Tanycytes Present in the Adult Rat Mediobasal Hypothalamus Support the Regeneration of Monoaminergic Axons

EXPERIMENTAL NEUROLOGY ARTICLE NO.

151, 1–13 (1998)

EN986784

Tanycytes Present in the Adult Rat Mediobasal Hypothalamus Support the Regeneration of Monoaminergic Axons Norbert Chauvet, Monica Prieto, and Gerard Alonso INSERM U336, University of Montpellier II, Montpellier, France Received August 11, 1997; accepted January 14, 1998

INTRODUCTION We have recently shown that tanycytes present in the median eminence (ME) constitute a preferential support for the regeneration of lesioned neurohypophysial oxytocinergic and vasopressinergic axons. However, although tanycytes are particularly abundant in the ME, they are also present along the third ventricle wall. This study was thus undertaken to determine whether tanycytes present in the mediobasal hypothalamus overlying the ME were also able to support the regeneration of the numerous monoaminergic axons innervating this region. Using confocal laser scanning microscopy combined with double or triple fluorescence immunostaining, we have compared the relationships occurring between glial cells and lesioned catecholaminergic and serotonergic axons at the levels of surgical cuts placed in the dorsomedial hypothalamus devoid of tanycytes or in the ventromedial hypothalamus containing numerous tanycyte processes. In dorsal lesions, catecholaminergic and serotonergic transected fibers were found to abut onto the scar formed along the surgical cut and composed of closely inderdigitating astrocyte processes strongly immunoreactive for both glial fibrillary acidic protein (GFAP) and vimentin (VIM). In ventral lesions, the lesional scar was composed of GFAP-immunoreactive (IR) and VIM-IR astrocyte processes and of VIM-IR but GFAP-negative processes that were identified as tanycytic processes. In all the ventral lesions examined, numerous catecholaminergic and serotonergic fibers were found to regenerate into the surgical cut in association with the VIM-IR, GFAP-negative tanycyte processes. On the other hand, such regenerating fibers were never found in scar portions containing only GFAP-IR astrocytic structures. These data indicate that, like in the ME, tanycytes present in the mediobasal hypothalamus of adult rat provide a substrate that favors the regeneration of lesioned axons.

In contrast with the peripheral nervous system (PNS), in the central nervous system (CNS) of adult mammals neurons do not spontaneously regenerate after injury. However, it has been clearly demonstrated that, when these central neurons are provided with PNS tissue, they have the intrinsic ability to regenerate their axons over long distances (1). It is thus accepted that the failure of adult CNS neurons to regenerate axons mainly results from an inhibition of the regenerating capacities by their microenvironment. Reactive astrocytes and activated microglia, which are the major cellular components of glial scar that develop around CNS lesion, have been proposed to be the major cell types responsible for such an adverse environment. While it is likely that the density of the glial scar creates a structural barrier that impedes axon regrowth (34), a recent concept suggests that abortive regeneration of lesioned neurons mostly result from molecular properties of astrocytes and oligodendrocytes (8, 40, 42). With the exception of the olfactory system (22), the neurohypophysial system of adult mammals appears unique with respect to its robust regenerative capacities. Indeed, it has long been known that hypothalamoneurohypophysial neurons are capable to regenerate axons when transected at the level of the median eminence (16). The ability of this system to quickly regenerate following axotomy may in large part be due to the presence of a specialized type of glial cells located in the median eminence (ME): the so-called tanycytes. Indeed, tanycytes represent the most abundant cell type of this organ, with numerous cell bodies lining the floor of the third ventricle and processes extending perpendicularly toward the external layer of the organ where they ramify and terminate around the local capillary plexus. We have recently shown that, in the ME, the regenerative sprouting of oxytocin and to a lesser extent of vasopressin magnocellular axons always occurs in close association with tanycyte processes (12). In addition to the ME itself, tanycytes are

r 1998 Academic Press

Key Words: astrocytes; tyrosine hydroxylase; serotonin; immunocytochemistry; confocal microscopy.

1

0014-4886/98 $25.00 Copyright r 1998 by Academic Press All rights of reproduction in any form reserved.

2

CHAUVET, PRIETO, AND ALONSO

also present in a large portion of the mediobasal hypothalamus. Tanycyte cell bodies are located on or just beneath the ventricular surface all along the ventral half portion of the third ventricle where they project long arch-shaped processes lateroventrally toward the ventral surface of the hypothalamus (18). The aim of the present study was to determine whether, similar to the tanycytes present in the ME, these tanycytes projecting throughout the mediobasal hypothalamus were able to support the regeneration of axons innervating this region. For this, two types of surgical lesions were made through the hypothalamus of adult rats either: (i) a lesion through the ventromedial hypothalamic region containing numerous tanycyte processes or (ii) a lesion through the dorsomedial hypothalamic region which is devoid of tanycyte cell bodies and processes. Among other axons types, these two regions contain a particularly dense network of catecholaminergic and serotonergic fibers. Confocal microscopy combined with double or triple immunocytochemical labeling was thus used to study the postlesional responses of catecholaminergic and serotonergic axons and their relationships with glial cells at the level of both ventral and dorsal intrahypothalamic lesions. MATERIALS AND METHODS

