Search

Cortical interneurons establish inhibitory microcircuits throughout the neocortex and their dysfunction has been implicated in epilepsy and neuropsychiatric diseases. Developmentally, interneurons migrate from a distal progenitor domain in order to populate the neocortex - a process that occurs at a slower rate in humans than in mice. In this study, we sought to identify factors that regulate the rate of interneuron maturation across the two species. Using embryonic mouse development as a model system, we found that the process of initiating interneuron migration is regulated by blood vessels of the medial ganglionic eminence (MGE), an interneuron progenitor domain. We identified two endothelial cell-derived paracrine factors, SPARC and SerpinE1, that enhance interneuron migration in mouse MGE explants and organotypic cultures. Moreover, pre-treatment of human stem cell-derived interneurons (hSC-interneurons) with SPARC and SerpinE1 prior to transplantation into neonatal mouse cortex enhanced their migration and morphological elaboration in the host cortex. Further, SPARC and SerpinE1-treated hSC-interneurons also exhibited more mature electrophysiological characteristics compared to controls. Overall, our studies suggest a critical role for CNS vasculature in regulating interneuron developmental maturation in both mice and humans., Competing Interests: MG, DA, GC, PD, AM, MQ, AM, SB, HF, CZ, PC, GH, DA, JG, EA No competing interests declared, (© 2021, Genestine et al.)

Melanin-concentrating hormone (MCH)-expressing neurons are key regulators of energy and glucose homeostasis. Here, we demonstrate that they provide dense projections to the median eminence (ME) in close proximity to tanycytes and fenestrated vessels. Chemogenetic activation of MCH neurons as well as optogenetic stimulation of their projections in the ME enhance permeability of the ME by increasing fenestrated vascular loops and enhance leptin action in the arcuate nucleus of the hypothalamus (ARC). Unbiased phosphoRiboTrap-based assessment of cell activation upon chemogenetic MCH neuron activation reveals MCH-neuron-dependent regulation of endothelial cells. MCH neurons express the vascular endothelial growth factor A (VEGFA), and blocking VEGF-R signaling attenuates the leptin-sensitizing effect of MCH neuron activation. Our experiments reveal that MCH neurons directly regulate permeability of the ME barrier, linking the activity of energy state and sleep regulatory neurons to the regulation of hormone accessibility to the ARC., Competing Interests: Declaration of Interests The authors declare no competing interests., (Copyright © 2020 The Authors. Published by Elsevier Inc. All rights reserved.)

The arcuate nucleus (Arc) integrates circulating hormonal and metabolic signals to control energy expenditure and intake. One of the most important routes that enables the Arc to sense circulating molecules is through the median eminence (ME), which lacks a typical blood-brain barrier. However, the mechanism by which circulating molecules reach the Arc neurons remains unclear. This review focuses on what is known to date regarding the special structure and permeability of the ME vasculature and active transport of circulating molecules from the ME to the Arc. Recent studies have demonstrated that the ME displays angiogenic behavior that is expected to provide high vascular permeability. Parenchymal diffusion of circulating molecules from the ME vasculature is size-dependent, and tanycytes actively transport circulating molecules from the ME to the Arc. Finally, we highlight structural plasticity of the Arc and ME as playing an important role in maintaining energy balance homeostasis., (Copyright © 2019 Elsevier Ltd. All rights reserved.)

The neocortex contains glutamatergic excitatory neurons and γ-aminobutyric acid (GABA)ergic inhibitory interneurons. Extensive studies have revealed substantial insights into excitatory neuron production. However, our knowledge of the generation of GABAergic interneurons remains limited. Here we show that periventricular blood vessels selectively influence neocortical interneuron progenitor behavior and neurogenesis. Distinct from those in the dorsal telencephalon, radial glial progenitors (RGPs) in the ventral telencephalon responsible for producing neocortical interneurons progressively grow radial glial fibers anchored to periventricular vessels. This progenitor-vessel association is robust and actively maintained as RGPs undergo interkinetic nuclear migration and divide at the ventricular zone surface. Disruption of this association by selective removal of INTEGRIN β1 in RGPs leads to a decrease in progenitor division, a loss of PARVALBUMIN and SOMATOSTATIN-expressing interneurons, and defective synaptic inhibition in the neocortex. These results highlight a prominent interaction between RGPs and periventricular vessels important for proper production and function of neocortical interneurons., (Copyright © 2016 Elsevier Inc. All rights reserved.)

To maintain homeostasis, hypothalamic neurons in the arcuate nucleus must dynamically sense and integrate a multitude of peripheral signals. Blood-borne molecules must therefore be able to circumvent the tightly sealed vasculature of the blood-brain barrier to rapidly access their target neurons. However, how information encoded by circulating appetite-modifying hormones is conveyed to central hypothalamic neurons remains largely unexplored. Using in vivo multiphoton microscopy together with fluorescently labeled ligands, we demonstrate that circulating ghrelin, a versatile regulator of energy expenditure and feeding behavior, rapidly binds neurons in the vicinity of fenestrated capillaries, and that the number of labeled cell bodies varies with feeding status. Thus, by virtue of its vascular connections, the hypothalamus is able to directly sense peripheral signals, modifying energy status accordingly.

Although the water channel protein aquaporin-1 (AQP1) is widely observed outside the rat brain in continuous, but not fenestrated, vascular endothelia, it has not previously been observed in any endothelia within the normal rat brain and only to a limited extent in the human brain. In this immunohistochemical study of rat brain, AQP1 has also been found in microvessel endothelia, probably of the fenestrated type, in all circumventricular organs (except the subcommissural organ and the vascular organ of the lamina terminalis): in the median eminence, pineal, subfornical organ, area postrema and choroid plexus. The majority of microvessels in the median eminence, pineal and choroid plexus, known to be exclusively fenestrated, are shown to be AQP1-immunoreactive. In the subfornical organ and area postrema in which many, but not all, microvessels are fenestrated, not all microvessels are AQP1-immunoreactive. In the AQP1-immunoreactive microvessels, the AQP1 probably facilitates water movement between blood and interstitium as one component of the normal fluxes that occur in these specialised sensory and secretory areas. AQP1-immunoreactive endothelia have also been seen in a small population of blood vessels in the cerebral parenchyma outside the circumventricular organs, similar to other observations in human brain. The proposed development of AQP1 modulators to treat various brain pathologies in which AQP1 plays a deleterious role will necessitate further work to determine the effect of such modulators on the normal function of the circumventricular organs.

