Immunocytochemistry of neuroendocrine systems: vasopressin, somatostatin, luliberin

Ultrastructural changes in the circumventricular organs after experimental subarachnoid hemorrhage

OBJECTIVES

Circumventricular organs (CVOs) are fine, periventricular, neurotransmitter-rich structures that are devoid of a blood-brain barrier and are known for their secretory role controlling fluid and electrolyte balance, thirst and even reproduction. Common pathologies of the brain such as trauma or bleeding affect CVOs, and hence their function. However, at what stage of these disease processes are CVOs affected and the time sequence of their recovery is still not clear. The aim of this study was to detect the morphological changes in CVOs using electron microscopy after experimental subarachnoid hemorrhage (SAH).

METHODS

Experimental SAH was induced by transclival puncture of the basilar artery. Both scanning and transmission electron microscopic examination of the representive sections from each CVO was undertaken.

RESULTS

Electron microscopy has shown that after SAH, the cells that form the CVOs exhibit signs of cellular necrosis with margination of the nucleus as well as cytoplasmic, mitochondrial and axonal edema. The subfornicial organ and organum vasculosum lamina terminalis appear to be more vulnerable to the effects of SAH than the median eminence or area postrema.

DISCUSSION

Considering the fact that the experimental SAH model we have used is very similar to the momentary rupture of an aneurysm with secondary reflex spasm to seal the hole, it will not be unrealistic to consider that similar effects may also take place in the clinical setting.

Purification of the main somatostatin-degrading proteases from rat and pig brains, their action on other neuropeptides, and their identification as endopeptidases 24.15 and 24.16

The main somatostatin-degrading proteases were purified from rat and pig brain homogenates and characterized as thiol- and metal-dependent endoproteases. Two types of proteases with apparent native and subunit molecular masses of 70 kDa and 68 kDa could be differentiated in both species. Beside somatostatin, both hydrolyzed several other neuropeptides with chain lengths between 8 and 30 amino acid residues. Cleavage sites were generally similar or identical, but some clear exceptions were observed for enzymes from both species which could be used to differentiate between the two proteases. The 68-kDa protease cleaved somatostatin at three bonds (Asn5-Phe6, Phe6-Phe7 and Thr10-Phe11) and neurotensin only at the Arg8-Arg9 bond, whereas the 70-kDa protease digested somatostatin at only two bonds (Phe6-Phe7 and Thr10-Phe11) and neurotensin as well as acetylneurotensin-(8-13) additionally (pig protease) or almost exclusively (rat protease) at the Pro10-Tyr11 bond. Relative rates for the digestions of various peptides were, however, more dependent on the species than on the type of protease. Cleavage sites for angiotensin II, bradykinin, dynorphin, gonadoliberin and substance P were, apart from different rates, identical for both proteases. In both species the 68-kDa protease was found to be mainly, but not exclusively, soluble and not membrane-associated, whereas the inverse was detected for the 70-kDa protease. Based on distinct molecular and catalytic properties, the 68-kDa protease is supposed to be congruent with the endopeptidase 24.15 (EC 3.4.24.15), the 70-kDa protease with endopeptidase 24.16 (EC 3.4.24.16, neurotensin-degrading endopeptidase). This investigation demonstrates that both proteases hydrolyze various neuropeptides with similar cleavage sites, but with species-dependent activity. Species-independent distinctions are the exclusive action of endopeptidase 24.16 on acetylneurotensin-(8-13) and liberation of free Phe from somatostatin only by endopeptidase 24.15.

Somatostatin binding sites on rat diencephalic astrocytes. Light-microscopic study in vitro and in vivo

Using a somatostatin-gold conjugate of known biological activity, high affinity binding sites for this neuropeptide were visualized at cellular resolution on cultured diencephalic astrocytes and on frozen sections of the rat diencephalon. Binding could be completely suppressed in competition experiments with surplus unlabeled somatostatin. On sections, the ligand was displaced from its binding sites by 10 microM guanosine triphosphate indicating a functional significance of the labeled structures. As with the native peptide, a surplus of the analog SMS 201-995 suppressed nearly all staining. The ligand was bound to distinct populations of astrocytes, namely to those in subependymal and perivascular positions, to astrocytes in somatostatin-innervated hypothalamic nuclei in the mid-sagittal plane and to borderline regions of circumventricular organs. A general mismatch between the distribution of somatostatin-immunoreactive terminals and the pattern of binding of the ligand does not exist. This, together with the competition experiments, suggests a functional relationship between the somatostatin-releasing neurons and associated astrocytes.

