Brain Research 958 (2002) 488–496www.bres-interactive.com

Interactive report

E strogen regulates GFAP-expression in specific subnuclei of thefemale rat interpeduncular nucleus: a potential role for estrogen

receptorba b c a ,*Attila Zsarnovszky , Todd Smith , Ferenc Hajos , Scott M. Belcher

aDepartment of Pharmacology and Cell Biophysics, University of Cincinnati College of Medicine, 231 Albert Sabin Way, PO Box 670575,Cincinnati, OH 45267-0575,USA

bArkansas Cancer Research Center–Partners in Research Program, Little Rock AR, 75225,USAcDepartment of Anatomy and Histology, Faculty of Veterinary Sciences, Szent Istvan University, Budapest, Hungary

Accepted 7 October 2002

Abstract

We previously demonstrated that in rat, astrocytic glial fibrillary acidic protein- (GFAP) expression in the interpeduncular nucleus(IPN) was responsive to testosterone and in females the intensity of GFAP-immunoreactivity (IR) followed the periodic hormonalchanges of the estrous cycle. The aim of this study was to test whether 17b-estradiol (E ), in the absence of other ovarian hormones, can2

influence GFAP-expression within individual subnuclei of the IPN and to determine the cellular distribution of estrogen receptorb (ERb)in the IPN. Quantitative surface-density analysis was used to compare the intensity of GFAP-IR at different anterio-posterior (AP) levelsof the IPN in ovariectomized female rats 24 h after treatment with E or vehicle. Estrogen-treatment resulted in a significant increase in2

GFAP-IR in the rostrolateral subnucleus of the IPN at AP:25.60, in the lateral-, dorsolateral-, dorsomedial- and central subnuclei at26.04 and in the lateral subnucleus at26.72. No significant differences were observed at25.80 and26.30. These results indicate thatE , in the absence of other ovarian hormones, modulates GFAP-expression within select IPN subnuclei and that these affects are2

dependent on position along the AP axis. To determine whether ERb was a possible mediator of the observed estrogenic effects, adjacentsection pairs of the IPN were immunostained for ERb or GFAP. Using the ‘mirror’ method, ERb-IR was detected in the cytoplasm ofGFAP-immunopositive astroglia and in the nuclei of GFAP-immunonegative neurons. These findings suggest that in the IPN, E may2

directly modulate GFAP-expression through ERb-mediated mechanisms. 2002 Elsevier Science B.V. All rights reserved.

Theme: Endocrine and autonomic regulation

Topic: Neuroendocrine regulation: other

Keywords: Astrocyte; Estrogen; Estrogen receptor beta; Glia, glial fibrillary acidic protein; Immunohistochemistry; Interpeduncular nucleus

1 . Introduction lordosis behavior [27,40] and functions associated withhippocampal and subcortical theta-type EEG activity

The anatomical connectivity of the interpeduncular [20,46,47]. It is well established that steroid hormonesnucleus (IPN) with other limbic regions and other parts of regulate the physiological and morphological properties ofthe central nervous system is well described [1,8,14]. As a neuroendocrine brain regions, however, accumulating datalimbic relay center, the IPN is involved in the regulation of indicate that these effects are not restricted to neuroen-a number of important physiological processes that include docrine brain centers. For example, a number of previous

studies have shown that the intensity of astrocytic GFAP-immunoreactivity (IR) displays a parallelism with changes

*Corresponding author. Tel.:11-513-558-1721; fax:11-513-558-in plasma levels of gonadal hormones in the hypothalamus1169.[11,12,28], the hippocampus [10,45] and the IPN [15,16].E-mail addresses: http: / /www.med.uc.edu/pharmacology/ faculty /

sm Belcher.ucm(S.M. Belcher),scott.belcher@uc.edu(S.M. Belcher). In the IPN of male rats, plastic changes in astroglial]

0006-8993/02/$ – see front matter 2002 Elsevier Science B.V. All rights reserved.PI I : S0006-8993( 02 )03771-X

489A. Zsarnovszky et al. / Brain Research 958 (2002) 488–496

GFAP-IR were evoked by experimental manipulation of changes (23mg estradiol-benzoate/100 g body weight),plasma testosterone levels [15], and in the IPN of females while the remaining four received an equal volume in-the intensity of GFAP-immunostaining varied during the jection of sesame oil vehicle. At 24 h after treatment theestrous cycle [16]. These findings indicate that similar to animals were anesthetized with 200 mg/kg ketamine/6.6other brain regions, steroid hormones regulate astroglial mg/kg xylazin i.m. and then transcardially perfused withGFAP-expression in the IPN. Because some data suggest physiological saline followed by 4% paraformaldehyde/that progesterone may also influence the function of the 3% acrolein in 0.1 M phosphate-buffer (PB). FollowingIPN [27,43], the question of whether changes in GFAP-IR perfusion brains were dissected and postfixed for 2 h in 2%during the estrous cycle can be attributed to changes in paraformaldehyde at 48C.circulating E concentrations remains unanswered.2

