Zeitplan Slot

Tumors are classified as T cell-inflamed (“hot”) or immunologically inert (“cold”) based on immune contexture, particularly CD8+ T cell infiltration. While hot tumors exhibit enhanced leukocyte entry via tumor-associated blood vessels, the role of tumor-associated lymphatic vessels (TA-LVs) in regulating immune responses remains poorly defined. This study investigated how lymphatic endothelial cells (LECs) shape leukocyte dynamics in hot (YUMMER1.7) versus cold (YUMM1.7) melanoma.

Photoconvertible Kikume Green-Red mice bearing YUMM1.7 or YUMMER1.7 tumors were used to isolate tumor-derived LECs and track photoconverted leukocytes from tumors to draining lymph nodes (LNs). Single-cell RNA sequencing (scRNA-seq) was performed to define LEC heterogeneity and identify candidate pathways involved in leukocyte egress. Adhesion molecule expression was validated by flow cytometry. To determine the functional role of lymphatic ICAM-1 and VCAM-1 in tumor control, Prox1-CreERT2; Icam1^fl/fl; Vcam1^fl/fl mice with LEC-specific deletion were grafted with YUMMER1.7 and MC38 tumor models.

scRNA-seq identified eight LEC subpopulations, including capillary, precollector, collector, valve, and proliferative subsets. Comparative analysis of LECs from normal skin, YUMM1.7, and YUMMER1.7 tumors revealed distinct inflammatory programs. While LECs from both tumor types upregulated NF-κB–associated genes, YUMMER1.7 LECs exhibited a stronger IFN-γ response signature. ICAM-1 and VCAM-1 were transcriptionally expressed in LECs from both tumor models and were elevated compared to normal skin. At the protein level, both tumor types showed increased expression relative to skin, with the highest levels observed in YUMMER1.7-derived LECs, consistent with enhanced activation in an immunogenic microenvironment. Functionally, loss of ICAM-1 and VCAM-1 on LECs led to significantly increased tumor growth in both YUMMER1.7 and MC38 models, indicating a role in supporting anti-tumor immunity. This effect is likely linked to impaired migration of antigen-bearing dendritic cells to tumor-draining LNs and reduced T cell priming.

Tumor immunogenicity is associated with distinct LEC activation states and differential adhesion molecule expression. Lymphatic ICAM-1 and VCAM-1 contribute to tumor control, potentially by regulating dendritic cell trafficking and adaptive immune responses.

Peyer’s patches (PPs) are secondary lymphoid organs that monitor the intestinal lumen and initiate adaptive immune responses against pathogens. The efferent lymphatic vessels of PPs serve as the primary pathway for lymphocyte egress from the gut, however little is known about their molecular identity and development.

We performed scRNA-seq of mouse intestinal lymphatic endothelial cells (LECs) from gut and PP samples and used a combination of bioinformatic analysis, immunofluorescence and multiparameter flow cytometry to characterize the different gut LEC populations. We also utilized 3D imaging to examine steps of PP lymphatic development and exploited a nanoindentation technique to provide a map of tissue stiffness across the PP.

PPs harbor a molecularly distinct population of LECs with upregulated gene expression programs related to cell adhesion, complement system, lipid metabolism processes and leukocyte migration. We found four subtypes characterized by unique molecular signatures and spatial distributions, which exhibit both similarities to and differences from previously described lymph node LECs.

We provide a three-stage model of PP lymphatic formation during postnatal development, which includes an initial phase of VEGF-C driven lymphatic expansion, followed by VEGF-C independent steps of remodeling and specialization.

Furthermore, PPs feature a spatial gradient in stiffness, attributed to the presence of diverse immune cell types. We also highlight how these differences emerge during development due to the presence of various mechanical cues, which coincide with PP lymphatic remodeling.

