SOX9 reactivation in cancer vessels shapes the tumor micro-environment through hypoxia and immune depletion promoting tumor growth and metastasis

SOX9 reactivation in cancer vessels shapes the tumor micro-environment through hypoxia and immune depletion promoting tumor growth and metastasis | 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 SOX9 reactivation in cancer vessels shapes the tumor micro-environment through hypoxia and immune depletion promoting tumor growth and metastasis Kiarash Khosrotehrani, Ghazaleh Hashemi, Haiming Li, Samuel X Tan, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5457583/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract The development of new vascular structures is a pre-requisite for tumor growth and spread. This process is often disorganised and produces immature and leaky vessels and relies at least in part on the activity of endovascular progenitor cells (EVPs), residing in vessel walls and giving rise to mature endothelial cells in de novo blood vessel networks in the tumor. Sox9 is a transcription factor that is playing an important role in stem cell self-renewal and fate choice and is highly upregulated in EVPs. In this study, we aimed to explore how Sox9 activity in the endothelium affects tumor vascularisation, microenvironment, and metastasis. Indeed, Sox9 expression was upregulated in tumor endothelial cells of mice harbouring melanomas. Similarly, we observed the up regulation of SOX9 in human endothelial cells exposed to melanoma cell co-culture or conditioned medium resulting in increased colony formation and reduced maturity as revealed in tube formation assays. Endothelial-specific conditional knockout of Sox9 (Sox9fl/fl/Cdh5CreERt2/Rosa-YFP) resulted in a significant reduction in total endothelial cells in B16-F0 or HcMel12 melanoma tumors inoculated intradermally in both flow-cytometry, lineage tracing and immunostaining of tumor sections. Functionally, there was a significant reduction in tumour size and lung metastases after Sox9 deletion in the endothelium. Importantly, despite a major reduction in the number and area of CD31 + vessels there was a significant increase in pericyte coverage suggesting increased maturity of the remaining vessels upon Sox9 deletion in the endothelium. These changes in the endothelium translated into a reduction in hypoxia as demonstrated by decreased GLUT1 expression and reduced nuclear localisation of HIF1α. RNA sequencing of sorted tumor cells as well as spatial transcriptomics of tumor sections with endothelial-specific deletion of Sox9 versus controls confirmed the reduction in hypoxia and showed dramatic increases in CD4 and CD8 immune T cell infiltration in the centre of tumors as confirmed by immunostaining. In summary, endothelial-specific Sox9 deletion resulted in fewer and more mature de novo vessels in the centre of the tumor and reduced metastatic dissemination, suggesting strategies that target this pathway may restore the normal function of blood vessels in tumors and prevent disease progression. Biological sciences/Cancer/Tumour angiogenesis Biological sciences/Cancer/Cancer microenvironment Melanoma Vascularisation Metastasis Endothelial Progenitor Cells Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Introduction Melanoma is a potentially aggressive skin cancer with a high risk of metastasis. In melanoma as in many other cancers, vascularisation is pivotal for the growth of cancer cells at the primary site and accelerates the spread of malignant cells 1 . Indeed, tumors need to establish a blood vessel supply to provide them with oxygen, nutrients and remove the waste 2 . In contrast to normal tissues, tumor vascularization exhibits a disordered labyrinth of dysfunctional and malformed vessels, resulting in structural and functional destabilization. This vascular network, characterized by high permeability and leakage, facilitates the intravasation of cancer cells and the formation of metastases. 3 In addition, recent studies have highlighted the role of abnormal tumor vasculature in shaping a protumorigenic and immunosuppressive tumor microenvironment (TME), hindering effective drug delivery and compromising therapeutic responses 4 . These novel observations provide an alternative vision to the role of tumor vasculature resulting from the excessive activity of VEGF as initially proposed by Folkman 5 . Indeed, in many cancers, targeted anti-VEGF therapy failed to reduce tumor vascularisation and improve patient outcomes 6 . Specifically, in patients with melanoma, no significant difference in distant metastases was observed 7 . Aside from angiogenesis driven by VEGF, other modes of vessel formation as well as consequences of the abnormal vessel structure may explain why anti-angiogenic therapy is not universally effective 8 . Previously, we reported the existence and activity of progenitor populations in the endothelium in human 9 and mouse 10 models allowing us to distinguish a hierarchy from endovascular progenitors (EVP) to differentiated cells (D). These populations differed functionally based on their proliferation, vessel formation and self-renewal capacity but also molecularly regarding their gene expression and surface markers. Moreover, lineage tracing experiments have shown that EVPs infiltrate tumors early on, developing vascular networks in the centre of melanoma tumors contributing to a variety of vascular beds 11 . Transcriptional profiling of EVPs versus D cells in the aorta or tumor vasculature identified Sox9 as a potential driver of EVP activity 10 – 12 . Indeed, Sox9 has been recently reported to drive EVP self-renewal and fate decision between endothelial or mesenchymal differentiation during wound healing suggesting it is an essential gene for progenitor function 13 . Moreover, S ox9 has similar important developmental roles in endocardial development of valves highlighting its ability to divert endothelial cells from their normal vascular function 14 . In the present study, we hypothesized that Sox9 activity in tumor vasculature would drive immature vessels formed by progenitors and would be responsible for the abnormal tumor vasculature. We have therefore examined the role of Sox9 in endothelial cells that contribute to melanoma vascular formation. The conditional deletion of Sox9 in the endothelium resulted in a dramatic reduction in tumor growth and metastasis through the normalization of tumor vessels, a modification of the pro-tumorigenic tumor microenvironment, and the restoration of immune cell infiltration providing a new understanding of the abnormal tumor vasculature and opening new avenues for anti-neovascularisation therapies. Results Expression of Sox9 is induced in Tumor Endothelial Cells in Human or Mouse Although the expression of transcription factor Sox9 in EVPs through bulk 10 and single-cell 12 RNA sequencing is well established, the presence and level of SOX9 protein in endothelial cells remains uncertain. B16-F0 melanoma cells were inoculated intradermally into the flank of C57BL/6J mice and resulting tumors were collected 14 days later (Fig. 1 A). Immunofluorescent staining of tumor sections confirmed the expression of SOX9-positive nuclei within the endothelial cells identified by CD31 co-staining (Fig. 1 B). No primary antibody was used as a control and hair follicles epidermal cells that express high levels of Sox9 , anatomically situated in the hair follicle bulge, were used as a positive control also ensuring the specificity of the staining (Supplemental Fig. 1). Similarly, we explored human primary cutaneous melanoma samples. Immunofluorescent staining revealed the presence of SOX9-positive cells in the endothelium of human melanoma tumor vessels (Fig. 1 C). Of note, SOX9 is known to be expressed in melanoma cells, and the nuclear staining in these tumor cells served as an internal positive control. To have a more quantitative assessment of SOX9 expression in endothelial cells, we next used Cdh5-Cre ER RosaYFP mice to genetically label all endothelial cells upon tamoxifen administration. After 5 consecutive days of tamoxifen administration, HcMel12 mCherry+ cells, a murine melanoma cell line derived from a primary melanoma tumor in Tyr::Hgf-Cdk4 R 24 C mice, were introduced via intradermal inoculation into the flank of mice 15 . Three weeks after tumor inoculation, mice were sacrificed, and tumors and aortas were collected (Fig. 1 D). Single YFP + Lineage − mCherry − endothelial cells were flow-sorted onto slides to verify SOX9 expression in the tumor and aorta (Fig. 1 E). SOX9 protein expression was commonly identified in the YFP + CD31 + endothelial cells in the tumor, confirming the expression of SOX9 in tumor endothelial cells (Fig. 1 E). Moreover, the proportion of SOX9-positive endothelial cells was significantly higher in the tumor compared to the aorta, suggesting that the tumor environment increased SOX9 expression in the endothelial cells (Fig. 1 F). Finally, to examine the relevance of these findings in other cancers, we used established and publicly available single cell RNA sequencing data of stromal cells from lung, colorectal, breast, and ovarian cancer focusing on the expression of SOX9 in tumor endothelial cells 16 . It was observed that SOX9 was upregulated in endothelial cells in tumors compared to corresponding normal tissue (Average Log2FC 3.07, P = 6.47e-09) (Fig. 1 G-H). In particular, tumor tissues exhibited a higher fraction of endothelial cells expressing SOX9 , with 1.2% in tumors versus 0.1% in adjacent normal tissues. Further stratification of endothelial cell populations indicated that SOX9 expression was predominantly localized to tip and venous endothelial cells, including high endothelial venules (HEVs). Collectively, these findings suggested that SOX9 is upregulated in a subset of tumor-associated endothelial cells in both mice and humans. This observation raised important questions about the functional role of SOX9 in tumor vascularization, particularly its influence on endothelial cell behavior within the tumor microenvironment. Melanoma-Conditioned Media Induces Sox9-Driven Progenitor Self-Renewal in ECFCs and Disrupts Tube Formation Capacity. To explore the impact of tumor-derived factors on SOX9 expression and endothelial progenitor cell function, we exposed primary human endothelial colony-forming cells (ECFCs) isolated from three donors to conditioned media (CM) from human melanoma cell lines carrying distinct mutations (WM164, MM96, DO4; see Supplemental Table 6) for five days (Fig. 2 A). Quantitative PCR analysis revealed a significant upregulation of SOX9 mRNA in ECFCs treated with WM164 CM (P = 0.021, Fig. 2 D), which was corroborated by western blot analysis showing an increase in SOX9 protein expression (P = 0.0248). A similar trend was observed when ECFCs were exposed to MM96 and DO4 CM (P = 0.07 and P = 0.0724, respectively; Fig. 2 B, C). To more closely mimic the tumor-endothelial interaction, GFP + ECFCs were co-cultured with mCherry + WM164 melanoma cells. After five days of co-culture, endothelial and melanoma cells were FACS sorted based on GFP and mCherry fluorescence respectively (Supplemental Fig. 2A, B). SOX9 expression in GFP + ECFCs was significantly upregulated following co-culture with melanoma cells (9-fold increase, P = 0.0006; Supplemental Fig. 2C), further highlighting the direct effect of tumor cells on SOX9 expression in endothelial progenitors. Next, a single-cell colony formation assay was performed to assess the functional impact of melanoma-conditioned media on ECFC self-renewal capacity. Flow-sorted ECFCs were plated 1 cell per well in endothelial growth medium (EGM2) and monitored for colony formation over 14 days. Colonies were categorized based on proliferative potential: high proliferative potential (HPP) colonies (> 500 cells), low proliferative potential (LPP) colonies (> 250–500 cells), and endothelial clusters (< 50 cells) (Fig. 2 E). Exposure to melanoma-conditioned media from WM164, MM96, and DO4 significantly increased the total proportion of colonies formed (P = 0.0058, P < 0.0001, and P < 0.0001, respectively; Fig. 2 F). ECFCs from the control group formed a higher proportion of endothelial clusters (80–90%) compared to CM-treated cells, which showed a 1.5 to 2-fold reduction (P < 0.0001 vs Control, n = 4). Interestingly, CM-treated ECFCs exhibited an increased proportion of HPP colonies, suggesting enhanced self-renewal and proliferative capabilities (P < 0.0001 vs Control ECFCs, n = 4; Fig. 2 G). To investigate the differentiation capacity of ECFCs under the influence of melanoma-conditioned media, we performed a matrigel capillary assay to assess their ability to form tube-like structures. The number of meshes, segment lengths, and nodes were significantly reduced in ECFCs treated with CM compared to the control group (P < 0.001; Fig. 2 H, I). Previous studies have shown that immature endothelial colony-forming cells (ECFCs) exhibit a reduced capacity for tube formation compared to differentiated ECFCs that have been primed with mesenchymal stem cells (MSCs) 17 . These results suggest that the upregulation of SOX9, induced by melanoma-conditioned media, was associated with an immature state of ECFCs, prioritizing self-renewal over differentiation, which inhibited their capacity to carry out specialized tasks like tube formation, characteristic of more differentiated endothelial cells. Reduced numbers but increased maturation of tumor vessels upon Sox9 deletion in the endothelium Building on these in vitro observations, we next explored the role of Sox9 in the tumor endothelium using a mouse model that enabled us to delete Sox9 specifically in endothelial cells while labelling them with YFP. We generated a conditional knock-out mouse model, Sox9 eKO (Sox9 fl/fl /Cdh5-Cre ER RosaYFP) , in comparison to Sox9 eWT (Sox9 +/+ /Cdh5-Cre ER RosaYFP) mice. To validate the efficiency of Sox9 deletion in the endothelium, we administered tamoxifen intraperitoneally for five consecutive days prior to the subcutaneous inoculation of B16-F0 melanoma cells (Supplemental Fig. 3A). Fourteen days post-inoculation, tumors were collected and YFP + CD34 + Lineage − and Lineage + cells were flow-sorted (Supplemental Fig. 3B) for genomic DNA analysis. As expected, deletion of exons 2 and 3 of the Sox9 gene was confirmed in endothelial cells from Sox9 eKO mice but not in lineage − positive cells or Sox9 eWT controls (Supplemental Fig. 3C). To further confirm the loss of Sox9 in the tumor endothelium, B16-F0 tumor sections from Sox9 eWT and Sox9 eKO mice were analysed by immunofluorescent staining. No SOX9 protein was observed in the YFP + cells in the Sox9 eKO mouse model (Supplemental Fig. 3D). Having established the above model, we next investigated the impact of Sox9 deletion on the development and structure of tumor-associated blood vessels. To evaluate changes in endothelial populations, we first conducted a flow cytometry analysis on B16-F0 primary tumors (Fig. 3 A). We applied our previously established gating strategy to identify the endothelial hierarchy, which ranges from progenitors (EVPs) to mature differentiated endothelial cells (D cells), based on CD31 expression levels, from low in EVPs to high in D cells (Fig. 3 B). Sox9 eKO tumors exhibited a significant 3.5-fold reduction in the total number of endothelial cells (YFP + CD34 + Lin − ) compared to Sox9 eWT (P = 0.0286, Fig. 3 C). This reduction was accompanied by a significant decrease in the percentage of EVPs and a concomitant increase in fully differentiated D cells (Fig. 3 D), consistent with previous reports in wounds and aorta 13 . This result suggested that Sox9 deletion in the endothelium promoted the maturation of progenitors toward a more differentiated endothelial phenotype. Immunofluorescence analysis of tumor sections further supported these findings, revealing a significant reduction in both the number and area fraction of YFP + and CD31 + vessels in Sox9 eKO tumors (Supplemental Fig. 4A, B). These observations align with the reduced endothelial cell population observed by flow cytometry, indicating that Sox9 deletion reduces overall vessel density within tumors. To assess whether this phenotype was reproducible across tumor models, we analyzed HcMel12 tumors. As in the B16-F0 model, a significant reduction in the percentage area fraction of YFP + (P = 0.015) and CD31 + (P = 0.007) vessels was observed in tumors from Sox9 eKO mice (Fig. 3 F, G). The absolute number of YFP + and CD31 + vessels also decreased significantly following Sox9 deletion (P = 0.015 and P = 0.007, respectively; Fig. 3 H). Given the reduced vessel density, we next evaluated the impact of Sox9 deletion on vessel maturation by quantifying pericyte coverage, a key indicator of vessel stability and maturation. Immunofluorescent staining for NG2 + pericytes revealed a trend towards increased area fraction covered by NG2 + vessels in Sox9 eKO tumors (P = 0.309, Fig. 3 I), although the total number of NG2 + vessels decreased in parallel with the overall reduction in vessel numbers (P = 0.055, Fig. 3 J). Importantly, the proportion of NG2 + vessels, indicative of vessels with pericyte coverage, was significantly higher in the Sox9 eKO group (P = 0.031, Fig. 3 K). This suggests that although fewer vessels were present in the absence of Sox9 , those that remained were more mature, characterized by increased pericyte coverage and thus greater vessel stability. Sox9 deletion in the endothelium reduces melanoma progression. We aimed to assess the effect of Sox9 deletion in the endothelium on melanoma outcomes. To investigate this, we used HcMel12 melanoma cells given its ability to become highly metastatic upon in vivo passaging, leading to spontaneous lung metastases in mice. This dual approach was particularly valuable, as it allowed us to delineate the distinct mechanisms underlying primary tumor development from those driving metastatic dissemination. Additionally, we reproduced the experiments using B16-F0 melanoma cells for primary tumor analysis and B16-F10 cells injected intravenously for metastasis assessment. HcMel12 tumor cells (not passaged in vivo ; P0) were inoculated intradermally into Sox9 eKO and Sox9 eWT animals. Tumor growth was monitored over 21 days, and tumors were weighed at the time of collection (Fig. 4 A). A significant reduction in tumor weight (Fig. 4 B; P = 0.0029) and volume (Fig. 4 C; P < 0.0001) was observed between Sox9 eWT and Sox9 eKO . Given the effect on the HcMel12 primary tumor progression, in vivo passaged HcMel12 were employed to assess the effect of Sox9 deletion in the endothelium on metastasis. To reflect the most common clinical scenario, the primary tumors were surgically resected, and mice were monitored for metastasis three weeks post-surgery (Fig. 4 D). Histological analysis with H&E staining was performed on lung and liver sections, revealing metastases exclusively in the lungs (Fig. 4 E). A significant reduction in the number of metastatic lung nodules was observed in Sox9 eKO mice (P < 0.035, Fig. 4 F). Notably, the size of the nodules, as reflected by their area, was also significantly decreased (P < 0.0104, Fig. 4 G). To further visualize lung metastases, micro-CT imaging was performed (Fig. 4 G and H; Sox9eWT; Video 1, Sox9eKO; Video 2), reinforcing that Sox9 deletion in the endothelium effectively reduced metastatic spread to the lungs from the primary tumor. Next, we asked if the reduction of metastasis was a direct consequence of changes to the vascularisation of the primary tumor or if it was secondary to alterations in the lung vasculature, the metastatic niche. Therefore, we used the same mouse models and injected the B16-F10 melanoma cells intravenously as an experimental model of metastasis (Supplemental Fig. 5A). This approach allowed tumor cells to disseminate through the bloodstream and directly engraft in the lung. Mice were sacrificed 10 days after tumor cell injection, and lung metastases were visible macroscopically. H&E staining of lung sections (Supplemental Fig. 5B) showed a trend towards a reduction in total metastatic area (P = 0.161, Supplemental Fig. 5C) and the number of nodules (P = 0.258, Supplemental Fig. 5D) in Sox9 eKO mice, although these differences were not statistically significant. To confirm the metastatic nature of the nodules, lung sections were stained with MITF, a melanoma marker (Supplemental Fig. 5F). Looking closer at the location of the nodules, in the proximal regions of the lungs, there was a significant reduction in both the area (P = 0.018, Supplemental Fig. 5C) and the number of nodules (P = 0.0008, Supplemental Fig. 5D), contrasted by a significant increase in nodule formation at the lung periphery. Despite these regional differences, the overall effects of Sox9 deletion in the endothelium on B16-F10 cell engraftment and lung metastasis formation were moderate, suggesting that Sox9 deletion primarily exerts its effects in the vasculature of the primary tumor, where its expression is elevated upon tumor establishment. Overall, deletion of Sox9 in the tumor vasculature had a dramatic effect on tumor outcomes and in metastatic dissemination prompting us to understand its effect at cellular and molecular level. Sox9 deletion in the endothelium results in global gene expression changes across the tumor and its microenvironment. To have a better global understanding of changes in primary tumors resulting in improved outcomes upon Sox9 endothelial deletion, we subjected HcMel12 primary tumors to spatial transcriptomics analysis. HcMel12 tumors (P0) were inoculated to Sox9 eKO and Sox9 eWT mice (n = 4 per group). On day 21, mice were sacrificed, and tumors were collected and processed for spatial transcriptomic analysis. Tissue spots with ≥ 500 genes and genes expressed in ≥ 3 tissue spots across the aggregate dataset were included for downstream analysis. Within this filtered series, the median number of genes and unique genes per spot were 2009 and 1238 respectively. Spot-level cell type annotation was achieved through robust cell type decomposition (RCTD), with reference to a public single-cell RNA dataset by Davidson et al [3] (Supplemental Figs. 6A, B). Ten cell types were identified and projected onto spot-level gene expression data at 55µm resolution (Fig. 5 A, Supplemental Figs. 6C, D). Notably, UMAP plots of tissue spots’ normalized gene expression matrices did not demonstrate spatial clustering by either cell type (Supplemental Fig. 6E) or endothelial Sox9 status (Supplemental Fig. 6F), substantiating the need for fractional cell type deconvolution through RCTD. In these murine tumor sections, melanoma cells invariably demonstrated the highest proportion in every tissue spot and were thus the primary cell type. Nonetheless, secondary cell