Cellular localization of nerve growth factor-like immunoreactivity in adult rat brain: Quantitative and immunohistochemical study

Neuroscience Vol. 60, No. I, pp. 67-84, 1994

Pergamon

0306-4522(93)E0088-8

Elsevier ScienceLtd Copyright 0 1994IBRO Printed in Great Britain. All rights reserved 0306-4522/94$6.00+ 0.00

CELLULAR LOCALIZATION OF NERVE GROWTH FACTOR-LIKE IMMUNOREACTIVITY IN ADULT RAT BRAIN: QUANTITATIVE AND IMMUNOHISTOCHEMICAL STUDY T. NISHIO,*~ S. FURUKAWA,~ I. AKIGUCHI,t N. OKA,? K. OHNISHI,~ H. TOMIMOTO,t S. NAKAMURA~ and J. KIMURAt TDepartment of Neurology, Faculty of Medicine, Kyoto University, Sakyo-ku, Kyoto 606, Japan fDepartment of Molecular Biology, Gifu Pharmaceutical University, Gifu 502, Japan Abstract-To elucidate the role and the mechanism of action of nerve growth factor in the adult central nervous system, we investigated the localization of nerve growth factor-like immunoreactivity in adult rat brain, both quantitatively and immunohistochemically, using polyclonal anti-nerve growth factor immunoglobulin G. We raised rabbit polyclonal anti-mouse nerve growth factor antibody with an extremely high titer as 10m9determined by an enzyme immunoassay. The affinity-purified anti-nerve growth factor immunoglobuhn G specifically recognized nerve growth factor with no cross-reaction to recombinant brain-derived neurotrophic factor and neurotrophin-3 evaluated by an enzyme immunoassay. We quantified nerve growth factor content in each layer of the adult rat cerebral cortex and in each small piece (0.225 mg wet weight tissue) of the diencephalon, brainstem and cerebellum with a highly sensitive two-site enzyme immunoassay. Nerve growth factor content was unevenly distributed in the cerebral cortex (dense in layers II/III and V/VI and sparse in layers I and IV). Moderate to high levels of nerve growth factor were registered in the habenular nuclei, zona incerta, ventral tegmental area, substantia nigra, locus coeruleus, ventral cochlear nucleus, trapezoid body, lateral vestibular nucleus, cerebellar nuclei and paraflocculus. Immunohistochemically, the nerve growth factor-like immunoreactivity was found in the cell bodies, dendrites and axons of adult rat central neurons, not only in the cerebral cortex, hippocampus and basal forebrain, but also in the diencephalon, brainstem and cerebellum. The population of neurons with nerve growth factor-like immunoreactivity was limited, but unexpectedly widespread, and the density of these cells correlated well with the content determined by an enzyme immunoassay in the present and a previous study [Nishio T. et al. (1992) Expl Neurol. 116, 76-841. The monoamine neurons, including dopaminergic, noradrenergic and serotonergic neurons, showed intense nerve growth factor-like immunoreactivity, indicating that the central monoaminergic neuronal system may also be involved in the nerve growth factor trophic system. To visualize nerve growth factor transported in the axons and to enhance the immunostaining in the nerve growth factor-producing cells, we injected colchicine, a potent inhibitor of microtubule polymerization and a blocker of axoplasmic transport, into the lateral ventricle of adult Wistar rat brain. Colchicine treatment enhanced the intensities of nerve growth factor-like immunoreactivity in the axons and cell bodies, especially in the axon hillocks and the proximal axons of the nerve growth factor-producing neurons. This observation may suggest the existence of an orthograde axonal transport system for nerve growth factor in the central neurons. These results suggest that nerve growth factor may exist in widespread populations of adult central neurons, including the monoaminergic neurons, and may exert a trophic effect on these neurons through not only a retrograde manner but also an orthograde manner of transport.

Nerve growth factor (NGF) is a prototype of targetderived neurotrophic factors and exerts trophic effects not only on peripheral neurons but also on central neurons, such as the magnocellular cholinergic neurons (MCNs) in the basal forebrain (reviewed in Thoenen et a1.6’ and Whittemore and Seiger6s). Recently, brain-derived neurotrophic factor (BDNF) was isolated4 and its molecular clones” revealed that the amino acid sequence of BDNF was about 50% homologous to that of NGF. More recently, with the use of polymerase chain reaction cloning techniques, additional NGF-like neurotrophic factors have been identified: neurotrophin3’3~“~38,5’ and neurotrophin-4/5.5.20,27 These five factors

*To whom correspondence should be addressed. Present address: Department of Neurology, National Center of Neurology and Psychiatry, 4-1-1, Ogawa-Higashi, Kodaira. Tokvo 187. Jaoan. Abbreviatiok: BDNF, ‘brain-derived neurotropic factor; BSA, bovine serum albumin; CA, Ammon’s horn; CHO, Chinese hamster ovary; CM, conditioned medium; dthr, dihydrofolate reductase; DMEM, Dulbecco’s minimal essential medium; EDTA, ethylenediaminetetra-acetate; EIA, enzyme immunoassay; FCS, fetal calf serum; IgG, immunoglobulin G; MCN, magnocellular cholinergic neurons; NGF, nerve growth factor; NGF-LI, NGF-like immunoreactivity; NGFR, NGF receptor; PBS, phosphate-buffered saline; PBS-Triton, PBS containing 0.3% Triton X-100; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis. 67

68

T. NISHIOPI ul.

NGF that was isolated from submaxillary glands of adult are collectively referred to as neurotrophins and the male mice (ICR mice; Shim& Laboratory Supplies Co., study of their trophic roles in the CNS is one of the Ltd, Kyoto, Japan; male, I2 weeks old. body weight 42 g. major topics in neuroscience. The demonstration of rz = 200) according to the method of Varon and coretrograde axonal transport of NGF from the ncoworkersM Antiserum against mouse NGF was raised in New Zealand White rabbits as described previously.‘h In cortex to the basal forebrains,55,56the presence of high brief, purified NGF (100 #g protein/kg body weight) with levels of NGF and NGF mRNA in the projection complete Freund’s adjuvant was injected intradermallv at area of the MCNS,‘~.~“,“~.~~.~*,~~,~’ and the identification SO-100 sites on the shaven back of rabbits, and subseqbent of high-affinity NGF receptors (NGFRs) in the cell iniections (ZOtir! NGFikg bodv weight) with incomolete bodies of the MCNs4’:” suggest a role for NGF as a Fieund’s adjuiant were performed every’six to eight \h;eeks for two years, collectively 15 times. The antibody titer was trophic factor on MCNs in the basal forebrain. monitored at two weeks after every injection, and elevated Although reports are not as prevalent as in the case gradually. Two weeks after the last injection, whole blood of MCNs, striatal intrinsic chofinergic neurons was collected. The antibody titer was determined by an should also be included in this category.39,47,49 EIA.2’ To confirm the specificity of the antibody, we Is NGF responsiveness limited only to these neurdetermined the titer of the antibody preadsorbed with an ons in the CNS? In the rat brainstem, moderate levels excess of NGF by EIA. Murine b-NGF-specific IgG was of NGF mRNA’“**8,67and high-a~nity NGFRs~‘.~~ affinity purified from the antiserum using a pure NGFbound Affi-Gel 10 column (Bio-Rad, CA). have been detected. A low level of NGF protein is Specificity c>f the a~nit~-puri~ed anti-nerve growth Jactor unambiguously present in the cerebellum,33.35.4’6’ immunoglobuiin G. The specificity of the affinity-purified anti-NGF IgG was evaluated by sodium dodecyl sulfatealthough no cholinergic neurons arc found in this polyacrylamide gel efectrophoresis (SDS-PAGE) and Westregion. Furthermore, the distribution of NGFR-like ern blotting. Briefly, mouse p-NGF (5 pg) and mouse immunoreactivity in the adult rat CNS is widesubmaxillary gland extract (IO pg) were loaded and electrospread. 9~‘2~32*44.45.53.70 These facts suggest that NGF acts phoresed through a 15%polyacrylamide separating gel. The proteins were visualized on Western blots using the affinityas a neurotroph~c molecule on other central neurons purified antibody (0.1 pg/ml). For the preadsorption test, in addition to the MCNs and striatal neurons. Rethe antibody was preadsorbed with an excess of NGF cently, using a highly sensitive enzyme immunoassay (0.05 pg/ml) for 24 h. We also evaluated the specificity of the (EIA), we have described in detail the distribution of affinity-purified anti-NGF IgG using immunohistochemNGF in the adult rat brain.4’ In addition to preistry with mouse submaxillary glands and an EIA with other neurotrophins. Briefly for the EIA method, 5irl of NGF. viously described NGF-responsive neurons, we found recombinant BDNF (produced as follows), recombii~ant the candidate neurons which would respond to NGF neurotrophin-3 (commercially obtained from Pepro Tech in the anterior and medial thalami, pontine reticular Inc., Rocky Hill, NJ) and cytochrome C (from horse heart, nuclei, superior olive, ventral cochlear nuclei and Sigma) (20pg/ml it] 0. I M Tris-HC1 buffer, pH 7.6) was coated on the center of a U-bottomed well of a 96-multiwell cerebellar Purkinje cell layer. Although histological polystyrene plate (Falcon). After blocking the non-occupied visualization of NGF protein and mRNA in the CNS space with 0.05% (v/v) bovine milk in EIA buffer [O.I M may help to understand the role of NGF in the CNS, Tris-HCl buffer (pH 7.6) containing 0.4 M NaCl, 0.1 oiu the immunohistochemical localization of NGFJ~‘t~‘S.h7 bovine serum albumin (BSA), 0.02% NaN, and I mM shows no obvious correlations to the quantitative MgCl,), 20~1 of the serially diluted affinity-puri~ed antiNGF rabbit IgG were applied on each well and incubated data determined by a reliable EIA.31.35.41.65 In situ hybridization has been successful for visualization of for 2 h at room temperature. After two washings, affinitypurified biotinylated anti-rabbit IgG (10 q/ml, Boehringer) NGF mRNA only in regions such as the hippowas applied and incubated for 1 h at room temperature. The campus or cerebral cortex, where the NGF mRNA following procedures using streptoavidin-galactosidase conNegative results do jugates and enzyme substrate were the same as those in the level is relatively high. 2~3~17,48s9.6y two-site EIA for NGF. In the case of ~combinant BDNF, not necessarily imply the absence of NGF but rather we checked BDNF molecules reliably bound on the well reflect the low Ievels of NGF protein and mRNA in bottom by other antibodies shown below. because we could the CNS.” not use BDNF in a pure form. It was determined by the antibodies prepared against peptide MDSKKRIG (NOS In the present study, we visualized the localization 92 99 of mature mouse BDNF amino acid sequence) that of NGF-like immunoreactivity (NGF-LI) in una sufIicient level of BDNF was certainly bound on the treated and colchicine-treated adult rat brains, using well bottom. (Furukawa S. and co-workers; unpublished an a~nity-purified polyclonal immunoglobulin G ~)bservations). (IgG) to /3-NGF, which is specific and with an extremely high titer, and specifically designed fixation Biologically active recombinant human neurotrophin-3 and immunohistochemical protocols. Furthermore, were purchased from Pepro Tech Inc. (Rocky Hill, NJ). we tested the reliability of the immunohistochemical Mouse BDNF was prepared as described below. The exdata by comparing them to the NGF contents deterpression vector of BDNF was constructed with insertions mined by an EIA. into pGEM-7Zf (Promega) of dihydrofolate reductase (dfhrt gene (BRLI and BDNF gene, which was cloned by EXPERIMENTAL Antibody