Our findings identify distinct subtypes PP-specific LECs with an insight into subset-specific marker genes, location and functions. We propose a role of mechanical forces in PP postnatal lymphatic remodeling and specialization. We aim to provide a better understanding on the roles of PP lymphatics in the promotion of beneficial mucosal and systemic immune surveillance during postnatal development and in pathological conditions.

Lymphatic vessels (LVs) are essential for fluid homeostasis and immune regulation in the lung. In Chronic Obstructive Pulmonary Disease (COPD), lymphatic remodeling is associated with disease progression, but the interplay between LVs, blood vessels (BVs) and immune cells remains poorly understood due to limitations in imaging and human-relevant models. This project combines 3D light-sheet microscopy (LSM) with advanced in vitro systems to characterize lung vascular networks. We aim to (i) analyze LVs in healthy and COPD lungs, (ii) develop LV-on-chip models, (iii) integrate BV and LV networks, and (iv) investigate inflammatory vascular–immune crosstalk in COPD.

Human non-COPD and COPD lung tissues were obtained from patients undergoing lung cancer surgery. Samples were either paraffin-embedded for histology (hematoxylin & eosin, H&E; immunofluorescence, IF) or punched with 3–5 mm biopsie puncher for staining and clearing using an adapted iDISCO protocol (Bauer A et al., J Exp Med, 2026). BV and LV were identified using vWF (BV marker), LYVE-1 and PDPN (LV markers), and α-SMA to distinguish LV collectors (α-SMA+/LYVE-1+) from capillaries (α-SMA−/LYVE-1+/PDPN+) and large BVs (vWF+/α-SMA+) from capillaries (vWF+/α-SMA−). Samples were imaged by LSM or confocal microscopy. In parallel, a LV-on-chip 3D model was developed using needle patterning to generate hollow channels in fibrin- or collagen-based hydrogels and seeded with primary lymphatic ECs.

Preliminary work established the experimental workflow. Clinical data, including COPD status, were collected. H&E and IF were used to assess tissue architecture, inflammation, LV presence, and antibody performance. Human lung clearing protocols are being optimized by adjusting tissue size and processing to enable whole-tissue visualization of LV and BV architecture. Initial staining identified α-SMA, vWF, PDPN or LYVE-1-positive structures, although uniform labeling throughout the whole tissue volume has not yet been achieved. In parallel, the LV-on-chip model is being optimized.

This work overcomes key technical challenges in studying human lung lymphatics by improving tissue clearing for LSM and establishing primary lymphatic EC-based microphysiological models. These advances provide foundation for human relevant models to investigate LV and BV remodeling and vascular–immune interactions in COPD.

Milroy disease is a congenital primary lymphedema characterized by chronic swelling of lower extremities. Pathogenic variants in lymphangiogenic receptor VEGFR3 and its ligand VEGF-C cause Milroy or Milroy-like lymphedema (Gordon Syndrome), respectively. Here, we report RPL10A encoding a component of 60S ribosomal subunit as a novel causative gene for Milroy-like lymphedema.

We identified six heterozygous RPL10A variants in seven families of patients with Milroy-like lymphedema. Molecular modelling and cellular analyses revealed that the truncated RPL10A variant retaining the longest portion of the C-terminus is unable to interact with rRNA and is rapidly degraded and excluded from nucleoli, the sites of ribosome assembly, indicating a loss of function.

We further found that RPL10A deficiency causes p53-dependent proliferation arrest and p53-independent migration defect in LECs.

While ribosomes are ubiquitously used, the association of heterozygous pathogenic RPL10A variants with congenital lymphatic defects provides a first example of vascular-related ribosomopathy and suggests an unrecognized function of RPL10A in lymphatic vessel development.

Alzheimer’s disease (AD) is a neurodegenerative disorder defined by amyloid-β (Aβ) plaque deposition and tau pathology. Increasing evidence implicates the adaptive immune system in AD, including the presence of clonally expanded effector memory CD8⁺ T cells in the cerebrospinal fluid and blood of patients. The goal of the project is to elucidate how T cells access the central nervous system (CNS) in AD, a disease associated with progressive alterations of brain barrier function.

