Pan‐cancer landscape of tumour endothelial cells pinpoints insulin receptor as a novel antiangiogenic target and predicts immunotherapy response

Dear Editor, An in-depth study of tumour vasculature can reveal the mechanisms and consequences of dysregulated tumour angiogenesis and influence new treatment strategies.1-3 Here, we establish the first pan-cancer tumour endothelial cell (TEC) atlas, characterize the shared and cancer-specific phenotypes of TECs among diverse cancer types and explore novel antiangiogenic targets. Neo-angiogenesis is an important step in tumour development and metastasis to meet the metabolic remodelling requirements of tumours.4 Anti-angiogenic therapies have been devised to inhibit pathological angiogenesis, but their limited efficiency suggests that the urgency of more effective agents is needed.5 Deconstructing TECs at the single-cell level could assist the development of new antiangiogenic drugs, but the current studies are limited by the narrowed cancer types, small sample sizes, and the relatively low abundance of endothelial cells (ECs) in tissue.1-3 Currently, there is a continuous lack of systematic studies to resolve TECs across different cancer types and in large sample sizes. We established a pan-cancer TEC atlas from the large-scale data of a total of 1.24 million cells derived from 381 high-quality samples (Figure S1A–G and Tables S1 and S2). A detailed description of this atlas is shown in the Supporting Information. Up to 63 320 ECs comprising 37 367 TECs and 25 953 normal ECs (NECs) were aggregated (Figure 1A–C). Clusters were annotated according to the biological functions of seven common EC lineages and the highest-ranking marker genes in TECs and NECs,3 and 22 functional clusters were identified (Figure 1C–E). Hepatocellular carcinoma had a high arterial infiltration, and ovarian cancer had an angiogenic infiltration (Figures 1G and S2). Differentially expressed genes and pathways were used to validate the annotations (Figure S3). This pan-cancer TEC atlas revealed a profoundly altered construction of TECs, with markedly reduced functional components such as capillary ECs and significantly elevated angiogenic components, including tip and immature ECs (Figures 2A,B and S4A,B,I). Enrichment analyses demonstrated that TECs were enriched in pathways associated with angiogenesis, EC development, and EC proliferation (Figure 2C,D). Further investigations confirmed a consistent trend of decreasing capillary ECs and increasing angiogenic ECs across almost all cancers (Figure 2E–J). Scavenging FCGR2B+ capillary ECs were abundant in normal liver and lung6 but were substantially diminished in liver and lung tumours (Figure S4I–K). In addition, the proportion of tip lymphatic EC phenotypes was increased in tumours, indicating enhanced lymph-angiogenesis in tumours (Figure 2G–J). Endothelial tip cells, the leading cells at the tips of vascular angiogenic sprouts, were universally increased in tumours (Figure 2); thus, comprehensive analyses of sprouting tip ECs were performed (Figure 3). Two novel tip EC clusters were newly defined, insulin receptor (INSR)+ tip ECs (C1) and PGF+ tip ECs (C2), which were verified in two independent cohorts (Figures 3A and S5). INSR+ tip ECs (C1), comprising the majority (67.6%) of all tip ECs, were significantly elevated in eight of the 12 cancer types (Figure 3B,C). The marker gene of this cluster, INSR, is a critical metabolic gene.7 INSR+ tip ECs expressed peaked levels of VEGFR1 and VEGFR2 (Figure 3D), suggesting their relation with angiogenesis. The expression of INSR in tumour vasculature was examined across diverse cancer types. Immunohistochemistry staining in 20 types of cancers exhibited a consistent amount of robust enrichment of INSR in TECs (Figure S6B,C). Further multiplex immunofluorescence staining of CD31, INSRβ, and VEGFR2 in 20 types of cancer tissues verified the uniform colocalization of INSRβ staining with CD31/VEGFR2 staining (Figure 3E). This universal pan-cancer expression of INSR bolstered the credibility of INSR as a pan-cancer angiogenic marker. The enrichment analyses of INSR+ tip ECs yielded pathways related to angiogenesis and IGF binding that were upregulated (Figure 4A,B). In the IGF pathway, both IGF1 and IGF2 were significantly correlated with classical EC markers in multiple cancers, but the correlations between IGF1R or IGF2R and EC markers were weak (Figures 4C and S6A). INSR, another high-affinity receptor of IGFs,8 was expressed exclusively in tumour ECs and identified as a novel angiogenic marker (Figure 4D,E). INSR strongly correlated with EC markers in multiple