production

PROCEDURES

and characterization

Production of afinity-purified anti-nerve grow)th .fhctor immunoglobulin G with a very high titer. Mouse B-NGF was

purified by CM-Sephadex C-50 c~ron~atography from 7s

boly&e;ase dhain reaction from mouse brain cDNA library. Chinese hamster ovary (CHO) cells (dfhr -; American Type Culture Collection63) were transfected with the expression vector and dfhr+ CHO cells were selected by culturing in the media supplemented with dialysed fetal calf serum (FCS). Further selection of gene-amplified CHO cells was

69

Cellular localization of NGF in adult rat brain performed by cloning of surviving cells in the presence of increasing concentrations of methotrexate (Sigma). Finally, one of the CHO cell clones selected with 5.0 pM methotrexate was used for BDNF production. The cells were grown in a Dulbecco’s minimal essential medium (DMEM) supplemented with 10% non-dialysed FCS until confluency was achieved. Then, the medium was replaced with serum-free DMEM, and the cells cultured for another three days. The conditioned medium (CM) was pooled and stored at -80°C until use. The CM (21-31).was diluted with a equal volume of 0.1 M acetate buffer. aH 5.0 (buffer A) and aDDlied to a column of CM52 cellulose whatman) ‘equilidrated with buffer A. The column was washed and eluted with buffer A containing 0.6 M NaCl, and the eluent contained most of the BDNF activity. Further, the active fractions were concentrated with pressure dialysis (Amicon PM10 filter) and applied to a column of Sephadex G-150 equilibrated with buffer A. The active fractions were pooled, concentrated to nearly 5 ml and used for evaluation-of antibody specificity. The nuritv of BDNF nrotein was nearlv 20% (50 ue BDNF/mlj, when judged from densitometiic scan of gel plates after SDS-PAGE. The content of NGF dete~ined by the two-site EIA was below 10 ng/ml, negligible for practical use. Biological activities of neurotrophins were assessed in terms of survival activity of dissociated sensory neurons from eight-day-old chick embryo dorsal root Concentrations of BDNF employed were deterganglia. 24,37 mined by a densitometric scan of protein bands after SDS-PAGE with NGF used as the standard. Determination of nerve growthfactor contents in detail in the cerebral cortex, diencephalon, brainstem and cerebellum by a highly sensitive enzyme immuno~say

We have previously reported the quantitative map of NGF contents in the adult rat brain determined by a highly sensitive EIA.4’ In this study, we determined a more detailed dist~bution of NGF contents using the EIA, in relation to the cortical layers in the cerebral cortex and in relation to the special populations of the neuronal nuclei in the diencephalon, brainstem and cerebellum. Preparatjon of 120 p-sectioned cortical slices. Cerebral cortices from seven-week-old Wistar rats (n = 10; Shimizu Laboratory Supplies Co., Ltd, Kyoto, Japan) were unfolded into a single plane and a square (4.0 x 4.0 mm) containing the frontoparietal cortices was cut out with a razor blade (from bregma to 4 mm posteriorly, and from the median line laterally at 2.0 mm to 6.0 mm). The cortical tissue sample was sectioned into 16 slices, each with a thickness of 120 pm (4.0 x 4.0 x 0.12 mm), in parallel with the cortical surface cut by a cryostat at -10°C. Preparation of brain samples sizedO.

x 0.5 x 0.5 mm. The

brains of seven-week-old Wistar rats (n = 2) were coronally sectioned into 0.9-mm-thick slices with a cryostat at - 10°C. Each coronal slice was cut perpendicularly into squares (0.5 x 0.5 mm) with a stainless steel razor blade. Brain samples of 0.9 x 0.5 x 0.5 mm (0.225 mm’) were prepared. The sectioning of the coronal slices, adjusted by cutting the brain initially at the plane including bregma and sectioning into 0.5 x 0.5 mm squares, was adjusted by cutting them in the same manner as the initial dissection, which was photographed. The anatomical orientation of each piece was critically checked by reference to an anatomical map and the dissections of brain areas were reliably reproducible. Measuring nerve growth factor contents in the brain samples by a two-site immunoassay. NGF content in each

brain sample was determined by a highly sensitive two-site EIA described previously.4’ Briefly, NGF was extracted by a freeze-thaw procedure from the brain samples in an extraction buffer [O.1M Tris-HCl buffer (pH 7.6), containing 2% (w/v) BSA, 2% (w/v) gelatin, 1.0 M NaCl, 2 mM EDTA and 80 trypsin-inhibitory units of aprotinin/l]. After coating the anti-mouse NGF rabbit IgG and blocking the non-occupied space with 0.05% (v/v) bovine milk in EIA

buffer, the sample solution prepared in the extraction buffer was applied and incubated for 4 h at room temperature. After two washings, a~nity-purify biotinylated antimouse NGF IgG (10 ng/ml) was applied and incubated for 12 h at 4°C. After another two washings, /I-D-galactosidaseconjugated streptavidin (Biogenex Lab.) was applied and incubated for 1 h at room temperature. The enzyme reaction was initiated at room temperature by the addition of 4-methylumbelliferyl-lY-D-galactoside and terminated by the addition of 0.1 ml of 0.1 M glycine_NaOH buffer (pH 10.3). The amount of 4-methylum~lIiferone formed was measured with a spectrofluorometer (model 65060, Hitachi). The wavelength for excitation and emission were 360 and 450 nm, respectively. Purified mouse NGF was used as a standard for the EIA. Fluorescence values of the standard and samples were corrected by subtraction of background values obtained with normal &G-coated wells or non-coated wells. NGF contents in the samples were evaluated by reference to the standard curve. The values obtained from 10 animals were averaged. Immunohisto~hemistry