Using the ArcAβ mouse model, we show that Aβ deposition progressively increases in the brain between 7 and 20 months of age, while the spinal cord remains plaque-free, indicating region-specific pathology. By combining ArcAβ mice with brain border reporter lines (Aqp4-mRuby3; Cdh5-GFP), we mapped T cell localization across distinct barrier compartments.

CD3⁺ (CD4⁺ and CD8⁺) T cells and CD45⁺ immune cells were detected at vascular and perivascular sites, in the choroid plexus, and within the leptomeninges of both ArcAβ and control mice. In contrast, infiltration of CD8⁺ T cells into the brain parenchyma was strongly dependent on age and Aβ burden, suggesting that amyloidosis selectively affects T cell entry beyond CNS interfaces. In ongoing work, we investigate how T cell-mediated neuroinflammation influences AD pathology in the ArcAβ model and employ two-photon intravital microscopy to visualize the anatomical routes of neuro-antigen-specific or control T cells into the brain and spinal cord.

Together, these studies aim to define how alterations of brain barriers regulate T cell recruitment to the CNS in AD and to set the stage for novel therapeutic strategies targeting brain barrier function and immune cell trafficking.

As the body’s first line of defence, skin utilises dermis-resident dendritic cells (DCs) to capture antigens. These cells then migrate through afferent lymphatic vessels (LVs) to lymph nodes, initiating adaptive immune responses. While in mice, the molecular and cellular mechanisms of DC migration into and within lymphatics have been studied extensively, this process remains much less well understood in humans. To bridge this translational gap, we investigate the migration of DCs directly in human skin.

We established a model leveraging the observation that, upon in vitro incubation, dermal DCs get activated and migrate into the lymphatic vasculature. We initially utilised confocal microscopy on smaller skin punch biopsies as a proof of concept. To accurately sample the heterogeneous human skin explants, we increased the acquisition volume by switching to light sheet microscopy. To address manual quantification inefficiencies, we collaborated with computational scientists to develop an AI-based algorithm for LV segmentation and are currently working on deep-learning DC detection to streamline analysis. To compare DC migration in human and murine systems, we utilise exogenous DC crawl-in assays in human and murine tissues.

qPCR analysis demonstrated that upon incubation of skin biopsies, inflammatory cytokines and DC maturation markers, such as fascin1 and CCR7, were upregulated. Initial use of our model with confocal microscopy of smaller explants revealed a profound tissue heterogeneity, with only ~10% of all dermal DCs successfully entering LVs within 40 h. This migration was blocked by adding pertussis toxin during cell culture. Further, we were able to develop a staining approach against both CD11c and HLA-DR to better identify DCs by light sheet microscopy.

Our experimental framework will enable us to track sites of lymphatic entry and capture individual cellular behaviours in highly heterogeneous tissue environments at high throughput. In the future, we aim to uncover the molecular mechanisms governing DC migration and lymphatic entry by using pharmacological inhibitors or activators of DC migration through lymphatics.
Overall, this work advances our understanding of lymphatic trafficking and may inform the design of therapies or novel adjuvants.

The lymphatic system is hierarchically organized. Lymphatic capillaries absorb lymph and drain into larger lymphatic collecting vessels, which transport the fluid back to the blood circulation through contractions of vessel-surrounding lymphatic smooth muscle cells (LMCs). As the largest lymphatic collector, the thoracic duct transports around 75% of the body’s lymph. LMC and their interaction with the adjacent lymphatic endothelial cell (LEC) layer remain poorly understood, partly due to the lack of human LMC-specific markers. Although impaired collecting vessel contractility is increasingly linked to a range of diseases, therapies targeting lymphatic drainage functions are still limited. To address these research gaps, this project aims to study the cellular interactions between the LEC layer and the LMC layer in human lymphatic collectors.