Somatostatin-binding sites on rat telencephalic astrocytes. Light- and electron-microscopic studies in vitro and in vivo

Using a somatostatin-gold conjugate as ligand, high-affinity binding sites for this neuropeptide were demonstrated at three levels: (i) cultured astrocytes from rat cortex, (ii) hippocampal slice cultures, and (iii) frozen tissue sections of rat telencephalon. The conjugate proved as active as the native peptide in competing for the binding sites. Light-microscopic visualization of bound ligand was achieved by silver intensification of the colloidal gold. This method is faster and yields superior resolution compared with autoradiography. Cultured astrocytes from cortex and hippocampus could be labeled by the ligand. At the light- and electron-microscopic level, astrocytes could be double-labeled by the somatostatin-gold conjugate and immunostaining for glial fibrillary acidic protein (GFAP). In hippocampal slice cultures, the conjugate did not penetrate into the neuropil because of a covering glial layer. However, a portion of this completely GFAP-positive covering glia reacted with the somatostatin ligand. In frozen brain sections, apart from delicate punctate structures, two types of labeled glia cells were seen: single stellate astrocytes and perivascular glia cells.

Peering through the windows of the brain

These seven specialized circumventricular structures of the mammalian brain represent windows with individualized structural characteristics permitting intimate contact between blood and cerebrospinal fluid, neurones and specialized ependyma-glia. These “Seven Windows of the Brain”, like the seven lucky deities of Japan, may each have a specific patron of body -brain function which they serve.1

Development of serotoninergic neurons from ventricular cells of the mouse neural plate in vitro

Cephalic neural plates and neural tubes of mice (pros- and rhombencephalic anlagen), developmental stages Theiler 11-18 [Th 11-18; embryonic day 7 1/2-11 (E7 1/2-11)], were prepared and cultured in a plasma clot with horse serum-containing MEM medium. Differentiation of the ventricular cells was studied in order to investigate the expression of serotoninergic properties. Serotoninergic neurons were not detected in preparations derived from neural plates of stage Th 11 (E7 1/2), but were demonstrated in increasing numbers from the early stage Th 12 (E8) onwards. The exclusively originated from the rhombencephalic floor caudal to the mesencephalic flexure. The serotoninergic neurons developed from these areas, irrespective of whether being cultured in their natural position within the neural plate, or separated as microcultures, or transplanted into the prosencephalic anlage. Every other region of the neural plate remained free of serotoninergic neurons. The in vitro findings are highly reproducible due to the following properties: the morphological and immunocytochemical peculiarities of the serotoninergic neurons, their tendency to appear in increasing numbers with age, their localization within the cultured neural plates and their appearance in all cultures from stage Th 12 (E8) on. Due to these findings it is considered possible that the progenitor cells of serotoninergic neurons might already have been determined within distinct areas in the mouse neural plates as early as stage Th 12 (E8).

Immunocytochemical studies of somatostatin neurons in brain

Immunohistochemical studies with antisera to somatostatin have, in many instances, led the way to our present understanding of the peptidergic nervous system. Somatostatin was among the first of the hypophysiotropic hormones shown to be contained in diverse neuronal circuits outside of the hypothalamus. For example, somatostatin is found within neurons ranging in location from the cerebral cortex to primary sensory neurons to enteric neurons within the gut wall. Somatostatin was also the first neuropeptide demonstrated to coexist within vertebrate neurons that also produce a classical neurotransmitter. Since this initial demonstration in sympathetic ganglionic neurons, somatostatin and numerous other neuropeptides have been demonstrated to coexist with a variety of classical neurotransmitters. The "rules" for coexistence are not clear, since somatostatin coexist in some instances with norepinephrine, in other cases with GABA, and probably with other classical transmitters as well. In some neurons, somatostatin also coexists with certain other neuropeptides. Finally, the specificity of immunohistochemical localizations of somatostatin has now been confirmed by virtue of the co-staining of somatostatin neurons with antisera to other portions of the biosynthetic precursor to somatostatin.

