controlling cellular adhesion, migration, and invasion—key hallmarks of metastatic progression\u003cstrong\u003e\u003csup\u003e[22]\u003c/sup\u003e\u003c/strong\u003e. Within the MSECs, we observed a strong positive correlation between UBE2H and LAMC2 expression (\u003cstrong\u003eFigure 3A\u003c/strong\u003e, \u003cstrong\u003eFigure S4A\u003c/strong\u003e). Spatial transcriptomics of primary CRC lesions further revealed a distinct epithelial subpopulation with high co-expression of both genes, predominantly localized at the invasive front (\u003cstrong\u003eFigure 3B\u003c/strong\u003e, \u003cstrong\u003eFigure S4D\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eTo evaluate their functional relevance, we silenced UBE2H and LAMC2 expression using small interfering RNA (siRNA) in two CRC cell lines, SW1116 and SW480. Knockdown of either gene significantly impaired CRC cell proliferation as measured by in vitro assays (\u003cstrong\u003eFigure 3C\u003c/strong\u003e, \u003cstrong\u003eFigure S4E\u003c/strong\u003e). To validate these findings in clinical specimens, immunohistochemistry (IHC) was performed on CRC tissues with and without lymph node metastasis. Both UBE2H and LAMC2 showed elevated protein expression in lymph node-positive tumors (\u003cstrong\u003eFigure 3D\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eFurther in vitro functional assays demonstrated that knockdown of either UBE2H or LAMC2 led to a marked reduction in cell migration and invasion capabilities (\u003cstrong\u003eFigure 3E\u003c/strong\u003e, \u003cstrong\u003eFigure S4F\u003c/strong\u003e). Mechanistically, gene silencing resulted in increased expression of the epithelial marker E-cadherin, and decreased levels of mesenchymal markers N-cadherin, Vimentin, and the EMT-inducing transcription factor Snail, indicating reversal of the EMT phenotype (\u003cstrong\u003eFigure 3F\u003c/strong\u003e, \u003cstrong\u003eFigure S4G\u003c/strong\u003e). Consistently, pathway enrichment analysis of our scRNA-seq data confirmed a strong positive association between LAMC2, UBE2H, and EMT-related signaling pathways (\u003cstrong\u003eFigure 3G\u003c/strong\u003e, \u003cstrong\u003eFigure S4B\u003c/strong\u003e). Kaplan–Meier survival analysis of CRC patient cohorts revealed that elevated expression of UBE2H and LAMC2 was significantly associated with reduced disease-free survival (DFS) (\u003cstrong\u003eFigure 3H\u003c/strong\u003e, \u003cstrong\u003eFigure S4C\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eCollectively, these results establish UBE2H and LAMC2 as key mediators of the metastatic phenotype in CRC. Their co-expression within MSECs not only defines a transcriptional state of high stemness and invasiveness but also contributes functionally to enhanced tumor aggressiveness, suggesting their potential as biomarkers and therapeutic targets in metastatic CRC.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSpatial transcriptomics reveals localization and interactions of metastatic stem-like cells\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSpatial context is essential for decoding tumor architecture and cellular interactions, yet conventional scRNA-seq lacks positional information. To overcome this limitation, we performed spatial transcriptomics (ST-seq) on five CRC samples with confirmed lymph node metastasis (LN+) to map in situ gene expression patterns and spatial organization. Based on hematoxylin and eosin (H\u0026amp;E) staining, spatial spots were annotated into tumor, stromal, and epithelial regions (\u003cstrong\u003eFigure 4A\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eCNV analysis revealed the highest CNV burden in tumor-designated regions, consistent with their malignant identity (\u003cstrong\u003eFigure S5A, B\u003c/strong\u003e). To determine cellular composition within these regions, we applied Robust Cell Type Decomposition (RCTD)\u003cstrong\u003e\u003csup\u003e[23-24]\u003c/sup\u003e\u003c/strong\u003e, a reference-based deconvolution algorithm. As expected, stromal regions were enriched with endothelial cells, while epithelial and tumor regions were dominated by epithelial cells. Notably, fibroblasts were highly concentrated along the tumor–stroma boundary, suggesting a spatially defined fibroblast population at the invasive front (\u003cstrong\u003eFigure 4A\u003c/strong\u003e), potentially shaping the local microenvironment. Across five CRC tissue sections, we obtained transcriptomic data from 20,859 spatial barcodes (\u003cstrong\u003eFigure 4B\u003c/strong\u003e) and evaluated EMT scores at tumor margins\u003cstrong\u003e\u003csup\u003e[25]\u003c/sup\u003e\u003c/strong\u003e. EMT scoring revealed significantly higher EMT activity at the tumor front compared to the tumor core (\u003cstrong\u003eFigure 4C\u003c/strong\u003e), supporting the notion that MSECs preferentially localize at the invasive edge—a finding that corroborates our scRNA-seq results (\u003cstrong\u003eFigure 4D, Figure S5D, E\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eTo explore intercellular interactions within this spatial context, we conducted ligand-receptor interaction analysis and identified significant enrichment of the COL1A1–SDC4 axis in fibroblas-epithelial communication (\u003cstrong\u003eFigure 4E\u003c/strong\u003e). This interaction was predominantly observed at the tumor–stroma interface, where MSECs and myofibroblast-like CAFs (myoCAFs) co-localize. Notably, both COL1A1 and SDC4 have previously been implicated in promoting colorectal cancer metastasis by facilitating tumor cell adhesion, motility, invasion, and maintenance of stem-like traits\u003cstrong\u003e\u003csup\u003e[26-27]\u003c/sup\u003e\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003eIn summary, ST-seq analysis provided critical spatial insight into the localization of MSECs within the tumor microenvironment. These cells are enriched at the tumor front, where they interact with surrounding myoCAFs through pro-metastatic signaling pathways such as COL1A1–SDC4, reinforcing their role in metastatic progression.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEnhanced interactions between myoCAFs and metastatic stem-like cells promote lymph node metastasis in CRC\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCancer metastasis is not solely driven by tumor-intrinsic mechanisms but is profoundly influenced by stromal components within the TME, particularly CAFs\u003cstrong\u003e\u003csup\u003e[28-29]\u003c/sup\u003e\u003c/strong\u003e. Recent advances in single-cell RNA sequencing have enabled high-resolution characterization of CAF heterogeneity across multiple malignancies\u003cstrong\u003e\u003csup\u003e[30-31]\u003c/sup\u003e\u003c/strong\u003e. In our dataset, fibroblasts were subdivided into five transcriptionally distinct populations based on specific marker gene expression, and their proportional distributions were analyzed across sample types (\u003cstrong\u003eFigure 5A-C\u003c/strong\u003e). Further classification revealed two dominant CAF phenotypes: epithelial-like CAFs (eCAFs) and myoCAFs (\u003cstrong\u003eFigure 5D, E\u003c/strong\u003e), the latter being associated with pro-metastatic functions.