Pieces of human thoracic duct were obtained from esophageal cancer patients undergoing tumor resection. The tissue was characterized at the protein level using immunofluorescence on paraffin sections and cleared tissue analyzed by light-sheet microscopy. Using spatial transcriptomics, we plan to take an exploratory approach to identifying novel LEC markers and a specific LMC marker. In vitro assays will be set up using primary LECs and LMCs isolated from human thoracic duct tissue to validate promising targets from the transcriptomic dataset.

Two-dimensional and three-dimensional immunofluorescence staining revealed the characteristic structure of the human thoracic duct, which had a diameter of approximately 3-5 mm. The LEC layer was CD31/Prox-1 double-positive but CD34 negative and was surrounded by α-SMA/MCAM double-positive LMCs. Furthermore, human thoracic duct tissue was cleared and analyzed with light-sheet microscopy. Following enzymatic digestion, PDPN/CD31/Prox-1 positive primary human LECs were successfully isolated and enriched in-vitro.

Preliminary data of the project showed the structural organization of the human thoracic duct and the feasibility of a subsequent transcriptomics-based characterization. Furthermore, primary LECs and LMCs isolated from the thoracic duct will provide a platform for future functional assays. Overall, this project aims to improve our understanding of LEC-LMCs interaction and lay the groundwork for lymphatic-targeted therapies.

Blood vessels are remarkably specialized structures that exhibit pronounced morphological, functional, and transcriptional heterogeneity. Endothelial cells are highly glycolytic, and cellular metabolism is a critical determinant of their angiogenic capacity. At the same time, vessel specification and maturation occur in parallel, during which endothelial cells gradually adopt a more metabolically quiescent state. This suggests that, even within the same nutrient-rich environment, distinct metabolic programs can coexist side by side. In this project, we investigate how endothelial cells remodel their metabolism during the transition from a pro-proliferative to a pro-quiescent state and how this defines EC fate.

We use the postnatal mouse retina as a model system of vessel development. Previous studies have shown that the pool of proliferating ECs that expands the growing vascular network derives from venous endothelium. ECs that get exposed to the highest levels of VEGF are subsequently selected to become tip cells. Arterial ECs are specified at the tip cell position from where former tip ECs migrate into the artery or capillary system, but rarely back into the vein. Using a tip-cell specific Cre line (Esm1-CreERT2) combined with a RosamTmG reporter, we fate traced tip ECs and their progeny over time. Using scRNA-sequencing combined with computational single cell flux estimation analysis (scFEA), we generated a high-resolution and timely resolved metabolic map of the developing mouse vasculature and identified EC glycolysis as a key pathway that governs arteriovenous fate which we further investigated by targeting the key glycolysis regulator PFKFB3.

While PFKFB3 deletion promoted the migration of tip ECs into the arteries, those arteries were overall less complex and lacked a proper arterial transcription signature. Overexpression of PFKFB3 in tip ECs did not result in any vascular phenotype, as PFKFB3 protein was rapidly degraded via the proteasome. Injection of a proteasome inhibitor (MG132) successfully accumulated PFKFB3, but tip ECs failed to properly segregate into the artery and were increasingly recruited into the vein.

Our data suggests that both transcriptional and translational mechanisms regulate metabolic heterogeneity and quiescence during vessel maturation and that metabolic states co-define arteriovenous fate.