Immunohistochemical distribution of somatostatin-like immunoreactivity in the central nervous system of the adult rat

The localization and distribution of somatostatin (growth hormone release-inhibiting hormone; somatotropin release-inhibiting factor) have been studied with the indirect immunofluorescence technique of Coons and collaborators and the immunoperoxidase method of Sternberger and coworkers using specific and well-characterized antibodies to somatostatin, providing semiquantitative, detailed maps of somatostatin-immunoreactive cell profiles and fibers. Our results demonstrate a widespread occurrence of somatostatin-positive nerve cell bodies and fibers throughout the central nervous system of adult, normal or colchicine-treated, albino rats. The somatostatin cell bodies varied in size from below 10 micron up to 40 micron in diameter and could have only a few or multiple processes. Dense populations of cell somata were present in many major areas including neocortex, piriform cortex, hippocampus, amygdaloid complex, nucleus caudatus, nucleus accumbens, anterior periventricular hypothalamic area, ventromedial hypothalamic nucleus, nucleus arcuatus, medial to and within the lateral lemniscus, pontine reticular nuclei, nucleus cochlearis dorsalis and immediately dorsal to the nucleus tractus solitarii. Extensive networks of nerve fibers of varying densities were also found in most areas and nuclei of the central nervous system. Both varicose fibers as well as dot- or "dust-like" structures were seen. Areas with dense or very dense networks included nucleus accumbens, nucleus caudatus, nucleus amygdaloideus centralis, most parts of the hypothalamus, nucleus parabrachialis, nucleus tractus solitarii, nucleus ambiguus, nucleus tractus spinalis nervi trigemini and the dorsal horn of the spinal cord. One exception is the cerebellum which only contained few somatostatin-positive cell bodies and nerve fibers. It should be noted that somatostatin-positive cell bodies and fibers did not always conform to the boundaries of the classical neuroanatomical nuclei, but could often be found in areas between these well-established nuclei or occupying, in varying concentrations, only parts of such nuclei. It was difficult to identify with certainty somatostatin-immunoreactive axons in the animals studied. Some pathways could, however, be demonstrated, but further experimental studies are necessary to elucidate the exact projections of the somatostatin-immunoreactive neurons in the rat central nervous system.

Fine structural cytology of the rat subfornical organ during ontogenesis

The main developmental events in the subfornical organ take place between 17 fetal days (fd) and 5 days post natum (dpn) at which time it possesses most of its mature fine structural characteristics. The surface regional characteristics of ependymal cells differentiate primarily during this time as well, while the ependymal cellular fine structure, shape and relationship with neurons and the vascularity are well established prior to birth. Undifferentiated neurons contain glycogen prior to 19 fd and then differentiate by developing processes and organelles characteristic of neurons. By 5 dpn, the various types of neurons found in the mature subfornical organ are all present, except for giant vacuolated cells. Synapses containing only electron-lucent vesicles are first present at 20 fd, those containing additional electron-dense vesicles at 3 dpn. Microglial cells are first identifiable at 17 fd, and the first protoplasmic astrocytes are recognizable at 21 fd, while fibrous astrocytes are not detectable prior to 7 dpn. By 5 dpn, the cytological elements of the subfornical organ are all in place, and further developmental changes leading to adult fine structural characteristics by 30 dpn are essentially quantitative in nature.