cancers (Figure S7A), indicating the pro-angiogenic effects of IGFs via INSR. Further in vitro matrigel angiogenesis assays demonstrated that overexpression of INSR significantly enhanced the tube-forming capacity (Figure 4F,H), while INSR knockdown impaired the formation of tubes (Figure 4G,I), supporting the enhancement of angiogenic capacity via INSR. Then, we examined the in vivo therapeutic efficacy of BMS-754807, a small-molecule inhibitor of INSR,9 in subcutaneous tumour models of two mouse cell lines. Both LLC and CT26 tumours responded significantly to INSR inhibition, whether or not it was combined with chemotherapy. The efficacy reached maximum when chemotherapy, VEGFR2 antibody and INSR inhibitor were administered simultaneously (Figures 4J–Q and S7B,C). We further used an orthotopic LLC model followed by flow cytometry to investigate the impacts of the INSR inhibitor on EC abundance (Figure 4N–Q). The INSR inhibitor significantly reduced the abundance of ECs in orthotopic tumours, regardless of whether chemotherapy was combined (Figures S7H and 4O). In addition, the INSR inhibitor could significantly lower the expression levels of VEGFR2 and INSR on ECs, and the effects were enhanced when αVEGFR2 was combined (Figure 4P,Q). Intriguingly, chemotherapy increased the EC abundance in orthotopic tumours, and the effects could be reversed by the INSR inhibitor (Figure 4O). When comparing all EC clusters, we discovered that mature venous ECs expressed the highest level of MHC-II molecules (Figure S8A,B). In two scRNA-seq immunotherapy cohorts, we observed remarkably enriched MHC-II+ mature venous ECs in patients with a better response to immunotherapy (Figure S8C–H). In addition, MHC-II+ mature venous ECs were also linked with a significantly favourable prognosis after immunotherapy (Figure S8I–N). In conclusion, we systematically characterized TECs across malignancies. Dissection of TECs revealed a novel INSR-expressing tip TEC cluster as the predominant angiogenic driver. The pervasive presentation on angiogenic TECs implicates INSR as an angiogenic hallmark. The in vitro and in vivo experiments further verified INSR as a novel antiangiogenic target. Furthermore, MHC-II+ mature venous ECs were linked with a better immunotherapy response. Our study of pan-cancer TECs may facilitate the development of innovative antiangiogenic agents. Dr. Ze-Xian Liu, Rui-Hua Xu and Huai-Qiang Ju had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Dr. Yongqiang Zheng, Yi-Qian Pan and Kun Liao contributed equally to this work as co-first authors. Conception and design: Rui-Hua Xu and Ze-Xian Liu. Supervision: Rui-Hua Xu and Ze-Xian Liu. Acquisition, analysis, or interpretation of data: Yongqiang Zheng, Kai Yu, Yanxing Chen and Ze-Xian Liu. Drafted the manuscript: Yongqiang Zheng, Yi-Qian Pan, Yuqing Deng, Huai-Qiang Ju, Ze-Xian Liu and Rui-Hua Xu. Experiments: Yi-Qian Pan, Kun Liao, Yuqing Deng, Qinian Wu, Hui Sun and Huai-Qiang Ju. Critical revision of the manuscript for important intellectual content: All authors. Software: Yongqiang Zheng, Kai Yu, Yanxing Chen and Hengying Pu. Statistical analysis: Yongqiang Zheng, Kai Yu and Ze-Xian Liu. Obtained funding: Ze-Xian Liu and Rui-Hua Xu. We are grateful to all the authors who shared their data online. The flow chart in Figure 1A and the Graphical Abstract were optimized with the help of Colouring (Guangzhou) Co., Ltd. This study was supported by the National Key R&D Program of China (2021YFA1302100), National Natural Science Foundation of China (81972239, 82321003, 81930065, 82173128 and 32370698), Science and Technology Program of Guangzhou (202206080011), Program for Guangdong Introducing Innovative and Entrepreneurial Teams (2017ZT07S096), Tip-Top Scientific and Technical Innovative Youth Talents of Guangdong Special Support Program (2019TQ05Y351), Young Talents Program of Sun Yat-sen University Cancer Center (YTP-SYSUCC-0029), Guangdong Basic and Applied Basic Research Foundation (2023B1515040030), CAMS Innovation Fund for Medical Sciences (CIFMS) (2019-I2M-5-036), Chih Kuang Sholarship for Outstanding Young Physician-Scientists of Sun Yat-sen University Cancer Center (CKS-SYSUCC-2023001), and Cancer Innovative Research Program of Sun Yat-sen University Cancer Center (CIRP-SYSUCC-0004). The authors declare no conflicts of interest. This study wasapproved by the Institutional Review Board of the Sun Yat-sen University CancerCentre (No. B2023-074-01). The animal experiments in this study were approvedby the Animal Experiments Committee of Sun Yat-sen University (No.L025501202303007). Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.