- Email: [email protected]
Nerve-dependent expression of high polysialic acid neural cell adhesion molecule in neurohypophysial astrocytes of adult rats
0306-4522/93 $6.00 + 0.00 Pergamon Press Ltd 0 1993 IBRO
Neuroscience Vol. 53, No. 1, pp. 213-221, 1993 Printed in Great Britain
NERVE-DEPENDENT EXPRESSION OF HIGH POLYSIALIC ACID NEURAL CELL ADHESION MOLECULE IN NEUROHYPOPHYSIAL ASTROCYTES OF ADULT RATS J. Z. KISS,*~ C. WANG* and G. ROUGON$ *Institute of Histology and Embryology, University of Geneva Medical School, 1 rue Michel Servet, CH- 12I 1 Geneva 4, Switzerland SBiologie de la differentiation cellulaire, Ura 179 CNRS, Fact&6 des Sciences de Luminy, Case 901, F-13288 Marseille Cedex 9, France Airstract--The adult hypothalamo-neurohypophysial system of the rat retains the capacity to express highly polysialylated isoforms of neural cell adhesion molecule normally expressed in developing tissues. Here we report that the expression of these isoforms in neurohypophysial astrocytes fpituicytes) may be regulated by neurosecretory cells. In the intact neurohypophysis a strong and homogeneously distributed immunostaining for the “embryonic”, highly sialylated form of neural cell adhesion molecules was detected by light-microscopic immunocytochemistry. By electron-microscopy, both neurosecretory axons and pituicytes were immunoreactive for this isoform. However, in contrast to the rather uniform staining on nerve fibres, polysialic acid immunolabelling on glial surfaces was uneven: immunostaining could be observed on ghal surfaces facing neuronal elements, but not at contact sites between pituicytes. In addition, most glial and neuronal elements were heavily and evenly labelled with the polyclonal antibody recognizing “total” neural cell adhesion molecule. Surgical transection of the hypophysial stalk, a procedure that eliminates descending neurosecretory axons from the neurohypophysis, resulted in the complete disappearance of polysialic immunoreactivity from the neurohypophysis. The electron-microscopic analysis confirmed that cell surfaces of pituicytes lacked this immunoreactivity after the lesion. When residual neurosecretory axons were observed following an incomplete lesion, immunoreactivity on axons and glial processes was maintained. Transection did not affect the distribution of “total” neural cell adhesion molecule. We postulate that the presence of neurosecretory axons in the neurohypophysis is necessary to maintain the capacity of pituicytes to express immunoreactivity for the polysialylated isoforms of neural cell adhesion molecule but riot the neural cell adhesion molecule itself since immunoreactivity for “total” neural cell adhesion molecule was unaltered after hypophysial stalk transection. This constitutes an example of how neuron-glia interactions could influence properties of the glial cell surface, which, in turn, could be a critical determinant of plasticity and regenerative capacities of the hypothalamo-neurohypophysial system.
The hypothalamo-neurohypophysial system (HNS), that produces the neurohormones vasopressin and oxytocin, has become a model system to study morphological plasticity. This is related to the recognition of two specific features of the system. First, it is now well established that a reversible reorganization of the structure of the HNS is induced by diverse physiological stimuli.*,‘7,30 For instance, under basal condition of hormone release, pituicytes, the resident astroglia in the neural lobe of the pituitary gland (NL),“s~~ surround neurosecretory axons and axonal endings and interpose their processes between the neurosecretory ending and the basal lamina. The glial isolation of nerve terminals may represent a physical barrier to hormone secretion. Under conditions of stimulated hormone release, for example parturition, lactation or osmotic challenge, tTo whom correspondence should be addressed. HNS, hypothalamo-neurohypophysial system; N-CAM, neural cell adhesion molecule; NL, neural lobe of the pituitary gland; F’SA, polysialic acid.
Abbreoiutions:
glial coverage of neurosecretory profiles markedly diminishes, resulting in an increase in the extent of neurovascular contacts. These changes are rapid and reversible since glial processes reoccupy the perivascular space and reappear between the neurosecretory elements once stimulation ceases.**17*30 Second, neurosecretory neurons were shown to exhibit a pronounced capacity for regeneration in the adult mammal. For example, spontaneous regeneration of neurosecretory axons was seen after axotomy in the median eminence or hypophysial stalk in the rat.24 Moreover, evidence suggests that neurosecretory axons regenerate abundantly into the NL or connective tissue grafted in contact with the proximal tips of these axons. ‘o-12Since the regeneration of neurosecretory fibres was limited when fibres were transected in the hypothalamus, a specific, permissive role of the microenvironment of the neurohypophysis has been hypothesized.rO*‘r Understanding the cellular and molecular mechanisms which permit the structural dynamism and regeneration of the HNS may provide clues leading to 213
214
1. L.
Kiss (‘I trl.
the discovery of conditions which promote regeneration and functional recovery in the adult mammalian brain. Many different molecules arc currently attributed a role in neuronal-glial interactions under-
lying regeneration and plasticity (for reviews, see Refs 7, 16). Of these, the neural cell adhesion molecules (N-CAM) appear to be critical since they participate in cellcell and cell-substrate interactions, thereby regulating tissue architecture. N-CAMS are expressed by neurons and glial cells and exist in a variety of isoforms, differing in their protein and carbohydrate moieties.‘4,‘5 In developing tissues, N-CAM contains more than 30% sialic acid, present in a polymeric form termed polysialic acid (PSA), in which the residues are c(2-8 linked (for review, see Ref. 13). This isoform is gradually replaced by several “adult” isoforms with a lower content of sialic acid.5,6,9In the rodent brain, the conversion takes place during the late perinatal or early postnatal periods but its timing differs in different regions, as they acquire their definitive structure.4*25 The timing of the conversion and the fact that the ability of N-CAM to promote cell adhesion is attenuated by its PSA content have suggested that the PSA moiety is essential for normal growth of the developing nervous system. Indeed, PSA-N-CAM expression may be a characteristic of motile cells or of cells undergoing reshaping (for review, see Refs 13,27). The recent demonstration of the high PSA containing, “embryonic” form of N-CAM in the NHS led to the suggestion that this molecule has a role in specific neuron-glia interactions which permit adult morphological plasticity and regeneration of the system.13 In the adult NL, high PSA N-CAM has been demonstrated on the surface of both neurosecretory axons and pituicytes.” Since the ability of glial cells to support the reversible remodelling and regeneration may depend on the presence of PSA, it is important to understand mechanisms which regulate the expression of this molecule in glial cell. To appreciate the role of neurosecretory axons in regulating PSA expression we surgically lesioned descending neurosecretory axons in the stalk region of the pituitary gland. This lesion was followed by immunocytochemical detection of NCAM and its PSA isoforms on the cell surface of pituicytes. EXPERIMENTAL
PROCEDURES
Male Sprague-Dawley rats (Sivz, Zurich, Switzerland) weighing 2W25Og were kept under standard laboratory conditions: 12-12 h lightdark cycle, and free access to food and water. To interrupt fibres travelling to the neurohypophysis, the pituitary stalk was transected by means of a Halasz (bayonet shaped) knife with a radius of 1 mm and height of 1.2 mm, as described previously.2i The animals’ heads were fixed in a Kopf stereotaxic instrument (O” head position) and all surgery was performed under pentobarbital (Nembutal) anaesthesia. The knife was inserted in a rostra1 position into the brain, at 2.5 mm behind the bregma until the tip of the blade touched the floor. Then, it was rotated
