Compartmentalized VEGF receptor expression in hypothalamic tanycytes reveals a novel non-endothelial axis of VEGF signaling: Tanycytes as a novel non-endothelial target of VEGF signaling
Ombeline Desruelle, Manon Leclerc, Sreekala Nampoothiri, Daniela Fernandois, Claude-Alain Maurage, Markus Schwaninger, Vincent Prevot, Ines Martinez-Corral
TL;DR
This study shows that VEGF signaling occurs in brain cells called tanycytes, revealing a new non-endothelial signaling system in the hypothalamus.
Contribution
The paper identifies a novel, spatially compartmentalized VEGF signaling system in hypothalamic tanycytes, distinct from endothelial cells.
Findings
VEGFR2 is preferentially expressed in ARH-tanycytes, while VEGFR1 is confined to VMH/DMH-tanycytes.
VEGFR2 expression becomes refined postnatally and declines with aging, while VEGFR1 is stable from birth.
VEGFA is broadly expressed in all tanycytes, supporting a paracrine signaling model.
Abstract
Vascular endothelial growth factors (VEGFs) and their receptors (VEGFRs) are critical regulators of angiogenesis and vascular homeostasis. While VEGF signaling has been extensively studied in endothelial cells, emerging evidence suggests it also plays roles in non-endothelial brain cells. However, its spatial and cell-type-specific function within the hypothalamus, and more specifically at the level of the blood/CSF barrier, remains poorly defined. In particular, little is known about VEGF receptor expression in tanycytes, a specialized glial population that lines the third ventricle and regulates body-brain communication within the median eminence (ME), a key neurovascular interface located at the tuberal region of the hypothalamus. We used a multi-modal approach including single-cell RNA sequencing (scRNA-seq) reanalysis, RNAscope in situ hybridization, immunohistochemistry,…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9- —https://doi.org/10.13039/501100000781European Research Council
- —https://doi.org/10.13039/501100001665Agence Nationale de la Recherche
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsNeurogenesis and neuroplasticity mechanisms · Barrier Structure and Function Studies · Single-cell and spatial transcriptomics
Introduction
Vascular endothelial growth factor receptors (VEGFRs) are membrane-bound receptor tyrosine kinases essential for angiogenesis, vascular permeability, lymphangiogenesis, and the regulation of cell proliferation and differentiation. The VEGFR family comprises three main receptors—VEGFR1 (Flt-1), VEGFR2 (Kdr), and VEGFR3 (Flt-4)—which interact with five structurally related VEGF ligands (VEGFA, VEGFB, VEGFC, VEGFD, and placental growth factor [PlGF]). In addition, VEGFs bind with high affinity to neuropilin co-receptors (NRP1 and NRP2) and to heparan sulfate proteoglycans (HSPGs) [1]. Each ligand–receptor interaction activates distinct signaling pathways that are essential for vascular and lymphatic system development and maintenance [1, 2]. However, although historically associated with vascular biology, VEGF/VEGFR signaling also plays diverse roles within the central nervous system (CNS), contributing to blood–brain barrier (BBB) regulation, neurogenesis, synaptic plasticity, and astrocyte function [2–4].
Among the VEGFRs, VEGFR1 has an affinity for VEGFA and is involved in monocyte migration and hematopoiesis. This receptor also acts as a decoy receptor to control the availability of VEGF ligands and is expressed on the surface of monocytes, macrophages, and endothelial cells. VEGFR2 is a critical regulator of VEGF induced angiogenesis, and it is predominantly localized on the surface of endothelial cells, facilitating key signaling processes essential for vascular development. This receptor has a strong affinity for VEGFA and promotes endothelial cell migration, proliferation, differentiation, and survival. On the other hand, VEGFR3 is mainly expressed in lymphatic endothelial cells exhibiting a strong affinity for VEGFC and VEGFD [1]. Importantly, both VEGFR1 and VEGFR2 are also expressed in non-endothelial cells [4–6] including neurons and glia, which can produce and respond to VEGFs [6–8]. Despite their increasing non-endothelial central roles in the brain, the spatial and cell-type-specific expression of VEGFRs in the healthy brain remains poorly characterized.
One region of particular interest in the study of VEGFRs is the median eminence (ME) at the tuberal hypothalamus, a circumventricular organ (CVO) that serves as a critical interface for neuroendocrine regulation. Specifically, the pituitary portal blood vessels in the ME contain endothelial cells that are fenestrated and lack blood-brain barrier properties, allowing circulating molecules to freely diffuse into the brain parenchyma. At this specific location, the blood-barrier functions are instead mediated by tanycytes, which are specialised ependymoglial cells lining the walls and floor of the third ventricle [9, 10]. Tanycytes form a tight junction-based blood–cerebrospinal fluid (CSF) barrier, also known as the tanycytic barrier, and extend long processes that contact both fenestrated vessels in the ME and BBB-protected vessels of adjacent hypothalamic nuclei [11, 12]. Their strategic positioning and contact with different nuclei of the hypothalamus enable them to regulate neurovascular exchange, sense peripheral metabolic cues, and control the entry of circulating signals into hypothalamic neural circuits [13–17] (Fig. 1A). Tanycytes are a heterogeneous population; they have been historically classified into distinct subtypes (α1, α2, β1, and β2) based on their dorsoventral location, molecular identity, and anatomical projections [9]. These different subtypes are believed to play specialized roles in barrier regulation, hormone transport, neurogenesis, and metabolic signaling [9, 18]. Despite this functional heterogeneity, our understanding of how VEGF signaling pathways are differentially distributed across tanycyte subtypes remains limited. A few studies from us and others have reported that VEGF expression in the hypothalamus could change in response to metabolic changes [6, 7], increasing ME’s permeability and access of circulating cues in the area. However, it remains unclear whether and which tanycytic populations express VEGFRs. Additionally, the role of VEGF signaling in tanycyte physiology or neuroendocrine function is unknown.
Fig. 1. Single cell RNA sequencing of VEGFR family in tanycytes. (A) On the left: Schematic of the ME representing its privileged localisation in the hypothalamus. On the right: Correlation between tanycyte subtypes and their corresponding hypothalamic nuclei. Tanycytes are color coded according to the UMAP in B. (B) UMAP showing clusters representative of four main tanycytic subpopulations generated by the reanalysis of scRNA-seq ME dataset from [18]. (C) Feature plots showing VEGFRs (Kdr, Flt1 and Flt4), ligands (Vegfa, Vegfb and Vegfc) and co-receptors (Nrp1 and Nrp2) in tanycytic subclusters. (D) Dot plot showing the expression of the different members of the family across the tanycytic subclusters
This knowledge gap is of particularly relevant in light of the widespread clinical use of anti-angiogenic therapies, including VEGF-neutralizing antibodies and VEGFR2 inhibitors, in conditions ranging from cancer to age-related macular degeneration [19–21]. While these drugs are designed to target peripheral vasculature, their potential effects on VEGF-sensitive glial cells in the brain have not been systematically addressed. In this study, we combined single-cell RNA sequencing (scRNA-seq), fluorescent in situ hybridization, immunofluorescence, and quantitative PCR to map the expression of VEGFR1, VEGFR2, and VEGF ligands across tanycyte subtypes in the ME of mouse and human brain. Our findings uncover a non-endothelial VEGF signaling axis in hypothalamic tanycytes, with distinct spatial and temporal regulation. These insights not only broaden novel roles of VEGF signaling in non-endothelial cells in the brain but also raise far-reaching considerations for the central effects of systemic VEGF-targeted therapies, particularly in the context of aging and metabolic disease in neuroendocrine regulation.