Lymphatic vessels contribute to anti-tumor immunity by transporting tumor antigens and antigen-presenting cells to tumor-draining lymph nodes where immune responses are initiated. Conversely, cancer cells can exploit these vessels to form distant metastases, and recent studies further show that lymphatic endothelial cells (LECs) can influence immune responses by acting as unconventional antigen-presenting cells and expressing immunomodulatory molecules. However, the underlying mechanisms, particularly the receptors and ligands expressed by LECs and their functional consequences, remain poorly understood. Therefore, this project aims to identify and functionally characterize LEC-expressed molecules that may regulate anti-tumor immune responses.
Fibrinogen-like protein 2 (Fgl2) exists in soluble (sFgl2) and membrane-bound (mFgl2) forms with distinct biological functions. Soluble Fgl2 has been described to exert immunosuppressive effects by modulating dendritic cells, T cells, macrophages, and natural killer cells. In contrast, mFgl2 acts as a prothrombinase and may promote fibrin deposition and coagulation, thereby affecting lymphatic homeostasis.

To identify LEC-expressed molecules that may regulate anti-tumor immune responses, we analyzed a single-cell RNA sequencing (scRNA-seq) dataset generated by our laboratory. This dataset comprised LECs and leukocytes isolated from healthy murine skin, as well as from immunologically inert (YUMM1.7) and T-cell-inflamed (YUMMER1.7) syngeneic melanoma models. Following differential gene expression analysis, we employed a series of bioinformatic analyses and literature research to identify suitable candidates, which were then experimentally validated. We are currently establishing functional in vitro assays and generating a lymphatic-specific knockout mouse to determine the functional role of our candidate.

Based on the scRNA-seq data, we identified 1792 differentially expressed genes in LECs across the three conditions. Among these, Fgl2 emerged as a highly expressed candidate. Using Western blotting and ELISA, we found both mFgl2 and sFgl2 in cell lysates and supernatant from cultured LECs. Furthermore, using immunofluorescence, we detected Fgl2 on LECs in mouse ear skin whole mounts, tumor and lymph node sections.

Fgl2 represents a promising LEC-expressed candidate with potential roles in immunomodulation and coagulation.

The endothelial glycocalyx is a carbohydrate-rich surface layer expected to play key roles in vascular biology but its functions are scarcely known. We have previously shown that sialylated O-glycans on lymphatic endothelial cells (LEC) play an important role in postnatal development of the lymph node (LN) macrophage populations. Here, we aim to further investigate the functional roles of such glycans (i) in the organisation of the glycocalyx of LECs and (ii) the mechanisms underlying the observed changes in the composition of the LEC glycocalyx, as well as their consequences for cell-cell interactions of LN LEC with immune cells.

Deletion of Cosmc (C1Galt1C1) and thus functionality of T-synthase, a key enzyme in O-glycan biosynthesis, was induced in either newborn or adult mice. LN of CosmcΔLEC and wild-type littermates as well as cultured LEC derived from LN were analysed by flow cytometry, qRT-PCR, Western blot, and immunofluorescence.

Investigating the impact of Cosmc deletion on the organisation of the glycocalyx of sinusoidal LECs, we observed that cultured LN LEC derived from CosmcΔLEC mice and LECs forming the LN sinus displayed only truncated O-glycans and an altered glycoprotein composition of their glycocalyx not paralleled by changes in the pertaining transcripts. To elucidate the mechanism underlying these changes, we hypothesised that enhanced posttranslational proteolysis and ectodomain shedding of glycoproteins with truncated O-glycans may play a role. Analysis of WT and CosmcΔLEC LEC by Western blot revealed that prominent LEC O-glycoproteins such as CD44 and podoplanin had indeed undergone proteolysis in the LEC from CosmcΔLEC mice whereas other, non-O-glycosylated cell surface proteins such as ICAM-1 remained unchanged.

Our study has defined key functional roles for sialylated O-glycans in shaping the molecular composition of the glycocalyx of lymphatic endothelia. When inducing a functional loss of T-synthase, we observed that LN LEC not only lost their ability to synthesize sialylated core 1 O-glycans but also displayed strongly reduced abundances of several key O-glycosylated glycoproteins, including podoplanin, Lyve-1 and CD44, likely due to increased proteolysis. Further research will identify the mechanisms involved and determine the effects of the altered LN-LEC glycocalyx on cell-cell interactions with monocytes/macrophages.