90 to the left and right. In controls. the knife was stopped 200.300pm short of the basal surface. Light-microscopic immunocytochenli.rrrJ, Rats were kept two to seven days postoperatively; they were anaesthetized with ether and fixed by transaortic perfusion of ice cold (4°C) 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2. Pituitaries were removed and placed in the same fixative overnight. then washed in several changes in phosphate buffer and immersed in 15% sucrose solution in 0.1 M phosphate buffer. Sections (8 )irn) from controls and lesioned rats were cut with a cryostat, fixed on gelatine-coated slides and processed immediately for immunocytochemistry. Sections were incubated with the primary antibody at 4°C for 24 h. A mouse monoclonal IgM antibody (anti Men B)16recognizing polymers of 2,glinked sialic acid units was used to visualize the PSA isoform of N-CAM. Dilution of ascites fluid was I :400. The site-directed rabbit polyclonal antibody recognizing the NH,-terminus of N-CAM (anti-total N-CAM)26 was diluted I : 1000. Monoclonal antibodies, PS 44 and PS 38, a gift from H. Gainer, were used to stain vasopressin and oxytocin-associated neurophysins, respectively.3.” Dilution of supematants was 1: 200. Triton-X 100 (0.3%) was used in the solution of first antibodies. After rinsing, sections were incubated with fluorescein-conjugated secondary antibodies (diluted I : 100 in 0. I M phosphate buffer pH 7.4.) for 2 h at room temperature. To reveal the anti-Men B, PS 44 and PS 38 monoclonal antibodies, affinity-purified anti-mouse IgM and IgG antibodies (Boehringer Mannheim, Rotkreuz, Switzerland) were used. Anti-total N-CAM antibody was visualized with an anti-rabbit IgG antibody (Boehringer Mannheim, Rotkeuz, Switzerland). The sections were stained with Evan’s Blue and mounted under coverslips. Sections were examined with a Zeiss Axiophot fluorescence microscope. The quality of the lesion was determined on the basis of (i) visual inspections of the lesion site on the ventral surface of the brain, and (ii) the absence of neurophysin immunoreactivity in the NL of the pituitary gland. Electron-microscopic immunocytochrmistry For electron microscopy, rats were anaesthetized and fixed by transaortic perfusion of ice-cold (4°C) 4% paraformaldehyde and 0.1% glutaraldehyde in sodium phosphate buffer. Pituitaries were removed and coronal sections of 40pm thickness were cut with a Vibratome (Oxford) and processed by the pre-embedding immunoperoxidase technique as described previously.% Briefly, sections were incubated with anti-Men-B (1:406) or with-anti-total N-CAM (1: 1000) at 4°C for 24 h: washed in nhosuhatebuffered saline, incubated in anti-mouse IgG and I&I or anti-rabbit IgG conjugated to peroxidase (Boehringer Mannheim, Rotkeuz, Switzerland) for 1 h. After washing in phosphate-buffered saline, sections were immersed in 3,3diaminobenzidine tetrahydrochloride (Sigma). Triton-X 100 was not used in any of the solutions. Control sections were treated in the same way, except that the first antibody was replaced with phosphate buffer. Sections to be used for electron microscopy were osmicated, dehydrated and embedded in Epon. Thin sections (without counterstaining with uranyl acetate or lead citrate) were examined in a Jeol- 1OOCX 11 microscope. RESULTS
Light-microscopic
immunocytochemistry
Our observations concerning the normal pituitary gland are in agreement with those reported in the recent paper of Theodosis et al.)’ In the NL, a strong, homogeneously distributed PSA immunolabelling was detected (Fig. 1A). At the light microscopic level,
Nerve-dependent
215
expression of PSA-N-CAM
Fig. 1. Immunocytochemical localization of PSA and “total” N-CAM in the pituitary gland of control and hypophysial stalk-transected animals. (A) PSA immunoreactivity in a control hypophysis. Note the strong labelling in the NL and the few labelled cells in the anterior lobe (AL). (B) “Total” N-CAM immunoreactivity in a control hypophysis. (C) PSA immunoreactivity after lesion. Note the absence of immunoreactivity in the NL, and the presence of labelled cells in the anterior lobe. (D) “Total” N-CAM immunoreactivity after lesion. All three lobes of the hypophysis are heavily labelled. This staining pattern is similar in control and lesioned animals. IL, intermediate lobe. x 188.
it was impossible to distinguish between immunolabelled neuronal and glial elements. In the intermediate and anterior lobes (Fig. lA), only a few scattered immunopositive cells were present. In contrast to this, all three lobes were intensely and homogeneously stained with the polyclonal antibody against total N-CAM (Fig. 1B). Seven days after hypophysial stalk transection, practically all neurosecretory fibres were degenerated as evidenced by the absence of vasopressin- and oxytocin-associated neurophysin staining in the posterior lobe (data not shown). Concomitantly, PSA immunostaining disappeared completely from the NL, whereas the scattered PSA staining in the other two lobes was still present after the lesion (Fig. 1C). Moreover, in some cases, more immunopositive cells were observed in the anterior lobe in lesioned than in control animals (Fig. 1C). Thus, the disappearance of PSA staining in the NL was not due to some nonspecific effect of the lesion which would perturb the capacity of the pituitary gland to express PSA. We also observed that the lesion did not markedly change the strong PSA immunoreactivity of neurosecretory cell bodies in the supraoptic and para-
ventricular nuclei and that of neurosecretory fibres in the median eminence (data not shown). In contrast to PSA immunoreactivity, total N-CAM immunoreactivity in the NL as well as in the other two lobes appeared unaltered following the lesion (Fig. 1D). It appears therefore, that the adult form of N-CAM immunoreactivity is continually expressed in the NL as well as in other parts of the pituitary gland after the lesion. Interestingly, when lesions were incomplete, as evidenced by the presence of a few residual fibres still revealed by neurophysin staining, neurosecretory fibres and their immediate environment were found to be immunolabelled for PSA (Fig. 2AB). The distribution of PSA was investigated by electronmicroscopy. Electron-microscopic
immunocytochemistry
On tissue sections stained by the preembedding immunoperoxidase technique, PSA immunoreactivity was found on the surface of neurosecretory fibres, varicosities, as well as bouton terminals in the normal NL (Figs 3, 4A). Immunostaining was consistently detectable on unmyelinated, fasciculating axons (Fig. 3A). PSA immunoreactivity was also present on
Fig. 2. Immunofluorescent localization of vasopressinioxytocin-associated neurophysin and PSA m the hypophysis after lesion. (A) Neurophysin immunoreactivity in the NL of a lesioned animal. In this animal the lesion was incomplete, a substantial number of residual fibres are still visible (arrow). (B) Adjacent section of (A) stained with anti PSA antibody. Note that in the posterior lobe intensive labelling for PSA is present in the region where neurophysin staining is concentrated (arrow) (see A). AL, anterior lobe: IL. intermediate lobe. x 188.
the cell surface of glial processes and ceil bodies (Fig. 3B, 4A). In contrast to the rather uniform staining on nerve fibres, immunostaining on glial surface was uneven, and regions of high immunoreactivity and low immunoreactivity could be observed. In general, PSA immunoreactivity was highest on glial surfaces facing neuronal elements (Fig. 4A). Immunostaining was virtually absent on pituicyte surface facing other pituicytes (Fig. 4A). A region also consistently negative was the membrane of axon terminals, as well as that of pituicyte endfeet contacting the basement membrane of fenestrated capillaries (Fig. 4A). In lesioned animals, PSA immunoreactivity was abolished, except when fibres escaped transection. In this case, axons and surrounding glial surface appeared distinctly immunolabelled (Fig. 5). No diaminobenzidine reaction product was detected in control incubations, in which the primary antibody was omitted. “Total” N-CAM immunolabelling was found by electron microscopy on neuron-neuron, neuron-glia, glia-glia interfaces as well as (Fig. 4B) on the neuronal or glial areas facing the basement membrane of capillaries. This pattern of immunolabelling was not modified by the lesion (data not shown).