Male adult Wistar rats (seven weeks of age, n = 5; Shimizu Laboratory Supplies Co. Ltd, Kyoto, Japan) were anesthetized with 50 mg/kg sodium pentobarbital and transcardially perfused with 50ml cold (4°C) phosphatebuffered saline (PBS), followed by 250ml of cold 2% paraformaldehyde and 0.2% parabenzoquinone in 0.1 M phosphate buffer (pH 7.6). The brains were removed, sliced into 3 mm segments, then postfixed in the latter solution for 2 h and immediately transferred to 30% sucrose in 0.I M phosphate buffer for two days prior to sectioning on a freezing microtome. Whole brain sections were cut into 20pm widths and treated with PBS containing 0.3% Triton X-100 (PBS-T&on) for 48 h. Thev were then incubated in a free-fioating state with the affinity-purified anti-mouse NGF IgG (0.5 gg/ml) in PBS-Triton for 48 h at 4°C. After blocking endogenous peroxidase with 0.3% H,O,, bound antibodies were detected by incubating sections in 1 fig/ml biotinylated goat anti-rabbit IgG antibody (Vector Lab&atories, Burlingame, CA) in PBS-Triton for 12 h at 4°C. followed by incubation for 1 h at room temperature in thd avidin-biotin-peroxidase reagent (1: 1000 dilution ABC kit, Vector Laboratories, Burlingame, CA). After rinsing for IOmin in PBS-Triton, peroxidase labeling was visualized by incubation in 0.05 M TrisCl buffer (pH 7.6) containing 0.02% diaminobenzidine tetrahydrochloride, 0.6% nickel ammonium sulfate and 0.0002% H,O, (chemicals were purchased from Dojin Chemicals and Wako Pure Chemical Industry, Japan). The staining specificity of the polyclonal anti-NGF IgG was assessed by (i) preadsorption of the antibody with an excess (0.2 @g/ml) of NGF for 24 h, (ii) omission of the primary antibody and {iii) incubation with preimmune rabbit serum. Colchicine-treated adult rat brain. Fifty micrograms of colchicine (Nakarai Chemicals, Ltd, Kyoto, Japan), dissolved in 0.01 M phosphate buffer (pH 7.6), were stereotaxitally injected into the Iateral ventricle of the brain of a seven-week-old adult Wistar rat (n = 5; Japan SLC, Inc., Shizuoka, Japan). Twenty-four hours later. the brain was processed for immun~histo~hemi~l investigation as described above. RESULTS

Characterization factor

of the polyclonai

anti-nerve

growth

antibody

The titer of the rabbit antiserum against mouse NGF was extremely high, 10m9 (Fig. la, open squares) as determined by an EIA. Binding was blocked by preadso~tion of the antiserum with NGF (filled squares). Immunoblotting (Fig. lb) of purified

70

T. NISHIOet al.

mouse NGF (lane 1) or crude submaxillary gland extract (lane 2) revealed only a single band with an M, of 13,060, as expected for the NGF monomer. This immunoblotting was completely eliminated by

incubation with an excess of purified NGF (lane No visible bands were detected in homogenates of brain (data not shown), because the level of NGF the brain was too low (under 2 rig/g tissue) to

a

3). rat in be

b 1000

Mol. wt

1 2 3 xw3 .”

-97.4

-66.2

). .= t.0

2

E

-14.4 -13.0

10

n

-9

*--“Q-l -8

’ “““‘I

’ ““.“I -7

’ --*-6

-5

(log ,ox)

Dilution ratio of Antiserum

Fig. 1. Characterization of the antibody. (a) Quantification of the anti-mouse NGF IgG. The antibody titer was determined by an EIA as IOe9 (open squares) and the specific signals were blocked by preadsorption of the antibody with an excess of NGF (filled squares). (b) Western blot analysis of the purified mouse NGF (lane 1) and the mouse submaxillary gland extract (lanes 2, 3) immunos~in~ with affinity-purist anti-NGF IgG (iancs I, 2) and with IgG that had been preadsorbed with an excess of NGF (lane 3). (c, d) Photomicrographs of the adult mouse submaxiilary gland immunostained with the anti-NGF IgG (c; magnification x 571) and with the preadsorbed IgG (d; x 571). Note that the IgG specifically recognizes NGF monomer, and the immunoreactivity is specifically localized in the secretory tubular cells.

Cellular localization of NGF in adult rat brain

71

a circle represents the piece containing NGF levels less than 5OOpg/g tissue. In the thalamus, the regions containing the habenular nuclei, mediodorsal nucleus and medial forebrain bundle showed moderate levels and the regions containing the zona incerta showed high levels of NGF (Fig. 4a). In the mesencephalon, high levels of NGF were observed in the regions containing the ventral tegmental area and substantia nigra, and moderate levels were noted in the regions including the central gray and superior colliculus (Fig. 4b). In the pons, high levels were noted in the regions containing the locus coeruleus, ventral cochlear nucleus, trapezoid body and lateral vestibular nucleus, and moderate levels were observed in the regions including the superior olive, caudal pontine reticular nucleus, nucleus raphe magnus and medial vestibular nucleus (Fig. 4c, d). In the cerebellum, high levels were recognized in the paragocculus and moderate levels in the cerebellar nuclei (Fig. 4d). without

;

10

100

1000

Conwnttntion of Anti-NGF IgG

10000

WW

Fig. 2. Specificity of the anti-NGF IgG evaluated by a sensitive EIA. Note that the polyclonal anti-NGF IgG reacts specifically with NGF even at concentrations lower than 10 ng/ml and has no cross-reaction with recombinant BDNF, recombinant neurotrophin-3 (NT-3) and cytochrome C, even at concentrations higher than 1&ml.

detected by Western blotting.67 In the immunohistochemistry of the adult male mouse submaxillary glands with the antibody, the immunoreactivity observed in the secretory tubular cells (Fig. lc) was also eliminated by preadsorption with NGF (Fig. Id). The affinity-purified anti-NGF IgG showed no crossreaction with recombinant BDNF, recombinant neurotrophin3 and cytochrome C, even at concentrations higher than 1 pg/ml (Fig. 2). Therefore, the anti-NGF IgG used specifically recognized NGF.@ Detailed distribution of nerve growth factor contents in the cerebral cortex, diencephalon, brainstem and cerebe~~urndeterm~d by enzyme imm~o~say Cerebral cortex. A schematic representation of the preparation of samples is shown in Fig. 3a. The distribution of NGF contents in 16 cortical slices, each with a thickness of 120 pm, is shown in Fig. 3b. The NGF content in each slice obtained from each animal deviated by lO-30% from the mean value. Each bar represents the mean value (n = 10) of NGF content for each cortical slice. The bars are shown in order of increasing distance from the cortical surface. NGF was unevenly distributed in the cortical layers of the adult rat neocortex; it was distributed densely in cortical layers II/III and V/VI, and sparsely in layers I and IV. Diencephalon, brainstem and cerebellum. The distribution map of NGF contents in the pieces (0.9 x 0.5 x 0.5 mm) from the diencephalon, brainstem and cerebellum is shown in Fig. 4. Each circle at the center of a square represents the mean value (n = 2) of NGF content per brain piece. The square

Nerve growth factor-like immunoreactivity in the adult rat brain

NGF-LI was observed exclusively in the neurons, which were limited but unexpectedly widespread (Table I), although reactive astrocytes observed after colchicine treatment showed positive NGF-LI. Some neurons contained NGF-LI in their cell bodies, dendrites and/or axons, while others included NGF-LI only in their somata. The immunolabeling was completely eliminated by preadsorption of the antibody with an excess of NGF (0.2 pg/ml) for 24 h (Figs 3d, 6b, d). Low-power photomicrographs illustrating the overall distribution of NGF-LI in the adult rat brain are shown in Fig. 5a-f. Cerebral cortex. NGF-LI was localized in the cell bodies, proximal dendrites and axons of some neurons (Fig. 6e). Moderate immunolabeling was observed in the pyramidal and non-pyramidal neurons in layers II/III and V/VI (Figs 3c, 6e). The density of NGF-LI-positive cells was high in layers II and VI and low in layer V. The localization of NGF-LI in the adult rat cerebral cortex correlated well with the quantitative data determined by EIA (Fig. 3b, c). The cingulate gyrus and pyriform cortex showed higher densities of NGF-LI-positive neurons (Fig. 5b, c), and this was also consistent with the quantitative findings?’ Hippocampal formation. Moderate NGF-LI was recognized mainly in the pyramidal cell layer in the CAI-CA4 area, and mild immunoreactivity was observed in the dentate granule cell layer (Fig. 6a). Intensely immunostained neurons were sparsely scattered among moderately labeled neurons in the pyramidal cell layer (Fig. 6f). The dentate hilus contained a higher density of intensely labeled pyramidal neurons than the other areas in the hippocampus (Fig. 6a). OIfactory bulb. Moderate immunoreactivity was

•

.