Blood-borne hormones acting in the mediobasal hypothalamus, like those controlling food intake, require relatively direct access to target chemosensory neurons of the arcuate nucleus (ARC). An anatomical substrate for this is a permeable microvasculature with fenestrated endothelial cells in the ARC, a system that has awaited comprehensive documentation. Here, the immunofluorescent detection of endothelial fenestral diaphragms in the rat ARC allowed us to quantitate permeable microvessels throughout its rostrocaudal extent. We have determined that permeable microvessels are part of the subependymal plexus irrigating exclusively the ventromedial (vm) ARC from the subadjacent neuroendocrine median eminence. Unexpectedly, permeable microvessels were concentrated proximal to the pituitary stalk. This marked topography strongly supports the functional importance of retrograde blood flow from the pituitary to the vmARC, therefore making a functional relationship between peripheral long-loop, pituitary short-loop, and neuroendocrine ultra-short loop feedback, altogether converging for integration in the vmARC (formerly known as the hypophysiotrophic area), thereby so pivotal as a multicompetent brain endocrinostat.

The blood-brain barrier (BBB) plays an important role in controlling the access of substances to the brain. Of the circumventricular organs (CVO), i.e. areas that lack a BBB, the median eminence and its close relationship with the hypothalamic arcuate nucleus plays an important role in controlling the entry of blood-borne substances to neurons of the mediobasal hypothalamus. In order to clarify the nature of the BBB in the median eminence-arcuate nucleus complex, we have used immunohistochemistry and antisera to protein components of the BBB-(1) tight junctions, claudin-5 and zona occludens-1 (ZO-1); (2) endothelial cells: (a) all endothelial cells: rat endothelial cell antigen-1 (RECA-1), (b) endothelial cells at BBB: endothelial barrier antigen (EBA), glucose transporter 1 (GLUT1) and transferrin receptor (TfR), and (c) endothelial cells at CVOs: dysferlin; (3) basal lamina: laminin; (4) vascular smooth muscle cells: smooth muscle actin (SMA); (5) pericytes: chondroitin sulfate proteoglycan (NG2); (6) glial cells: (a) astrocytes: glial fibrillary acidic protein (GFAP), (b) tanycytes: dopamine- and cAMP-regulated phosphoprotein of 32kDA (DARPP-32), (c) microglia: CD11b. Neuronal cell bodies located in the ventromedial aspect of the arcuate nucleus were visualized by antiserum to agouti-related protein (AgRP). The study provides a detailed analysis on the cellular localization of BBB components in the mediobasal hypothalamus. Some vessels in the ventromedial aspect of the arcuate nucleus lacked the BBB markers EBA and TfR, suggesting an absence of an intact BBB. These vessels may represent a route of entry for circulating substances to a subpopulation of arcuate nucleus neurons.

Activation of the G-protein-coupled receptor GPR54 by kisspeptins during normal puberty promotes the central release of gonadotropin-releasing hormone (GnRH) that, in turn, leads to reproductive maturation. In humans and mice, a loss of function mutations of GPR54 prevents the onset of puberty and leads to hypogonadotropic hypogonadism and infertility. Using electrophysiological, morphological, molecular, and retrograde-labeling techniques in brain slices prepared from vGluT2-GFP and GnRH-GFP mice, we demonstrate the existence of two physiologically distinct subpopulations of GnRH neurons. The first subpopulation is comprised of septal GnRH neurons that colocalize vesicular glutamate transporter 2 and green fluorescent protein and is insensitive to metabotropic glutamate receptor agonists, but is exquisitely sensitive to kisspeptin which closes potassium channels to dramatically initiate a long-lasting activation in neurons from prepubertal and postpubertal mice of both sexes. A second subpopulation is insensitive to kisspeptin but is uniquely activated by group I metabotropic glutamate receptor agonists. These two physiologically distinct classes of GnRH cells may subserve different functions in the central control of reproduction and fertility.

Recent studies have localized the glutamatergic cell marker type-2 vesicular glutamate transporter (VGLUT2) to distinct peptidergic neurosecretory systems that regulate hypophysial functions in rats. The present studies were aimed to map the neuronal sources of VGLUT2 in the median eminence and the posterior pituitary, the main terminal fields of hypothalamic neurosecretory neurons. Neurons innervating these regions were identified by the uptake of the retrograde tract-tracer Fluoro-Gold (FG) from the systemic circulation, whereas glutamatergic perikarya of the hypothalamus were visualized via the radioisotopic in situ hybridization detection of VGLUT2 mRNA. The results of dual-labeling studies established that the majority of neurons accumulating FG and also expressing VGLUT2 mRNA were located within the paraventricular, periventricular and supraoptic nuclei and around the organum vasculosum of the lamina terminalis and the preoptic area. In contrast, only few FG-accumulating cells exhibited VGLUT2 mRNA signal in the arcuate nucleus. Dual-label immunofluorescent studies of the median eminence and posterior pituitary to determine the subcellular location of VGLUT2, revealed the association of VGLUT2 immunoreactivity with SV2 protein, a marker for small clear vesicles in neurosecretory endings. Electron microscopic studies using pre-embedding colloidal gold labeling confirmed the localization of VGLUT2 in small clear synaptic vesicles. These data suggest that neurosecretory neurons located mainly within the paraventricular, anterior periventricular and supraoptic nuclei and around the organum vasculosum of the lamina terminalis and the preoptic area secrete glutamate into the fenestrated vessels of the median eminence and posterior pituitary. The functional aspects of the putative neuropeptide/glutamate co-release from neuroendocrine terminals remain to be elucidated.

Following systemic injection, several different dyes and markers are found to accumulate rapidly in cells in the arcuate nucleus and median eminence, and the capillaries in this region appear specialised for exchange of molecules. The present study used hydroxystilbamidine (FluoroGold equivalent) to identify cells that take up molecules from the circulation in these regions; 2-6 h following injection, uptake was seen in the external and intermediate zones of the median eminence and the adjacent ventral part of the arcuate nucleus, but not in other regions of the hypothalamus. The labelled cells were small; double-labelling experiments revealed that they expressed glial fibrillary acid protein (GFAP), but not NeuN, Agouti-related protein (AgRP) or beta-endorphin. They had the morphology of astrocytes and were readily distinguished from tanycytes by staining for vimentin. Many of these labelled astrocytes also expressed leptin receptors and neuropeptide Y Y1 receptors. The surrounding neurons that expressed these receptors did not take up this dye. This demonstrates that astrocytes take up molecules from the circulation in the median eminence and adjacent arcuate nucleus, and may have a significant signalling role in regulation of food intake., ((c) 2005 Wiley-Liss, Inc.)