The blood-brain barrier (BBB), formed by the brain microvasculature, protects the brain against neurotoxins, pathogens, and circulating immune cells. In multiple sclerosis (MS), an autoimmune disease of the central nervous system (CNS), a compromised BBB is a characteristic early hallmark diagnosed by contrast enhancing lesions observed by MRI. Individuals with radiologically isolated syndrome (RIS), have incidental MRI findings characteristic of MS but show no clinical manifestations of MS. Since approximately half of these individuals later develop MS, individuals with RIS provide a unique opportunity to investigate early disease mechanisms at the level of the BBB.
This study aims to investigate transcriptional changes at the BBB that may be associated with early disease mechanism and the transition from RIS to MS.

Using human induced pluripotent stem cells (hiPSC) differentiated into brain microvascular endothelial cells (BMECs), bulk RNA sequencing (bulk RNA-seq) was performed from two individuals, each sampled longitudinally at two time points: following the diagnosis of RIS, and subsequently after the diagnosis of MS. The sequencing data were processed by performing quality control, normalization, differential expression analysis, and pathway enrichment. Pairwise differential expression analysis was performed between RIS and MS time points using the paired samples. This longitudinal design enabled the investigation of early disease mechanisms and the transcriptional changes associated with disease progression.

The differential gene expression analysis identified over 800 differentially expressed genes (adjusted p-value, FDR < 0.05) of which more than half were upregulated in MS compared to RIS. The additional pathway enrichment analysis identified pathways involved in the inflammatory response as activated in MS, whereas the genes involved in mesenchymal transition pathways are enriched in RIS.

Although limited by the small cohort size, this longitudinal design provides preliminary insights into transcriptional changes at the level of the BBB that are associated with disease progression. These findings suggest that BBB transcriptional alterations are already present in RIS and the progression from RIS to MS is potentially associated with an activated inflammatory response.

The blood-brain barrier (BBB) is a protective cellular interface that tightly controls the exchange of molecules between the bloodstream and the brain, thus preserving central nervous system (CNS) homeostasis. This barrier is established by biochemically specialized brain microvascular endothelial cells (BMECs), which also regulate the trafficking of immune cells into the CNS. However, barrier function does not arise from BMECs alone, it depends on continuous interaction with pericytes and astrocytes, together forming the neurovascular unit (NVU). BBB integrity becomes compromised in neurodegenerative and neuroinflammatory conditions such as Alzheimer's disease (AD) and multiple sclerosis (MS). This project employs human induced pluripotent stem cells (hiPSCs) derived from AD and MS patients to model the BBB and NVU in disease contexts.
Apolipoprotein E4 (ApoE4), the primary genetic risk factor for AD, has recently been implicated in BBB dysfunction. We have developed protocols for differentiating hiPSCs into BMECs, pericytes, and astrocytes, enabling the generation of fully isogenic NVU models. Using this approach, we established isogenic NVU models from hiPSCs of healthy controls (HC) homozygous for ApoE3 and AD patients homozygous for ApoE4, in order to examine how ApoE4 affects BBB integrity.

Permeability assays, TEER measurements, immunofluorescence staining, multi-color flow cytometry, organ-on-a-chip

So far we have successfully differentiated and charachterized BMECs from all four clones. Independently of donor genotype cells exhibit characteristic tight and adherens junction expression, form a functional barrier and respond to inflammatory stimuli.
This is aligned with the existing data and literature, as in NVU APOE is mainly expressed by astrocytes and pericytes. Thus, the addition of these cells to our model is essential.

These NVU models will next be integrated into microscale devices featuring nanoporous ultrathin silicon nitride membranes (μSiM), whose modular design supports co-culture of all three cell types for in vitro NVU modeling. The μSiM platform will be further enhanced with a flow module to examine how shear stress influences BMEC behavior and immune cell interactions with the NVU under physiologically relevant flow conditions.