:-[."

•

>

n=l

~l,z:

~

ot .

¢2

~

O ' - Z

>

0

0

c,.) ¢,)

c~ ~

o

.~.~_

~

.

t

.

. ....-

...

.

.'.1

::', " , ' : - , . "

? ...

- .

:

$

*,

.. ..:

.

: .

-

~ o o = =-~-

.

::

-.

©

"o

0

~.~

~

~

~

"'-=

~

i., 0 c~

¢,~

o ~"~

=.= = o , . ~ o

o

¢.;.i . ~

0

'¢-

o~.=

:::k ~_, . e l ~D X:)

Z ' ~ ^

C_~ or. o

~-~

_2c_~

o 0

0

~.o~ b= =~~o~

m

.==oo~.

O ~

0

~

0

¢.) ' , o

~

O0

~"

l..O

(.0

I~

O0

0")

0

"r-

~

eO

~

l.O

t-..O

a3!l~ I1~3.11JO~) pguo!13a~-mfl 0~:[ ~Z~>~ 0'~ 8> o

,'~

=

= .(D

._~ ~

. ~ . ~ "~ ~ ,.~ ¢)

,~

o

¢n

" - z : o >,.o .~b 0

c~ XD

"-~ .= ~ o =, t2 o= = .=-"~ 04.= ¢n

~1

~ o o - ~ -o ..9. o

73

Celfufar localization of NGF in aduh rat brain

nigra (A9) and ventral tegmental area (AlO) showed intense NGF-LX (Figs Sd, 7a, b,) while the neurons in the mesencephalic trigeminal nucleus were mildly immunore~cti~e. The fibers in the medial forebrain bundle were moderately immunolabeled. Pons. Moderate NGF-LI was noted in the neurons in the trapezoid body (Figs Se, f, 8e, f) and ventral cochlear nucleus (Figs Se, f, Bc, d), while mild NGFLI was noted in the pontine nucleus, lateral lemniscus, pontine reticular nucleus, superior o%vc: (Figs 5e, f, 8e) and lateral vestibuiar nucleus (Fig. Sf, 8a). Intense NGF-LI was recognized in the serotonergic neurons in the nucleus raphe pontis (B5) (Fig. 7c, d) and the catecholaminergic neurons in A4-A8, including the locus coeruleus {A) (Figs Se, 7e, f j.

noted in the glomerular cell layer but the granule cell layer was not stained. Basal forebrain and striarum. The neurons in the anterior olfactory nucleus, medial septum, diagonal band of Broca, basal nucleus, nucleus innominata, lateral amygdalaid nucleus and striatum contained NGF-LI in their cell bodies as discrete punctate labeling (Fig. 6g). Diencephnlon. Mild immunoreactivity was recog nized in the neuroxls of the medial habenular nucleus (Fig. 5b, c), medial thalamic nuclei, hypothalamic supraoptic and suprachiasmatic paraventricular, nuclei. Intensely immunoreactive neurons were noted in the zona incerta (A13) (Fig. Sb) and paraventricular region (All, 12, 14). Mesencepkabn. The neurons in the substantia

d

&I 40

l l&l so2500 p&tissue

Fig. 4. Distribution map of NGF contents in the diencephalon, brainstem and cerebellum determined by a highly scnsitivc EIA. Each square represents a brain pkxx sized 0.9 x 0.5 x 0.5 mm, from which NGF was extracted. Each circle at the center of a square represents the mean value (n = 2) of NGF content per brain pkce. The square without circle represents the piece containing NGF levels less than 500 pg/g tissue. Note that a mild to high amount of NGF is observed in the habcnuiar nuclei, me&dorsal nucleus, medial forebrain bundle, zona incerta la), ventral tegmental area, substantia n&a, central gray. superior colliculus (b), locus coeruleus, ventral cochlear nucleus, trapezoid body, lateral vestibular nucleus, superior olive, caudal poniine reticular nucleus, nucleus raphe magnus, medial vestibular nucleus {d), paraflocculus and the cerebellar nuclei @). atv, ventral tegmental area; cov. ventral cochleaearnucleus; g, gelatinose thalamic nucleus; hm, medial habenular nucleus; i, interpositus nucleus; ip, in~~~uncular nucleus; 1, lateral cerebellar nucleus; m, medial webellar nucleus; npV, principal scnsrrrytrigeminal nucleus; nV, trigeminsl motor nucleus; nVII, facial motor nucleus; os, superior olive; WI, parafocculus; r, red nucIeus; rgi, gigantacellular reticular nucleus; rm, nucleus raphe magnus; rpc, parvicellular reticular nucleus; rpoc, caudal pontine rsticuiar nucieus: SC,superior collicuius; snc, substantia nigra, compact zone; snr, substantia nigra, reticular zone; tcl, centrolatcral thalamic nucleus; tlp, lateral posterior thalamic nucleus; tpm, ventroposterior medial Wilamic nucleus; tpl, ventroposterior lateral nucleus; tsV, nucleus of spinal trigemifial tract; tvm, ventromedial thahmic nucleus; vl, lateral vcstibular nucleus; vm, medial vestivular au&us: d, U3na k&a.

74

T. NISHIOet al.

Table 1. Nerve growth factor-like immunoreactivity in the adult rat brain. Intensity of immunoreactiv~ty Cerebral cortex layer I layer II layer III layer IV layer V layer VI Hip~ampus pyramidal cell layer granule cell layer other layers Olfactory bulb glomerular layer Basal forebrain anterior olfactory nucleus diagonal band of Broca mediat septal nucleus

basal nucleus Amygdaloid nucleus (lateral) Caudate-putamen Globus pallidus Thalamus habenular nucieus other thalamic nuclei Hypothalamus paraventricular nucleus supraoptic nucleus Mesencephalon mesencephalic trigeminal raphe nucleus oculomotor nucleus superior coIliculus Pons pontine nucleus pontine raphe nucleus lateral lemniscus pontine reticular nucleus trapezoid body superior olive ventral cochlear nucleus lateral vestibular nucleus Medulla inferior olive spinal trigeminal nucleus hypoglossal nucieus external cuneate nucleus medullary reticular nucleus Cerebellum Purkinje cell layer granule cell layer cerebellar nuclei Catecholamine neurons Al-Al4

++ ++ + ++ +t + f-f

(Fig. 6c, h), and it was denser in the paraflocculus than in the other lobes (Fig. 5f). The cerebellar granule cell layer and the white matter showed no immunola~ling. The neurons in the cerebeilar nuclei contained mild levels of NGF-LI (Fig. 8a, b). Nerve growth factor-like

immunoreactivity in colchi-

tine-treated rat brain

+ + + *

++ t +

+ + 4 t + -I2 + + + ++ii t + “t-t+ + + ++ + -I-+ + + + f + + ++ + +4+

Key: negative or barely detectable, - ; slightly stained or clearly immunopositive neurons sparsely scattered, & ; mildly stained, + ; moderately stained, + + ; intensely stained, + + +

~eduIZa. Mild NGF-LI was noted in the neurons of the inferior olive, spinal trigeminal nucleus, hypoglossal nucleus, external cuneate nucleus and medullary reticular nucleus. Intense NGF-LI was seen in the catecholaminergic neurons in Al-A3. Cerebellum. Moderate NGF-LI was noted in the cell bodies and dendrites of the Purkinje cells

After colchicine treatment, NGF-LI was observed in the axons in the gray and white matter, including the corpus callosumt fimbria, anterior commissure and pencil fibers in the striatum (Figs 9a, b, IOa, f) and was enhanced in the cell bodies and the proximal axons of the pyramidal neurons in the cerebral cortex and hippocampal CAl-CA4 (Fig. 1Oa-e). The intensity of immunolabeling decreased in the dentate granule cell layer after colchicine treatment (Figs 6a, 9b, l@d). NGF-LI was increased in the pyramidal tracts in the brainstem and the molecular layer of the cerebellum (Fig. log). Colchicine treatment attenuated the intensity of NGF-LI in the basal forebrain neurons and a slight decrease was noted in the substantia nigra. Reactive astrocytosis was observed in the cerebral cortex 48 h after colchicine treatment and reactive astrocytes showed moderate immunoreactivity. DISCUSSION