Glial and endothelial cells interact throughout the brain to define specific functional domains. Whether endothelial cells convey signals to glia in the mature brain is unknown but is amenable to examination in circumventricular organs. Here we report that purified endothelial cells of one of these organs, the median eminence of the hypothalamus, induce acute actin cytoskeleton remodeling in isolated ependymoglial cells and show that this plasticity is mediated by nitric oxide (NO), a diffusible factor. We found that both soluble guanylyl cyclase and cyclooxygenase products are involved in this endothelial-mediated control of ependymoglia cytoarchitecture. We also demonstrate by electron microscopy that activation of endogenous NO release in the median eminence induces rapid structural changes, allowing a direct access of neurosecretory axons containing gonadotropin-releasing hormone (GnRH) (the neuropeptide controlling reproductive function) to the portal vasculature. Local in vivo inhibition of NO synthesis disrupts reproductive cyclicity, a process that requires a pulsatile, coordinated delivery of GnRH into the hypothalamic-adenohypophyseal portal system. Our results identify a previously unknown function for endothelial cells in inducing neuroglial plasticity and raise the intriguing possibility that endothelial cells throughout the brain may use a similar signaling mechanism to regulate glial-neuronal interactions.

The median eminence (ME) is considered as the final common pathway connecting the nervous and endocrine systems. In this neurohemal structure, dynamic interactions among nerve terminals, tanycytes, and astrocytes determine through plastic processes the neurohormones access to the portal blood. Because brain-derived neurotrophic factor (BDNF) is involved in plastic changes, we investigated its presence and that of its receptors (TrkB) in the different cellular types described in the ME. Using in situ hybridization and immunohistochemical techniques, we demonstrated that BDNF immunoreactivity was essentially located in the astrocytes and to a lesser extent in tanycytes. By contrast, BDNF was not detected in nerve terminals reaching the external layer of the ME. TrkB antibodies recognizing the extracellular receptor domain labeled all of these different cell types, suggesting an autocrine or paracrine action of BDNF at this level. More selective antibodies showed that TrkB.FL immunostaining was found in tanycytes and nerve endings, whereas TrkB.T1 immunostaining was localized in all cellular types. Immobilization stress increased BDNF mRNA and BDNF immunoreactivity patterns and induced biphasic BDNF release from the ME, as analyzed by push-pull perfusion. In addition, we observed that 60-min stress intensified BDNF immunoreactivity in the internal layer and also its colocalization with glial fibrillary acidic protein. Stress also accentuated BDNF immunostaining in the perivascular space in elements that were not labeled with antibodies recognizing fibroblast or endothelial cells. These data disclosed a novel location of BDNF and its receptors in the ME, which are presumably involved in dynamic processes such as hormone release.

Recent studies from our laboratory suggested that the vascular endothelium of the median eminence was involved via nitric oxide secretion in the modulation of GnRH release during the estrous cycle. To further investigate that issue, we studied the variations of endothelial nitric oxide synthase protein and mRNA in the median eminence of female rats killed at different time points of the day and/or of the estrous cycle. Endothelial nitric oxide synthase protein levels were measured by Western blot, and endothelial nitric oxide synthase mRNA analysis was performed with semiquantitative RT-PCR (for each time point, n = 4). The results revealed that endothelial nitric oxide synthase synthesis varied markedly across the estrous cycle. Indeed, endothelial nitric oxide synthase protein (n = 20) and mRNA (n = 16) levels increase significantly on 0800 h and 1600 h proestrus compared with 1400 h diestrus II. In a second step, quantification analysis were made in median eminence obtained from ovariectomized and ovariectomized, E2 benzoate primed rat. The results show a significant increase in expression of endothelial nitric oxide synthase protein as well as endothelial nitric oxide synthase mRNA in ovx-E2 primed rat median eminence. Concurrently, the levels of the cav-1 protein, a specific endogenous inhibitor of endothelial nitric oxide synthase, were measured in median eminence during estrous cycle and in ME from ovx and ovx-E2 primed rats. A significant decrease of median eminence cav-1 was noted on 1600 h proestrus and in ovx-E2 primed rats when compared with 1400 h diestrus II and ovx, respectively. Altogether, these results strongly suggest that high NO release from median eminence observed on proestrus may be due to an increase of endothelial nitric oxide synthase expression and a decrease of the cav-1 protein levels. These findings demonstrate that E2 is able to modulate endothelial nitric oxide synthase and cav-1 expression both during the estrous cycle and in experimental conditions and consequently reinforce the idea that nitric oxide acting on GnRH release, is essentially endothelial in origin. These results may also imply that variations of endothelial nitric oxide synthase expression are essential for the pulsatile/cyclic nitric oxide median eminence release observed in a previous study.

The cell bodies of hypothalamic secretory neurons are localized in areas protected by the blood-brain barrier (BBB), whereas their axon terminals are localized in the median eminence, which lacks a BBB. This implies a complex barrier system, allowing neurons of the central nervous system to secrete into the blood stream without making the BBB leaky. In the present study, three experimental protocols were applied to clarify certain relevant aspects of the barriers operating in the medial basal hypothalamus of the rat. We established that the milieu of the arcuate nucleus is exposed to both the ventricular and the subarachnoidal cerebrospinal fluid (CSF). The median eminence milieu, the perivascular space of the portal vessels, and the subarachnoid space appear to be in open communication; also, beta2-tanycytes establish an efficient barrier between the median eminence milieu and the ventricular CSF. Similarly, beta1-tanycytes establish a lateral barrier, separating the intercellular space of the median eminence from that of the arcuate nucleus. We also found that the glucose transporter I (GLUT I), a BBB marker, is localized throughout the whole plasma membrane of beta1-tanycytes, but is missing from beta2-tanycytes. Expression of GLUT I by tanycytes progressively develops during the first postnatal weeks; while the degree of damage of the arcuate nucleus by administration of monosodium glutamate, at different postnatal intervals, parallels that of the GLUT I immunoreactivity of beta1-tanycytes. An explanation is offered for the selective destruction of the arcuate neurons by the parenteral administration of monosodium glutamate to infant rats.