Animals Male adult Sprague–Dawley rats (Iffa-Credo, L’Arbresle, France) were used. They were kept in light (12 h light:12 h dark)- and temperature (24 6 1°C)-controlled rooms and had free access to standard dry food and tap water. Surgical Lesions Lesions were performed according to the method originally described by Hala`sz and Pupp (23). After

deep anaesthesia with equithesin (3 ml/kg), the animals were fixed in a stereotaxic device, and knife lesions were placed bilaterally in the hypothalamus according to the stereotaxic atlas of Paxinos and Watson (33). The knife used consisted of an L-shaped knife with an obtuse angle of 110° (radius 1.2 mm) made from a 25-gauge 3.5 spinal needle. The tip was oriented backwards and placed at 6.5 mm anterior to the interaural line. It was then lowered through the midline to 8.5 (n 5 5) or to 10 mm (n 5 7) below the surface of the skull and turned 360°. This was found to produce two types of horizontal lesions, one through the dorsomedial and the other through the ventromedial hypothalamus (Fig. 1). Immunocytochemical Procedures All lesioned animals were killed 3 weeks after surgery. Under deep anesthesia with sodium pentobarbital (60 mg/kg), rats were perfused through the ascending aorta with phosphate-buffered saline (PBS) followed by 500 ml of fixative composed of 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4. The brain was dissected and fixed by immersion in the same fixative for 2–4 days. The hypothalamus was then cut frontally with a vibratome into 30- to 40-µm-thick sections, which were carefully rinsed in PBS and subsequently treated for double or triple fluorescence immunostaining. The antibodies used included: (1) mouse IgG monoclonal antibodies against vimentin (VIM) and glial fibrillary acidic protein (GFAP); (2) mouse IgM monoclonal antibodies against VIM; and (3) rabbit IgG polyclonal antibodies against GFAP, tyrosine hydroxylase (TH), and 5-hydroxytryptamine (5HT). Series of adjacent vibratome sections were treated: (1) for double immunofluorescence labeling by incubating them with two primary antibodies, combining a rabbit polyclonal antibody and a mouse IgG monoclonal antibody; (2) for triple immunofluorescence labeling by

FIG. 1. Schematic representation of the knife used and of a frontal section through the hypothalamus showing the anatomical locations of the surgical lesions performed at the level of the dorsomedial (1) and the ventromedial (2) hypothalamus. AR, arcuate nucleus; DM, dorsomedial hypothalamus; Fx, fornix; ME, median eminence; OC, optic chiasma; V, third ventricle.

TANYCYTES AND MONOAMINERGIC AXON REGENERATION

TABLE 1

3

RESULTS

Antibodies Used in the Present Study Antigen

Source, type, isotype

Dilution

Source

Reference

Vimentin Vimentin GFAP GFAP 5HT TH

Mouse, monoclonal, IgG Mouse, monoclonal, IgM Mouse, monoclonal, IgG Rabbit, polyclonal, IgG Rabbit, polyclonal, IgG Rabbit, polyclonal, IgG

1:1000 1:200 1:1000 1:1000 1:800 1:3000

Sigma Sigma Sigma Dako Immunotech Jacques Boy Institut

(32) (45) (15) (53) (48) (4)

incubating them with three primary antibodies, combining a rabbit polyclonal antibody, a mouse IgG monoclonal antibody and a mouse IgM monoclonal antibody. They were incubated 48–72 h at 4°C with primary antibodies in PBS containing 0.1% Triton X-100, 1% normal goat serum, and 1% bovine serum albumin. After rinsing in PBS, they were incubated for 2 h at 4°C with two secondary antibodies (for double fluorescence): an anti-rabbit IgG conjugated with Cy3 (Jackson Laboratories, West Grove, PA) and an anti-mouse IgG conjugated with fluorescein (Sigma, St Louis, MO); or three secondary antibodies (for triple immunofluorescence): an anti-rabbit IgG conjugated with Cy3, an anti-mouse IgG conjugated with Cy5 (Jackson Laboratories), and an anti-mouse IgM conjugated with fluorescein (Sigma). The secondary antibodies were diluted in PBS containing 0.1% Triton X-100, 1% normal goat serum, and 1% bovine serum albumin. After careful rinsing, sections were mounted in Mowiol (Calbiochem, La Jolla, CA) and observed under a MRC-600 confocal laser scanning microscope (Bio-Rad) equipped with a krypton/argon-mixed gas laser. Three laser lines emitting at 488, 568, and 645 nm were used for the fluorescein-, Cy3-, and Cy5-conjugated secondary antibodies, respectively, providing minimum overlap of the emission spectra of the three fluorochromes. The background noise of each confocal image was reduced by averaging six image inputs. The organization of the immunofluorescent labelings was studied on reconstructed thick sections made by projecting a series of 20–30 consecutive confocal images 1 µm apart in the z plane, collected through the thickness of the vibratome section. The specificity, origin, and dilution of the antibodies are given in Table 1. Controls consisted of (1) omitting the primary antibodies and applying the secondary antibodies alone, (2) applying each primary antibody sequentially and then adding an inappropriate secondary antibody, and (3) exciting each fluorochromes by the two inappropriate laser lines. This allowed us to confirm that the three secondary antibodies used did not induce artifactual fluorescent labeling and that there was no overlap of the emission spectra of the three fluorochromes.