Signz$cance of the present results with respect to the previous investigations on nerve growth factor-like immunoreactivity in the CNS

Others have reported the localization of NGF-LI within fiber pathways throughout the brain in developing’ and adult rats. ” However, the immunolabeling appeared to be a non-specific interaction between IgG molecules and a component of selected fibers, because similar results were obtained using preimmune IgG or secondary antibody alone.” Recently, Conner and co-workers” have reported NGF-LI in the adult rat basal forebrain and hippocampal formation. Like us, they used affinity-purified polyclonal antibody and specifically designed fixation and immunohistochemical protocols for localizing NGF-LI in the adult rat CNS. Though the “discrete punctate labeling” in the basal forebrain neurons they observed was similar to that seen in the present study, the immunostaining pattern in the hippocampal formation, cerebral cortex and other CNS areas was different from ours. The difference may result from the difference of the titer of the antibody used. We also designed the immunization protocol as described in the Experimental Procedures section in order to obtain an extremely high titer antibody. In their study, the neurons in the cerebral cortex or olfactory bulb showed no immunolabeling, while other investigators have found moderate levels of NGF in the adult rat cerebral cortex and olfactory bulb.30,32.3*~“2 Moreover, the distribution of NGF

Cellular localization of NGF in adult rat brain

75

sparsely scattered in all areas of the hippocampus. mRNA in the hippocampal formation determined by Both of these patterns of NGF mRNA expression in situ hybridization (NGF mRNA being expressed in both the pyramidal cell and granular cell laymay exist in the hippocampus, as we found both er&3.17,48,59@) was quite different from the dist~bution intensely labeled neurons which were scattered and of NGF-LI in their study (NGF-LI being localized in moderately labeled neurons which were numerous. the mossy fibers of the granular cells but not in the Senut et 01.” described a distribution of NGF precurpyramidal cells”). Recently, Emfors and co- sor-like immunoreactivity in the adult rat hippoworkers” found a more restricted hybridization patcampus that was very similar to that described by us tern than that described in previous studies,2~3*‘7A8*5g*69 for NGF. Thus, all of these reports support the in that the cells expressing NGF mRNA were results of the present study.

Fig. 5. Low-power photomicrographs illustrating the overall distribution of NGF-like immunoreactivity in the adult rat brain. Magnification x 7.

Fig. 6. NGF-like immunoreacrivity in the hippocampus, cerebellum, cortical and hippocampal pyramidal neurons, magnocelhdar neurons in the basal forebrain and cerebellar Purkinje cells in the adult rat brain. Photomicrographs of a frontal section of the hippocampus (a, b; magnification x 23), cerebellum (c, d; x 33) and the pyramidal neurons in layer V of the cerebral cortex (e; x 685), the pyramidal neurons in the CA2 hippocampal subfield (f; x 457), the magnocellular neurons in the nucleus basalis (g; x 685), and the in the cerebellar Purkinje cells (h; x 457). Note the NGF-LI in the pyramidal cell layer in area CAI-CA4. granular cell layer, in the dentate gyms (a) and in the cerebellar Purkinje celi layer (c). The pyramidal cell layer in the CAI-CA4 hippocampal subfield shows stronger NGF-LI than the granule cell layer in the dentate gyrus (a). The cerebellar granule cell layer and the white matter show no immunolabeling (c). The positive NGF-LI is eliminated by preadsorption of the antibody with an excess of NGF (b. d). NGF-LI is localized to the cell bodies, proximal dendrites and axons in the layer V pyramidal neurons of the cerebral cortex (3). NGF-LI is noted in the cell bodies and proximal dendrites in the hippocampal pyramidal neurons. and intensely immunostained neurons are sparsely scattered in the pyramidal ceil layer among moderately immunola~ied neurons (f). Discrete pun&ate labeling is observed in the cell bodies of the ma~noceiluIar neurons in the basal forebrain(g). Note the moderate NGF-LI in thecell bodies and dendrites ofthe Purkinje cells and the absence of labeling in the cerebellar granule cell layer (h).

Cellular

localization

of NGF

SigniJicance of the present results with respect to other neurotrophins Recently, other members of the NGF family (BDNF, neurotrophins-3, -4, and -5) have been identified.4~5~‘3~20~24~27~37~38~5’ Since these molecules have approximately 50% amino acid sequence homology to NGF, anti-NGF antibodies may cross-react with these molecules. However, several lines of evidence argue against the possibility that the NGF-LI that we observed was due to cross-reaction with other neurotrophins. First, five hydrophilic regions presumed to be potential antigenic determinants for NGFM are not

in adult rat brain

II

necessarily homologous in their amino acid sequence with these neurotrophins.5~20~24~37 Second, the polyclonal anti-mouse /I-NGF antibody does not cross-react with BDNF or neurotrophin-3.s2 Third, mRNA for BDNF and neurotrophin-3, but not NGF mRNA,’ have been detected in the cerebellar granule The present study revealed no specific cell layer.‘4*23,42 labeling in the cerebellar granule cell layer. Fourth, NGF mRNA is expressed in the glomerular region, while BDNF mRNA is predominantly localized to the granule cell layer in the adult rat olfactory system.” In the present study, NGF-LI was not observed in the granule cell layer but in the glomerular layer of the

Fig. 7. NGF-like immunoreactivity in the monoaminergic neurons. The dopaminergic neurons in the substantia nigra (A9) and ventral tegmental area (AIO) (a, b), the serotonergic neurons in the nucleus raphe pontis (BS) (c, d) and the noradrenergic neurons in the locus coeruleus (A6) (e, f). Magnification of a, c, e, x 26; b, x 570; d, x 250; f, x 285. Note that intense NGF-LI is observed in the cell bodies, dendrites and axons in these monoaminergic neurons.

78

T. NESHIOet al

Fig. 8. NGF-like immunoreactivity in the cerebellar nuclei (a, b), ventral cochlear nucleus (c, d) and trapezoid body (e, f). Magnification of a, c, e, x 26; b, x 520; d, x 125; f, x 520. Note that mild NGF-LI is observed in the neurons and their neuropils in the cerebellar nuclei (a, b) and moderate NGF-LI is recognized in the neurons of the ventral cochlear nucleus (c. d) and trapezoid body (e, f) and their neuropils.

olfactory system. Fifth, the expression of neurotrophin-3 mRNA is restricted to the dentate granule cell layer and the pyramidal cell layer in the medial CA1 and CA2,r4” while we found NGF-LI in the dentate granule cell layer and the pyramidal layer in CAl-CA4. Sixth, BDNF-like immunoreactivity is localized to the nuclei of the hippocampal neurons,” while NGF-LI, in the present study, was observed in the neuronal cell bodies and dendrites. Finally, as a direct answer to the question, the anti-NGF antibody in use did not cross-react with BDNF and neurotrophin-3 evaluated with a sensitive EIA (Fig. 2). Moreover, we investigated the NGF-LI not only by immunohistochemistry,

but also by a two-site EIA, which is more reliable as it requires at least two epitopes for detection, Therefore, our anti-NGF antibodies likely did not cross-react with other neurotrophins. The present study is the first to report on the localization and distribution of NGF-LI in the adult rat CNS, as far as a thorough description of the whole brain is concerned. Comparison between the localization of nerve growth factor-like immunoreactivity and quantitative distributton determined by enzyme immunoassay The localization of NGF-LI in the present study was generally consistent with the detailed distribution

Cellular

localization

of NGF

of NGF in the adult rat brain determined by EIA.4’ In the cerebral cortex, NGF-LI was localized in neurons mainly in layers II/III and V/VI, while NGF levels were high in cortical layers II/III and V/VI and were low in layers I and IV. The cingulate gyrus and pyriform cortex contained higher densities of NGFLI-positive cells, and NGF levels in these regions were moderately high (500-1000 pg/g tissue). In the hippocampus, NGF-LI was localized mainly in the pyramidal cell layer and dentate granule cell layer,

in adult

rat brain

19

and the former showed stronger staining than the latter, while the piece of 1.5mm-square containing the subfields CA3- CA4 and dentate gyrus showed higher NGF levels than that comprising area CAlCA2. This is probably because the density of the NGF-LI-positive cells per 1.5-mm-square was higher in the areas containing the subfields CA3-CA4 and dentate gyrus than in the areas comprising the subfields CAl-CAZ (Fig. 3a). In the diencephalon, the result of high NGF levels in the havenular nuclei

Fig. 9. Low-power photomicrographs of the rat brain illustrating NGF-LI enhancement after colchicine treatment. Fifty micrograms of colchicine were injected into the lateral ventricle (left side in the photomicrographs) of the rat brain. Magnification of a and b; x 10. Note that a great enhancement is observed in fibers of the corpus callosum, pencil fibers of the striatum, and in the neurons of the cerebral cortex and hippocampus.