Ontogenetic appearance of somatostatin-containing nerve terminals in the median eminence of rats

The development of immunoreactive (ir) somatostatin-containing nerve terminals in the rat median eminence (ME) has been examined electron-microscopically. Nerve fibers containing ir particles scattered throughout the axoplasm are first seen in the external layer of the ME on day 18.5 of gestation, and, on day 21.5 appear to terminate on the basement membrane of the perivascular space of the portal vessels. After birth, the fiber terminals contain several membrane-limited granules, which are labeled with ir PAP particles. Ultrathin, Epon-embedded sections of ME, treated by the protein A gold-labeling method for somatostatin, demonstrate positively labeled granules in the nerve fibers in the postnatal ME, but in the prenatal tissue, no specific gold-labeling is found. These findings show that, in the external layer of the ME, somatostatin storing occurs in the granules in the axonal terminals after birth.

Immunocytochemistry and ultrastructure of the neuropil located ventral to the rat supraoptic nucleus

The neuropil located ventral to the SON was investigated by the use of immunoperoxidase staining for neurophysins, oxytocin and vasopressin, and electron microscopy. The study was performed in six groups of rats: 1) control; 2) infusion of isotonic saline into the CSF; 3) infusion of hypertonic saline into the CSF; 4) drinking hypertonic saline for 4 days; 5) same as group 4 but injection of colchicine into the CSF on second day of dehydration; 6) salt loading for 3 months. In the control rats the ventral neuropil contained a few immunoreactive processes, the general morphology of which was completely different from that of the neurosecretory axons emerging from the SON at its dorsal aspect. In rats of groups 3 to 6 the ventral processes (VP) became loaded with neurosecretory granules, whereas the perikarya and axons were depleted. Based on their general morphology and reactivity pattern it is suggested that the VP are dendrites. Most of these "dendrites" were embedded in a glial cushion formed by the processes of a particular type of marginal glia. Some of these "dendrites" enveloped an arteriole penetrating the optic tract. All VP were rich in synaptic contacts. The possibility that the VP of neurosecretory cells may be functionally related to the subarachnoid CSF and the arteriolar blood flow is discussed.

Differing postnatal development of the somatostatin- and luliberin- systems in the male and female rat

By means of light-microscopic immunohistochemistry the perikarya of the luliberin-(LRF-) and somatostatin systems of neonate rats were found to be in differing stages of development. At a time point when the LRF-producing neurons had obviously attained their final shape and size, the somatostatin-immunoreactive perikarya were still in a postnatal phase of maturation. Whereas the number of the latter perikarya increases with advancing age, the number of LRF-immunoreactive perikarya decreases significantly from postnatal day 7 onward. Both peptide-hormone systems do not project concomitantly and to the same extent to their principal neurohemal regions in the organum vasculosum laminae terminalis (OVLT) and the median eminence (ME). In all presently studied stages of development, despite considerable individual variations in one age group, among the components of the LRF-system the OVLT displays a more intense immunoreactivity than the ME. The somatostatin system, however, projects to the OVLT with a conspicuous temporal delay compared to the ME, and, furthermore, in the OVLT the pattern of immunoreactivity characteristic of adult rats is not yet attained at postnatal day 21. Evidence for differences in the immunoreactivity between male and female animals was restricted to the LRF-system. Finally, the results obtained on the stria terminalis speak in favour of the fact that the long-range extrahypothalamic projections of the somatostatin system also undergo postnatal maturation. In the stria terminalis, somatostatin-immunoreactive fibers can be demonstrated initially on postnatal day 7. They attain their full immunoreactivity on postnatal day 21. Furthermore, in the bed nucleus of the stria terminalis an intermittent cytoplasmic immunoreactivity is observed, which is limited to the animals of postnatal day 7 and disappears completely during the further course of development.

Identification and immunocytochemical localization of a human adult brain-specific antigen (HABSA)

A new protein with a molecular weight of 669,000, identified in brain extracts from 4 to 69 years old subjects has been isolated and immunochemically characterized. The antigen is found in human adult brain but not in the brains of human fetuses and newborn infants or in the brains of several other species tested. Immunocytochemically, using the PAP method, the antigen is localized at the surface of some nerve cells and on astrocytes and oligodendrocytes of the cerebral cortex, corpus striatum, pons and medulla. The Golgi epithelial cells with Bergmann's fibers, and the velate and ordinary astrocytes in the cerebellum show immunoreactivity as well.