Results
Spatially distinct VEGFR1/VEGFR2 expression in tanycyte subtypes in males and females
To assess the cell-type specificity of Vegfr2 and Vegfr1 expression within the ME, we re-analysed publicly available single-cell RNA sequencing (scRNA-seq) datasets from the adult male mouse hypothalamus [18]. Clustering analysis identified distinct tanycyte subtypes, including β1/β2- and α1/α2-tanycytes, that will be referred to hereafter by their anatomical localization as ME-tanycytes (β2), arcuate nucleus (ARH)-tanycytes (β1, α2), ventromedial hypothalamic (VMH)-tanycytes (α2, α1), and dorsomedial hypothalamic (DMH)-tanycytes (α1)^17^ (Fig. 1 and Supp Fig. 1). Our re-analysis revealed that both VEGFRs, Vegfr2 (Kdr) and Vegfr1 (Flt1), are differentially enriched across tanycyte subtypes. Vegfr2 expression is enriched in ARH-tanycytes, with minimal to no expression in ME-tanycytes and low expression in VMH and DMH-tanycytes. In contrast, Vegfr1 displays a more defined expression pattern, being predominantly expressed in VMH/DMH-tanycytes, while largely absent in ME- and ARH-tanycytes. This spatially segregated expression of Vegfr1 and Vegfr2 was confirmed using a dot plot (Fig. 1D). Interestingly, the third member of the VEGFR family, Vegfr3 (Flt4), was not detected in any tanycytic subpopulations (Fig. 1B-D and Supp Fig. 1). We also examined the expression of co-receptors Nrp1 and Nrp2. Both were expressed across all tanycyte subtypes, although Nrp2 showed relatively lower expression in ME-tanycytes than Nrp1 (Fig. 1B-D and Supp Fig. 1).
In parallel, to investigate the potential source and anatomical distribution of VEGF signaling components within the ME, we analysed the expression patterns of VEGF family ligands. VEGFA mRNA was robustly expressed across all tanycytic populations, with particularly high levels in ME-tanycytes (Fig. 1C-D), which notably lack Vegfr1 and Vegfr2, and in DMH-tanycytes, which express high levels of Vegfr1 (Fig. 1C-D). Other VEGF ligands were also evaluated. Vegfb was expressed in all tanycytic populations, while Vegfc showed only low-level expression if any (Fig. 1C).
We validated the scRNA-seq results by performing in-situ-hybridization (RNAscope) analysis of male coronal brain slices. Our results confirm the differential expression of Vegfr1/Vegfr2 and Vegfa across the various tanycytic populations. Vegfr2 mRNA was prominently localized to tanycytes facing the ARH, while Vegfr1 mRNA was restricted to the VMH/DMH-tanycytes along the ventricle wall (Fig. 2A, Sup Fig. 2A). Notably, Vegfr1 expression and compartmentalization appeared stronger than that of Vegfr2. Furthermore, we observed overlapping expression of Vegfa and Vegfr1 in regions where Vegfr1 was highly expressed (Fig. 2B, Supp Fig. 2B). As expected, based on the scRNA-seq results, Vegfr3 was not detected in tanycytes (Supp Fig. 2E). These spatial patterns were consistent across biological replicates. Signals for Vegfr1,* Vegfr2* and Vegfr3 were also detected, as expected according to the singlecell sequencing data, in endothelial cells of both fenestrated and non-fenestrated blood–brain barrier vessels. Notably, Vegfr2 was more highly expressed in fenestrated vessels than in BBB- vessels, whereas Vegfr1 exhibited a trend toward higher expression in BBB vessels compared to fenestrated vessels, although this difference did not reach statistical significance (Fig. 2C, Sup Fig. 1B, Supp Fig. 2G). In addition, signals for Vegfa were also found in the choroid plexus as previously described [22] (Supp. Figure 2B). These results suggest that VEGFRs and ligand expression are spatially compartmentalized among tanycyte subtypes (Fig. 2E), with ME-tanycytes potentially acting as signaling hubs despite lacking VEGFR expression themselves. This spatial architecture supports a model of directional, inter-subtype communication and suggests a sophisticated VEGF signaling landscape within the ME.
Fig. 2VEGFR1 and VEGFR2 are differentially expressed in tanycytes. (A) In situ hybridization images for Vegfr1 (red) and Vegfr2 (green) in coronal sections of the hypothalamic tuberal region of adult male mice. Vimentin immunoreactivity is shown in white and nuclei counterstaining (DAPI) in blue. Images are low-magnification views of the 3 V at the level of the ME. ME and VMH/DMH regions are magnified on the right. Blue arrows point to ME tanycytes showing no expression of Vegfr1 and Vegfr2. Note that ARH tanycytes express only Vegfr2, while VMH/DMH tanycytes express high Vegfr1 and low levels of Vegfr2. Blood vessels are indicated by asterisks. (B) In situ hybridization images for Vegfr1 (red), Vegfr2 (green) and Vegfa (yellow). High magnification of the boxed region is shown on the right. Individual channels are shown for Vegfr2/Vegfr1 and Vegfa. Note that Vegfa is highly expressed at the level of the ME tanycytes (blue arrows) and VMH/DMH tanycytes (red arrows). (C) Immunofluorescence of ME coronal sections for VEGFR1 (left panel, in red) and VEGFR2 (right panel, in green). Magnification of ME, ARH and VMH/DMH are shown for each immunoreactivity showing VEGFR1 predominantly in the VMH/DMH region while VEGFR2 in the ARH region (white arrowheads). (D) Quantification is shown for VEGFR1 and VEGFR2 protein expression in tanycytes. (E) Schematic image and representation of the ME illustrating the compartmentalization of the VEGFR1, VEGFR2 and VEGFA. Data is represented as a percentage of intensity and corrected to background from ME tanycytes (n = 3). Data are represented as mean ± standard error of the meand (SEM). Statistical analysis was performed using ANOVA and Kruskal-wallis test. n = 3. * = p < 0.05. Scale bar = 100 μm in all panels. ARH, arcuate nucleus of the hypothalamus; ME, median eminence; VMH, ventromedial nucleus of the hypothalamus; DMH, dorsomedial nucleus of the hypothalamus; 3V, third ventricle
To further characterize the spatial architecture of VEGFR expression along the antero-posterior axis of the ME, which extends over 1.2 mm, we performed RNAscope across serial coronal sections spanning from the anterior to posterior ME (Bregma − 1.34 to Bregma − 2.54), creating a detailed anatomical expression atlas (Supp Figs. 3 and 4). This analysis revealed that the compartmentalization of Vegfr1 and Vegfr2 in tanycytes is most clearly defined in the central portion of the ME, where the spatial segregation between ARH- and VMH/DMH-tanycytes is sharply maintained. In contrast, the anterior and posterior ME regions displayed more overlapping expression domains, with some co-localization of Vegfr1 and Vegfr2 mRNA in tanycytes. This rostro-caudal gradient highlights additional heterogeneity within the ME and suggests that VEGFR-mediated signaling may be differentially modulated along its longitudinal axis (Supp Figs. 3 and 4).