type visualization identified a heterogeneous tumour milieu, with broad spatial variation in proportions of endothelial cells, T/B lymphocytes, natural killer cells, antigen-presenting cells, and fibroblasts (Fig. 5 A). Sample-level cell type proportions were then compared between Sox9 eKO and Sox9 eWT tumors (Fig. 5 B). Relative to Sox9 eWT , Sox9 eKO tumors demonstrated a trend towards reduced endothelial cells (P = 0.067), but higher levels of T cells (P = 0.172) and macrophages (P = 0.357). To further explore the behaviour of Sox9 eKO versus Sox9 eWT tumors, we applied cell type-weighted gene set variation analysis to identify between-group differences in composite gene pathways for melanoma cells, endothelial cells, and lymphocytes. As compared to melanoma cells in Sox9 eWT tumors, melanoma cells in Sox9 eKO tumors demonstrated reduced expression of genes relating to the hypoxia, glycolysis, TGFβ signalling and epithelial-mesenchymal transition (EMT) pathways (adjusted P = 0.0373 for all; Fig. 5 C). Similarly, endothelial cells in Sox9 eKO tumors demonstrated downregulation of genes corresponding to fatty acid oxidation (Acetyl-CoA thiolase), EMT, TGFβ signalling, vascular morphogenesis but also hypoxia, and glycolysis (adjusted P = 0.0406 for all) (Fig. 5 D). On ligand-receptor interaction analysis through stLearn 18 we identified a number of important changes in endothelium-enriched tissue spots (Fig. 5 F). In Sox9 eWT tumors, Hepatocyte Growth Factor (HGF) produced by tumor cells could be responsible for the abnormal tumor endothelial phenotype by acting on endothelial MET receptors (Fig. 5 G). In contrast, endothelial cells in Sox9 eKO tumors were enriched for interactions between the platelet-derived growth factor beta (PDGFβ) and platelet-derived growth factor receptor beta (PDGFRβ) in fibroblasts/pericytes suggesting better pericyte coverage of blood vessel. Similarly, there was increased interaction between the intercellular adhesion molecule 2 (ICAM2) on endothelial cells and integrin alpha M (ITGAM) on immune cells (Fig. 5 H). These observations collectively indicate that deletion of Sox9 in the endothelium induces major and broad changes in the tumor environment. Reduction in Glycolysis and Hypoxia in melanoma primary tumors after deletion of Sox9 in the endothelium In addition to the observed changes in the tumor vasculature, the spatial transcriptomics analysis suggested gene expression changes in tumor cells. To validate these findings, bulk RNA sequencing was performed on mcherry-positive HcMel12 melanoma cells (non in vivo passaged; P0) that were FACS-sorted from Sox9 eKO versus Sox9 eWT tumors (n = 5 mice per group, Supplemental Fig. 7). We identified 323 differentially expressed (DE) genes with an unadjusted p value 0.5; Red), and 222 downregulated in Sox9 eKO mice (logFC < -0.5; Blue) (Fig. 6 A). Gene Set Enrichment Analysis (GSEA) utilizing the MsigDB mouse gene set collections revealed several significantly enriched and under-enriched gene sets associated with critical biological pathways in Sox9 eKO melanoma tumor cells (Fig. 6 B). Notably, the analysis identified a significant downregulation of several pathways, including Glycolysis (248 genes; P = 2.75e-04), Hypoxia (256 genes; P = 1.37e-09), MTORC1 (279 genes; P = 1.75e-05), and Cholesterol Homeostasis (94 genes; P = 7.65e-05). Conversely, G2M checkpoint (285 genes; P = 0.0015), mitotic spindle (312 genes; P = 0.007), and E2F targets (262 genes; P = 0.008) were found to be enriched. Among the top three most differentially downregulated pathways in the Sox9 eKO (Fig. 6 C), Hypoxia (NES − 2.8, FDR 6.8e-08), Glycolysis (NES − 2.25, FDR 3.4e-03), and mTORC1(NES-2.29, FDR 4.3e-04) signalling emerged as particularly relevant, given their well-established roles in mediating the tumor's metabolic adaptation to hypoxic conditions. The observed alterations in these pathways in our model may be directly linked to the vascular changes induced by Sox9 deletion. Hypoxia is a hallmark of the tumor microenvironment, arising from the imbalance between the rapid proliferation of tumor cells and inadequate vascular supply. Under hypoxic conditions, cancer cells activate an adaptive response mediated primarily by Hypoxia-Inducible Factor 1 (HIF1α). This transcription factor drives the upregulation of glycolytic enzymes, including GLUT1 (Glucose Transporter 1), a key regulator of glycolysis in cancer. GLUT1 facilitates increased glucose uptake into tumor cells, enabling them to maintain high glycolytic activity, which in turn supports their accelerated proliferative capacity and enhanced survival under oxygen-limited conditions. To confirm the downregulation of the hypoxia and glycolysis pathways identified in tumor cells from both the spatial transcriptomics and the bulk RNA sequencing data, we assessed the protein expression levels of GLUT1 and HIF1α through immunofluorescence staining (Fig. 6 D; F). A significant reduction in the intensity and area of both markers was observed (P = 0.0411 and 0.0043 respectively, Fig. 6 E; G), indicating decreased activation of these metabolic pathways in Sox9 eKO tumors. These findings, combined with the observed reduction in vessel number along with their maturation, suggested improved tumor perfusion and vessel normalization. Restoration of immune infiltration in primary tumors after deletion of Sox9 in the endothelium Ultimately, the spatial transcriptomics data indicated an increase in immune cell infiltration following the deletion of Sox9 in the endothelium. To validate this finding, tumor sections from the Sox9 eWT and Sox9 eKO HcMel12 tumors were stained with CD4 and CD8 antibodies to assess T lymphocyte infiltration (Figs. 7 A;B). Immunofluorescence staining revealed a significant two-fold increase in CD4 + cell infiltration in the Sox9 eKO tumors compared to the controls (P = 0.0411; Fig. 7 C). Notably, when comparing the core and periphery of the tumors, significant changes were observed exclusively in the core region (Figs. 7 D;E), which exhibited a complete absence of visible CD4 + cells in the Sox9 eWT group. This pattern was particularly pronounced for CD8 staining, which also demonstrated a significant two-fold increase in infiltration in the core of the Sox9 eKO tumors (Figs. 7 F-H). Discussion Tumor vascularisation is one of the hallmarks of cancer and an imperative step for tumor progression and metastasis 19 , 20 . The tumor vasculature is poorly formed and destabilised which leads to cancer cell intravasation, migration, and metastasis 21 and for many cancers anti-angiogenic drugs have only partially addressed this issue 22 . A better understanding of the origins of vessels in the tumor and the role of endothelial cells in the tumor microenvironment may therefore lead to better therapeutic options for patients. Our lab demonstrated that endovascular progenitor cells (EVPs) which reside in the blood vessel walls, contribute to the formation of a de novo blood vessel network in the tumor. Initial gene expression characteristic of the EVP population identified transcription factor Sox9 as an essential marker distinguishing tissue resident murine endothelial progenitor cells from mature endothelial cells in various tissue beds 10 including tumors 11 . The present study revealed that SOX9 expression is augmented in endothelial cells exposed to a tumor environment both in vivo and in vitro , in mice and human cells. This was done mostly in melanoma but also identified in other tumour types such as human colorectal, lung, ovarian, and breast cancer samples. In all these examples the expression of SOX9 was increased in tumour endothelium compared to the endothelial cells of normal tissue, suggesting that SOX9 fulfils a significant function in tumor endothelial cells. Indeed, upon conditional deletion of Sox9 in the endothelium, there was a reduction in tumour vessel numbers and surface area in parallel with an increase in pericyte coverage suggesting an increase in maturation. This was paralleled with a reduction in hypoxia, glycolysis and an increase in immune cell infiltration leading to reduced tumor size and metastasis. These findings support an essential role of Sox9 in the development of the abnormal tumour vasculature and its deletion as an essential step in vessel normalisation. One possible explanation for the decrease in metastasis from the primary tumor could be the reduction in the total number of endothelial cells and vessels in the centre of the primary tumor. By providing fewer escape routes to tumor cells, metastasis may be reduced 23 . An alternative possibility, however, is that vessel normalisation improves tumor oxygenation, increases immune cell infiltration, and reduces cancer cell intravasation. Vessel leakiness reduces blood flow through an increase in intratumoral pressure, causing hypoxia in the tumor microenvironment and surrounding tissue. In return, hypoxia promotes metastasis of melanoma cells by downregulating melanocyte differentiation markers and enhancing their invasion in a HIF-1α dependent manner. The activation of Snail and Twist by HIF-1α triggers EMT, promoting increased metastatic potential. Furthermore, HIF-1α facilitates the migration of cancer cells by regulating the expression of collagen fibres, integrins and deposition of ECM 24 . Moreover, low oxygen levels in the tumor microenvironment promote the induction of pro-tumor regulatory T cells while hampering the function of anti-tumor CD8 + T cells 25 . Additionally, hypoxia leads to higher expression of programmed death-ligand 1 (PD-L1) by myeloid-derived suppressor cells, which hinders the function of immune cells 24 . Many aspects of our observations were in line with vessel normalisation as described in these previous studies. The reduction in the total number of endothelial cells and blood vessels triggered by the deletion of Sox9 was accompanied by increased pericytes coverage, confirmed by NG2 immunofluorescent staining. Pericytes play a pivotal role in maintaining the stability of blood vessels by covering endothelial cells and forming a supportive and protective network. This strategic positioning of pericytes along the vascular wall prevents the abnormal leakage of vessels 26 . For instance, in aged mice, microvasculature is deprived of a supportive structure, pericytes coverage, and adherent junctions, resulting in leaky vessels 27 . Coutelle and colleagues showed that targeting both ANG2 and VEGF in a xenograft model of colorectal cancer cells led to improvement in pericyte coverage, vascular integrity, and leaky vessels 28 . Moreover, the study by Keskin et al. revealed that depleting pericytes and targeting NG2 signalling repaired vascular stability in various model systems, reducing tumor growth and metastasis 29 . Conversely, other studies showed that pericyte depletion was linked to increased hypoxia, EMT, and metastasis in breast cancer 30 . There is ongoing debate regarding whether leaky blood vessels have advantages 31 over normalised vessels. Indeed tumor cells proliferate rapidly, and the surrounding tissue limits tumor development, suppressing blood flow and making the vessels leaky. In addition, insufficient lymphatic drainage increases fluid flux and elevates interstitial fluid pressure, which contributes to enhanced vessel permeability, disorganised vascular pressure, and intravasation of cancer cells. This restricts drug delivery, implying normalisation might be advantageous over leaky vessels 32 . Our study consistently demonstrated the reduction of hypoxia and glycolysis upon Sox9 conditional deletion using a variety of techniques and platforms such as immunostaining, bulk RNA sequencing and spatial transcriptomics. These studies further support the role of Sox9 in driving the abnormal vessel structure as its deletion may restore normalised and mature vessels and reduce hypoxia even when the tumour has a larger size, 3 weeks after inoculation. Studies have shown that inhibition of glycolysis suppressed melanoma cell proliferation by inducing apoptosis 33 34 . The observation of smaller tumors in the Sox9 eKO group aligns with these findings. Moreover, exposure to a hypoxic environment prompting HIF-1α activation suppresses melanocytic markers and enhances the invasive potential of melanoma cells 35 . More importantly, our spatial transcriptomics findings suggested that the normalisation of vessels through the deletion of Sox9 in endothelial cells was accompanied by profound changes in the tumor microenvironment. This involved a change in immune cell infiltration that was particularly dramatic in the centre of tumors. As indicated by ligand-receptor interactions, this could be part of the tumour vessel normalisation process, providing better expression of adhesion molecules on endothelial cells that allow the initiation of the various steps leading to the entry of immune cells in the tumour tissue. Alternatively, the reduction of hypoxia and glycolysis may lead in significant reduction of byproducts of glycolysis such as lactate accumulation which has been described to dampen immune responses 36 . Finally, despite the lack of clear indication in our spatial transcriptomics findings, one cannot exclude the production of immunomodulatory mediators by endothelial cells upon SOX9 expression. Indeed, PDL1 expression on the endothelium has been reported and can affect T cell effector functions 37 . The direct immunomodulatory function of the endothelium in this context warrants further investigation. Most of our insight into the developmental functions of Sox9 has originated from studies involving the male reproduction system, chondrogenesis, and heart valve development. Recent studies revealed the role of SOX9 in endothelial to mesenchymal transition by analysis of single-cell chromatin in HUVECs 38 and wound healing assay 13 . In these previous studies, Sox9 has distinct roles in self-renewal and stem cell fate decision favouring a mesenchymal fate 39 . However, there are no studies on the role of Sox9 in endothelial cells in tumor vascularisation and metastasis. despite our findings of increased self-renewal in vitro , it is also questionable if Sox9 , in this new context of tumor vascularisation, is a strict marker for endothelial progenitors or EVPs. Other efforts to reduce progenitor function have resulted in similar results. Donovan et al. showed that conditional depletion of Rbpj , a transcription factor highly upregulated in EVPs, resulted in the depletion of EVPs and significantly reduced metastases from primary tumors. Similarly to Sox9 , Rbpj depletion has major implications in a range of target genes beyond its importance for progenitor function. Overall, it is difficult to completely dissociate the impact of such genes on progenitor function as opposed to their alternative effects on the endothelium. In conclusion, as in many other situations, Sox9 expression is upregulated in the context of tumor endothelium where it drives a developmental program including an increase in progenitor numbers and activity, resulting in immature vessels that promote hypoxia, tumor growth and metastasis. We have argued and demonstrated that its depletion normalises vessels despite reducing their numbers, resulting in less hypoxia, better tumor immune infiltration, drastically changing the tumour microenvironment resulting in less metastases (Fig. 8 ). Adoption of such developmental program in the cancer context is not novel and further emphasises the need for development of targeted therapies. Material and Method Animal Model and Ethics All animal procedures were conducted in accordance with the University of Queensland Animal Ethics Committee (AEC) guidelines and approvals (#2022/AE000150 and #2022/AE000335). Both male and female mice, aged 10–14 weeks, were included in this study. Endothelial-specific Sox9 gene knockout was achieved by breeding Sox9 fl/fl mice, which had been backcrossed to a C57BL/6J background, with Cdh5Cre ERt2 /Rosa-YFP mice, resulting in the generation of Sox9 fl/fl/ Cdh5Cre ERt2 /Rosa-YFP ( Sox9 eKO ) triple-transgenic mice. Cdh5Cre ERt2 /Rosa-YFP ( Sox9 eWT ) mice were used as controls. Wild-type C57BL/6J mice (WT) were sourced from the Animal Resources Centre (Perth, Western Australia). For conditional knockout of Sox9 , tamoxifen (Sigma-Aldrich, USA) was administered to activate Cre recombinase. Tamoxifen was prepared by dissolving it in a solution of 10% ethanol and 90% corn oil, achieving a final concentration of 20 mg/ml. Each mouse received an intraperitoneal injection of 2 mg of tamoxifen (100 µl) every 24 hours for up to five consecutive days prior to melanoma cell inoculation. Tumor Cell Culture and innoculation B16-F0 murine melanoma cells were cultured in RPMI 1640 medium (Gibco, ThermoFisher Scientific, USA) supplemented with 10% fetal bovine serum (FBS). HcMel12 and HcMel12 mCherry+ murine melanoma cells were maintained in RPMI 1640 medium with 10% FBS, 2 mM l-glutamine, 1 mM HEPES, and 10 mM non-essential amino acids (Gibco, ThermoFisher Scientific, USA). For subcutaneous cell injection, mice were anesthetized with isoflurane, and hair was removed using an electric clipper. Suspensions of 5×10 5 B16-F0 cells or 1×10 6 HcMel12 and HcMel12 mCherry+ cells in 100 µl saline were injected subcutaneously. Serial Tumor Transplantation HcMel12 tumors develop spontaneous lung metastasis following repeated subcutaneous transplantation in C57BL/6J mice. For the initial transplantation, 1×10 6 cultured HcMel12 cells were injected subcutaneously into the flank of C57BL/6J mice. Tumor growth was monitored through visual inspection and palpation, with measurements taken using a sliding vernier caliper. Tumors were excised 21 days post-inoculation and mechanically dissociated using scissors. Cells were filtered through a 70 µm cell strainer (BD Biosciences, USA), washed with PBS, and 1×10 6 HcMel12 cells in 100 µL saline were injected subcutaneously into new C57BL/6J mice. This process was repeated over five passages until lung metastases were observed macroscopically. Resection of HcMel12 Tumors Subcutaneous HcMel12 tumors were surgically resected 21 days post-inoculation. Mice were placed on a heat mat and anesthetized with 2% isoflurane via a nose cone. The skin around the surgical site was cleansed with a cotton swab and a topical antiseptic agent (70% ethanol or chlorhexidine). For the procedure, mice were positioned in lateral recumbency, and the tumor was gently pinched with tweezers and excised using sterile dissecting scissors. The incision site was closed with N604 sutures. Following the procedure, tumor weight was recorded on a scale, and tumor volume was measured with a sliding vernier caliper by recording diameters along two axes. Tumor volume was calculated as V = L×W 2 , where V is the tumor volume, L is the tumor length, and W is the tumor width. Tissue Processing of Murine Tumors Following University of Queensland ethical guidelines, mice were sacrificed, and melanoma tumors along with other tissues were collected. Tumors were fixed in 4% paraformaldehyde (PFA) for 2 hours at room temperature, then washed twice with 1×PBS (Gibco, ThermoFisher Scientific, USA) and sequentially immersed in 10%, 20%, and 30% sucrose (Chem-Supply, AUS) for cryoprotection. Fixed samples were embedded in OCT compound (Sakura Finetek, USA) and stored at -80˚C for long-term storage. Tumors were sectioned at 8 µm thickness on a Cryostat Leica CM1950, with sections collected on Superfrost Plus slides (Thermo Scientific, Germany) and spaced 30 µm apart. Hematoxylin and Eosin (H&E) staining For H&E staining, lungs were placed in ProSciTech histology cassettes and stored in 70% ethanol. Tissues were embedded in paraffin and were sectioned 100 µm apart from each other. The Translational Research Institute core facility carried out the H&E staining. Images were taken with an Olympus Slide Scanner VS120 microscope (20x objective). The area and number of nodules were analysed using Olympus OlyVIA software (Version 3.1). Immunofluorescence - mouse Cryosections were permeabilized with 0.5% Triton X-100 (Chem Supply, Australia) and blocked with 20% normal goat serum. Primary antibodies used in this study for mouse tissue sections included anti-CD31, anti-SOX9, anti-GFP, anti-CD34, anti-GLUT1, and anti-HIF1alpha (Supplemental Table 1). Unbound antibodies were removed by washing sections three times for 5 minutes in 1× PBS/0.1% Tween-20 (Amresco, USA). Secondary antibodies conjugated with Alexa Fluor 488, 568, and 647 (Invitrogen, USA) were applied for fluorescence detection, with sections incubated at room temperature for 40 minutes. Following incubation, sections were washed three times with PBS/0.1% Tween-20. Nuclear staining was achieved using ProLong® Gold mounting medium with DAPI (Invitrogen, USA). Microscopy All confocal images were acquired at the Translational Research Institute Microscopy facility. High-magnification images for quantification were captured using an Olympus FV3000 Laser Scanning Confocal Microscope and a Nikon Spinning Disk Confocal Microscope (both Tokyo, Japan). Images were obtained at 10×, 20×, 40×, and 60× magnifications. Blood Vessels density Tumor sections were stained with CD31, NG2, and anti-GFP (for endogenous YFP). Vessel counts and area fractions in tumors were quantified using NIS-Element AR analysis software (Version 5.01.0064). Look-Up Tables (LUTs) were standardized across samples and compared with no-secondary control samples to prevent false measurements. Statistical Analysis Statistical analyses were conducted using GraphPad Prism v8 software. Non-parametric data were analysed using the Mann-Whitney test, and parametric data were analysed using 2-way ANOVA with Bonferroni correction. A p-value < 0.05 was used to determine statistical significance. Flow Cytometry All flow cytometry was conducted at the Translational Research Institute Flow Cytometry Facility. Tissues were excised, dissociated, and transferred to pre-warmed Hanks’ Balanced Salt Solution (HBSS, Gibco, ThermoFisher Scientific, USA) for digestion with 1 mg/ml collagenase I, 1 mg/ml dispase (both from Gibco, ThermoFisher Scientific, USA), and 150 U/ml DNase-I (Sigma-Aldrich, USA) in a 37˚C water bath for 30 minutes. The dissociated tissue was filtered through a 70 µm nylon strainer using a flat plunger, washed with PBS, and pelleted at 380 rcf for 5 minutes. The pellet was treated with a hypertonic or RBC lysis buffer for 2 minutes to remove erythrocytes, then diluted with PBS and centrifuged again. After discarding the supernatant, cells were prepared in unstained, single-color, and Fluorescence-Minus-One (FMO) controls for gating. Each pellet was resuspended in FACS buffer (0.5% BSA, 2 mM EDTA, 1×PBS) and stained for 30 minutes at 4°C. Antibodies are listed in Supplemental Tables 3 and 4. Cytospin of FACS-sorted cells Cells were FACS sorted directly into FBS, washed with 1×PBS, and resuspended in 150µL of PBS. A Cytospin™ 4 Cytocentrifuge was employed to distribute cells in suspension onto SuperFrost Plus™ slides. Adhered cells were fixed with ice-cold 2% PFA for 10 minutes on ice, and the standard immunofluorescent staining protocol was carried out. gDNA Extraction and PCR Genomic DNA was extracted from sorted YFP + CD34 + Lin − endothelial cells from the B16-F0 tumor to validate the conditional deletion of Sox9 in endothelial cells. Following the manufacturer’s specifications (Qiagen, USA), the QIAamp DNA Mini Kit was utilised. The polymerase chain reaction was performed with 100 ng of gDNA. The PCR products were run on a 1.5% agarose gel mixed with ethidium bromide. The gel was imaged using a Gel Doc XR + imaging system (Bio-Rad Laboratories, USA). Primers sequence are listed in Supplemental Table 5. RNA Sequencing and Data Analysis RNA was extracted from the FACS-sorted HcMel12 mCherry+ cells using a QIAGEN mini kit (Qiagen, USA) following manufacturer instructions. RNA sequencing was conducted by the Institute for Molecular Bioscience (IMB, University of Queensland). The total RNA library was prepared with TruSeq standard mRNA kit (Illumina, USA) to be used in the Sanger/Illumina 1.9. RNA sequencing analysis was conducted on the Galaxy Australia server by QFAB. QFAB performs initial QC, reads mapped to a reference genome and read counts for the annotated genes. Reads (FASTQ files) were trimmed with Trimmomatic, mapped to a reference genome assembly with HiSAT2, and reads were counted against the GenCode gene annotation with featureCounts. Read count normalisation and differential gene expression analysis were performed in the lab using DESeq2 Galaxy Version 1.22.3. Data processing and differential expression analysis across different samples were performed using the standard Limma-Voom pipeline 40 . Gene set enrichment analysis was performed using the fgsea package). Hallmark gene sets were downloaded from the Molecular Signatures Database (MSigDB) 41 . Normalized enrichment score (NES) and adjusted p value were calculated for each hallmark gene set. Human Cell Culture Human placenta foetal ECFCs were isolated according to our previously published protocol 42 . The human ethics boards of The University of Queensland and the Royal Brisbane and Women’s Hospital (#HREC/09/QRBW/14) granted the use of human tissue. ECFCs were cultured on rat tail collagen (Gibco, ThermoFisher Scientific, USA) coated tissue-culture T75 flasks in Endothelial Growth Medium (EGM-2, Lonza Group, Switzerland). For conditioned media experiments, WM164, D04 and MM96 human melanoma cells (Supplemental Table 6) were seeded at 40% confluency with RPMI (Gibco, ThermoFisher Scientific, USA) and 2% FBS for 72 hours. CM was collected and centrifuged for 10 minutes at 1000 g (4 ˚C) and filtered with a 0.45 µm filter (Merck Millipore). ECFCs were treated with conditioned media for 5 consecutive days. RNA Extraction, cDNA synthesis and qPCR RNA was extracted from cells using a QIAGEN mini kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions. RNA quality and concentration were assessed using A260nm/A280nm spectroscopy on the Nanodrop ND-1000 (Thermo-scientific, Langenselbold, Germany). 