\u003c/p\u003e\n\u003cp\u003eTo elucidate the molecular crosstalk between CAFs and MSECs, we performed ligand-receptor interaction analysis. The COL1A1-SDC4 ligand-receptor pair emerged as a dominant signaling axis between myoCAFs and MSECs, consistent with our findings from spatial transcriptomic analysis (\u003cstrong\u003eFigure 5G\u003c/strong\u003e). Notably, COL1A1-SDC4-mediated signaling was significantly more pronounced in myoCAF-MSEC interactions compared to those involving other tumor epithelial populations \u003cstrong\u003e(Figure S6G\u003c/strong\u003e). In addition, we observed that MSECs exhibited the highest frequency and strength of fibroblast-mediated intercellular interactions. Among these, NOTCH signaling—an established pathway implicated in tumor progression and stemness—was markedly activated within MSECs engaged by myoCAFs (\u003cstrong\u003eFigure 5F, I and Figure S6E\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eBased on these findings, we hypothesized that MSECs and myoCAFs are spatially co-localized within the tumor niche. Using RCTD-based spatial mapping, we observed partial but significant co-localization of these two cell populations in the primary tumor (\u003cstrong\u003eFigure 5H, Figure S6F\u003c/strong\u003e). This spatial association was further validated through multiplex immunohistochemistry (mIHC), which demonstrated significantly elevated COL1A1 and SDC4 co-expression in lymph node metastasis-positive (Tumor/LN\u003csup\u003e+\u003c/sup\u003e) samples compared to non-metastatic (Tumor/LN\u003csup\u003e-\u003c/sup\u003e) counterparts (\u003cstrong\u003eFigure 5J\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eCollectively, these findings reveal an intensive, spatially organized interaction between MSECs and myoCAFs, primarily mediated through the COL1A1-SDC4 axis and associated with NOTCH activation, supporting a cooperative role in facilitating CRC lymph node metastasis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDistinct fibroblast states determined by spatial proximity to tumor epithelial cells\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCAFs are pivotal components of the solid tumor microenvironment, known to modulate tumor growth, invasion, and metastasis\u003cstrong\u003e\u003csup\u003e[32-33].\u003c/sup\u003e\u003c/strong\u003e Given their dynamic interactions with tumor epithelial cells and their impact on clinical outcomes, we further dissected the molecular features of fibroblasts in CRC based on spatial localization relative to the tumor epithelium. Using spatial transcriptomics (ST) data, we mapped tumor epithelial regions and quantified the proximity of fibroblasts to epithelial compartments to assess their spatial organization (\u003cstrong\u003eFigure 6A, Figure S7A\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eWe observed that fibroblast localization significantly influenced the composition of neighboring cells within theTME\u003cstrong\u003e\u003csup\u003e[34]\u003c/sup\u003e\u003c/strong\u003e. Specifically, immune cell density—particularly B cells and T cells—increased with greater fibroblast distance from the tumor epithelium, whereas tumor cell density was highest in proximity to fibroblasts and declined with increasing distance (\u003cstrong\u003eFigure 6B, Figure S7B\u003c/strong\u003e), suggesting spatially dependent immune modulation. Fibroblasts were subsequently stratified into two categories based on their median spatial distance from the tumor epithelium: epithelial-adjacent fibroblasts and epithelial-distant fibroblasts. Gene expression profiling revealed that epithelial-distant fibroblasts expressed markers such as ZG16, OLFM4, CA2, PLA2G2A, DUOX2, IGHA1, PPBP, GUCA2A, FABP1, and CEACAM7, predominantly enriched in fibroblast-rich peripheral zones of the TME (\u003cstrong\u003eFigure 6C\u003c/strong\u003e). In contrast, epithelial-adjacent fibroblasts showed elevated expression of FN1, COL5A1, IGHG1, AEBP1, COL1A2, THBS2, COL1A1, COL3A1, MMP2, and C3, localizing mainly to central tumor regions with dense fibroblast infiltration (\u003cstrong\u003eFigure 6D, Figure S7C\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eTo further explore the functional implications of this spatial dichotomy, we computed enrichment scores for epithelial-distant and epithelial-adjacent fibroblast signatures within our scRNA-seq dataset. Projection of these scores revealed clustering patterns that were highly consistent with our prior CAF subtyping (\u003cstrong\u003eFigure 6E\u003c/strong\u003e). Notably, epithelial-adjacent fibroblasts closely aligned with myoCAF populations, while epithelial-distant fibroblasts were more closely associated with eCAFs.\u003c/p\u003e\n\u003cp\u003eThese results indicate that fibroblast populations exhibit distinct spatially dependent states, which correspond to divergent molecular programs and functional phenotypes. Proximity to tumor epithelial cells appears to dictate CAF identity and behavior, potentially influencing tumor-stroma crosstalk, immune cell exclusion, and metastatic progression in CRC.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMacrophages exhibit immunosuppressive and tumor-promoting phenotypes in CRC\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe identified a total of 3,021 macrophages across the single-cell RNA sequencing dataset and further classified them into seven transcriptionally distinct subpopulations (\u003cstrong\u003eFigure 7A\u003c/strong\u003e). Comparative analysis revealed that primary tumors harbored a higher abundance of DCSTAMP\u003csup\u003e+\u003c/sup\u003e, LYPD2\u003csup\u003e+\u003c/sup\u003e, and CCDC141\u003csup\u003e+\u003c/sup\u003e macrophages, whereas CCL20\u003csup\u003e+\u003c/sup\u003e macrophages were relatively less frequent in the primary site compared to metastatic lesions (\u003cstrong\u003eFigure 7B\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eTo delineate macrophage differentiation trajectories, we conducted RNA velocity analysis. This revealed that BCL11B\u003csup\u003e+\u003c/sup\u003e macrophages likely serve as the progenitor population, giving rise to two major branches: one differentiating toward CCL20\u003csup\u003e+\u003c/sup\u003e and P2RY6\u003csup\u003e+\u003c/sup\u003e macrophages, and the other toward CCDC141\u003csup\u003e+\u003c/sup\u003e and LYPD2\u003csup\u003e+\u003c/sup\u003e subsets (\u003cstrong\u003eFigure 7C\u003c/strong\u003e). This trajectory was further validated by pseudotime analysis using Monocle2, which consistently positioned BCL11B\u003csup\u003e+\u003c/sup\u003e macrophages at the root of the lineage tree, generating multiple transcriptionally distinct endpoints (\u003cstrong\u003eFigure 7D\u003c/strong\u003e). Gene expression profiling demonstrated that BCL11B\u003csup\u003e+\u0026nbsp;\u003c/sup\u003emacrophages exhibited low expression of inflammation- and immunity-associated genes, indicative of an immature or quiescent state. In contrast, CCL20\u003csup\u003e+\u003c/sup\u003e, P2RY6\u003csup\u003e+\u003c/sup\u003e, and CCDC141\u003csup\u003e+\u003c/sup\u003e macrophages showed robust upregulation of pro-inflammatory and immune-modulatory genes, albeit with distinct transcriptional programs, suggesting divergent functional roles within the tumor microenvironment (\u003cstrong\u003eFigure 7E\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eInterestingly, all macrophage subsets expressed elevated levels of immune checkpoint molecules. Specifically, CCL20\u003csup\u003e+\u003c/sup\u003e macrophages upregulated TIGIT and CTLA4, while P2RY6\u003csup\u003e+\u0026nbsp;\u003c/sup\u003ecells showed high expression of CD274 (PD-L1) and VSIR (VISTA). LYPD2\u003csup\u003e+\u003c/sup\u003e and CCDC141\u003csup\u003e+\u0026nbsp;\u003c/sup\u003esubsets displayed the highest levels of multiple checkpoint molecules, consistent with their terminal differentiation phenotypes. These data indicate that macrophages may contribute to immune evasion through checkpoint-mediated suppression.