The median eminence (ME) is a small brain area forming both the structural and functional bridge between the hypothalamus and the hypophysis. It is supplied by a variety of neurohormones and neuropeptides which are delivered to the ME by different hypothalamic and extrahypothalamic pathways. These biologically active substances may act in the ME locally influencing the activity of secretion of the neighbouring terminals or, after being released from the neuronal endings into the network of fenestrated capillaries and transported to the hypophysis, they may be involved in the regulation of secretion of adenohypophyseal hormones. Recent demonstrations of extensive colocalizations of these biologically active substances in individual axonal endings in the ME with wide spectrum of biological actions further emphasizes the ME as an important place involved in the neuroendocrine regulatory processes.

Luteinizing hormone-releasing hormone (LHRH) release is required for ovulation in mammals. Although evidence for the direct action of gonadal steroids on LHRH neurons has been minimal, their importance in inducing the preovulatory surge of LHRH is unequivocal. We have identified a subgroup of LHRH neurons with progestin receptors in guinea pigs. Given their central position, these neurons may constitute foci of initial activity, which are amplified throughout the population of LHRH neurons, resulting in increased LHRH neurosecretion on the afternoon of proestrus. Additionally, gonadal steroids may regulate LHRH secretion at the level of the terminals. Using immunoelectron microscopy and image analysis, we have illustrated the dramatic influence of gonadal steroids on individual LHRH terminals in the median eminence of rats. Indirectly, gonadal steroids may modulate LHRH release by modulating glial elements. Using double-label fluorescence confocal microscopy, we illustrate that LHRH terminals in the median eminence are encased by end-feet of tanycytes. Acting on glial elements, gonadal steroids may regulate access of LHRH terminals to the basal lamina and influence the amount of the neuropeptide reaching the portal vessels. We propose that during the preovulatory surge, LHRH release is coordinated by synergistic mechanisms operating at the level of particular subgroups of neuronal perikarya and/or discrete regions of the median eminence. These synergistic actions may ensure that LHRH is released in a precipitous fashion, to induce the surge of LH from the pituitary, required for ovulation.

Microvascular corrosion casts of the pituitary gland of one to nine-day-old chickens (Gallus domesticus, white leghorn hybrids) were analysed with the scanning electron microscope. Results show that the chicken median eminence is supplied by branches of the infundibular and the neural-lobe arteries. They form a flat two-dimensional primary capillary plexus, which lacks any capillary loops and is continuous with the plexus in the neural lobe. The capillaries occupy about 60% of the total area of the median eminence. A subependymal plexus is present, showing no contact with the plexus of the median eminence. The chicken neural lobe consists of many hollow buds. These buds are supplied by branches of the neural-lobe artery, which gives rise to a two-dimensional capillary meshwork similar to that of the median eminence. An anterior group of portal vessels, comprising 14-16 vessels with a mean diameter of 37 microns, and a posterior group of portal vessels, comprising 2-4 shorter and slightly thinner vessels, arise from the median eminence are the sole blood supply for the adenohypophysis (distal lobe). Anterior portal vessels supply the cephalic lobe and the most rostral area of the caudal lobe, and posterior portal vessels supply the caudal lobe of the distal lobe. There are no short portal vessels connecting the neural lobe with the distal lobe. The (sinusoidal) capillary bed of the distal lobe is laminated. The chicken hypophysis drains into the cavernous sinus, which empties into the cerebral carotid veins. Within the period studied (days 1-9 after hatching) no age-related changes were found.

The capillary endothelial cells of the median eminence represent a potential site for the degradation/modification of both circulating and hypothalamic peptides passing through the hypophysial portal system toward the pituitary. This study examines endothelial cell peptidase expression in vitro by monitoring the metabolism of gonadotropin-releasing hormone (GnRH) by cultured endothelial cells from sheep median eminence. Cleavage of GnRH by median eminence endothelial cell membranes generated GnRH1-5 as the primary stable product, which was then degraded to GnRH1-3 and free amino acids. Degradation of GnRH was completely inhibited by TPCK, ZnCl2 and N-ethylmaleimide, and partially inhibited by EDTA and by a specific inhibitor of the metalloendopeptidase EC 3.4.24.15, CFP-AAY-pAB. Interestingly, an increase in GnRH1-9 production was seen with the latter inhibitors, suggesting a two-step mechanism of GnRH degradation involving a primary cleavage at the Pro9-Gly10-NH2 bond, inhibitable by TPCK, ZnCl2, and NEM, followed by cleavage by EC 3.4.24.15 to generate GnRH1-5. Phosphoramidon and angiotensin converting enzyme inhibitors (as well as other non-specific inhibitors) were without effect, indicating that endopeptidase EC 3.4.24.11 and angiotensin converting enzyme are not involved. Neither bovine aortic endothelial cell nor AtT-20 cell membranes exhibited this pattern of peptidase activity. Degradation of GnRH by intact median eminence endothelial cells in culture was also observed, suggesting an extracellular orientation for these enzymes; the potential role of such peptidases in the fine regulation of both pituitary function and local blood flow is currently under investigation.

The interaction between the brain microvascular endothelium and bovine serum albumin complexed with insulin and colloidal gold (insulin-BSA-gold) was studied in adult and newborn mice. The results suggest: (a) the modification of albumin enhances its binding to the luminal front of the endothelial cells, as compared to unmodified albumin used in previous studies from this laboratory; (b) the binding density of insulin-BSA-gold complex to blood-brain barrier microvessels is approximately 2.5 times higher in newborn than in adult mice; (c) in adult mice, fenestrated endothelia of the median eminence and choroid plexus demonstrate the highest binding capacity (over five and two times higher, respectively, than in blood-brain barrier endothelia); (d) in the median eminence only, the gold-labelled tracer particles may be transported across the vessel wall. Our observations offer new ultrastructural evidence that: (1) the modification of BSA molecules by complexing with insulin does not enhance the transport of BSA across the blood-brain barrier in mouse brain, and (2) insulin-BSA-gold complex appears adequate for ultrastructural localization of blood-brain barrier insulin receptors but is of questionable value as a tracer for demonstration of increased transendothelial transport in blood-brain barrier microvasculature.

Image alignment is an absolute requirement for three-dimensional (3-D) reconstruction from serial sections, and Fourier correlation is the most powerful way to compute alignments. The rotational and translational components of misalignment can be corrected by an iterative correlation procedure, but for images having significant differences, alignment can fail with a likelihood proportional to the extent of the differences. We found that translational correction was determined much more reliably when low-pass filters were applied to the product transforms from which the correlations were calculated. Rotational corrections based on polar analyses of the auto-correlations of the images instead of on the images directly contributed to more accurate alignments. These methods were used to generate 3-D reconstructions of brain capillary modules from serial-section mosaics of digitized transmission electron micrographs.