Fig. 3. Tanycytic VEGFRs in female mice. (A) In situ hybridization (RNAscope) images for Vegfr1 (red), Vegfr2 (green) and Vegfa (yellow) in coronal sections of the hypothalamic tuberal region of adult female mice. The boxed region is amplified on the right. Note that Vegfa is highly expressed at the level of the ME tanycytes (blue arrows) and VMH tanycytes (red arrows). (B) Left: Schematic representing the qPCR sorting strategy. Right: qPCR analysis of Vegfr2 (Kdr) and Vegfr1 (Flt1) variation during the estrus cycle (n = 6–7). Data are represented as mean ± standard error of the mean (SEM). Statistical analysis was performed using ANOVA and Fishers’ LSD parametric test. * = p < 0.05. Scale bar = 100 μm in all panels. ARH, arcuate nucleus of the hypothalamus; ME, median eminence; VMH, ventromedial nucleus of the hypothalamus; DMH, dorsomedial nucleus of the hypothalamus; 3V, third ventricle; E, estrus: D, diestrus; P, proestrus
Fig. 4. The VEGFRs expression is exclusive to ME tanycytes. (A) Schematic representation of the localization of the different CVOs. (B) In situ hybridization images for Vegfr1 (red) and Vegfr2 (yellow) in coronal sections of different circumventricular organs: Median eminence (ME), Subfornical organ (SFO), Subcommisural organ (SCO), Area postrema (AP) and Organum vasculosum of lamina terminalis (OVLT). The Choroid plexus (CP) is also shown. Collagen IV immunoreactivity is shown in green to identify blood vessels and nuclei counterstaining (DAPI) in blue. Scale bar = 50 μm in all panels. Dotted line in white limits the region of the CVO. The yellow square marks the region magnified in (C). Note that only tanycytes in the ME express both receptors (white arrowheads). Low levels of Vegfr2 can also be detected at the level of the SCO and the epithelial cells of the CP, not overlapping with blood vessels (orange arrowheads). Collagen IV deposits (fractones [53]) are marked with asterisks
Protein expression was further confirmed using immunofluorescence. VEGFR1 and VEGFR2 proteins were detected predominantly at the level of the tanycytic cell body, facing the ventricular lumen, with little to no signal observed along their basal processes. VEGFR2 protein is enriched in ARH tanycytes and VEGFR1 protein is detected mainly at the level of VMH/DMH tanycytes (Fig. 2C-D, Supp Fig. 2A). For VEGFR2, co-staining with the tight junction marker ZO-1, which delineates the ventricular interface, indicated that VEGFR2 signal is enriched at the ventricular-facing side of tanycytic cell bodies. For VEGFR1, receptor signal was assessed in GFP-labeled tanycytes, confirming prominent expression in tanycytic cell bodies at the ventricular wall (Supp Fig. 2C-D). The expression levels of both receptors were substantially lower in tanycytes compared to the strong vascular expression observed in the surrounding vasculature (Fig. 2C, Supp Fig. 2F).
Considering that the VEGFR family can be modulated by estrogens [23, 24], and estradiol can influence VEGFR2 expression by stimulating angiogenesis through VEGFR2 upregulation [24], we next investigated whether the observed compartmentalization pattern is conserved in female mice (Fig. 3). In situ hybridization (RNAscope) on female mice brain sections at estrus and proestrus confirmed the receptor compartmentalization observed in males (Vegfr2 at the level of the ARH and Vegfr1 in VMH/DMH tanycytes) (Fig. 3A, Supp Fig. 5A). We observed a reduced spatial overlap between Vegfr1 and Vegfa expression domains in females, suggesting sex-specific differences in ligand-receptor interactions or signaling microenvironments. To better analyse the levels of Vegfrs in female tanycytes at the different stages of the estrus cycle, we used a qPCR-based approach in isolated tanycytes. R26-tdTomato female mice were stereotaxically injected with AAV1/2-Dio2-Cre into the lateral ventricle to drive Cre-dependent tdTomato expression in tanycytes [25] (Fig. 3B). Two weeks post-injection, animals were sacrificed at defined stages of the estrous cycle (estrus, proestrus and diestrus), and Tomato^+^ tanycytes were isolated by fluorescence-activated cell sorting (FACS). qPCR analysis was then performed to assess Vegfr1,* Vegfr2 and ligand* mRNA expression. The qPCR results revealed no significant fluctuations in Vegfrs or ligand expression in tanycytes during the estrous cycle (Fig. 3B and Supp Fig. 5B).
Fig. 5VEGFA is expressed in all the circumventricular organs. (A–B) Representative RNAscope in situ hybridization images showing Vegfr1 (red) and Vegfa (yellow) mRNA expression across coronal sections of several CVOs: the median eminence (ME), subfornical organ (SFO), subcommissural organ (SCO), and area postrema (AP). The choroid plexus (CP) is also included for comparison. Blood vessels are visualized by Collagen IV immunoreactivity (green), and nuclei are counterstained with DAPI (blue). White dotted lines delineate the boundaries of each CVO. Scale bar = 50 μm for all panels. (B) Higher magnification of boxed regions from (A) to highlight subcellular localization. Yellow arrowheads indicate strong Vegfa expression in ME tanycytes, the SFO, and the CP. Other CVOs (SCO, AP) also exhibit Vegfa signal, though at lower levels. Collagen IV deposits (fractones) are marked by asterisks
VEGF receptor expression patterns are specific to median eminence tanycytes
All the circumventricular organs (CVOs) but one, the subcommisural organ (SCO), are characterized by the presence of a fenestrated endothelium, and tanycyte-like cells forming a barrier (blood-cerebrospinal fluid barrier) [13, 26]. Each CVO is specialized in distinct physiological functions, raising the important question of whether the VEGFR expression patterns identified in the ME are conserved across other CVOs containing tanycyte-like cells. We extended our analysis to include the organum vasculosum of the lamina terminalis (OVLT), subfornical organ (SFO), area postrema (AP), and subcommissural organ (SCO) (Fig. 4A). We also included the choroid plexus (CP) in our analysis, given its epithelial specialization and high secretory activity, despite it not being a CVO in the strict sense [27].
Using in situ hybridization, we evaluated the expression on coronal brain sections, of Vegfr1 (Flt1),* Vegfr2 (Kdr)*, and Vegfa, alongside Collagen IV immunostaining as a marker of vascular extracellular matrix to distinguish vascular from extravascular expression. Surprisingly, neither Vegfr1 nor Vegfr2 was detected in tanycyte-like cells of the OVLT, SFO, SCO, or AP. As expected, endothelial expression of both receptors was evident in blood vessels across all CVOs, confirmed by co-localization with collagen IV, validating the specificity of the RNAscope probes (Fig. 4B-C). We observed, however, low Vegfr2 expression in epithelial cells of the choroid plexus (Fig. 4B-C).
On the other hand, Vegfa was robustly expressed in ventricular and parenchymal cells of both the SFO and OVLT, suggesting that ligand availability alone is not sufficient to drive receptor expression in tanycytes. Additionally, we observed strong Vegfa expression in the CP, consistent with previous reports [22, 28] (Fig. 5A-B). These findings suggest that the coordinated, subtype-specific expression of VEGFA, VEGFR1, and VEGFR2 in tanycytes is a unique feature of the ME, likely reflecting the specialized neurovascular and neuroendocrine functions of this region. This regional specificity supports the notion that tanycyte-like populations are molecularly and functionally heterogeneous across CVOs, and that VEGF signaling plays a particularly specialized role in ME tanycyte-mediated vascular and endocrine regulation.
