Professional Documents
Culture Documents
Research report
Abstract
Glial cell line derived neurotrophic factor (GDNF) is a potent survival factor for several types of neurons. GDNF binds with high
affinity to GDNF-family receptor a-1 (GFRa-1). This receptor is expressed in different areas of the brain, including the hippocampus and
dentate gyrus. By using in situ hybridization and immunohistochemistry, we found that 19% to 37% of glutamic acid decarboxylase
(GAD) expressing neurons co-expressed GFRa-1 in the hippocampus. GFRa-1 / GAD co-expression was found mainly in the stratum (s)
pyramidale (29–37%) and s. oriens (20–25%). Further characterization of GFRa-1 expressing interneurons, based on their calcium-
binding protein immunoreactivity, demonstrated that many parvalbumin (PV) immunoreactive neurons express GFRa-1 in the s.
pyramidale of CA1 (72%), CA2 (70%) and CA3 (70%) subfields of the hippocampus. GFRa-1 / PV double labeled neurons were also
detected in the s. oriens of CA1 (52%), CA2 (27%) and CA3 (36%) subfields. The expression of GFRa-1 in principal neurons and in a
specific sub-population of GABAergic neurons (PV-containing neurons) suggest that GDNF might modulate, in a selective manner,
functions of the entire adult hippocampus. 2000 Elsevier Science B.V. All rights reserved.
0006-8993 / 00 / $ – see front matter 2000 Elsevier Science B.V. All rights reserved.
PII: S0006-8993( 00 )02682-2
A. Sarabi et al. / Brain Research 877 (2000) 262 – 270 263
those brain areas originally shown to be responsive to hybridized at 558C for 16 h with 10 7 cpm / ml sense or
GDNF (i.e. motoneurons and dopaminergic cells). As antisense [ 35 S]-labeled GFRa-1 probes (corresponding to
GFRa-1 is widely expressed in the adult brain [27,40,42], nucleotides 273–1741 of the rat GFRa-1 cDNA, original
it has been suggested that GDNF may activate several GFRa-1 plasmid was kindly provided by Dr. Andreas
sub-types of cells. GFRa-1 is expressed in both embryonic Tomac, NIDA). Sections were treated with 4 mg / ml RNase
and adult rodent hippocampus [21,27,40,42]. RT-PCR A at 378C for 1 h, washed in 13SSC, 50% formamide at
measurements demonstrate that the highest levels of 558C for 2 h, and in 0.13SSC at 688C for 1 h. Sections
hippocampal GFRa-1 expression are reached between were mounted on coated slides, air dried, dipped in nuclear
postnatal days 0 and 14 [21], corresponding to crucial track emulsion and exposed for several weeks prior to
periods for cellular migration and synaptogenesis, and development.
implying that GDNF through GFRa-1 might participate in
the development and maturation of the hippocampus.
2.3. Double in situ hybridization
Additional studies have demonstrated that GFRa-1 expres-
sion in the hippocampus is highly regulated by physiologi-
For double in situ hybridization material was processed
cal changes. For example, up-regulation of GFRa-1 ex-
as described above using a [ 35 S]-labeled antisense GFRa-1
pression is triggered in the adult hippocampus by kainic
probe and digoxigenin-labeled antisense probes for the two
acid treatment [33,42] and kindling-evoked stimulation
isoforms of rat glutamic acid decarboxylase (GAD 65 and
[19], as well as after global forebrain ischemia ([19];
GAD 67 , generously provided by Dr. Allan Tobin, Uni-
unpublished observations) and cortical injury (unpublished
versity of California Los Angeles, Los Angeles, CA). After
observations).
last SSC wash, sections were rinsed with TBS buffer (20
While hippocampal expression of GFRa-1 has been
mM TRIS, 0.5 M NaCl), pH 8.2. Sections were incubated
documented [27,40,42], the phenotypic properties of hip-
with an alkaline phosphatase-conjugated antibody against
pocampal cells which express GFRa-1 have not been
digoxigenin (Dig, Boehringer Mannheim; Indianapolis,
determined. Within the hippocampus, there are gluta-
IN) overnight at 48C; alkaline phosphatase reaction was
matergic and GABAergic neurons. Moreover, hippocampal
developed with nitroblue tetrazolium and 5-bromo-4-chlo-
GABAergic neurons constitute a heterogeneous population
ro-3-indolyl phosphate (Life Technologies; Gaithersburg,
with regard to their morphology, biochemical composition
MD). Sections were dipped in nuclear track emulsion and
and synaptic connections [8–10,16,20,37]. Sub-populations
exposed for several weeks prior to development.