5-100ng of RNA was used for cDNA synthesis using the Invitrogen SuperScript III First Strand Kit (Life Technologies, USA). To quantify relative changes in key gene expression, Real Time-Quantitative Polymerase Chain Reaction (qPCR) was conducted on cDNA using SYBR Green Master Mix reagent (Applied Biosystems, United Kingdom). Reactions were run in triplicates on a QuantStudio Flex 7 (ThermoFisher Scientific, USA). Fold change of gene expression was determined using the delta delta Ct method and normalized to the housekeeper gene. Primer sequences are listed in Supplemental Table 7. Western Blot Cells were washed with 1×PBS and lysed in RIPA buffer with 1× Complete Protease Inhibitor (Sigma-Aldrich, USA), followed by sonication. Protein concentration was determined using a BCA assay (ThermoFisher, USA). Equal amounts of protein were mixed with Laemmli buffer, boiled at 100°C for 5 minutes, and loaded onto an SDS-PAGE gel. Proteins were transferred to a membrane, blocked overnight at 4°C, and probed with rabbit anti-SOX9 (1:1000, Merck Millipore, USA) and mouse anti-β-actin (1:5000, Sigma-Aldrich, USA). Detection was performed using IRDye® 800CW goat anti-rabbit and IRDye® 680RN goat anti-mouse secondary antibodies (1:2500, LI-COR Biosciences, USA) on an Odyssey CLx imaging system (LI-COR Biosciences, USA). Spatial Transcriptomics Sample preparation Animals were euthanized according to University of Queensland Ethics guidelines. Tissues were snap-frozen in Tissue Tek OCT compound (Sakura Finestek, USA) using isopentane and liquid nitrogen, stored at -80°C. For sectioning, samples were equilibrated at -15°C in a cryostat, sectioned at 8 µm, placed on dry ice, and stored at -80°C. Visium Gene Expression Procedure Tissues were fixed, and stained with hematoxylin and eosin. Permeabilization was performed for 25 minutes, followed by cDNA synthesis (reverse transcription, second-strand synthesis, release), qPCR to determine cycle number, and cDNA clean-up using SPRIselect. cDNA quality was assessed using bioanalyzer trace and quantification. Fragmentation, end repair, A-tailing, and SPRIselect were performed according to the 10× Genomics protocol. The final product was sequenced at the Institute of Molecular Bioscience sequencing facility. Data preparation Base call sequence files were aligned to the mm10 (mouse) 43 reference assembly using the Space Ranger software package (v3.0; 10x Genomics, Pleasanton). Count matrices were aggregated as a single object in R 4.4.1 through the Seurat package (5.1.0) 44 and spatially oriented to 55µm tissue spots spaced 100µm apart in a hexagonal configuration. Tissue spots with < 500 transcripts and genes present in less than < 3 tissue spots were excluded from downstream analyses. For some analyses, the SCTransform (SCT) function of Seurat was applied to perform variance-stabilizing transformation of sample-level count matrices, using default parameters (variable.features = 3000, residual.features = NULL). Other analyses performed independent normalization procedures and therefore employed raw count matrices; these are specified when they occur. Cell type composition Under the Visium platform, each tissue spot can contain over 25 distinct cells, and standard clustering methods are unsuited to cell type classification. Therefore, we applied cell type deconvolution with reference to a publicly available scRNA murine melanoma dataset by Davidson et al [ArrayExpress: E-MTAB-7427] 45 . To ensure concordance with our samples, this scRNA dataset was restricted to primary tumour cells only, and the original series of 15 cell type clusters was collapsed into 9 (Supplemental Figs. 6A, B). An additional melanoma cell cluster was developed by performing gene set variation analysis (GSVA) 46 on a standard series of melanocyte markers (Pmel, Mitf, Mlana, Sox10, Tyrp1, Dct) 47 , 48 , then including previously-unclassified cells with a GSVA z-score of ≥ 2. Spot-level cell type annotation was subsequently performed through robust cell type deconvolution (RCTD) 49 on spot-level raw count matrices, which returned a proportion vector for each tissue spot and cell type. The primary and secondary cell types of each tissue spot were overlaid on the original H&E image through proprietary R code. Differences in cell type proportions between Sox9eKO and Sox9eWT samples were assessed using the Wilcoxon rank-sum test with Benjamini-Hochberg correction at a false discovery rate (FDR) of 0.05. Differential gene expression Differential expression (DE) of captured transcripts was evaluated with respect to specific cell types. To achieve this, spot-level raw count matrices were first weighted by their corresponding cell type proportions. These were pseudobulked into a cell type-specific count matrix for each sample, then normalized for library size and gene-specific variance through the limma-voom model 50 . Cell type-specific DE between Sox9eKO and Sox9eWT samples was evaluated using empirical Bayes-moderated t-testing 51 with Benjamini-Hochberg correction (FDR 0.05). Gene pathway analyses Gene pathway analyses were performed using GSVA on curated murine gene sets from the Molecular Signatures Database (MSigDB) 52 , including the Hallmark 41 and Gene Ontology 53 collections. Using SCT count matrices, cell type-agnostic GSVA was performed on all tissue spots. Sample-level cell type-specific GSVA measures were obtained by weighting the GSVA score for each spot by the corresponding cell type proportion, then pseudobulking these across the sample. Between-group differences were assessed using the Wilcoxon rank-sum test with Benjamini-Hochberg correction (FDR 0.05). Ligand-receptor analyses Ligand-receptor (LR) interactions were performed using the stLearn package (0.4.12) for Python 3.8.19 18 , with reference to the ConnectomeDB LR database 54 . stLearn was applied to raw spot-level count matrices to quantify LR interactions for each spot and its six immediate hexagonal neighbours. Spot-level LR scores with significant enrichment over background pairs (p < 0.05) were pseudobulked into a sample-level LR matrix; this was conducted both overall, as well as with weighting by cell type. Between-group differences were then normalized using the limma-voom pipeline and statistically evaluated through empirical Bayes-moderated t-testing with Benjamini-Hochberg 51 correction (FDR 0.05). LR pairs with the lowest adjusted p-value for between-group differences were visualized in chord diagrams through the circlize package 55 . Declarations Data Availability RNA-Sequencing and Spatial transcriptomic data were deposited into the Gene Expression Omnibus database under accession number GSE277775 (Spatial transcriptomic) and GSE278207 (RNA-seq) and are available at the following URL: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE277775 https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE278207 Acknowledgements This research was carried out at the Translational Research Institute, Woolloongabba, QLD 4102, Australia. The Translational Research Institute is supported by a grant from the Australian Government. Author Contributions G.H. and L.S. executed most experiments, evaluated results and wrote the manuscript. K.K. and L.S. conceived and designed the study. C.Z. performed analysis of the public scRNA-seq dataset. S.T. analyzed the Spatial Transcriptomics data, quantified CD4 and CD8 T cell infiltration in tumor immunofluorescent stainings and processed the RNA-seq data. H.L. contributed to in vitro experiments. K.C.L. performed immunofluorescent staining of CD4 and CD8 T cells. Q.N. provided expertise on Spatial Transcriptomics. J.D. contributed to in vivo experiments. E.R. reviewed the manuscript and figures. K.K. provided overall supervision and mentorship throughout the project, guided the interpretation of results, and contributed to the critical revision of the manuscript. All co-authors have reviewed and approved this manuscript for publication. Competing Interests statement All authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript. References Emmett MS, Dewing D, Pritchard-Jones RO (2011) Angiogenesis and melanoma - from basic science to clinical trials. Am J Cancer Res 1:852–868 Potente M, Gerhardt H, Carmeliet P (2011) Basic and therapeutic aspects of angiogenesis. Cell 146:873–887. 10.1016/j.cell.2011.08.039 Jain RK (2005) Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science 307:58–62. 10.1126/science.1104819 Schaaf MB, Garg AD, Agostinis P (2018) Defining the role of the tumor vasculature in antitumor immunity and immunotherapy. Cell Death Dis 9:115. 10.1038/s41419-017-0061-0 Folkman J (1971) Tumor angiogenesis: therapeutic implications. N Engl J Med 285:1182–1186. 10.1056/NEJM197111182852108 Bergers G, Hanahan D (2008) Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer 8:592–603. 10.1038/nrc2442 Corrie PG et al (2014) Adjuvant bevacizumab in patients with melanoma at high risk of recurrence (AVAST-M): preplanned interim results from a multicentre, open-label, randomised controlled phase 3 study. Lancet Oncol 15:620–630 Hashemi G, Dight J, Khosrotehrani K, Sormani L (2022) Melanoma Tumour Vascularization and Tissue-Resident Endothelial Progenitor Cells. Cancers (Basel) 14. 10.3390/cancers14174216 Shafiee A, Patel J, Hutmacher DW, Fisk NM, Khosrotehrani K (2018) Meso-Endothelial Bipotent Progenitors from Human Placenta Display Distinct Molecular and Cellular Identity. Stem Cell Rep 10:890–904. 10.1016/j.stemcr.2018.01.011 Patel J et al (2017) Functional Definition of Progenitors Versus Mature Endothelial Cells Reveals Key SoxF-Dependent Differentiation Process. Circulation 135:786–805. 10.1161/CIRCULATIONAHA.116.024754 Donovan P et al (2019) Endovascular progenitors infiltrate melanomas and differentiate towards a variety of vascular beds promoting tumor metastasis. Nat Commun 10:18. 10.1038/s41467-018-07961-w Lukowski SW et al (2019) Single-Cell Transcriptional Profiling of Aortic Endothelium Identifies a Hierarchy from Endovascular Progenitors to Differentiated Cells. Cell Rep 27, 2748–2758 e2743, 10.1016/j.celrep.2019.04.102 Zhao J et al (2021) Sox9 and Rbpj differentially regulate endothelial to mesenchymal transition and wound scarring in murine endovascular progenitors. Nat Commun 12:2564. 10.1038/s41467-021-22717-9 Akiyama H et al (2004) Essential role of Sox9 in the pathway that controls formation of cardiac valves and septa. Proc Natl Acad Sci U S A 101:6502–6507. 10.1073/pnas.0401711101 Bald T et al (2014) Ultraviolet-radiation-induced inflammation promotes angiotropism and metastasis in melanoma. Nature 507:109–113. 10.1038/nature13111 Wang W, Wang L, She J, Zhu J (2021) Examining heterogeneity of stromal cells in tumor microenvironment based on pan-cancer single-cell RNA sequencing data. Cancer Biol Med 19:30–42. 10.20892/j.issn.2095-3941.2020.0762 Shafiee A et al (2017) Priming of endothelial colony-forming cells in a mesenchymal niche improves engraftment and vasculogenic potential by initiating mesenchymal transition orchestrated by NOTCH signaling. FASEB J 31:610–624. 10.1096/fj.201600937 Pham D et al (2023) Robust mapping of spatiotemporal trajectories and cell-cell interactions in healthy and diseased tissues. Nat Commun 14:7739. 10.1038/s41467-023-43120-6 Hanahan D, Folkman J (1996) Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86:353–364. 10.1016/s0092-8674(00)80108-7 Hanahan D, Weinberg RA (2011) Hallmarks of cancer: the next generation. Cell 144:646–674. 10.1016/j.cell.2011.02.013 Nagy JA, Chang SH, Dvorak AM, Dvorak HF (2009) Why are tumour blood vessels abnormal and why is it important to know? Br J Cancer 100:865–869. 10.1038/sj.bjc.6604929 Simon T, Gagliano T, Giamas G (2017) Direct Effects of Anti-Angiogenic Therapies on Tumor Cells: VEGF Signaling. Trends Mol Med 23:282–292. 10.1016/j.molmed.2017.01.002 Duong T et al (2012) Genetic ablation of SOX18 function suppresses tumor lymphangiogenesis and metastasis of melanoma in mice. Cancer Res 72:3105–3114. 10.1158/0008-5472.CAN-11-4026 Martin JD, Fukumura D, Duda DG, Boucher Y, Jain RK (2016) Reengineering the Tumor Microenvironment to Alleviate Hypoxia and Overcome Cancer Heterogeneity. Cold Spring Harb Perspect Med 6. 10.1101/cshperspect.a027094 Martin JD, Seano G, Jain RK (2019) Normalizing Function of Tumor Vessels: Progress, Opportunities, and Challenges. Annu Rev Physiol 81:505–534. 10.1146/annurev-physiol-020518-114700 Gerhardt H, Semb H (2008) Pericytes: gatekeepers in tumour cell metastasis? J Mol Med (Berl) 86:135–144. 10.1007/s00109-007-0258-2 Jeong JH et al (2017) Microvasculature remodeling in the mouse lower gut during inflammaging. Sci Rep 7:39848. 10.1038/srep39848 Coutelle O et al (2015) Dual targeting of Angiopoetin-2 and VEGF potentiates effective vascular normalisation without inducing empty basement membrane sleeves in xenograft tumours. Br J Cancer 112:495–503. 10.1038/bjc.2014.629 Keskin D et al (2015) Targeting vascular pericytes in hypoxic tumors increases lung metastasis via angiopoietin-2. Cell Rep 10:1066–1081. 10.1016/j.celrep.2015.01.035 Cooke VG et al (2012) Pericyte depletion results in hypoxia-associated epithelial-to-mesenchymal transition and metastasis mediated by met signaling pathway. Cancer Cell 21:66–81. 10.1016/j.ccr.2011.11.024 Sounni NE et al (2010) Stromal regulation of vessel stability by MMP14 and TGFbeta. Dis Model Mech 3:317–332. 10.1242/dmm.003863 Viallard C, Larrivee B (2017) Tumor angiogenesis and vascular normalization: alternative therapeutic targets. Angiogenesis 20:409–426. 10.1007/s10456-017-9562-9 Gao C et al (2016) Jolkinolide B induces apoptosis and inhibits tumor growth in mouse melanoma B16F10 cells by altering glycolysis. Sci Rep 6:36114. 10.1038/srep36114 Shriwas P et al (2021) A small-molecule pan-class I glucose transporter inhibitor reduces cancer cell proliferation in vitro and tumor growth in vivo by targeting glucose-based metabolism. Cancer Metab 9:14. 10.1186/s40170-021-00248-7 Widmer DS et al (2013) Hypoxia contributes to melanoma heterogeneity by triggering HIF1alpha-dependent phenotype switching. J Invest Dermatol 133:2436–2443. 10.1038/jid.2013.115 Wang ZH, Peng WB, Zhang P, Yang XP, Zhou Q (2021) Lactate in the tumour microenvironment: From immune modulation to therapy. EBioMedicine 73:103627. 10.1016/j.ebiom.2021.103627 Rodig N et al (2003) Endothelial expression of PD-L1 and PD-L2 down-regulates CD8 + T cell activation and cytolysis. Eur J Immunol 33:3117–3126. 10.1002/eji.200324270 Fuglerud BM et al (2022) SOX9 reprograms endothelial cells by altering the chromatin landscape. Nucleic Acids Res 50:8547–8565. 10.1093/nar/gkac652 Zhao J et al (2024) Distinct roles of SOX9 in self-renewal of progenitors and mesenchymal transition of the endothelium. Angiogenesis 27:545–560. 10.1007/s10456-024-09927-7 Law CW et al (2016) RNA-seq analysis is easy as 1-2-3 with limma, Glimma and edgeR. F1000Res 5, 10.12688/f1000research.9005.3 Liberzon A et al (2015) The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst 1:417–425. 10.1016/j.cels.2015.12.004 Patel J et al (2013) Prospective surface marker-based isolation and expansion of fetal endothelial colony-forming cells from human term placenta. Stem Cells Transl Med 2:839–847. 10.5966/sctm.2013-0092 Church DM et al (2011) Modernizing reference genome assemblies. PLoS Biol 9:e1001091. 10.1371/journal.pbio.1001091 Hao Y et al (2021) Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587 e3529, 10.1016/j.cell.2021.04.048 Davidson S et al (2020) Single-Cell RNA Sequencing Reveals a Dynamic Stromal Niche That Supports Tumor Growth. Cell Rep 31:107628. 10.1016/j.celrep.2020.107628 Hanzelmann S, Castelo R, Guinney J (2013) GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics 14. 10.1186/1471-2105-14-7 Karras P et al (2022) A cellular hierarchy in melanoma uncouples growth and metastasis. Nature 610:190–198. 10.1038/s41586-022-05242-7 Hodis E et al (2022) Stepwise-edited, human melanoma models reveal mutations' effect on tumor and microenvironment. Science 376:eabi8175. 10.1126/science.abi8175 Cable DM et al (2022) Robust decomposition of cell type mixtures in spatial transcriptomics. Nat Biotechnol 40:517–526. 10.1038/s41587-021-00830-w Law CW, Chen Y, Shi W, Smyth GK (2014) voom: Precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol 15:R29. 10.1186/gb-2014-15-2-r29 Ritchie ME et al (2015) limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43:e47. 10.1093/nar/gkv007 Liberzon A et al (2011) Molecular signatures database (MSigDB) 3.0. Bioinformatics 27:1739–1740. 10.1093/bioinformatics/btr260 Ashburner M et al (2000) Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25:25–29. 10.1038/75556 Ramilowski JA et al (2015) A draft network of ligand-receptor-mediated multicellular signalling in human. Nat Commun 6:7866. 10.1038/ncomms8866 Gu Z, Gu L, Eils R, Schlesner M, Brors B (2014) circlize Implements and enhances circular visualization in R. Bioinformatics 30:2811–2812. 