\u003c/p\u003e\n\u003cp\u003eTo further assess functional polarization, we computed M1/M2 macrophage polarization scores\u003cstrong\u003e\u003csup\u003e[35]\u003c/sup\u003e\u003c/strong\u003e. The majority of macrophages displayed an M2-like immunosuppressive phenotype, with the exception of BCL11B\u003csup\u003e+\u003c/sup\u003e macrophages, which showed low scores for both M1 and M2 signatures (Figure 7F-G). This suggests that most macrophage subsets in the CRC microenvironment adopt a pro-tumorigenic, immune-suppressive state, potentially contributing to metastatic progression and therapeutic resistance.\u003c/p\u003e\n\n\n\n\n\n\n\n\n"},{"header":"Discussion","content":"\u003cp\u003eIntratumoral heterogeneity in CRC plays a fundamental role in shaping tumor progression and metastatic potential. While scRNA-seq has offered valuable insights into the cellular diversity and transcriptomic states of primary CRC and liver metastases\u003cstrong\u003e\u003csup\u003e[36]\u003c/sup\u003e\u003c/strong\u003e, the molecular and spatial mechanisms governing lymph node metastasis remain largely unresolved. In particular, the specific epithelial subpopulations responsible for initiating metastasis have not been clearly defined. In this study, we integrated scRNA-seq with ST to characterize MSECs within both primary CRC tumors and paired metastatic lymph nodes, revealing their spatial distribution, temporal dynamics, and potential functional significance.\u003c/p\u003e\u003cp\u003eCSCs have been widely implicated as critical drivers of metastatic dissemination and tumor recurrence, with mounting evidence indicating that only a subset of malignant cells possess the capacity to propagate tumors at distant sites\u003cstrong\u003e\u003csup\u003e[6,37-39]\u003c/sup\u003e\u003c/strong\u003e. For example, ablation of LGR5\u003csup\u003e+\u003c/sup\u003e CSCs in orthotopically xenografted CRC organoid models suppresses tumor growth, whereas their re-emergence restores proliferative capacity during early dissemination\u003cstrong\u003e\u003csup\u003e[40]\u003c/sup\u003e\u003c/strong\u003e. In our study, MSECs were localized predominantly along tumor-invasive boundaries and exhibited molecular features consistent with stemness and metastatic competence. Notably, we observed significant overexpression of UBE2H and LAMC2 in MSECs, two genes previously associated with aggressive tumor phenotypes. UBE2H, a ubiquitin-conjugating enzyme, has been shown to be upregulated in CRC and linked to poor prognosis and metastatic progression, consistent with reports in lung adenocarcinoma\u003cstrong\u003e\u003csup\u003e[22]\u003c/sup\u003e\u003c/strong\u003e. Similarly, LAMC2 has been implicated in promoting local invasion and lymphatic spread in CRC and may serve as a prognostic biomarker\u003cstrong\u003e\u003csup\u003e[41]\u003c/sup\u003e\u003c/strong\u003e. Functional assays further demonstrated that silencing UBE2H and LAMC2 markedly reduced CRC cell proliferation, migration, invasion, and EMT activity, underscoring their pro-metastatic roles. Both scRNA-seq and ST data confirmed the enrichment of UBE2H\u003csup\u003e+\u003c/sup\u003e/LAMC2\u003csup\u003e+\u003c/sup\u003e MSECs within primary tumors, suggesting that lymph node metastases may arise from a spatially distinct, transcriptionally defined subpopulation within the primary site. Although previous studies have utilized scRNA-seq to profile distant metastases in CRC\u003cstrong\u003e\u003csup\u003e[42-44]\u003c/sup\u003e\u003c/strong\u003e, lymphatic dissemination often constitutes the earliest step in the metastatic cascade. Our study provides the first spatially resolved single-cell atlas of this early metastatic event, revealing the molecular and metabolic characteristics of disseminating tumor cells. Importantly, our work was conducted using patient-matched primary CRC tumors and metastatic lymph nodes, enabling a clinically relevant reconstruction of tumor evolution. These findings offer new insights into the cellular origins of lymph node metastasis in CRC and suggest potential molecular targets for intercepting early-stage metastatic progression.\u003c/p\u003e\u003cp\u003eMetastatic progression involves a cascade of genetic, epigenetic, and transcriptional alterations that endow tumor cells with enhanced invasive and adaptive capacities\u003cstrong\u003e\u003csup\u003e[45]\u003c/sup\u003e\u003c/strong\u003e. Recent advances in high-resolution sequencing technologies have enabled deeper exploration of these evolutionary trajectories. Our study identifies MSECs as a key tumor subpopulation driving metastasis, and underscores the critical role of tumor-microenvironment interactions in this process\u003cstrong\u003e\u003csup\u003e[46]\u003c/sup\u003e\u003c/strong\u003e. Notably, MSECs were preferentially localized at the tumor-host interface, a region often enriched for early disseminating tumor cells. The surrounding microenvironment at this invasive front appears to modulate intercellular communication, reducing direct interactions between tumor cells and stromal components\u003cstrong\u003e\u003csup\u003e[47]\u003c/sup\u003e\u003c/strong\u003e. In squamous cell carcinoma, for example, tumor-specific clusters at the invasive margin engage with multiple stromal lineages to suppress local invasion. Similarly, our CellChat analysis of scRNA-seq data revealed that the NOTCH signaling pathway is highly active in MSECs, particularly through communication with myoCAFs. The NOTCH pathway has a context-dependent role in tumorigenesis and is known to promote tumor stem cell proliferation and self-renewal\u003cstrong\u003e\u003csup\u003e[48-50]\u003c/sup\u003e\u003c/strong\u003e. Additionally, we identified a critical interaction mediated by the SDC4-COL1A1 ligand-receptor pair, which enhances CRC invasiveness and metastatic competence. These findings suggest that therapeutic targeting of the SDC4-COL1A1 axis or its downstream NOTCH signaling may represent a promising strategy for metastatic CRC intervention.\u003c/p\u003e\u003cp\u003eMacrophages are essential components of the tumor immune microenvironment and exhibit multifaceted roles in tumor initiation, progression, and metastatic dissemination. In our dataset, macrophages within lymph node metastatic CRC displayed predominantly M2-like immunosuppressive phenotypes, characterized by elevated expression of immune checkpoint and other immunoregulatory genes. This macrophage subset contributes to immune evasion, extracellular matrix (ECM) remodeling, and promotion of metastasis\u003cstrong\u003e\u003csup\u003e[51-54]\u003c/sup\u003e\u003c/strong\u003e. Spatial transcriptomics-based CellChat analysis further confirmed that SDC4-COL1A1 interactions were spatially enriched between tumor epithelial cells and adjacent fibroblasts, reinforcing the notion that the TME is spatially organized and functionally compartmentalized. Moreover, we found that myoCAFs consistently localize in proximity to tumor epithelial cells, a pattern that was robustly observed across both scRNA-seq and spatial datasets. Collectively, these findings underscore the spatial heterogeneity of the TME and its influence on tumor cell behavior, particularly in the metastatic cascade of CRC.