The first aim of this study is to characterize the effects 2 .2. Tissue preparation and immunohistochemistryof E on GFAP-expression within the IPN of ovariectom-2

ized (ovx) rats and in ovx rats following a single subcuta- Vibratome sections, 60mm thick, were cut and rinsedneous injection of E (ovx/E ). Because the IPN is a 3310 min in PB. To block unbound aldehydes, sections2 2

mosaic of subnuclei with well-defined neural connectivity, were incubated for 20 min in 1% sodium-borohydride,we investigated whether individual subnuclei are differen- followed by 637-min washes in PB. Washed sections weretially responsive to E . Densitometric measurements of incubated overnight at room temperature with mouse-anti-2

GFAP-IR were used to show that in specific subnuclei GFAP monoclonal antibody G-A-5 (6.4mg/ml; Sigma, St.GFAP is expressed at significantly higher levels in E - Louis, MO). To remove unbound antibodies, sections were2

treated animals as compared to controls. washed 3310 min in PB, and washed sections were thenThe mechanism through which E increases GFAP- incubated at room temperature for 2 h with biotinylated2

expression in the IPN is presently unknown. Previous horse anti-mouse IgG antisera (diluted 1:250 in PB; Vectorstudies have not detected estrogen receptora (ERa) Laboratories, Burlingame, CA). Sections were then washedexpression in the IPN; however, ERb mRNA is expressed 3310 min in PB and bound antibodies were visualizedin the IPN at moderate levels [30,42]. The presence of with nickel-intensified diamino-benzidine by the avidin–ERb in the IPN suggests that the effects of E on GFAP- biotin peroxidase complex method following standard2

expression in astrocytes of the IPN could be mediated protocols (Vectastain ABC Kit; Vector Laboratories, Burl-through ERb. The cellular localization of ERb in the IPN ingame, CA).however, has not been established, and it is therefore For the co-localization of GFAP and ERb, adjacentunknown whether this receptor is expressed in astrocytes sections of the IPN from five normal cycling female ratsof the adult IPN. Therefore, immunohistochemical tech- were grouped in pairs and were immunolabeled for eitherniques were used to determine the cellular localization of GFAP or ERb. The tissue processing and immunohisto-ERb protein in the IPN. The results of those experiments chemistry were performed as described above with theindicate that in many IPN astrocytes, ERb- and GFAP-IR exception that ERb-IR was detected with rabbit anti-ERbwere co-localized, and that ERb-IR was localized to the antisera Z8P (2.5mg/ml; Zymed Laboratories, San Fran-nucleus of numerous GFAP-immunonegative neurons. cisco, CA) and biotinylated goat anti-rabbit secondaryThese results suggest that E may influence GFAP-expres- antisera (Vector Laboratories, Burlingame, CA). Specificity2

sion in astroglia of the IPN through ERb-mediated mecha- of this antiserum for ERb has been demonstrated bynisms and that E may have additional, neuron-specific Western blot analysis [41] and independently confirmed in2

actions in the IPN. our laboratory (unpublished observation). In negativecontrol experiments, omission of the primary antibodiesresulted in no specific immunostaining for either GFAP or

2 . Materials and methods ERb.

2 .1. Animals 2 .3. Densitometric measurements and data analysis

All animal procedures were done in accordance with Five different AP levels (AP:25.6,25.8,26.04,26.30approved Institutional Animal Care and Use Committee and26.72 according to the coordinates by Paxinos andprotocols and followed NIH guidelines. A total of ten Watson [36]) of the IPN were examined from each animal.normal cycling female Sprague–Dawley rats (230–250 g) Focusing the microscope to the upper surface of thewere ovariectomized and maintained under standard lab- sections, digital images were captured from the surface oforatory conditions (12-h dark–light cycles, standard rat each section. To ensure the reliability of densitometricchow and tap water ad libitum). Two weeks post-surgery measurements, all images from a given anterioposteriorsix of these animals received a single subcutaneous level were captured during the same microscopy sessioninjection of estradiol-benzoate in sesame oil at a dose using identical microscope- and computer settings. Therelevant to examine estrogen-induced neuromorphological anatomically corresponding subnuclei of the IPN were

A. Zsarnovszky et al. / Brain Research 958 (2002) 488–496490

identified as follows: rostral (R), rostrolateral (RL), dor-somedial (DM), central (C), dorsolateral (DL), apical (A)and intermediate (I) subnuclei. Mean pixel density per unitarea within each subnuclei was determined using Kodak1D Image Analysis Software (version 3.5). Areas occupiedby vessels were excluded from all measurements. Meanpixel density values from three sections representing thesame AP level from each animal were determined; theresulting mean values were used as representative of thatanimal. A paired Student’st-test was used to assess thelevel of significance between control and treatment groupvalues obtained from the same anatomical subnuclei andthe same AP level (Prism v3.01; GraphPad). An inves-tigator blinded to experimental treatment performed alldata collection and analysis.