Fig. 10. Higher-power photomicrographs of the rat brain illustrating NGF-LI enhancement after colchicine treatment. The cerebral cortex and hippocampal formation [a-e: magnification of a, x 34; b, cortical layer II, x 300; c, cortical layer V, x 300; d, CA2 hippocampal subfield, x 150, e, dentate hilus (CA4 subfield), x 3001, f, fimbria, x 75 and g, cerebellum, x IS0 are shown Note that NGF-LI is enhanced in the axons, the proximal axons (and/or the axon hillocks) and the cell bodies of the pyramidal neurons in the cerebral cortex and hippocampus after colchicine treatment. The immunoreactivity in the dentate granule cell layer decreases. Varicose dot-like structures with NGF-LI are observed mainly in the molecular layer of the cerebellum.

Cellular localization of NGF in adult rat brain and zona incerta revealed in the present study (Fig. 4a) was consistent with that of relatively high levels in the medial thalamus noted in the previous investigation4’ and was also consistent with immunohist~hemical data reveaiing the neurons in the habenular nuclei and zona incerta (Al 3) with moderate to intense NGF-LI. In the brainstem and cerebellum, the result of moderate to high NGF levels in the substantia nigra, ventral tegmental area, raphe pontis, locus coeruleus, ventral cochlear nucleus, trapezoid body, superior olive, pentine reticular nucleui, lateral vestibular nucleus, cerebellar nuclei and paraflocculus (Fig. 4b-d) was also consistent with immunohistochemical findings showing neurons in these areas with moderate to intense NGF-LI, although we could not include the substantia nigra, ventral tegmentai area, raphe pontis and locus coeruleus as regions with high NGF levels in the previous study41 because the brain pieces examined were larger (0.9 x 1.5 x 1Smm) than those in this study. Generally, the density of NGF-LI-positive cells correlated well with the quantity of NGF. ~~g~I~c~nee of the nerve growth factor-like immunoreactive cells : nerve growth factor -responsive newtons or nerve growth factor synthesizing cells

81

terminal, retrogradely transported in the axon and hence to the cell body of the NGF-responsive neurons would be confirmed. From the previously reported distribution of and the cellular localization of NGFR ~75 g~12~32~44~45~53~70 NGF-LI shown in this study, the MCNs in the basal forebrain, striatal and pallidal neurons, hypothalamic paraventricular and supraoptic neurons, A 13 neurons, mesencephalic trigeminal neurons, superior collicular neurons, ventral cochlear neurons, raphe neurons, reticular neurons, medial vestibular neurons, neurons in the trapezoid body and cerebellar Purkinje cells would be NGF-responsive neurons in the CNS. On the other hand, NGF mRNA has been detected in the glomerular layer in the olfactory bulb, in the pyramidal cell layer and granule cell layer of the hip~campus and in layers II, III and VI of the cerebral cortex.2,3J7.48.59,69 The present study revealed positive NGF-LI in all of these neurons, and hence they would actually synthesize NGF protein. Monoamine neurons and nerve growth factor

The initial jnvestjgatjons on searching for NGFresponsive neurons in the CNS were focused on catecholaminergic neurons, because the NGF-responsive neurons in the peripheral nervous system had been NGF-LI was localized exclusively to neurons in the established to be the sympathetic neurons and neural adult CNS on non-pathological conditions, although crest-derived sensory neurons.6’*68 Bjiirklund et al. reactive astrocytes observed after colchicine treatment observed effects of centrally administers NGF in showed positive NGF-LI. The NGF-LI-positive neurstimulating central catecholamine and indolamine ons could be NGF-synthesizing cells or, alternatively, neuronal ingrowth into irides transplanted into adult NGF-responsive cells. NGF-responsive cells should express the NGFRs: a low-affinity receptor, ~7.5,*~.~~ host brain6.’ However, subsequent studies could not confirm these conclusions and the NGF-responsive and a high-affinity receptor, trk p140.22,2’3’*34The neurons in the CNS have been considered to be cholincellular distribution of NGFR ~7.5in the CNS, studied using the monoclonal antibody l92-IgG, appears to be ergic.6’.68 studies limited, but widespread. 9.12.32,44.4SS3~70 Recent The present study indicated that catecholaminergic have disclosed that NGFR p75 is also a low-affinity and indolaminergic neurons showed positive NGF-LI, receptor for the other neurotrophins: BDNF, and which was also quantitatively confirmed by a two-site neurotrophins-3 and neurotrophins-4.20,50@ Thus, the EIA. The immunolabeling pattern observed in the neurons expressing NGFR ~75 may be the neurons monoamine neurons was intense and distinct from responsive to these neurotrophins. Further underthat seen in other neurons. The slight decrease of standing of the role of NGF in the CNS requires NGF-LI in the neurons after colchicine treatment information on the distribution of trk ~140. However, suggests that the monoaminergic neurons may relittle has been published in this area except for the spond to NGF. The nigrostriatal and mesolimbic report of the basal forebrain neurons labeled with trk dopaminergic systems may be involved in the NGF ~140 mRNA. 25 The present study provides further trophic system. However, this remains to be chalinformation regarding the NGF-responsive neurons in lenged because there have been some reports arguing the CNS. The discrete punctate labeling, which has that the nigral dopaminergjc neurons respond not to been reported by Conner et al.,” was observed in the NGF but to BDNF in vitro,26or that these neurons do NGF-responsive neurons in the basal forebrain and not contain NGFR.32@,45 Another possibility is that striatum. They also demonstrated that NGF-LI was the monoaminergic neurons may produce NGF. co-localized exclusively with NGFR p75-positive Cerebeilar Purkinje cells and nerve growth factor neurons. The punctate distribution of the immunoreaction product in these neurons may denote a vesicular During development, the cerebeilar Purkinje localization of the NGF-like antigen, although further cells express not only low-affinity NGFRs but electron microscopic evaluation will be needed to also biologically active, high-affinity NGFRs.g,i2,43*s3.70 determine the precise intracellular distribution, If the Recently, it has been reported that NGF and depolarpunctate immunoreaction product is indeed localized izing stimuli or excitatory neurotransmitters enhance to endosomes, then the hypothesis that the the survival of cultured Purkinje cells.‘O We NGF-NGFR complex is internalized at the nerve have reported that the cerebellar nuclei and lateral

T. NISHIOet al

82

vestibular nucleus, which receive nerve terminals from the Purkinje cells, contain relatively high levels of NGF, and hence the cerebellar Purkinje cells may be NGF responsive.4’ We confirmed this in the present study. After colchicine treatment, varicose dot-like structures with NGF-LI were observed mainly in the molecular layer of the cerebellum, which may indicate a transport of NGF in the dendrites of the Purkinje cells or in the parallel fibers of the granular cells. Axonal transport of nerve growth factor: orthograde or retrograde:? NGF, a prototype of target-derived factors, is transported retrogradely in the axons from the peripheral target organs to the cell bodies of the NGFresponsive neurons.6’ Since central neurons form a neuron-to-neuron network, the target neurons of NGF-responsive neurons synthesize and secrete NGF, which is subsequently taken up by NGF-responsive neurons. Colchicine treatment not only blocks the axonal transport but also up-regulates NGF mRNA in the central neurons8 Therefore, upon colchicine application we expected that NGFLI in the NGF-synthesizing neurons would be enhanced uniformly in the cell bodies. However. colchicine treatment markedly enhanced NGF-LI in the axon hillocks and the proximal axons (Fig. lOc+) of the pyramidal neurons in the cerebral cortex and the hippocampus. These neurons also express NGF mRNA.2,‘.‘7.48.59.69This observation may suggest the existence of an orthograde axonal transport system for NGF in central neurons. Recently, the possibility of an orthograde axonal transport system for other neurotrophins (BDNF and neurotrophin-3) has been

reported.54 While the retrograde transport of NGF has been well established in the peripheral nervous system,6’ its physiological roles and the mechanisms of action, including autocrine/paracrine regulation of neuronal survival, remain to be elucidated in the CNS. The present study clearly demonstrated that NGFLI was enhanced in axons after colchicine treatment, indicating that NGF may indeed be transported in the axons in the CNS. As expected, NGF-LI was clearly enhanced in the fimbria (Fig. 10f) and corpus callosum (Figs 9a, b, 10a). However, the enhancement of NGF-LI was unexpectedly observed in the pencil fibers in the striatum (Fig. 9a). This suggests that NGF may be transported through striatopetal or striatofugal pathways.

