Endothelial cells (ECs) not only form a semi-permeable monolayer with anti-thrombotic properties, but also release angiocrine factors orchestrating organ development, regeneration, and metabolism [1]. ECs form organ-specific niches that control organ homeostasis [2]. This concept of angiocrine signaling revolutionized our understanding of blood vessels during the past decade. Notably, EC-released factors not only act locally, but can also act systemically when released into the blood stream [3]. As such the endothelium, which represents one of largest internal surfaces of the human body, plays crucial roles in the maintenance of body homeostasis, but is also critically involved in the pathogenesis of life-threatening diseases such as cancer [1, 2, 3, 4, 5].

In recent years, remarkable insights into the function of ECs have been made, for example demonstrating that endothelial dysfunction is a central pathophysiological mechanism of coronavirus disease (COVID-19) [6]. Single-cell technologies revealed a fascinating picture of organ-specific endothelial heterogeneity [7]. Most notably, a landmark study demonstrated that preventing capillary rarefaction dramatically prolonged the lifespan and improved health in aged mice [8], indicating unexpected roles of ECs for preserving organ function throughout life.

This brief review article addresses recent publications that broadened our understanding of vascular ECs in the context of cancer progression, metastasis, and therapy resistance. It will not address the important work on lymphatic ECs due to space limitations.

The concept of tumor angiogenesis [9] led to the discovery of cancer-secreted endothelial growth factors such as VEGF, and thus to the development of anti-angiogenic drugs targeting VEGF or its receptors, which are usually given in combination with chemotherapy [10]. Their effect on overall survival is limited mostly because of primary or acquired therapy resistance. Recent clinical trials showed that anti-angiogenic drugs can strongly enhance responsiveness to immunotherapy in several cancer entities. Surprisingly, the mode of action is still unclear. There is evidence that anti-angiogenic drugs lead to regression of immature vessels in tumors, thereby reducing vascular leakiness and blood shunting. This process called “vessel normalization” allows better tissue perfusion and thereby a better delivery of anti-cancer drugs [11]. In addition, abnormal tumor blood vessels promote an immunosuppressive cancer microenvironment, while normalized tumor blood vessels favor a more immunostimulatory condition [10, 11, 12]. Consequently, blocking the vascular destabilizing factor angiopoietin-2 for example enhances vascular integrity in the tumor periphery, leading to effector T cell infiltration into the tumor core and a subsequent better response to checkpoint inhibitor immunotherapy [13]. This indicates that targeting other angiogenic factors than VEGF could offer new therapeutic possibilities.

It seems possible that therapies targeting ECs could not only lead to vessel regression and/or normalization but also to additional beneficial effects. In pancreatic cancer mouse models, vascular remodeling sensitizes to chemotherapy and immunotherapy [14]. Recent studies identified specific EC subtypes with immunomodulatory potential which may contribute to immunosuppression in cancer [15], but their function will have to be investigated before making a definitive statement. Anti-angiogenic drugs in combination with immune-modulating therapies promote trans-differentiation of postcapillary venules into inflamed high-endothelial venules (HEVs) [16]. This is important as HEVs enable influx of T cells into the tumor mass, which strongly improves immunotherapy [17].

Also, outside of HEVs, tumor ECs play a crucial role for the entry of immune cells from the blood stream into the tumor microenvironment and tumor ECs also execute immunomodulatory functions [18]. ECs interact with immune cells by secreting cytokines or by direct ligand–receptor interactions (Figure 1). For instance, ECs regulate leukocyte extravasation through expression of adhesion molecules (selectins, ICAM1, VCAM1, PECAM1) and through weakening of endothelial cell–cell contacts allowing transmigration of immune cells [4,19,20]. ECs influence not only immune cell recruitment but also immune cell activation within the tumor microenvironment. For example, tumor ECs control polarization of tumor-infiltrating monocytes into immunosuppressive macrophages through CXCL2 secretion [21]. In glioblastoma, the tumor endothelium secretes IL-6, which generates an anti-inflammatory and pro-tumorigenic macrophage activation [22]. This clearly suggests that tumor ECs play numerous roles in protecting cancer cells from the immune attack. Consequently, during immunotherapy, ECs can limit T cell activation [23]. As such, simultaneous targeting the vasculature during immunotherapy can be beneficial [14], and better understanding the mutual interactions between ECs and immune cells will be key to further improve immunotherapy in the future.