Organization of Glial Cells in the Intact Hypothalamus The observation of sections double immunostained for VIM and GFAP indicated that each of the two immunostainings was associated with specific cell types. VIM-immunostaining was found to be essentially associated with cells surrounding the lumen of the third ventricle including, (1) ependymocytes, present along the dorsal portion of the ventricular surface, that exhibited a regular cubicoidal shape (Figs. 2 and 3B), and (2) tanycytes, present along the ventral portion of the ventricular surface, the cell bodies of which were very similar to that of ependymocytes, but which exhibited long processes projecting into the periventricular parenchyma (Figs. 2, 3D, and 3F). A transitional area containing VIM-immunoreactive (IR) tanycytes and ependymocytes was observed approximately midway along the lateral wall of the third ventricle. VIM-IR tanycyte processes penetrated the neuropil portion of the mediobasal hypothalamus with different courses according to their location. In the ME, the processes projected perpendicularly to the main axis of the organ toward the outer limit of the external layer. In the ventrolateral arcuate nucleus, the processes arched in the neuropil and reached the ventral surface of the brain at a distance up to 1 mm from the ventricular lumen. The more dorsally located tanycytes generally exhibited shorter processes that projected perpendicularly to the ventricular surface and terminated around blood vessels. In these regions, in contrast with the ME, the tanycyte processes were poorly ramified. GFAP immunostaining was associated with cells showing the typical morphology of astrocytes, i.e., stellate cells with several processes radiating from the cell body. Such GFAP-IR cells were observed throughout the different regions of the hypothalamus in which they appeared to be homogeneously dispersed (Figs. 3A, 3C, and 3E). The careful examination of the sections double immunostained for VIM and GFAP indicated that no colocalization could be detected throughout the different dorsoventral hypothalamic regions (Fig. 3). Organization of the Glial Cells in the Lesioned Hypothalamus The distribution of the glial cells was studied on frontal vibratome sections of lesioned hypothalami double immunostained for GFAP and VIM. In some animals, tissue necrosis forming cavities up to 50 µm wide was associated with the lateral portion of the dorsomedial lesions.

4

CHAUVET, PRIETO, AND ALONSO

FIG. 2. Montage of confocal images of a frontal section through the hypothalamus immunostained for VIM. VIM immunostaining is associated with both (i) ependymocytes represented as cubicoidal cells located along the dorsal portion of the third ventricle surface and (ii) tanycytes represented as periventricular cells exhibiting more or less elongated processes located along the ventralmost periventricular region. V, third ventricle; ME, median eminence.

Dorsomedial hypothalamic lesions. A typical gliosis reaction, characterized by a marked increase in both the number and staining intensity of GFAP-IR structures, was observed along the knife cut (Fig. 4A). These GFAP-IR structures were generally densely aggregated all along the lesion borders forming a typical lesional

scar of about 200–300 µm wide extending on both sides of the surgical cut. Reactive astrocytes exhibiting large cell bodies and numerous extended processes as well as intense GFAP immunostaining were also detected in the area surrounding the lesion. Numerous VIM-IR structures were also detected in close vicinity to the lesion (Fig. 4B). However, their distribution was not as widespread as the GFAP-IR structures, most of them being restricted to the border of the knife cut. The simultaneous visualization of the two immunostainings clearly showed that VIM and GFAP were frequently colocalized within structures closely surrounding the lesion. In contrast, reactive astrocytes intensely GFAP-IR, located at a distance from the lesion (more than 150 µm), were not or only faintly stained by the VIM-antibody. As in the intact hypothalami, all the VIM-IR ependymocytes located along the lateral wall of the third ventricle were GFAP-negative (Figs. 4A and 4B). Ventromedial hypothalamic lesions. The organization of the GFAP-IR structures present in the vicinity of this lesion was very similar to that observed in the dorsomedial lesion (Figs. 4C and 4E). Namely, a typical glial scar formed by the aggregation of astrocytes was observed along the knife cut. Compared with the dorsal lesions, however, VIM-IR structures showed a more complex organization pattern (Figs. 4D and 4F). The observation of double-immunostained sections showed that the region closely surrounding the lesion contained two distinct types of glial structures including (i) cell bodies and their connected processes that were intensely immunostained for both VIM and GFAP and (ii) elongated processes that were VIM-IR but GFAPnegative. Although this latter type of glial structure was observed throughout the different portions of the lesions, they were always much more numerous in those portions of the lesion located close to the third ventricle wall. In some lesioned rats, a bridge joining the two lateral walls of the third ventricle was formed at the level of the lesion. Such bridges were found to contain a majority of processes that were VIM-IR- and GFAP-negative and some cell bodies and processes that were positive for both GFAP and VIM. Postlesional Responses of Hypothalamic Axons The organization of lesioned axons was studied on frontal vibratome sections of lesioned hypothalami double or triple immunostained for: (1) a neuronal marker including either TH, an enzyme present in all catecholaminergic neurons, or 5HT, the transmitter of serotonergic neurones, and (2) one or two glial markers including GFAP and/or VIM. Dorsomedial hypothalamic lesions. For the two types of axons considered here, a large number of highly immunostained fibers was detected both dorsally and ventrally to the surgical cut. Moreover, for the

TANYCYTES AND MONOAMINERGIC AXON REGENERATION

5

FIG. 3. Paired confocal images of sections through the intact hypothalamus double-immunostained for GFAP (A, C, E) and for VIM (B, D, F). Throughout the dorsal (A, B), median (C, D), and ventral (E, F) hypothalamus, GFAP-IR and VIM-IR are associated with distinct anatomical structures. GFAP is associated with astrocyte-like cell bodies and processes present within the three regions. VIM-IR is associated with ependymocytes lining the dorsal ventricular border of the third ventricle (B) and with tanycyte-like cells and their elongated processes in the ventralmost hypothalamic regions (D, F). V, third ventricle; ME, median eminence.