The differentiated cytology, cytochemistry, and functions within subdivisions of the tuber cinereum prompted this morphometric and physiological investigation of capillaries in the medium eminence and arcuate nucleus of albino rats. Morphometric studies established that the external zone of the median eminence had 3-5 times the number and surface area of true and sinusoidal capillaries than the internal or subependymal median eminence zones, or either of two subdivisions examined in the arcuate nucleus. Type-I true capillaries, around which Virchow-Robin spaces comprise 1% of arcuate tissue area, were situated proximally to the median eminence border. This finding is consistent with a premise that confluent pericapillary spaces enable infiltration of arcuate neurons by factors from capillary blood from the median eminence or Virchow-Robin spaces. Physiologically, the rate of penetration across the median eminence capillaries by blood-borne [14C]alpha-amino-isobutyric acid (a neutral amino acid used as a capillary permeability tracer) was 142 times greater than for capillaries in the distal arcuate nucleus within 12 s of tracer administration. A new finding was that the proximal arcuate nucleus had a permeability x surface area product of 69 microliters g-1 min-1, 34 times greater than that in more distal aspects of the tuber where blood-brain barrier properties exist. We also found that the microcirculatory transit time of a plasma space marker, [14C]sucrose, was considerably longer (1.2 s) in the median eminence and proximal arcuate nucleus than in the distal arcuate or ventromedial nucleus (0.4 s). By virtue of its high capillary permeability and extensive blood-tissue surface area, including the wide Virchow-Robin spaces, the median eminence external zone could be a gateway for flooding other tuberal compartments with blood-borne factors. This effect may be compounded by capillary bed specializations in the proximal arcuate nucleus where Type-I true capillaries, Type-III sinusoids, and pericapillary spaces are confluent with those in the median eminence. The results indicate that the proximal arcuate parenchyma could be exposed to circulating neuroactive substances on a moment-to-moment basis.

Three-dimensional reconstructions from serial section images require the accurate registration of those images. Image correlation is the most powerful computed alignment method and its performance on identical images, or parts thereof, has been thoroughly studied. Correlation alignments of complex, dissimilar images can fail, however, with a likelihood proportional to the magnitude of the differences. We report that alignments can be computed more reliably and more accurately (higher-valued correlation coefficients) by the combined use of lowpass-filtered product transforms (from which the correlation functions are formed), autocorrelation correction of rotational misalignment, and covariance correction of translation misalignment. A simple rule is proposed for the lowpass filter cutoff radius depending on measures of the images' differences. These methods are demonstrated with a reconstruction of a capillary loop in the median eminence of the hypothalamus.

Most circumventricular organs (CVOs) have unusually dense and permeable capillary networks that facilitate secretion of or tissue penetration by circulating substances, unlike other nervous system structures wherein blood-brain barrier (BBB) properties of the capillary endothelium limit solute permeability. In this brief review, I shall discuss new facts from recent experiments, and draw on interpretations from previous studies, to illustrate how capillary systems vary both between and within some CVOs, how closely microvascular properties coincide topographically with the distribution and density of neuropil, transmitter substances and hormonal binding sites, and how physiological data can be combined with morphological descriptions of capillary beds to accent specialized processes of blood-brain solute exchange in individual CVOs. The emphasis of this paper is on exchange microvessels of the rat area postrema (AP), subfornical organ (SFO) and median eminence (ME) which are regions of dense binding for several hormones and contain appreciable numbers of neurons (AP and SFO) or neural terminations that may be part of the sensing apparatus for humoral messengers of homeostatic systems. The work is intended to highlight established concepts about the process of blood monitoring by CVOs, summarize new morphological and physiological characteristics of their capillaries, and provide clues to novel research that could foster further understanding of these curious sentinel and secretory organs of the brain.

Electron microscopy of the median eminence (ME) of the Mongolian gerbil (Meriones unguiculatus) revealed that, unlike most other mammalian species, abundant neurohaemal contacts were present not only in the external zone (EZ), but also in the internal zone (IZ) up to the subependymal layer. In the IZ, nerve terminals with dense core vesicles and/or small clear vesicles abutted on the outer basal lamina of the perivascular space of portal capillaries, alternating with tanycyte processes. In addition to these neurohaemal contacts, several layers of vesicle-filled varicosities surrounded the portal vasculature. An analysis of serial thin sections showed that the latter varicosities could also reach the perivascular basal lamina or contact it through small extensions in other planes of section. Apparently at least some of the nerve terminals making neurohaemal contacts were en passant in nature. A correlative investigation of synaptophysin (a major integral membrane protein of small synaptic vesicles) immunoreactivity at the light microscopical level demonstrated a conspicuously dense immunostaining around portal capillaries in both EZ and IZ of the proximal and distal ME (neural stalk). Since this perivascular accumulation of immunoreactivity coincides precisely with the ultrastructural accumulation of vesicle-filled axons which establish numerous neurohaemal contacts, it is concluded that this pattern of synaptophysin immunostaining indicates sites of neurohaemal contacts at the light microscopical level. During postnatal development, the perivascular concentration of synaptophysin immunoreactivity in the IZ appeared concomitantly with the early postnatal expansion of long portal capillary loops into the IZ.(ABSTRACT TRUNCATED AT 250 WORDS)

The aim of these studies is to determine the sources of vascularization, the vascular areas, the appearance and density of the median eminence capillary network in prenatal human brains of both sexes. The brain blood vessels under examination were filled with a mixture of India ink and gelatin. Serial paraffin sections 20-40 microns thick, cleared after Spaltehol. Two arterial stems take part in median eminence vascularization together with their branches: a. hypophisialis superior and a. hypophisialis inferior. In order to quantify the density of the capillary network, the authors used the standard stereologic parameters--volume density and mean radius of blood vessels. No statistically significant sex-dependent differences were established regarding the appearance and density of the median eminence capillary network of the human brain in the course of the comparative testing of the mean values of these parameters in the male and female sex.