DISCUSSION
Our study confirms recent findings3’ that the adult HNS retains the capacity to express the embryonic form of N-CAM. Under normal conditions, pituicytes as well as neurosecretory axons are intensively labelled for PSA-N-CAM in the NL. Hypophysial stalk transection, that eliminates all descending neurosecretory fibres from the NL, results in a complete disappearance of PSA immunoreactivity, but not that of total N-CAM immunoreactivity, from cell surfaces of pituicytes. This supports the view that the
presence of PSA immunoreactivity on pituicytes depends on interactions between neurosecretory axons and pituicytes. The present data, as well as a recent report from Theodosis et al.,” indicate that under normal circumstances pituicytes are able to express PSA immunoreactivity on their cell surface. It is very likely that the immunoreactivity observed using the monoclonal antibody reacting with a 2-8 polysialosyl units,‘” is borne by the N-CAM molecule. A recent published study suggests that sodium channels carry homopolymers of z 2-&linked PSA.“’ However, the discrepancy between the distribution pattern of PSA’““’ and sodium channels” immunoreactivities in the rat brain makes it unlikely that sodium channels are tnajor carriers of PSA in the HNS. We cannot formally exclude, that PSA units are part of a highly sialylated glycolipid, though this is unlikely. since we had been unable to detect PSA immunoreactivity in ganglioside extracts prepared either from developing or adult mouse brain.” That astroglia in general, are able to express the PSA-N-CAM during the development is well established.‘.“,” However, in most regions of the adult brain, differentiated astrocytes, such as Bergman glia in the cerebellar cortex, do not express detectable astrocytes levels of PSA-N-CAM. ” PSA-positive have been described only in three discrete areas of the adult brain: the olfactory system,” the HNS” and the optic nerve.’ Observations on PSA-N-CAM in the optic nerve and retina indicated that immunoreactivity disappears from ganglion cell axons during development and persists in adults only on Miiller cells and astrocytes.’ This data, as well as observations that reactive astroglia in the limbic system following kainic acid treatment also transiently express PSA immunoreactivity** suggest that in some situations. astrocytes are indeed able to synthesize PSA-rich N-CAM in the adult.
Nerve-dependent
expression of PSA-N-CAM
Fig. 3. (A) Electron-microscopic images of PSA labelled neurosecretory axons from an intact NL. The area shown at higher power in inset is indicated by fine arrows (x 36,600; inset; x 66,770). (B) Electronmicroscopic images of PSA-labelled neuronal and glial elements in a control neurohypophysis. Note the presence of immunoreactivity on the surface of axonal (A) profiles as well as pituicytes (P). Capillary endothelium, basal membrane and the surface of terminals contacting the basal membrane are negative (arrows). PS, pericapillary space. x 34,090.
217
Fig. 4. Electron-microscopic images of immunoreactive elements from an intact neurohypophysis (A) PSA-labelled pituicytes (P) and axonal profiles (A) adjacent to a capillary. Note that g&al membrane facing neuronal profiles is PSA positive. but not that facing another glial surface (arrow heads). Glial as well as neuronal membranes contacting the basement membrane are immunonegative (arrow). (B) N-CAM total immunoreactivity at contact site (arrowheads) between two pituicytes. N. cell nucleus: L, lipid inclusions; PS, pericapillary space. x 31.100.
Nerve-dependent
expression of PSA-N-CAM
219
Fig. 5. Thin section of a neurohypophysis from a lesioncd animal. The majority of pituicyte surfaces (P) are negative for PSA. However, glial cell surfaces forming a direct environment of an immunolabelled, residual fibre (A) are selectively labelled. x 27,200.
Our ~l~tron-micro~opic study of the NL shows that PSA irnrnuno~~~~ty on gliai cell surface is unevenly distributed: labelling was mainly concentrated on membranes facing neuronal elements. This is in contrast to the rather homogeneous distribution pattern of total N-CAM immunoreactivity. A further example of polarized topography of the N-CAM molecule is provided by astrocytic endfeet in the retina which are immunoreactive all over their surface membrane except in contact sites with the basement membrane.2 In the present study, a similar distribution pattern of PSA-N-CAM was observed on membranes of neurosecretory terminals contacting the basement membrane of fenestrated capillaries. These observations should be interpreted with caution, since false-negative staining due to the limited penetration of the monoclonal anti-PSA antibody cannot be excluded. Nevertheless, it is possible that mechanisms exist which establish a polarized topog raphy for PSA-N-CAM. Such a polarized topography of N-CAM molecules on glial cells might result from selective targeting, surface modulation of the molecule or from some other unknown mechanisms. Whatever the underlying mechanism is, our results indicate that neurons have a regulatory role in these processes. The identi~~tion of the putative regulatory mechanisms and the description of the molecular
machinery involved (cell membrane contact or diffusible factor(s) released by neurons) will require the use of in vitro model systems. Finally, an alternative hypothesis to explain the results of our lesion study would be that pituicytes of the adult rat are in fact unable to synthesize and express PSA-N-CAM. According to this model, in the intact pituitary gland PSA immunoreactivity would be released from nerve fibres and subsequently tied on glial cell surface receptors. This hypothesis is consistent with the specific “neuron-related’” topography of PSA-N-CAM on pituicyte membrane. It is unlikely that PSA immunoreacti~ty revealed on glial surface is an artefact due to the liberation of this molecule from neuronal surface during tissue processing. Indeed, immunoreactivity was not detectable in the extracellular space or on surfaces of basement membranes, endothelial cells and pericytes located in the neighbourhood of neurosecretory axons. Different membraneassociated isoforms of the N-CAM molecules have been described (for review, see Ref. 13). One of them, the 120,000mol. wt isoform, has been shown to be anchored to membranes by a complex glycanphosphatidylinositol group and to exist in a soluble form under physiological conditions.‘8,32 This isoform is detectable in the adult HNS.31 Further studies
are needed to clarify if this N-CAM &form can be released from neurosecretory fibres or alternatively, whether sialic acid polymers can be released independently of the N-CAM protein core. The possible transfer of sialic acid (alone or conjugated to N-CAM) from neurosecretory axons to glial surface would be a new, previously unrecognized mechanism by which neurons may modify the glial environment
so that this became permissive regeneration of nerve fibres.
for plastictty
and
Acknowledgements-We wish to thank Mrs Esther Sutter for technical assistance and Mr J.-Ph. Gerber for photographic work. This work was supported by the Swiss National Foundation grant 31-32745.91 to J.Z.K. and grants from Association Francaise contre la Mvopathie and Association de Recherche contre le Cancer to 6.R.