2 .4. Co-localization of ERb with GFAP

The ‘mirror’ technique [51] was applied to unambigu-ously determine the cellular co-localization of ERb andGFAP in cells of the IPN. Adjacent 60-mm sections werecut and arranged in pairs; one section of each pair wasimmunostained for GFAP while the other section wasstained for ERb as described above. Stained sections werethen mounted with their matching surfaces oriented up-wards, dehydrated through increasing concentrations ofethanol and coverslipped. Digital images of the uppersurface of each section were captured and the image pairsoverlaid using Adobe Photoshop software (version 5.5).Cells that were cut in half were identified and examinedfor the presence of GFAP-IR, ERb-IR or both.

3 . Results

Consistent with previous descriptions [16], GFAP-IRwas observed in the cell bodies and processes of astrocytesin the IPN. The distribution and density of the astroglia (asidentified by the presence of perivascular GFAP-IR and

Fig. 1. Representative photomicrographs of GFAP-immunoreactivitycharacteristic nuclear morphology) were similar in control (GFAP-IR) in the interpeduncular nucleus (IPN) of (A) ovariectomizedand E -treated groups. (OVX) and (B) ovariectomized and estrogen-treated (OVX1E ) rats at2 2

As compared to control, a general increase in GFAP-IR AP 25.6. Increased intensity in GFAP-IR was found in all the subnucleiof the IPN. (C) Densitometric analysis revealed that the GFAP-expressionwas observed in the E -treated group. The comparison of2was significantly higher in only the rostrolateral subnucleus (RL). Valuesindividual subnuclei of the IPN revealed that significantare expressed as mean pixel density (6S.E.M.). For OVX,n54 and for

changes in the intensity of GFAP-IR were confined to OVX1E , n56. *Value is significantly different from the anatomically2specific subnuclei. The rostral portion of the rostrolateral matching subnucleus of the control (P,0.05). AP, anterioposterior; C,subnucleus displayed a robust increase in GFAP-expres-central; DM, dorsomedial; R, rostral. Scale bar represents 200mm.

sion in response to E -treatment (Fig. 1). However, in2

more caudal regions this difference was statistically in-significant. At AP level25.80 no significant differences detected. In the rostral and intermediate subnuclei abetween the E -treated and control groups were detected in reproducible, but statistically insignificant increase was2

any subnuclei. At26.04, a significantly higher level of observed. At26.30 no significant differences in GFAP-IRGFAP-expression in the lateral-, dorsolateral-, dorsome- were detected, whereas, in the lateral subnucleus at26.72dial- and central subnuclei of the E -treated group was (the subnucleus that extends to the caudal-most part of the2

observed (Fig. 2); however, in the rostral aspect of these IPN) the most pronounced increase in GFAP-immuno-subnuclei no significant changes in GFAP-expression were staining was detected (Fig. 3).

491A. Zsarnovszky et al. / Brain Research 958 (2002) 488–496

Fig. 2. Representative photomicrographs of GFAP-IR in the IPN of (A)OVX and (B) OVX1E rats at AP26.04. (C) Densitometric analysis2

revealed that GFAP-expression was significantly increased in the lateral,dorsolateral, dorsomedial and central subnuclei. Values are expressed asmean pixel density (6S.E.M.). For OVX,n54 and for OVX1E , n56.2 Fig. 3. Representative photomicrographs of GFAP-IR in the IPN of (A)*Value is significantly different from the anatomically matching subnu-

OVX and (B) OVX1E rats at AP26.72. (C) Densitometric analysis2cleus of the control (P,0.05). AP, anterioposterior; C, central; DL,revealed that GFAP-expression was significantly increased in the lateraldortsolateral; DM, dorsomedial; I, intermediate; L, lateral; R, rostral.(L) subnucleus. Values are expressed as mean pixel density (6S.E.M.).Scale bar represents 200mm.For OVX, n54 and for OVX1E , n56. *Value is significantly different2

from the anatomically matching subnucleus of the control (P,0.05). A,apical; AP, anterioposterior; C, central; DM, dorsomedial; I, intermediate.Scale bar represents 200mm.