6

CHAUVET, PRIETO, AND ALONSO

FIG. 4. Paired confocal images of sections double immunostained for GFAP (A, C, E) and for VIM (B, D, F) at the level of the dorsomedial (A, B) and ventromedial (C-F) hypothalamic lesions. (A, B) In a lesion through the dorsomedial hypothalamus, most of the GFAP-IR structures surrounding the cut also exhibit VIM-IR, whereas a number of GFAP-IR astrocytes located at distance from the lesion appear to be deprived of VIM-IR (arrows). (C–F) In a lesion through the ventromedial hypothalamus, GFAP-IR and VIM-IR structures are concentrated along the border of the cut and, as in the dorsal lesion, most of the GFAP-IR located at a distance from the lesion are VIM-negative. Moreover, a large number of VIM-IR structures extending long processes throughout the cut and the surrounding area are devoid of GFAP-immunostaining (arrows in E and F). Note that in C and D, numerous GFAP- and VIM-IR structures located at the level of the surgical cut project into the third ventricular lumen. Arrowheads indicate the location of the surgical cut. V, third ventricle.

TANYCYTES AND MONOAMINERGIC AXON REGENERATION

7

FIG. 5. Paired confocal images of sections double-immunostained for GFAP (A, C) and 5HT (B) or TH (D) at the level of the dorsomedial hypothalamic lesions. A dense lesional scar formed by tightly packed GFAP-IR astrocytes is observed along the surgical cut of the dorsal lesion (A, C). 5HT-IR (B) or TH-IR (D) fibers do not cross the lesioned area but rather appear to abut onto or to run parallel to the GFAP-IR scar on both side of the lesion. Arrowheads indicate the location of the surgical cut. V, third ventricle.

two types of axons, the fiber portions located in close vicinity to the surgical cut were always found to exhibit increased intensity of immunostaining. The observation of sections traited for multiple immunostaining clearly indicated that 5HT or TH intensely immunostained axonal fibers rarely penetrated into the lesioned scar consisting of densely packed astrocytes intensely immunoreactive for GFAP and VIM. On the other hand, numerous such fibers were frequently observed either to abut perpendicularly on the border of the scar, or to run parallel to it (Fig. 5).

Ventromedial hypothalamic lesions. In contrast with the dorsal lesions, numerous fibers immunoreactive for 5HT or TH were found to penetrate the lesioned area. In the lesioned rats in which a glial bridge had been formed between the two lateral walls of the ventricle, a large number of 5HT- and TH-IR axons were found to pass through the bridge and to penetrate the contralateral periventricular area. In some cases, the site of the surgical cut was difficult to visualize when observing the immunostained axons only. However, the observation of sections double immunostained for one of the

8

CHAUVET, PRIETO, AND ALONSO

neuronal markers and for either VIM or GFAP generally allowed us to clearly visualize numerous axonal fibers that had entered into and frequently appeared to cross through the glial scar surrounding the surgical cut (Fig. 6). The observation of sections triple immunostained for one of the neuronal markers and for both VIM and GFAP, further indicated that, throughout the lesion scar and particularly in its portions located close to the wall of the third ventricle, such regenerating axonal fibers were preferentially associated with VIM-IR, GFAP-negative profiles. Moreover, all along the lesioned area, 5HT-IR and TH-IR axonal fibers always appeared to be preferentially concentrated in those portions of the glial scar that exclusively contained VIM-IR- and GFAP-negative structures (Fig. 7). In contrast, throughout the scar, only scarce 5HT- or TH-IR fibers were detected in those portions showing a dense accumulation of GFAP-IR profiles (Fig. 7F).

adrenergic (50) axonal fibers arising from neurons located in mesencephalic or medullary nuclei and dopaminergic axons arising from dopaminergic cell bodies located in the dorsal and ventral hypothalamic regions (30). Since TH-IR-transected fibers were detected both dorsally and ventrally to the dorsal and ventral surgical cut, it is likely that catecholaminergic axons of various origins were transected by the two types of lesions performed here. Taken together, these data indicate that no clear discrimination can be made with regards to the cellular origin of the serotonergic and catecholaminergic fibers lesioned by the dorsal and the ventral surgical cuts performed here. It is thus concluded that the differential postlesional responses observed here cannot be related to properties intrinsic to the lesioned neurons. An alternative explanation is, therefore, that such differences mainly result from the differential organization of the glial scars that are formed along the dorsal and ventral surgical cuts.