Developmental refinement of VEGF receptor compartmentalization
Electron microscopy studies from the 1970s demonstrated that the external zone of the ME begins to form during late embryogenesis and continues to mature throughout the first postnatal weeks [29–31] In mice, capillary loops of the fenestrated vessels only penetrate the ME after birth, first appearing towards the end of the first postnatal week and becoming more numerous during the second week [31]. In parallel, tanycytes undergo progressive postnatal differentiation starting around postnatal day 4 (P4) and continuing through P14, with full transcriptional maturity typically reached between P14 and P20 [32, 33]. Given this temporal window of structural and cellular maturation, we wondered whether VEGFR expression is also acquired postnatally. To address this, we investigated the developmental timing of VEGFR expression and compartmentalization within the ME. Specifically, we performed and quantified RNAscope for Vegfr1,* Vegfr2*, and Vegfa in coronal brain sections at different mouse postnatal stages (P0, P4, P8, P12, P16, and P21 just before puberty onset), and compared them to the adult mouse brain (3 months old). (Figs 6 and 7). We observed that both, Vegfr1 and Vegfr2, are already excluded from ME tanycytes at P0. Moreover, Vegfr1 mRNA was already robustly expressed in VMH/DMH-tanycytes at P0, and this expression remained spatially restricted throughout postnatal development, indicating early and stable compartmentalization of Vegfr1. In contrast, the developmental trajectory of Vegfr2 was slightly different; Vegfr2 is expressed in ARH and VMH/DMH tanycytes at P0, indicating an early-established exclusion from the ME tanycyte subtype, however, the preferential expression of Vegfr2 in ARH-tanycytes became more evident in adulthood. This coincided with a slight downregulation of its expression in VMH/DMH-tanycytes of adult mice (Fig. 7B-C). Thus, the full spatial segregation of Vegfr2 among tanycyte subtypes appears to be a postnatal refinement process. As expected, Vegfr1 and Vegfr2 expression in the vasculature was detectable from P0, as confirmed by their presence in vascular structures throughout development (Fig. 6, white arrows).
Fig. 6. Differential expression of Vegfr1, Vegfr2 and Vegfa during postnatal development. In situ hybridization images showing mRNA expression of Vegfr1 (red), Vegfr2 (yellow), and Vegfa (green) in coronal sections of the hypothalamic tuberal region from male mice at postnatal days P0, P4, P8, P12, P16, and P21, compared to adult (3-month-old) animals. The third ventricle (3V) is labeled in all the images for spatial orientation. Note the absence of Vegfr1 and Vegfr2 expression in ME-tanycytes (indicated by a white discontinuous line) at all time points, including at birth, and the emergence of high Vegfa expression in non-tanycytic cells within the ME region beginning around P8 (red arrowheads). White arrowheads at P0 indicate blood vessels. Scale bar = 100 μm in all panels. n = 3 animals per time point. ARH, arcuate nucleus of the hypothalamus; ME, median eminence; VMH, ventromedial nucleus of the hypothalamus; DMH, dorsomedial nucleus of the hypothalamus; 3V, third ventricle
Fig. 7. Quantification of Vegfr1, Vegfr2, and Vegfa during postnatal development. (A) Representative RNAscope in situ hybridization images showing Vegfa (green), Vegfr1 (red), and Vegfr2 (yellow) mRNA expression in the tuberal hypothalamus. White dotted boxes indicate the regions selected for quantification: median eminence (ME), arcuate nucleus (ARH), and ventromedial/dorsomedial hypothalamic regions (VMH/DMH). Scale bar = 100 μm. (B) Quantification of Vegfr1 (red), Vegfr2 (yellow), and Vegfa (green) expression across developmental stages (P0, P4, P8, P12, P16, P21, and adult) in each of the three regions. Expression is represented as signal area (pixels/µm²). Red lines illustrate relative changes across regions for each time point, highlighting spatiotemporal compartmentalization. (C) Summary of expression levels at P0, P21, and adult stages across the three regions. Notably, Vegfa expression shows a marked increase in adult samples, particularly in the ME and VMH/DMH regions, suggesting enhanced ligand availability in mature hypothalamic tissue. n = 3 animals per time point. Statistical analysis was performed using two-way ANOVA followed by Fisher’s LSD post hoc test. p < 0.05 was considered significant
Regarding the ligand, Vegfa mRNA was present in all tanycytes from P0, with a higher expression in ME and VMH/DMH tanycytes starting from P4 (Fig. 6, blue arrows, Fig. 7). Notably, in P8 mice, we identified discrete clusters of cells expressing high levels of Vegfa in the ME parenchyma, suggesting either local signaling shifts or cellular recruitment/remodeling events during this critical period (indicated in Fig. 6 with red arrows). Our findings suggest that Vegfa expression is established early during postnatal development, preceding the full maturation of VEGFR2 compartmentalization. This temporal sequence supports a model in which VEGF ligand availability primes the microenvironment, while receptor expression is dynamically regulated to fine-tune signaling specificity during early postnatal remodeling of the ME.
To contextualize these transcriptional changes with morphological development, we performed vimentin immunofluorescence to visualize tanycyte architecture [12] at the same developmental stages. We observed that at early postnatal time points (P0–P4), tanycytes in the ME lacked the characteristic branched structure shown in adult animals. Beginning at P8, we observed the presence of branched tanycytic endfeet in close apposition to the fenestrated capillaries of the ME, while a single basal process became evident in other tanycyte populations (Supp Fig. 6). This morphological maturation aligns with previous observations [33], and supports the idea that VEGFR dynamics may be linked to tanycyte specialization and vascular interaction.
Age-dependent regulation of VEGF receptor expression
The VEGF system undergoes an age-related decline in signaling efficiency, resulting in reduced angiogenesis, impaired tissue regeneration, and increased susceptibility to age-associated diseases [34]. Dysregulation of this pathway has also been implicated in neurodegenerative disorders such as Alzheimer’s disease, where altered VEGFA signaling contributes to neuroinflammation, vascular dysfunction, and cognitive decline [35–37]. In light of these findings, we sought to explore whether VEGFR signaling in tanycytes is similarly regulated across the lifespan. To this end, we conducted qPCR analysis on FACS-isolated Tomato⁺ tanycytes from microdissected ME explants from male mice at 3, 6, 9, 12, and 18 months of age. Using our AAV approach, we specifically label cells lining the 3V, including all tanycytic subtypes [14]. The Tomato⁻ fraction, which represents a heterogeneous population including but not limited to endothelial cells, was analyzed in parallel as a reference population (Supp Fig. 7). Our qPCR results revealed that, tanycytic Vegfr2 (Kdr) expression is age-regulated, with the highest levels at 3 months of age, corresponding to early adulthood, followed by an abrupt decline starting from 6 months of age (Fig. 8A, Supp Fig. 7). This age-related decrease suggests that VEGFR2-mediated signaling becomes attenuated in tanycytes during aging, which could impact their interactions with the vasculature or influence their neuroendocrine regulatory functions. In contrast, Vegfr1 (Flt1) expression remained stable throughout aging, with a slight increase by 18 months of age, indicating that VEGFR1 may serve a more constitutive or maintenance-related function in tanycytes that is less sensitive to age-related cues. As in previous experiments, Vegfr3 (Flt4) was barely detected in tanycytes at any age, reinforcing the conclusion that VEGFR3 is not an essential component of tanycyte-VEGF signaling, even under potential age-induced stress or remodelling.