Immunohistochemical analysis of ERb-expression in theIPN revealed that ERb-IR was localized to the cytoplasmor the nucleus of GFAP-immunopositive astrocytes and tothe nucleus of GFAP-immunonegative neuron-like cells portions of glial processes (Fig. 4), whereas glial sheaths(Fig. 4). In astrocytes, cytoplasmic ERb-IR was localized or pods that were closely associated with neurons mostprimarily to perinuclear cytoplasm and the proximal frequently lacked ERb-IR (Fig. 4, black arrows). In lateral

A. Zsarnovszky et al. / Brain Research 958 (2002) 488–496492

Fig. 4. Co-localization of GFAP- and ERb-immunoreactivity. Representative photomicrographs of GFAP- (A) and ERb- (B) immunostainedinterpeduncular nuclei at AP25.8. Nuclear ERb-IR was frequently found in the central part of the IPN (white-framed inset, black arrowheads), whereascytoplasmic ERb-IR was more frequent in the lateral part of the nucleus (black-framed inset, white arrowheads). Scale bar on upper panel represents 100mm. Paired adjacent sections were immunostained for either GFAP (C) or ERb (D) and digital images were overlaid. Cells that were vibratome-cut intotwo halves were identified and the co-localization of GFAP with ERb was determined. White arrows point to cells co-expressing GFAP and ERb. Blackarrows point to a cell expressing nuclear ERb and surrounded by ERb-negative but GFAP-positive astroglial processes. Black arrowheads show a cell thatis ERb-immunonegative that is in close proximity to a glial sheath immunopositive for both ERb and GFAP. Scale bar on lower panel represents 20mm.C, central; DM, dorsomedial; R, rostral; RL, rostrolateral.

regions of the IPN, however, some of the GFAP- and 4 . DiscussionERb-immunopositive glial sheaths were observed associ-ated with the neuronal somata of these GFAP- and ERb- In previous studies, Hajos et al. demonstrated thatimmunonegative cells (Fig. 5). While numerous GFAP- testosterone increased GFAP-expression in the IPN ofimmunonegative cells (most probably neurons) were found male rats [15] and in females the different phases of theto express ERb, the majority of those cells were also estrus cycle influenced the pattern of GFAP-immunostain-ERb-immunonegative. The neuron-like cells that did ex- ing in the IPN [16]. From those results it was concludedpress ERb were concentrated within central regions and that elevated plasma E -levels were associated with in-2

more caudally in the dorsomedial- and apical subnuclei creased expression of GFAP within the IPN. However,(Fig. 6). In these neurons, the cellular pattern of ERb- since progesterone can also influence IPN-related functionsimmunostaining was consistent with ERb localized pri- [27], it was necessary to determine whether E , in the2

marily within the nucleus. absence of changing serum progesterone concentrations,

493A. Zsarnovszky et al. / Brain Research 958 (2002) 488–496

Fig. 5. ERb-immunonegative neurons in close opposition to ERb-immunopositive astroglial processes. Two representative photomicrographs of ERb-IRastrocytes (white arrows) from the lateral part of the IPN whose distal processes (black arrowheads) are embracing (A) or unilaterally in close association(B) with ERb-immunonegative neuronal perikarya (black arrow). Scale bar represents 15mm.

can modulate GFAP-expression within the IPN. In the astroglia or neurons) in regions of significant changes inpresent study we examined the effect of a single subcuta- GFAP-expression are responsive to E and must therefore2

neous injection of E on the expression of GFAP in the possess E -responsive signaling mechanisms; and that the2 2

IPN of ovariectomized female rats. Our findings support IPN has a rostrocaudal segmentation pattern of E -sen-2

the conclusion that increased E concentrations were sitivity.2

sufficient to cause increased GFAP-expression in the IPN. Although these data indicate that E can modulate2

Additionally, detailed analysis of the changes in the pattern GFAP-expression in the IPN, the mechanism responsibleof GFAP-IR revealed that E differentially affects in- remains unclear. One potential pathway through which E2 2

dividual IPN subnuclei. These observations suggest that may regulate astrocytic GFAP-expression is the directthe anatomical structures (neuronal afferents and/or local activation of ER-mediated mechanisms in the astrocytes.

Previous studies have demonstrated that transcripts encod-ing ERb, but not ERa, were expressed in the IPN [30,42].However, a determination of the cell-type specific patternof ERb-expression is not possible from those earlierstudies. Our results indicate that ERb co-localized withGFAP in a considerable number of cells in the IPN,suggesting that astroglia are directly influenced by E2

through ERb-mediated mechanisms. The role of ERb asthe only ER expressed in astrocytes is supported byprevious reports demonstrating that ERb was expressed inastroglia of the arcuate nucleus, hippocampus and thecerebellum [6,7,24].