DISCUSSION

Differential Postlesional Responses of Transected Axons Among the large variety of neuronal systems that innervate the periventricular hypothalamus, axons immunoreactive for 5HT or TH represent the most prominent input yet identified. The present data indicate that both the dorsal and ventral surgical cuts performed here lesioned a large number of axons of each type and that their postlesional responses markedly differed depending on the site of lesion: only scarce axons were found to grow into the glial scar formed around the dorsal lesion, whereas numerous axons were found to penetrate into the ventral lesion. One possible explanation for such differential postlesional responses is that the two types of surgical cuts lesioned distinct neuronal systems exhibiting different capacities for regeneration. Although some cells containing 5HT have been reported in the dorsomedial hypothalamus of the rat (19), it is generally admitted that, like the rest of the brain, the large majority of 5HT-IR fibers innervating the periventricular hypothalamus arise from serotonergic neurons located in the raphe nuclei (46, 47). It can thus be assumed that the 5HT fibers transected by the dorsal or ventral surgical cuts mostly arise from the same mesencephalic neuronal system. TH-IR axons innervating the periventricular hypothalamus arise from several intrahypothalamic and extrahypothalamic neuronal systems (11, 26, 52). TH, the enzyme that converts tyrosine into DOPA, is present in all the central catecholaminergic neurons, including dopaminergic, noradrenergic, and adrenergic neurons. The TH-IR fibers in the periventricular hypothalamus thus comprise noradrenergic (37, 38) and

Differential Organization of the Dorsomedial and Ventromedial Glial Scars The three main components of the glial scars are astrocytes, oligodendrocytes, and activated microglia. Previous studies of Schwab and colleagues indicate that oligodendrocytes and CNS myelin of higher vertebrates contain potent membrane-bound inhibitors of neurite growth (9, 10, 36, 41). However, the hypothalamus is a brain region that contains only scarce oligodendrocytes and myelinated axons (3). Microglia cells also participate actively in the CNS tissue reaction (20, 21, 28), and their activation depends on the type of insult (17). However, it is unlikely that marked differences exist between the activated microglia cells present in the vicinity of the dorsal and ventral lesions performed here. The present data indicate that similar astrocytic reactions were associated with the surgical lesions placed in the dorsomedial or the ventromedial hypothalamus. They were generally characterized by an increased density of the GFAP-IR astrocytes exhibiting both hypertrophy and a marked increase in their GFAP-immunoreactivity and by the temporary expression of VIM in those astrocytes located close to the lesion (7, 39, 51). However, there is increasing evidence that astrocytes are a highly heterogeneous class of cell (24, 25). It is thus possible that different astrocyte types are present in the dorsomedial and ventromedial hypothalamus. In a previous study, we described the existence of several differences between the glial scars formed around surgical lesions placed in dorsolateral or mediobasal hypothalamus regions (3). In dorsolateral scars, astrocytes are strongly immunoreactive to embryonic NCAM and intense laminin immunoreactivity is observed over large patches included in the scar. More-

TANYCYTES AND MONOAMINERGIC AXON REGENERATION

9

FIG. 6. Paired confocal images of sections double-immunostained for VIM (A, C) or GFAP (E) and 5HT (B, D, F) at the level of the ventromedial hypothalamus lesions. (A, B) Numerous 5HT-IR fibers are detected throughout the lesional area containing numerous closely intermingled VIM-IR structures and appear to cross the glial scar. (C, D) At higher magnification 5HT-IR fibers appear continuously distributed through the lesional scar containing both VIM-IR cell bodies and numerous VIM-IR elongated processes. (E, F) The 5HT-IR axons appear to cross the dense GFAP-IR astroglial scar surrounding the lesion. In B, note that a large number of 5HT-IR axons pass through the bridge formed between the two lateral walls of the third ventricle. Arrowheads indicate the location of the surgical cut. V, third ventricle.

10

CHAUVET, PRIETO, AND ALONSO

FIG. 7. Colored confocal images of a section triple immunostained for VIM (green, A), GFAP (blue, B), and TH (red, C) at the level of a ventromedial hypothalamic lesion. Paired superpositions of these images clearly show that (1) in addition to structures immunoreactive for both GFAP and VIM, the lesional scar contains a number of structures immunoreactive for VIM (green) but GFAP-negative (arrows in D), and (2) TH-IR fibers mostly cross the lesional scar in close apposition with VIM-IR structures (Yellow, E), in portions of the scar that are devoid of GFAP-IR structures (arrows in F). Small arrowheads in C pointed to TH-IR cell bodies located in the arcuate nucleus. Arrowhead in D indicates the location of the surgical cut.