10.1093/bioinformatics/btu393 Additional Declarations There is NO Competing Interest. Supplementary Files SupplementalTablesandFigures.docx Cite Share Download PDF Status: Under Review 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-5457583","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":383449012,"identity":"ff5e18ac-21d0-4d8b-b8a9-4af81c791cad","order_by":0,"name":"Kiarash Khosrotehrani","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABB0lEQVRIie3RMUvDQBTA8RcCl+Xi/EIw+QoNB9JB6Ve5EGiXOgmOEig8F+t+UL9DRkddmuXo3Eni3kJWMYJnKXRoz9JN5P7Tu+N+3MEBuFx/MATGtiPfrJPdfIzwLREnE8jLYyQq2bBpu7dkEEzn7cdzf1TV0mvWBGnv5TCJgdWZohvB+aJQU43XlW787IkgqywkgYDisJT5A44FhGTIUrI4JPB+JbwzJF0J74tw1FvK4NOQgY2Yh81jzn5u4cI3t0hDmBkgt5FowoaRIim4Hgv/nDBT+n0SzRZYKAvBmi6w7WQS3GvhrekuPauL13Z1e3n1aCHg729tvgYt5w8nTzrtcrlc/79vzFdSosyDrvwAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-6406-4076","institution":"Diamantina Institute / University of Queensland","correspondingAuthor":true,"prefix":"","firstName":"Kiarash","middleName":"","lastName":"Khosrotehrani","suffix":""},{"id":383449013,"identity":"89878f92-d786-4481-83db-97c565e32e6f","order_by":1,"name":"Ghazaleh Hashemi","email":"","orcid":"https://orcid.org/0000-0002-1326-8992","institution":"The University of Queensland Frazer Institute","correspondingAuthor":false,"prefix":"","firstName":"Ghazaleh","middleName":"","lastName":"Hashemi","suffix":""},{"id":383449014,"identity":"96ff53b7-c1ae-4d9e-a2c0-9793fec3bbd8","order_by":2,"name":"Haiming Li","email":"","orcid":"https://orcid.org/0000-0002-6991-8114","institution":"The University of Queensland Frazer Institute","correspondingAuthor":false,"prefix":"","firstName":"Haiming","middleName":"","lastName":"Li","suffix":""},{"id":383449015,"identity":"6ecbf8f0-4b2a-4b50-a275-8e27e70214da","order_by":3,"name":"Samuel X Tan","email":"","orcid":"","institution":"The University of Queensland Frazer Institute","correspondingAuthor":false,"prefix":"","firstName":"Samuel","middleName":"X","lastName":"Tan","suffix":""},{"id":383449016,"identity":"657bb74e-6c8e-4517-a771-75db5f496bfa","order_by":4,"name":"Chenhao Zhou","email":"","orcid":"https://orcid.org/0000-0001-7702-3436","institution":"Frazer Institute / University of Queesland","correspondingAuthor":false,"prefix":"","firstName":"Chenhao","middleName":"","lastName":"Zhou","suffix":""},{"id":383449017,"identity":"f131c0a0-0d35-42cd-ba78-b787685708fc","order_by":5,"name":"James Dight","email":"","orcid":"","institution":"The University of Queensland Frazer Institute","correspondingAuthor":false,"prefix":"","firstName":"James","middleName":"","lastName":"Dight","suffix":""},{"id":383449018,"identity":"2c8079fa-3e45-4312-98f2-4f4ac5e4d95c","order_by":6,"name":"Kwong Ching Li","email":"","orcid":"","institution":"The University of Queensland Frazer Institute","correspondingAuthor":false,"prefix":"","firstName":"Kwong","middleName":"Ching","lastName":"Li","suffix":""},{"id":383449019,"identity":"2455d502-3bad-46b3-b442-25691c8c5564","order_by":7,"name":"Quan Nguyen","email":"","orcid":"https://orcid.org/0000-0001-7870-5703","institution":"Institute for Molecular Bioscience / University of Queensland","correspondingAuthor":false,"prefix":"","firstName":"Quan","middleName":"","lastName":"Nguyen","suffix":""},{"id":383449020,"identity":"3b242050-2b66-49cc-9712-164c6d2ee108","order_by":8,"name":"Edwige Roy","email":"","orcid":"https://orcid.org/0000-0003-1991-5844","institution":"University of Queensland","correspondingAuthor":false,"prefix":"","firstName":"Edwige","middleName":"","lastName":"Roy","suffix":""},{"id":383449021,"identity":"a6c53aea-9e97-4149-8731-b9a2e6ff7054","order_by":9,"name":"laura sormani","email":"","orcid":"https://orcid.org/0000-0002-0838-2088","institution":"The University of Queensland","correspondingAuthor":false,"prefix":"","firstName":"laura","middleName":"","lastName":"sormani","suffix":""}],"badges":[],"createdAt":"2024-11-15 04:50:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5457583/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5457583/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":70171869,"identity":"549e7dd4-ded7-4e4f-88ab-4c06bab454b3","added_by":"auto","created_at":"2024-11-29 06:48:42","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":211807,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSox9 expression is induced in Tumor Endothelial Cells. A. \u003c/strong\u003eC57BL/6J mice were injected intradermally with B16-F0 melanoma cells and sacrificed 14 days post-tumor inoculation. \u003cstrong\u003eB\u003c/strong\u003e. Immunofluorescence staining of endothelial cells in B16-F0 melanoma tumorsusing CD31 (AF568) and SOX9 (AF647) antibodies, with nuclei counterstained by DAPI (AF 405). \u003cstrong\u003eC. \u003c/strong\u003eImmunofluorescence staining of endothelial cells in human melanoma tissue, utilizing CD31 (AF568) and SOX9 (AF647) antibodies, with nuclei counterstained by DAPI (AF 405). \u003cstrong\u003eD\u003c/strong\u003e. \u003cem\u003eCdh5CreERt2/Rosa-YFP \u003c/em\u003elineage tracing mice were induced with tamoxifen for five consecutive days. Following tamoxifen treatment, HcMel12 P0 cells were injected intradermally, and tumorswere collected 21 days later. \u003cstrong\u003eE\u003c/strong\u003e. Immunofluorescence staining of sorted endothelial cells (YFP+ Lineage- mCherry-) from melanoma tumorsand aorta tissue using CD31 (AF568) and SOX9 (AF647) antibodies. \u003cstrong\u003eF\u003c/strong\u003e. Quantification of YFP+ CD31+ SOX9+ cells. Data were analyzedusing the Mann-Whitney test and are presented as mean ± SD (n=4 mice). \u003cstrong\u003eG\u003c/strong\u003e. Unsupervised clustering revealed four major endothelial cell subtypes, each encompassing cells from various cancer types (BC: Breast cancer; CRC: Colorectal cancer; LC: Lung cancer; OvC: Ovarian cancer). \u003cstrong\u003eH\u003c/strong\u003e.UMAP plots display single-cell expression patterns of SOX9 across different endothelial cell clusters, pooled from tumor and normal tissue-derived cells (left) and separately from tumor and normal tissue (right). The colorof the dots indicates high (red) or low (grey) expression levels of SOX9.\u003c/p\u003e","description":"","filename":"Slide1.png","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/d5916e1654f6c718edb87b36.png"},{"id":70173098,"identity":"85cd8ac3-fb1a-4328-a5d2-40acb7d97113","added_by":"auto","created_at":"2024-11-29 06:56:42","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":156545,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMelanoma-Conditioned Media Induces Sox9-Driven Progenitor Self-Renewal in ECFCs and Disrupts Tube Formation Capacity. A\u003c/strong\u003e. Schematic representation of the experimental procedure. \u003cstrong\u003eB\u003c/strong\u003e. Western blot analysis of SOX9 protein expression in endothelial colony-forming cells (ECFCs) from three different donors treated with melanoma-conditioned media (WM164, MM96 and Do4 cells). \u003cstrong\u003eC.\u003c/strong\u003e Quantification of SOX9 protein expression normalized to β-actin. Data were analyzed using 2-way ANOVA and are presented as mean ± SD (n=3 donors). \u003cstrong\u003eD\u003c/strong\u003e. qPCR analysis of SOX9 mRNA expression in ECFCs from three different donors treated with melanoma-conditioned media (WM164, MM96 and Do4 cells). \u003cstrong\u003eE\u003c/strong\u003e. Single-cell colony-forming assay (scCFA): Quantification of the proportion of total colonies formed. \u003cstrong\u003eF\u003c/strong\u003e. Representative images of the 3 different types of colonies observed in the scCFA, including endothelial cells (EC), low proliferative potential colonies (LPP), and high proliferative potential colonies (HPP). \u003cstrong\u003eG\u003c/strong\u003e. Quantification of the different types of colonies formed. \u003cstrong\u003eH\u003c/strong\u003e. Tube formation assay evaluating the ability of ECFCs to form capillary-like structures. \u003cstrong\u003eI\u003c/strong\u003e. Quantification of tube formation parameters using the ImageJ Angiogenesis Analyzer software, including Node (red spot): Pixels with at least three neighboring connections; Junction (dark blue edge): Groups of nodes forming bifurcations; Segment (yellow): Linear structures connecting two junctions; Branch (green): A linear structure connecting a junction and an extremity; Mesh (light blue): Enclosed areas formed by segments.\u003c/p\u003e","description":"","filename":"Slide2.png","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/1eeb75a27d6d479ce442b07f.png"},{"id":70171665,"identity":"d8d878ee-9d10-43a9-904b-b2cd1a58924a","added_by":"auto","created_at":"2024-11-29 06:40:42","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":216435,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSox9 Deletion Reduces Tumor Endothelial Cells and Vessel Density While Enhancing Vessel Maturation.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA. \u003c/strong\u003eSox9fl/fl/Cdh5CreERt2/Rosa-YFP (\u003cem\u003eSox9eKO\u003c/em\u003e) and Cdh5CreERt2/Rosa-YFP (\u003cem\u003eSox9eWT\u003c/em\u003e) mice were subcutaneously injected with B16-F0 cells and sacrificed 14 days post-tumor inoculation. \u003cstrong\u003eB\u003c/strong\u003e. Flow cytometry analysis defined the endothelial cell hierarchy by gating the CD34+Lin− population within YFP+ cells. Endovascular Progenitors (EVP, light blue), Transit Amplifying (TA, purple), and Differentiated (D, navy) populations were further characterized based on CD31 and VEGFR2 expression: EVP as CD31lowVEGFR2low, TA as CD31intVEGFR2low, and Differentiated as CD31highVEGFR2high, respectively. \u003cstrong\u003eC\u003c/strong\u003e. Quantification of total endothelial cells (CD34+Lin-) per 5x10^5 cells. \u003cstrong\u003eD\u003c/strong\u003e. Flow cytometry analysis comparing the endothelial hierarchy between Sox9eWT and Sox9eKO tumors. Results were analyzedusing 2-way ANOVA and are presented as mean ± SD (P \u0026lt; 0.0001, n=4). \u003cstrong\u003eE\u003c/strong\u003e. Sox9fl/fl/Cdh5CreERt2/Rosa-YFP (Sox9eKO) and Cdh5CreERt2/Rosa-YFP (Sox9eWT) mice were subcutaneously injected with HcMel12 P0 cells and sacrificed 21 days post-tumor inoculation. \u003cstrong\u003eF\u003c/strong\u003e. Immunofluorescence staining of HcMel12 P0 tumorsshowing YFP+ (yellow, Alexa-Fluor 488), CD31+ (red, Alexa-Fluor 568) vessels, NG2+ (green, Alexa-Fluor 647) pericytes, and nuclei stained with DAPI (blue, Alexa-Fluor 405). Scale bar = 50 µm. \u003cstrong\u003eG\u003c/strong\u003e. Quantification of YFP+ and CD31+ vessel area fractions (YFP: P = 0.015; CD31: P = 0.007). Results were analyzedusing the Mann-Whitney test and presented as mean ± SD (n=5). \u003cstrong\u003eH\u003c/strong\u003e. Quantification of YFP+ and CD31+ vessel counts. Results were analyzedusing the Mann-Whitney test and presented as mean ± SD (n=5). \u003cstrong\u003eI\u003c/strong\u003e. Quantification of NG2+ area fraction and vessel number (area: P = 0.3095; vessel count: P = 0.0556). \u003cstrong\u003eJ\u003c/strong\u003e. Quantification of number of NG2+ vessels. \u003cstrong\u003eK\u003c/strong\u003e. Quantification of NG2 coverage per vessel (P = 0.031). Results were analyzedusing the Mann-Whitney test and presented as mean ± SD (n=5).\u003c/p\u003e","description":"","filename":"Slide3.png","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/c8ce5ac1702e3ffa64ebf330.png"},{"id":70174918,"identity":"9e64e2cb-5da3-48ea-85c6-a4b9221ae4b3","added_by":"auto","created_at":"2024-11-29 07:12:42","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":176862,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSox9 Deletion in Endothelial Cells Impairs Melanoma Tumor Growth and Metastasis.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA.\u003c/strong\u003eSchematic representation of the experimental protocol. \u003cem\u003eSox9fl/fl/Cdh5CreERt2/Rosa-YFP \u003c/em\u003e(\u003cem\u003eSox9eKO\u003c/em\u003e) and \u003cem\u003eCdh5CreERt2/Rosa-YFP \u003c/em\u003e(\u003cem\u003eSox9eWT\u003c/em\u003e) mice were treated with tamoxifen for five consecutive days to induce gene deletion. HcMel12 cells were then injected subcutaneously, and mice were sacrificed 21 days post-tumor inoculation.\u003cbr\u003e\n \u003cstrong\u003eB.\u003c/strong\u003eTumor weight of HcMel12 tumors. Data were analyzed using the Mann-Whitney test and are presented as mean ± SD (n=8 mice; P=0.0029). \u003cstrong\u003eC.\u003c/strong\u003eTumor volume of HcMel12 tumors measured on days 10, 14, 18, and 21 post-inoculation. Data were analyzed using 2-way ANOVA and are presented as mean ± SD (n=8 mice; P\u0026lt;0.0001). \u003cstrong\u003eD\u003c/strong\u003e.Schematic representation of the experimental protocol. \u003cem\u003eSox9eKO\u003c/em\u003e and \u003cem\u003eSox9eWT\u003c/em\u003e mice were treated with tamoxifen for five consecutive days. HcMel12 cells were then injected intradermally, and primary tumors were surgically resected 21 days post-tumor inoculation. Day 42 post tumour inoculation, lungs were collected for H \u0026amp; E staining analysis.\u003cstrong\u003e E.\u003c/strong\u003e H\u0026amp;E staining of the lungs. Bar scale represents 500 µM, 40X magnification \u003cstrong\u003eF. \u003c/strong\u003eQuantification of lung nodules. Results were analysed using the Mann-Whitney test and presented as mean ± SD (n=8; P=0.035). \u003cstrong\u003eG. \u003c/strong\u003eQuantification of the area of nodules in the lungs. Results were analysed using the Mann-Whitney test and presented as mean ± SD (n=8; P=0.010). \u003cstrong\u003eH. \u003c/strong\u003eMicro-CT images of lungs.\u003c/p\u003e","description":"","filename":"Slide4.png","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/b979ba81d078e06ba9268eb0.png"},{"id":70171672,"identity":"4534b4ba-57cc-44de-a99d-db8b45a4d4b9","added_by":"auto","created_at":"2024-11-29 06:40:42","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":275117,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSpatial Whole-Transcriptome Mapping Reveals Distinct Changes in the \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eSox9eKO\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003eTumor Microenvironment. A\u003c/strong\u003e. Cell type maps for \u003cem\u003eSox9eWT\u003c/em\u003e (n = 4) and \u003cem\u003eSox9eKO\u003c/em\u003e (n = 4) murine melanoma sections; robust cell type deconvolution applied using public single-cell RNA dataset on murine melanoma from Davidson \u003cem\u003eet al. \u003c/em\u003eAll included spots had a primary cell type of melanoma; therefore, visualization indicates the cell type with the second-highest proportion for each spot. B. Cell composition plot presenting the mean proportion of each cell type by Sox9 status. \u003cstrong\u003eC-E. \u003c/strong\u003eGene set variation analysis (GSVA) displaying differential pathway enrichment across 40 selected Hallmark and Gene Ontology MSigDB murine gene sets. Spot-level GSVA values were obtained by random walk, then weighted according to the proportion of C) melanoma cells, D) endothelial cells, and E) lymphocytes [comprising T cells, B cells, and NK cells] in each spot. Blue and red points respectively indicate reduced and increased activity in \u003cem\u003eSox9eKO\u003c/em\u003esamples relative to Sox9eWT samples. Vertical purple line indicates FDR = 0.05 threshold following Benjamini-Hochberg correction. \u003cstrong\u003eF\u003c/strong\u003e. Ligand-receptor analysis through stLearn. Blue and red points respectively indicate reduced and increased abundance of specific ligand-receptor interactions in \u003cem\u003eSox9eKO\u003c/em\u003esamples relative to Sox9eWT samples. \u003cstrong\u003eG,H. \u003c/strong\u003eChord diagrams of ligand-receptor interactions enriched in F) Sox9eWT and G) \u003cem\u003eSox9eKO\u003c/em\u003e samplesrespectively. Arrows indicate directionality from ligands to receptors. Arrow coloursdenote the cell type the ligand-receptor interaction was most frequently observed in, while segment colours denote the cell type the transcript was most frequently observed in.\u003c/p\u003e","description":"","filename":"Slide5.png","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/4113cb5dd5ca7891d0014295.png"},{"id":70173369,"identity":"b91d3bcb-c502-4f74-85a8-408bc6245a47","added_by":"auto","created_at":"2024-11-29 07:04:42","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":447707,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSox9 deletion in endothelial cells reverses hypoxia-induced metabolic rewiring in tumors.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e. Volcano plot illustrating differentially expressed (DE) genes from RNA sequencing of FACS-sorted mcherry+ HcMel12 melanoma cells isolated from \u003cem\u003eSox9eWT \u003c/em\u003e(n = 5) and \u003cem\u003eSox9eKO\u003c/em\u003e (n = 5) tumors. Genes with an unadjusted p-value \u0026lt; 0.05 are shown, with upregulated genes (logFC\u0026gt; 5) in red and downregulated genes (logFC\u0026lt; -0.5) in blue. \u003cstrong\u003eB. \u003c/strong\u003eGene Set Enrichment Analysis (GSEA) Hallmark analysis displaying enriched gene sets (FDR \u0026lt; 0.05), ranked by normalized enrichment score (NES). Red bars indicate positive NES values, representing enrichment in the Sox9eKO group, while blue bars indicate negative NES values, representing enrichment in the Sox9eWT group. \u003cstrong\u003eC. \u003c/strong\u003eEnrichment plots for the top 3 gene sets enriched in the GSEA Hallmark analysis, showing the running ES score profiles. Heat maps show the top 20 genes for each pathway, with gene expression levels represented by a colorscale ranging from red (high expression) to dark blue (lowest expression). \u003cstrong\u003eD. \u003c/strong\u003eImmunofluorescence staining of HcMel12 mCherry+ tumors, showing mCherry-positive melanoma cells (red), GLUT1-positive cells (green), and DAPI-stained nuclei (blue). Scale bar = 20 μm. \u003cstrong\u003eE\u003c/strong\u003e.Quantification of GLUT1-positive cells (P = 0.0411, n = 6, Mann-Whitney statistical test). \u003cstrong\u003eF\u003c/strong\u003e. Immunofluorescence staining of HcMel12 mCherry+ tumors, showing mCherry-positive melanoma cells (red), HIF1-α-positive cells (green), and DAPI-stained nuclei (blue). Scale bar = 20 μm. \u003cstrong\u003eG\u003c/strong\u003e.Quantification of HIF1-α-positive cells (P = 0.0043, n = 6, Mann-Whitney statistical test).\u003c/p\u003e","description":"","filename":"Slide6.png","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/8dd4eaabeef82add9f525f29.png"},{"id":70171689,"identity":"3e4e5934-d115-4b29-9a0e-6c0157f8bfe5","added_by":"auto","created_at":"2024-11-29 06:40:43","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":443867,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSox9 Deletion Restores Lymphocytes Infiltration in the Tumor Core.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e. Immunofluorescence staining of CD4+ cells in Sox9eWT and Sox9eKO HcMel12 tumours (n=6 tumors per group) using a Rat anti-mouse CD4 FITC antibody. Nuclei were stained with DAPI (AF 405). \u003cstrong\u003eB\u003c/strong\u003e. Immunofluorescence staining of CD8+ cells in Sox9eWT and Sox9eKO HcMel12 tumours (n=6 tumors per group) using a Rat anti-mouse CD8 FITC antibody. Nuclei were stained with DAPI (AF 405). \u003cstrong\u003eC\u003c/strong\u003e. Quantification of the percentage of CD4+ cells per tumor, in the periphery and in the core of the tumor. (n=6 tumorsper group; Mann-Whitney statistical test). \u003cstrong\u003eD\u003c/strong\u003e. Quantification of the percentage of CD8+ cells per tumor, in the periphery (red) and in the core (yellow) of the tumor. (n=6 tumorsper group; Mann-Whitney statistical test).\u003c/p\u003e","description":"","filename":"Slide7.png","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/33db6b56c58252697e02bc7c.png"},{"id":70171696,"identity":"48e68ad6-6a9c-4016-856b-dfc66d1c5170","added_by":"auto","created_at":"2024-11-29 06:40:43","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":168841,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSOX9-Driven Vascular Program Blockade in Cancer Vessels Normalizes the Tumor Microenvironment and Limits Metastatic Progression.\u003c/strong\u003eEndothelial cell-specific conditional knockout of Sox9 reduces primary tumor vascularization, normalizes vessel structure, and decreases glycolysis and hypoxia, thereby inhibiting tumor growth and metastasis.\u003c/p\u003e","description":"","filename":"Slide8.png","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/3b9be8e2d81a26a91910ddcc.png"},{"id":70174920,"identity":"68094dbc-dd49-4fac-9a37-6d5abef00b19","added_by":"auto","created_at":"2024-11-29 07:12:53","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3247244,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/010e412f-3e06-4fe3-8467-bc0168b67b5c.pdf"},{"id":70173094,"identity":"85357340-337a-4b5b-94e7-bf68922169a2","added_by":"auto","created_at":"2024-11-29 06:56:42","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3305956,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementalTablesandFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-5457583/v1/94363081a4b53c1aaef5eab6.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"SOX9 reactivation in cancer vessels shapes the tumor micro-environment through hypoxia and immune depletion promoting tumor growth and metastasis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMelanoma is a potentially aggressive skin cancer with a high risk of metastasis. In melanoma as in many other cancers, vascularisation is pivotal for the growth of cancer cells at the primary site and accelerates the spread of malignant cells \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Indeed, tumors need to establish a blood vessel supply to provide them with oxygen, nutrients and remove the waste \u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. In contrast to normal tissues, tumor vascularization exhibits a disordered labyrinth of dysfunctional and malformed vessels, resulting in structural and functional destabilization. This vascular network, characterized by high permeability and leakage, facilitates the intravasation of cancer cells and the formation of metastases. \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e In addition, recent studies have highlighted the role of abnormal tumor vasculature in shaping a protumorigenic and immunosuppressive tumor microenvironment (TME), hindering effective drug delivery and compromising therapeutic responses \u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. These novel observations provide an alternative vision to the role of tumor vasculature resulting from the excessive activity of VEGF as initially proposed by Folkman\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Indeed, in many cancers, targeted anti-VEGF therapy failed to reduce tumor vascularisation and improve patient outcomes \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. Specifically, in patients with melanoma, no significant difference in distant metastases was observed \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Aside from angiogenesis driven by VEGF, other modes of vessel formation as well as consequences of the abnormal vessel structure may explain why anti-angiogenic therapy is not universally effective \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003ePreviously, we reported the existence and activity of progenitor populations in the endothelium in human \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e and mouse \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e models allowing us to distinguish a hierarchy from endovascular progenitors (EVP) to differentiated cells (D). These populations differed functionally based on their proliferation, vessel formation and self-renewal capacity but also molecularly regarding their gene expression and surface markers. Moreover, lineage tracing experiments have shown that EVPs infiltrate tumors early on, developing vascular networks in the centre of melanoma tumors contributing to a variety of vascular beds \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Transcriptional profiling of EVPs versus D cells in the aorta or tumor vasculature identified \u003cem\u003eSox9\u003c/em\u003e as a potential driver of EVP activity \u003csup\u003e\u003cspan additionalcitationids=\"CR11\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Indeed, \u003cem\u003eSox9\u003c/em\u003e has been recently reported to drive EVP self-renewal and fate decision between endothelial or mesenchymal differentiation during wound healing suggesting it is an essential gene for progenitor function \u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. Moreover, S\u003cem\u003eox9\u003c/em\u003e has similar important developmental roles in endocardial development of valves highlighting its ability to divert endothelial cells from their normal vascular function\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. In the present study, we hypothesized that \u003cem\u003eSox9\u003c/em\u003e activity in tumor vasculature would drive immature vessels formed by progenitors and would be responsible for the abnormal tumor vasculature. We have therefore examined the role of \u003cem\u003eSox9\u003c/em\u003e in endothelial cells that contribute to melanoma vascular formation. The conditional deletion of \u003cem\u003eSox9\u003c/em\u003e in the endothelium resulted in a dramatic reduction in tumor growth and metastasis through the normalization of tumor vessels, a modification of the pro-tumorigenic tumor microenvironment, and the restoration of immune cell infiltration providing a new understanding of the abnormal tumor vasculature and opening new avenues for anti-neovascularisation therapies.