Additionally, ERb was found localized to the nucleus ofa subset of IPN neurons suggesting that E may directly2

influence some IPN neurons through ERb-mediated trans-activation of E responsive gene expression. Thus, the2

function of a subset of IPN neurons may be differentiallyregulated through ERb-mediated changes in estrogenresponsive gene expression. The cell body of the ERb-

Fig. 6. Differential density of ERb-IR in the IPN. In caudal portions of expressing neurons was found surrounded by ERb-im-the IPN (AP26.72) nuclear, presumably neuronal, ERb-IR is primarily munonegative glial sheathes, whereas some ERb-immuno-confined to the dorsomedial and apical subnuclei, whereas in the lateral negative neurons were found surrounded by the infrequent-subnucleus, scattered ERb-IR astroglia are observed. Only scattered

ly observed ERb-immunopositive astrocytic sheaths. Be-ERb-immunopositive cells can be seen in the central and intermediatecause glial /neuronal interactions at the synapse are knownsubnuclei. A, apical; C, central subnucleus; DM, dorsomedial; I, inter-

mediate; L, lateral. Scale bar represents 150mm. to play a significant role in the regulation of synaptic

A. Zsarnovszky et al. / Brain Research 958 (2002) 488–496494

activity [23,44], those observations suggest that the distal including those subnuclei that were not significantlyprocesses of ERb-immunopositive astrocytes may play a influenced by E -treatment. Further, the central, inter-2

role in the modulation of synaptic structure and activity of mediate and lateral subnuclei of the IPN are innervated byERb-immunonegative neurons in response to changing E the E -sensitive lateral hypothalamus and the nucleus of2 2

concentrations. These observations suggest that neuronal the diagonal band of Broca, yet in those subnuclei signifi-activity of the IPN may respond to changing E con- cant changes in GFAP-expression were not consistently2

centrations in two independent and temporally distinct observed. In the rostral, lateral, dorsomedial and dorsola-fashions: (i) by rapidly altering synaptic transmission teral subnuclei, which are receptive fields of dorsal teg-directly at the synapse of ERb-immunonegative neurons; mental area afferents [1,8,14,33,42], only the lateral subnu-and (ii) more slowly through the regulation of estrogen cleus showed pronounced changes in GFAP-immunostain-responsive gene expression in ERb-immunopositive neu- ing.rons. Although the functional relevance of the segmented

The direct action of E on neurons and astroglia of the pattern of variable estrogen responsiveness detected along2

IPN can account for the observed changes in GFAP- the long axis of the IPN is not yet known, the finding of aexpression; nevertheless, the E -dependent release of compartmentalized pattern of E -responsiveness in the IPN2 2

cytokines and growth factors can also influence GFAP- raises many interesting possibilities. Because estrogen canexpression [13,29,31,35], and thus, cannot be ruled out as induce plastic synaptic changes in the hypothalamus andindirectly contributing mechanisms. Similarly, the potential hippocampus that result in well-described changes ininfluence of afferent mediated changes in synaptic struc- estrogen-regulated functions (e.g. the regulation of theture coordinately regulating GFAP-expression in associ- hypothalamic gonadotrophin secretion or the facilitation ofated astrocytes of the IPN can also not be ruled out. hippocampal long-term potentiation), it is possible that the

Fluctuations in GFAP-expression are a hallmark of the contrasted estrogenic conditions applied in this studysynaptic remodeling that occurs during the estrous cycle in reveal not only a morphological compartmentalization, butthe mediobasal hypothalamus and in the hippocampus also functionally compartmentalized units along the long[11,12,45]. Many forebrain areas, such as the nucleus of axis of the IPN. Similar functional and morphologicthe diagonal band of Broca, the lateral hypothalamus, and patterns of spinal cord segmentation and modularity of thethe supramamillary nucleus, are major centers of E - cerebral cortex are well described, and the lamellar organi-2

dependent regulation of ‘limbic’ functions, and their zation of hippocampal excitatory pathways was proposedprojection fields include the mediobasal hypothalamus, the based on structural and functional data [4]. While thehippocampus, and also the IPN [1,8]. Therefore, it remains original hypothesis of hippocampal organization is sup-possible that the synaptic activity of these forebrain ported by numerous studies, the exact nature of hippocam-afferents is being altered by changes in E -concentration, pal compartmentalization appears to be more complex and2

and as in the hypothalamus and hippocampus, those remains unclear [2,3], and the relationship between thealterations in synaptic activity may in turn influence compartmental organization of the hippocampus and theGFAP-expression in the IPN. For example, at certain AP IPN is not at all well characterized. To the best of ourlevels, the dorsomedial, dorsolateral and rostrolateral sub- knowledge, an AP compartmentalization of the IPN hasnuclei, GFAP-IR was increased in response to E -treat- not previously been reported; however, neurons of the IPN2

ment. Each of those subnuclei receive afferents from the at different rostrocaudal locations are known to be gener-dorsal raphe nucleus, a mesencephalic region known to ated on different embryonic days [32] and numerousexpress ERb [41,42]. Thus, in the dorsomedial, dorsolater- studies indicate that in the IPN there are marked region-al and rostrolateral subnuclei, the regulation of GFAP- specific differences in the cytoarchitecture, synaptology,expression through estrogen-modulated changes in afferent connectivity, and neuropeptide and biogenic amine contentsynaptic activity cannot be excluded. [9,18,19,26,32,37]. Considering the morphological and