The brain is protected by three meningeal layers. The outermost layer is the dura mater, next the arachnoid mater lines the roof of the cerebrospinal fluid (CSF)-filled subarachnoid space (SAS) and contains an outer arachnoid barrier (AB) layer in addition to an inner arachnoid layer. The pia mater covers the brain parenchyma. Recent studies identified the regions where bridging veins (BVs) cross the AB to join the dural venous sinuses, as “arachnoid cuff exit” (ACE) points, allowing for bidirectional CSF exchange and immune cell trafficking.

2P-IVM, through skull imaging, immunofluorescence staining, electron microscopy.

Using fluorescent brain barriers reporter mice, we here show that the BVs first traverse the inner arachnoid cell layer and only cross the layer of AB fibroblasts closer to the dural sagittal sinus. We further confirm in vivo the previously described endothelial phenotypic transition of BVs - from claudin-5 to PLVAP expression- and show it occurs only after these enter the dura and are no longer ensheathed by E-cadherin expressing AB cells. Following cisterna magna infused tracer by through skull imaing we first observed the tracer at 10-30 minutes in the dorsal dura mater prior to reaching the ACE points, instead of reaching them via the SAS. Furthermore, observing the distribution of 10 kDa fluorescent vascular tracers by two-photon intravital microscopy (2P-IVM) in VE-cadherin-GFP reporter mice showed that following their diffusion into the dura mater they fail to cross into the SAS at ACE points within one hour.
ACE points have also been proposed as sites of myeloid cell trafficking during neuroinflammation, however these observations only showed transmigration across the inner arachnoid cell layer. Using 2P-IVM, we did not observe T cells crossing the VE-cadherin-GFP expressing arachnoid mater fibroblasts at ACE points under healthy or neuroinflammatory conditions. Ongoing studies applying exogenous chemokines aim to determine if immune cell trafficking across ACE points can be initiated.

Together, our findings support a revised model in which the arachnoid forms a tightly sealed cuff around bridging veins. Rather than contsituting ACE points, these structures function as arachnoid cuff seals (ACSs) that preserve barrier integrity, limit fluid exchange and immune cell passage between the SAS and dura mater.

The induction of microvascular networks plays crucial roles for tissue regeneration. Since vascular invasion by the host is a slow process, it is attractive to integrate a vascular compartment in vitro before in vivo implantation (pre-vascularization), as well as to generate engineered tissues of physiologically relevant complexity. However, a significant challenge remains in achieving in vivo long-term survival and functional performance of such pre-formed microvascular networks.

Based on principles derived from vascular biology, we generated a bioactive fibrin hydrogel engineered with a cross-linkable form of Vascular Endothelial Growth Factor (TG-VEGF) to provide a sustained angiogenic signaling microenvironment to vascular cells.

An optimized dose of TG-VEGF significantly promoted in vitro microvascular self-assembly of human endothelial and pericyte-like cells, as efficiently as by supplementing soluble VEGF in the culture. However, upon in vivo implantation soluble VEGF could not sustain the pre-assembled networks and led to their significant regression, similarly to the no-VEGF condition. In contrast, TG-VEGF decoration effectively supported survival and expansion of the human network. Further, only TG-VEGF could stimulate ingrowth by the mouse vasculature and promote hybrid vessel formation, with rapid anastomosis to the host circulation and functional perfusion of the human pre-assembled networks. Mechanistically, we found that TG-VEGF could sustain active pERK signaling in the pre-assembled endothelium and also specifically recruit a pro-stabilizing population of Neuropilin-1-expressing monocytes.

In conclusion, engineering of the signaling microenvironment is a powerful tool to support the in vivo persistence and functional inosculation of self-assembled microvascular networks. In perspective, the flexibility afforded by this vascularization module could be exploited both for in vivo regeneration and in vitro disease models with physiological complexity.