CONCLUSIONS

In summary, our results suggest the presence of NGF in certain populations of neurons in the adult rat brain. Neurons with NGF-LI were limited, but more widely distributed than expected. Colchicine treatment markedly enhanced the NGF-LI in the axon hillocks and the proximal axons of the pyramidal neurons in the cerebral cortex and the hippocampus, which also express NGF mRNA, suggesting the presence of an orthograde axonal transport system for NGF in the CNS. NGF may have a trophic effect on the neurons in the adult human brain and help to preserve their functions or repair their disorders. NGF deficits and/or NGF administration may be relevant to human CNS pathologies and regeneration.

REFERENCES

Ayer-Lelievre C. S., Ebendal

T., Olson L. and Seiger A. (1983) Localization of NGF-like immunoreactivity in rat neurons tissue. Med. Biol. 61, 296-304. Ayer-Lelievre C. S., Olson L., Ebendal T., Seiger A and Persson H. (1988) Expression of the p-nerve growth factor gene in hippocampal neurons. Science 240, 1339-1341. Bandtlow C. E., Meyer M., Lindholm D., Spranger M., Heumann R. and Thoenen H. (1990) Regional and cellular distribution of interleukin l/I and nerve growth factor mRNA in the adult rat brain: possible relationship to the regulation of nerve growth factor synthesis. J. Cell Biol. 111, 1701~1711. Barde Y.-A., Edgar D. and Thoenen H. (1982) Purification of a new neureotrophic factor from mammalian brain. Eur. molec. Biol. Org. J. 1, 549-553. Berkemeier L. R., Winslow J. W., Kaplan D. R., Mikolics K., Goeddel D. V. and Rosenthal A. (1991) Neurotrophin-5: a novel neurotrophic factor that activates trk and trkB. Neuron 7, 857-866. Bjerre B., Bjijrklund A. and Stenevi U. (1973) Stimulation of growth of new axonal sprouts from lesioned monoamine neurons in adult rat brain by nerve growth factor. Brain Res. 60, 161-176. Bjarklund A. and Stenevi U. (1972) Nerve growth factor: stimulation of regenerative growth of central noradrenergic neurons. Science 17, 1251-1253. Ceccatelli S., Ernfors P., Villar M. J., Persson H. and Hakfelt T. (1991) Expanded distribution of mRNA for nerve growth factor, brain-derived neurotrophic factor, and neurotrophin 3 in the rat brain after colchicine treatment. Proc. natn. Acad. Sci. U.S.A. 88, 10352-10356. Cohen-Cory S., Dreyfus C. F. and Black I. B. (1989) Expression of high- and low-affinity nerve growth factor receptors by Purkinje cells in the developing rat cerebellum. Expl. Neural. 105, 104~-109. Cohen-Cory S., Dreyfus C. F. and Black 1. B. (1989) NGF and excitatory neurotransmitters regulate survival and morphogenesis of cultured cerebellar Purkinje cells. J. Neurosci. 11, 462471. Conner J. M., Mui S., Varon S., Hagg H. and Manthorpe M. (1992) The localization of nerve growth factor-like immunoreactivity in the adult rat basal forebrain and hippocampal formation. J. camp. Neural. 319, 454462. Eckenstein F. (1988) Transient expression of NGF-receptor-like immunoreactivity in postnatal rat brain and spinal cord. Brain Res. 446, 149%154.

Cellular localization of NGF in adult rat brain

83

13. Ernfors P., Inabez C. F., Ebendal ‘I., Olson L. and Persson H. (1990) Molecular cloning and neurotrophic activities ofa protein with structural similarities to nerve growth factor: developmental and topographical expression in the brain. Proc. natn. Acad. Sci. U.S.A. 87, 5454-5458. 14. Ernfors P., Wetmore C., Olson L. and Persson H. (1990) Identification

of cells in rat brain and peripheral tissues expressing mRNA for members of the nerve growth factor family. Neuron 5, 511-526. 15. Finn P. J., Ferguson I. A., Wilson P. A., Vahaviolos J. and Rush R. A. (1987) Immunohistochemical evidence for the distribution of nerve growth factor in the embryonic mouse. J. Neurocytol 16, 639647. 16. Furukawa S., Kamo I., Furukawa Y., Akazawa S., Satoyoshi E., Itoh K. and Hayashi K. (1983) A highly sensitive enzyme immunoassay for mouse b nerve growth factor. J. Neurochem. 40, 734744. 17. Gall C. M. and Isackson P. J. (1989) Limbic seizures increase neuronal production of messenger RNA for nerve growth factor. Science 240, 1339-1341. 18. Goedert M., Fine A., Hunt S. P. and Ullrich A. (1986) Nerve growth factor mRNA in peripheral and central tissues and in the human central nervous system: lesion effects in the rat brain and levels in Alzheimer’s disease. Molec. Brain Res. 1, 85-92. 19. Guthrie K. M. and Gall C. M. (1991) Differential expression of mRNAs for the NGF family of neurotrophic factors in the adult central olfactory system. J. camp. Neurol. 313, 955102. 20. Hallbook F., Ibanez C. F. and Persson H. (1991) Evolutionary studies of the nerve growth factor family reveal a novel member abundantly expressed in Xenopus ovary. Neuron 6, 845-858. 21. Harlow E. and Lane D. (1988) Antibodies, A Laboratory Manual. Cold Spring Harbor Laboratory, U.S.A. 22. Hempstead B. L., Martin-Zanca D., Kaplan D. R., Parada L. F. and Chao M. V. (1991) High-affinity NGF binding requires coexpression of the trk proto-oncogene and the low-affinity NGF receptor. Nature 350, 678-683. 23. Hofer M., Pagliusi S. R., Hohn A., Leibrock J. and Barde Y.-A. (1990) Regional distribution of brain-derived neurotrophic factor mRNA in the adult mouse brain. Eur. molec. Biol. Org. J. 9, 2459-2464. 24. Hohn A.. Leibrock J.. Bailev K. and Barde Y.-A. (1990) Identification and characterization of a novel member of the nerve growth factor/brain derived neurotrophic factor family. Nature 334, 339-341. 25. Holtzman D. M., Li Y., Parada L. F., Kinsman S., Chen C. K., Valletta J. S., Zhou J., Long J. B. and Mobley W. C. (1992) ~140”~ mRNA marks NGF-responsive forebrain neurons: evidence that trk gene expression is induced by NGF. Neuron 9, 465478. 26. Hyman C., Hofer M., Barde Y.-A., Juhasz M., Yancopoulos G. D. and Squint0 S. P. (1991) BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature 350, 230-232. 27. Ip N. Y., Ibanez C. F., Nye S. H., McClain J., Jones P. F., Gies D. R., Bellucio L., LeBeau M. M., Espinosa R. III, Squint0 S. P., Persson H. and Yancopoulos G. D. (1992) Mammalian neurotrophin-4: structure, chromosomal localization, tissue distribution, and receptor specificity. Proc. natn. Acad. Sci. U.S.A. 89, 306&3064. 28. Johnson D., Lanahan A., Buck C. R., Sehgal A., Morgan C., Mercer E., Bothwell M. and Chao M. (1986) Expression and structure of the human NGF receptor. Cell 47, 5455554. 29. Kaplan D. R., Hempstead B. L., Martin-Zanca D., Chao M. V. and Parada L. F. (1991) The trk protooncogene product: a signal transducing receptor for nerve growth factor. Science 252, 554558. 30. Klein R., Conway D., Parada L. F. and Barbacid M. (1990) The trkB tyrosine protein kinase gene codes for a second neurogenic receptor that lacks the catalytic kinase domain. Cell 61, 647656. _ 31. Klein R.. Jine S.. Nanduri V.. O’Rourke E. and Barbacid M. (1991) The trk oroto-oncoaene encodes a receptor for ~ nerve growth-factor. Cell 65, ‘1899197. 32. Koh S., Oyler G. and Higgins G. (1989) Localization of nerve growth factor receptor messenger RNA and protein in the adult rat brain. Expl. Neurol. 106, 209-221. 33. Korsching S., Auburger G., Heumann R., Scott J. and Thoenen H. (1985) Levels of nerve growth factor and its mRNA in the central nervous system of the rat correlate with cholinergic innervation. Eur. molec. Biol. Org. J. 4, 1389-1393. 34. Lambelle F., Klein R. and Barbacid M. (1991) trkC, a new member of the trk family of tyrosine protein kinases, is a receptor for neurotrophin-3. Cell 66, 967-979. 35. Large T. H., Bodary S. C., Clegg D. O., Weskamp G., Otten U. and Reichardt L. F. (1986) Nerve growth factor gene expression in the developing rat brain. Science 234, 352-355. 36. Larkfors L., Ebendal T., Whittemore S. R., Persson H., Hoffer B. and Olson L. (1987) Decreased level of nerve growth factor (NGF) and its messenger RNA in the aged rat brain. Brain Res. 427, 55-60. 37. Leibrock J., Lottspeich F., Hohn A., Hofer M., Hangerer B., Masiakowski P., Thoenen H. and Barde Y.-A. (1989) Molecular cloning and expression of brain-derived neurotrophic factor. Nature 341, 149-152. 38. Maisonpierre P. C., Belluscio L., Friedman B., Anderson R. F., Wiegand S. J., Furth M. E., Lindsay R. M. and Yancopoulos G. D. (1990) NT-3, BDNF, and NGF in the developing rat nervous system: parallel as well as reciprocal pattern of expression. Neuron 5, 501-509. 39. Martinez H. J., Dreyfus C. F., Jonakait G. M. and Black I. B. (1985). Nerve growth factor promotes cholinergic development in brain striatal cultures. Proc. natn. Acad. Sci. U.S.A. 82, 7777-7781. 40. Meier R., Becker-Andre M., Gotz R., Heumann R., Shaw A. and Thoenen H. (1986) Molecular cloning of bovine and chick nerve growth factor (NGF): delineation of conserved and unconserved domains and their relationship to the biological activity and antigenicity of NGF. Eur. molec. Biol. Org. J. 5, 1489-1493. 41. Nishio T., Akiguchi I. and Furukawa S. (1992) Detailed distribution of nerve growth factor determined by a highly sensitive enzyme immunoassay. Expl. Neurol. 116, 76-84. 42. Phillips H. S., Hains J. M., Laramee G. R., Rosenthal A. and Winslow J. W. (1990) Widespread expression of BDNF but not NT3 by target areas of basal forebrain cholinergic neurons. Science 250, 290-294. 43. Pioro E. P. and Cue110 A. C. (1988) Purkinje cells of adult rat cerebellum express nerve growth factor receptor immunoreactivity: light microscopic observations. Brain Res. 455, 82-186. 44. Pioro E. P. and Cueho A. C. (1990) Distribution of nerve growth factor receptor-like immunoreactivity in the adult rat central nervous system. Effect of colchicine and correlation with the cholinergic system-I. Forebrain. Neuroscience I