\u003c/p\u003e\u003cp\u003eThere are several limitations to this study. First, the substantial inter-patient heterogeneity poses challenges in identifying universal features across samples. Second, the sample size is limited—we performed scRNA-seq and spatial transcriptomics on only five CRC patients—which restricts the generalizability of our findings. To strengthen the robustness of our conclusions, we applied an integrative strategy combining both technologies to improve data confidence and spatial resolution. Third, while early dissemination of CRC can occur via both hematogenous and lymphatic routes, the detection of early hematogenous metastasis remains technically difficult and inconsistently characterized. As a result, we focused on lymphatic dissemination in this study, while the molecular mechanisms underlying early hematogenous spread warrant future investigation.\u003c/p\u003e\u003cp\u003eOur results delineate the spatiotemporal characteristics of MSECs, demonstrating their spatial enrichment at the tumor–host interface and close association with myoCAFs at both single-cell and spatial levels. Importantly, we found that fibroblasts exhibit distinct transcriptional states depending on their spatial proximity to tumor epithelial cells, highlighting the dynamic evolution of CRC cells during early metastatic dissemination. These findings lay a foundation for further exploration of early dissemination mechanisms and offer potential prognostic and therapeutic implications.\u003c/p\u003e\u003cp\u003eIn summary, we conducted scRNA-seq and spatial transcriptomics to systematically investigate the cellular heterogeneity and tumor ecosystems of paired primary and metastatic lesions in CRC. Through the integrated analysis of 84,735 single cells from five patients, we identified nine epithelial subpopulations, among which clusters C3 and C4 were enriched for stem-like features and metastatic potential. MSECs were characterized by elevated expression of UBE2H and LAMC2, which promoted EMT and metastatic dissemination and were associated with shorter progression-free survival. Spatial analysis confirmed that MSECs localize to the tumor periphery, where they engage myoCAFs via the SDC4-COL1A1 ligand-receptor axis, shaping an immunosuppressive microenvironment and limiting immune infiltration. Additionally, tumor-associated macrophages within metastatic lymph nodes exhibited M2-like polarization and expressed immunoregulatory molecules such as PD-L1 and VISTA, further facilitating immune evasion.\u003c/p\u003e\u003cp\u003eTogether, this study presents a comprehensive single-cell and spatial atlas of lymph node metastatic CRC, identifying MSECs as key drivers of dissemination. We demonstrate that UBE2H and LAMC2 promote the metastatic phenotype of CRC cells, and that the SDC4-COL1A1-NOTCH axis mediates critical tumor-stroma interactions. These findings provide new mechanistic insights into lymphatic dissemination in CRC and suggest actionable therapeutic targets to prevent metastatic progression.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003eEthical statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was reviewed and approved by the Institutional Review Board of Tongji Hospital, Tongji University. All patients provided written informed consent prior to the collection of samples.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHuman specimens\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study involved five patients who underwent surgerical treatment at the Department of General Surgery, Tongji Hospital, Tongji University. A total of ten fresh surgical specimens were collected, consisting of five primary tumors and five paired metastatic lymph nodes. The presence of lymph node metastases was confirmed by pathologists through intraoperative cytologic assessments and postoperative analysis of paraffin-embedded sections. Clinical data, including demographic and clinicopathologic characteristics of the tumors and staging at diagnosis according to the American Joint Committee on Cancer System, are detailed in\u003cstrong\u003e\u0026nbsp;Supplementary Table 1\u003c/strong\u003e,\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell culture\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwo CRC cell lines, SW1116 and SW480, were obtained from the Chinese Academy of Sciences (Shanghai, China). The cells were cultured in 5% CO\u003csub\u003e2\u003c/sub\u003e atmosphere at 37℃ using DMEM or RPMI-1640 medium (Corning) supplemented with 1% penicillin-streptomycin and 10% fetal bovine serum (FBS). The cell lines were confirmed to be mycoplasma-free and subjected to monthly authentication via PCR analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePreparation of single-cell suspensions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePrior to tissue processing, adipose tissue and visible blood vessels were removed. Fresh normal mucosa and CRC tissues were rinsed with ice-cold PBS, and minced into small fragments. Normal mucosal tissues were incubated in 10 mL of EDTA-containing buffer (15 mM HEPES, 5 mM EDTA, 1 mM DTT and 10% FBS-supplemented PBS) for 1 h at 37°C. Tumor tissues were treated with 10 mL of DTT (65 mM)-supplemented PBS (with 10% FBS) for 15 min at 37'C with continuous shaking. Following incubation, EDTA and DTT were removed by washing with PBS twice. The tissue fragments were then digested with collagenase VIII (0.38 mg/ml) and DNase I (0.1 mg/ml) in complete RPMI-1640 medium (with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin) for 1 h at 37°C. Following digestion, the mixture was vigorously shaken for 5 min and cells were mechanically dissociated using 21-gauge syringes. The resulting cell suspension were filtered through a 100\u0026nbsp;µm filter, pelleted and washed twice with PBS to prepare freshly isolated cell suspensions for scRNA-seq and flow cytometry staining.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSingle-cell RNA sequencing\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFreshly prepared cell suspensions were processed immediately according to the manufacturer’s protocol of the 10X Chromium 3’ v3 kit (10x Genomics, Pleasanton, CA). Libraries were prepared and sequenced on the NovaSeq 6000 platform (Illumina, Inc., San Diego, CA) at GENERGY BIO (Shanghai, China).