TANYCYTES AND MONOAMINERGIC AXON REGENERATION

over, these astrocytes appear to be tightly packed with the occurrence of extended gap junctions between their limiting plasma membrane. In mediobasal hypothalamic lesion, in contrast, the scar contains only slight immunostaining to embryonic NCAM and laminin and gap junctions are less frequent and involve shorter portions of adjacent membranes. Our data further indicate that additional differences exist concerning the glial cell types included within the scars formed around the dorsomedial and ventromedial hypothalamic lesions. The double-immunostaining experiments indeed clearly showed that the ventromedial lesions always contained VIM-IR- and GFAP-negative processes that can be identified as tanycyte processes. In the intact hypothalamus, apart from the ependymocytes lining the superior part of the third ventricle, structures immunoreactive to VIM always exhibited the anatomical organization of tanycytes. Tanycytes are specific ependymal cells with basal processes which radiate into the underlying neuropil (27). Although they have been described in different regions of the ventricular system, these cells are particularly abundant in the mediobasal hypothalamus and the median eminence (6, 18) but are completely absent in the dorsal hypothalamus. Although tanycytes express both VIM and GFAP (5, 14, 35), their content in GFAP is far lower than that of astrocytes (12, 13, 29). As previously reported (12, 13), under the immunocytochemical conditions used in this study, tanycytes thus generally appeared VIM-IR and GFAP-negative. The appearance of a perilesional area devoid of GFAP-IR astrocytes but containing VIM-IR cells has been previously described (17, 31, 49). However such lesional areas were found to be repopulated by GFAP-IR reactive astrocytes as soon as 1 to 2 weeks after the lesion, i.e., at postlesional delays far shorter than that used in the present study. Although VIM immunostaining was also associated with reactive astrocytes included in the ventromedial lesional scar, tanycytic processes could be easily distinguished by their lack of immunostaining for GFAP and by their typical elongated shape. Moreover, our data indicate that axonal sprouting was essentially found in those parts of the lesion where such elongated VIM-IR tanycyte processes were present. In contrast, as previously described in the median eminence (12), regenerating axonal sprouts were rarely located in that part of the ventromedial lesion containing a high density of GFAP-immunoreactive astrocytic structures. This strongly suggests that, as in other brain regions, reactive astrocytes located in the intrahypothalamic lesional scars do impede axonal regeneration and that tanycyte processes present in the ventralmost lesions are able to support the regeneration of various axon types. Clearly, such an interpretation fully fits our previous observations showing that (i) neuropeptide-Y contain-

11

ing neurons of the arcuate nucleus regenerate axons through the astroglial scar produced by the mediobasal hypothalamic lesion (2), (ii) tanycytes of the median eminence constitute a preferential support for the regeneration of lesioned neurohypohysial oxytocinergic and vasopressinergic axons (12), and (iii) cultured tanycytes have the capacity to support survival and neurite outgrowth of a large variety of cocultured postnatal CNS neurons (13). In conclusion, the data in the present study provide additional support to the idea that tanycytes located in the adult rat mediobasal hypothalamus play a major role in the capacity for regeneration of various axons types innervating this region. One possibility that has already been proposed is that tanycytes produce specific molecules that exert a direct beneficial effect on the survival and axonal outgrowth of injured neurons (13). Another nonexclusive possibility is, however, that tanycytes support indirectly the regeneration of lesioned axons by influencing the organization of the reactive astrocytes constituting the glial scar. Interestingly enough, such a role is now admitted for immature astrocytes (43, 44) that present a striking number of morphological and biochemical similarities with adult tanycytes. In this line, it would be interesting to examine the levels of inhibitory cell surface or extracellular matrix molecules in the two different lesions performed here, in order to determine if the nonpermissive nature of reactive astrocytes could have been influenced by tanycytes. In vitro and in vivo investigations are now in progress to determine whether tanycytes are able to influence reactive astrocytes of various CNS regions and as such to sustain the postlesional regeneration of mature neurons in extrahypothalamic CNS regions. ACKNOWLEDGMENTS This work was supported by IRME and Inversiones Cavdeca, C.A.

REFERENCES 1.

Aguayo, A. J. 1985. Axonal regeneration from injured neurons in the adult mammalian central nervous system. In Synaptic Plasticity (C. W. Cotman, Ed.), pp. 457–484. Guilford Press, New York. 2. Alonso, G., and A. Privat. 1993. Neuropeptide Y-producing neurons of the arcuate nucleus regenerate axons after surgical deafferentation of the mediobasal hypothalamus. J. Neurosci. Res. 34: 510–522. 3. Alonso, G., and A. Privat. 1993. Reactive astrocytes involved in the formation of lesional scars differ in the mediobasal hypothalamus and in other forebrain regions. J. Neurosci. Res. 34: 523–538. 4. Arluison, M., M. Dietl, and J. Thibault. 1984. Ultrastructural morphology of dopaminergic synapses in the striatum of the rat using tyrosine hydroxylase immunocytochemistry: A topographical study. Brain Res. Bull. 13: 269–285.

12 5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16. 17.

18. 19.

20.

21.

22.

23.

24.