While E -modulated synaptic plasticity outside of the functional ties between the IPN and the hippocampal2

IPN can potentially account for some of the observed complex, it would not be surprising if future studieschanges in GFAP-expression, careful comparison of the identified a rostrocaudal modularity in the function of theconnectivity of individual subnuclei of the IPN IPN that is important for understanding the numerous[17,18,21,22] with the overall pattern of the observed higher functions that are mediated through this brainsubnuclear specific changes in GFAP-expression reveal nucleus. Because of the important role of the IPN inthat regions of the IPN innervated by well known E - regulating memory and learning-related higher brain func-2

responsive forebrain structures did not consistently display tions that are associated with cortical and subcorticalchanges in GFAP-expression. For example, the most theta-type EEG-activity [39,46,47,49], the finding that E2

substantial innervation of the IPN is from the habenulae affects the structure of distinct IPN subnuclei and that both[25,34]. Although ERs are present in the habenular nuclei astrocytes and neurons of the IPN express ERb may be of[42,48,50], the habenular projections through the fas- particular significance with regard to understanding es-ciculus retroflexus innervate virtually all of the IPN [14], trogen’s influence on higher brain functions.

495A. Zsarnovszky et al. / Brain Research 958 (2002) 488–496

[17] G.S. Hamill, N.J. Lenn, A morphological and morphometric analysisA cknowledgementsof the subnuclear structure of the interpeduncular nucleus in the rat,Soc. Neurosci. Abstr. 7 (1981) 83.

This work was supported by NINDS/NIH grants [18] G.S. Hamill, N.J. Lenn, The subnuclear organization of the ratNS37795 and NS42798 (S.M.B.). T.S. was the recipient of interpeduncular nucleus: a light and electron microscopic study, J.a ‘Partners in Research’ undergraduate research fellowship Comp. Neurol. 222 (1984) 396–408.

[19] G.S. Hamill, J.A. Olschowka, N.J. Lenn, D.M. Jacobowitz, Thefrom the Arkansas Cancer Research Center that wassubnuclear distribution of substance P, cholecystochinin, vasoactivefunded through NCI/NIH grant R25 CA 49425.intestinal peptide, somatostatin, leu-enkephalin, dopamine-b-hy-droxylase, and serotonin in the rat interpeduncular nucleus, J. Comp.Neurol. 226 (1984) 580–596.

[20] F. Haun, T.C. Eckenrode, M. Murray, Habenula and thalamus cellR eferencestransplants restore normal sleep behaviors disrupted by denervationof the interpeduncular nucleus, J. Neurosci. 12 (1992) 3282–3290.

[1] S. Ahlenius, W.J.H. Nauta, Afferent and efferent connections of the [21] L.M. Hemmendinger, R.Y. Moore, The interpeduncular nucleus ininterpeduncular nucleus in the rat as revealed by neuroanatomical the rat: cytoarchitecture and cytochemistry, Soc. Neurosci. Abstr. 8tracer techniques, Neurosci. Lett. Suppl. 5 (1980) 188. (1982) 665.

[2] D.G. Amaral, Emerging principles of intrinsic hippocampal organi- [22] L.M. Hemmendinger, R.Y. Moore, Interpeduncular organization inzation, Curr. Opin. Neurobiol. 3 (1993) 225–229. the rat: cytoarchitecture and histochemical analysis, Brain Res. Bull.

[3] D.G. Amaral, M.P. Witter, The three-dimensional organization of the 13 (1984) 163–179.hippocampal formation: a review of anatomical data, Neuroscience [23] M. Iino, K. Goto, W. Kakegawa, H. Okado, M. Sudo, S. Ishiuchi, A.31 (1989) 571–591. Miwa, Y. Takayasu, I. Saito, K. Tsuzuki, S. Ozawa, Glia-synapse

21[4] P. Andersen, V.P. Bliss, K.K. Skrede, Lamellar organization of interaction through Ca -permeable AMPA receptors in Bergmannhippocampal excitatory pathways, Exp. Brain Res. 13 (1971) 222– glia, Science 292 (2001) 926–929.238. [24] R.L. Jakab, J.K. Wong, S.M. Belcher, Estrogen receptorb immuno-