Blood flow was measured in the ovine median eminence and neural lobe before and after the intravenous infusion of dopamine (n = 7), the D1 agonist SKF 38393 (n = 4), the D2 agonist bromocriptine (n = 4), and the dopamine antagonist haloperidol (n = 5). It was also measured before and after the intracarotid infusion of dopamine into eight naive sheep and seven sheep pretreated with phenoxybenzamine. Radiolabeled microspheres were used to determine regional cerebral and regional neurohypophysial blood flows (RNHBF) in these 35 adult female sheep anesthetized with pentobarbital sodium. Samples for serum prolactin measurement by radioimmunoassay were obtained before and after drug infusion. Intravenous dopamine infusion did not change median eminence or neural lobe blood flow (RNHBF) but increased renal and choroid plexus blood flow. Intravenous haloperidol caused a significant fall in RNHBF and blood flow in choroid plexus, caudate nucleus, and kidneys. Intracarotid dopamine infusion decreased RNHBF but increased choroid plexus blood flow. RNHBF was significantly greater in the seven sheep pretreated with phenoxybenzamine than in the eight naive sheep. These findings do not support a role for dopamine in the regulation of median eminence blood flow. The last observation does add support to the hypothesis that norepinephrine or epinephrine interaction with alpha-receptors plays a role in the control of ovine median eminence blood flow and hence in the regulation of delivery of humoral messages from the brain to the anterior pituitary gland.

The development of the pituitary gland and its blood vessels is described in rat embryos (gestational day 12 through day 21) injected with India-ink via the umbilical vessels. The vascularization of all components of the pituitary gland develops from the surface network covering the prosencephalic vesicle. However, vascular connections exist between the prospective median eminence and the anterior pituitary gland in the earliest stages examined (day 12) but are not augmented by vessels from the stomatodeal roof until day 13. Primary portal veins appear initially on day 13, the vascularization of the pars distalis is visible on day 15. The "Mantelplexus" covering the floor of the diencephalon is discernible on day 16. Large-caliber portal veins appear immediately before birth, but otherwise there is no significant change in the vascular pattern during the last five days of gestation. The pars intermedia and the median eminence-pituitary stalk region remain avascular throughout embryonic life.

Microdissection techniques for isolated removal of the various regions of the hypothalamus as well as the individual hypothalamic nuclei are detailed. Recent development of biochemical microassays have made it possible that the concentrations of neurohormones, neuropeptides, neurotransmitters and their related enzymes could be detected in such a small volume of brain tissue than the hypothalamic nuclei. Data available of the hypothalamic distribution of above substances are summarized. The possible role and origin of intra- and extrahypothalamic neurohormones as well as the existence of the so-called "hypophysiotrophic area" are discussed.

The median eminence (ME) of hypophysectomized rats was studied by means of light and electron microscopy. Paraldehyde-fuchsin (PAF)-positive material is seen in the external zone (EZ) of the ME 2--5 days after the operation. Its amount gradually increases especially in the caudal part of the ME during the following few days. Some PAF-positive fibers make contact with the subependymally located blood capillaries. In the most caudal region of the recessus infundibuli they penetrate into the third ventricle. PAF-positive material decreases markedly from the ME of rats two months after hypophysectomy and exposure to a 1% salt load. Fibers of types A1, A2 and B containing granules of 120--220 nm, 100--150 nm and 80--100 nm in diameter, respectively, are seen in the EZ of the ME in hypophysectomized rats, although almost exclusively A2- and B-type structures make contact with the primary portal capillaries in intact animals. All types of neurosecretory fibers establish contact with the subependymal nonfenestrated blood capillaries and penetrate the recessus infundibuli. Some neurosecretory terminals of different types make direct contact with the glandular cells of the pars tuberalis or are separated from them by a thin basal lamina. It is assumed that mainly neurosecretory fibers of types A2 and B are permanently connected with the primary portal capillaries in the EZ of the ME in intact mammals, while the overwhelming majority of fibers of A1-type shows ingrowth during the course of postoperative reparation. The possible physiological significance of the described changes is discussed.

Human fetuses, 11.5 to 16.8 weeks' gestational age, were perfused with silicone rubber compound to study the hypothalamo-hypophyseal portal system. Impregnated vessels were observed throughout the median eminence, infundibular stalk, pars tuberalis and both lobes of the pituitary gland in all specimens. The data suggest that an intact portal system for the transport of hypothalamic releasing factors to the pituitary gland is established as early as 11.5 weeks' gestation.

Ventriculo-cisternal perfusion of horseradish peroxidase (HRP) in the mouse brain has demonstrated that a brain-blood barrier exists at the microvascular endothelium in brain parenchyma but not in the median eminence of the hypothalamus. The brain-blood barrier is similar to the blood-brain barrier in that: tight junctions prevent the movement of protein between endothelial cells, HRP taken into the endothelial cells is directed to lysosomal dense bodies, and, contrary to the literature, a vesicular transendothelial transport of HRP from brain to blood does not occur under normal conditions. The endocytosis of ventricular injected HRP from the abluminal side of the endothelium is demonstrably less than the endocytosis of intravenous injected HRP from the luminal side; hence, the cerebral endothelium expresses a degree of polarity regarding the internalization of its cell surface membrane and extracellular protein. The passage of cerebrospinal fluid-borne or blood-borne HRP between some ependymal cells of the median eminence is not precluded by tight junctions. These patent extracellular channels offer a direct pathway for the exchange of substances between cerebrospinal fluid in the third ventricle and fenestrated capillaries in the median eminence.

The formation of interrelations of the axons of neurosecretory cells and of ependyme cells with the capillaries of primary portal plexus in rats from the 14th day of embryogenesis till the 9th day of postnatal life was studied using the light and electron microscope methods. During the whole period under study, the basal processes of the ependyme cells reach the primary portal plexus of the capillaries. The terminals of the basal processes are usually separated from the endothelium of the capillaries by two basal membranes and enclosed pericapillary space. After the birth, some basal process penetrate in the pericapillary space and terminate on the endothelium. The surface of contact of the ependyme cell processes with the external basal membrane increases with the age, this being accompanied by the increase of pinocytotic activity. The neurosecretory axons are found in the median eminence already on the 14th day of embryogenesis, but by the 20th day only they reach the external basal membrane and penetrate sometimes in the pericapillary space. After the birth, the number of axons reaching the external basal membrane and the surface of contact between them increase gradually with, apparently, a concomitant intensification of the transport of neurohormones in the portal circulatory system of the hypophysial-hypothalamic complex.