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDimension reduction and clustering analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData were scaled using the top 2000 most variable genes via FindVariableFeatures function in the Seurat v3 R package. Principal component analysis (PCA) was performed using variable genes, followed by neighbor identification with FindNeighbors to facilitate graph-based clustering. Cell subtypes were delineated using FindClusters, and UMAP (uniform manifold approximation and projection) visualization was employed to illustrate cell distribution. To mitigate the batch effects, the Harmony algorithm was utilized prior to clustering. Cells were first partitioned into epithelial, stromal, myeloid, T, and B cell categories, followed by a more granular classification into distinct subtypes across samples. In the first step, clusters were scored for the predefined gene signatures [3], including epithelial cells (EPCAM, KRT8, KRT18), stromal cells (COL1A1, COL1A2, COL6A1, COL6A2, VWF, PLVAP, CDH5, S100B), myeloid cells (CD68, XCR1, CLEC9A, CLEC10A, CD1C, S100A8, S100A9, TPSAB1, and OSM), T cells (NKG7, KLRC1, CCR7, FOXP3, CTLA4, CD8B, CXCR6, and CD3D), and B cells (MZB1, IGHA1, SELL, CD19, and AICDA). Signature scores were calculated as the mean log\u003csub\u003e2\u003c/sub\u003e(LogNormalizedUMI+1) across all genes in each signature. Each cluster was assigned to the compartment of its maximal score and all cluster assignments were manually verified to ensure the accurate partitioning. In the second step, the Harmony algorithm was applied before clustering analysis to correct for batch effects, followed by further application of FindNeighbors and FindClusters in Seurat to obtain cell subtypes. Additionally, we defined 58 cell types in CRC based on the gene signatures and known lineage markers for each cell type.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInferCNV analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInitial copy number variations (CNVs) for each region were estimated using inferCNV R package (version 1.0.4). The CNV for each cell type was calculated according to expression level derived from single-cell sequencing data, employing a cutoff of 0.1. Genes were sorted according to their chromosomal locations, and a moving average of gene expression was computed using a window size of 101 genes. The expression values were then centered to zero by subtracting the mean. Epithelial cells were selected as malignant cells, while other cells were considered as normal cells. De-noising procedures were applied to refine the CNV profiles, resulting in the final CNV datasets for anlaysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell Cycle analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo calculate cell cycle scores for G1, S and G2/M phases, the Cyclone function of the scran R package (version 1.14.3) was applied. Cells were classified as follows: G1 phase (if the G1 score is above 0.5 and greater than the G2/M score); G2/M phase (if the G2/M score is above 0.5 and greater than the G1 score); S phase (if neither score exceeds 0.5). The results were visualized using UMAP to show the proportions of different cell cycle across various clusters or subgroups.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMonocle2 Pseudotime Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe developmental pseudotime was determined using the Monocle2 package (version 2.9.0). The raw counts were first converted from Seurat object into Cell Data Set object by the importCDS function of Monocle. The differentialGeneTest function was employed to select ordering genes (with q-value \u0026lt; 0.01), which were likely to be informative for ordering cells along the pseudotime trajectory. Dimensional reduction clustering analysis was performed using the reduceDimension function, followed by trajectory inference with the orderCells function using default parameters. Changes in gene expression over pseudotime were visualized with plot_genes_in_pseudotime function,\u0026nbsp;allowing for the tracking of developmental changes throughout the pseudotime trajectory.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003escVelo Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo perform the RNA velocity analysis, the spliced reads and unspliced reads were recounted using the Python script velocyto.py (https://github.com/velocyto-team/velocyto.py) on the output folder from Cell Ranger. Following this, the likelihood-based dynamical model and the velocity graph were constructed using scVelo ( https://scvelo.readthedocs.io/). This process enabled to calculate the RNA velocities, which include rates of transcription, splicing and degradation for individual single cells. The resulting velocity fields were then projected onto the UMAP of Seurat, allowing for visualization of cellular dynamics in the context of the identified clusters.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCytoTRACE Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCytoTRACE (version 0.3.3) was used to predict the differentiation states from scRNA-seq data. The raw count data was first loaded, and the CytoTRACE function was utilized with the enableFast = TRUE parameter to determine the recommended time order of cells. Visualization of the results was achieved using the plotCytoTRACE function.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003escMetabolism Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003escMetabolism (version 0.2.1) was performed to quantify the metabolic activity from scRNA-seq data. The metabolic pathway activity for each cell was assessed using VISIONalgorithm. This analysis leveraged preset preset metabolic gene sets, including 85 KEGG pathways and 82 REACTOME pathways, to score the metabolic activity comprehensively across the sampled cells.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell subtype similarity analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the similarity of cell subtypes, the following steps were performed: (1) the top 1,000 highly variable genes were identified across different cell subtypes; (2) the mean expression values of these top 1,000 highly variable genes were calculated for each cell subtype; (3) the hierarchical clustering was performed using a distance metric defined by (1-Pearson correlation coefficient)/2.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDifferential expression analysis\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo identify differentially expressed genes for each cell subtype, the functions “FindAllMarkers” (multiple condition comparisons) and “FindMarkers” (two condition comparisons) from the Seurat package were used with default parameters. Genes were considered as differentially expressed if they met the criteria of a p-value \u0026lt; 0.05 and a log\u003csub\u003e2\u003c/sub\u003e(fold change, FC) \u0026gt; 0.25.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePlasmid constructions and transduction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHuman UBE2H and LAMC2 were amplified using the reverse-transcribed cDNA from SW1116 and other cell lines. The siRNA targeting UBE2H and LAMC2 were purchased from Sangon Biotech.\u0026nbsp;The specific siRNA sequences are as follows:\u003c/p\u003e\n\u003cp\u003esiUBE2H:\u003c/p\u003e\n\u003cp\u003eForward Primer TTATTGGCCTATCCTAACCCCA\u003c/p\u003e\n\u003cp\u003eReverse Primer TGCTTGTATTCTTCTGGTCGGT\u003c/p\u003e\n\u003cp\u003esiLAMC2:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eForward Primer GACAAACTGGTAATGGATTCCGC\u003c/p\u003e\n\u003cp\u003eReverse Primer TTCTCTGTGCCGGTAAAAGCC\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e10x Visium spatial RNA-seq data analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe FASTQ files were processed and aligned to GRch38 huma reference genome using Space Ranger software (version 2.0.1) from 10x Genomics, with unique molecular identifier (UMI) counts summarized for each barcode. To distinguish tissue overlaying spots from the background, tissue overlaying spots were detected according to the images. The filtered UMI count matrix was then analyzed\u0026nbsp;using Seurat (version 4.1.0) R package.