[6] I. Azcoitia, A. Sierra, L.M. Garcia-Segura, Localization of estrogen reactivity in differentiating cells of the developing rat cerebellum, J.receptor beta-immunoreactivity in astrocytes of the adult rat brain, Comp. Neurol. 430 (2001) 396–409.Glia 26 (1999) 260–267. [25] K.Y. Kataoka, Y. Nakamura, R. Hassler, Habenulo-interpeduncular

[7] G.P. Cardona-Gomez, L. Don Carlos, L.M. Garcia-Segura, Insulin- tract: a possible cholinergic neuron in the rat brain, Brain Res. 62like growth factor I receptors and estrogen receptors colocalize in (1973) 264–267.female rat brain, Neuroscience 99 (2000) 751–760. [26] M.D. Kawaja, B.A. Flumerfelt, S.P. Hunt, A.W. Hrycyshyn, Sub-

[8] A. Contestabile, B.A. Flumerfelt, Afferent connections of the stance P immunoreactivity in the rat interpeduncular nucleus:interpeduncular nucleus and the topographic organization of the synaptic interactions between substance P-positive profiles andhabenulo-interpeduncular pathway: an HRP study in the rat, J. choline acetyltransferase- or glutamate decarboxylase-immuno-Comp. Neurol. 196 (1981) 253–270. reactive structures, Neuroscience 42 (1991) 739–755.

[9] A. Contestabile, F. Fonnum, Topography of cholinergic and sub- [27] M. Kawakami, T. Akema, S. Ando, Electrophysiological studies onstance P pathways in the habenulo-interpeduncular system: an the neural networks among estrogen and progesterone effectiveimmunocytochemical and microchemical approach, Neuroscience 21 brain areas on lordosis behavior of the rat, Brain Res. 169 (1979)(1987) 253–270. 287–301.

[10] J.R. Day, N.J. Laping, M. Lampert-Etchells, S.A. Brown, J.P. [28] S.G. Kohama, J.R. Goss, T.H. McNeill, C.E. Finch, Glial fibrillaryO’Callaghan, T.H. McNeill, C.E. Finch, Gonadal steroids regulate acidic protein mRNA increases at proestrus in the arcuate nucleus ofthe expression of glial fibrillary acidic protein in the adult male rat mice, Neurosci. Lett. 183 (1995) 164–166.hippocampus, Neuroscience 55 (1993) 435–443. [29] F. Labombarda, S. Gonzalez, P. Roig, A. Lima, R. Guennoun, M.

[11] L.M. Garcia-Segura, I. Torres-Aleman, F. Naftolin, Astrocytic shape Schumacher, A.F. De Nicola, Modulation of NADPH-diaphoraseand glial fibrillary acidic protein immunoreactivity are modified by and glial fibrillary acidic protein by progesterone in astrocytes fromestradiol in primary rat hypothalamic cultures, Dev. Brain Res. 47 normal and injured rat spinal cord, J. Steroid Biochem. Mol. Biol.(1989) 298–302. 73 (2000) 159–169.

[12] L.M. Garcia-Segura, S. Luquin, A. Parducz, F. Naftolin, Gonadal [30] N. Laflamme, R.E. Nappi, G. Drolet, C. Labrie, S. Rivest, Expres-hormone regulation of glial fibrillary acidic protein immunoreactivi- sion and neuropeptidergic characterization of estrogen receptorsty and glia ultrastructure in the rat neuroendocrine hypothalamus, (ERalpha and ERbeta) throughout the rat brain: anatomical evidenceGlia 10 (1994) 59–69. of distinct roles of each subtype, J. Neurobiol. 36 (1998) 357–378.

[13] F.C. Gomes, D. Paulin, V. Moura Neto, Glial fibrillary acidic protein [31] N.J. Laping, B. Teter, N.R. Nichols, I. Rozovsky, C.E. Finch, Glial(GFAP): modulation by growth factors and its implication in fibrillary acidic protein: regulation by hormones, cytokines, andastrocyte differentiation, Braz. J. Med. Biol. Res. 32 (1999) 619– growth factors, Brain Pathol. 4 (1994) 259–275.631. [32] N.J. Lenn, S.A. Bayer, Neurogenesis in subnuclei of the rat

[14] H.J. Groenewegen, S. Ahlenius, S.N. Haber, N.W. Kowall, W.J.H. interpeduncular nucleus and medial habenula, Brain Res. Bull. 16Nauta, Cytoarchitecture, fiber connections, and some histochemical (1986) 219–224.aspects of the interpeduncular nucleus in the rat, J. Comp. Neurol. [33] E.R. Marchand, J.N. Riley, R.Y. Moore, Interpeduncular nucleus249 (1986) 65–102. afferents in the rat, Brain Res. 193 (1980) 339–352.