The correlation of dopamine (DA)-, noradrenaline (NA)- or serotonin (5HT)-containing neurons and thyrotropin releasing hormone (TRH)-containing neurons in the median eminence of the rat, as well as the coexistence of monoamines (MA) and TRH in the neurons, were examined by subjecting ultrathin sections to a technique that combines MA autoradiography and TRH immunocytochemistry. The distribution and localization of silver grains after 3H-MA injection were examined by application of circle analysis on the autoradiographs. TRH-like immunoreactive nerve terminals containing the immunoreactive dense granular vesicles were found to have an intimate contact with monoaminergic terminals labeled after 3H-DA, 3H-NA or 3H-5HT infusion in the vicinity of the primary portal capillaries in the median eminence. Synapses between TRH-like immunoreactive axons and MA axons labeled with silver grains, however, have not been observed to date. Findings suggesting the coexistence of TRH and MA in the same nerve terminals or the uptake of 3H-MA into TRH-like immunoreactive nerve terminals, where silver grains after 3H-MA injection were concurrently localized in TRH-like immunoreactive nerve terminals, were rarely observed in the median eminence. Percentages of the nerve terminals containing both immunoreactive granular vesicles and silver grains after 3H-MA injection to total nerve terminals labeled after 3H-MA infusion silver grains were equally very low in 3H-DA, 3H-NA or 3H-5HT, amounting to less than 6.1%.

Using immunohistochemistry at the light and electron microscopic level some blood vessels along the median eminence were shown to be surrounded by dense networks of vasoactive intestinal polypeptide/peptide histidine isoleucine (VIP/PHI)-positive fibers. VIP and PHI released from these fibers may contribute to the elevated levels of these two peptides measured in portal blood as compared to peripheral blood by radioimmunoassay. VIP and PHI may also be important in the control of blood flow through the median eminence.

The vascularization of the hypothalamo-adenohypophysial region of the toad Bufo bufo (L.) (Amphibia, Anura) was studied by means of light- and scanning electron microscopy. Special attention was given to the portal vascular system of the median eminence and the pars distalis. Course and arrangement of these vessels are described.

The concept of neurosecretion as the mechanism by which neural control of adenohypophyseal function is accomplished was based on the observation that long capillary loops penetrate deeply into the supraopticohypophyseal tract as it passes through the median eminence internal zone. However, neural contact upon these capillary loops has not been demonstrated in the mammalian median eminence. The present transmission electron microscopic investigation of the rabbit median eminence demonstrates neurohemal contact in the median eminence internal zone. Axons containing small lucent vesicles 53.3 +/- 3.28 nm in diameter (mean +/- SEM) or small lucent and large granular vesicles with a mean diameter of 122.4 (+/- 3.28 nm) in their terminals make neurohemal contact with capillary loops in the internal zone and form a cuff about them. These terminals resemble terminals found in the external zone. Intravenous injection of the false neurotransmitter 5-hydroxydopamine (5-OH-DA) renders small lucent vesicles granular in both the external and internal zone. The effect of 5-OH-DA injection is abolished by concurrent reserpine administration. Whereas large granular vesicles in many terminals become lucent after reserpine administration, in others they remained electron dense. Viewed in the light of previous studies our findings suggest that the internal plexus arises from the external plexus and invaginates the neuropil carrying connective tissue and parvicellular axon terminals of aminergic and peptidergic systems from the external zone into the internal zone, that some elements making neurohemal contact with long capillary loops are terminals of the noradrenergic reticular infundibular tract arising outside the hypothalamus in the brainstem, and that long capillary loops form a system of repeating microvascular modules which markedly increase the surface available for neurohemal contact.

Vascular casts of the pituitary-median eminence complex from seventeen adult female rabbits were examined with the scanning electron microscope. The results of this study confirm the presence of a single capillary bed common to the entire neurohypophysis. Arterial supply to the rabbit pituitary is only to the neurohypophysis. A direct supply to adenohypophysis was not found. Within the median eminence there are an external and internal capillary plexus. The internal capillary plexus is directed toward the infundibular recess of the third ventricle. It does not receive a direct arterial supply but derives its blood supply from the external plexus before draining to the adenohypophysis. Vessels of the posterior median eminence are confluent with vessels of the infundibular stem. On the basis of these studies, it is proposed that the entire neurohypophysis, not simply the median eminence, serves as the final common pathway to the glandular pituitary. It is also proposed that in the median eminence, vessels are organized to deliver blood containing hypothalamic releasing and inhibiting hormones as well as posterior lobe neural hormones (antidiuretic hormone and oxytocin) to the ventricular surface for subsequent transport to cerebrospinal fluid and distribution to the brain.

The direction of pituitary blood flow, the amount of pituitary blood flow, its regional control, and the role of the median eminence microcirculation are the subjects of this review. Present concepts of pituitary blood flow are focused almost entirely on its direction and arouse from studies of pituitary vascular anatomy performed almost 50 years ago. The development of new anatomic techniques has led to a reappraisal of pituitary angioarchitecture, stimulated physiological studies to clarify the pattern of blood flow within the entire gland, and led to a reappraisal of accepted concepts of directional pituitary blood flow. The availability of techniques to accurately measure organ blood flow has permitted study of pituitary blood flow; and, when combined with knowledge of pituitary anatomy, the application of these techniques promises to provide a means to develop insight into control of the mechanisms by which chemical messengers are delivered to the pituitary to control its function. New anatomic techniques promise to develop new understanding of the three-dimensional arrangement of median eminence microvasculature and yield new concepts of blood flow regulation within the median eminence that can be tested by physiological means.

The objectives were to (a) determine the age in development when GnRH is first detectable in the brain and (b) observe the distribution of GnRH throughout the fetal and early postnatal period. GnRH was localized immunohistochemically in fetal (15, 16, 17 and 19 days of gestation) and early postnatal (1- and 7-day-old) mice with the peroxidase-antiperoxidase (PAP) method of Sternberger. In the organum vasculosum of the lamina terminalis (OVLT) and in the median eminence of the fetus, GnRH was first detected at 17 days of gestation. In the OVLT, GnRH was found ventral to the preoptic recess of the third ventricle near the ventral surface of the brain. In addition, GnRH was located adjacent to the superficial portal capillaries near the surface of the median eminence. At 19 days of gestation, the distribution of GnRH was similar to that observed at 17 days and there was a marked increase in amount. In the newborn mouse, GnRH was undetectable in the OVLT and its content in the median eminence was decreased as compared to that observed in the fetus. By the seventh postnatal day, a considerable accumulation of GnRH had occurred in the OVLT and median eminence. In the OVLT, it was associated with capillaries ventral to the preoptic recess, and its distribution in the median eminence was similar to that in the adult mouse. In both the OVLT and median eminence of the fetal and early postnatal mouse GnRH appeared to be stored in axons and axon endings, but was not detectable in nerve cell bodies or ependymal cells. These observations suggest that the potential for neuroendocrine control of gonadotropin secretion exists in the fetal mouse early as 17 days of gestation.