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBright-field imaging\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBright-field images were captured by a whole slide scanner (Panoramic MIDI FL, 3DHISTECH) at a 20× resolution. H\u0026amp;E-stained sections of each sample were carefully reviewed by two experienced pathologists to confirm the pathology. Following this, a trained pathologist manually annotated the images to identify distinct regions, including tumor, stromal, and normal regions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eST barcoded microarray slide Information:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe library preparation slides utilized in this study were obtained from the ST team (https://www.spatialtranscriptomics.com). Each slide features an array of spots measuring 55 μm in diameter, spaced 100 μm apart from center to center, covering a total area of 6.5×6.5 mm². Each slide contains four capture zones, with each zone comprising approximately 5,000 unique gene expression spots. This design enables high-resolution spatial transcriptomics analysis, allowing for in-depth examination of gene expression patterns within tissue samples, which is crucial for understanding tissue architecture and cellular interactions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRCTD Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo infer the cell-type composition of each spot, RCTD (version\u0026nbsp;1.1.0) was employed. The analysis utilized\u0026nbsp;the default parameters in the creat. RCTD function, with specific modifications: a minimum number of cell \u0026gt; 1 per cell type and a minimum of unique molecular identifier (UMI) count \u0026gt; 1 per pixel. Additionally, the doublet_mode parameter was configured to FALSE in the run.RCTD function.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCellChat Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe cell communication analysis was conducted using the CellChat R package (version 1.1.3). The processs began by importing the normalized expression matrix to create a cellchat object using the create CellChat function. Subsequently, the data was preprocessed through several embedded functions with the default parameters: identify Over Expressed Genes, identify Over Expressed Interactions and project Data. Potential ligand-receptor interactions were assessed using the compute Commun Prob, filter Communication (with a minimum of 10 cells) and compute Commun Prob Pathway functions. Finally, the cell communication network was aggregated using the aggregate Net function, allowing for a comprehensive understanding of the interactions between different cell types within the tissue.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell Communication Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCell communication analysis was conducted using CellPhoneDB (version 4.1.0) to identify biologically relevant ligand-receptor interactions from scRNA-seq data. A ligand or receptor was considered “expressed” in a specific cell type if at least 10% of the cells exhibited non-zero read counts for the corresponding gene. Statistical significance was assessed by randomly shuffling cluster labels, generating a null distribution through 1,000 permutations, and calculating p-values based on a normal distribution curve derived from permuted interaction scores. Cell-cell communication networks were constructed by linking cell types where ligands were expressed in one type and corresponding receptors in another, and these networks were visualized using the Igraph (version 1.2.4.1) and Circlize (version 0.4.8) R packages. In spatial communication analysis, intra-spots represented closely interacting cells of the same type, inter-spots included other cell types within two layers of proximity, and distal-spots indicated minimal interaction. This classification helped elucidate spatial communication dynamics within the tissue. Additionally, spatial co-localization analysis was performed by selecting a specific cell type as the center spot, calculating Euclidean distances, and generating density plots to infer co-localization patterns at varying distances.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eCRC: Colorectal Cancer; LNM: Lymph Node Metastasis; TME: Tumor Microenvironment; scRNA-seq: Single-Cell RNA sequencing; ST: Spatial Transcriptomics; MSECs: Metastatic Stem-like Epithelial Cells; EMT: Epithelial-Mesenchymal Transition; myoCAFs: myofibroblast fibroblasts; DFS: Disease-Free Survival; CSCs: Cancer stem cells; ECM: Extracellular matrix; CAFs: Cancer-Associated Fibroblasts; IHC: Immunohistochemistry; mIHC: Multiplex immunohistochemical; UBE2H: Uiquitin-conjugating enzyme E2 H; LAMC2: Laminin subunit gamma-2; COL1A1: Collagen Type I Alpha 1 Chain; SDC4: Syndecan-4; UMAP: Manifold Approximation and Projection; CNV: copy number variation; siRNA: small interfering RNA; H\u0026amp;E: hematoxylin and eosin; RCTD: Robust Cell Type Decomposition.\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was financially supported by the following grants: Jiangsu Provincial Research Hospital Project (YJXYY202204-ZD06), Jiangsu Commission of Health (K2024011), Postdoctoral Research Funding Project of Jiangsu Province (2021K012A), Science and Technology Bureau of Nantong City (JC2023108), Macao Science and Technology Development Fund (0133/2024/RIA2 and 0069/2021/AFJ), Shanghai Science and Technology Commission (23Y11902400), Shanghai Municipal Health Commission (202340231), Tongji University Scientific Research Project (22120240386), Ganquan New Star Talent Program of Shanghai Tongji Hospital (HBRC2104), Shanghai Tongji Hospital Research Funding (RCOD2102, ITJ(ZD)2308, and GJPY2111).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and/or analyzed during the current study are in the process of being uploaded and will be made available to the journal and the public upon request as soon as the deposition is complete.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;Human colorectal cancer tumor and lymph node samples were collected from patients at Shanghai Tongji Hospital in accordance with institutional ethical guidelines under protocol number 2021-116, approved by the Ethics Committee of Shanghai Tongji Hospital. Informed consent was obtained from all participants prior to sample collection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003cbr\u003e\u0026nbsp;The authors declare no conflicts of interest and no competing financial interests. All participants provided written informed consent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eXinxing Li and Feng Wang conceived and supervised the study. Jinran Wu, Xiaomao Yin, and Jiexuan Wang performed the single-cell and spatial transcriptomic experiments. Lin Zhu, Xin Yang, and Jingjing Qian were responsible for sample processing, quality control, and library preparation. Wenqiang Wang, Liangchen Zhu, Xuan Dai, and Zekun Zhao contributed to clinical sample collection and annotation. Dongsheng Li, Siyuan Yin, and Runqi Hong performed bioinformatics analysis and data interpretation. Kai Xu, Zhiqian Hu, and Io Nam Wong assisted with computational modeling and pathway analysis. Yan Wang and Jiahui Yin contributed to figure preparation and literature review. Jinran Wu and Xiaomao Yin drafted the manuscript with input from all authors. Xinxing Li, Io Nam Wong, and Feng Wang critically revised the manuscript and provided overall project coordination. All authors read and approved the final manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eCHEN K, COLLINS G, WANG H, et al. Pathological Features and Prognostication in Colorectal Cancer [J]. Current oncology (Toronto, Ont), 2021, 28(6): 5356-83.