The major hurdle in anti-angiogenic cancer treatment is primary or acquired therapy resistance. The underlying processes (Figure 2) are still not completely understood and there is a lack of suitable and widely accepted biomarkers [10]. Single-cell RNA sequencing (scRNAseq) revealed a remarkable endothelial heterogeneity across many organs [24], which becomes altered in diseases [25]. scRNAseq of human and murine lung cancer samples revealed that anti-angiogenic therapy only targets a subset of tumor ECs [26] which might explain the limited therapy efficiency. Single-cell technologies may help to explore novel targets and vulnerabilities of the tumor endothelium. In breast cancer, distinct subsets of tumor ECs coordinate cancer metabolism by lipid handling and transport towards cancer cells [27]. Several recent publications demonstrated that ECs fulfill critical roles in coordinating the transport of nutrients from blood plasma to parenchymal cells [5] and it can be foreseen that this also plays a major role in coordinating cancer metabolism. Lastly, cellular metabolism also differs in tumor vs. non-tumor ECs. For example, glucose uptake and glycolysis are higher in tumor vs. non-tumor ECs [28,29]. Therefore, it is conceivable that endothelial metabolic vulnerabilities could serve as potential therapeutic targets. However, this field is still poorly understood.

Tumors may acquire resistance towards anti-VEGF drugs by secreting other pro-angiogenic factors [10]. However, it is also important to note that generally not all tumor blood vessels are generated by VEGF-driven angiogenesis since some tumors and in particular metastases grow by co-opting pre-existing vessels (vessel co-option). Such co-opted vessels are widely considered to be resistant against anti-VEGF therapy [30]. A recent publication shed some light into the resistance mechanism and showed striking differences in the transcriptional landscape, WNT signaling activation and cellular metabolism of ECs in newly formed vs. co-opted blood vessels of liver metastases [31], a finding that may be used to develop treatment strategies to prevent the latter. In certain tumors, it was also assumed that cancer cells may form tube-like structures which get connected to blood vessels (vascular mimicry), and such structures would be resistant against anti-VEGF therapy [10].

Interestingly, also the stiffness of the tumor tissue appears to play a major role in resistance towards anti-angiogenic therapy, and targeting cancer-associated fibrosis may restore response rates towards to antiangiogenic therapy [32]. Further research is needed to not only decipher the molecular mechanisms of anti-angiogenic therapy resistance but also to identify robust biomarkers for early detection and eventually ways to overcome resistance mechanisms.

ECs dynamically provide secreted factors involved in orchestrating their microenvironment [1]. Alterations of the angiocrine landscape in tumor ECs (Figure 3) contributes to intravasation and extravasation of cancer cells as well as to the cancer stem cell phenotype, epithelial-to-mesenchymal plasticity, and immunosuppression [4]. Recent work showed that in glioblastoma, tumor ECs secrete extracellular vesicles as drivers of proneural-to-mesenchymal reprogramming of cancer cells, thus facilitating their invasiveness and chemoresistance [33].

The endothelium in the bone marrow is a major component of the hematopoietic niche. Specialized ECs control the differentiation of hematopoietic stem cells by angiocrine factors [1]. Recently, a study showed that a subset of ECs in the bone marrow specifically promotes differentiation of monocyte-dendritic cell progenitors by secreting colony-stimulating factor-1 [34]. ETS, SOX, and nuclear hormone receptor families are essential transcription factors specifying ECs within such niches [35]. A general notion is that cancer cells seem to hijack such endothelial niches and thus acquire, for example, a cancer stem cell phenotype and therapy resistance [36, 37, 38∗, 39]. This is of utmost importance as the bone marrow is a frequent site of metastasis. While several signaling cascades play an important role in the intimate EC-tumor cell cross talk, Notch signaling appears to be central to coordinating both endothelial and cancer cell properties [21,33,36,40, 41, 42, 43]. Further research is needed to explore whether targeting Notch signaling in the context of tumor angiocrine signaling is beneficial.