The hypophyseal portal vessels were studied in forty nine species of birds. The primary capillary plexus in the median eminence is single or divided into an anterior and a posterior plexus. Irrespective of whether the primary capillary plexus is single or divided, distinct, non-interconnected anterior and posterior groups of portal vessels are present in all the species investigated. The anterior group of portal vessels originates in the anterior region of the median eminence and breaks up into capillaries in the cephalic lobe of the pars distalis; the posterior group of portal vessels originates in the posterior region of the median eminence and breaks up into capillaries in the caudal lobe of the pars distalis. This type of regional distribution of portal vessels appears to be of general occurrence in the avian pituitary. The median eminence in the species investigated shows an AF-positive anterior region and an AF-negative posterior region. The pars distalis is differentiated into histologically distinct cephalic and caudal lobes. The arrangement of the portal vessels into anterior and posterior groups provides morphological basis for the view that the functions of the cephalic lobe may be controlled by the anterior median eminence, whereas those of the caudal lobe may be controlled by the posterior median eminence. However, experimental data available to date do not suggest a physiological significance to the widespread incidence of the regional distribution of portal vessels in the avian pituitary.

The development of neurocytes of arcuate (ABC) and suprachiasmatic (SC) and the median eminence (ME) together with portal capillaries during embryogenesis and post-natal periods in female Wistar rats were studied at light and electron microscope level. Neurocytes of ARC nucleus producing secretory granules was observed first at 20 day of embryogenesis. After birth the number of such neurocytes producing secretory granules was greatly increased. In SC nucleus the neurocytes begin producing secretory granules in 5...7 day old animals. The results of the karyometric study also reveal that the development of ARC nucleus neurocyte proceeds the SC nucleus. This asynchronous development of tonic (ARC nucleus) and cyclic centre (SC nucleus) probably determines the critical period of sexual programming in female rats. Development of ME including the differentiation of portal capillaries is structurally completed during 3-4 weeks after birth. The significance of these findings in relation to the role of ARC and SC nuclei in sexual cycle regulation in female rats has been pointed out and discussed.

The ultrastructure of the median eminence of neonatal (newborn) 1-, 3-, 5-, and 10-day old) and adult guinea pigs was studied to determine the dynamic changes occurring in this structure during early life. At birth the portal vasculature consists of the Mantelplexus and a few, non-fenestrated capillary loops. The number of ansae and the degree of fenestration increase rapidly after birth. The abundance of cytoplasmic and ciliary projections into the ventricular recess and the large numbers of organelles indicate that the ependymal cells are more active in the neonatal period than in the adult male. Moreover, the ependymal endfeet cover most of the surface area of the primary portal plexus during this time. The neuronal layers of the median eminence are difficult to distinguish at birth due to the lack of myelinated fibers in the zona interna. Significant myelination appears on day 3 but is not complete until day 10. There is a progressive increase in the numbers of Herring bodies and large neuro-secretory granules (1,500-1,700 A) during this same time period. In the zona externa, few nerve terminals abut on the perivascular space until day 3. Increases in numbers of granules per axon profile were noted for each day after birth. Despite the relatively long gestation period of the guinea pig (68-72 days), the morphologic appearance of the median eminence at birth suggests that the neurovascular link controlling anterior pituitary function is not yet complete.

Ependymal junctions in both the ventro-lateral wall and infundibular floor of the 3rd ventricle were examined in adult, young, neonatal, and fetal rats in freeze-fracture preparations. During late fetal and early postnatal life, tight junctions developed in both ventro-lateral and floor ependyma with, first, the appearance of an area of waffle-like or honeycomb texture in the membrane where tight junctions will appear; second, the decoration of P-face crests of this honeycomb with individual particles; and lastly, fusion of these particles into smooth strands. In adults, tight junctions were rare in the lateral ependyma, but persisted on the floor as a loose network. Many small ependymal gap junctions were already present at the earliest fetal stage (17 days) examined. They occupied significantly larger proportions of the membranes on the ventro-lateral wall than on the floor throughout the time course due to an increase in their size but not in their numbers. The smallest gap junctions were regarded as the newest in a process of renewal throughout life. The development of both junctional types parallels what is known on the onset of neuroendocrine functions in the median eminence. In the rat, this is apparently just before, at, or just after birth.

The size of the neuro-vascular contact surface was estimated along the median eminence (ME) proximal pituitary stalk (PS) of adult rats using serially cut 1 micrometer thick plastic sections. It was found 26% larger in female animals. The dorsal surface of the pars tuberalis was used as reference surface, and its size was estimated in the same way. The average rate of increase of the contact surface with respect to the reference surface was 3.05 and 2.98 in females and 2.34 and 2.35 in males. The reference surface was divided into small components and also the regional distribution of the rate of increase of the contact surface was analyzed. Very folded contact surface areas were found scattered in the rostral half of the ME. Folded areas made up the ME on both sides of the midline and the dorsal and ventral surfaces of the PS, whereas relatively smooth areas are localized along the 2 margins of the ME and the 2 sides of the PS. The results were correlated with the known termination sites of dopamine and various hypothalamic hormone containing pathways.

The organization of the system of perivascular space around the capillaries in the neurohypophysis was studied in the adult and developing laboratory mouse by the use of histological silver impregnation and electron microscopical techniques. In the median eminence short and long extensions, arising mainly from the shallow space around capillary loops of the primary plexus of the portal system, formed radiations into the adjacent neural tissue of the external zone. The tissue of the neural lobe was separable into non-vascular regions dominated by undilated portions of neurosecretory nerve fibres and pituicytes, and neurovascular regions with perivascular space extensions forming an extensive system of connections between neighbouring capillaries. In the median eminence, the system of extensions of the perivascular space was estimated to increase the neurovascular contact surface area by at least 50%, implying an increased efficiency of the organ without a notable increase of its volume. The possibility that the ramifications of the perivascular space imply an enhanced uptake rate into the bloodstream and a subsequent increased concentration of the neurohormones in the portal blood, was discussed. During development of the median eminence, differentiation of perivascular space extensions of the adult type started in the juvenile of about three weeks of age, when shallow capillary loops had been formed. In the neural lobe, perivascular space ramifications were already present when the internal capillaries were formed and were fairly frequent in ten-day young. At the age of three to four weeks the organization of the system was similar to that of the adult animal.