[15] F. Hajos, K. Halasy, B. Gerics, F. Szalay, Glial fibrillary acidic [34] R. Mitchell, Connections of the habenula and the interpeduncularprotein (GFAP)-immunoreactivity is reduced by castration in the nucleus in the cat, J. Comp. Neurol. 121 (1963) 441–457.interpeduncular nucleus of male rats, Neuroreport 10 (1999) 2229– [35] W. Ochiai, M. Yanagisawa, T. Takizawa, K. Nakashima, T. Taga,2233. Astrocyte differentiation of fetal neuroepithelial cells involving

[16] F. Hajos, K. Halasy, B. Gerics, F. Szalay, E. Michaloudi, G.C. cardiotrophin-1-induced activation of STAT3, Cytokine 14 (2001)Papadopoulos, Ovarian cycle-related changes of glial fibrillary 264–271.acidic protein (GFAP) immunoreactivity in the rat interpeduncular [36] G. Paxinos, C. Watson, The Rat Brain in Stereotaxic Coordinates,nucleus, Brain Res. 862 (2000) 43–48. Academic Press, 1986, 2nd edn.

A. Zsarnovszky et al. / Brain Research 958 (2002) 488–496496

[37] J.C. Pearson, Immunocytochemical localization of GABA in the rat [46] A. Valjakka, J. Vartiainen, L. Tuomisto, J.T. Tuomisto, H. Olkkonen,interpeduncular nucleus, Soc. Neurosci. Abstr. 14 (1988) 472.15. M.M. Airaksinen, The fasciculus retroflexus controls the integrity of

[39] S. Raghavachari, M.J. Kahana, D.S. Rizzuto, J.B. Caplan, M.P. REM sleep by supporting the generation of hippocampal thetaKirschen, B. Bourgeois, J.R. Madsen, J.E. Lisman, Gating of human rhythm and rapid eye movements in rats, Brain Res. Bull. 47 (1998)theta oscillations by a working memory task, J. Neurosci. 21 (2001) 171–184.3175–3183. [47] V. Varga, A. Kekesi, G. Juhasz, B. Kocsis, Reduction of the

[40] P. Riskind, R.L. Moss, Effects of lesions of putative LHRH-con- extracellular level of glutamate in the median raphe nucleustaining pathways and midbrain nuclei on lordotic behavior and associated with hippocampal theta activity in the anaesthetized rat,luteinizing hormone release in ovariectomized rats, Brain Res. Bull. Neuroscience 84 (1998) 49–57.11 (1983) 493–500. [48] C.K. Wagner, A.J. Silverman, J.I. Morrell, Evidence for estrogen

[41] P.J. Shughrue, I. Merchenthaler, Distribution of estrogen receptor receptor in cell nuclei and axon terminals within the lateral habenulabeta immunoreactivity in the rat central nervous system, J. Comp. of the rat: regulation during pregnancy, J. Comp. Neurol. 392 (1998)Neurol. 436 (2001) 64–81. 330–342.

[42] P.J. Shughrue, M.V. Lane, I. Merchenthaler, Comparative distribution [49] S.P. Wiebe, U.V. Staubli, Recognition memory correlates of hip-of estrogen receptor-alpha and -beta mRNA in the rat central pocampal theta cells, J. Neurosci. 21 (2001) 3955–3967.nervous system, J. Comp. Neurol. 388 (1997) 507–525. [50] M. Yokosuka, H. Okamura, S. Hayashi, Postnatal development and

[43] L.K. Takahashi, R.D. Lisk, Intracranial sites regulating the biphasic sex difference in neurons containing estrogen receptor-alpha im-action of progesterone in estrogen-primed golden hamsters, Endo- munoreactivity in the preoptic brain, the diencephalon, and thecrinology 119 (1986) 2744–2754. amygdala in the rat, J. Comp. Neurol. 389 (1997) 81–93.

[44] D.T. Theodosis, D.A. Poulain, Contribution of astrocytes to activity- [51] A. Zsarnovszky, T.L. Horvath, F. Naftolin, C. Leranth, AMPAdependent structural plasticity in the adult brain, Adv. Exp. Med. receptors colocalize with neuropeptide-Y- and galanin-containing,Biol. 468 (1999) 175–182. but not with dopamine, neurons of the female rat arcuate nucleus: a

[45] P.A. Tranque, I. Suarez, G. Olmos, B. Fernandez, L.M. Garcia- semiquantitative immunohistochemical colocalization study, Exp.Segura, Estradiol-induced redistribution of glial fibrillary acidic Brain Res. 133 (2000) 532–537.protein immunoreactivity in the rat brain, Brain Res. 406 (1987)348–351.