of GABAergic neurons can be identified according to their
content of calcium binding protein; parvalbumin (PV),
calbindin (CB) and calretinin (CR) [3]. In the present 2.4. In situ hybridization and immunocytochemistry
communication, we have used double labeling techniques
to determine the cellular phenotype of hippocampal GFRa- In situ hybridization combined with immunolabeling
1 expressing neurons to better understand GDNF target was performed as previously described [25]. After pre-
cells in this brain structure. hybridization, sections were hybridized with [ 35 S]-labeled
antisense GFR a-1 probe at 558C for 16 h. Sections were
treated with 4 mg / ml RNase A at 378C for 1 h, washed in
2. Material and methods 13SSC, 50% formamide at 558C for 2 h, and in 0.13SSC
at 688C for 1 h. Material was rinsed with PB prior to
2.1. Tissue preparation processing for immunocytochemistry, incubated in 4%
bovine serum albumin (BSA) supplemented with 0.3%
Adult Sprague–Dawley male rats (200–250 g body Triton X-100 in PB for 1 h, followed by incubation with
weight) were anesthetized with chloral hydrate (35 mg / the corresponding primary antibody for 24 h at 48C.
100 g) and perfused transcardially with a solution of 4% Primary antibody dilutions were done in PB containing 1%
paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.3. BSA and 0.3% Triton X-100. The following antibodies
Brains were then removed from the skull, postfixed from 2 were used at 1:2000 dilutions: anti-parvalbumin (Incstar,
to 15 h at 48C, rinsed with PB and sequentially transferred Sillwater, MI), anti-calbindin mouse monoclonal and anti-
to 12, 14 and 18% sucrose solutions. Brains were frozen calretinin rabbit polyclonal antibodies (Swant, Switzer-
on dry ice and sectioned on a cryostat to obtain coronal land). After rinsing 3310 min in PB, sections were
sections of 40 mm in thickness. All procedures were processed with an ABC kit (Vector Laboratories, Burling-
approved by NIDA Animal Care and Use Committee. ame, CA). Material was incubated in a 1:200 dilution of
the biotinylated secondary antibody. After rinsing with PB,
2.2. In situ hybridization sections were incubated with avidin-biotinylated horserad-
ish peroxidase for 2 h. Samples were rinsed and the
In situ hybridization was performed as previously peroxidase reaction was developed with 0.05% 3,3-
described [25]. After pre-hybridization, sections were diaminobenzidine-4 HCl (DAB) and 0.003% hydrogen
264 A. Sarabi et al. / Brain Research 877 (2000) 262 – 270
peroxide (H 2 O 2 ). Sections were mounted on coated slides, the identity of interneurons was established by their
air dried, dipped in nuclear track emulsion and exposed for expression of GAD mRNA. Co-expression of GFRa-1 and
several weeks prior to development. All used antibodies GAD transcripts were found mainly in s. oriens and s.
resulted in specific pattern of labeling as previously pyramidale (Fig. 2A–B). Semi-quantitative analysis (Table
described for PV [3,20], CR [3,11,14] and CB [2,3,39]. 1) indicated that 29%62 (sem, n5554) of the total
population of GAD expressing neurons co-express GFRa-
2.5. Data analysis 1 in s. pyramidale of CA1. Co-expression was also seen in
s. pyramidale of CA2 (27%62, SEM, n5194) and CA3
Sections processed for in situ hybridization and im- (37%62.9, SEM, n5592). Lower co-expression was
munocytochemistry were analyzed and photographed with found in s. oriens of CA1 (20%62.2 SEM, n5389), CA2
bright field or polarized microscopy. Observation of the (25%63.3; SEM, n544) and CA3 (19%62.0, SEM, n5
entire hippocampus indicated lack of GFRa-1 expressing 250) subfields of the hippocampus.
in either CR- or CB-immunoreactive neurons, thus, quan-
titative analysis was omitted for this material. The GFRa-1 3.2. Expression of GFRa -1 in PV-immunoreactive
expressing cells, and neurons containing GAD mRNA or interneurons of the hippocampus and dentate gyrus
PV-immunoreactivity were counted in random sections of
the CA1–CA3 subfields of the hippocampus and dentate We sought to determine if any of the Ca 21 binding
gyrus. For each region, the percentage of neurons con- proteins known to be present in GABAergic neurons were
taining GFRa-1 and GAD in the total population of GAD detected in GFRa-1 expressing interneurons. While no CB
expressing neurons was calculated. Similarly, the per- either CR immunoreactivity was observed in GFRa-1
centage of neurons containing GFRa-1 mRNA and PV- expressing interneurons, GFRa-1 / PV double-labeled neu-
immunoreactivity in the total population of PV-immuno- rons were frequently found in the s. pyramidale of the
labeled neurons was calculated. A neuron was considered CA1–CA3 sub-fields (Fig. 3). GFRa-1 / PV double-labeled
double labeled when its soma was brown (PV-immuno- neurons were also found in the s. oriens of CA1–CA3
label) or purple (GAD-expression) and contained an (Fig. 3) subfields but were not detected in either the s.