REFERENCES
1. Bartsch U., Kirchhoff F. and Schachner M. (1989) lmmunohisto1ogical
lo~li~tion of the adhesion molecules LI, N-CAM, and MAG in the developing and adult optic nerve of mice. J. romp. Neural. 2S4, 451462. 2. Bartsch U., Kirchhoff F. and Schachner M. (19%) Highly sialylated N-CAM is expressed in adult mouse optic nerve and retina. J. Neurocytol. 19, 550---565. 3. Ben Barak Y., Russell J. ‘I., Whitnall M. H., Ozato K. and Gainer H. (1985) Neurophysin in the hypothalamoneurohypophysial system. 1. Production and characterization of monoclonal antibodies. J. Neurosci. 5, 81.-97. 4. Boisseau S., Nedelec J., Poirier V., Rougon G. and Simonneau M. (1991) Analysis of high PSA N-CAM expression during mammalian spinal cord and peripheral nervous system development. Development 112, 69-82. 5. Breen K. C. and Regan C. M. (1988) Developmental control of N-CAM sialylation state by Golgi sialytransferase isoforms. Development 104, 147-.I 54. 6. Breen K. C. and Regan C. M. (1988) Differentiation-dependent sialylation of individual neural cell adhesion molecule polypeptides during postnatal development. J. Neurochem. 50, 712-716. 7. Bunge R. P. and Hopkins J. M. (t990) The role of peripheral and central neuroglia in neural regeneration in vertebrates. Spmirturs Neurosci. 2, 509-.5 18. 8. Chapman D. B., Theodosis D. T., Montagnese C., Poulain D. A. and Morris J. F. (1986) Osmotic stimulation causes structural plasticity of neurone-glia reIationships of the oxytocin but not vasopressin secreting neurones in the hypothalamic supraoptic nucleus. ~~ros~~e~e~ 17, 679686. 9. Cunningham B. A., Hemperly J. J., Murray B. A., Prediger E. A., Brackenbury R. and Edelman G. M. (1987) Neural celt adhesion molecule: structure, immunoglobulin-like domains, ceil surface modulation, and alternative RNA splicing. Science 234, 7999806. 10. Dellmann H. D., Lue L. F. and Bellin S. 1. (1987) Neurosecretory axon regeneration into intrahypothalamic nctrral lobe allografts: neurophysin immunohistochemistry and fine structure. Expl Bruin Res. 67, 543-555. II. Dellmann H. D., Lue L. F. and Bellin S. I. (1987) Fine structural characteristics of neurophysin-positive perivascular plexus that develop in the rat hypothalamus following interruption of the hypothalamo-neurohypophysial tract. Cell Tiss. Res. 247, 137. 143. 12. Dellmann H. D., Lue L. F., Bellin S. I. and Oussat M. (1988) An immunohistochemical and fine-structural analysis of peptidergic hypothalamic neurosecretory axon regeneration into the leptomeninges of the rat. Brain Rex. 450, 181-189. 13. Edelman G. M. (1986) Cell adhesion molecules in the regulation of animal form and tissue pattern. A. Rev. Cell Bioi. 2, 81ll16. 14. Finne J. (1982) Occurrence of unique polysialosyl carbohydrate units in glycoproteins of developing brain. 1. hio!. C&m. 257, 11,966-l 1,970. 15. Finne .I., Finne U., Deagostini Bazin H. and Goridis C. (1983) Occurrence of alpha 2-8 linked polysialosyl units in a neural cell adhesion molecule. Biochem. biophys. Res Commun. 112,482487. 16. Hatten M. E., Liem R. K., Shelanski M. L. and Mason C. A. (1991) Astroglia in CNS injury. Gliu 4, 233 ,243. 17. Hatton G. I. (1990) Emerging concepts of structure-function dynamics in adult brain: the hypothalamoneurohypophysial system. Prog. Neurobiol. 34, 437-504. 18. He H. T., Finne J. and Goridis C. (1987) Biosynthesis, membrane association, and release of N-CAM-120, a phosphatidylinositol-linked form of the neural cell adhesion molecule. J. Cell Biol. 105, 2489-2500. 19. Hekmat A., Bitter Suermann D. and Schachner M. (1990) Immunocytological localization of the highly polysialylated form of the neural cell adhesion molecule during development of the murine cerebellar cortex. J. camp. Neural. 291, 457467. 20. Kiss J. 2. and Williams T. H. (1983) ACTH-immunoreactive boutons form synaptic contacts in the hypothalamic arcuate nucleus of rat: evidence for local opiocortin connections. Brain Res. 263, 142-146. 21. Kiss J. Z., Williams T. H. and Palkovits M. (1984) Distribution and projections of choiecystokinin-immunor~~ctive neurons in the hypothalamic paraventricular nucleus of rat. J. camp. Neurof. 227, l73-I 81. 22. Le-Gal-La-Salle G., Rougon G. and Valin A. (t 992) The embryonic form of neural celf surface molecute (E-NCAM) in the rat hip~campus and its reexpression on glial cells following kainic acid-induced status epilepticus. J. Nenrosci. 12, 872.--882. 23. Miragali F., Kadmon G. and Schachner M. (1989) Expression of Ll and N-CAM cell adhesion molecules during development of the mouse olfactory system. Devi Biof. 135, 272-286. 24. Raisman G. (1973) Electronmicroscopic studies of the development of new neurohemal contacts in the median eminence of rat after hypophysectomy. Brairf Res. 55, 245~-262. 25. Rougon G., Deagostini Bazin H., Hirn M. and Goridis C. (1982) Tissue- and developmental stage-specific forms of a neural ceil surface antigen linked to differences in glycosylation of a common polypeptide. Eur. molec. Biol. Org. J. 1, 1239.-1244 26. Rougon G., Dubois C., Buckley N., Magnani J. L. and Zollinger W. (1986) A monoclonal antibody against meningococcus group B polysaccharides distinguishes embryonic from adult N-CAM. J. Cell Biol. 103, 2429-2437. 27. Rutishauser U. and Landmesser L. (1991) Polysialic acid on the surface of axons regulates patterns of normal and activity-dependent innervation. Trends Nruro,sci. 14, 528 ~532
Nerve-dependent
expression of PSA-N-CAM
221
28. Salm A. K., Hatton G. I. and Nilaver G. (1982) Imfnunoreactive glial fibrillary acidic protein in pituicytes of the rat neurohypophysis. Brain Res. 236, 471476. 29. Suess U. and Pliska V. (1981) Identification of the pituicytes as astroglial cells by indirect immunofluorescence-staining for the glial fibrillary acidic protein. Brain Res. 221, 27-33. 30. Theodosis D. T. and Poulain D. A. (1984) Evidence for structural plasticity in the supraoptic nucleus of the rat hypothalamus in relation to gestation and lactation. Neuroscience 11, 183-193. 31. Theodosis D. T., Rougon G. and Poulain D. A. (1991) Retention of embryonic features by an adult neuronal system capable of plasticity: polysialylated neural cell adhesion molecule in the hypothalamo-neurohypophysial system. Proc. natn. Acad. Sci. U.S.A. 88, 54945498. 32. Theveniau M., Durbec P., Gennarini G., Wood J. and Rougon G. (1992) Expression and release of phosphatidylinositol anchored cell surface molecules by a cell line derived from sensory neurons. J. cell. Biochem. 48, 61-72. 33. Westenbroek R. E., Merrick D. K. and Catteral W. A. (1989) Differential subcellular localization of the Rl and RI 1
Na+ channel subtypes in central neurons. Neuron 3, 695-704. 34. Whitnall M. H., Key S., Ben Barak Y., Ozato K. and Gainer H. (1985) Neurophysin in the hypothalamoneurohypophysial system. II. Immunocytochemical studies of the ontogeny of oxytocinergic and vasopressinergic neurons. J. Neurosci. 5, 98-109. 35. Zuber C., Lackie P. M., Catteral W. A. and Roth J. (1992) Polysialic acid is associated with sodium channels and the neural cell adhesion molecule N-CAM in the adult rat brain. J. biol. Chem. 267, 9965-9971. (Accepted 24 September
1992)
Neuroscience Vol. 53, No. 1, pp. 213-221, 1993 Printed in Great Britain