Fig. 8. Quantification of Vegfr1, Vegfr2 and Vegfa during postnatal development. (A) qPCR analysis of Vegfr1, Vegfr2, and Vegfr3 mRNA expression levels in tanycytes isolated from male mice aged 3, 6, 9, 12, and 18 months. Notably, Vegfr2 expression shows a progressive decline with age, while Vegfr1 and Vegfr3 remain stable. (B) qPCR analysis of VEGFR ligands. Vegfa and Vegfc remain relatively stable across age, while Vegfb shows a modest upregulation in older mice. (C) Expression of VEGF co-receptors Nrp1 and Nrp2 in sorted tanycytes. While Nrp1 remains unchanged, Nrp2 expression declines with age, paralleling the pattern of Vegfr2. qPCR data are normalized to housekeeping genes 18S and Actb, and presented as mean ± SEM; n = 5–6 mice per time point. Statistical analysis was performed using one-way ANOVA with Fisher’s LSD post hoc test. p < 0.05 was considered significant
We next assessed expression of VEGF ligands (Fig. 8B), which are critical for autocrine and paracrine signaling. Vegfa exhibited relatively stable expression across all time points, with a modest increase in older mice, potentially reflecting a compensatory response to declining receptor levels or vascular alterations with age. Vegfc, the main ligand for VEGFR3, remained largely unchanged. Strikingly, Vegfb expression showed a robust and significant increase from 3 to 18 months, suggesting that VEGFB–VEGFR1 signaling may become more prominent in the ventricular wall of aged mice, possibly as a mechanism to maintain vascular or metabolic homeostasis in the face of Vegfr2 decline. Finally, we evaluated the expression of VEGF co-receptors Nrp1 and Nrp2. Nrp2, which modulates VEGF signaling [38], exhibited a clear age-dependent downregulation, paralleling the decline in Vegfr2 itself, and reinforcing the idea that Vegfr2/Nrp2-dependent pathways are progressively silenced with age. In contrast, Nrp1 expression remained unchanged, suggesting differential regulatory control of neuropilin family members in tanycytes (Fig. 8C).
VEGF receptors are expressed in human tanycytes
Tanycytes are also found in humans, with similar characteristics to the ones present in mice [39, 40]. Thus, we wanted to assess if VEGFR expression is conserved in human ME tanycytes. To that end we decided to analyse the human spatial transcriptomics data from [41] for the different members of the VEGF family (Fig. 9). We observed that indeed, VEGFA and VEGFR2 are highly expressed in the human ME, while VEGFB and VEGFR1 show a more dispersed expression. As in mice, almost no expression was detected for VEGFR3 (Fig. 9A). We further confirmed our observations using RNAscope in situ hybridization in post-mortem hypothalamic tissue from healthy individuals. After validating the specificity of the technique (Supp Fig. 8), tissue samples from the region surrounding the third ventricle were examined for VEGFR1 and VEGFR2 expression. Consistent with our findings in mice, VEGFR2 mRNA was detected in vimentin-positive cells lining the third ventricle, at the level of the ME and ARH (Fig. 9B). In contrast, VEGFR1 mRNA was not detected in this vimentin-positive population (at the level of the ME or dorsally in the ventricle), suggesting a potential species-specific difference in receptor expression or regulatory control. Importantly, both VEGFR1 and VEGFR2 mRNA were readily detected in vascular endothelial cells, confirming the specificity and functionality of the probes used for this analysis.
Fig. 9VEGFRs are expressed in human tanycytes. (A) Spatial transcriptomics data from human hypomap [41] for VEGF receptors, co-receptors and ligands. Arrows point to regions with detected expression (B) RNAscope in situ hybridization for VEGFR1 (yellow) and VEGFR2 (red) in coronal sections of the hypothalamic tuberal region from an adult human male. Immunofluorescence for vimentin (green) was used to identify tanycyte-like cells along the third ventricle. Boxed regions are magnified on the right. VEGFR2 mRNA puncta were observed in vimentin⁺ cells, consistent with tanycyte identity, whereas VEGFR1 expression was restricted to blood vessels (asterisks) and not detected in periventricular vimentin⁺ cells. White arrow points to ME tanycytes, where we observe clear VEGFR2 puncta. Scale bar 100 μm for all panels
To further substantiate these observations, we analyzed human single-nucleus RNA sequencing data from hypothalamic tissue [42]. This independent dataset confirmed robust VEGFR2 expression in β2-tanycyte populations, whereas VEGFR1 transcripts were detectable at a lower level in both β2- but also α2/ β1- tanycytes (Sup. Figure 8B-C). Notably, both VEGFA and VEGFB were highly expressed across all identified human tanycyte populations, consistent with the widespread ligand expression observed in our spatial and RNAscope analyses.
Together, these complementary analyses indicate that VEGFR expression in median eminence tanycytes differs between mice and humans. Vegfr2 is absent or barely detectable in the mouse median eminence, whereas in adult human median eminence tanycytes the receptor is highly abundant, as revealed by spatial transcriptomics, RNAscope, and single-nucleus RNA sequencing. In contrast, Vegfr1 is highly expressed in mouse VMH/DMH tanycytes but shows low expression in human tanycytes. VEGF ligands, including VEGFA and VEGFB, are broadly expressed across human tanycyte populations. Our findings suggest that, while ligand availability is conserved, VEGF receptor utilization by tanycytes differs between species, potentially reflecting differences in median eminence organization, cellular composition, or regulatory context rather than fundamental divergence of the VEGF signaling pathway.
Discussion
Here, we identify a compartmentalized VEGF signaling axis across distinct tanycyte subtypes in the median eminence (ME). VEGFR2 is enriched in β1-tanycytes that interface with the arcuate nucleus (ARH), whereas VEGFR1 is predominantly expressed in α-tanycytes projecting toward dorsomedial and ventromedial hypothalamic regions. In contrast, β2-tanycytes, which directly contact the fenestrated vasculature in the ME, show the highest VEGFA expression yet lack detectable VEGFRs. This ligand–receptor segregation reveals an unexpectedly complex arrangement of autocrine and paracrine signaling within the tanycyte population and underscores the functional specialization of ME tanycytes relative to other CVOs.
The preferential localization of VEGFR1 and VEGFR2 proteins at the ventricular wall, rather than along the basal tanycytic processes, suggests that VEGF signaling in tanycytes is spatially oriented toward the cerebrospinal fluid interface. This polarized distribution is consistent with a role for tanycytes as sensors or integrators of CSF-borne VEGF signals, rather than as mediators of parenchymal VEGF signaling along their distal processes. While our imaging approaches do not allow precise, the confinement of VEGFR signal to the ventricular-facing domain argues for a directional organization of VEGF responsiveness in tanycytes. Defining the ultrastructural localization and signaling context of these receptors will require higher-resolution approaches, but the observed ventricular enrichment provides an important framework for interpreting tanycytic VEGF function.
In this context, our findings suggest that VEGF signaling in tanycytes may extend beyond its well-established endothelial roles in angiogenesis, vascular remodeling, and permeability. Such spatial organization raises the possibility that VEGFR1, enriched in α-tanycytes at the upper ventricular wall, limits the availability of VEGFA in the CSF, thereby indirectly shaping VEGFR2-dependent responses in ARH-facing tanycytes. Alternatively, VEGFR1⁺ or VEGFR2⁺ tanycytes may respond to VEGFA secreted locally by neurons projecting from the lateral area of the hypothalamus (notably MCH-neurons [6]) or astrocytes [8]. These models remain speculative but underscore the potential for region- and subtype-specific VEGF signaling to regulate hypothalamic neurovascular communication.