CHAUVET, PRIETO, AND ALONSO Basco, E., P. L. Woodhams, F. Hajos, and R. Balazs. 1981. Immunocytochemical demonstration of glial fibrillary acidic protein in mouse tanycytes. Anat. Embryol. 162: 217–222. Bruni, J. E., M. R. Del Bigio, and R. E. Clattenburg. 1985. Ependyma: Normal and pathological. A review of the literature. Brain Res. 356: 1–19. Calvo, J. L., A. L. Carbonell, and J. Boya. 1991. Co-expression of glial fibrillary acidic protein and vimentin in reactive astrocytes following brain injury in rats. Brain Res. 566: 333–336. Carbonetto, S. 1991. Facilitatory and inhibitory effects of glial cells and extracellular matrix in axonal regeneration. Curr. Opin. Neurobiol. 1: 407–413. Caroni, P., and M. E. Schwab. 1989. Codistribution of neurite growth inhibitors and oligodendrocytes in rat CNS: appearance follows nerve fiber growth and precedes myelination. Dev. Biol. 136: 287–295. Caroni, P., and M. E. Schwab. 1988. Two membrane protein fractions from rat central nervous myelin with inhibitory properties for neurite outgrowth and fibroblast spreading. J. Cell Biol. 106: 1281–1288. Chan-Palay, V., L. Zaborszky, C. Ko¨hler, M. Goldstein, and S. L. Palay. 1984. Distribution of tyrosine-hydroxylase-immunoreactive neurons in the hypothalamus of rats. J. Comp. Neurol. 227: 467–496. Chauvet, N., M. L. Parmentier, and G. Alonso. 1995. Transected axons of adult hypothalamo-neurohypophysial neurons regenerate along tanycytic processes. J. Neurosci. Res. 41: 129–144. Chauvet, N., A. Privat, and G. Alonso. 1996. Aged median eminence glial cell cultures promote survival and neurite outgrowth of cocultured neurons. Glia 18: 211–223. De Vitry, F., R. Picart, C. Jacquet, and A. Tixier-Vidal. 1981. Glial fibrillary acidic protein. A cellular marker of tanycytes in the mouse hypothalamus. Dev. Neurosci. 4: 457–460. Debus, E., K. Weber, and M. Osborn. 1983. Monoclonal antibodies specific for glial fibrillary acidic (GFA) protein and for each of the neurofilament triplet polypeptides. Differentiation 25: 192– 203. Dellmann, H. D. 1973. Degeneration and regeneration of neurosecretory systems. Int. Rev. Cytol. 36: 215–315. Fernaud-Espinosa, I., M. Nieto-Sampedro, and P. Bovolenta. 1993. Differential activation of microglia and astrocytes in anisomorphic and isomorphic gliotic tissue. Glia 8: 277–291. Flament-Durand, J., and J. P. Brion. 1985. Tanycytes: Morphology and functions: A review. Int. Rev. Cytol. 96: 121–155. Frankfurt, M., J. M. Lauder, and E. C. Azmitia. 1981. The immunocytochemical localization of serotonergic neurons in the rat hypothalamus. Neurosci. Lett. 24: 227–232. Giulian, D., J. Chen, J. E. Ingeman, J. K. George, and M. Noponen. 1989. The role of mononuclear phagocytes in wound healing after traumatic injury to adult mammalian brain. J. Neurosci. 9: 4416–4429. Giulian, D., K. Vaca, and M. Corpuz. 1993. Brain glia release factors with opposing actions upon neuronal survival. J. Neurosci. 13: 29–37. Graziadei, P. P. C., and G. A. Monti Graziadei. 1978. The olfactory system: a model for the study of neurogenesis and axon regeneration in mammals. In Neuronal Plasticity (C. W. Cotman, Ed.), pp. 131–153. Raven Press, New York. Halasz, B., and L. Pupp. 1965. Hormone secretion of the anterior pituitary gland after physical interruption of all nervous pathways to the hypophysiotropic area. Endocrinology 77: 553–562. Hatten, M. E., R. K. H. Liem, M. L. Shelanski, and C. A. Mason. 1991. Astroglia in CNS injury. Glia 4: 233–243.

25. 26.

27. 28. 29.

30.

31.

32. 33. 34.

35.

36.

37.

38.

39.

40. 41.

42. 43.

44.