NERVE-DEPENDENT EXPRESSION OF HIGH POLYSIALIC ACID NEURAL CELL ADHESION MOLECULE IN NEUROHYPOPHYSIAL ASTROCYTES OF ADULT RATS J. Z. KISS,*~ C. WANG* and G. ROUGON$ *Institute of Histology and Embryology, University of Geneva Medical School, 1 rue Michel Servet, CH- 12I 1 Geneva 4, Switzerland SBiologie de la differentiation cellulaire, Ura 179 CNRS, Fact&6 des Sciences de Luminy, Case 901, F-13288 Marseille Cedex 9, France Airstract--The adult hypothalamo-neurohypophysial system of the rat retains the capacity to express highly polysialylated isoforms of neural cell adhesion molecule normally expressed in developing tissues. Here we report that the expression of these isoforms in neurohypophysial astrocytes fpituicytes) may be regulated by neurosecretory cells. In the intact neurohypophysis a strong and homogeneously distributed immunostaining for the “embryonic”, highly sialylated form of neural cell adhesion molecules was detected by light-microscopic immunocytochemistry. By electron-microscopy, both neurosecretory axons and pituicytes were immunoreactive for this isoform. However, in contrast to the rather uniform staining on nerve fibres, polysialic acid immunolabelling on glial surfaces was uneven: immunostaining could be observed on ghal surfaces facing neuronal elements, but not at contact sites between pituicytes. In addition, most glial and neuronal elements were heavily and evenly labelled with the polyclonal antibody recognizing “total” neural cell adhesion molecule. Surgical transection of the hypophysial stalk, a procedure that eliminates descending neurosecretory axons from the neurohypophysis, resulted in the complete disappearance of polysialic immunoreactivity from the neurohypophysis. The electron-microscopic analysis confirmed that cell surfaces of pituicytes lacked this immunoreactivity after the lesion. When residual neurosecretory axons were observed following an incomplete lesion, immunoreactivity on axons and glial processes was maintained. Transection did not affect the distribution of “total” neural cell adhesion molecule. We postulate that the presence of neurosecretory axons in the neurohypophysis is necessary to maintain the capacity of pituicytes to express immunoreactivity for the polysialylated isoforms of neural cell adhesion molecule but riot the neural cell adhesion molecule itself since immunoreactivity for “total” neural cell adhesion molecule was unaltered after hypophysial stalk transection. This constitutes an example of how neuron-glia interactions could influence properties of the glial cell surface, which, in turn, could be a critical determinant of plasticity and regenerative capacities of the hypothalamo-neurohypophysial system.
The hypothalamo-neurohypophysial system (HNS), that produces the neurohormones vasopressin and oxytocin, has become a model system to study morphological plasticity. This is related to the recognition of two specific features of the system. First, it is now well established that a reversible reorganization of the structure of the HNS is induced by diverse physiological stimuli.*,‘7,30 For instance, under basal condition of hormone release, pituicytes, the resident astroglia in the neural lobe of the pituitary gland (NL),“s~~ surround neurosecretory axons and axonal endings and interpose their processes between the neurosecretory ending and the basal lamina. The glial isolation of nerve terminals may represent a physical barrier to hormone secretion. Under conditions of stimulated hormone release, for example parturition, lactation or osmotic challenge, tTo whom correspondence should be addressed. HNS, hypothalamo-neurohypophysial system; N-CAM, neural cell adhesion molecule; NL, neural lobe of the pituitary gland; F’SA, polysialic acid.
Abbreoiutions:
glial coverage of neurosecretory profiles markedly diminishes, resulting in an increase in the extent of neurovascular contacts. These changes are rapid and reversible since glial processes reoccupy the perivascular space and reappear between the neurosecretory elements once stimulation ceases.**17*30 Second, neurosecretory neurons were shown to exhibit a pronounced capacity for regeneration in the adult mammal. For example, spontaneous regeneration of neurosecretory axons was seen after axotomy in the median eminence or hypophysial stalk in the rat.24 Moreover, evidence suggests that neurosecretory axons regenerate abundantly into the NL or connective tissue grafted in contact with the proximal tips of these axons. ‘o-12Since the regeneration of neurosecretory fibres was limited when fibres were transected in the hypothalamus, a specific, permissive role of the microenvironment of the neurohypophysis has been hypothesized.rO*‘r Understanding the cellular and molecular mechanisms which permit the structural dynamism and regeneration of the HNS may provide clues leading to 213
214
1. L.
Kiss (‘I trl.
the discovery of conditions which promote regeneration and functional recovery in the adult mammalian brain. Many different molecules arc currently attributed a role in neuronal-glial interactions under-
lying regeneration and plasticity (for reviews, see Refs 7, 16). Of these, the neural cell adhesion molecules (N-CAM) appear to be critical since they participate in cellcell and cell-substrate interactions, thereby regulating tissue architecture. N-CAMS are expressed by neurons and glial cells and exist in a variety of isoforms, differing in their protein and carbohydrate moieties.‘4,‘5 In developing tissues, N-CAM contains more than 30% sialic acid, present in a polymeric form termed polysialic acid (PSA), in which the residues are c(2-8 linked (for review, see Ref. 13). This isoform is gradually replaced by several “adult” isoforms with a lower content of sialic acid.5,6,9In the rodent brain, the conversion takes place during the late perinatal or early postnatal periods but its timing differs in different regions, as they acquire their definitive structure.4*25 The timing of the conversion and the fact that the ability of N-CAM to promote cell adhesion is attenuated by its PSA content have suggested that the PSA moiety is essential for normal growth of the developing nervous system. Indeed, PSA-N-CAM expression may be a characteristic of motile cells or of cells undergoing reshaping (for review, see Refs 13,27). The recent demonstration of the high PSA containing, “embryonic” form of N-CAM in the NHS led to the suggestion that this molecule has a role in specific neuron-glia interactions which permit adult morphological plasticity and regeneration of the system.13 In the adult NL, high PSA N-CAM has been demonstrated on the surface of both neurosecretory axons and pituicytes.” Since the ability of glial cells to support the reversible remodelling and regeneration may depend on the presence of PSA, it is important to understand mechanisms which regulate the expression of this molecule in glial cell. To appreciate the role of neurosecretory axons in regulating PSA expression we surgically lesioned descending neurosecretory axons in the stalk region of the pituitary gland. This lesion was followed by immunocytochemical detection of NCAM and its PSA isoforms on the cell surface of pituicytes. EXPERIMENTAL
PROCEDURES
Male Sprague-Dawley rats (Sivz, Zurich, Switzerland) weighing 2W25Og were kept under standard laboratory conditions: 12-12 h lightdark cycle, and free access to food and water. To interrupt fibres travelling to the neurohypophysis, the pituitary stalk was transected by means of a Halasz (bayonet shaped) knife with a radius of 1 mm and height of 1.2 mm, as described previously.2i The animals’ heads were fixed in a Kopf stereotaxic instrument (O” head position) and all surgery was performed under pentobarbital (Nembutal) anaesthesia. The knife was inserted in a rostra1 position into the brain, at 2.5 mm behind the bregma until the tip of the blade touched the floor. Then, it was rotated