Notably, we observed co-expression of VEGFA and VEGFR1 within α-tanycytes, a configuration that could bias VEGF signaling toward VEGFR1-specific functions or reflect a local regulatory arrangement distinct from canonical endothelial signaling. The preferential expression of VEGFR2 in ARH-facing tanycytes and VEGFR1 in VMH/DMH-associated tanycytes is conceptually reminiscent of tip- and stalk-cell dynamics during sprouting angiogenesis, where VEGFR2 promotes tip-cell responses to VEGFA while VEGFR1 in stalk cells can limit VEGFR2 activation [43]. However, this comparison is meant to highlight possible functional parallels rather than establish a shared mechanism. By analogy, VEGFR1⁺ tanycytes may serve to buffer VEGFA availability in the CSF, whereas VEGFR2⁺ tanycytes could help maintain the barrier properties of the ventricular wall. Importantly, our data is descriptive and do not directly address functional roles; functional studies will therefore be required to determine whether VEGFR1⁺ and VEGFR2⁺ tanycytes perform roles analogous in any way to endothelial cells.
Our observations suggest the possibility that VEGFR1 and VEGFR2 contribute to distinct and spatially segregated roles of tanycyte subtypes, that may include from region-specific vascular interactions, regulation of barrier permeability, to modulation of endocrine access to hypothalamic circuits. Importantly, the confinement of VEGFR1 expression to α- but not β- tanycytes, further emphasizes that VEGF signaling is tightly compartmentalized, perhaps reflecting the unique physiological demands of different hypothalamic regions. From a translational perspective, uncovering a pathway that regulates tanycytic–vascular interactions at both fenestrated and BBB-protected vessels could provide new opportunities to modulate barrier permeability, for example to facilitate drug delivery. At the same time, recognizing new non-endothelial functions of this pathway may help avoid off-target effects of anti-angiogenic therapies while also opening the possibility to exploit tanycytic VEGF signaling for CNS-directed interventions.
Another striking observation that makes tanycytic VEGF/VEGFR signaling even more interesting is that VEGFR expression pattern is unique to the ME and absent in other CVOs that contain tanycyte-like cells, such as the OVLT, SFO, SCO, and AP^13^. While these regions also contain fenestrated and BBB-protected vessels, their tanycytic interfaces appear molecularly distinct. The high VEGFA expression in OVLT and SCO, but not in the AP, suggests that each CVO employs tailored VEGF-based mechanisms adapted to its physiological roles. The exclusivity of VEGFR1/2 expression to ME tanycytes may therefore reflect the central role of this structure in metabolic sensing and hypothalamic regulation. It is tempting to speculate that specific metabolic cues trigger VEGF release from hypothalamic cells, which then act in a paracrine manner on neighbouring tanycytes to regulate vascular exchange or other, as yet unidentified, tanycytic functions.
Tanycytes express estrogen receptor-α, and cyclical changes in estrogen have been shown to remodel ME barrier properties [44]. Estradiol regulates communication between tanycytes and endothelial cells, while also modulating VEGF expression and VEGFR2 signaling in the brain vasculature [24, 45]. Given these established interactions, we anticipated sex differences in VEGFR expression. Although RNAscope analyses suggested reduced spatial overlap between Vegfr1 and Vegfa in females compared to males, qPCR measurements revealed substantial inter-individual variability across estrous cycle stages, particularly during estrus and proestrus. It is important to note that, given the spatial segregation of VEGF receptors across distinct tanycyte subtypes, qPCR performed on the entire tanycyte population may underestimate subtype-specific fluctuations, as expression changes occurring in a limited subset of cells could be diluted at the population level. Further studies using more selective, spatially resolved approaches will be required to dissect these sex- and stage-dependent regulatory mechanisms in greater detail. Even if preliminary, our observations raise intriguing questions about how hormonal states may intersect with VEGF signaling in tanycytes. In particular, the dramatic decline in circulating estrogen during menopause could influence tanycytic VEGFR function and thereby impact hypothalamic barrier regulation. Because women spend nearly one-third of their lives in a post-menopausal state, which is characterized by the loss of ovarian steroids, understanding whether this dramatic and permanent decrease in circulating estrogens reshapes tanycytic VEGF signaling has broad physiological and clinical implications. Future studies in ovariectomized mouse models, or in hormonally manipulated animals, could directly test whether estradiol regulates VEGFR1/2 expression or function in tanycytes. Such work would clarify whether sex hormones act as modulators of this newly identified VEGF signaling axis and whether menopause represents a critical window for altered tanycytic–vascular communication.
Tanycytes derive from radial glia–like progenitors and mature postnatally into subtype-specific identities [33]. Our data acquired during postnatal development reveal distinct temporal trajectories of VEGFR expression. Vegfr1 is present from birth (P0) and remains confined to VMH/DMH-tanycytes, suggesting early specification of this subtype. In contrast, the restriction of Vegfr2 expression to ARH-tanycytes only becomes obvious in adulthood, consistent with later maturation of this domain. Both receptors are excluded from ME-tanycytes from the earliest stages, indicating that compartmentalization is not a consequence of postnatal differentiation. These temporal patterns support a paracrine model in which ligand and receptor are spatially segregated. Of interest is that clusters of high Vegfa-expressing cells appeared around P8 in the ME parenchyma, coinciding with the first vascular loops described in the ME, raising the possibility that local VEGF signaling contributes to vascular remodeling. Notably, this time window also corresponds to morphological maturation of tanycytes at around P8. Our findings suggest a coordinated program in which VEGF signaling integrates with tanycyte differentiation and vascular development to establish specialized hypothalamic barriers during early postnatal life, remarkable coordination for a small region with such essential roles in metabolism and reproduction. It will also be important to determine whether other barrier systems in the ME, such as the lateral diffusion barrier or perineuronal net barrier [46], participate in this coordinated maturation.
Systemic studies have placed VEGF deficiency at the center of multiorgan aging, and sustained VEGF activity has been shown to promote healthier lifespan [34]. In line with this, we observed a progressive decline in VEGFR2 expression in tanycytes beginning at 6 months, accompanied by a parallel decrease in its co-receptor Nrp2. This temporal pattern is consistent with observations in the vasculature, where aging is associated with reduced VEGF signaling and decreased VEGFR2 expression in endothelial cells [34, 47]. By contrast, VEGFR1 expression remained stable, suggesting a shift from a VEGFR2- to a VEGFR1-dominant axis in tanycytes over time. Whether this transition is adaptive, helping to preserve barrier integrity or tanycytic function, or maladaptive, contributing to impaired neuroendocrine communication, remains an open question. Given that VEGF dysregulation has been linked to Alzheimer’s disease and other disorders of vascular and metabolic dysfunction [35–37]. Tanycytic VEGFR2 downregulation could represent one mechanism linking hypothalamic aging to systemic pathology. These observations highlight tanycytic VEGFR as a potential therapeutic target: enhancing VEGFR2 activity could improve barrier properties or metabolic resilience, whereas systemic anti-angiogenic therapies [48, 49] may inadvertently impair tanycytic function. Indeed, we observed tanycyte fragmentation in Alzheimer’s disease patients [50], consistent with emerging evidence implicating VEGFA–VEGFR2 in neurodegeneration [36, 51].
Importantly, we detected VEGFR2 expression in human tanycytes lining the floor of the third ventricle at the level of the median eminence. In human samples, VEGFR2 and VEGFA were expressed within the same tanycytic populations, whereas VEGFR1 was barely detected. In contrast, mouse median eminence tanycytes show a distinct pattern of VEGFR expression, underscoring species- and region-specific differences in receptor utilization. These findings indicate that, while VEGF ligand availability is broadly maintained, the composition of VEGF receptors expressed by tanycytes differs between mice and humans. Such differences are likely influenced by variations in median eminence organization, cellular context, or regulatory environment rather than fundamental divergence of the VEGF signaling pathway. Together, our results highlight the need for deeper functional characterization of VEGFR signaling in tanycytes to understand how these cells integrate vascular cues and contribute to neuroendocrine regulation at the brain–blood interface.