\u003c/li\u003e\n\u003cli\u003eSIEGEL R L, MILLER K D, FUCHS H E, et al. Cancer statistics, 2022 [J]. CA: a cancer journal for clinicians, 2022, 72(1): 7-33.\u003c/li\u003e\n\u003cli\u003eBENSON A B, VENOOK A P, AL-HAWARY M M, et al. Colon Cancer, Version 2.2021, NCCN Clinical Practice Guidelines in Oncology [J]. Journal of the National Comprehensive Cancer Network : JNCCN, 2021, 19(3): 329-59.\u003c/li\u003e\n\u003cli\u003eFRANK M H, WILSON B J, GOLD J S, et al. Clinical Implications of Colorectal Cancer Stem Cells in the Age of Single-Cell Omics and Targeted Therapies [J]. Gastroenterology, 2021, 160(6): 1947-60.\u003c/li\u003e\n\u003cli\u003eBAYIK D, LATHIA J D. Cancer stem cell-immune cell crosstalk in tumour progression [J]. Nature reviews Cancer, 2021, 21(8): 526-36.\u003c/li\u003e\n\u003cli\u003eBATLLE E, CLEVERS H. Cancer stem cells revisited [J]. Nature medicine, 2017, 23(10): 1124-34.\u003c/li\u003e\n\u003cli\u003eEUN K, HAM S W, KIM H. Cancer stem cell heterogeneity: origin and new perspectives on CSC targeting [J]. BMB reports, 2017, 50(3): 117-25.\u003c/li\u003e\n\u003cli\u003eWU J, JING X, DU Q, et al. Disruption of the Clock Component Bmal1 in Mice Promotes Cancer Metastasis through the PAI-1-TGF-\u0026beta;-myoCAF-Dependent Mechanism [J]. Advanced science (Weinheim, Baden-Wurttemberg, Germany), 2023, 10(24): e2301505.\u003c/li\u003e\n\u003cli\u003eSHIBUE T, WEINBERG R A. EMT, CSCs, and drug resistance: the mechanistic link and clinical implications [J]. Nature reviews Clinical oncology, 2017, 14(10): 611-29.\u003c/li\u003e\n\u003cli\u003e LAMBERT A W, WEINBERG R A. Linking EMT programmes to normal and neoplastic epithelial stem cells [J]. Nature reviews Cancer, 2021, 21(5): 325-38.\u003c/li\u003e\n\u003cli\u003e CA\u0026ntilde;ELLAS-SOCIAS A, CORTINA C, HERNANDO-MOMBLONA X, et al. Metastatic recurrence in colorectal cancer arises from residual EMP1(+) cells [J]. Nature, 2022, 611(7936): 603-13.\u003c/li\u003e\n\u003cli\u003e WEI C, SUN W, SHEN K, et al. Delineating the early dissemination mechanisms of acral melanoma by integrating single-cell and spatial transcriptomic analyses [J]. Nature communications, 2023, 14(1): 8119.\u003c/li\u003e\n\u003cli\u003e QUAH H S, CAO E Y, SUTEJA L, et al. Single cell analysis in head and neck cancer reveals potential immune evasion mechanisms during early metastasis [J]. Nature communications, 2023, 14(1): 1680.\u003c/li\u003e\n\u003cli\u003e ZHANG S, FANG W, ZHOU S, et al. Single cell transcriptomic analyses implicate an immunosuppressive tumor microenvironment in pancreatic cancer liver metastasis [J]. Nature communications, 2023, 14(1): 5123.\u003c/li\u003e\n\u003cli\u003e KARIMI E, YU M W, MARITAN S M, et al. Single-cell spatial immune landscapes of primary and metastatic brain tumours [J]. Nature, 2023, 614(7948): 555-63.\u003c/li\u003e\n\u003cli\u003e LIU Y, HE M, TANG H, et al. Single-cell and spatial transcriptomics reveal metastasis mechanism and microenvironment remodeling of lymph node in osteosarcoma [J]. BMC medicine, 2024, 22(1): 200.\u003c/li\u003e\n\u003cli\u003e ZILIONIS R, ENGBLOM C, PFIRSCHKE C, et al. Single-Cell Transcriptomics of Human and Mouse Lung Cancers Reveals Conserved Myeloid Populations across Individuals and Species [J]. Immunity, 2019, 50(5): 1317-34.e10.\u003c/li\u003e\n\u003cli\u003e PURAM S V, TIROSH I, PARIKH A S, et al. Single-Cell Transcriptomic Analysis of Primary and Metastatic Tumor Ecosystems in Head and Neck Cancer [J]. Cell, 2017, 171(7): 1611-24.e24.\u003c/li\u003e\n\u003cli\u003e LAWSON D A, KESSENBROCK K, DAVIS R T, et al. Tumour heterogeneity and metastasis at single-cell resolution [J]. Nature cell biology, 2018, 20(12): 1349-60.\u003c/li\u003e\n\u003cli\u003e XING H, LIU D, LI J, et al. TERTp Mutation and its Prognostic Value in Glioma Patients Under the 2021 WHO Classification: A Real-World Study [J]. Cancer medicine, 2025, 14(2): e70533.\u003c/li\u003e\n\u003cli\u003e GULATI G S, SIKANDAR S S, WESCHE D J, et al. Single-cell transcriptional diversity is a hallmark of developmental potential [J]. Science (New York, NY), 2020, 367(6476): 405-11.\u003c/li\u003e\n\u003cli\u003e YEN M C, WU K L, LIU Y W, et al. Ubiquitin Conjugating Enzyme E2 H (UBE2H) Is Linked to Poor Outcomes and Metastasis in Lung Adenocarcinoma [J]. Biology, 2021, 10(5).\u003c/li\u003e\n\u003cli\u003e CABLE D M, MURRAY E, ZOU L S, et al. Robust decomposition of cell type mixtures in spatial transcriptomics [J]. Nature biotechnology, 2022, 40(4): 517-26.\u003c/li\u003e\n\u003cli\u003e CHEN J, LIU W, LUO T, et al. A comprehensive comparison on cell-type composition inference for spatial transcriptomics data [J]. Briefings in bioinformatics, 2022, 23(4).\u003c/li\u003e\n\u003cli\u003e BYERS L A, DIAO L, WANG J, et al. An epithelial-mesenchymal transition gene signature predicts resistance to EGFR and PI3K inhibitors and identifies Axl as a therapeutic target for overcoming EGFR inhibitor resistance [J]. Clinical cancer research : an official journal of the American Association for Cancer Research, 2013, 19(1): 279-90.\u003c/li\u003e\n\u003cli\u003e LI R, LIU X, HUANG X, et al. Single-cell transcriptomic analysis deciphers heterogenous cancer stem-like cells in colorectal cancer and their organ-specific metastasis [J]. Gut, 2024, 73(3): 470-84.\u003c/li\u003e\n\u003cli\u003e LI, X., PAN, J., LIU, T., YIN, W., MIAO, Q., ZHAO, Z., GAO, Y., ZHENG, W., LI, H., DENG, R., HUANG, D., QIU, S., ZHANG, Y., QI, Q., DENG, L., HUANG, M., TANG, P. M., CAO, Y., CHEN, M., YE, W. \u0026amp; ZHANG, D. 2023. Novel TCF21(high) pericyte subpopulation promotes colorectal cancer metastasis by remodelling perivascular matrix. \u003cem\u003eGut,\u003c/em\u003e 72\u003cstrong\u003e,\u003c/strong\u003e 710-721.\u003c/li\u003e\n\u003cli\u003e LI, X., SUN, Z., PENG, G., XIAO, Y., GUO, J., WU, B., LI, X., ZHOU, W., LI, J., LI, Z., BAI, C., ZHAO, L., HAN, Q., ZHAO, R. C. \u0026amp; WANG, X. 2022. Single-cell RNA sequencing reveals a pro-invasive cancer-associated fibroblast subgroup associated with poor clinical outcomes in patients with gastric cancer. \u003cem\u003eTheranostics,\u003c/em\u003e 12\u003cstrong\u003e,\u003c/strong\u003e 620-638.\u003c/li\u003e\n\u003cli\u003e LIN, S., MA, L., MO, J., ZHAO, R., LI, J., YU, M., JIANG, M. \u0026amp; PENG, L. 2024. Immune cell senescence and exhaustion promote the occurrence of liver metastasis in colorectal cancer by regulating epithelial-mesenchymal transition. \u003cem\u003eAging (Albany NY),\u003c/em\u003e 16\u003cstrong\u003e,\u003c/strong\u003e 7704-7732.\u003c/li\u003e\n\u003cli\u003e LIOTTA, L. A. \u0026amp; KOHN, E. C. 2001. The microenvironment of the tumour-host interface. \u003cem\u003eNature,\u003c/em\u003e 411\u003cstrong\u003e,\u003c/strong\u003e 375-9.\u003c/li\u003e\n\u003cli\u003e LIU, X., QIN, J., NIE, J., GAO, R., HU, S., SUN, H., WANG, S. \u0026amp; PAN, Y. 2023. ANGPTL2+cancer-associated fibroblasts and SPP1+macrophages are metastasis accelerators of colorectal cancer. \u003cem\u003eFront Immunol,\u003c/em\u003e 14\u003cstrong\u003e,\u003c/strong\u003e 1185208.