34. 57-87. 45. Pioro E. P. and Cuello A. C. (1990) Distribution

of nerve growth factor receptor-like immunoreactivity in the adult rat central nervous system. Effect of colchicine and correlation with the cholinergic system--II. Brainstem, cerebellum, and spinal cord. Neuroscience 34, 89-l 10.

84

T. NISHIOet al.

46.

Radeke M. J., Misko T. P., Hsu C., Herzenberg L. A. and Shooter E. M. (1987) Gene transfer and molecular cloning of the rat nerve growth factor receptor. Nat& 325, 593-597. 47. Raivich G. and Kreutzbere G. W. (1987) The localization and distribution of high athnitv B-nerve growth factor bindine. sites in the central nervous-system of theadult rat, a light microscopic autoradiographic %udy using [i2sI]fi-nerve growth factor. N~roscie~~e 20, 23-36. 48. Rennert P. D. and Heinrich G. (1986) Nerve growth factor mRNA in brain: localization by in situ hybridizatio~l. Biochem. biophys. Res. Corn. 138, 813-S 19. 49. Richardson P. M., Verge-Issa V. M. K. and Riopelle R. J. (1986) Distribution of neuronal receptors for nerve growth factor in the rat. J. Neurosci. 6, 2312.-2321.

50. Roadriguez-Tebar A., Dechant G. and Barde Y.-A. (1990) Binding of brain-derived neurotrophic factor to the nerve growth factor receptor. Neuron 4, 4877492. 51. Rosenthal A., Goeddel D. V., Nguyen T., Lewis M., Shih A., Laramie G. R., Nikolics K. and Winslow J. W. (1990) Primary structure and biological activity of a novel human neurotrophic factor. Neuron 4, 767.-773. 52. Ruit K. G., Elliott J. L., Osborne P. A., Yan Q. and Snider W. D. (1992) Selective dependence of mammalian dorsai root ganglion neurons on nerve growth factor during embryonic development. Neuron 8, 573-587. 53. Schatteman G. C., Gibbs L., Lanahan A. A., Claude P. and Bothwell M. (1988) Expression of NGF receptor in the developing and adult primate central nervous system. J. Neurosci. 8, 860-873. 54. Schecterson L. C. and Bothwell M. (1992) Novel roles for neurotronhins are suggested by BDNF and NT-3 mRNA -_ expression in developing neurons. N&on 9, 449463. 55. Schwab M. E., Otten U., Agid Y. and Thoenen H. (1979) Nerve growth factor (NGF) in the rat CNS: absence of specific retrograde axonal transport and tyrosine hydroxylase induction in locus coeruleus and substantia nigra. Brain Res. 168, 473483. 56. Seiler M. and Schwab M. E. (1984) Specific retrograde transport of nerve growth factor (NGF) from neocortex to nucleus basalis in the rat. Bruin Res. 300, 33-39. 57. Senut M. C., Lamour Y., Lee J., Brachet P. and Dicou E. (1990) Neuronal localization of the nerve growth factor precursor-like immunoreactivity in the rat brain. IPII. J. devl Neurosci. 8, 65-80. 58. Shelton D. L. and Reichardt L. F. (1986) Studies on the expression of the nerve growth factor (NGF) gene in the central

nervous system: level and regional distribution of NGF mRNA suggest that NGF functions as a trophic factor for several distinct populations of neurons. Proc. natn. Acud. U.S.A. 83, 2714-2718. 59. Spillantini M. G., Aloe E., Alleva E., De Simone R., Goedert M. and Levi-Montalcini R. (1989) Nerve growth factor mRNA and protein increase in hypothalamus in a mouse model of aggression. Proc. natn. Acad. Sci. U.S.A. 84, 8555-8559. 60. Squint0 S. P., Stitt T. N., Aldrich T. H., Davis S., Bianco S. M., Radziejewski C., Glass D. J., Masiakowski P.. Furth

61. 62. 63. 64. 65. 66.

M. E., Valenzuela D. M., Di Stefano P. S. and Yancopoulos G. D. (1991) trkB encodes a functional receptor for BDNF and NT-3 but not NGF. Celt 65, 885-893. Theonen H., Bandtiow C. and Heunlann R. (1987) The physiological function of nerve growth factor in the central nervous system: comparison with the periphery. Rev. Physiol. Biochem. Pharmac. 109, 146178. Ullrich A., Gray A., Berman C. and Dull ‘I’. J. (1983) Human beta-nerve growth factor gene sequence highly homologous to that of mouse. Nature 303, 821-825. Urlaub G. and Chasin L. A. (1980) Isolation of Chinese hamster cell mutants deficient in dihydrofolate reductase activity. Proc. natn. Acad. U.S.A. 77, 4216-1220. Varon S., Nomura J. and Shooter E. M. (1967) The isolation of the mouse nerve growth factor protein in a high molecular weight form. Bjochemjs~r~ 6, 2202-2209. Weskamp G. and Otten U. (1987) An enzyme-linked immunoassay for nerve growth factor (NGF): a tool for studying regulatory mechanisms involved in NGF production in brain and peripheral tissues. J. Neurochem. 48, 1779-1786. Wetmore C., Cao Y., Petterson R. F. and Olson L. (1991) Brain-derived neurotrophic factor: subcellular compartmentalization and interneuronal transfer as visualized with anti-peptide antibodies. Proc. natn. Acad. Sci. U.S.A. 88,

9843-9847. 67. Whittemore S. R., Ebendal T., Larkfors L., Olson L., Seiger A., Stromberg I. and Persson H. (1986) Developmental

and regional expression of /? nerve growth factor messenger RNA and protein in the rat central nervous system. Proc. natn. &ad. 5% U.S.A. 83, 817-821. 68. Whittemore S. R. and Seiger A. (1987) The expression, localization and functional significance of p-nerve growth factor in the central nervous system. Brain Res. Rev. 12, 439464. 69. Whittemore S. R., Friedman P. L., Larhammar D., Persson H., Gonzalez-Carvajal M. and Holets V. R. (1988) Rat b-nerve growth factor sequence and site of synthesis in the adult hippocampus. J. Neurosci. Res. 20, 403410. 70. Yan Q, and Johnson E. M. (1988) An immunohistochemical study of the nerve growth factor receptor in developing rats. J. Neurosci. 8, 3481-3498. (Accepted 8 December 1993)