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e \u003cb\u003eExpression of\u003c/b\u003e \u003cb\u003eSox9\u003c/b\u003e \u003cb\u003eis induced in Tumor Endothelial Cells in Human or Mouse\u003c/b\u003e\u003c/p\u003e \u003cp\u003eAlthough the expression of transcription factor \u003cem\u003eSox9\u003c/em\u003e in EVPs through bulk \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e and single-cell \u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e RNA sequencing is well established, the presence and level of SOX9 protein in endothelial cells remains uncertain. B16-F0 melanoma cells were inoculated intradermally into the flank of C57BL/6J mice and resulting tumors were collected 14 days later (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Immunofluorescent staining of tumor sections confirmed the expression of SOX9-positive nuclei within the endothelial cells identified by CD31 co-staining (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB). No primary antibody was used as a control and hair follicles epidermal cells that express high levels of \u003cem\u003eSox9\u003c/em\u003e, anatomically situated in the hair follicle bulge, were used as a positive control also ensuring the specificity of the staining (Supplemental Fig.\u0026nbsp;1). Similarly, we explored human primary cutaneous melanoma samples. Immunofluorescent staining revealed the presence of SOX9-positive cells in the endothelium of human melanoma tumor vessels (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC). Of note, SOX9 is known to be expressed in melanoma cells, and the nuclear staining in these tumor cells served as an internal positive control.\u003c/p\u003e \u003cp\u003eTo have a more quantitative assessment of SOX9 expression in endothelial cells, we next used \u003cem\u003eCdh5-Cre\u003c/em\u003e\u003csup\u003e\u003cem\u003eER\u003c/em\u003e\u003c/sup\u003e \u003cem\u003eRosaYFP\u003c/em\u003e mice to genetically label all endothelial cells upon tamoxifen administration. After 5 consecutive days of tamoxifen administration, HcMel12\u003csup\u003emCherry+\u003c/sup\u003e cells, a murine melanoma cell line derived from a primary melanoma tumor in Tyr::Hgf-Cdk4\u003csup\u003eR\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003eC\u003c/sup\u003e mice, were introduced via intradermal inoculation into the flank of mice\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Three weeks after tumor inoculation, mice were sacrificed, and tumors and aortas were collected (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). Single YFP\u003csup\u003e+\u003c/sup\u003e Lineage\u003csup\u003e\u0026minus;\u003c/sup\u003e mCherry\u003csup\u003e\u0026minus;\u003c/sup\u003e endothelial cells were flow-sorted onto slides to verify SOX9 expression in the tumor and aorta (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE). SOX9 protein expression was commonly identified in the YFP\u003csup\u003e+\u003c/sup\u003eCD31\u003csup\u003e+\u003c/sup\u003e endothelial cells in the tumor, confirming the expression of SOX9 in tumor endothelial cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE). Moreover, the proportion of SOX9-positive endothelial cells was significantly higher in the tumor compared to the aorta, suggesting that the tumor environment increased SOX9 expression in the endothelial cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF).\u003c/p\u003e \u003cp\u003eFinally, to examine the relevance of these findings in other cancers, we used established and publicly available single cell RNA sequencing data of stromal cells from lung, colorectal, breast, and ovarian cancer focusing on the expression of \u003cem\u003eSOX9\u003c/em\u003e in tumor endothelial cells\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. It was observed that \u003cem\u003eSOX9\u003c/em\u003e was upregulated in endothelial cells in tumors compared to corresponding normal tissue (Average Log2FC 3.07, P\u0026thinsp;=\u0026thinsp;6.47e-09) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eG-H). In particular, tumor tissues exhibited a higher fraction of endothelial cells expressing \u003cem\u003eSOX9\u003c/em\u003e, with 1.2% in tumors versus 0.1% in adjacent normal tissues. Further stratification of endothelial cell populations indicated that \u003cem\u003eSOX9\u003c/em\u003e expression was predominantly localized to tip and venous endothelial cells, including high endothelial venules (HEVs).\u003c/p\u003e \u003cp\u003eCollectively, these findings suggested that \u003cem\u003eSOX9\u003c/em\u003e is upregulated in a subset of tumor-associated endothelial cells in both mice and humans. This observation raised important questions about the functional role of SOX9 in tumor vascularization, particularly its influence on endothelial cell behavior within the tumor microenvironment.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eMelanoma-Conditioned Media Induces Sox9-Driven Progenitor Self-Renewal in ECFCs and Disrupts Tube Formation Capacity.\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo explore the impact of tumor-derived factors on SOX9 expression and endothelial progenitor cell function, we exposed primary human endothelial colony-forming cells (ECFCs) isolated from three donors to conditioned media (CM) from human melanoma cell lines carrying distinct mutations (WM164, MM96, DO4; see Supplemental Table\u0026nbsp;6) for five days (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Quantitative PCR analysis revealed a significant upregulation of \u003cem\u003eSOX9\u003c/em\u003e mRNA in ECFCs treated with WM164 CM (P\u0026thinsp;=\u0026thinsp;0.021, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD), which was corroborated by western blot analysis showing an increase in SOX9 protein expression (P\u0026thinsp;=\u0026thinsp;0.0248). A similar trend was observed when ECFCs were exposed to MM96 and DO4 CM (P\u0026thinsp;=\u0026thinsp;0.07 and P\u0026thinsp;=\u0026thinsp;0.0724, respectively; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, C). To more closely mimic the tumor-endothelial interaction, GFP\u003csup\u003e+\u003c/sup\u003e ECFCs were co-cultured with mCherry\u0026thinsp;+\u0026thinsp;WM164 melanoma cells. After five days of co-culture, endothelial and melanoma cells were FACS sorted based on GFP and mCherry fluorescence respectively (Supplemental Fig.\u0026nbsp;2A, B). \u003cem\u003eSOX9\u003c/em\u003e expression in GFP\u003csup\u003e+\u003c/sup\u003e ECFCs was significantly upregulated following co-culture with melanoma cells (9-fold increase, P\u0026thinsp;=\u0026thinsp;0.0006; Supplemental Fig.\u0026nbsp;2C), further highlighting the direct effect of tumor cells on SOX9 expression in endothelial progenitors.\u003c/p\u003e \u003cp\u003eNext, a single-cell colony formation assay was performed to assess the functional impact of melanoma-conditioned media on ECFC self-renewal capacity. Flow-sorted ECFCs were plated 1 cell per well in endothelial growth medium (EGM2) and monitored for colony formation over 14 days. Colonies were categorized based on proliferative potential: high proliferative potential (HPP) colonies (\u0026gt;\u0026thinsp;500 cells), low proliferative potential (LPP) colonies (\u0026gt;\u0026thinsp;250\u0026ndash;500 cells), and endothelial clusters (\u0026lt;\u0026thinsp;50 cells) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE). Exposure to melanoma-conditioned media from WM164, MM96, and DO4 significantly increased the total proportion of colonies formed (P\u0026thinsp;=\u0026thinsp;0.0058, P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001, and P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001, respectively; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF). ECFCs from the control group formed a higher proportion of endothelial clusters (80\u0026ndash;90%) compared to CM-treated cells, which showed a 1.5 to 2-fold reduction (P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001 vs Control, n\u0026thinsp;=\u0026thinsp;4). Interestingly, CM-treated ECFCs exhibited an increased proportion of HPP colonies, suggesting enhanced self-renewal and proliferative capabilities (P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001 vs Control ECFCs, n\u0026thinsp;=\u0026thinsp;4; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eG). To investigate the differentiation capacity of ECFCs under the influence of melanoma-conditioned media, we performed a matrigel capillary assay to assess their ability to form tube-like structures. The number of meshes, segment lengths, and nodes were significantly reduced in ECFCs treated with CM compared to the control group (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eH, I). Previous studies have shown that immature endothelial colony-forming cells (ECFCs) exhibit a reduced capacity for tube formation compared to differentiated ECFCs that have been primed with mesenchymal stem cells (MSCs) \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThese results suggest that the upregulation of SOX9, induced by melanoma-conditioned media, was associated with an immature state of ECFCs, prioritizing self-renewal over differentiation, which inhibited their capacity to carry out specialized tasks like tube formation, characteristic of more differentiated endothelial cells.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eReduced numbers but increased maturation of tumor vessels upon\u003c/b\u003e \u003cb\u003eSox9\u003c/b\u003e \u003cb\u003edeletion in the endothelium\u003c/b\u003e\u003c/p\u003e \u003cp\u003eBuilding on these \u003cem\u003ein vitro\u003c/em\u003e observations, we next explored the role of \u003cem\u003eSox9\u003c/em\u003e in the tumor endothelium using a mouse model that enabled us to delete \u003cem\u003eSox9\u003c/em\u003e specifically in endothelial cells while labelling them with YFP. We generated a conditional knock-out mouse model, \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e \u003cem\u003e(Sox9\u003c/em\u003e\u003csup\u003e\u003cem\u003efl/fl\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e/Cdh5-Cre\u003c/em\u003e\u003csup\u003e\u003cem\u003eER\u003c/em\u003e\u003c/sup\u003e \u003cem\u003eRosaYFP)\u003c/em\u003e, in comparison to \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e \u003cem\u003e(Sox9\u003c/em\u003e\u003csup\u003e\u003cem\u003e+/+\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e/Cdh5-Cre\u003c/em\u003e\u003csup\u003e\u003cem\u003eER\u003c/em\u003e\u003c/sup\u003e \u003cem\u003eRosaYFP)\u003c/em\u003e mice. To validate the efficiency of \u003cem\u003eSox9\u003c/em\u003e deletion in the endothelium, we administered tamoxifen intraperitoneally for five consecutive days prior to the subcutaneous inoculation of B16-F0 melanoma cells (Supplemental Fig.\u0026nbsp;3A). Fourteen days post-inoculation, tumors were collected and YFP\u003csup\u003e+\u003c/sup\u003e CD34\u003csup\u003e+\u003c/sup\u003e Lineage\u003csup\u003e\u0026minus;\u003c/sup\u003e and Lineage\u003csup\u003e+\u003c/sup\u003e cells were flow-sorted (Supplemental Fig.\u0026nbsp;3B) for genomic DNA analysis. As expected, deletion of exons 2 and 3 of the \u003cem\u003eSox9\u003c/em\u003e gene was confirmed in endothelial cells from \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e mice but not in lineage\u003csup\u003e\u0026minus;\u003c/sup\u003epositive cells or \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e controls (Supplemental Fig.\u0026nbsp;3C). To further confirm the loss of \u003cem\u003eSox9\u003c/em\u003e in the tumor endothelium, B16-F0 tumor sections from \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e and \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e mice were analysed by immunofluorescent staining. No SOX9 protein was observed in the YFP\u003csup\u003e+\u003c/sup\u003e cells in the \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e mouse model (Supplemental Fig.\u0026nbsp;3D).\u003c/p\u003e \u003cp\u003eHaving established the above model, we next investigated the impact of \u003cem\u003eSox9\u003c/em\u003e deletion on the development and structure of tumor-associated blood vessels. To evaluate changes in endothelial populations, we first conducted a flow cytometry analysis on B16-F0 primary tumors (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). We applied our previously established gating strategy to identify the endothelial hierarchy, which ranges from progenitors (EVPs) to mature differentiated endothelial cells (D cells), based on CD31 expression levels, from low in EVPs to high in D cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB). \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors exhibited a significant 3.5-fold reduction in the total number of endothelial cells (YFP\u003csup\u003e+\u003c/sup\u003eCD34\u003csup\u003e+\u003c/sup\u003eLin\u003csup\u003e\u0026minus;\u003c/sup\u003e) compared to \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e (P\u0026thinsp;=\u0026thinsp;0.0286, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). This reduction was accompanied by a significant decrease in the percentage of EVPs and a concomitant increase in fully differentiated D cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD), consistent with previous reports in wounds and aorta\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. This result suggested that \u003cem\u003eSox9\u003c/em\u003e deletion in the endothelium promoted the maturation of progenitors toward a more differentiated endothelial phenotype. Immunofluorescence analysis of tumor sections further supported these findings, revealing a significant reduction in both the number and area fraction of YFP\u003csup\u003e+\u003c/sup\u003e and CD31\u003csup\u003e+\u003c/sup\u003e vessels in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors (Supplemental Fig.\u0026nbsp;4A, B). These observations align with the reduced endothelial cell population observed by flow cytometry, indicating that \u003cem\u003eSox9\u003c/em\u003e deletion reduces overall vessel density within tumors.\u003c/p\u003e \u003cp\u003eTo assess whether this phenotype was reproducible across tumor models, we analyzed HcMel12 tumors. As in the B16-F0 model, a significant reduction in the percentage area fraction of YFP\u003csup\u003e+\u003c/sup\u003e (P\u0026thinsp;=\u0026thinsp;0.015) and CD31\u003csup\u003e+\u003c/sup\u003e (P\u0026thinsp;=\u0026thinsp;0.007) vessels was observed in tumors from \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e mice (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF, G). The absolute number of YFP\u003csup\u003e+\u003c/sup\u003e and CD31\u003csup\u003e+\u003c/sup\u003e vessels also decreased significantly following Sox9 deletion (P\u0026thinsp;=\u0026thinsp;0.015 and P\u0026thinsp;=\u0026thinsp;0.007, respectively; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eH).\u003c/p\u003e \u003cp\u003eGiven the reduced vessel density, we next evaluated the impact of \u003cem\u003eSox9\u003c/em\u003e deletion on vessel maturation by quantifying pericyte coverage, a key indicator of vessel stability and maturation. Immunofluorescent staining for NG2\u003csup\u003e+\u003c/sup\u003e pericytes revealed a trend towards increased area fraction covered by NG2\u003csup\u003e+\u003c/sup\u003e vessels in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors (P\u0026thinsp;=\u0026thinsp;0.309, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eI), although the total number of NG2\u003csup\u003e+\u003c/sup\u003e vessels decreased in parallel with the overall reduction in vessel numbers (P\u0026thinsp;=\u0026thinsp;0.055, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eJ). Importantly, the proportion of NG2\u003csup\u003e+\u003c/sup\u003e vessels, indicative of vessels with pericyte coverage, was significantly higher in the \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e group (P\u0026thinsp;=\u0026thinsp;0.031, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eK). This suggests that although fewer vessels were present in the absence of \u003cem\u003eSox9\u003c/em\u003e, those that remained were more mature, characterized by increased pericyte coverage and thus greater vessel stability.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eSox9\u003c/b\u003e \u003cb\u003edeletion in the endothelium reduces melanoma progression.\u003c/b\u003e\u003c/p\u003e \u003cp\u003eWe aimed to assess the effect of \u003cem\u003eSox9\u003c/em\u003e deletion in the endothelium on melanoma outcomes. To investigate this, we used HcMel12 melanoma cells given its ability to become highly metastatic upon \u003cem\u003ein vivo\u003c/em\u003e passaging, leading to spontaneous lung metastases in mice. This dual approach was particularly valuable, as it allowed us to delineate the distinct mechanisms underlying primary tumor development from those driving metastatic dissemination. Additionally, we reproduced the experiments using B16-F0 melanoma cells for primary tumor analysis and B16-F10 cells injected intravenously for metastasis assessment.\u003c/p\u003e \u003cp\u003eHcMel12 tumor cells (not passaged \u003cem\u003ein vivo\u003c/em\u003e; P0) were inoculated intradermally into \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e and \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e animals. Tumor growth was monitored over 21 days, and tumors were weighed at the time of collection (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). A significant reduction in tumor weight (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB; P\u0026thinsp;=\u0026thinsp;0.0029) and volume (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC; P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001) was observed between \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e and \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eGiven the effect on the HcMel12 primary tumor progression, \u003cem\u003ein vivo\u003c/em\u003e passaged HcMel12 were employed to assess the effect of \u003cem\u003eSox9\u003c/em\u003e deletion in the endothelium on metastasis. To reflect the most common clinical scenario, the primary tumors were surgically resected, and mice were monitored for metastasis three weeks post-surgery (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). Histological analysis with H\u0026amp;E staining was performed on lung and liver sections, revealing metastases exclusively in the lungs (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE). A significant reduction in the number of metastatic lung nodules was observed in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e mice (P\u0026thinsp;\u0026lt;\u0026thinsp;0.035, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF). Notably, the size of the nodules, as reflected by their area, was also significantly decreased (P\u0026thinsp;\u0026lt;\u0026thinsp;0.0104, Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG). To further visualize lung metastases, micro-CT imaging was performed (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG and H; Sox9eWT; Video 1, Sox9eKO; Video 2), reinforcing that \u003cem\u003eSox9\u003c/em\u003e deletion in the endothelium effectively reduced metastatic spread to the lungs from the primary tumor.\u003c/p\u003e \u003cp\u003eNext, we asked if the reduction of metastasis was a direct consequence of changes to the vascularisation of the primary tumor or if it was secondary to alterations in the lung vasculature, the metastatic niche. Therefore, we used the same mouse models and injected the B16-F10 melanoma cells intravenously as an experimental model of metastasis (Supplemental Fig.\u0026nbsp;5A). This approach allowed tumor cells to disseminate through the bloodstream and directly engraft in the lung. Mice were sacrificed 10 days after tumor cell injection, and lung metastases were visible macroscopically. H\u0026amp;E staining of lung sections (Supplemental Fig.\u0026nbsp;5B) showed a trend towards a reduction in total metastatic area (P\u0026thinsp;=\u0026thinsp;0.161, Supplemental Fig.\u0026nbsp;5C) and the number of nodules (P\u0026thinsp;=\u0026thinsp;0.258, Supplemental Fig.\u0026nbsp;5D) in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003emice, although these differences were not statistically significant. To confirm the metastatic nature of the nodules, lung sections were stained with MITF, a melanoma marker (Supplemental Fig.\u0026nbsp;5F). Looking closer at the location of the nodules, in the proximal regions of the lungs, there was a significant reduction in both the area (P\u0026thinsp;=\u0026thinsp;0.018, Supplemental Fig.\u0026nbsp;5C) and the number of nodules (P\u0026thinsp;=\u0026thinsp;0.0008, Supplemental Fig.\u0026nbsp;5D), contrasted by a significant increase in nodule formation at the lung periphery. Despite these regional differences, the overall effects of \u003cem\u003eSox9\u003c/em\u003e deletion in the endothelium on B16-F10 cell engraftment and lung metastasis formation were moderate, suggesting that \u003cem\u003eSox9\u003c/em\u003e deletion primarily exerts its effects in the vasculature of the primary tumor, where its expression is elevated upon tumor establishment.\u003c/p\u003e \u003cp\u003eOverall, deletion of \u003cem\u003eSox9\u003c/em\u003e in the tumor vasculature had a dramatic effect on tumor outcomes and in metastatic dissemination prompting us to understand its effect at cellular and molecular level.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eSox9\u003c/b\u003e \u003cb\u003edeletion in the endothelium results in global gene expression changes across the tumor and its microenvironment.\u003c/b\u003e\u003c/p\u003e \u003cp\u003eTo have a better global understanding of changes in primary tumors resulting in improved outcomes upon \u003cem\u003eSox9\u003c/em\u003e endothelial deletion, we subjected HcMel12 primary tumors to spatial transcriptomics analysis. HcMel12 tumors (P0) were inoculated to \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e and \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e mice (n\u0026thinsp;=\u0026thinsp;4 per group). On day 21, mice were sacrificed, and tumors were collected and processed for spatial transcriptomic analysis. Tissue spots with \u0026ge;\u0026thinsp;500 genes and genes expressed in \u0026ge;\u0026thinsp;3 tissue spots across the aggregate dataset were included for downstream analysis. Within this filtered series, the median number of genes and unique genes per spot were 2009 and 1238 respectively.