In conclusion, beyond their well-established roles in vascular biology, our findings demonstrate that VEGFRs are expressed in a spatially organized manner in the distinct hypothalamic tanycyte populations, revealing a previously underappreciated non-endothelial site of VEGF signaling in the brain. This organization adds a new layer of complexity to how systemic VEGF cues may be interpreted at the brain–body interface and underscores the importance of considering central nervous system targets, such as tanycytes, when developing or administering VEGF-targeted therapies.
Methods
Animals
R26-tdTomato^loxP−STOP−loxP^reporter mice (JAX#007914) were obtained from Jackson Laboratory. Wild-type C57BL/6J mice were purchased from Charles River Laboratories. Age details for each experimental group is provided in the corresponding figure legend. Mice were housed in a temperature-controlled facility (21 °C) under a 12 h light/dark cycle, with ad libitum access to water and standard chow (20% of energy from protein, 67% from carbohydrates, and 12% from fat by dry weight). For developmental studies, pups were collected at postnatal day (P) 0, P4, P8, P12, P16, P21, and in adulthood (3 months of age). For aging experiments, adult male mice were sacrificed at 3, 6, 12, and 18 months. All experiments presented in the paper were performed using 3–6 animals per group depending on the experiment. All animal procedures were conducted in accordance with the European Union Directive 2010/63/EU for animal experiments and approved by the Institutional Ethics Committee for the Care and Use of Experimental Animals of the University of Lille and the French Ministry of National Education, Higher Education and Research APAFIS#29172-2020121811279767 v5.
Single-cell RNA sequencing data reanalysis
Single-cell RNA sequencing data from the Gene Expression Omnibus (GEO; accession number GSE90806, PMID: 28166221) were retrieved to analyze the expression profile of VEGF receptors and their ligands in cells of the ARH/ME region, particularly tanycytes. Pre-processed gene counts and metadata were obtained as a SingleCellExperiment object via the CampbellBrainData() function of the scRNAseq package and reanalyzed using Seurat v5.1.0, a toolkit for single-cell genomics in R v4.4.2. Data integration was performed with the Harmony workflow [52] within Seurat. Standard Seurat functions were applied for clustering, identification of cluster-specific markers, and visualization. Major clusters were annotated into distinct cell types using previously reported marker genes [18]. Tanycytes were subset, reclustered (dimensions = 1:16; resolution = 0.6) and annotated to define the four subpopulations based on established marker genes [18]. hUMAP_1 and hUMAP_2 represent the two-dimensional embedding coordinates of cells grouped by similarity in their gene expression profiles.
Human samples
Human brain tissue was obtained in accordance with French regulations (Good Practice Concerning the Conservation, Transformation and Transportation of Human Tissue to be Used Therapeutically, published on December 29, 1998). Dissected blocks of adult human brain tissue containing the hypothalamus were immersion-fixed in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS), pH 7.4, at 4 °C for 1 week. Tissues were then cryoprotected in 30% sucrose in PBS at 4 °C until they sank, embedded in Tissue-Tek OCT compound (Sakura Finetek), frozen on dry ice, and stored at − 80 °C until sectioning. Coronal cryosections were cut at 20 μm thickness for downstream analyses.
Stereotaxic injection into the lateral ventricle
Recombinant AAV1/2-Dio2-iCre (1.25 × 10⁹ genomic particles/µL) was produced as previously described [25]. For female analysis experiments, 2 µL of AAV1/2-Dio2-iCre was stereotaxically injected into the lateral ventricle (anteroposterior: −0.3 mm; mediolateral: ±1.0 mm; dorsoventral: −2.5 mm from bregma) of isoflurane-anesthetized R26-tdTomato mice. Three to four weeks post-injection, mice were sacrificed and the ME was microdissected for downstream cell dissociation and fluorescence-activated cell sorting (FACS) analysis. For aging studies, 2 µL of AAV1/2-CAG-TdTomato (2.5 × 10⁹ genomic particles/µL,VectorBuilder) was similarly infused into the lateral ventricle of wild-type C57BL/6J mice under isoflurane anesthesia, four weeks prior to the age of sacrifice. For all stereotaxic surgeries, mice were anesthetized with isoflurane and placed on a heating pad to maintain body temperature. Eyes were protected with ophthalmic gel (Ocrygel). Injections were performed using a Kopf 963/962 stereotaxic frame, following the coordinates − 0.3 mm posterior to bregma, ± 1.0 mm lateral, and − 2.5 mm ventral from the skull surface. A total volume of 2 µL was infused at a rate of 0.3 µL/min using a KD Scientific Legato 130 syringe pump and a 2 µL Neuro-Hamilton syringe. After injection, the skin was sutured and mice were allowed to recover on a heating pad before being returned to clean cages.
Estrous cycle
To determine the stage of the estrous cycle, vaginal cytology was performed by flushing the vaginal canal with a small volume of sterile saline using a pipette, every day at the same time each day, for at least 2 weeks. The cell suspension was collected and a few drops were placed on a clean glass microscope slide for immediate examination under a light microscope. Estrous cycle phases were identified based on the relative proportions of leukocytes, cornified epithelial cells, and nucleated epithelial cells. The stages were categorized as proestrus, estrus, or diestrus.
Tissue processing and immunostaining
For immunostaining, brains were fixed in 4% PFA/PBS overnight at 4 °C before cryoprotection in 30% sucrose in PBS overnight at 4 °C. Cryoprotected brains were frozen in OCT on dry ice and stored at -80 °C until analysis. For RNA in situ hybridization, fresh brains were immediately frozen in OCT after harvesting and stored at -20 °C. Frozen brains were placed at -20 °C overnight for equilibration and coronally cut on a Leica CM3050S cryostat at 20 μm (sections on slides). Slides were kept at -20 °C until further processing.
Selected sections were dried for 30 min at room temperature before fixing in cold acetone/ethanol (50%v/v) for 1 min. Then, after 3 washes of 5 min with PBS-Triton 0.1%, sections were blocked in incubation solution (ICS, 1% BSA in PBS-Triton 0.3% pH 7.4) for 1 h. Blocking was followed by primary antibody incubation in ICS for 24–48 h at 4 °C. Primary antibodies were then rinsed out before incubation in fluorophore-coupled secondary antibodies for 1 h in ICS at room temperature. Secondary antibodies were washed, and sections counterstained with DAPI (D9542, Sigma). Finally, after a last wash of 10 min in PBS 1X, sections were mounted on slides with Mowiol, left to dry on the bench and stored at 4 C until image analysis.
TargetConcentrationReferenceSupplierRabbit anti-vimentin[1:4000]AB92547AbcamGoat anti-collagen IV[1:500]AB769AbcamGoat anti-mVEGF R1 (Flt-1)[1:100]AF471R&D SystemGoat anti-mVEGF R2 (Flk-1)[1:50]AF644R&D SystemDonkey anti-rabbit (488)[1:300]A21206InvitrogenDonkey anti-goat (488)[1:300]A11055InvitrogenDonkey anti-goat (568)[1:300]A11057Invitrogen
FACs sorting and quantitative PCR
For fluorescence-activated cell sorting (FACS) of hypothalamic tanycytes, mice were injected with either AAV-Dio2-iCre into R26-tdTomato mice or AAV-CAG-tomato into wild-type C57BL/6J mice. Three to four weeks post-injection, median eminences (MEs) were microdissected and enzymatically dissociated using the Papain Dissociation System (Worthington Biochemical Corporation, Lakewood, NJ) to obtain single-cell suspensions.