aggregation of silver particles clearly above the immedi- radiatum or s. lacunosum moleculare. Within the dentate
ately surrounding background. Typically a minimum of gyrus, GFRa-1 / PV double-labeled neurons were present in
15–20 silver grains represented weak label above back- the s. granulosum (Fig. 4). A semiquantitative analysis of
ground. Background was also evaluated from slides hy- GFRa-1 / PV double-labeled cells (Table 2) indicated that
bridized with sense probes. Double-labeled cells were most of these cells are concentrated in the s. pyramidale of
counted in 4–6 random sections from 4 rats. CA1 (72%62.3, SEM, n5508), CA2 (70%62.9, SEM,
n5148) and CA3 (70%62.4, SEM, n5358) sub-fields of
the hippocampus. Lower co-localization of PV-immuno-
3. Results reactivity and GFRa-1 transcripts was observed in the s.
oriens of CA1 (52%62.7, SEM, n5152), CA2 (27%63.2,
We confirmed and extended previous reports SEM, n526) and CA3 (36%64.3, SEM, n529) sub-fields
[21,27,40,42] showing expression of GFRa-1 in the hip- of the hippocampus and s. granulosum (25%61.3, SEM,
pocampus (Fig. 1A–B). GFRa-1 transcripts were observed n5130) of the dentate gyrus. Within the s. pyramidale, PV
mainly in the s. oriens and s. pyramidale of CA1–CA3 immunoreactive neurons clearly showed higher concen-
sub-fields of the hippocampus (Fig. 1A–B). The signal for trations of silver grains than the surrounding pyramidal
GFRa-1 expression was uniformly moderate in the s. neurons; while most CA1 pyramidal neurons contained
pyramidale (Fig. 1A–B), indicating for the presence of between 5–18 silver grains, adjacent PV-immunoreactive
GFRa-1 in principal glutamatergic neurons. However, we neurons had more than 50 silver grains.
frequently observed single cells with strong label in the s.
oriens (Fig. 1A–B) and s. pyramidale (Fig. 1A–B),
suggesting that GFRa-1 might be expressed not only in 4. Discussion
principal glutamatergic neurons but also in interneurons.
Thus, we sought to determine the phenotype of these By using double labeling techniques, we have demon-
putative interneurons expressing GFRa-1 mRNA. strated that GFRa-1 is expressed in both pyramidal and
GABAergic neurons in the hippocampus. Hippocampal
3.1. Expression of GFRa -1 in hippocampal interneurons GFRa-1 expressing interneurons constitute 19–37% of the
total population of GABAergic neurons in the s.
We used double in situ hybridization histochemistry to pyramidale and s. oriens of CA1–CA3 subfields of the
investigate the phenotype of GFRa-1 expressing neurons hippocampus. Consistently, the levels of GFRa-1 expres-
in the hippocampus. In this procedure, GFRa-1 transcripts sion were lower in pyramidal neurons when compared with
were detected with a radioactive antisense riboprobe while interneurons. While these results suggest that under normal
A. Sarabi et al. / Brain Research 877 (2000) 262 – 270 265
Fig. 1. Distribution of GFRa-1 expressing cells in CA1 (A) and CA3 (B) subfields of the hippocampus. (A–B). Moderate levels of GFRa-1 are seen in the
s. pyramidale of the CA1 and CA3 subfields. Note additional highly expressing cells scattered in s. pyramidale (large arrows) and s. oriens (small arrows).
SP, s. pyramidale. SO, s. oriens. SR, s. radiatum. Scale bar5100 mm.