\u003c/p\u003e \u003cp\u003eSpot-level cell type annotation was achieved through robust cell type decomposition (RCTD), with reference to a public single-cell RNA dataset by Davidson et al [3] (Supplemental Figs.\u0026nbsp;6A, B). Ten cell types were identified and projected onto spot-level gene expression data at 55\u0026micro;m resolution (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA, Supplemental Figs.\u0026nbsp;6C, D). Notably, UMAP plots of tissue spots\u0026rsquo; normalized gene expression matrices did not demonstrate spatial clustering by either cell type (Supplemental Fig.\u0026nbsp;6E) or endothelial Sox9 status (Supplemental Fig.\u0026nbsp;6F), substantiating the need for fractional cell type deconvolution through RCTD.\u003c/p\u003e \u003cp\u003eIn these murine tumor sections, melanoma cells invariably demonstrated the highest proportion in every tissue spot and were thus the primary cell type. Nonetheless, secondary cell type visualization identified a heterogeneous tumour milieu, with broad spatial variation in proportions of endothelial cells, T/B lymphocytes, natural killer cells, antigen-presenting cells, and fibroblasts (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). Sample-level cell type proportions were then compared between \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e and \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e tumors (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). Relative to \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e, \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors demonstrated a trend towards reduced endothelial cells (P\u0026thinsp;=\u0026thinsp;0.067), but higher levels of T cells (P\u0026thinsp;=\u0026thinsp;0.172) and macrophages (P\u0026thinsp;=\u0026thinsp;0.357).\u003c/p\u003e \u003cp\u003eTo further explore the behaviour of \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e versus \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e tumors, we applied cell type-weighted gene set variation analysis to identify between-group differences in composite gene pathways for melanoma cells, endothelial cells, and lymphocytes. As compared to melanoma cells in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e tumors, melanoma cells in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors demonstrated reduced expression of genes relating to the hypoxia, glycolysis, TGFβ signalling and epithelial-mesenchymal transition (EMT) pathways (adjusted P\u0026thinsp;=\u0026thinsp;0.0373 for all; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). Similarly, endothelial cells in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors demonstrated downregulation of genes corresponding to fatty acid oxidation (Acetyl-CoA thiolase), EMT, TGFβ signalling, vascular morphogenesis but also hypoxia, and glycolysis (adjusted P\u0026thinsp;=\u0026thinsp;0.0406 for all) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD).\u003c/p\u003e \u003cp\u003eOn ligand-receptor interaction analysis through stLearn\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e we identified a number of important changes in endothelium-enriched tissue spots (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF). In \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e tumors, Hepatocyte Growth Factor (HGF) produced by tumor cells could be responsible for the abnormal tumor endothelial phenotype by acting on endothelial MET receptors (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG). In contrast, endothelial cells in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors were enriched for interactions between the platelet-derived growth factor beta (PDGFβ) and platelet-derived growth factor receptor beta (PDGFRβ) in fibroblasts/pericytes suggesting better pericyte coverage of blood vessel. Similarly, there was increased interaction between the intercellular adhesion molecule 2 (ICAM2) on endothelial cells and integrin alpha M (ITGAM) on immune cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH). These observations collectively indicate that deletion of \u003cem\u003eSox9\u003c/em\u003e in the endothelium induces major and broad changes in the tumor environment.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eReduction in Glycolysis and Hypoxia in melanoma primary tumors after deletion of\u003c/b\u003e \u003cb\u003eSox9\u003c/b\u003e \u003cb\u003ein the endothelium\u003c/b\u003e\u003c/p\u003e \u003cp\u003eIn addition to the observed changes in the tumor vasculature, the spatial transcriptomics analysis suggested gene expression changes in tumor cells. To validate these findings, bulk RNA sequencing was performed on mcherry-positive HcMel12 melanoma cells (non \u003cem\u003ein vivo\u003c/em\u003e passaged; P0) that were FACS-sorted from \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e versus \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e tumors (n\u0026thinsp;=\u0026thinsp;5 mice per group, Supplemental Fig.\u0026nbsp;7).\u003c/p\u003e \u003cp\u003eWe identified 323 differentially expressed (DE) genes with an unadjusted p value\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Among these genes, 101 were upregulated (logFC\u0026thinsp;\u0026gt;\u0026thinsp;0.5; Red), and 222 downregulated in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e mice (logFC \u0026lt; -0.5; Blue) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). Gene Set Enrichment Analysis (GSEA) utilizing the MsigDB mouse gene set collections revealed several significantly enriched and under-enriched gene sets associated with critical biological pathways in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e melanoma tumor cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). Notably, the analysis identified a significant downregulation of several pathways, including Glycolysis (248 genes; P\u0026thinsp;=\u0026thinsp;2.75e-04), Hypoxia (256 genes; P\u0026thinsp;=\u0026thinsp;1.37e-09), MTORC1 (279 genes; P\u0026thinsp;=\u0026thinsp;1.75e-05), and Cholesterol Homeostasis (94 genes; P\u0026thinsp;=\u0026thinsp;7.65e-05). Conversely, G2M checkpoint (285 genes; P\u0026thinsp;=\u0026thinsp;0.0015), mitotic spindle (312 genes; P\u0026thinsp;=\u0026thinsp;0.007), and E2F targets (262 genes; P\u0026thinsp;=\u0026thinsp;0.008) were found to be enriched. Among the top three most differentially downregulated pathways in the \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC), Hypoxia (NES \u0026minus;\u0026thinsp;2.8, FDR 6.8e-08), Glycolysis (NES \u0026minus;\u0026thinsp;2.25, FDR 3.4e-03), and mTORC1(NES-2.29, FDR 4.3e-04) signalling emerged as particularly relevant, given their well-established roles in mediating the tumor's metabolic adaptation to hypoxic conditions. The observed alterations in these pathways in our model may be directly linked to the vascular changes induced by \u003cem\u003eSox9\u003c/em\u003e deletion.\u003c/p\u003e \u003cp\u003eHypoxia is a hallmark of the tumor microenvironment, arising from the imbalance between the rapid proliferation of tumor cells and inadequate vascular supply. Under hypoxic conditions, cancer cells activate an adaptive response mediated primarily by Hypoxia-Inducible Factor 1 (HIF1α). This transcription factor drives the upregulation of glycolytic enzymes, including GLUT1 (Glucose Transporter 1), a key regulator of glycolysis in cancer. GLUT1 facilitates increased glucose uptake into tumor cells, enabling them to maintain high glycolytic activity, which in turn supports their accelerated proliferative capacity and enhanced survival under oxygen-limited conditions.\u003c/p\u003e \u003cp\u003eTo confirm the downregulation of the hypoxia and glycolysis pathways identified in tumor cells from both the spatial transcriptomics and the bulk RNA sequencing data, we assessed the protein expression levels of GLUT1 and HIF1α through immunofluorescence staining (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD; F). A significant reduction in the intensity and area of both markers was observed (P\u0026thinsp;=\u0026thinsp;0.0411 and 0.0043 respectively, Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE; G), indicating decreased activation of these metabolic pathways in \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors. These findings, combined with the observed reduction in vessel number along with their maturation, suggested improved tumor perfusion and vessel normalization.\u003c/p\u003e \u003cp\u003e \u003cb\u003eRestoration of immune infiltration in primary tumors after deletion of\u003c/b\u003e \u003cb\u003eSox9\u003c/b\u003e \u003cb\u003ein the endothelium\u003c/b\u003e\u003c/p\u003e \u003cp\u003eUltimately, the spatial transcriptomics data indicated an increase in immune cell infiltration following the deletion of \u003cem\u003eSox9\u003c/em\u003e in the endothelium. To validate this finding, tumor sections from the \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e and \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e HcMel12 tumors were stained with CD4 and CD8 antibodies to assess T lymphocyte infiltration (Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA;B). Immunofluorescence staining revealed a significant two-fold increase in CD4\u003csup\u003e+\u003c/sup\u003e cell infiltration in the \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors compared to the controls (P\u0026thinsp;=\u0026thinsp;0.0411; Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eC). Notably, when comparing the core and periphery of the tumors, significant changes were observed exclusively in the core region (Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD;E), which exhibited a complete absence of visible CD4\u003csup\u003e+\u003c/sup\u003e cells in the \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e group. This pattern was particularly pronounced for CD8 staining, which also demonstrated a significant two-fold increase in infiltration in the core of the \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e tumors (Figs.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eF-H).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eTumor vascularisation is one of the hallmarks of cancer and an imperative step for tumor progression and metastasis \u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. The tumor vasculature is poorly formed and destabilised which leads to cancer cell intravasation, migration, and metastasis \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e and for many cancers anti-angiogenic drugs have only partially addressed this issue \u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. A better understanding of the origins of vessels in the tumor and the role of endothelial cells in the tumor microenvironment may therefore lead to better therapeutic options for patients. Our lab demonstrated that endovascular progenitor cells (EVPs) which reside in the blood vessel walls, contribute to the formation of a \u003cem\u003ede novo\u003c/em\u003e blood vessel network in the tumor. Initial gene expression characteristic of the EVP population identified transcription factor \u003cem\u003eSox9\u003c/em\u003e as an essential marker distinguishing tissue resident murine endothelial progenitor cells from mature endothelial cells in various tissue beds \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e including tumors \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe present study revealed that \u003cem\u003eSOX9\u003c/em\u003e expression is augmented in endothelial cells exposed to a tumor environment both \u003cem\u003ein vivo\u003c/em\u003e and \u003cem\u003ein vitro\u003c/em\u003e, in mice and human cells. This was done mostly in melanoma but also identified in other tumour types such as human colorectal, lung, ovarian, and breast cancer samples. In all these examples the expression of \u003cem\u003eSOX9\u003c/em\u003e was increased in tumour endothelium compared to the endothelial cells of normal tissue, suggesting that \u003cem\u003eSOX9\u003c/em\u003e fulfils a significant function in tumor endothelial cells. Indeed, upon conditional deletion of \u003cem\u003eSox9\u003c/em\u003e in the endothelium, there was a reduction in tumour vessel numbers and surface area in parallel with an increase in pericyte coverage suggesting an increase in maturation. This was paralleled with a reduction in hypoxia, glycolysis and an increase in immune cell infiltration leading to reduced tumor size and metastasis. These findings support an essential role of \u003cem\u003eSox9\u003c/em\u003e in the development of the abnormal tumour vasculature and its deletion as an essential step in vessel normalisation.\u003c/p\u003e \u003cp\u003eOne possible explanation for the decrease in metastasis from the primary tumor could be the reduction in the total number of endothelial cells and vessels in the centre of the primary tumor. By providing fewer escape routes to tumor cells, metastasis may be reduced \u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. An alternative possibility, however, is that vessel normalisation improves tumor oxygenation, increases immune cell infiltration, and reduces cancer cell intravasation. Vessel leakiness reduces blood flow through an increase in intratumoral pressure, causing hypoxia in the tumor microenvironment and surrounding tissue. In return, hypoxia promotes metastasis of melanoma cells by downregulating melanocyte differentiation markers and enhancing their invasion in a HIF-1α dependent manner. The activation of \u003cem\u003eSnail\u003c/em\u003e and \u003cem\u003eTwist\u003c/em\u003e by HIF-1α triggers EMT, promoting increased metastatic potential. Furthermore, HIF-1α facilitates the migration of cancer cells by regulating the expression of collagen fibres, integrins and deposition of ECM \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. Moreover, low oxygen levels in the tumor microenvironment promote the induction of pro-tumor regulatory T cells while hampering the function of anti-tumor CD8\u003csup\u003e+\u003c/sup\u003e T cells \u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Additionally, hypoxia leads to higher expression of programmed death-ligand 1 (PD-L1) by myeloid-derived suppressor cells, which hinders the function of immune cells \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. Many aspects of our observations were in line with vessel normalisation as described in these previous studies.\u003c/p\u003e \u003cp\u003eThe reduction in the total number of endothelial cells and blood vessels triggered by the deletion of \u003cem\u003eSox9\u003c/em\u003e was accompanied by increased pericytes coverage, confirmed by NG2 immunofluorescent staining. Pericytes play a pivotal role in maintaining the stability of blood vessels by covering endothelial cells and forming a supportive and protective network. This strategic positioning of pericytes along the vascular wall prevents the abnormal leakage of vessels \u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. For instance, in aged mice, microvasculature is deprived of a supportive structure, pericytes coverage, and adherent junctions, resulting in leaky vessels \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. Coutelle and colleagues showed that targeting both ANG2 and VEGF in a xenograft model of colorectal cancer cells led to improvement in pericyte coverage, vascular integrity, and leaky vessels \u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. Moreover, the study by Keskin \u003cem\u003eet al.\u003c/em\u003e revealed that depleting pericytes and targeting NG2 signalling repaired vascular stability in various model systems, reducing tumor growth and metastasis \u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. Conversely, other studies showed that pericyte depletion was linked to increased hypoxia, EMT, and metastasis in breast cancer \u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. There is ongoing debate regarding whether leaky blood vessels have advantages \u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e over normalised vessels. Indeed tumor cells proliferate rapidly, and the surrounding tissue limits tumor development, suppressing blood flow and making the vessels leaky. In addition, insufficient lymphatic drainage increases fluid flux and elevates interstitial fluid pressure, which contributes to enhanced vessel permeability, disorganised vascular pressure, and intravasation of cancer cells. This restricts drug delivery, implying normalisation might be advantageous over leaky vessels \u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eOur study consistently demonstrated the reduction of hypoxia and glycolysis upon \u003cem\u003eSox9\u003c/em\u003e conditional deletion using a variety of techniques and platforms such as immunostaining, bulk RNA sequencing and spatial transcriptomics. These studies further support the role of \u003cem\u003eSox9\u003c/em\u003e in driving the abnormal vessel structure as its deletion may restore normalised and mature vessels and reduce hypoxia even when the tumour has a larger size, 3 weeks after inoculation. Studies have shown that inhibition of glycolysis suppressed melanoma cell proliferation by inducing apoptosis \u003csup\u003e33 34\u003c/sup\u003e. The observation of smaller tumors in the \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e group aligns with these findings. Moreover, exposure to a hypoxic environment prompting HIF-1α activation suppresses melanocytic markers and enhances the invasive potential of melanoma cells \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. More importantly, our spatial transcriptomics findings suggested that the normalisation of vessels through the deletion of \u003cem\u003eSox9\u003c/em\u003e in endothelial cells was accompanied by profound changes in the tumor microenvironment. This involved a change in immune cell infiltration that was particularly dramatic in the centre of tumors. As indicated by ligand-receptor interactions, this could be part of the tumour vessel normalisation process, providing better expression of adhesion molecules on endothelial cells that allow the initiation of the various steps leading to the entry of immune cells in the tumour tissue. Alternatively, the reduction of hypoxia and glycolysis may lead in significant reduction of byproducts of glycolysis such as lactate accumulation which has been described to dampen immune responses\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. Finally, despite the lack of clear indication in our spatial transcriptomics findings, one cannot exclude the production of immunomodulatory mediators by endothelial cells upon SOX9 expression. Indeed, PDL1 expression on the endothelium has been reported and can affect T cell effector functions\u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. The direct immunomodulatory function of the endothelium in this context warrants further investigation.\u003c/p\u003e \u003cp\u003eMost of our insight into the developmental functions of \u003cem\u003eSox9\u003c/em\u003e has originated from studies involving the male reproduction system, chondrogenesis, and heart valve development. Recent studies revealed the role of \u003cem\u003eSOX9\u003c/em\u003e in endothelial to mesenchymal transition by analysis of single-cell chromatin in HUVECs \u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e and wound healing assay\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. In these previous studies, \u003cem\u003eSox9\u003c/em\u003e has distinct roles in self-renewal and stem cell fate decision favouring a mesenchymal fate\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. However, there are no studies on the role of \u003cem\u003eSox9\u003c/em\u003e in endothelial cells in tumor vascularisation and metastasis. despite our findings of increased self-renewal \u003cem\u003ein vitro\u003c/em\u003e, it is also questionable if \u003cem\u003eSox9\u003c/em\u003e, in this new context of tumor vascularisation, is a strict marker for endothelial progenitors or EVPs.\u003c/p\u003e \u003cp\u003eOther efforts to reduce progenitor function have resulted in similar results. Donovan \u003cem\u003eet al.\u003c/em\u003e showed that conditional depletion of \u003cem\u003eRbpj\u003c/em\u003e, a transcription factor highly upregulated in EVPs, resulted in the depletion of EVPs and significantly reduced metastases from primary tumors. Similarly to \u003cem\u003eSox9\u003c/em\u003e, \u003cem\u003eRbpj\u003c/em\u003e depletion has major implications in a range of target genes beyond its importance for progenitor function. Overall, it is difficult to completely dissociate the impact of such genes on progenitor function as opposed to their alternative effects on the endothelium.\u003c/p\u003e \u003cp\u003eIn conclusion, as in many other situations, \u003cem\u003eSox9\u003c/em\u003e expression is upregulated in the context of tumor endothelium where it drives a developmental program including an increase in progenitor numbers and activity, resulting in immature vessels that promote hypoxia, tumor growth and metastasis. We have argued and demonstrated that its depletion normalises vessels despite reducing their numbers, resulting in less hypoxia, better tumor immune infiltration, drastically changing the tumour microenvironment resulting in less metastases (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e). Adoption of such developmental program in the cancer context is not novel and further emphasises the need for development of targeted therapies.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Material and Method","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\n \u003ch2\u003eAnimal Model and Ethics\u003c/h2\u003e\n \u003cp\u003eAll animal procedures were conducted in accordance with the University of Queensland Animal Ethics Committee (AEC) guidelines and approvals (#2022/AE000150 and #2022/AE000335). Both male and female mice, aged 10\u0026ndash;14 weeks, were included in this study. Endothelial-specific \u003cem\u003eSox9\u003c/em\u003e gene knockout was achieved by breeding \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003efl/fl\u003c/em\u003e\u003c/sup\u003e mice, which had been backcrossed to a C57BL/6J background, with \u003cem\u003eCdh5Cre\u003c/em\u003e\u003csup\u003e\u003cem\u003eERt2\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e/Rosa-YFP\u003c/em\u003e mice, resulting in the generation of \u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003efl/fl/\u003c/em\u003e\u003c/sup\u003e\u003cem\u003eCdh5Cre\u003c/em\u003e\u003csup\u003e\u003cem\u003eERt2\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e/Rosa-YFP\u003c/em\u003e (\u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeKO\u003c/em\u003e\u003c/sup\u003e) triple-transgenic mice. \u003cem\u003eCdh5Cre\u003c/em\u003e\u003csup\u003e\u003cem\u003eERt2\u003c/em\u003e\u003c/sup\u003e\u003cem\u003e/Rosa-YFP\u003c/em\u003e (\u003cem\u003eSox9\u003c/em\u003e\u003csup\u003e\u003cem\u003eeWT\u003c/em\u003e\u003c/sup\u003e) mice were used as controls. Wild-type C57BL/6J mice (WT) were sourced from the Animal Resources Centre (Perth, Western Australia).