Generalized lipidosis in newborn rats and Guinea pigs induced during prenatal development by administration of amphiphilic drugs to pregnant animals

Pregnant rats and guinea pigs were treated throughout the second half of gestation with amphiphilic drugs (chlorphentermine, chlorcyclizine, chloroquine) known to induce generalized lipidosis. The offspring were sacrificed immediately after birth, and several tissues (lung, liver, kidney, spleen, pituitary gland, adrenal gland, spinal cord, hypothalamus) were examined by electron microscopy. Generalized lipidosis was found in the offspring of both species, albeit of lesser degree than in the mothers. The results show that fetal and adult tissues respond to lipidosis-inducing drugs in a qualitatively similar way; the quantitative differences found may be related to pharmacokinetic and cellular factors.

Estrogen-dependent changes in the functional interrelationships among neurons, ependymal cells and glial cells of the arcuate nucleus. Cytometric studies in the female albino mouse

The synchronizing effect of ethinylestradiol (4 micrograms/g b.w.) on neurons of the arcuate nucleus 700-950 micrometers caudal to the posterior edge of the optic chiasma was studied by karyometry in 6-week-old albino mice during proestrus. The caudal portion of the arcuate nucleus was identified as the most estrogen-sensitive subdivision; all neurons showed an increase in their nuclear area (mean transect, profile area of the nucleus) 1 h following administration of ethinylestradiol. This hypothalamic region was selected for the subsequent electron-microscopic cytometric study to analyze functional interrelationships among neurons, ependymal cells and glial cells. Six and 12 days after ovariectomy no significant change in the nuclear area of neurons and ependymal cells was found 850-950 micrometers behind the posterior slope of the optic chiasma, but the neurons exhibited a decrease in the number of polyribosomes, the volume fraction (Vvmi) and the surface density of the inner membrane of mitochondria (Svmi). A similar decrease in Vvmi and Svmi was measured in the apical part of ependymal cells and in the pericapillary profiles of ependymal and glial cells, which was accompanied by a reduction in the surface density of ependymal processes extending into the ventricular lumen. In addition, no change of Vvmi and Svmi was seen in the basal subnuclear part of ependymal cells. This bipolar functional reaction of ependymal cells after ovariectomy is discussed as an indicator of ependymal control of neuronal activity by sequestering biologically active agents, e.g., transmitters of neurohormones, in their apical and basal extensions facing the ventricular surface or the pericapillary space.

Electron microscopic immunocytochemical investigation on the postnatal development of the vasopressin system in the rat

The present ultrastructural results indicate that, in the rat, the vasopressin-synthesizing perikarya of the supraoptic nucleus (NSO) attain a certain degree of maturity earlier than those of the paraventricular nucleus (NPV). In the neonate rat, the stainability of the nuclear areas is very weak; in the perikarya of the NSO a few labeled granules can be found, whereas the perikarya of the NPV often display only a labeled Golgi area, the cytoplasm being devoid of granules. At the end of the first (NSO) and the second (NPV) postnatal weeks, the filling of the neurosecretory granules with vasopressin is inhomogeneous with irregular spots of reaction product distributed on the granules. This feature is less obvious during the following week and has nearly disappeared after the third and fourth postnatal weeks. Already in the neonate two types of vasopressin-positive fibers are observed in the median eminence, characterized by the different diameters of their granules and by their typical location in the internal and the external pericapillary contact zone. Especially in one and two week-old animals, in the internal zone of the median eminence and, to a lesser degree in the neural lobe, the immuncytochemical reaction product is deposited on an axonal tubular network. Judging from the presence of very few vasopressin-negative fibers in the neural lobe of the neonate, the development of the oxytocin system appears to be delayed. A characteristic relationship between pituicytes and the neurosecretory fibers can be observed during the first two postnatal weeks. After the third postnatal week the immunocytochemical features of the vasopressin system correspond approximately to that in adult rats.