\u003c/li\u003e\n\u003cli\u003e PETITPREZ F, DE REYNI\u0026egrave;S A, KEUNG E Z, et al. B cells are associated with survival and immunotherapy response in sarcoma [J]. Nature, 2020, 577(7791): 556-60.\u003c/li\u003e\n\u003cli\u003e SHEN T, LIU J L, WANG C Y, et al. Targeting Erbin in B cells for therapy of lung metastasis of colorectal cancer [J]. Signal transduction and targeted therapy, 2021, 6(1): 115.\u003c/li\u003e\n\u003cli\u003e MA, C., YANG, C., PENG, A., SUN, T., JI, X., MI, J., WEI, L., SHEN, S. \u0026amp; FENG, Q. 2023. Pan-cancer spatially resolved single-cell analysis reveals the crosstalk between cancer-associated fibroblasts and tumor microenvironment. \u003cem\u003eMol Cancer,\u003c/em\u003e 22\u003cstrong\u003e,\u003c/strong\u003e 170.\u003c/li\u003e\n\u003cli\u003e AZIZI E, CARR A J, PLITAS G, et al. Single-Cell Map of Diverse Immune Phenotypes in the Breast Tumor Microenvironment [J]. Cell, 2018, 174(5): 1293-308.e36.\u003c/li\u003e\n\u003cli\u003e WANG F, LONG J, LI L, et al. Single-cell and spatial transcriptome analysis reveals the cellular heterogeneity of liver metastatic colorectal cancer [J]. Science advances, 2023, 9(24): eadf5464.\u003c/li\u003e\n\u003cli\u003e YIN W, WANG J, JIANG L, et al. Cancer and stem cells [J]. Experimental biology and medicine (Maywood, NJ), 2021, 246(16): 1791-801.\u003c/li\u003e\n\u003cli\u003e BABAEI G, AZIZ S G, JAGHI N Z Z. EMT, cancer stem cells and autophagy; The three main axes of metastasis [J]. Biomedicine \u0026amp; pharmacotherapy = Biomedecine \u0026amp; pharmacotherapie, 2021, 133: 110909.\u003c/li\u003e\n\u003cli\u003e LAMBERT A W, PATTABIRAMAN D R, WEINBERG R A. Emerging Biological Principles of Metastasis [J]. Cell, 2017, 168(4): 670-91.\u003c/li\u003e\n\u003cli\u003e SHIMOKAWA M, OHTA Y, NISHIKORI S, et al. Visualization and targeting of LGR5(+) human colon cancer stem cells [J]. Nature, 2017, 545(7653): 187-92.\u003c/li\u003e\n\u003cli\u003e CHENG L, LI X, DONG W, et al. LAMC2 regulates the proliferation, invasion, and metastasis of gastric cancer via PI3K/Akt signaling pathway [J]. Journal of cancer research and clinical oncology, 2024, 150(5): 230.\u003c/li\u003e\n\u003cli\u003e SHEN T, LIU J L, WANG C Y, et al. Targeting Erbin in B cells for therapy of lung metastasis of colorectal cancer [J]. Signal transduction and targeted therapy, 2021, 6(1): 115.\u003c/li\u003e\n\u003cli\u003e LI R, LIU X, HUANG X, et al. Single-cell transcriptomic analysis deciphers heterogenous cancer stem-like cells in colorectal cancer and their organ-specific metastasis [J]. Gut, 2024, 73(3): 470-84.\u003c/li\u003e\n\u003cli\u003e LIU X, QIN J, NIE J, et al. ANGPTL2+cancer-associated fibroblasts and SPP1+macrophages are metastasis accelerators of colorectal cancer [J]. Frontiers in immunology, 2023, 14: 1185208.\u003c/li\u003e\n\u003cli\u003e GUI P, BIVONA T G. Evolution of metastasis: new tools and insights [J]. Trends in cancer, 2022, 8(2): 98-109.\u003c/li\u003e\n\u003cli\u003e ZHANG Z, WANG Y, ZHANG J, et al. COL1A1 promotes metastasis in colorectal cancer by regulating the WNT/PCP pathway [J]. Molecular medicine reports, 2018, 17(4): 5037-42.\u003c/li\u003e\n\u003cli\u003e ZHENG H, LIU H, GE Y, et al. Integrated single-cell and bulk RNA sequencing analysis identifies a cancer associated fibroblast-related signature for predicting prognosis and therapeutic responses in colorectal cancer [J]. Cancer cell international, 2021, 21(1): 552.\u003c/li\u003e\n\u003cli\u003e CHU X, TIAN W, NING J, et al. Cancer stem cells: advances in knowledge and implications for cancer therapy [J]. Signal transduction and targeted therapy, 2024, 9(1): 170.\u003c/li\u003e\n\u003cli\u003e LI Y, LIN C, CHU Y, et al. Characterization of Cancer Stem Cells in Laryngeal Squamous Cell Carcinoma by Single-cell RNA Sequencing [J]. Genomics, proteomics \u0026amp; bioinformatics, 2024, 22(4).\u003c/li\u003e\n\u003cli\u003e SHI Q, XUE C, ZENG Y, et al. Notch signaling pathway in cancer: from mechanistic insights to targeted therapies [J]. Signal transduction and targeted therapy, 2024, 9(1): 128.\u003c/li\u003e\n\u003cli\u003e DAKAL T C, BHUSHAN R, XU C, et al. Intricate relationship between cancer stemness, metastasis, and drug resistance [J]. MedComm, 2024, 5(10): e710.\u003c/li\u003e\n\u003cli\u003e MANTOVANI A, LOCATI M. Tumor-associated macrophages as a paradigm of macrophage plasticity, diversity, and polarization: lessons and open questions [J]. Arteriosclerosis, thrombosis, and vascular biology, 2013, 33(7): 1478-83.\u003c/li\u003e\n\u003cli\u003e QIAN B Z, POLLARD J W. Macrophage diversity enhances tumor progression and metastasis [J]. Cell, 2010, 141(1): 39-51.\u003c/li\u003e\n\u003cli\u003e RUFFELL B, COUSSENS L M. Macrophages and therapeutic resistance in cancer [J]. Cancer cell, 2015, 27(4): 462-72.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Colorectal cancer, Metastatic stem-like epithelial cells, Myofibroblastic CAFs, Tumor microenvironment, Spatial transcriptomics","lastPublishedDoi":"10.21203/rs.3.rs-6803290/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6803290/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Lymph node metastasis (LNM) is a major determinant of poor prognosis in colorectal cancer (CRC), yet the cellular and spatial mechanisms driving metastatic initiation remain poorly defined. Here, we performed integrative single-cell RNA sequencing (scRNA-seq; 84,735 cells) and spatial transcriptomics (ST; 20,859 spots) on paired primary tumors and metastatic lymph nodes from five CRC patients. We identified a metastatic stem-like epithelial cell population (MSECs), marked by LGR5, AXIN2, UBE2H, and LAMC2, enriched at invasive fronts and metastatic niches. MSECs co-localized with myofibroblastic cancer-associated fibroblasts (myoCAFs) and communicated via a COL1A1-SDC4 axis that activated NOTCH signaling and promoted epithelial–mesenchymal transition (EMT). Functional assays demonstrated that UBE2H or LAMC2 knockdown impaired CRC cell migration, invasion, and EMT, while their high expression predicted poor disease-free survival. Spatial analysis further revealed that myoCAFs formed a stromal barrier facilitating immune exclusion and metastatic expansion. These findings highlight the MSEC–myoCAF crosstalk as a key driver of CRC metastasis and nominate UBE2H, LAMC2, and COL1A1-SDC4 as potential therapeutic targets.","manuscriptTitle":"Single-cell and Spatial Transcriptomic Mapping Reveals MSEC-myoCAF Interactions Driving Lymph Node Metastasis in Colorectal Cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-05 15:09:41","doi":"10.21203/rs.3.rs-6803290/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"b50271be-9c5d-4927-8822-b2f349e2f62e","owner":[],"postedDate":"June 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":49538633,"name":"Biological sciences/Cancer/Gastrointestinal cancer/Colorectal cancer"},{"id":49538634,"name":"Biological sciences/Cancer/Gastrointestinal cancer/Colorectal cancer/Rectal cancer"},{"id":49538635,"name":"Biological sciences/Cancer/Cancer microenvironment"},{"id":49538636,"name":"Biological sciences/Cancer/Tumour heterogeneity"}],"tags":[],"updatedAt":"2025-06-21T07:45:34+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-05 15:09:41","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6803290","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6803290","identity":"rs-6803290","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}