\u003c/p\u003e\n \u003cp\u003eFor conditional knockout of \u003cem\u003eSox9\u003c/em\u003e, tamoxifen (Sigma-Aldrich, USA) was administered to activate Cre recombinase. Tamoxifen was prepared by dissolving it in a solution of 10% ethanol and 90% corn oil, achieving a final concentration of 20 mg/ml. Each mouse received an intraperitoneal injection of 2 mg of tamoxifen (100 \u0026micro;l) every 24 hours for up to five consecutive days prior to melanoma cell inoculation.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eTumor Cell Culture and innoculation\u003c/h3\u003e\n\u003cp\u003eB16-F0 murine melanoma cells were cultured in RPMI 1640 medium (Gibco, ThermoFisher Scientific, USA) supplemented with 10% fetal bovine serum (FBS). HcMel12 and HcMel12\u003csup\u003emCherry+\u003c/sup\u003e murine melanoma cells were maintained in RPMI 1640 medium with 10% FBS, 2 mM l-glutamine, 1 mM HEPES, and 10 mM non-essential amino acids (Gibco, ThermoFisher Scientific, USA). For subcutaneous cell injection, mice were anesthetized with isoflurane, and hair was removed using an electric clipper. Suspensions of 5\u0026times;10\u003csup\u003e5\u003c/sup\u003e B16-F0 cells or 1\u0026times;10\u003csup\u003e6\u003c/sup\u003e HcMel12 and HcMel12\u003csup\u003emCherry+\u003c/sup\u003e cells in 100 \u0026micro;l saline were injected subcutaneously.\u003c/p\u003e\n\u003ch3\u003eSerial Tumor Transplantation\u003c/h3\u003e\n\u003cp\u003eHcMel12 tumors develop spontaneous lung metastasis following repeated subcutaneous transplantation in C57BL/6J mice. For the initial transplantation, 1\u0026times;10\u003csup\u003e6\u003c/sup\u003e cultured HcMel12 cells were injected subcutaneously into the flank of C57BL/6J mice. Tumor growth was monitored through visual inspection and palpation, with measurements taken using a sliding vernier caliper. Tumors were excised 21 days post-inoculation and mechanically dissociated using scissors. Cells were filtered through a 70 \u0026micro;m cell strainer (BD Biosciences, USA), washed with PBS, and 1\u0026times;10\u003csup\u003e6\u003c/sup\u003e HcMel12 cells in 100 \u0026micro;L saline were injected subcutaneously into new C57BL/6J mice. This process was repeated over five passages until lung metastases were observed macroscopically.\u003c/p\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003eResection of HcMel12 Tumors\u003c/h2\u003e\n \u003cp\u003eSubcutaneous HcMel12 tumors were surgically resected 21 days post-inoculation. Mice were placed on a heat mat and anesthetized with 2% isoflurane via a nose cone. The skin around the surgical site was cleansed with a cotton swab and a topical antiseptic agent (70% ethanol or chlorhexidine). For the procedure, mice were positioned in lateral recumbency, and the tumor was gently pinched with tweezers and excised using sterile dissecting scissors. The incision site was closed with N604 sutures. Following the procedure, tumor weight was recorded on a scale, and tumor volume was measured with a sliding vernier caliper by recording diameters along two axes. Tumor volume was calculated as \u003cem\u003eV\u0026thinsp;=\u0026thinsp;L\u0026times;W\u003c/em\u003e\u003csup\u003e\u003cem\u003e2\u003c/em\u003e\u003c/sup\u003e, where \u003cem\u003eV\u003c/em\u003e is the tumor volume, \u003cem\u003eL\u003c/em\u003e is the tumor length, and \u003cem\u003eW\u003c/em\u003e is the tumor width.\u003c/p\u003e\n\u003c/div\u003e\n\u003ch3\u003eTissue Processing of Murine Tumors\u003c/h3\u003e\n\u003cp\u003eFollowing University of Queensland ethical guidelines, mice were sacrificed, and melanoma tumors along with other tissues were collected. Tumors were fixed in 4% paraformaldehyde (PFA) for 2 hours at room temperature, then washed twice with 1\u0026times;PBS (Gibco, ThermoFisher Scientific, USA) and sequentially immersed in 10%, 20%, and 30% sucrose (Chem-Supply, AUS) for cryoprotection. Fixed samples were embedded in OCT compound (Sakura Finetek, USA) and stored at -80˚C for long-term storage. Tumors were sectioned at 8 \u0026micro;m thickness on a Cryostat Leica CM1950, with sections collected on Superfrost Plus slides (Thermo Scientific, Germany) and spaced 30 \u0026micro;m apart.\u003c/p\u003e\n\u003ch3\u003eHematoxylin and Eosin (H\u0026amp;E) staining\u003c/h3\u003e\n\u003cp\u003eFor H\u0026amp;E staining, lungs were placed in ProSciTech histology cassettes and stored in 70% ethanol. Tissues were embedded in paraffin and were sectioned 100 \u0026micro;m apart from each other. The Translational Research Institute core facility carried out the H\u0026amp;E staining. Images were taken with an Olympus Slide Scanner VS120 microscope (20x objective). The area and number of nodules were analysed using Olympus OlyVIA software (Version 3.1).\u003c/p\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003eImmunofluorescence - mouse\u003c/h2\u003e\n \u003cp\u003eCryosections were permeabilized with 0.5% Triton X-100 (Chem Supply, Australia) and blocked with 20% normal goat serum. Primary antibodies used in this study for mouse tissue sections included anti-CD31, anti-SOX9, anti-GFP, anti-CD34, anti-GLUT1, and anti-HIF1alpha (Supplemental Table 1). Unbound antibodies were removed by washing sections three times for 5 minutes in 1\u0026times; PBS/0.1% Tween-20 (Amresco, USA). Secondary antibodies conjugated with Alexa Fluor 488, 568, and 647 (Invitrogen, USA) were applied for fluorescence detection, with sections incubated at room temperature for 40 minutes. Following incubation, sections were washed three times with PBS/0.1% Tween-20. Nuclear staining was achieved using ProLong\u0026reg; Gold mounting medium with DAPI (Invitrogen, USA).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n \u003ch2\u003eMicroscopy\u003c/h2\u003e\n \u003cp\u003eAll confocal images were acquired at the Translational Research Institute Microscopy facility. High-magnification images for quantification were captured using an Olympus FV3000 Laser Scanning Confocal Microscope and a Nikon Spinning Disk Confocal Microscope (both Tokyo, Japan). Images were obtained at 10\u0026times;, 20\u0026times;, 40\u0026times;, and 60\u0026times; magnifications.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n \u003ch2\u003eBlood Vessels density\u003c/h2\u003e\n \u003cp\u003eTumor sections were stained with CD31, NG2, and anti-GFP (for endogenous YFP). Vessel counts and area fractions in tumors were quantified using NIS-Element AR analysis software (Version 5.01.0064). Look-Up Tables (LUTs) were standardized across samples and compared with no-secondary control samples to prevent false measurements.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\n \u003ch2\u003eStatistical Analysis\u003c/h2\u003e\n \u003cp\u003eStatistical analyses were conducted using GraphPad Prism v8 software. Non-parametric data were analysed using the Mann-Whitney test, and parametric data were analysed using 2-way ANOVA with Bonferroni correction. A p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was used to determine statistical significance.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\n \u003ch2\u003eFlow Cytometry\u003c/h2\u003e\n \u003cp\u003eAll flow cytometry was conducted at the Translational Research Institute Flow Cytometry Facility. Tissues were excised, dissociated, and transferred to pre-warmed Hanks\u0026rsquo; Balanced Salt Solution (HBSS, Gibco, ThermoFisher Scientific, USA) for digestion with 1 mg/ml collagenase I, 1 mg/ml dispase (both from Gibco, ThermoFisher Scientific, USA), and 150 U/ml DNase-I (Sigma-Aldrich, USA) in a 37˚C water bath for 30 minutes. The dissociated tissue was filtered through a 70 \u0026micro;m nylon strainer using a flat plunger, washed with PBS, and pelleted at 380 rcf for 5 minutes. The pellet was treated with a hypertonic or RBC lysis buffer for 2 minutes to remove erythrocytes, then diluted with PBS and centrifuged again. After discarding the supernatant, cells were prepared in unstained, single-color, and Fluorescence-Minus-One (FMO) controls for gating. Each pellet was resuspended in FACS buffer (0.5% BSA, 2 mM EDTA, 1\u0026times;PBS) and stained for 30 minutes at 4\u0026deg;C. Antibodies are listed in Supplemental Tables 3 and 4.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\n \u003cdiv id=\"Sec20\" class=\"Section3\"\u003e\n \u003ch2\u003eCytospin of FACS-sorted cells\u003c/h2\u003e\n \u003cp\u003eCells were FACS sorted directly into FBS, washed with 1\u0026times;PBS, and resuspended in 150\u0026micro;L of PBS. A Cytospin\u0026trade; 4 Cytocentrifuge was employed to distribute cells in suspension onto SuperFrost Plus\u0026trade; slides. Adhered cells were fixed with ice-cold 2% PFA for 10 minutes on ice, and the standard immunofluorescent staining protocol was carried out.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec21\" class=\"Section2\"\u003e\n \u003ch2\u003egDNA Extraction and PCR\u003c/h2\u003e\n \u003cp\u003eGenomic DNA was extracted from sorted YFP\u003csup\u003e+\u003c/sup\u003e CD34\u003csup\u003e+\u003c/sup\u003e Lin\u003csup\u003e\u0026minus;\u003c/sup\u003e endothelial cells from the B16-F0 tumor to validate the conditional deletion of \u003cem\u003eSox9\u003c/em\u003e in endothelial cells. Following the manufacturer\u0026rsquo;s specifications (Qiagen, USA), the QIAamp DNA Mini Kit was utilised. The polymerase chain reaction was performed with 100 ng of gDNA. The PCR products were run on a 1.5% agarose gel mixed with ethidium bromide. The gel was imaged using a Gel Doc XR\u003csup\u003e+\u003c/sup\u003e imaging system (Bio-Rad Laboratories, USA). Primers sequence are listed in Supplemental Table 5.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e\n \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e\n \u003ch2\u003eRNA Sequencing and Data Analysis\u003c/h2\u003e\n \u003cp\u003eRNA was extracted from the FACS-sorted HcMel12\u003csup\u003emCherry+\u003c/sup\u003e cells using a QIAGEN mini kit (Qiagen, USA) following manufacturer instructions.\u003c/p\u003e\n \u003cp\u003eRNA sequencing was conducted by the Institute for Molecular Bioscience (IMB, University of Queensland). The total RNA library was prepared with TruSeq standard mRNA kit (Illumina, USA) to be used in the Sanger/Illumina 1.9. RNA sequencing analysis was conducted on the Galaxy Australia server by QFAB. QFAB performs initial QC, reads mapped to a reference genome and read counts for the annotated genes. Reads (FASTQ files) were trimmed with Trimmomatic, mapped to a reference genome assembly with HiSAT2, and reads were counted against the GenCode gene annotation with featureCounts. Read count normalisation and differential gene expression analysis were performed in the lab using DESeq2 Galaxy Version 1.22.3.\u003c/p\u003e\n \u003cp\u003eData processing and differential expression analysis across different samples were performed using the standard Limma-Voom pipeline\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Gene set enrichment analysis was performed using the fgsea package). Hallmark gene sets were downloaded from the Molecular Signatures Database (MSigDB)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. Normalized enrichment score (NES) and adjusted p value were calculated for each hallmark gene set.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec24\" class=\"Section2\"\u003e\n \u003ch2\u003eHuman Cell Culture\u003c/h2\u003e\n \u003cp\u003eHuman placenta foetal ECFCs were isolated according to our previously published protocol \u003csup\u003e\u003cspan class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. The human ethics boards of The University of Queensland and the Royal Brisbane and Women\u0026rsquo;s Hospital (#HREC/09/QRBW/14) granted the use of human tissue. ECFCs were cultured on rat tail collagen (Gibco, ThermoFisher Scientific, USA) coated tissue-culture T75 flasks in Endothelial Growth Medium (EGM-2, Lonza Group, Switzerland).\u003c/p\u003e\n \u003cp\u003eFor conditioned media experiments, WM164, D04 and MM96 human melanoma cells (Supplemental Table 6) were seeded at 40% confluency with RPMI (Gibco, ThermoFisher Scientific, USA) and 2% FBS for 72 hours. CM was collected and centrifuged for 10 minutes at 1000 g (4 ˚C) and filtered with a 0.45 \u0026micro;m filter (Merck Millipore). ECFCs were treated with conditioned media for 5 consecutive days.\u003c/p\u003e\n \u003cdiv id=\"Sec26\" class=\"Section3\"\u003e\n \u003ch2\u003eRNA Extraction, cDNA synthesis and qPCR\u003c/h2\u003e\n \u003cp\u003eRNA was extracted from cells using a QIAGEN mini kit (Qiagen, Valencia, CA) according to the manufacturer\u0026rsquo;s instructions. RNA quality and concentration were assessed using A260nm/A280nm spectroscopy on the Nanodrop ND-1000 (Thermo-scientific, Langenselbold, Germany). 5-100ng of RNA was used for cDNA synthesis using the Invitrogen SuperScript III First Strand Kit (Life Technologies, USA). To quantify relative changes in key gene expression, Real Time-Quantitative Polymerase Chain Reaction (qPCR) was conducted on cDNA using SYBR Green Master Mix reagent (Applied Biosystems, United Kingdom). Reactions were run in triplicates on a QuantStudio Flex 7 (ThermoFisher Scientific, USA). Fold change of gene expression was determined using the delta delta Ct method and normalized to the housekeeper gene. Primer sequences are listed in Supplemental Table 7.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec28\" class=\"Section2\"\u003e\n \u003ch2\u003eWestern Blot\u003c/h2\u003e\n \u003cp\u003eCells were washed with 1\u0026times;PBS and lysed in RIPA buffer with 1\u0026times; Complete Protease Inhibitor (Sigma-Aldrich, USA), followed by sonication. Protein concentration was determined using a BCA assay (ThermoFisher, USA). Equal amounts of protein were mixed with Laemmli buffer, boiled at 100\u0026deg;C for 5 minutes, and loaded onto an SDS-PAGE gel. Proteins were transferred to a membrane, blocked overnight at 4\u0026deg;C, and probed with rabbit anti-SOX9 (1:1000, Merck Millipore, USA) and mouse anti-\u0026beta;-actin (1:5000, Sigma-Aldrich, USA). Detection was performed using IRDye\u0026reg; 800CW goat anti-rabbit and IRDye\u0026reg; 680RN goat anti-mouse secondary antibodies (1:2500, LI-COR Biosciences, USA) on an Odyssey CLx imaging system (LI-COR Biosciences, USA).\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec29\" class=\"Section2\"\u003e\n \u003ch2\u003eSpatial Transcriptomics\u003c/h2\u003e\n \u003cdiv id=\"Sec30\" class=\"Section3\"\u003e\n \u003ch2\u003eSample preparation\u003c/h2\u003e\n \u003cp\u003eAnimals were euthanized according to University of Queensland Ethics guidelines. Tissues were snap-frozen in Tissue Tek OCT compound (Sakura Finestek, USA) using isopentane and liquid nitrogen, stored at -80\u0026deg;C. For sectioning, samples were equilibrated at -15\u0026deg;C in a cryostat, sectioned at 8 \u0026micro;m, placed on dry ice, and stored at -80\u0026deg;C.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec31\" class=\"Section2\"\u003e\n \u003ch2\u003eVisium Gene Expression Procedure\u003c/h2\u003e\n \u003cp\u003eTissues were fixed, and stained with hematoxylin and eosin. Permeabilization was performed for 25 minutes, followed by cDNA synthesis (reverse transcription, second-strand synthesis, release), qPCR to determine cycle number, and cDNA clean-up using SPRIselect. cDNA quality was assessed using bioanalyzer trace and quantification. Fragmentation, end repair, A-tailing, and SPRIselect were performed according to the 10\u0026times; Genomics protocol. The final product was sequenced at the Institute of Molecular Bioscience sequencing facility.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec32\" class=\"Section2\"\u003e\n \u003ch2\u003eData preparation\u003c/h2\u003e\n \u003cp\u003eBase call sequence files were aligned to the mm10 (mouse)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e reference assembly using the Space Ranger software package (v3.0; 10x Genomics, Pleasanton). Count matrices were aggregated as a single object in R 4.4.1 through the Seurat package (5.1.0)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e and spatially oriented to 55\u0026micro;m tissue spots spaced 100\u0026micro;m apart in a hexagonal configuration. Tissue spots with \u0026lt;\u0026thinsp;500 transcripts and genes present in less than \u0026lt;\u0026thinsp;3 tissue spots were excluded from downstream analyses. For some analyses, the SCTransform (SCT) function of Seurat was applied to perform variance-stabilizing transformation of sample-level count matrices, using default parameters (variable.features\u0026thinsp;=\u0026thinsp;3000, residual.features\u0026thinsp;=\u0026thinsp;NULL). Other analyses performed independent normalization procedures and therefore employed raw count matrices; these are specified when they occur.\u003c/p\u003e\n \u003cdiv id=\"Sec33\" class=\"Section3\"\u003e\n \u003ch2\u003eCell type composition\u003c/h2\u003e\n \u003cp\u003eUnder the Visium platform, each tissue spot can contain over 25 distinct cells, and standard clustering methods are unsuited to cell type classification. Therefore, we applied cell type deconvolution with reference to a publicly available scRNA murine melanoma dataset by Davidson et al [ArrayExpress: E-MTAB-7427]\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. To ensure concordance with our samples, this scRNA dataset was restricted to primary tumour cells only, and the original series of 15 cell type clusters was collapsed into 9 (Supplemental Figs.\u0026nbsp;6A, B). An additional melanoma cell cluster was developed by performing gene set variation analysis (GSVA)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e on a standard series of melanocyte markers (Pmel, Mitf, Mlana, Sox10, Tyrp1, Dct)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e47\u003c/span\u003e,\u003cspan class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e, then including previously-unclassified cells with a GSVA z-score of \u0026ge;\u0026thinsp;2.\u003c/p\u003e\n \u003cp\u003eSpot-level cell type annotation was subsequently performed through robust cell type deconvolution (RCTD)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e on spot-level raw count matrices, which returned a proportion vector for each tissue spot and cell type. The primary and secondary cell types of each tissue spot were overlaid on the original H\u0026amp;E image through proprietary R code. Differences in cell type proportions between Sox9eKO and Sox9eWT samples were assessed using the Wilcoxon rank-sum test with Benjamini-Hochberg correction at a false discovery rate (FDR) of 0.05.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec34\" class=\"Section3\"\u003e\n \u003ch2\u003eDifferential gene expression\u003c/h2\u003e\n \u003cp\u003eDifferential expression (DE) of captured transcripts was evaluated with respect to specific cell types. To achieve this, spot-level raw count matrices were first weighted by their corresponding cell type proportions. These were pseudobulked into a cell type-specific count matrix for each sample, then normalized for library size and gene-specific variance through the limma-voom model\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. Cell type-specific DE between Sox9eKO and Sox9eWT samples was evaluated using empirical Bayes-moderated t-testing\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e with Benjamini-Hochberg correction (FDR 0.05).\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003ch3\u003eGene pathway analyses\u003c/h3\u003e\n\u003cp\u003eGene pathway analyses were performed using GSVA on curated murine gene sets from the Molecular Signatures Database (MSigDB)\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e, including the Hallmark \u003csup\u003e\u003cspan class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e and Gene Ontology\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e collections. Using SCT count matrices, cell type-agnostic GSVA was performed on all tissue spots. Sample-level cell type-specific GSVA measures were obtained by weighting the GSVA score for each spot by the corresponding cell type proportion, then pseudobulking these across the sample. Between-group differences were assessed using the Wilcoxon rank-sum test with Benjamini-Hochberg correction (FDR 0.05).\u003c/p\u003e\n\u003ch3\u003eLigand-receptor analyses\u003c/h3\u003e\n\u003cp\u003eLigand-receptor (LR) interactions were performed using the stLearn package (0.4.12) for Python 3.8.19 \u003csup\u003e18\u003c/sup\u003e, with reference to the ConnectomeDB LR database\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e54\u003c/span\u003e\u003c/sup\u003e. stLearn was applied to raw spot-level count matrices to quantify LR interactions for each spot and its six immediate hexagonal neighbours. Spot-level LR scores with significant enrichment over background pairs (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were pseudobulked into a sample-level LR matrix; this was conducted both overall, as well as with weighting by cell type. Between-group differences were then normalized using the limma-voom pipeline and statistically evaluated through empirical Bayes-moderated t-testing with Benjamini-Hochberg\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e correction (FDR 0.05). LR pairs with the lowest adjusted p-value for between-group differences were visualized in chord diagrams through the circlize package\u003csup\u003e\u003cspan class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRNA-Sequencing and Spatial transcriptomic data were deposited into the Gene Expression Omnibus database under accession number GSE277775 (Spatial transcriptomic) and GSE278207 (RNA-seq) and are available at the following URL:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ehttps://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE277775\u003c/p\u003e\n\u003cp\u003ehttps://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE278207\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research was carried out at the Translational Research Institute, Woolloongabba, QLD 4102, Australia. The Translational Research Institute is supported by a grant from the Australian Government.