Ho¨ke, A., and J. Silver. 1994. Heterogeneity among astrocytes in reactive gliosis. Perspect. Dev. Neurobiol. 2: 269–274. Ho¨kfelt, T., O. Johansson, K. Fuxe, M. Goldstein, and D. Park. 1976. Immunohistochemical studies on the localization and distribution of monoamine neuron systems in the rat brain. I. Tyrosine hydroxylase in the mes- and diencephalon. Med. Biol. 54: 427–453. Horstmann, E. 1954. Die faserglia des selachiergehirns. Z. Zellforrsch. Microsk. Anat. 39: 588–617. Kreutzberg, G. W. 1996. Microglia: a sensor for pathological events in the CNS. Trends Neurosci. 19: 312–318. Meister, B., T. Ho¨kfelt, Y. Tsuruo, H. Hemmings, C. Ouimet, P. Greengard, and M. Goldstein. 1988. DARPP-32, a dopamineand cyclic AMP-regulated phosphoprotein in tanycytes of the mediobasal hypothalamus: distribution and relation to dopamine and luteinizing hormone-releasing hormone neurons and other glial elements. Neuroscience 27: 607–622. Moore, K. E., and K. J. Lookingland. 1995. Dopaminergic neuronal systems in the hypothalamus. In Psychopharmacology: The Fourth Generation of Progress (F. E. Bloom and D. J. Kupfer, Eds.), pp. 245–256. Raven Press, New York. Ogawa, M., M. Araki, I. Nagatsu, and M. Yoshida. 1989. Astroglial cell alteration caused by neurotoxins: Immunohistochemical observations with antibodies to glial fibrillary acidic protein, laminin, and tyrosine hydroxylase. Exp. Neurol. 106: 187–196. Osborn, M., E. Debus, and K. Weber. 1984. Monoclonal antibodies specific for vimentin. Eur. J. Cell Biol. 34: 137–143. Paxinos, G., and C. Watson. 1982. The Rat Brain in Stereotaxic Coordinates. Academic Press, London. Reier, P. J., L. J. Stensaas, and L. Guth. 1983. The astrocytic scar as an impediment to regeneration in the central nervous system. In Spinal Cord Reconstruction (C. C. Kao, R. P. Bunge, and P. J. Reier, Eds.), pp. 163–196. Raven Press, New York. Roessmann, U., M. E. Velasco, S. D. Sindley, and P. Gambetti. 1980. Glial fibrillary acidic protein (GFAP) in ependymal cells during development: An immunohistochemical study. Brain Res. 200: 13–21. Rubin, B. P., I. Dusart, and M. E. Schwab. 1994. A monoclonal antibody (in-1) which neutralizes neurite growth inhibitory proteins in the rat CNS recognizes antigens localized in CNS myelin. J. Neurocytol. 23: 209–217. Sawchenko, P. E., and L. W. Swanson. 1981. Central noradrenergic pathways for the integration of hypothalamic neuroendocrine and autonomic responses. Science 214: 685–687. Sawchenko, P. E., and L. W. Swanson. 1982. The organization of noradrenergic pathways from the brainstem to the paraventricular and supraoptic nuclei. Brain Res. Rev. 4: 275–322. Schiffer, D., M. T. Giordana, A. Migheli, G. Giaccone, S. Pezzotta, and A. Mauro. 1986. Glial fibrillary acidic protein and vimentin in the experimental glial reaction of the rat brain. Brain Res. 374: 110–118. Schwab, M. E. 1996. Molecules inhibiting neurite growth: A minireview. Neurochem. Res. 21: 755–761. Schwab, M. E., and P. Caroni. 1988. Oligodendrocytes and CNS myelin are nonpermissive substrates for neurite growth and fibroblast spreading in vitro. J. Neurosci. 8: 2381–2393. Silver, J. 1994. Inhibitory molecules in development and regeneration. J. Neurol. 242: S22–S24. Smith, G. M., R. H. Miller, and J. Silver. 1986. Changing role of forebrain astrocytes during development, regenerative failure, and induced regeneration upon transplantation. J. Comp. Neurol. 251: 23–43. Smith, G. M., and J. Silver. 1988. Transplantation of immature

TANYCYTES AND MONOAMINERGIC AXON REGENERATION and mature astrocytes and their effect on scar formation in the lesioned central nervous system. Prog. Brain Res. 78: 353–361. 45. Stathopoulos, E., G. S. Ne´ve, C. R. Taylor, and A. L. Epstein. 1989. LN-6: A monoclonal antibody to vimentin expressed in non-hematopoietic mesenchymal cells and derived tumors and reactive in B5-fixed, paraffin-embedded tissues. J. Histochem. Cytochem. 37: 1363–1370. 46. Steinbusch, H. W. M. 1981. Distribution of serotonin-immunoreactivity in the central nervous system of the rat. Cell bodies and terminals. Neuroscience 6: 557–618. 47. Steinbusch, H. W. M., and R. Nieuwenhuys. 1981. Localization of serotonin-like immunoreactivity in the central nervous system and pituitary of the rat, with special references to the innervation of the hypothalamus. In Serotonin: Current Aspects of Neurochemistry and Function (B. Haber, S. Gabay, M. R. Issidorides, and S. G. A. Alivesatos, Eds.), pp. 7–36. Adv. Exp. Med. Biol., Vol. 133. Plenum, New York. 48. Steinbusch, H. W. M., A. A. J. Verhofstad, and H. W. J. Joosten. 1978. Localization of serotonin in the central nervous system by immunohistochemistry. Neuroscience 3: 811–819.

49.

50.

51.

52.

53.

13

Stichel, C. C., and H. W. Muller. 1994. Relationship between injury-induced astrogliosis, laminin expression and axonal sprouting in the adult rat brain. J. Neurocytol. 23: 615–630. Swanson, L. W., and B. K. Hartman. 1975. The central adrenergic system. An immunofluorescence study of the location of cell bodies and their efferent connections in the rat utilizing dopamine-B-hydroxylase as a marker. J. Comp. Neurol. 163: 467– 506. Takamiya, Y., S. Kohsaka, S. Toya, M. Otani, and Y. Tsukada. 1988. Immunohistochemical studies on the proliferation of reactive astrocytes and the expression of cytoskeletal proteins following brain injury in rats. Dev. Brain Res. 38: 201–210. van den Pol, A. N., R. S. Herbst, and J. F. Powell. 1984. Tyrosine hydroxylase-immunoreactive neurons of the hypothalamus: A light and electron microscopic study. Neuroscience 13: 1117– 1156. Viale, G., M. Gambacorta, G. Coggi, P. Dell’Orto, M. Milani, and C. Doglioni. 1991. Glial Fibrillary acidic protein immunoreactivity in normal and diseaded human breast. Virchows Arch. A Pathol. Anat. Histopathol. 418: 339–348.