Metastasis is the major cause of death in cancer patients. Cancer cells interact with ECs, which may alter their function to facilitate metastasis e.g., by transmigration into the bloodstream. One recently described example of this is the release of double-stranded RNA from tumor cells, which activates the RNA-sensing receptor TLR3 to induce expression of endothelial SLIT2 expression. This in turn acts as a chemokine to promote cancer cell migration across the vessel wall and into the blood stream [44].

It is of utmost importance to notice that not only within the primary tumor, but also at distant sites, ECs also play a major role in cancer cell extravasation, survival, and proliferation [45]. Mechanistically, activation of endothelial Notch signaling plays a crucial role in the entire metastatic cascade [21,33,36,40, 41, 42, 43]. Furthermore, cancer cells may also interfere with endothelial tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) at the premetastatic niche (before arrival of circulating cancer cells) which activates local ECs [46]. The activation of the quiescent endothelium occurs e.g., during wound healing and inflammation. Activated ECs alter their angiocrine landscape, weaken cell-to-cell contacts and express leukocyte adhesion factors, thus recruiting immune cells and responding well to growth factors that enable angiogenesis. Cancer cells hijack this system and, through EC activation, cancer cells transmigrate through the endothelium within the primary tumor to enter the bloodstream, adhere to the endothelial surface and extravasate during colonization of distant organs [10]. In this process, tumor-induced expression of the leukocyte adhesion molecules VCAM1 and ICAM1 on ECs play a major role in the transmigration within the primary tumor mass to reach to bloodstream but also at distant sites, which promotes extravasation and survival of circulating cancer cells [43,47].

It should be noted that many cancers cause a systemic pro-thrombotic pro-coagulant condition [3], and that blood coagulation can drive cancer cell arrest in the circulation, thereby enabling subsequent extravasation and metastasis [48,49]. Besides blood clotting or manipulating the endothelium at distant sites, circulating cancer cells may also induce necroptotic EC death during colonization [50,51]. For this, circulating cancer cells do not need to pass the endothelial barrier through cell-to-cell junctions, they simply generate large gaps in the vessel wall. Whether this mechanism plays a major role in metastatic spreading in human patients needs further investigation.

The majority of extravasated cancer cells stay in a dormant state within a vascular niche for long time and it is poorly understood what leads to their activation and proliferation. Upon outgrowth of an overt metastasis, remodeling of the local vasculature occurs. A recent publication describes how breast tumor cells lead to rapid expansion of certain blood vessels within the bone marrow which subsequently allow expansion of metastases [52]. The formation of vascular niches that support survival of disseminated cancer cells can further be influenced by local macrophages responding to the presence of cancer cells. Such macrophages signal to the vasculature by the release of tenascin-C leading to EC activation and overt outgrowth of metastases [53].

Besides the well documented local interactions of disseminated cancer cells with ECs within the metastatic niche there is convincing evidence that tumors already prime the pre-metastatic niche before the arrival of cancer cells. This occurs through the release of cytokines, nucleic acids, and a plethora of extracellular vesicles [54], which e.g., change systemic metabolism [55]. It can be expected that such factors first impact the endothelium at distant sites to prepare them for the arrival of CTCs. Notably, the primary tumor reprograms the endothelium in the entire body, affecting thereby a large surface area, amplifying tumor signals and facilitating cancer cell colonization [3]. This is achieved e.g., by the induction of leucine-rich alpha-2-glycoprotein-1 (LRG1) in ECs [56]. The protein LRG1 appears as a highly interesting target as it causes ab-normalization of the blood vessel architecture, promotes colonization of cancer cells, and inhibits immunotherapy. Recent breakthrough publications showed that all of this can be prevented by treatment with neutralizing LRG1 substances [56,57]. Interestingly, even microorganisms can be involved in conditioning of the premetastatic niche. Gut bacteria, which enter the blood stream due to disturbed vascular barriers in colorectal cancer, can alter the liver vasculature to favor the colonization by circulating colorectal cancer cells [58]. Putting this into context one might wonder whether generating a pro-inflammatory endothelial micromilieu is—independent of its initiating factor—a principal mechanism in forming the premetastatic niche. However, this is speculation at this moment.

In summary, recent breakthrough discoveries elucidated cancer-induced vascular changes both within the primary tumor and the (pre-)metastatic niche and these will serve as valuable basis for future clinical studies aimed at reducing metastatic spreading, the major cause of suffering and death in patients with cancer.