90 to the left and right. In controls. the knife was stopped 200.300pm short of the basal surface. Light-microscopic immunocytochenli.rrrJ, Rats were kept two to seven days postoperatively; they were anaesthetized with ether and fixed by transaortic perfusion of ice cold (4°C) 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.2. Pituitaries were removed and placed in the same fixative overnight. then washed in several changes in phosphate buffer and immersed in 15% sucrose solution in 0.1 M phosphate buffer. Sections (8 )irn) from controls and lesioned rats were cut with a cryostat, fixed on gelatine-coated slides and processed immediately for immunocytochemistry. Sections were incubated with the primary antibody at 4°C for 24 h. A mouse monoclonal IgM antibody (anti Men B)16recognizing polymers of 2,glinked sialic acid units was used to visualize the PSA isoform of N-CAM. Dilution of ascites fluid was I :400. The site-directed rabbit polyclonal antibody recognizing the NH,-terminus of N-CAM (anti-total N-CAM)26 was diluted I : 1000. Monoclonal antibodies, PS 44 and PS 38, a gift from H. Gainer, were used to stain vasopressin and oxytocin-associated neurophysins, respectively.3.” Dilution of supematants was 1: 200. Triton-X 100 (0.3%) was used in the solution of first antibodies. After rinsing, sections were incubated with fluorescein-conjugated secondary antibodies (diluted I : 100 in 0. I M phosphate buffer pH 7.4.) for 2 h at room temperature. To reveal the anti-Men B, PS 44 and PS 38 monoclonal antibodies, affinity-purified anti-mouse IgM and IgG antibodies (Boehringer Mannheim, Rotkreuz, Switzerland) were used. Anti-total N-CAM antibody was visualized with an anti-rabbit IgG antibody (Boehringer Mannheim, Rotkeuz, Switzerland). The sections were stained with Evan’s Blue and mounted under coverslips. Sections were examined with a Zeiss Axiophot fluorescence microscope. The quality of the lesion was determined on the basis of (i) visual inspections of the lesion site on the ventral surface of the brain, and (ii) the absence of neurophysin immunoreactivity in the NL of the pituitary gland. Electron-microscopic immunocytochrmistry For electron microscopy, rats were anaesthetized and fixed by transaortic perfusion of ice-cold (4°C) 4% paraformaldehyde and 0.1% glutaraldehyde in sodium phosphate buffer. Pituitaries were removed and coronal sections of 40pm thickness were cut with a Vibratome (Oxford) and processed by the pre-embedding immunoperoxidase technique as described previously.% Briefly, sections were incubated with anti-Men-B (1:406) or with-anti-total N-CAM (1: 1000) at 4°C for 24 h: washed in nhosuhatebuffered saline, incubated in anti-mouse IgG and I&I or anti-rabbit IgG conjugated to peroxidase (Boehringer Mannheim, Rotkeuz, Switzerland) for 1 h. After washing in phosphate-buffered saline, sections were immersed in 3,3diaminobenzidine tetrahydrochloride (Sigma). Triton-X 100 was not used in any of the solutions. Control sections were treated in the same way, except that the first antibody was replaced with phosphate buffer. Sections to be used for electron microscopy were osmicated, dehydrated and embedded in Epon. Thin sections (without counterstaining with uranyl acetate or lead citrate) were examined in a Jeol- 1OOCX 11 microscope. RESULTS
Light-microscopic
immunocytochemistry
Our observations concerning the normal pituitary gland are in agreement with those reported in the recent paper of Theodosis et al.)’ In the NL, a strong, homogeneously distributed PSA immunolabelling was detected (Fig. 1A). At the light microscopic level,
Nerve-dependent
215
expression of PSA-N-CAM
Fig. 1. Immunocytochemical localization of PSA and “total” N-CAM in the pituitary gland of control and hypophysial stalk-transected animals. (A) PSA immunoreactivity in a control hypophysis. Note the strong labelling in the NL and the few labelled cells in the anterior lobe (AL). (B) “Total” N-CAM immunoreactivity in a control hypophysis. (C) PSA immunoreactivity after lesion. Note the absence of immunoreactivity in the NL, and the presence of labelled cells in the anterior lobe. (D) “Total” N-CAM immunoreactivity after lesion. All three lobes of the hypophysis are heavily labelled. This staining pattern is similar in control and lesioned animals. IL, intermediate lobe. x 188.
it was impossible to distinguish between immunolabelled neuronal and glial elements. In the intermediate and anterior lobes (Fig. lA), only a few scattered immunopositive cells were present. In contrast to this, all three lobes were intensely and homogeneously stained with the polyclonal antibody against total N-CAM (Fig. 1B). Seven days after hypophysial stalk transection, practically all neurosecretory fibres were degenerated as evidenced by the absence of vasopressin- and oxytocin-associated neurophysin staining in the posterior lobe (data not shown). Concomitantly, PSA immunostaining disappeared completely from the NL, whereas the scattered PSA staining in the other two lobes was still present after the lesion (Fig. 1C). Moreover, in some cases, more immunopositive cells were observed in the anterior lobe in lesioned than in control animals (Fig. 1C). Thus, the disappearance of PSA staining in the NL was not due to some nonspecific effect of the lesion which would perturb the capacity of the pituitary gland to express PSA. We also observed that the lesion did not markedly change the strong PSA immunoreactivity of neurosecretory cell bodies in the supraoptic and para-
ventricular nuclei and that of neurosecretory fibres in the median eminence (data not shown). In contrast to PSA immunoreactivity, total N-CAM immunoreactivity in the NL as well as in the other two lobes appeared unaltered following the lesion (Fig. 1D). It appears therefore, that the adult form of N-CAM immunoreactivity is continually expressed in the NL as well as in other parts of the pituitary gland after the lesion. Interestingly, when lesions were incomplete, as evidenced by the presence of a few residual fibres still revealed by neurophysin staining, neurosecretory fibres and their immediate environment were found to be immunolabelled for PSA (Fig. 2AB). The distribution of PSA was investigated by electronmicroscopy. Electron-microscopic
immunocytochemistry
On tissue sections stained by the preembedding immunoperoxidase technique, PSA immunoreactivity was found on the surface of neurosecretory fibres, varicosities, as well as bouton terminals in the normal NL (Figs 3, 4A). Immunostaining was consistently detectable on unmyelinated, fasciculating axons (Fig. 3A). PSA immunoreactivity was also present on
Fig. 2. Immunofluorescent localization of vasopressinioxytocin-associated neurophysin and PSA m the hypophysis after lesion. (A) Neurophysin immunoreactivity in the NL of a lesioned animal. In this animal the lesion was incomplete, a substantial number of residual fibres are still visible (arrow). (B) Adjacent section of (A) stained with anti PSA antibody. Note that in the posterior lobe intensive labelling for PSA is present in the region where neurophysin staining is concentrated (arrow) (see A). AL, anterior lobe: IL. intermediate lobe. x 188.
the cell surface of glial processes and ceil bodies (Fig. 3B, 4A). In contrast to the rather uniform staining on nerve fibres, immunostaining on glial surface was uneven, and regions of high immunoreactivity and low immunoreactivity could be observed. In general, PSA immunoreactivity was highest on glial surfaces facing neuronal elements (Fig. 4A). Immunostaining was virtually absent on pituicyte surface facing other pituicytes (Fig. 4A). A region also consistently negative was the membrane of axon terminals, as well as that of pituicyte endfeet contacting the basement membrane of fenestrated capillaries (Fig. 4A). In lesioned animals, PSA immunoreactivity was abolished, except when fibres escaped transection. In this case, axons and surrounding glial surface appeared distinctly immunolabelled (Fig. 5). No diaminobenzidine reaction product was detected in control incubations, in which the primary antibody was omitted. “Total” N-CAM immunolabelling was found by electron microscopy on neuron-neuron, neuron-glia, glia-glia interfaces as well as (Fig. 4B) on the neuronal or glial areas facing the basement membrane of capillaries. This pattern of immunolabelling was not modified by the lesion (data not shown).