FACS was performed on a SONY SH800 Sorter Cytometer device using a 70 μm sorting chip. In tdTomato expressing mice, tanycytes were identified based on tdTomato fluorescence (excitation: 561 nm; detection: 675 ± 20 nm). A sample of the cortex (that does not contain any fluorescence) was used as negative control for gating. For each animal, between 4,000/8,000 fluorescent-positive and -negative cells were sorted directly into 10/20 µL of lysis buffer (0.1% Triton X-100, Sigma-Aldrich; 0.4 U/µL RNaseOUT, ThermoFisher Scientific).
Starting with the same amount of sorted tanycytes, for gene expression analysis, total RNA from FACS-sorted cells was treated with DNase I (Invitrogen, ThermoFisher) to remove genomic DNA contamination and then reverse transcribed using SuperScript III Reverse Transcriptase (Invitrogen, ThermoFisher) according to the manufacturer’s protocol. A linear preamplification step was performed using TaqMan PreAmp Master Mix (Applied Biosystems™, Cat. No. 4488593). Quantitative PCR (qPCR) was subsequently performed using the TaqMan^®^ Universal Master Mix II (Applied Biosystems™, Cat. No. 4440049) on an Applied Biosystems QuantStudio 3 Real-Time PCR instrument (ref: A28131) and Applied Biosystems StepOnePlus Real-Time PCR using TaqMan Gene Expression Assays (Applied Biosystems, see list of probes below). Relative gene expression was normalized to housekeeping genes (18S and Actb) and analysed using the 2^-ΔΔCt method. The purity of the sorted cells was confirmed with endothelial (Pecam1), neuronal (Elav4) and tanycytic markers (GPR50). Data are presented as mean ± SEM, and statistical tests are specified in the relevant figure legends.
GeneReferenceActBMm02619580_g1Elavl4Mm01263580_mHGpr50Mm00439147_m1Nrp1Mm01253208_m1Nrp2Mm00803099_m1Pecam1Mm01242576_m1Rn18SMm03928990_g1VEGFaMm00437306_m1VEGFaMm01281449_m1VEGFbMm00442102_m1VEGFcMm00437310_m1VEGFR1 (Flt1)Mm00438992_m1VEGFR2 (Kdr)Mm01222421_m1VEGFR3 (Flt4)Mm01292604_m1
RNAscope
RNAscope was performed by following the user manual RNAscope^®^ Multiplex Fluorescent Reagents Kit v2 Assay (Advanced Cell Diagnostics - ACD) following the manufacturer’s instructions. Briefly, fresh frozen brain sections were first fixed in PFA for 15 min at 4 °C. Slides were dehydrated in different solutions of ethanol (50%, 70% and 100% ethanol). Hydrogen Peroxide was added to each slide in a humid chamber and rinsed. Protease IV was then added to each brain section in a humid chamber, followed by a wash step. Selected probes were prepared and applied to the tissue, then incubated for 2 h at 40 °C in a HybEZ oven. Following hybridization, slides were washed in RNAscope wash buffer and subjected to signal amplification using AMP 1, AMP 2, and AMP 3 reagents, each incubated for 30 min at 40 °C with washes in between. HRP-C1 and HRP-C2 channels were developed sequentially as per the kit instructions, with appropriate wash steps. For mouse, specific probes were used to detect Flt1 (415541-C1, NM_010228.3, target region 756–1663), Kdr (414811-C2, NM_010612.2, target region 1766–2673), and Vegfa (412261-C3, NM_001025257.3, target region 946–2156) mRNAs and Flt4 (481371-C1, NM_008029.3, target region 95–1058). For human, specific probes were used to detect Flt1 (560701-C2, NM_001160031.1, target region 817–1810) and Kdr (435371-C1, NM_002253.2, target region 1400–3515).
Following RNAscope, immunofluorescence was performed on the same sections. Sections were blocked for 1 h in PBS 1X + 0.5% Triton X-100 + 1% BSA and then incubated with the primary antibody in the blocking solution overnight at 4 °C. The following day, sections were rinsed 3 × 10 min in PBS 1X at room temperature. Sections were then incubated with secondary antibodies in blocking solution for 2 h at RT, and washed again 3 × 10 min in PBS 1X. Nuclear staining was performed with DAPI (1:10,000, 5 min), and slides were mounted using Mowiol mounting medium.
Image analysis and quantification
Images were acquired using the Leica Stellaris 5 inverted confocal microscope (Leica Microsystems), using the 20x/0.8 dry objective, the 40x/1.4 oil immersion objective or the 63x/1.4 oil immersion objective. Z-stack were acquired at 0.6 to 1 μm thick optical sections through the whole depth of the section. Tiled images were acquired and automatically stitched using LasX software (Leica Microsystems). Maximum intensity projections were systematically generated using the same software.
Image processing was done using Fiji (ImageJ) software. Regions of interest (ROIs) corresponding to specific tanycyte subtypes were manually annotated based on anatomical landmarks and Vimentin expression. For RNAscope quantification, images were first converted to 8-bit and thresholded to the same value for each image of the same channel. The area and number of pixels of each region were taken to correct the values, as the RNAscope method stipulates that the number of pixels is representative of the number of mRNA (https://acdbio.com/). For immunofluorescence quantification of the intensity of fluorescence, images were acquired with the same settings during confocal microscopy. Mean intensity was measured on hand-drawn ROIs.
Human snRNAseq and spatial transcriptomics data analysis
The expression of VEGFRs and ligands in tanycytes was analyzed using single nucleus RNA sequencing (snRNA-seq) data from human post-mortem brains. The mediobasal hypothalamus (MBH) samples were obtained from 5 different subjects, some of them controls, some of them diagnosed with Alzheimer's disease. Controls included a male subject diagnosed with chronic myelomonocytic leukemia and a female subject diagnosed with type-II diabetes/diffuse atheroma. Nuclei were isolated from the frozen human MBH, libraries were prepared and sequenced as described in ^42^. Following data pre-processing, integration and cluster annotation, the four tanycytic subtypes were subset based on known markers. Gene expression was visualized using scCustomize() and DotPlot() in the Seurat R package (v.5.3.0).
The preprocessed spatial transcriptomics data was retrieved from human hypomap [41] to map the spatial localization of VEGF pathway gene in human tanycytes. Feature plots were generated using the scCustomize package (RRID: SCR_024675).
Statistical analysis
All data are presented as mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism 10, using Student’s t-test, one-way ANOVA test, Kruskal Wallis test, or Fisher’s LSD parametric test. P-values under 0.05 were considered to be statistically significant. All experiments were independently replicated at least three times. The number of mice used per experiment is reported in the figure legends.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Lhomme T et al. Tanycytic networks mediate energy balance by feeding lactate to glucose-insensitive POMC neurons. J Clin Invest 2021;131.10.1172/JCI 140521 PMC 843961134324439 · doi ↗ · pubmed ↗
- 2Monroe BG, Paull WK. Ultrastructural changes in the hypothalamus during development and hypothalamic activity: the median eminence. In: Swaab DF, Schadé JP, editors. Progress in Brain Research. Vol. 41. Elsevier; 1974. p. 185–208.10.1016/S 0079-6123(08)61907-X 4614312 · doi ↗ · pubmed ↗
- 3Neuropilins in. physiological and pathological angiogenesis - Staton – 2007 - The Journal of Pathology - Wiley Online Library. https://pathsocjournals.onlinelibrary.wiley.com/doi/10.1002/path.218210.1002/path.218217503412 · doi ↗ · pubmed ↗