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eG.H. and L.S. executed most experiments, evaluated results and wrote the manuscript. K.K. and L.S. conceived and designed the study. C.Z. performed analysis of the public scRNA-seq dataset. S.T. analyzed the Spatial Transcriptomics data, quantified CD4 and CD8 T cell infiltration in tumor immunofluorescent stainings and processed the RNA-seq data. H.L. contributed to \u003cem\u003ein vitro\u003c/em\u003e experiments. K.C.L. performed immunofluorescent staining of CD4 and CD8 T cells. Q.N. provided expertise on Spatial Transcriptomics. J.D. contributed to \u003cem\u003ein vivo\u003c/em\u003e experiments. E.R. reviewed the manuscript and figures. K.K. provided overall supervision and mentorship throughout the project, guided the interpretation of results, and contributed to the critical revision of the manuscript. All co-authors have reviewed and approved this manuscript for publication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interests statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors certify that they have no affiliations with or involvement in any organization or entity with any financial interest or non-financial interest in the subject matter or materials discussed in this manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eEmmett MS, Dewing D, Pritchard-Jones RO (2011) Angiogenesis and melanoma - from basic science to clinical trials. Am J Cancer Res 1:852\u0026ndash;868\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePotente M, Gerhardt H, Carmeliet P (2011) Basic and therapeutic aspects of angiogenesis. Cell 146:873\u0026ndash;887. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2011.08.039\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2011.08.039\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJain RK (2005) Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science 307:58\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.1104819\u003c/span\u003e\u003cspan address=\"10.1126/science.1104819\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchaaf MB, Garg AD, Agostinis P (2018) Defining the role of the tumor vasculature in antitumor immunity and immunotherapy. Cell Death Dis 9:115. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41419-017-0061-0\u003c/span\u003e\u003cspan address=\"10.1038/s41419-017-0061-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFolkman J (1971) Tumor angiogenesis: therapeutic implications. N Engl J Med 285:1182\u0026ndash;1186. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1056/NEJM197111182852108\u003c/span\u003e\u003cspan address=\"10.1056/NEJM197111182852108\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBergers G, Hanahan D (2008) Modes of resistance to anti-angiogenic therapy. Nat Rev Cancer 8:592\u0026ndash;603. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nrc2442\u003c/span\u003e\u003cspan address=\"10.1038/nrc2442\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCorrie PG et al (2014) Adjuvant bevacizumab in patients with melanoma at high risk of recurrence (AVAST-M): preplanned interim results from a multicentre, open-label, randomised controlled phase 3 study. Lancet Oncol 15:620\u0026ndash;630\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHashemi G, Dight J, Khosrotehrani K, Sormani L (2022) Melanoma Tumour Vascularization and Tissue-Resident Endothelial Progenitor Cells. Cancers (Basel) 14. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3390/cancers14174216\u003c/span\u003e\u003cspan address=\"10.3390/cancers14174216\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShafiee A, Patel J, Hutmacher DW, Fisk NM, Khosrotehrani K (2018) Meso-Endothelial Bipotent Progenitors from Human Placenta Display Distinct Molecular and Cellular Identity. Stem Cell Rep 10:890\u0026ndash;904. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.stemcr.2018.01.011\u003c/span\u003e\u003cspan address=\"10.1016/j.stemcr.2018.01.011\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePatel J et al (2017) Functional Definition of Progenitors Versus Mature Endothelial Cells Reveals Key SoxF-Dependent Differentiation Process. Circulation 135:786\u0026ndash;805. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1161/CIRCULATIONAHA.116.024754\u003c/span\u003e\u003cspan address=\"10.1161/CIRCULATIONAHA.116.024754\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDonovan P et al (2019) Endovascular progenitors infiltrate melanomas and differentiate towards a variety of vascular beds promoting tumor metastasis. Nat Commun 10:18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-018-07961-w\u003c/span\u003e\u003cspan address=\"10.1038/s41467-018-07961-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLukowski SW et al (2019) Single-Cell Transcriptional Profiling of Aortic Endothelium Identifies a Hierarchy from Endovascular Progenitors to Differentiated Cells. \u003cem\u003eCell Rep\u003c/em\u003e 27, 2748\u0026ndash;2758 e2743, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.celrep.2019.04.102\u003c/span\u003e\u003cspan address=\"10.1016/j.celrep.2019.04.102\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao J et al (2021) Sox9 and Rbpj differentially regulate endothelial to mesenchymal transition and wound scarring in murine endovascular progenitors. Nat Commun 12:2564. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-021-22717-9\u003c/span\u003e\u003cspan address=\"10.1038/s41467-021-22717-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAkiyama H et al (2004) Essential role of Sox9 in the pathway that controls formation of cardiac valves and septa. Proc Natl Acad Sci U S A 101:6502\u0026ndash;6507. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1073/pnas.0401711101\u003c/span\u003e\u003cspan address=\"10.1073/pnas.0401711101\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBald T et al (2014) Ultraviolet-radiation-induced inflammation promotes angiotropism and metastasis in melanoma. Nature 507:109\u0026ndash;113. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nature13111\u003c/span\u003e\u003cspan address=\"10.1038/nature13111\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang W, Wang L, She J, Zhu J (2021) Examining heterogeneity of stromal cells in tumor microenvironment based on pan-cancer single-cell RNA sequencing data. Cancer Biol Med 19:30\u0026ndash;42. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.20892/j.issn.2095-3941.2020.0762\u003c/span\u003e\u003cspan address=\"10.20892/j.issn.2095-3941.2020.0762\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShafiee A et al (2017) Priming of endothelial colony-forming cells in a mesenchymal niche improves engraftment and vasculogenic potential by initiating mesenchymal transition orchestrated by NOTCH signaling. FASEB J 31:610\u0026ndash;624. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1096/fj.201600937\u003c/span\u003e\u003cspan address=\"10.1096/fj.201600937\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePham D et al (2023) Robust mapping of spatiotemporal trajectories and cell-cell interactions in healthy and diseased tissues. Nat Commun 14:7739. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41467-023-43120-6\u003c/span\u003e\u003cspan address=\"10.1038/s41467-023-43120-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanahan D, Folkman J (1996) Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 86:353\u0026ndash;364. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s0092-8674(00)80108-7\u003c/span\u003e\u003cspan address=\"10.1016/s0092-8674(00)80108-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanahan D, Weinberg RA (2011) Hallmarks of cancer: the next generation. Cell 144:646\u0026ndash;674. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2011.02.013\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2011.02.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNagy JA, Chang SH, Dvorak AM, Dvorak HF (2009) Why are tumour blood vessels abnormal and why is it important to know? Br J Cancer 100:865\u0026ndash;869. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/sj.bjc.6604929\u003c/span\u003e\u003cspan address=\"10.1038/sj.bjc.6604929\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSimon T, Gagliano T, Giamas G (2017) Direct Effects of Anti-Angiogenic Therapies on Tumor Cells: VEGF Signaling. Trends Mol Med 23:282\u0026ndash;292. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.molmed.2017.01.002\u003c/span\u003e\u003cspan address=\"10.1016/j.molmed.2017.01.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuong T et al (2012) Genetic ablation of SOX18 function suppresses tumor lymphangiogenesis and metastasis of melanoma in mice. Cancer Res 72:3105\u0026ndash;3114. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1158/0008-5472.CAN-11-4026\u003c/span\u003e\u003cspan address=\"10.1158/0008-5472.CAN-11-4026\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartin JD, Fukumura D, Duda DG, Boucher Y, Jain RK (2016) Reengineering the Tumor Microenvironment to Alleviate Hypoxia and Overcome Cancer Heterogeneity. Cold Spring Harb Perspect Med 6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1101/cshperspect.a027094\u003c/span\u003e\u003cspan address=\"10.1101/cshperspect.a027094\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMartin JD, Seano G, Jain RK (2019) Normalizing Function of Tumor Vessels: Progress, Opportunities, and Challenges. Annu Rev Physiol 81:505\u0026ndash;534. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1146/annurev-physiol-020518-114700\u003c/span\u003e\u003cspan address=\"10.1146/annurev-physiol-020518-114700\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGerhardt H, Semb H (2008) Pericytes: gatekeepers in tumour cell metastasis? J Mol Med (Berl) 86:135\u0026ndash;144. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s00109-007-0258-2\u003c/span\u003e\u003cspan address=\"10.1007/s00109-007-0258-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJeong JH et al (2017) Microvasculature remodeling in the mouse lower gut during inflammaging. Sci Rep 7:39848. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/srep39848\u003c/span\u003e\u003cspan address=\"10.1038/srep39848\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCoutelle O et al (2015) Dual targeting of Angiopoetin-2 and VEGF potentiates effective vascular normalisation without inducing empty basement membrane sleeves in xenograft tumours. Br J Cancer 112:495\u0026ndash;503. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/bjc.2014.629\u003c/span\u003e\u003cspan address=\"10.1038/bjc.2014.629\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKeskin D et al (2015) Targeting vascular pericytes in hypoxic tumors increases lung metastasis via angiopoietin-2. Cell Rep 10:1066\u0026ndash;1081. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.celrep.2015.01.035\u003c/span\u003e\u003cspan address=\"10.1016/j.celrep.2015.01.035\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCooke VG et al (2012) Pericyte depletion results in hypoxia-associated epithelial-to-mesenchymal transition and metastasis mediated by met signaling pathway. Cancer Cell 21:66\u0026ndash;81. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ccr.2011.11.024\u003c/span\u003e\u003cspan address=\"10.1016/j.ccr.2011.11.024\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSounni NE et al (2010) Stromal regulation of vessel stability by MMP14 and TGFbeta. Dis Model Mech 3:317\u0026ndash;332. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1242/dmm.003863\u003c/span\u003e\u003cspan address=\"10.1242/dmm.003863\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eViallard C, Larrivee B (2017) Tumor angiogenesis and vascular normalization: alternative therapeutic targets. Angiogenesis 20:409\u0026ndash;426. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10456-017-9562-9\u003c/span\u003e\u003cspan address=\"10.1007/s10456-017-9562-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao C et al (2016) Jolkinolide B induces apoptosis and inhibits tumor growth in mouse melanoma B16F10 cells by altering glycolysis. Sci Rep 6:36114. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/srep36114\u003c/span\u003e\u003cspan address=\"10.1038/srep36114\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShriwas P et al (2021) A small-molecule pan-class I glucose transporter inhibitor reduces cancer cell proliferation in vitro and tumor growth in vivo by targeting glucose-based metabolism. Cancer Metab 9:14. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s40170-021-00248-7\u003c/span\u003e\u003cspan address=\"10.1186/s40170-021-00248-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWidmer DS et al (2013) Hypoxia contributes to melanoma heterogeneity by triggering HIF1alpha-dependent phenotype switching. J Invest Dermatol 133:2436\u0026ndash;2443. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/jid.2013.115\u003c/span\u003e\u003cspan address=\"10.1038/jid.2013.115\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang ZH, Peng WB, Zhang P, Yang XP, Zhou Q (2021) Lactate in the tumour microenvironment: From immune modulation to therapy. EBioMedicine 73:103627. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ebiom.2021.103627\u003c/span\u003e\u003cspan address=\"10.1016/j.ebiom.2021.103627\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRodig N et al (2003) Endothelial expression of PD-L1 and PD-L2 down-regulates CD8\u0026thinsp;+\u0026thinsp;T cell activation and cytolysis. Eur J Immunol 33:3117\u0026ndash;3126. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/eji.200324270\u003c/span\u003e\u003cspan address=\"10.1002/eji.200324270\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFuglerud BM et al (2022) SOX9 reprograms endothelial cells by altering the chromatin landscape. Nucleic Acids Res 50:8547\u0026ndash;8565. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkac652\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkac652\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao J et al (2024) Distinct roles of SOX9 in self-renewal of progenitors and mesenchymal transition of the endothelium. Angiogenesis 27:545\u0026ndash;560. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10456-024-09927-7\u003c/span\u003e\u003cspan address=\"10.1007/s10456-024-09927-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLaw CW et al (2016) RNA-seq analysis is easy as 1-2-3 with limma, Glimma and edgeR. \u003cem\u003eF1000Res\u003c/em\u003e 5, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.12688/f1000research.9005.3\u003c/span\u003e\u003cspan address=\"10.12688/f1000research.9005.3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiberzon A et al (2015) The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst 1:417\u0026ndash;425. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cels.2015.12.004\u003c/span\u003e\u003cspan address=\"10.1016/j.cels.2015.12.004\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePatel J et al (2013) Prospective surface marker-based isolation and expansion of fetal endothelial colony-forming cells from human term placenta. Stem Cells Transl Med 2:839\u0026ndash;847. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5966/sctm.2013-0092\u003c/span\u003e\u003cspan address=\"10.5966/sctm.2013-0092\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChurch DM et al (2011) Modernizing reference genome assemblies. PLoS Biol 9:e1001091. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1371/journal.pbio.1001091\u003c/span\u003e\u003cspan address=\"10.1371/journal.pbio.1001091\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHao Y et al (2021) Integrated analysis of multimodal single-cell data. \u003cem\u003eCell\u003c/em\u003e 184, 3573\u0026ndash;3587 e3529, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.cell.2021.04.048\u003c/span\u003e\u003cspan address=\"10.1016/j.cell.2021.04.048\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavidson S et al (2020) Single-Cell RNA Sequencing Reveals a Dynamic Stromal Niche That Supports Tumor Growth. Cell Rep 31:107628. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.celrep.2020.107628\u003c/span\u003e\u003cspan address=\"10.1016/j.celrep.2020.107628\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHanzelmann S, Castelo R, Guinney J (2013) GSVA: gene set variation analysis for microarray and RNA-seq data. BMC Bioinformatics 14. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/1471-2105-14-7\u003c/span\u003e\u003cspan address=\"10.1186/1471-2105-14-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKarras P et al (2022) A cellular hierarchy in melanoma uncouples growth and metastasis. Nature 610:190\u0026ndash;198. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-022-05242-7\u003c/span\u003e\u003cspan address=\"10.1038/s41586-022-05242-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHodis E et al (2022) Stepwise-edited, human melanoma models reveal mutations' effect on tumor and microenvironment. Science 376:eabi8175. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1126/science.abi8175\u003c/span\u003e\u003cspan address=\"10.1126/science.abi8175\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCable DM et al (2022) Robust decomposition of cell type mixtures in spatial transcriptomics. Nat Biotechnol 40:517\u0026ndash;526. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41587-021-00830-w\u003c/span\u003e\u003cspan address=\"10.1038/s41587-021-00830-w\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLaw CW, Chen Y, Shi W, Smyth GK (2014) voom: Precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol 15:R29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/gb-2014-15-2-r29\u003c/span\u003e\u003cspan address=\"10.1186/gb-2014-15-2-r29\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRitchie ME et al (2015) limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43:e47. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkv007\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkv007\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiberzon A et al (2011) Molecular signatures database (MSigDB) 3.0. Bioinformatics 27:1739\u0026ndash;1740. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/bioinformatics/btr260\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btr260\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAshburner M et al (2000) Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25:25\u0026ndash;29. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/75556\u003c/span\u003e\u003cspan address=\"10.1038/75556\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamilowski JA et al (2015) A draft network of ligand-receptor-mediated multicellular signalling in human. Nat Commun 6:7866. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/ncomms8866\u003c/span\u003e\u003cspan address=\"10.1038/ncomms8866\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGu Z, Gu L, Eils R, Schlesner M, Brors B (2014) circlize Implements and enhances circular visualization in R. Bioinformatics 30:2811\u0026ndash;2812. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/bioinformatics/btu393\u003c/span\u003e\u003cspan address=\"10.1093/bioinformatics/btu393\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"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":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Melanoma, Vascularisation, Metastasis, Endothelial Progenitor Cells","lastPublishedDoi":"10.21203/rs.3.rs-5457583/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5457583/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe development of new vascular structures is a pre-requisite for tumor growth and spread. This process is often disorganised and produces immature and leaky vessels and relies at least in part on the activity of endovascular progenitor cells (EVPs), residing in vessel walls and giving rise to mature endothelial cells in \u003cem\u003ede novo\u003c/em\u003e blood vessel networks in the tumor. \u003cem\u003eSox9\u003c/em\u003e is a transcription factor that is playing an important role in stem cell self-renewal and fate choice and is highly upregulated in EVPs. In this study, we aimed to explore how \u003cem\u003eSox9\u003c/em\u003e activity in the endothelium affects tumor vascularisation, microenvironment, and metastasis. Indeed, \u003cem\u003eSox9\u003c/em\u003e expression was upregulated in tumor endothelial cells of mice harbouring melanomas. Similarly, we observed the up regulation of SOX9 in human endothelial cells exposed to melanoma cell co-culture or conditioned medium resulting in increased colony formation and reduced maturity as revealed in tube formation assays. Endothelial-specific conditional knockout of \u003cem\u003eSox9\u003c/em\u003e (Sox9fl/fl/Cdh5CreERt2/Rosa-YFP) resulted in a significant reduction in total endothelial cells in B16-F0 or HcMel12 melanoma tumors inoculated intradermally in both flow-cytometry, lineage tracing and immunostaining of tumor sections. Functionally, there was a significant reduction in tumour size and lung metastases after \u003cem\u003eSox9\u003c/em\u003e deletion in the endothelium. Importantly, despite a major reduction in the number and area of CD31\u003csup\u003e+\u003c/sup\u003e vessels there was a significant increase in pericyte coverage suggesting increased maturity of the remaining vessels upon \u003cem\u003eSox9\u003c/em\u003e deletion in the endothelium. These changes in the endothelium translated into a reduction in hypoxia as demonstrated by decreased GLUT1 expression and reduced nuclear localisation of HIF1α. RNA sequencing of sorted tumor cells as well as spatial transcriptomics of tumor sections with endothelial-specific deletion of \u003cem\u003eSox9\u003c/em\u003e versus controls confirmed the reduction in hypoxia and showed dramatic increases in CD4 and CD8 immune T cell infiltration in the centre of tumors as confirmed by immunostaining. In summary, endothelial-specific \u003cem\u003eSox9\u003c/em\u003e deletion resulted in fewer and more mature \u003cem\u003ede novo\u003c/em\u003e vessels in the centre of the tumor and reduced metastatic dissemination, suggesting strategies that target this pathway may restore the normal function of blood vessels in tumors and prevent disease progression.\u003c/p\u003e","manuscriptTitle":"SOX9 reactivation in cancer vessels shapes the tumor micro-environment through hypoxia and immune depletion promoting tumor growth and metastasis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-29 06:40:37","doi":"10.21203/rs.3.rs-5457583/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"954b6b25-6e8e-44b8-904c-0aeddf09edc4","owner":[],"postedDate":"November 29th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":40845556,"name":"Biological sciences/Cancer/Tumour angiogenesis"},{"id":40845557,"name":"Biological sciences/Cancer/Cancer microenvironment"}],"tags":[],"updatedAt":"2025-01-17T19:10:24+00:00","versionOfRecord":[],"versionCreatedAt":"2024-11-29 06:40:37","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5457583","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5457583","identity":"rs-5457583","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}