DISCUSSION
Our study confirms recent findings3’ that the adult HNS retains the capacity to express the embryonic form of N-CAM. Under normal conditions, pituicytes as well as neurosecretory axons are intensively labelled for PSA-N-CAM in the NL. Hypophysial stalk transection, that eliminates all descending neurosecretory fibres from the NL, results in a complete disappearance of PSA immunoreactivity, but not that of total N-CAM immunoreactivity, from cell surfaces of pituicytes. This supports the view that the
presence of PSA immunoreactivity on pituicytes depends on interactions between neurosecretory axons and pituicytes. The present data, as well as a recent report from Theodosis et al.,” indicate that under normal circumstances pituicytes are able to express PSA immunoreactivity on their cell surface. It is very likely that the immunoreactivity observed using the monoclonal antibody reacting with a 2-8 polysialosyl units,‘” is borne by the N-CAM molecule. A recent published study suggests that sodium channels carry homopolymers of z 2-&linked PSA.“’ However, the discrepancy between the distribution pattern of PSA’““’ and sodium channels” immunoreactivities in the rat brain makes it unlikely that sodium channels are tnajor carriers of PSA in the HNS. We cannot formally exclude, that PSA units are part of a highly sialylated glycolipid, though this is unlikely. since we had been unable to detect PSA immunoreactivity in ganglioside extracts prepared either from developing or adult mouse brain.” That astroglia in general, are able to express the PSA-N-CAM during the development is well established.‘.“,” However, in most regions of the adult brain, differentiated astrocytes, such as Bergman glia in the cerebellar cortex, do not express detectable astrocytes levels of PSA-N-CAM. ” PSA-positive have been described only in three discrete areas of the adult brain: the olfactory system,” the HNS” and the optic nerve.’ Observations on PSA-N-CAM in the optic nerve and retina indicated that immunoreactivity disappears from ganglion cell axons during development and persists in adults only on Miiller cells and astrocytes.’ This data, as well as observations that reactive astroglia in the limbic system following kainic acid treatment also transiently express PSA immunoreactivity** suggest that in some situations. astrocytes are indeed able to synthesize PSA-rich N-CAM in the adult.
Nerve-dependent
expression of PSA-N-CAM
Fig. 3. (A) Electron-microscopic images of PSA labelled neurosecretory axons from an intact NL. The area shown at higher power in inset is indicated by fine arrows (x 36,600; inset; x 66,770). (B) Electronmicroscopic images of PSA-labelled neuronal and glial elements in a control neurohypophysis. Note the presence of immunoreactivity on the surface of axonal (A) profiles as well as pituicytes (P). Capillary endothelium, basal membrane and the surface of terminals contacting the basal membrane are negative (arrows). PS, pericapillary space. x 34,090.
217
Fig. 4. Electron-microscopic images of immunoreactive elements from an intact neurohypophysis (A) PSA-labelled pituicytes (P) and axonal profiles (A) adjacent to a capillary. Note that g&al membrane facing neuronal profiles is PSA positive. but not that facing another glial surface (arrow heads). Glial as well as neuronal membranes contacting the basement membrane are immunonegative (arrow). (B) N-CAM total immunoreactivity at contact site (arrowheads) between two pituicytes. N. cell nucleus: L, lipid inclusions; PS, pericapillary space. x 31.100.
Nerve-dependent
expression of PSA-N-CAM
219
Fig. 5. Thin section of a neurohypophysis from a lesioncd animal. The majority of pituicyte surfaces (P) are negative for PSA. However, glial cell surfaces forming a direct environment of an immunolabelled, residual fibre (A) are selectively labelled. x 27,200.
Our ~l~tron-micro~opic study of the NL shows that PSA irnrnuno~~~~ty on gliai cell surface is unevenly distributed: labelling was mainly concentrated on membranes facing neuronal elements. This is in contrast to the rather homogeneous distribution pattern of total N-CAM immunoreactivity. A further example of polarized topography of the N-CAM molecule is provided by astrocytic endfeet in the retina which are immunoreactive all over their surface membrane except in contact sites with the basement membrane.2 In the present study, a similar distribution pattern of PSA-N-CAM was observed on membranes of neurosecretory terminals contacting the basement membrane of fenestrated capillaries. These observations should be interpreted with caution, since false-negative staining due to the limited penetration of the monoclonal anti-PSA antibody cannot be excluded. Nevertheless, it is possible that mechanisms exist which establish a polarized topog raphy for PSA-N-CAM. Such a polarized topography of N-CAM molecules on glial cells might result from selective targeting, surface modulation of the molecule or from some other unknown mechanisms. Whatever the underlying mechanism is, our results indicate that neurons have a regulatory role in these processes. The identi~~tion of the putative regulatory mechanisms and the description of the molecular
machinery involved (cell membrane contact or diffusible factor(s) released by neurons) will require the use of in vitro model systems. Finally, an alternative hypothesis to explain the results of our lesion study would be that pituicytes of the adult rat are in fact unable to synthesize and express PSA-N-CAM. According to this model, in the intact pituitary gland PSA immunoreactivity would be released from nerve fibres and subsequently tied on glial cell surface receptors. This hypothesis is consistent with the specific “neuron-related’” topography of PSA-N-CAM on pituicyte membrane. It is unlikely that PSA immunoreacti~ty revealed on glial surface is an artefact due to the liberation of this molecule from neuronal surface during tissue processing. Indeed, immunoreactivity was not detectable in the extracellular space or on surfaces of basement membranes, endothelial cells and pericytes located in the neighbourhood of neurosecretory axons. Different membraneassociated isoforms of the N-CAM molecules have been described (for review, see Ref. 13). One of them, the 120,000mol. wt isoform, has been shown to be anchored to membranes by a complex glycanphosphatidylinositol group and to exist in a soluble form under physiological conditions.‘8,32 This isoform is detectable in the adult HNS.31 Further studies
are needed to clarify if this N-CAM &form can be released from neurosecretory fibres or alternatively, whether sialic acid polymers can be released independently of the N-CAM protein core. The possible transfer of sialic acid (alone or conjugated to N-CAM) from neurosecretory axons to glial surface would be a new, previously unrecognized mechanism by which neurons may modify the glial environment
so that this became permissive regeneration of nerve fibres.
for plastictty
and
Acknowledgements-We wish to thank Mrs Esther Sutter for technical assistance and Mr J.-Ph. Gerber for photographic work. This work was supported by the Swiss National Foundation grant 31-32745.91 to J.Z.K. and grants from Association Francaise contre la Mvopathie and Association de Recherche contre le Cancer to 6.R.