266 A. Sarabi et al. / Brain Research 877 (2000) 262 – 270
Fig. 2. Simultaneous detection of GFRa-1 and GAD transcripts in the hippocampus. Under polarized light, GFRa-1 / GAD double-labeled cells appeared
as dark neurons (GAD expression) covered with green grains (GFRa-1 expression). (A). CA1. GFRa-1 / GAD double-labeled cells are found in the SP (big
arrows) or in its proximity (small arrow). Single GAD expressing cells (arrow heads) are seen in SO and SP. (B). CA3. GFRa-1 / GAD double-labeled cells
are present in the SP (big arrow) and SO (small arrows). Note a single GAD expressing cell (arrow head) in the SO. Scale bar525 mm for A and 32 mm for
B.
conditions, GABAergic interneurons might be the principal heterogeneous population with regard to their morphology,
sites of GFRa-1 expression, previous studies have shown biochemical composition and synaptic connections
that GFRa-1 expression can be up-regulated in principal [8,10,16,20,37]. Sub-populations of GABAergic neurons
neurons following neuronal activation [19,33,41]. For can be identified according to their content of calcium
instance, local [42] or systemic applications [33] of kainic binding protein (CB, CR and PV). We found that many of
acid increase expression of GFRa-1 in principal neurons of the GFRa-1 expressing interneurons contain parvalbumin,
CA1, CA3 and dentate gyrus. Similar increase in expres- these GFRa-1 / PV double-labeled cells were preferentially
sion was obtained after kindling [19], global forebrain distributed in the s. pyramidale, where about 70% of
ischemia ([19]; unpublished observations) and cortical PV-immunoreactive neurons expressed GFRa-1. Less co-
injury (unpublished observations). localization was found in the s. oriens of CA1 (52%) and
Hippocampal GABAergic interneurons constitute a s. granulosum (25%). Our data are consistent with previ-
A. Sarabi et al. / Brain Research 877 (2000) 262 – 270 267
Fig. 3. Simultaneous detection of GFRa-1 transcripts and PV immunoreactivity in the CA1 subfield of the hippocampus. Under polarized light,
GFRa-1 / PV double-labeled cells appear as brown neurons (PV-immunoreactivity) covered with green grains (GFRa-1 expression). GFRa-1 / PV
double-labeled cells are concentrated in s. pyramidale (arrows). PV single-labeled neuron is seen in s. oriens (arrow head). Scale bar525 mm.
268 A. Sarabi et al. / Brain Research 877 (2000) 262 – 270
Fig. 4. Simultaneous detection of GFRa-1 transcripts and PV immunoreactivity in the dentate gyrus. Under polarized light, GFRa-1 / PV double-labeled
cells appear as brown neurons (PV-immunoreactivity) covered with green grains (GFRa-1 expression). (A). Low magnification micrograph showing
PV-immunoreactive neurons in s. granulosum (SG). (B). At higher magnification, GFRa-1 / PV double-labeled cells with high (arrows) to low (arrow-head)
levels of GFRa-1 expression are concentrated in s. granulosum. H, hilus. Scale bar58 mm for A and 25 mm for B.
GFRa-1 transcripts are distributed in principal neurons and detected in hippocampus of adult rodent brains [1,21,42].
interneurons in the hippocampus and dentate gyrus, it However, increases in GDNF expression also occur as a
remains to be seen whether these two kind of cells use Ret consequence of kainate-induced excitation [13]. The pres-
dependent [7,15,40,41] or independent [31,43] signaling ence of GDNF transcripts within the hippocampus, suggest
mechanisms. Both pathways may be in place in the that this might be a source for GFRa-1 activation but an
hippocampus, since Ret is expressed in the hippocampus additional source might reside in hippocampus target
albeit in very low levels [21,42]. In response to kainate- regions, such as the septum which also has low levels of
induced excitation, scattered hippocampal neurons showed GDNF expression [1].
increased Ret expression [42], suggesting that Ret expres- We concluded that GFRa-1 is expressed in specific
sion is regulated by hippocampal physiological activity. subsets of hippocampal neurons suggesting a role for
Similarly, very low levels of GDNF transcripts have been GDNF in the adult hippocampal function.
A. Sarabi et al. / Brain Research 877 (2000) 262 – 270 269
Expression of nerve growth factor and neurotrophin-3 mRNAs in GABA-, cholecystokinin-, vasoactive intestinal polypeptide-, and
hippocampal interneurons: morphological characterization, levels of somatostatin-like immunoreactivity in the area dentate and hip-
expression, and colocalization of nerve growth factor and pocampus of the rat, J. Comp. Neurol. 256 (1987) 42–60.
neurotrophin-3, J. Comp. Neurol. 395 (1998) 73–90. [38] A. Tomac, E. Lindqvist, L.F. Lin, S.O. Ogren, D. Young, B.J.
[29] J.G. Pichel, L. Shen, H.Z. Sheng, A.C. Granholm, J. Drago, A. Hoffer, L. Olson, Protection and repair of the nigrostriatal dopa-
Grinberg, E.J. Lee, S.P. Huang, M. Saarma, B.J. Hoffer, H. Sariola, minergic system by GDNF in vivo, Nature 373 (1995) 335–339.