REFERENCES
1. Bartsch U., Kirchhoff F. and Schachner M. (1989) lmmunohisto1ogical
lo~li~tion of the adhesion molecules LI, N-CAM, and MAG in the developing and adult optic nerve of mice. J. romp. Neural. 2S4, 451462. 2. Bartsch U., Kirchhoff F. and Schachner M. (19%) Highly sialylated N-CAM is expressed in adult mouse optic nerve and retina. J. Neurocytol. 19, 550---565. 3. Ben Barak Y., Russell J. ‘I., Whitnall M. H., Ozato K. and Gainer H. (1985) Neurophysin in the hypothalamoneurohypophysial system. 1. Production and characterization of monoclonal antibodies. J. Neurosci. 5, 81.-97. 4. Boisseau S., Nedelec J., Poirier V., Rougon G. and Simonneau M. (1991) Analysis of high PSA N-CAM expression during mammalian spinal cord and peripheral nervous system development. Development 112, 69-82. 5. Breen K. C. and Regan C. M. (1988) Developmental control of N-CAM sialylation state by Golgi sialytransferase isoforms. Development 104, 147-.I 54. 6. Breen K. C. and Regan C. M. (1988) Differentiation-dependent sialylation of individual neural cell adhesion molecule polypeptides during postnatal development. J. Neurochem. 50, 712-716. 7. Bunge R. P. and Hopkins J. M. (t990) The role of peripheral and central neuroglia in neural regeneration in vertebrates. Spmirturs Neurosci. 2, 509-.5 18. 8. Chapman D. B., Theodosis D. T., Montagnese C., Poulain D. A. and Morris J. F. (1986) Osmotic stimulation causes structural plasticity of neurone-glia reIationships of the oxytocin but not vasopressin secreting neurones in the hypothalamic supraoptic nucleus. ~~ros~~e~e~ 17, 679686. 9. Cunningham B. A., Hemperly J. J., Murray B. A., Prediger E. A., Brackenbury R. and Edelman G. M. (1987) Neural celt adhesion molecule: structure, immunoglobulin-like domains, ceil surface modulation, and alternative RNA splicing. Science 234, 7999806. 10. Dellmann H. D., Lue L. F. and Bellin S. 1. (1987) Neurosecretory axon regeneration into intrahypothalamic nctrral lobe allografts: neurophysin immunohistochemistry and fine structure. Expl Bruin Res. 67, 543-555. II. Dellmann H. D., Lue L. F. and Bellin S. I. (1987) Fine structural characteristics of neurophysin-positive perivascular plexus that develop in the rat hypothalamus following interruption of the hypothalamo-neurohypophysial tract. Cell Tiss. Res. 247, 137. 143. 12. Dellmann H. D., Lue L. F., Bellin S. I. and Oussat M. (1988) An immunohistochemical and fine-structural analysis of peptidergic hypothalamic neurosecretory axon regeneration into the leptomeninges of the rat. Brain Rex. 450, 181-189. 13. Edelman G. M. (1986) Cell adhesion molecules in the regulation of animal form and tissue pattern. A. Rev. Cell Bioi. 2, 81ll16. 14. Finne J. (1982) Occurrence of unique polysialosyl carbohydrate units in glycoproteins of developing brain. 1. hio!. C&m. 257, 11,966-l 1,970. 15. Finne .I., Finne U., Deagostini Bazin H. and Goridis C. (1983) Occurrence of alpha 2-8 linked polysialosyl units in a neural cell adhesion molecule. Biochem. biophys. Res Commun. 112,482487. 16. Hatten M. E., Liem R. K., Shelanski M. L. and Mason C. A. (1991) Astroglia in CNS injury. Gliu 4, 233 ,243. 17. Hatton G. I. (1990) Emerging concepts of structure-function dynamics in adult brain: the hypothalamoneurohypophysial system. Prog. Neurobiol. 34, 437-504. 18. He H. T., Finne J. and Goridis C. (1987) Biosynthesis, membrane association, and release of N-CAM-120, a phosphatidylinositol-linked form of the neural cell adhesion molecule. J. Cell Biol. 105, 2489-2500. 19. Hekmat A., Bitter Suermann D. and Schachner M. (1990) Immunocytological localization of the highly polysialylated form of the neural cell adhesion molecule during development of the murine cerebellar cortex. J. camp. Neural. 291, 457467. 20. Kiss J. 2. and Williams T. H. (1983) ACTH-immunoreactive boutons form synaptic contacts in the hypothalamic arcuate nucleus of rat: evidence for local opiocortin connections. Brain Res. 263, 142-146. 21. Kiss J. Z., Williams T. H. and Palkovits M. (1984) Distribution and projections of choiecystokinin-immunor~~ctive neurons in the hypothalamic paraventricular nucleus of rat. J. camp. Neurof. 227, l73-I 81. 22. Le-Gal-La-Salle G., Rougon G. and Valin A. (t 992) The embryonic form of neural celf surface molecute (E-NCAM) in the rat hip~campus and its reexpression on glial cells following kainic acid-induced status epilepticus. J. Nenrosci. 12, 872.--882. 23. Miragali F., Kadmon G. and Schachner M. (1989) Expression of Ll and N-CAM cell adhesion molecules during development of the mouse olfactory system. Devi Biof. 135, 272-286. 24. Raisman G. (1973) Electronmicroscopic studies of the development of new neurohemal contacts in the median eminence of rat after hypophysectomy. Brairf Res. 55, 245~-262. 25. Rougon G., Deagostini Bazin H., Hirn M. and Goridis C. (1982) Tissue- and developmental stage-specific forms of a neural ceil surface antigen linked to differences in glycosylation of a common polypeptide. Eur. molec. Biol. Org. J. 1, 1239.-1244 26. Rougon G., Dubois C., Buckley N., Magnani J. L. and Zollinger W. (1986) A monoclonal antibody against meningococcus group B polysaccharides distinguishes embryonic from adult N-CAM. J. Cell Biol. 103, 2429-2437. 27. Rutishauser U. and Landmesser L. (1991) Polysialic acid on the surface of axons regulates patterns of normal and activity-dependent innervation. Trends Nruro,sci. 14, 528 ~532
Nerve-dependent
expression of PSA-N-CAM
221
28. Salm A. K., Hatton G. I. and Nilaver G. (1982) Imfnunoreactive glial fibrillary acidic protein in pituicytes of the rat neurohypophysis. Brain Res. 236, 471476. 29. Suess U. and Pliska V. (1981) Identification of the pituicytes as astroglial cells by indirect immunofluorescence-staining for the glial fibrillary acidic protein. Brain Res. 221, 27-33. 30. Theodosis D. T. and Poulain D. A. (1984) Evidence for structural plasticity in the supraoptic nucleus of the rat hypothalamus in relation to gestation and lactation. Neuroscience 11, 183-193. 31. Theodosis D. T., Rougon G. and Poulain D. A. (1991) Retention of embryonic features by an adult neuronal system capable of plasticity: polysialylated neural cell adhesion molecule in the hypothalamo-neurohypophysial system. Proc. natn. Acad. Sci. U.S.A. 88, 54945498. 32. Theveniau M., Durbec P., Gennarini G., Wood J. and Rougon G. (1992) Expression and release of phosphatidylinositol anchored cell surface molecules by a cell line derived from sensory neurons. J. cell. Biochem. 48, 61-72. 33. Westenbroek R. E., Merrick D. K. and Catteral W. A. (1989) Differential subcellular localization of the Rl and RI 1
Na+ channel subtypes in central neurons. Neuron 3, 695-704. 34. Whitnall M. H., Key S., Ben Barak Y., Ozato K. and Gainer H. (1985) Neurophysin in the hypothalamoneurohypophysial system. II. Immunocytochemical studies of the ontogeny of oxytocinergic and vasopressinergic neurons. J. Neurosci. 5, 98-109. 35. Zuber C., Lackie P. M., Catteral W. A. and Roth J. (1992) Polysialic acid is associated with sodium channels and the neural cell adhesion molecule N-CAM in the adult rat brain. J. biol. Chem. 267, 9965-9971. (Accepted 24 September
1992)