H. Westphal, Defects in enteric innervation and kidney development [39] K. Toth, T.F. Frund, Calbindin D28K-containing nonpyramidal cells
in mice lacking GDNF, Nature 382 (1996) 73–76. In the rat hippocampus: their Immunoreactivity for GABA and
[30] K. Pong, R.Y. Xu, W.F. Baron, J.C. Louis, K.D. Beck, Inhibition of projection to the medial septum, Neuroscience 49 (1992) 793–805.
phosphatidylinositol 3-kinase activity blocks cellular differentiation [40] J.J. Treanor, L. Goodman, F. de Sauvage, D.M. Stone, K.T. Poulsen,
mediated by glial cell line-derived neurotrophic factor in dopa- C.D. Beck, C. Gray, M.P. Armanini, R.A. Pollock, F. Hefti, H.S.
minergic neurons, Neurochem. 71 (1998) 1912–1919. Phillips, A. Goddard, M.W. Moore, A. Buj-Bello, A.M. Davies, N.
[31] D. Poteryaev, A. Titievsky, Y.F. Sun, J. Thomas-Crusells, M. Asai, M. Takahashi, R. Vandlen, C.E. Henderson, A. Rosenthal,
Lindahl, M. Billaud, U. Arumae, M. Saarma, GDNF triggers a novel Characterization of a multicomponent receptor for GDNF, Nature
ret-independent Src kinase family-coupled signaling via a GPI- 382 (1996) 80–83.
linked GDNF receptor alpha 1, FEBS Lett. 463 (1999) 63–66. [41] M. Trupp, E. Arenas, M. Fainzilber, A.S. Nilsson, B.A. Sieber, M.
[32] M.L. Price, B.J. Hoffer, A.C. Granholm, Effect of GDNF on fetal Grigoriou, C. Kilkenny, E. Salazar-Grueso, V. Pachnis, U. Arumae,
septal forebrain transplant in ocula, Exp. Neurol. 141 (1996) 181– H. Sariola, M. Saarma, C.F. Ibanez, Functional receptor for GDNF
189. encoded by the c-ret proto-oncogene product, Nature 381 (1996)
[33] M. Reeben, A. Laurikainen, J.O. Hiltunen, E. Castren, M. Saarma, 785–789.
The messenger RNAs for both glial cell line-derived neurotrophic [42] M. Trupp, N. Belluardo, H. Funakoshi, C.F. Ibanez, Complementary
factor receptors, c-ret and GDNFRalpha, are induced in the rat brain and overlapping expression of glial cell line-derived neurotrophic
in response to kainate-induced excitation, Neuroscience 83 (1998) factor (GDNF), c-ret proto-oncogene, and GDNF receptor-alpha
151–159. indicates multiple mechanisms of trophic actions in the adult rat
[34] C.E. Ribak, R. Nitsch, L. Seress, Proportion of parvalbumin-positive CNS, J. Neurosci. 17 (1997) 3554–3567.
basket cells in the GABAergic innervation of pyramidal and granule [43] M. Trupp, R. Scott, S.R. Whittemore, C.F. Ibanez, Ret-dependent
cells of the rat hippocampal formation, J. Comp. Neurol. 300 (1990) and -independent mechanisms of glial cell line-derived neurotrophic
449–461. factor signaling in neuronal cells, J. Biol. Chem. 274 (1999) 20885–
[35] N. Rocamora, M. Pascual, L. Acsady, L. de Lecea, T.F. Freund, E. 20894.
Soriano, Expression of NGF and NT3 mRNAs in hippocampal [44] D.H. Van Weering, J.L.J. Bos, Glial cell line-derived neurotrophic
interneurons innervated by the GABAergic septohippocampal path- factor induces Ret-mediated lamellipodia formation, J. Biol. Chem.
way, J. Neurosci. 16 (1996) 3991–4004. 272 (1997) 249–254.
[36] M.P. Sanchez, I. Silos-Santiago, J. Frisen, B. He, S.A. Lira, M. [45] J. Ylikoski, U. Pirvola, J. Virkkala, P. Suvanto, X.Q. Liang, E.
Barbacid, Renal agenesis and the absence of enteric neurons in mice Magal, R. Altschuler, J.M. Miller, M. Saarma, Guinea pig auditory
lacking GDNF, Nature 382 (1996) 70–73. neurons are protected by glial cell line-derived growth factor from
[37] R. Sloviter, G. Nilaver, Immunocytochemical localization of degeneration after noise trauma, Hear Res. 124 (1998) 17–26.